Arthritis

Knee Osteoarthritis

Knee Osteoarthritis

Arthritis

Knee Osteoarthritis

Copyright © 2011 Hindawi Publishing Corporation. All rights reserved.

This is a focus issue published in volume 2011 of “Arthritis.” All articles are open access articles distributed under the Creative Com-mons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original workis properly cited.

Editorial Board

Alejandro Balsa, SpainHenning Bliddal, DenmarkMarie-Christophe Boissier, FranceRuben Burgos-Vargas, MexicoDeh-Ming Chang, TaiwanMarco Amedeo Cimmino, ItalyMichel De Bandt, FranceKurt de Vlam, BelgiumChanghai Ding, AustraliaJorg Distler, GermanyDirk Elewaut, BelgiumJoao Eurico Fonseca, PortugalAnnamaria Iagnocco, ItalyTsuyoshi Kasama, JapanMarkku Kauppi, FinlandShinichi Kawai, JapanHerbert Kellner, Germany

George D. Kitas, UKShigeru Kotake, JapanBurkhard Leeb, AustriaFrederic Liote, FranceJeffrey R. Lisse, USAK. P. Machold, AustriaCharles J. Malemud, USABernhard J. Manger, GermanyBianca Marasini, ItalyMarco Matucci-Cerinic, ItalyNeil John McHugh, UKPeter McNair, New ZealandPaola Migliorini, ItalyPierre Miossec, FranceYuki Nanke, JapanJavier Narvaez, SpainKusuki Nishioka, Japan

Aleth Perdriger, FranceTimothy Radstake, The NetherlandsSusan Reisine, USAPascal Richette, FranceBruce M. Rothschild, USAAnne Rutjes, SwitzerlandMalcolm Smith, AustraliaM. Takei, JapanFrancesco Trotta, ItalyA. van der Helm-van Mil, The NetherlandsPeter M. van der Kraan, The NetherlandsJiri Vencovsky, Czech RepublicCornelis L. Verweij, The NetherlandsLucy R. Wedderburn, UKPierre Youinou, France

Contents

Quantitative Cartilage Imaging in Knee Osteoarthritis, Felix Eckstein and Wolfgang WirthVolume 2011, Article ID 475684, 19 pages

Patella Eversion Reduces Early Knee Range of Motion and Muscle Torque Recovery after Total KneeArthroplasty: Comparison between Minimally Invasive Total Knee Arthroplasty and Conventional TotalKnee Arthroplasty, Tokifumi Majima, Osamu Nishiike, Naohiro Sawaguchi, Kouichi Susuda,and Akio MinamiVolume 2011, Article ID 854651, 6 pages

Developmental Mechanisms in Articular Cartilage Degradation in Osteoarthritis, Elena V. TchetinaVolume 2011, Article ID 683970, 16 pages

The Self-Administered Patient Satisfaction Scale for Primary Hip and Knee Arthroplasty,Mahomed, Rajiv Gandhi, Lawrence Daltroy, and J. N. KatzVolume 2011, Article ID 591253, 6 pages

Serum Leptin Concentration Positively Correlates with Body Weight and Total Fat Mass inPostmenopausal Japanese Women with Osteoarthritis of the Knee, Jun Iwamoto, Tsuyoshi Takeda,Yoshihiro Sato, and Hideo MatsumotoVolume 2011, Article ID 580632, 6 pages

Current Surgical Treatment of Knee Osteoarthritis, Karolin Ronn, Nikolaus Reischl, Emanuel Gautier,and Matthias JacobiVolume 2011, Article ID 454873, 9 pages

Hindawi Publishing CorporationArthritisVolume 2011, Article ID 475684, 19 pagesdoi:10.1155/2011/475684

Review Article

Quantitative Cartilage Imaging in Knee Osteoarthritis

Felix Eckstein1, 2 and Wolfgang Wirth1, 2

1 Institute of Anatomy and Musculoskeletal Research, Paracelsus Medical University, Strubergaβe 21, 5020 Salzburg, Austria2 Chondrometrics GmbH, 83404 Ainring, Germany

Correspondence should be addressed to Felix Eckstein, felix.eckstein@pmu.ac.at

Received 19 August 2010; Accepted 25 October 2010

Academic Editor: Dirk Elewaut

Copyright © 2011 F. Eckstein and W. Wirth. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Quantitative measures of cartilage morphology (i.e., thickness) represent potentially powerful surrogate endpoints in osteoarthritis(OA). These can be used to identify risk factors of structural disease progression and can facilitate the clinical efficacy testingof structure modifying drugs in OA. This paper focuses on quantitative imaging of articular cartilage morphology in the knee,and will specifically deal with different cartilage morphology outcome variables and regions of interest, the relative performanceand relationship between cartilage morphology measures, reference values for MRI-based knee cartilage morphometry, imagingprotocols for measurement of cartilage morphology (including those used in the Osteoarthritis Initiative), sensitivity to changeobserved in knee OA, spatial patterns of cartilage loss as derived by subregional analysis, comparison of MRI changes withradiographic changes, risk factors of MRI-based cartilage loss in knee OA, the correlation of MRI-based cartilage loss with clinicaloutcomes, treatment response in knee OA, and future directions of the field.

1. Introduction

Magnetic resonance imaging (MRI) has revolutionized thefield of clinical research in osteoarthritis (OA) becauseit can directly visualize all diarthrodial tissues, includingcartilage, bone, menisci, ligaments, synovium, and others.As it has been recognized that OA is a disease of theentire joint, involving most (if not all) of the above tissues,MRI has substantial advantages over radiography, which canonly delineate the bone. Owing to its three-dimensionalcoverage of anatomical structures [1, 2] (Figure 1), MRIadditionally permits to obtain quantitative measures ofrelevant tissue structures (and their changes over time)in OA. Quantitative measures of cartilage morphology(i.e., thickness, volume, surface areas) represent potentiallypowerful surrogate endpoints in osteoarthritis (OA). Thesecan be used to identify risk factors of structural diseaseprogression and can facilitate the clinical efficacy testing ofdisease (or structure) modifying drugs in OA (DMOADs),which are not clinically available to date.

This paper will focus on the knee, as most of thequantitative cartilage imaging work has been performed inthat joint. It will further focus on the cartilage, as this

the tissue that has generated most interest in context ofquantitative measurement in OA using MRI. Last, we willfocus on quantitative cartilage morphology (i.e., thickness,surface areas, volume) but will not cover quantitativeMRI techniques measuring cartilage composition, such asdGEMRIC, T2, T1rho, and others [3].

Quantitative measurements of cartilage morphology(structure) fully exploit the 3D nature of MRI data sets [1, 2];their strength is that they are less observerdependent andmore objective than scoring methods, and that relativelysmall changes in cartilage thickness, which occur relativelyhomogeneously over larger areas may be detected over time,which are not apparent to the naked eye. This is important,as the progression of structural changes in OA has generallybeen shown to be slow, both when being evaluated byradiography [4–6] and MRI [6–10]. A recent study foundthat quantitative measures of cartilage morphology [11] weremore powerful in revealing relationships between local riskfactors (meniscus damage and malalignment) and knee car-tilage loss than a semi-quantitative approach using ordinalscales (i.e., whole organ MRI score) [12]. The disadvantageof quantitative measurement, however, is that it requiresspecialized software and is more time intensive because tissue

2 Arthritis

boundaries need to be tracked (i.e., segmented) throughoutlarge series of slices using trained technical personnel.Also, quantitative measurements are less sensitive to theoccurrences of small focal changes within larger structures(i.e., cartilage lesions), which may be readily picked upby an expert reader, particularly if the location within thelarger structure is variable from joint to joint. A recentstudy showed, for instance, that MRI-based semiquantitativescoring of cartilage status was able to differentiate betweenknees with and without early (i.e., Kellgren-Lawrence grade[KLG] 2) radiographic OA, whereas quantitative measuresof cartilage morphology displayed no or little differencebetween healthy and KLG2 knees [13]. It thus depends onthe context and on the specific research question, whetheror not quantitative cartilage assessment is better suited asan outcome measures for a particular study than semi-quantitative measures. Ideally, both approaches should beused in complimentary rather than competing fashion instudies assessing either the status or the progression of OA.

Focusing on quantitative imaging of cartilage morphol-ogy in the knee, this paper will sequentially address

(i) different cartilage morphology outcome variablesand regions of interest in the knee,

(ii) the relative performance and relationship betweencartilage morphology measures,

(iii) imaging protocols for measurement of cartilage mor-phology, including validation,

(iv) rates of change and sensitivity to change observed inknee OA.

(v) spatial patterns of cartilage loss in knee OA as derivedby subregional analysis,

(vi) comparison of MRI changes with radiographic chan-ges in knee OA,

(vii) risk factors of cartilage loss in the knee as identifiedby quantitative cartilage MR imaging,

(viii) the correlation of MRI-based cartilage loss with clin-ical outcomes, and treatment response in knee OA,

(xi) future directions of the field.

2. Cartilage Morphology Outcome Variablesand Regions of Interest in the Knee

A consensus-based nomenclature for the above-mentionedstructural (i.e., morphological metric labels) or composi-tional features as well as definitions for regions of interestin the knee (i.e., anatomical labels, see Table 1 and Figure 2)has been proposed by a group of experts [14]. Theabove nomenclature will be used throughout this paper,and important abbreviations for morphology metrics andanatomical regions of interests including recent extensions(i.e., statistical labels and subregional labels, i.e., [15]) aresummarized in Table 1. Cartilage morphology outcomes

(a)

(b)

Figure 1: 3D reconstruction of the knee cartilages after segmen-tation: (a) View from anteromedial with softtissues in grey (b)View from anterior-lateral, with the bone segmented and withthe cartilage thickness distribution in the patella displayed in falsecolors (red: thick cartilage; blue: thin cartilage). The cartilage of themedial tibia (MT) is depicted dark blue, that of the lateral tibia (LT)green), that of the medial weight-bearing femoral condyles (cMF)yellow, that of the lateral weight-bearing femoral condyles (cLF)red, that of the patella (P) magenta, and that of the femoral trochlea(TrF) turquoise. Segmentation was performed based on a 3D-DESSknee imaging data set from the Osteoarthritis Initiative (OAI),a public-private partnership funded by the National Institutesof Health and conducted by the OAI Study Investigators. Foranatomical (region of interest) labels, also see Figure 2 and Table 1.

commonly include the size of the total area of subchondralbone (tAB), the area of the cartilage surface (AC), thedenuded (dAB) and cartilage covered (cAB) area of subchon-dral bone, the cartilage thickness over the tAB (ThCtAB)or over the cAB (ThCcAB), the cartilage volume (VC),the cartilage volume normalized to the tAB (VCtAB), thecartilage signal intensity [16–18], and others (Table 1).

To obtain the above quantitative morphological mea-sures of cartilage, the relevant cartilage plates of a joint needto be segmented by a trained user with the choice of severalinput devices [19], with or without assistance from (semi-automated) segmentation software [20–28]. Because therelative performance of different segmentation algorithmshas been discussed in previous reviews [9], this point will

Arthritis 3

Table 1: Morphological (metrics), statistical, and anatomical (region of interest) labels commonly used in cartilage morphology publicationson the knee.

Abbreviation Explanation Unit

Morphological (metrics) label

VC volume of the cartilage (mm3/mL)

tAB total area of subchondral bone (cm2)

AC area of cartilage surface (cm2)

cAB area of tAB covered by AC (cm2)

dAB% percent of tAB denuded (not covered by AC) (%)

VCtAB volume of the cartilage divided by tAB (mm)

ThCtAB thickness of the cartilage over the entire tAB (mm)

ThCcAB thickness of cartilage over cAB (mm)

dAB% percent of tAB denuded (not covered by AC) (%)

VCtAB volume of the cartilage divided by tAB (mm)

ThCtAB thickness of the cartilage over the entire tAB (mm)

ThCcAB thickness of cartilage over cAB (mm)

Statistical labels

Me mean (i.e., thickness)

Max maximum (i.e., thickness)

Mavmaximal averaged, for example, mean of the top1% values

Min minimum (i.e., thickness)

Mivminimum averaged, for example, mean of thelowest 1% values

SD standard deviation (i.e., thickness)

CV% coefficient of variation (i.e., thickness)

c(Me, Mav)thickness measured from cartilage surface (AC) tobone interface (tAB)

b(Me, Mav)thickness measured from bone interface (tAB) tocartilage surface (AC)

a(Me, Mav) average of the two above (b, c)

Anatomical (region of interest) labels

Total cartilage plates

P Patella

MT Medial tibia

LT Lateral tibia

F Femur

TrF Femoral trochlea

MF Medial femoral condyle

cMF weight-bearing portion of MF

pMF posterior portion of MF

LF Lateral femoral condyle

cLF weight-bearing portion of LF

pLF posterior portion of LF

MFTC aggregate values for MT and cMF (MT + cMF)

LFTC aggregate values for LT and cLF (LT + cLF)

Subregions (to be combined with above total plate labels, i.e., cMT or ccMF)

c central

e external

i internal

a anterior

p posterior

For anatomical (region of interest) labels, also see Figures 1 and 2.

4 Arthritis

not be covered in depth in this chapter. Using the above tools,an operator needs to accurately trace both the bone-cartilageinterface (i.e., the subchondral bone surface; tAB), and thesurface of the cartilage, respectively (AC). The tracing of thetAB should include dABs, but not (peripheral) osteophytesurfaces. As there are various sources of artifacts on MRI, andbecause signal intensity and contrast may vary substantiallybetween baseline and followup acquisitions, there exists acurrent consensus that expert quality control is importantfor accurate analyses; the time required for segmentation ofthe cartilages or for the correction of computer-generatedsegmentation may take several hours per knee joint. Afterall slices of interest have been segmented, image analysissoftware can be used to compute the three-dimensionalmorphological features listed in Table 1.

Anatomical regions of interest in the knee are listedand explained in Table 1 and Figure 2. Since the weight-bearing (cMF, cLF) and posterior aspects (pMF and pLF)of the femoral condyles are continuous and lack a definiteanatomical border, different definitions for these ROIs havebeen proposed: Glaser et al. [32] used the projection of theposterior intercondylar bone bridge as a cutoff between theweight-bearing and posterior zone, whereas later studies [33]introduced a 60% distance criterion between the trochlearnotch and the most posterior aspects of both femoralcondyles as a cutoff between both regions (Figure 2). In aface-to-face comparison, the work in [34] reported the tABof the 60% ROI to be approximately 20% greater and lessvariable (between subjects) than that based on the bonebridge, with cartilage morphology metrics being generallymore reproducible in the 60% ROI. However, thicknessmeasures did not differ significantly between both ROIs, andthe longitudinal rate of change and standardized responsemean (SRM = mean change/SD of change, as a measure ofsensitivity to change) over two years were similar for bothROIs. Using sagittal images, both 60% [33] (Figure 2) and75% [35, 36] cutoffs have been used, the 60% cutoff assigning36 (±1.8%) of the tAB of MF to the weight bearing (cMF)and 64% to the posterior portion (pMF), whereas the 75%cutoff assigns 47% (±2.0%) of the tAB to cMF and 53% topMF, respectively [36]. Again, the mean change and SRMwere similar for both the 60% and the 75% ROI [36]. The(weight-bearing) medial femorotibial compartment (MT +cMF) is commonly addressed as MFTC, and the lateralcompartment (LT + cLF) as LFTC (Figure 2).

Quantitative measures of surface curvature and jointincongruity have also been determined from MR images[37] and were observed to discriminate between subjectswith various radiographic OA grades cross-sectionally at0.2 T [38, 39]. Curvature estimates at different scales (at0.2 T) were reported to be associated with the magnitude ofcartilage loss longitudinally [40] and cartilage homogeneity(quantified by measuring entropy from the distribution ofsignal intensities in tibial cartilage from 0.2 T gradient echoimages) was reported to discriminate between subjects with-out and with early radiographic OA [18]. This measure wasproposed to be particularly sensitive in peripheral regions,where the cartilage is covered by the meniscus [41]. Theseresults are surprising because other MRI techniques that

P

TrF cLFpLF

LT

LF

LFTC

(a)

TrFcMF

pMF

MT

MF

MFTC

(b)

Figure 2: Sagittal 3D DESS MR images showing anatomicalregions of interest commonly analyzed: (a) lateral femorotibialcompartment, (b) medial femorotibial compartment; P: patella,TrF: femoral trochlear, MT: medial tibia, MF: medial femoralcondyle, cMF: weight-bearing part of the medial femoral condyle,pMF: posterior part of the medial femoral condyle, MFTC: cMF+ MT; LT: lateral tibia, LF: lateral femoral condyle, cLF: weight-bearing part of the lateral femoral condyle, pMF: posterior part ofthe lateral femoral condyle, LFTC: cLF + LT; the magental line showsthe projection of the trochlear notch, the blue line the posteriorend of the medial and lateral femoral condyle, and the turquoiseline the 60% criterion (of the distance between the trochlear notchand the posterior ends of the condyles) used to separate cMFfrom pMF, and cLF from pLF, respectively. Images are from theOsteoarthritis Initiative (OAI), a public-private partnership fundedby the National Institutes of Health and conducted by the OAI StudyInvestigators. For anatomical (region of interest) labels, also seeFigure 1 and Table 1.

have been validated for targeting specific macromoleculesof the cartilage, such as collagen, proteoglycans, or water(T2 mapping, T1rho, dGEMRIC and others) [3, 42] haveoften been unsuccessful in discriminating between healthyknees and knees with early OA, and they have generally notbeen able to discriminate between different radiographic OAstages, in particular between early (preradiographic) OA andradiographic OA [6, 8, 43].

3. Relative Performance and Relationshipbetween Cartilage Morphology Measures

Most investigations dealing quantitatively with cartilagemorphology in OA have focused on the cartilage volume

Arthritis 5

ccLF−0.1%(−0.02)

icLF−0.3%(−0.07)

ecLF+0.3%(+0.11) ccMF

−3.5%(−0.39)

ecMF−2.7%(−0.33)

icMF−1.6%(−0.34)

(a)

MFTC−1.8%(−0.47)

LFTC−0.5%(−0.18)

cLF0.0%

(−0.00)

LT−0.8%(−0.35)

cMF−2.6%(−0.42)

MT−1.2%(−0.44)

(b)

pMT−0.7%(−0.12)

aMT−0.9%(−0.25)

cMT−1.7%(−0.44)

eMT−2.2%(−0.38)

iMT−0.5%(−0.19)

pLT−1.1%(−0.23)

aLT−0.2%(−0.07)

cLT−1.2%(−0.31)

eLT−0.5%(−0.14)

iLT−1.1%(−0.32)

(c)

Figure 3: Display of the rates of change (%/annum) and standardized response mean (SRM) in femorotibial cartilage compartments, platesand subregions. (a) View of the weight-bearing part of the medial (cMF) and lateral femoral condyle (cLF) from inferior. (b) View of theweight-bearing part of the cMF and cLF and of the medial (MT) and lateral tibia (LT) from posterior. (c) View of the MT and LT fromsuperior. For an explanation of the subregion abbreviations, please see Table 1. The data represent mean values from 3 studies: (i) the KLG3participants of the A 9001140 study (n = 28) [29], (ii) the high risk (BMI > 30; KLG ≥ 2) subcohort from a first release of OAI participants(n = 54) [30], (iii) knees with neutral alignment from the MAK study (n = 74) [31].

(VC), but this outcome measure has a number of pitfalls.The ability to discriminate between OA and healthy subjectsis limited, because cartilage volume is largely determined bybone size, which increases the intersubject variability andthus limits the ability to discriminate between subjects withand without cartilage loss [44]. This has led to misinterpre-tations in the literature, where it has been suggested thata high VC may be protective of OA, because men showhigher VCs than women, and women are more susceptibleto knee OA than men. However, men have mainly largerjoint surfaces than women (and hence also larger VC) [45]even after adjustment for body height and weight [46]; VCshould therefore not be directly compared between sexes.In longitudinal studies, the subchondral bone area has beenshown to increase with aging, both in healthy reference

subjects and in OA patients [47–49]. Such effects may maska reduction in cartilage thickness in OA when measuringVC, because of the simultaneous expansion of the bone andcartilage layer. Therefore, alternative outcomes have beenused, such as the VC normalized to the subchondral bonearea (VCtAB), or the cartilage thickness over the entiresubchondral bone area (ThCtAB) [44, 50].

In a recent study, Hudelmaier et al. [34] examined therelationship of the above parameters and their test-retest pre-cision (at 3 T) in a set of 33 subjects, both without and withsigns of radiographic osteoarthritis (reproducibility study).Further, they compared these parameters at baseline and at2-year followup in 28 subjects with advanced radiographicosteoarthritis (sensitivity study). They found that the AC waslarger than the tAB in all cartilage plates. In MT and LT,

6 Arthritis

the cartilage volume divided by the total bone area (VCtAB)was similar to the mean cartilage thickness over the totalbone area (ThCtAB.aMe), whereas in cMF and cLF theVCtAB was somewhat greater than the ThCtAB.aMe. Dif-ferent implementations of measuring the cartilage thickness(e.g., minimal distance from bone to cartilage, or minimaldistance from cartilage to bone, or the average of both)produced very similar values in all cartilage plates. Themaximal thickness over the total bone area (ThCtAB.Max)was found to be almost twice as high as the mean thickness(ThCtAB.Me) in the femorotibial plates. Reproducibilityerrors for cartilage volume divided by the tAB (VCtAB) weresimilar to those for the cartilage thickness over the totalbone area (ThCtAB) and tended to be smaller than thosefor cartilage volume (VC). The reproducibility errors werealso similar for different implementations of the thicknessmeasurements (see above). The maximal thickness overthe total bone area (ThCtAB.Max) and the average of thetop 1% greatest thickness values (ThCtAB.Mav) displayedlarger reproducibility errors than the averaged mean cartilagethickness over the total bone area (ThCtAB.Me) in allcartilage plates, but reproducibility errors for ThCtAB.Mavtended to be smaller than those for ThCtAB.Max. In terms ofthe rate of (and sensitivity to) change, the cartilage volumedivided by the total bone area (VCtAB), and the meancartilage thickness over the total bone area (ThCtAB.aMe)exhibited higher rates of change and greater SRMs (greatersensitivity to change) than cartilage volume (VC) in MT, butthe difference was only marginal in cMF. The rates of changeand SRMs for cartilage thickness over the covered bone area(ThCcAB) tended to be less than for cartilage thickness overthe total bone area (ThCtAB) and for cartilage volume (VC),independent of the specific implementation, but tended tobe greater than those for cartilage surface area (AC) andthe cartilage covered bone area (cAB). ThCtAB.Max andThCtAB.Mav showed low rates of change and SRMs, inparticular in cMF. Table 2 lists the percent change, the SRM,the significance level, and the precision error (test-retest)in MT and cMF (60% ROI) for different morphologicalvariables from this study [34]. In summary, the normalizedcartilage volume (VCtAB) and the mean cartilage thicknessover the entire subchondral bone area (ThCtAB.Me) tendedto be more reproducible and more sensitive to change (SRMup to −0.62) than cartilage volume (SRM up to −0.44),cartilage thickness over the cartilaginous area (ThCcAB;SRM up to −0.48) or maximal cartilage thickness (SRM upto −0.35) [34].

Other publications also reported that the sensitivityto change for ThCtAB or VCtAB was greater than forVC [30, 31], whereas others found comparable SRMs forthese variables [51, 52]. A recent paper [31] reported that,when cartilage loss was rapid (due to high mechanicalchallenge in mal-aligned knees), “horizontal” cartilage loss(i.e., an increase in denuded area = dAB) made a strongercontribution to the total cartilage loss (= reduction inThCtAB), whereas when cartilage loss was relatively slowin neutrally aligned knees, the “vertical” cartilage loss(reduction ThCcAB) made a stronger contribution. Thisfinding will need to be confirmed in other cohorts and

pathos-phyiological conditions. Wirth et al. [53] recentlyexplored the rate and sensitivity to change of the minimalcartilage thickness (ThCtAB.Min) and applied the measure-ment to central subregions of MT, LT, cMF, and cLF, respec-tively. In 156 participants of the Osteoarthritis Initiative(OAI), they found the one-year rate of the ThCtAB.Minchanges to be greater than those of ThCtAB.Me, but alsoreported a greater standard deviation, so that ThCtAB.Minwas found to be less sensitive to change than ThCtAB.Me.

A recent paper [54] investigated the mathematical rela-tionship between the above morphologic measurementsand explored whether a subset of the above variables fullyreflects differences observed in cartilage in cross-sectionaland longitudinal studies. The benefits of this reductionin variables are an increased statistical power due to lessmultiple comparison issues, an improved understanding ofrelationships between the morphologic measures of kneecartilage, and a greater efficiency in reporting the resultsin the literature. Buck et al. [54] used cross-sectional [55]and longitudinal (baseline to 2 year followup) 3T MR imagedata [29] from 152 women (77 healthy and 75 with kneeOA). They found that the total area of the subchondral bone(tAB), cartilage thickness (ThCtAB.tAB), and the percentageof denuded area of the subchondral bone (%dAB) explainedmore than 90% of the cross-sectional and longitudinalvariation in the full set of cartilage morphology measures,both in healthy and in osteoarthritic knees. The authorstherefore recommended these three variables as an efficientsubset for describing structural status and change in kneecartilage [54].

4. Reference Values for MRI-Based KneeCartilage Morphometry

Several groups have reported reference values of cartilagemorphology in healthy volunteers [44, 56, 57], includingtemplates/atlases for comparison of cartilage thickness distri-bution patterns between healthy reference subjects and OApatients [50, 58] and reference values for the radiographicjoint space width (JSW) [59]. Beattie et al. [59] foundthat measures of JSW did not significantly decrease withincreasing decade, but remained fairly constant throughoutthe lifespan in either sex; the same was observed for cartilagemorphometry measures. The authors suggested that theremay therefore be no need to differentiate a T- or Z-scorein OA diagnosis because cartilage thickness and JSW remainconstant throughout life in the absence of OA.

Recently, several authors have proposed the measure-ment of certain anatomically defined subregions withincartilage plates to determine the spatial pattern of cartilageloss [15, 60, 61] (Table 1). A recent analysis of a largepopulation-based cohort reported sex-specific normal valuesand potential maximal Z-scores for specific subregions ofthe femorotibial cartilage [62]. The authors studied 686Framingham participants (309 men, 377 women, age 62 ±8 years) without radiographic femorotibial OA (“normals”)and a subset of 376 Framingham participants (156 men,220 women) who additionally had no MRI features of

Arthritis 7

Table 2: Rate of change and sensitivity to change over 2 years in 28 participants with Kellgren-Lawrence grade [KLG] 3, and test-retestreproducibility in 33 participants with KLG0 to KLG3 for various cartilage morphology metrics and regions of interest in the medialfemorotibial compartment.

MT cMF 60%

MC% SRM P RMSCV% MC% SRM P RMSCV%

VC −2.3 −0.44 .03 2.5% −3.5 −0.32 .10 2.6%

tAB 0.5 0.37 .06 1.0% −0.1 −0.04 .84 1.1%

AC −0.9 −0.33 .09 1.0% −1.7 −0.22 .26 1.3%

cAB −1.0 −0.29 .14 1.0% −3.0 −0.36 .07 1.1%

VCtAB −2.7 −0.59 .00 1.9% −4.0 −0.33 .09 2.0%

ThCtAB.aMe −2.6 −0.58 .01 1.9% −3.6 −0.31 .12 1.7%

ThCtAB.bMe −2.8 −0.62 .00 1.9% −3.3 −0.29 .13 1.9%

ThCtAB.cMe −2.5 −0.56 .01 2.0% −3.9 −0.33 .10 1.7%

ThCcAB.aMe −1.4 −0.43 .03 1.9% −1.5 −0.18 .34 1.7%

ThCcAB.bMe −1.5 −0.48 .02 1.9% −1.2 −0.15 .43 1.9%

ThCcAB.cMe −1.3 −0.42 .04 2.0% −2.0 −0.23 .23 1.7%

ThCtAB.aMax −1.4 −0.27 .17 4.4% 0.0 0.00 .99 2.8%

ThCtAB.bMax −1.7 −0.30 .12 4.2% 0.7 0.10 .61 3.3%

ThCtAB.cMax −1.1 −0.18 .35 5.3% −0.5 −0.09 .65 3.2%

ThCtAB.aMav −1.4 −0.31 .11 3.8% −0.3 −0.05 .79 2.5%

ThCtAB.bMav −1.7 −0.35 .07 3.5% 0.0 0.00 .98 2.8%

ThCtAB.cMav −1.3 −0.25 .19 4.5% −0.5 −0.09 .65 2.8%

MC%: mean change in %, SRM: standardized response mean (= mean change/SD of change), P: level of significance of change using a paired t-test withoutadjustment for multiple comparisons; RMSCV%: root mean square coefficient of variation of test-retest acquisitions at baseline, with repositioning in betweenscans. For other abbreviations, please see Table 1. Note that values are given for the “long” femoral region of interest, that is, a 60% distance between thetrochlear notch and the posterior end of both femoral condyles.

cartilage lesions (“supernormals”). The Framingham par-ticipants had thinner cartilage in the medial (3.59 mm)than in the lateral femorotibial compartment (3.86 mm).Medially, the femur displayed thicker cartilage (1.86 mm)than the tibia (1.73 mm), and laterally the tibia thickercartilage (2.09 mm) than the femur (1.77 mm). The thickestcartilage was observed in central, and the thinnest in externalfemoro-tibial subregions. The mean values in Framingham“supernormals” and in non-exposed Osteoarthritis Initiative(http://www.oai.ucsf.edu/) reference participants (partici-pants without symptoms or risk factors of knee osteoarthritis(OA)) were very similar to those in Framingham “normals”.The authors concluded that adequate reference values couldbe obtained from populations without radiographic OA(independent of risk factors and their specific MRI lesionstatus), and that a cartilage thickness loss of approximately27% is required for attaining a Z-score of −2.

5. Imaging Protocols for Measurement ofCartilage Morphology Including Validation

Quantitative work performed on cartilage with MRI between1994 and 2006 has been summarized previously [8, 9, 63]and will not be repeated in this paper. Briefly, for quantifyingcartilage morphology, water-excitation (or fat-suppressed)T1-weighted spoiled gradient recalled echo acquisition inthe steady state (SPGR) or fast low angle shot (FLASH) at1.5 T or 3 T represent the current gold standard [9, 64, 65]for quantitative cartilage imaging. Double-echo steady-state

imaging (DESS) with water excitation has recently gainedinterest because of the faster acquisition time and lowerslice thickness that can be achieved (Figure 2) [33, 36, 66–68]. SPGR/FLASH sequences are readily available on almostall MRI scanners and do not require specific hard- orsoftware, whereas the DESS is currently only available fromone vendor [33]. Because the DESS acquires two separateimages with different echo times simultaneously, this addi-tionally provides potential opportunity to estimate T2 andto obtain morphological and compositional information ofthe cartilage from a single high-resolution data set [69]. Thisapproach is still undergoing validation.

The previously mentioned Osteoarthritis Initiative (OAI)(http://www.oai.ucsf.edu/) is a large research endeavor joint-ly sponsored by the National Institute of Health (NIH), theNational Institute of Arthritis and Musculoskeletal and SkinDiseases (NIAMS), and the pharmaceutical industry. Thisstudy in a cohort of 4800 participants is currently focusingon identifying imaging (and other) biomarkers for predictingand monitoring the onset and progression of symptomaticknee OA using-3 T MRI over a 4-year period (currentlybeing extended to 8-year followup). The OAI relies on anearly isotropic sagittal DESS sequence with water excitationin both knees for quantifying cartilage morphology andon a coronal FLASH sequence with water excitation inone knee of all participants [70]. Sagittal images have theadvantage that all cartilage plates of the knee (includingthe femoropatellar and femoro-tibial compartment) arevisualized, but suffer from partial volume effects in the

8 Arthritis

internal and external femoro-tibial subregions (Figure 3).Coronal images, in contrast, can delineate the femoro-tibialjoint and axial images visualize the patella with little partialvolume effects, but there is currently no consensus, whichof the above is the preferred orientation. A direct face-to-face comparison of 2 year changes measured in coronal andsagittal (SPGR) images revealed similar rates and patterns ofcartilage loss in the femoro-tibial joint [71].

The technical accuracy (validity) and test-retest precision(reproducibility) of quantitative cartilage measurements at1.5 T have been summarized in previous reviews [8, 9].Analyses based on 1.0 T images acquired with a dedicatedextremity scanner were found to be consistent with 1.5 Timaging, albeit less precise (reproducible) [72]. The use ofperipheral MRI scanners at lower field strengths potentiallypermits more widespread distribution of this technology,especially when access to high-field MRI is limited. Quantita-tive cartilage measurement at 0.2 T have also been proposed[18, 27, 28, 38–40] but have not been validated versusexternal standards or measurement at higher field strength.However, they were shown to display substantially largerprecision errors than measurements performed at higherfield strength. 3 T cartilage imaging has been cross-calibratedwith 1.5 T and lower precision errors than for 1.5 T imagingwere reported when acquiring thinner (coronal) slices of1.0 mm on a 3 T system [73]. Morphometric analysis fromDESS images, as acquired at 3 T in the OAI, was found tobe consistent with that from FLASH images and to displaysimilar test-retest precision errors as FLASH in the femoro-tibial joint, both using unpaired [33] and paired readingapproaches [67, 68]. In terms of sensitivity to change, Wirthet al. [36] performed a face-to-face comparison betweenFLASH and DESS over one year longitudinally in 80 knees.The study confirmed a high agreement between cartilagethickness measures as determined from FLASH and DESScross-sectionally [33] and a similar sensitivity to changeof coronal FLASH and sagittal DESS. Further, the studyrevealed a moderate correlation of the longitudinal one-year changes, indicating that it may be adequate to poolanalyses obtained with FLASH and DESS in larger statisticalanalyses [36]. Also, the authors found that analysis of every2nd slice (i.e., obtaining information every 1.4 mm) of thesagittal DESS displayed similar SRMs as compared withsegmentation of every 0.7 mm slice, both when either usingodd or even slice numbers [36]. Due to the near-isotropicresolution of the sagittal DESS, multiplanar reconstruction(MPR) in the coronal and axial planes is feasible [33, 67, 68].The rates of (and sensitivity to) change of coronal MPR DESSwas similar to that of coronal FLASH and sagittal DESS butdid not provide an advantage over the direct analysis of thesagittal DESS [36].

Generally, results from different vendors for cartilagemorphometry were shown to be comparable at 1.5 T [74]and at 3 T [75], although one study reported slight offsetsbetween different scanners and protocols from the samevendor [76]. At 3 T, precision errors of cartilage morphom-etry were observed to be similar for different vendors andscanners in a multicenter trial, and measurements wererelatively stable over a 3-month observation period [48].

The stability of geometric measurements over longer periodson phantoms was found to be satisfactory and comparablebetween several scanners of the same manufacturer over a 3-year period in the OAI [77].

Use of different coils has been evaluated at 3 T. Althoughthe test-retest precision was similar between a phased arrayand quadrature coil, certain offsets in cartilage morphologyoutcomes were observed [67]; these prohibit changes of thecoil between baseline and followup measurements. Cartilagemorphometry on images acquired 2 hours after intravenousGd-DTPA injection (for the purpose of simultaneous dGEM-RIC imaging) was reported to be highly correlated (r =0.85−0.95) with that on images obtained before the injectionof the contrast agent at baseline [78]. However, a 2-yearlongitudinal analysis in OA participants reported that thesensitivity to change of post-Gd-DTPA cartilage imaging wassubstantially less than that from images acquired prior tointravenous Gd-DTPA injection [79].

6. Rates of Change and Sensitivity toChange Observed in Knee OA

Numerous reports on longitudinal changes of cartilagemorphology in subjects with different grades of knee OAhave been published [8, 9, 29–31, 36, 52, 53, 61, 80–87].These studies have revealed variable results with regard tothe rates of cartilage loss and SRM [6, 8, 9] Two studiesreported almost no loss in cartilage volume over a 1-year [87]and 3-year period [80], respectively, whereas other studiesreported up to 7% annual cartilage loss in the femoro-tibial joint [82]. Reasons for this may include variabilityin imaging and image analysis technology, differences inrisk factor profiles between cohorts, differences in studyduration, experience and blinding of the readers, and others.A recent study [88] tested the hypothesis that “Proof ofConcept” studies with shorter durations may be achievablewith 3 T MRI, by selecting populations at high risk of rapidmedial femoro-tibial progression and using advanced imageanalysis techniques. Female participants with knee pain, abody mass index ≥25, and radiographic evidence of medialOA and varus mal-alignment were monitored over 3 and 6months, respectively, and anatomically corresponding ROIswere identified on each image by using a three-dimensionalstatistical shape model of the bone surface. The primaryoutcome was the change in cartilage thickness in the aspectof cMF that is exposed within the meniscus window duringarticulation, excluding the peripheral aspects of the femoralsurface. Despite these efforts, no change in ThCtAB wasdetected at P < .05 at 3 or 6 months followup; the meanchange at 3 months from a log-scale ANOVA model was−2.1% [95% confidence interval (CI) (−4.4%, +0.2%)] andthe change over 6 months was 0.0% [95% CI (−2.7%,+2.8%)]. Changes in the lateral tibia were significant at 6month followup (−1.5%), but only without correction formultiple comparisons. The authors concluded that the smallinconsistent compartment changes and the relatively highvariability in cartilage thickness changes seen in the studyprovided no confidence for a 3- or 6-month study, not even

Arthritis 9

based on a patient population selected for rapid progression[88].

Analyses of the first release of 160 participants of the OAIprogression cohort (baseline and year 1 followup data) foundsignificant change of up to 2% per annum, with substantiallyhigher rates of progression in the cMF than in the MT,and higher rates in LT than in cLF [30, 52, 53]. However,this pattern of change was not entirely consistent acrosscohorts, when focusing on the SRM rather than on the rate ofchange [10, 29, 31]. Several studies therefore have taken theapproach of reporting the aggregate thickness in the tibia andweight-bearing femur (MFTC or LFTC) [29–31, 36, 67, 89].One study suggested that longitudinal changes in VC in thetibia and in the weight-bearing femur are highly correlated[82], and that the measurement of only tibial cartilage istherefore sufficient. However, given that at least some cohortsappear to display larger changes and higher SRMs in theweight-bearing femur than in the tibia [30, 52, 53], thisapproach has limitations.

Medial and lateral femoro-tibial cartilage loss as well aspatellar cartilage loss were found to be not significantly asso-ciated with each other [90]. The ratio of medial versus lateralcartilage loss was reported to be 1.4 : 1 in knees with neutralbiomechanical alignment, consistent with higher mechanicalloads being transferred across the medial compartment inneutral knees [31]. In varus knees, the ratio was 3.7 : 1, andin valgus knees it was 1 : 6.0, confirming that knee alignmentis an important determinant of medial versus lateral rates ofcartilage loss [31].

After anterior cruciate ligament rupture, a reductionof cartilage volume and thickness was observed in thefemoral trochlea (TrF), while an increase was found inthe weight-bearing medial femur (cMF) [91]. The latterobservation may be consistent with cartilage swelling orhypertrophy observed as a sign of early OA in variousanimal models [92–96]. A recent cross-sectional study foundsignificantly thicker cartilage in the medial compartment ofwomen with medial radiographic KLG2 OA compared withhealthy knees [55] and significantly thinner cartilage in somesubregions in knees with medial radiographic OA with jointspace narrowing (JSN, i.e., KLG3). These observations wereconfirmed by a large cross-sectional analysis of more than1000 OA participants, in which the authors [97] confirmeda significantly greater cartilage thickness in KLG2 comparedto healthy knees, specifically in the external subregion of themedial femur (ecMF), both in men and in women. Thesefindings have suggested that there may be an initial phaseof cartilage swelling/hypertrophy in knee OA, particularlyat the KLG2 stage, which is characterized by osteophyteswithout a reduction in JSW. This has been supported byrecent longitudinal observations by Buck et al. [98] whoexplored whether the 2-year longitudinal change in cartilagethickness in femoro-tibial subregions (see below) of kneeswith radiographic osteoarthritis (ROA) differed from thatin healthy knees. Knees from 75 women with definitesigns of medial radiographic OA were compared with 77asymptomatic healthy controls without radiographic OA.A substantial portion of ROA knees were classified as havinglongitudinal cartilage thinning (28%) or thickening (20%)

in at least one medial femoro-tibial subregion comparedwith longitudinal changes in healthy knees, and only 5%showed both subregional thinning and thickening at thesame time, across (different) medial subregions. Whereas theestimated proportion of KLG3 knees with significant medialcartilage thinning (46%) was substantially greater than thatwith cartilage thickening (18%), the estimated percentagesof KLG2 knees with significant medial thinning (20%) andthickening (23%) were similar. The subregion in whichcartilage thickening was observed was ecMF in the majorityof the cases. The authors concluded that OA may not bea one-way road of cartilage loss and that particularly inearly radiographic OA, cartilage changes may occur in bothdirections simultaneously, that is, cartilage thinning andcartilage thickening. This may provide a reason why relativelysmall (and variable) rates of change have been observed inOA cohorts, and why short-term trials are challenging [88].

7. Spatial Patterns of Cartilage Loss inKnee OA as Derived by Subregional Analysis

As mentioned above, recent efforts have been focused onmeasuring anatomically defined subregions within cartilages[15, 60, 61], with the aim of elucidating spatial pattern ofcartilage thinning, and to potentially identify (sub) regionswith increased rates of (and sensitivity to) cartilage loss inintervention trials.

7.1. Cross-Sectional Studies. The previously mentionedcross-sectional study by Hellio Le Graverand et al. [55]reported that cartilage “thinning” in female knees withmedial JSN (KLG3) was most evident in the central sub-region of the cMF and in the external subregion of theMT, and in the internal subregion of the LT. This wasextended by the study of Frobell et al. [97], who reportedthat the external medial tibia showed the greatest reductionin cartilage thickness (z-scores−5.1/−5.6 in men/women) inknees with medial joint space narrowing (OARSI JSN) grade3 and the external lateral tibia (z scores −6.0 for both sexes)the greatest reduction in knees with lateral JSN grade 3. Theauthors, however, reported that at least 25% of the averagenormal cartilage thickness was maintained in all subregionsof end-stage ROA knees.

Although these differences were generally not affectedwhen possible effects of demographic covariates (height andBMI) were considered [55, 97], it is difficult to excludeconfounding by interperson differences in cross-sectionalstudies. Therefore, Eckstein et al. [99] performed a within-person, between-knee comparison in 80 participants of theOAI who displayed medial JSN in one knee, but no medialor lateral JSN in the contralateral knee. The strength ofthis approach is that it rules out confounding from person-specific demographic features, and that it is potentially moresensitive to detecting differences cross-sectionally, given themuch smaller magnitude of side differences between kneeswithin the same (healthy) person compared with differencesacross (healthy) subjects [100]. The authors estimated themagnitude of cartilage thickness reductions to be 190 μm

10 Arthritis

(5.2%) in the medial femoro-tibial compartment (MFTC)with JSN OARSI grade 1, 630 μm (18%) with OARSI grade2, and 1560 μm (44%) with OARSI grade 3 [101, 102]. Sidedifferences were greater in cMF than in MT, and greater inMT than in pMT [99]. Within MT the greatest differenceswere observed in the external and central subregions, andwithin MF the greatest differences were observed in thecentral subregion of the weight-bearing portion of MF. Whenevaluating A-P subregions in the MF [103], the greatestdifferences between mJSN and contra-lateral no-mJSN kneeswere observed in regions located between 30◦ and 75◦ at theMF.

7.2. Longitudinal Studies. Pelletier et al. [61] reported thatthe rate of change in cartilage morphology in the centralaspects of the femoro-tibial joint exceeded that in totalcartilage plates, but found that the SRM was not improvedbecause of the higher variability of subregional changes [51].Wirth et al. [53] found the sensitivity to change (SRM) in thecentral MT to be slightly greater than for the total MT in asubsample of OA Initiative participants, but this finding wasnot confirmed in the cMF.

Figure 3 summarizes the rates of change (%/annum) andthe sensitivity to change (SRM) for different subregions fromthree published studies.

(a) A 2-year multicenter study at 3T (Pfizer A 9001140).Because healthy reference participants and partici-pants with KLG 2 did not show significant changesin cartilage morphology [29], results from KLG 3participants were used (n = 28).

(b) The 2nd cohort included is a first release of baselineand year 1 followup data from the OAI progressionsubcohort [53]. Results of a subcohort with a highrisk of progression (BMI > 30; KLG ≥ 2) wereincluded (n = 54).

(c) The 3rd cohort included was from the MAK study[11, 31]. Data from a subcohort of participants withneutral knee alignment were included in the analysis(n = 74).

The central and external part of cMF, and the externaland central aspect of MT displayed the relatively greatestchange across subregions in the MFTC (Figure 3). Withthe exception of the external medial femur, these regionsconsistently displayed greater changes than the total cartilageplate across the studies. In the LFTC, the central, internal,and posterior LT displayed the relatively greatest changes,and no relevant average changes (across studies) wereobserved in the cLF (Figure 3). Rates of change in the centraland internal LT were consistently greater than those for thetotal cartilage plates. Please note that the patterns for thesensitivity to change (SRM; Figure 3) are similar to thoseof the rates of change, but not identical. Consistent withother observations in the literature [51], we found that thesensitivity to change in the subregions was not consistentlyhigher than in the total plates across the above three studies.However, analysis of the central aspects (subregions) of

the medial and lateral femoro-tibial compartments revealedconsistently greater SRMs than the analysis of the entireMFTC and LFTC, respectively.

Wirth et al. [103] recently presented a method whichextended the previously developed method of subregionsin the weight-bearing femoro-tibial joint [15] to anterior-posteriorly spaced subregions across the entire femoralcondyle. This method was applied to participants fromthe OAI and confirmed that cartilage thinning in theanterior (weight-bearing) region of the MF was greaterthan that in the posterior aspect of the MF. The authorsreported the greatest longitudinal changes (and SRM) tobe located at 30 to 60◦ (from the trochlear notch [0◦] tothe posterior/superior end of the MF (150◦), with a slightvariation between knees with different OARSI JSN grades.

7.3. Ordered Value Approach (Subregion Ranking). Buck etal. [104] analyzed patterns of subregional cartilage change[15] in individual knees and found highly variable patternsof change. To compare the rate of change between twogroups (i.e., ROA knees with healthy knees, or DMOADtreated knees with control knees) he therefore recommendedthe use of ordered values (OVs) or ranking system, inwhich the subregional changes (in MFTC) were assignedto ranked orders in each knee, that is, the subregionwith greatest magnitude of cartilage thinning to OV1, theone with the second greatest magnitude to OV2, and theone with the smallest magnitude of cartilage thinning (orwith the greatest magnitude of cartilage thickening) to thehighest rank order. When averaging longitudinal changesin cartilage thickness (ThCtAB) across these OVs (whichvary in location across subjects), the authors found that theminimal P value (Wilcoxon) for the differences in 2-yearchange in medial cartilage thickness in a relatively smallnumber of knees with radiographic OA and JSN (KLG3)versus healthy knees (KLG0) was P = .001, with OV1 toOV4 displaying significant differences between both groups.When averaging changes across compartments, plates, orsubregions (i.e., the conventional approach), in contrast,only one medial subregion displayed significant differences(in the rate of change) between KLG 3 and KLG0 knees(P = .037). Cartilage thickening was significantly greater inknees with radiographic OA (definite osteophytes) withoutJSN (KLG2) versus KLG0 knees in one medial subregionusing the conventional approach (P = .02), but in two OVsusing the ordered values approach (minimal P = .007).The authors concluded that the ordered values approach wasmore sensitive in detecting cartilage thinning in KLG3 versusKLG0, and cartilage thickening in KLG2 versus KLG0 knees,respectively. The authors also suggested that this methodwas particularly useful in the context of comparing a cohorttreated with a disease-modifying OA drug versus one treatedwith a placebo, or in detecting risk factors of OA progression.

Wirth et al. [105] recently extended this approach toinclude eight medial and eight lateral (n = 16) subregions.They reported significantly greater cartilage loss in KLG3than in KLG2 knees, the ordered value approach again dis-playing considerably smaller P values than the conventional

Arthritis 11

approach. This opens new possibilities of including partici-pants with medial and lateral OA (or with varus and valgusmal-alignment) into a study, without the need of definingcartilage thickness changes in a certain compartment, plateor subregion as the primary endpoint. The relevant questionwould then NOT be whether a certain risk factor is associatedwith or whether a drug can modify cartilage thicknesschanges in a given location (region), BUT whether the riskfactor is associated with or whether the drug can modifythe change in cartilage thickness wherever it occurs in anindividual knee.

8. Comparison of MRI Changes withRadiographic Changes in Knee OA

Several studies found only weak correlations between MRI-detected cartilage loss and OA progression in radiography[83, 87, 106]. However, a recent publication reported astronger correlation when the longitudinal reduction in JSWin radiographs was compared with cartilage loss in thecentral aspect of the MFTC [61]. Whereas some studiesfound a higher rate and sensitivity to change of MRI-based measurements of cartilage morphology compared withradiography [83, 85, 107], a recent analysis reported asomewhat greater SRM (−0.62) for fluoroscopy-based LyonSchuss radiography versus ThCtAB of the MT measured withMRI (−0.59) [29]. However, the authors found the SRM forfixed flexion radiography, a commonly used nonfluoroscopicprotocol that also is used in the OAI [108–112], to besubstantially less (SRM = −0.20) in the same study [29].The authors argued that the relatively high SRM of theminimal JSW measured by Lyon Schuss may be due tothe fluoroscopic guidance providing optimal alignment ofthe anterior and posterior tibial rim, and to radiographybeing performed under weight-bearing conditions wherethe cartilage tissue is compressed, while MRI is performedin a supine non weight-bearing position. Also, it mustbe kept in mind that radiographic assessment of JSWdepends also on meniscus extrusion, and not only oncartilage thickness [113–115] and that meniscus pathology,in particularly subluxation, can therefore cause changes inJSW over time in the absence of cartilage loss. Duryea etal. recently compared the responsiveness (=sensitivity tochange) of radiography with that of MRI in the first releaseof the OAI cohort (150 subjects) over 12 months [116].The radiographic JSW measurements relied on automatedsoftware to delineate the femoral and tibial margins [117,118]. Measures included the medial compartment minimumJSW and JSW at fixed locations that were compared topreviously published cartilage morphology measures [52].The SRM value for radiographic JSW measured at theoptimal fixed location was −0.32 compared to −0.39 for themost responsive MRI measure. For a subsample of KLG2 orKLG3 knees, the most responsive SRM values were −0.34and −0.42, respectively. The authors concluded that new(fixed distance) measures of JSW changes were superior toconventional minimal JSN measures and provide a similarsensitivity to change as MRI.

9. Risk Factors of Cartilage Loss inKnee OA as Identified by QuantitativeCartilage Imaging with MRI

Great interest is directed at identifying risk factors (predic-tors) of subsequent cartilage loss, both to understand thepathophysiology of the disease and to be able to identifyso-called fast progressors for inclusion in pharmacologicalintervention studies that attempt to show protection fromstructural change over relatively short periods (e.g., [88]).This paragraph will focus on studies that have reportedcorrelations between risk factors of progression and quan-titative measures of cartilage morphology, but not thosethat have relied on semi-quantitative scoring of MRI orquantitative measurement of JSW. The list of (potential) riskfactors for cartilage loss is not complete, but encompassesimportant examples examined both from cross-sectional andlongitudinal studies. Risk factors associated with higher ratesof progression were the following.

Advanced Radiographic OA and Low Cartilage Thickness atBaseline. Opposite to earlier assumptions and a synthesis ofevidence from radiographic studies [119], recent evidencesuggests that advanced radiographic OA (JSN) is a strong(if not the strongest) predictor of fast progression (i.e.,cartilage loss). There has been evidence that knees withhigher KL grades and increased JSN [29, 30, 53, 61] displaygreater rates of (and sensitivity to) change than those withlower KL grades and without baseline JSN. An analysisof specific radiographic features in a sample from theOAI found that osteophyte status (at baseline) was notassociated with medial cartilage loss over 12 months but thatknees with medial joint space narrowing showed a trendtowards higher rates of change than those without, andthat knees with medial femoral subchondral bone sclerosisdisplayed significantly greater rates of progression than thosewithout [120]. The same study also found that low baselinecartilage thickness was a strong predictor of longitudinalloss in cartilage thickness [120], whereas an earlier studyhad reported that higher baseline cartilage volume [81] wasstrongly associated with increased cartilage loss. A within-person, between-knee comparison in painful knees selectedfrom the OAI [35] recently reported that the cartilage losswas greater in knee with radiographic JSN than in contra-lateral knees without JSN in the same subjects, and thatthe side differences were greater with higher grades of JSN.Progression was particularly fast in the small subgroup withOARSI JSN grade 3 knees [35].

Meniscus Extrusion and Tears/Damage [61, 84, 85]. Meniscustears were found to be associated with greater tibial plateaubone area, but not with reduced tibial cartilage volume ina two year longitudinal study [121]. However, Sharma etal. [11] reported a significant relationship of cartilage losswith meniscus tears, albeit not with meniscus extrusion. Arecent analysis found site-specific relationships between localmeniscus tears and subregional cartilage loss, suggesting thata tear in the anterior horn, central part, or posterior horn of

12 Arthritis

the meniscus was associated with increased cartilage loss inadjacent tibial subregions [122]. Crema et al. [123] reportedgrade 2 and 3 medial meniscus lesions to be associatedwith greater cartilage loss in the femoro-tibial compartment,but not grade 1 lesions (=intrasubstance meniscus signalchanges). They concluded that the protective function ofthe meniscus was preserved in case of these early lesions.Recent evidence suggests that the meniscus may undergoa phase of hypertrophy in OA [124, 125]. Raynauld et al.[126] observed that selecting a subcohort of participantswith meniscus tears/extrusion did not improve the ability toidentify treatment effects of a potentially structure modifyingdrug, because of the larger standard deviation of the changein the participants with meniscus pathology.

Knee Malalignment and Adduction Moment. A strong rela-tionship was observed between (varus and valgus) mal-alignment and the ratio of cartilage loss in MFTC versusLFTC [11, 31, 127, 128]. After adjustment for meniscuschanges, the study by Sharma et al. [11] found that varusmalalignment and medial meniscus damage both predictedmedial tibial cartilage volume (and thickness) loss. Incontrast, medial-lateral joint laxity, measured with a deviceapplying a fixed varus and valgus load, was not found tohave consistent effects and was not a significant predictorof cartilage loss in models fully adjusted for alignmentand meniscus damage [11]. Teichtahl et al. [129] showedan increase in varus mal-alignment between baseline andfollowup to be associated with an increase in the rateof MT cartilage loss, whereas there was no significantcorrelation with the rate of cartilage loss of the LT. Theauthors concluded that methods to reduce progression ofvarus alignment may also delay the progression of medialfemoro-tibial OA. Frontal plane knee valgus mal-alignmentwas also correlated with patellar cartilage loss [130]. Ina largely nonarthritic cohort, in contrast, no correlationbetween cartilage loss and mal-alignment was identified[131]. A recent cross-sectional analysis revealed that a higherpeak knee adduction moment was observed in participantswith medial compared to those with lateral meniscus tears[132]. Participants with a higher knee adduction momentdisplayed a larger medial meniscus extrusion and lowermedial meniscus height, whereas the inverse relationship wasobserved for the lateral meniscus. A higher knee adductionmoment was also associated with a higher ratio of the medialto lateral tibial subchondral bone area, whereas cartilagethickness and denuded areas in the tibia and femur were notrelated to the knee adduction moment. Similar results werefound for the relationship between knee adduction angularimpulse and meniscus, cartilage, and bone morphology[132].

High BMI. In contrast with a synthesis from the radio-graphic literature [119], MRI-based studies on progressionhave found higher rates of cartilage loss in subjects with ahigh BMI [30, 53, 61, 83, 85, 90, 133]. This relationship wasalso suggested to exist in the patella in subjects without OA[134].

Bone Marrow Alterations [61, 85]. Raynauld et al. [126]reported that although bone marrow lesions and cysts didnot increase significantly in size over 24 months in anOA cohort, there was a significant correlation between sizechange of bone marrow lesions and cysts with the loss ofcartilage volume in the medial femoro-tibial compartment.A relationship between very large bone marrow lesions andlateral tibial cartilage loss was also reported in asymptomaticpersons [135, 136].

Focal Cartilage Lesions or Defects as Graded by Visual Scoringand Denuded Areas as Determined Quantitatively from MRI.Cartilage defects at baseline (visual scoring) appeared to beassociated with longitudinal measurement of quantitativecartilage loss in the same compartment in OA subjects [137,138], although the second of the two above studies [138] onlyfound a significant relationship in the femoro-patellar butnot in the femoro-tibial joint. Other studies reported that thepresence of cartilage defects predicted knee cartilage loss alsoin asymptomatic individuals without radiographic knee OA[139, 140]. It was hypothesized that tibial subchondral bonearea expansion may lead to the development of knee cartilagedefects (which are associated with future cartilage loss) and ispredictive of the need for knee joint replacement in subjectswith knee OA, independent of radiographic change [141].Morphometric studies have recently provided evidence thatareas of denuded subchondral bone (dABs), as determinedby segmentation at baseline [142], also predict subsequentcartilage loss [120, 143]. Hunter et al. [143] reported thatin a subsample of knees with no denuded area (at baseline)the SRM for subsequent cartilage volume loss was −0.25,whereas it was−0.30 in the knees with intermediate denudedareas and −1.0 in knees in knees with severe denudedareas. Denuded areas were observed to either originatefrom cartilage loss or from internal osteophytes [142]. In asubsample from the OAI, almost half of the men and a thirdof the women displayed dABs; 61% of the dABs representedinternal osteophytes. One of 47 knees with KLG0 displayedany dAB, whereas 29 of the 32 KLG4 knees were affected.There were significant relationships of dAB with increasingKL grades (P < .001) and with ipsi-compartimental JSN.Internal osteophytes were more frequent laterally (mainlyposterior tibia and internal femur), whereas full thicknesscartilage loss was more frequent medially (mainly externaltibia and femur).

Molecular Markers from Biological Fluids. Dam et al. [144]reported a significant association between baseline levels ofthe C-terminal telopeptide of type II collagen (CTXII) andcartilage loss in 158 study participants (36 with ROA at base-line) using low field (0.2 T) MRI–based cartilage loss. In thisstudy, elevated CTXII was also associated with radiographicprogression (by KLG or JSN, but did not reach statisticalsignificance [144]. Bruyere et al. [145] followed 62 patientswith knee OA using 1.5 Tesla MRI and found that baselinecartilage oligomeric matrix protein (COMP), C-terminaltelopeptide of type I collagen (CTXI), and CTXII didnot correlate with one-year changes in cartilage thickness,

Arthritis 13

but longitudinal increase in CTXII over three months did(P = .04). Pelletier et al. [146] reported higher baselinevalues of interleukin 6, C-reactive protein, and COMP tobe predictive of greater cartilage loss with MRI, whereasEckstein et al. reported a relatively large set (n = 16) of differ-ent molecular markers of bone formation, bone resorption,cartilage synthesis, cartilage degradation, and inflammationtake at baseline to be substantially less predictive of cartilageloss than simple radiographic measures, such as reducedJSW, or low baseline cartilage thickness [147].

Other Risk Factors. Some evidence has been provided thatsmoking may be associated with increased cartilage loss[140, 148, 149] in line with previous radiographic studies,but other factors such as age, sex, pain, function, physicalactivity levels, synovitis (effusion), sex hormone levels, andserum or urine biomarkers were not consistently found tobe associated with cartilage thinning measured quantitativelywith MRI and studies (including those with radiography)have produced partially contradictory results [119].

10. The Correlation of MRI-Based CartilageLoss with Clinical Outcomes and TreatmentResponse in Knee OA

Estimates of tibial cartilage loss over two years were suggestedto be correlated with those over 4.5 years, albeit the authorsdid not report the consistency of the longitudinal changesin the second versus the first observation period [86]. Moreimportantly, however, the rate of change in VC over 2years was significantly associated with total knee arthroplasty(TKA) at year 4 [150]. For every 1% increase in the rate ofcartilage loss there was a 20% increased risk of undergoingTKA and participants in the highest tertile of tibial cartilageloss had a 7.1 higher odds of TKA than those in the lowesttertile. In contrast, radiographic scores of OA did not predictTKA in the same study. A more recent study concludedfrom the same sample that when subchondral bone cystswere present, cartilage loss and risk of knee replacementwere higher than if only bone marrow lesions were present,suggesting that cysts identify those that may benefit mostfrom the prevention of structural disease progression [151].These are important findings as they link longitudinalchanges in cartilage morphology as a potential surrogatemeasure of disease progression, to a clinical outcome (i.e.,how a patient feels or functions, or how long the knee“survives” [TKA]).

Raynauld et al. [107] recently reported that licofelone(a drug that inhibits both cyclooxygenase and lipoxygenase)significantly reduced cartilage loss over time when aver-aged over both femoro-tibial compartment, and that MRIwas superior to radiographs in demonstrating a structuremodifying effect in this multicentre trial. Interestingly, theeffects were significant only in the lateral, but not in themedial compartment, although the participants had beenselected for medial femoro-tibial radiographic OA, and themedial compartment had thus been defined as the primaryendpoint. To date, no structure or disease-modifying drug

(SMOAD or DMOAD) has yet been approved by regulatoryagencies, neither based on radiographic nor on MRI-basedevidence of structure modification in knee cartilage.

11. Future Directions of the Field

In the Osteoarthritis Initiative, baseline, 12-, 24-, and 36-month followup clinical, radiographic, and MRI data havebeen made publicly available for approximately 4800 partic-ipants of the OAI cohort (http://www.niams.nih.gov/ne/oi/),and central readings of fixed flexion radiographs (OARSIatlas scores [101, 102]), quantitative measurements of radio-graphic joint space width (JSW) as well as quantitative car-tilage morphology outcomes from MRI are available for var-ious subsets that have been read/analyzed by expert readingfacilities. These and the results of other large epidemiologicalstudies will provide ample opportunity for collaborativeresearch and should allow the research community to makerapid progress in understanding the risk factors involvedin quantitative cartilage loss in OA. Most importantly, itwill allow one to determine which imaging biomarkers canbest predict clinical outcomes, such as real or virtual TKA[152]. This will be an important step in validating novelcartilage imaging biomarkers and approaches as surrogatemeasures of disease progression, particularly in therapeuticintervention trials. Once the clinical importance of theseimaging biomarkers are established, further improvementsin imaging hardware, coils, sequences, and image analysisalgorithms may foster a more automated analysis of cartilagemorphology, composition, and other articular tissues thancurrently possible. This will be of particular importance oncestructure- or disease-modifying drugs become available, asthis may require monitoring the treatment response in largesets of OA patients. Currently, quantitative MRI of articularcartilage represents a powerful research tool in experimental,epidemiological, and pharmacological intervention studies.Once structure- or disease-modifying drugs (SMOADs orDMOADs) will become available, quantitative MRI of thecartilage may also play a more important role in clinicaldecision making and practice.

Disclosures

F. Eckstein is co-owner and CEO of Chondrometrics GmbH,a company that licenses software to academic researchers andprovides image analysis service for academic researchers andthe pharmaceutical industry. F. Eckstein provides consultingservices to Merck Serono Inc. and Novartis Inc. and hasreceived research funding from Pfizer, Eli Lilly, MerckSerono,Glaxo Smith Kline, Centocor Research, Wyeth, and Novartis.W. Wirth has a part time appointment with ChondrometricsGmbH.

References

[1] C.-C. Gluer, R. Barkmann, H. K. Hahn et al., “Parametricbiomedical imaging—what defines the quality of quantitativeradiological approaches?” RoFo Fortschritte auf dem Gebiet

14 Arthritis

der Rontgenstrahlen und der Bildgebenden Verfahren, vol. 178,no. 12, pp. 1187–1201, 2006.

[2] P. Augat and F. Eckstein, “Quantitative imaging of muscu-loskeletal tissue,” Annual Review of Biomedical Engineering,vol. 10, pp. 369–390, 2008.

[3] D. Burstein, M. Gray, T. Mosher, and B. Dardzinski, “Mea-sures of molecular composition and structure in osteoarthri-tis,” Radiologic Clinics of North America, vol. 47, no. 4, pp.675–686, 2009.

[4] M.-P. H. Le Graverand, S. Mazzuca, M. Lassere et al., “Assess-ment of the radioanatomic positioning of the osteoarthriticknee in serial radiographs: comparison of three acquisitiontechniques,” Osteoarthritis and Cartilage, vol. 14, no. 1, pp.37–43, 2006.

[5] G. A. Cline, J. M. Meyer, R. Stevens, C. Buckland-Wright,C. Peterfy, and J. F. Beary, “Comparison of fixed flexion,fluoroscopic semi-flexed and MTP radiographic methods forobtaining the minimum medial joint space width of theknee in longitudinal osteoarthritis trials,” Osteoarthritis andCartilage, vol. 14, no. 1, pp. 32–36, 2006.

[6] A. Guermazi, D. Burstein, P. Conaghan et al., “Imaging inosteoarthritis,” Rheumatic Disease Clinics of North America,vol. 34, no. 3, pp. 645–687, 2008.

[7] D. J. Hunter, P. G. Conaghan, C. G. Peterfy et al.,“Responsiveness, effect size, and smallest detectable differ-ence of Magnetic Resonance Imaging in knee osteoarthritis,”Osteoarthritis and Cartilage, vol. 14, no. 1, pp. 112–115, 2006.

[8] F. Eckstein, D. Burstein, and T. M. Link, “Quantitative MRIof cartilage and bone: degenerative changes in osteoarthritis,”NMR in Biomedicine, vol. 19, no. 7, pp. 822–854, 2006.

[9] F. Eckstein, F. Cicuttini, J.-P. Raynauld, J. C. Waterton,and C. Peterfy, “Magnetic resonance imaging (MRI) ofarticular cartilage in knee osteoarthritis (OA): morphologicalassessment,” Osteoarthritis and Cartilage, vol. 14, no. 1, pp.46–75, 2006.

[10] F. Eckstein, A. Guermazi, and F. W. Roemer, “QuantitativeMR imaging of cartilage and trabecular bone in osteoarthri-tis,” Radiologic Clinics of North America, vol. 47, no. 4, pp.655–673, 2009.

[11] L. Sharma, F. Eckstein, J. Song et al., “Relationship of menis-cal damage, meniscal extrusion, malalignment, and jointlaxity to subsequent cartilage loss in osteoarthritic knees,”Arthritis and Rheumatism, vol. 58, no. 6, pp. 1716–1726,2008.

[12] C. G. Peterfy, A. Guermazi, S. Zaim et al., “Whole-organmagnetic resonance imaging score (WORMS) of the knee inosteoarthritis,” Osteoarthritis and Cartilage, vol. 12, no. 3, pp.177–190, 2004.

[13] S. Reichenbach, M. Yang, F. Eckstein et al., “Does cartilagevolume or thickness distinguish knees with and without mildradiographic osteoarthritis? The Framingham Study,” Annalsof the Rheumatic Diseases, vol. 69, no. 1, pp. 143–149, 2010.

[14] F. Eckstein, G. Ateshian, R. Burgkart et al., “Proposal for anomenclature for Magnetic Resonance Imaging based mea-sures of articular cartilage in osteoarthritis,” Osteoarthritisand Cartilage, vol. 14, no. 10, pp. 974–983, 2006.

[15] W. Wirth and F. Eckstein, “A technique for regional analysisof femorotibial cartilage thickness based on quantitativemagnetic resonance imaging,” IEEE Transactions on MedicalImaging, vol. 27, no. 6, Article ID 4359064, pp. 737–744,2008.

[16] J. Hohe, S. Faber, T. Stammberger, M. Reiser, K.-H.Englmeier, and F. Eckstein, “A technique for 3D in vivoquantification of proton density and magnetization transfer

coefficients of knee joint cartilage,” Osteoarthritis and Carti-lage, vol. 8, no. 6, pp. 426–433, 2000.

[17] J. Hohe, S. Faber, R. Muehlbauer, M. Reiser, K.-H. Englmeier,and F. Eckstein, “Three-dimensional analysis and visualiza-tion of regional MR signal intensity distribution of articularcartilage,” Medical Engineering and Physics, vol. 24, no. 3, pp.219–227, 2002.

[18] A. A. Qazi, J. Folkesson, P. C. Pettersen, M. A. Karsdal, C.Christiansen, and E. B. Dam, “Separation of healthy andearly osteoarthritis by automatic quantification of cartilagehomogeneity,” Osteoarthritis and Cartilage, vol. 15, no. 10,pp. 1199–1206, 2007.

[19] E. J. McWalter, W. Wirth, M. Siebert et al., “Use ofnovel interactive input devices for segmentation of articularcartilage from magnetic resonance images,” Osteoarthritisand Cartilage, vol. 13, no. 1, pp. 48–53, 2005.

[20] S. Solloway, C. E. Hutchinson, J. C. Waterton, and C.J. Taylor, “The use of active shape models for makingthickness measurements of articular cartilage from MRimages,” Magnetic Resonance in Medicine, vol. 37, no. 6, pp.943–952, 1997.

[21] T. Stammberger, F. Eckstein, M. Michaelis, K.-H. Englmeier,and M. Reiser, “Interobserver reproducibility of quantitativecartilage measurements: comparison of B-spline snakes andmanual segmentation,” Magnetic Resonance Imaging, vol. 17,no. 7, pp. 1033–1042, 1999.

[22] Z. A. Cohen, D. M. McCarthy, S. D. Kwak et al., “Knee carti-lage topography, thickness, and contact areas from MRI: in-vitro calibration and in-vivo measurements,” Osteoarthritisand Cartilage, vol. 7, no. 1, pp. 95–109, 1999.

[23] J. A. Lynch, S. Zaim, J. Zhao, A. Stork, C. G. Peterfy, andH. K. Genant, “Cartilage segmentation of 3D MRI scans ofthe osteoarthritic knee combining user knowledge and activecontours,” in Medical Imaging 2000: Image Processing, vol.3979 of Proceedings of SPIE, pp. 925–935, February 2000.

[24] P. M. M. Cashman, R. I. Kitney, M. A. Gariba, and M. E.Carter, “Automated techniques for visualization and map-ping of articular cartilage in MR images of the osteoarthriticknee: a base technique for the assessment of microdamageand submicro damage,” IEEE Transactions on Nanobioscience,vol. 1, no. 1, pp. 42–51, 2002.

[25] C. Kauffmann, P. Gravel, B. Godbout et al., “Computer-aidedmethod for quantification of cartilage thickness and volumechanges using MRI: validation study using a syntheticmodel,” IEEE Transactions on Biomedical Engineering, vol. 50,no. 8, pp. 978–988, 2003.

[26] S. D. Pathak, L. Ng, B. Wyman et al., “Quantitative imageanalysis: software systems in drug development trials,” DrugDiscovery Today, vol. 8, no. 10, pp. 451–458, 2003.

[27] J. Folkesson, E. Dam, O. F. Olsen, P. Pettersen, and C. Chris-tiansen, “Automatic segmentation of the articular cartilagein knee MRI using a hierarchical multi-class classificationscheme,” in Proceedings of the 8th International Conference onMedical Image Computing and Computer-Assisted Interven-tion (MICCAI ’05), vol. 3749 of Lecture Notes in ComputerScience, pp. 327–334, Palm Springs, Calif, USA, October2005.

[28] J. Folkesson, E. B. Dam, O. F. Olsen, P. C. Pettersen, and C.Christiansen, “Segmenting articular cartilage automaticallyusing a voxel classification approach,” IEEE Transactions onMedical Imaging, vol. 26, no. 1, pp. 106–115, 2007.

[29] M.-P. Hellio Le Graverand, R. J. Buck, B. T. Wyman et al.,“Change in regional cartilage morphology and joint spacewidth in osteoarthritis participants versus healthy controls:

Arthritis 15

a multicentre study using 3.0 Tesla MRI and Lyon-Schussradiography,” Annals of the Rheumatic Diseases, vol. 69, no. 1,pp. 155–162, 2010.

[30] F. Eckstein, S. Maschek, W. Wirth et al., “One year change ofknee cartilage morphology in the first release of participantsfrom the Osteoarthritis Initiative progression subcohort:association with sex, body mass index, symptoms andradiographic osteoarthritis status,” Annals of the RheumaticDiseases, vol. 68, no. 5, pp. 674–679, 2009.

[31] F. Eckstein, W. Wirth, M. Hudelmaier et al., “Patterns offemorotibial cartilage loss in knees with neutral, varus, andvalgus alignment,” Arthritis Care and Research, vol. 59, no.11, pp. 1563–1570, 2008.

[32] C. Glaser, R. Burgkart, A. Kutschera, K.-H. Englmeier, M.Reiser, and F. Eckstein, “Femoro-tibial cartilage metrics fromcoronal MR image data: technique, test-retest reproducibil-ity, and findings in osteoarthritis,” Magnetic Resonance inMedicine, vol. 50, no. 6, pp. 1229–1236, 2003.

[33] F. Eckstein, M. Hudelmaier, W. Wirth et al., “Double echosteady state magnetic resonance imaging of knee articularcartilage at 3 Tesla: a pilot study for the OsteoarthritisInitiative,” Annals of the Rheumatic Diseases, vol. 65, no. 4,pp. 433–441, 2006.

[34] M. Hudelmaier, W. Wirth, B. Wehr et al., “Femorotibialcartilage morphology: reproducibility of different metricsand femoral regions, and sensitivity to change in disease,”Cells Tissues Organs, vol. 192, no. 5, pp. 340–350, 2010.

[35] F. Eckstein, O. Benichou, W. Wirth et al., “Magnetic reso-nance imaging-based cartilage loss in painful contralateralknees with and without radiographic joint space narrowing:data from the osteoarthritis initiative,” Arthritis Care andResearch, vol. 61, no. 9, pp. 1218–1225, 2009.

[36] W. Wirth, M. Nevitt, M.-P. Hellio Le Graverand et al., “Sen-sitivity to change of cartilage morphometry using coronalFLASH, sagittal DESS, and coronal MPR DESS protocols—comparative data from the osteoarthritis initiative (OAI),”Osteoarthritis and Cartilage, vol. 18, no. 4, pp. 547–554, 2010.

[37] J. Hohe, G. Ateshian, M. Reiser, K.-H. Englmeier, and F.Eckstein, “Surface size, curvature analysis, and assessmentof knee joint incongruity with MRI in vivo,” MagneticResonance in Medicine, vol. 47, no. 3, pp. 554–561, 2002.

[38] E. B. Dam, J. Folkesson, P. C. Pettersen, and C. Christiansen,“Automatic morphometric cartilage quantification in themedial tibial plateau from MRI for osteoarthritis grading,”Osteoarthritis and Cartilage, vol. 15, no. 7, pp. 808–818, 2007.

[39] J. Folkesson, E. B. Dam, O. F. Olsen, and C. Christiansen,“Accuracy evaluation of automatic quantification of thearticular cartilage surface curvature from MRI,” AcademicRadiology, vol. 14, no. 10, pp. 1221–1228, 2007.

[40] J. Folkesson, E. B. Dam, O. F. Olsen, M. A. Karsdal, P.C. Pettersen, and C. Christiansen, “Automatic quantifica-tion of local and global articular cartilage surface curva-ture: biomarkers for osteoarthritis?” Magnetic Resonance inMedicine, vol. 59, no. 6, pp. 1340–1346, 2008.

[41] A. A. Qazi, E. B. Dam, M. Nielsen, M. A. Karsdal, P. C.Pettersen, and C. Christiansen, “Osteoarthritic cartilage ismore homogeneous than healthy cartilage: identification ofa superior region of interest colocalized with a major riskfactor for osteoarthritis,” Academic Radiology, vol. 14, no. 10,pp. 1209–1220, 2007.

[42] T. J. Mosher and B. J. Dardzinski, “Cartilage MRI T2 relax-ation time mapping: overview and applications,” Seminars inMusculoskeletal Radiology, vol. 8, no. 4, pp. 355–368, 2004.

[43] F. Eckstein, T. Mosher, and D. Hunter, “Imaging of kneeosteoarthritis: data beyond the beauty,” Current Opinion inRheumatology, vol. 19, no. 5, pp. 435–443, 2007.

[44] R. Burgkart, C. Glaser, S. Hinterwimmer et al., “Feasibilityof T and Z scores from magnetic resonance imaging data forquantification of cartilage loss in osteoarthritis,” Arthritis andRheumatism, vol. 48, no. 10, pp. 2829–2835, 2003.

[45] S. C. Faber, F. Eckstein, S. Lukasz et al., “Gender differencesin knee joint cartilage thickness, volume and articular surfaceareas: assessment with quantitative three-dimensional MRimaging,” Skeletal Radiology, vol. 30, no. 3, pp. 144–150,2001.

[46] I. G. Otterness and F. Eckstein, “Women have thinner carti-lage and smaller joint surfaces than men after adjustment forbody height and weight,” Osteoarthritis and Cartilage, vol. 15,no. 6, pp. 666–672, 2007.

[47] Y. Wang, C. Ding, A. E. Wluka et al., “Factors affectingprogression of knee cartilage defects in normal subjects over2 years,” Rheumatology, vol. 45, no. 1, pp. 79–84, 2006.

[48] F. Eckstein, R. J. Buck, D. Burstein et al., “Precision of 3.0Tesla quantitative magnetic resonance imaging of cartilagemorphology in a multicentre clinical trial,” Annals of theRheumatic Diseases, vol. 67, no. 12, pp. 1683–1688, 2008.

[49] F. Eckstein, M. Hudelmaier, S. Cahue, M. Marshall, and L.Sharma, “Medial-to-lateral ratio of tibiofemoral subchondralbone area is adapted to alignment and mechanical load,”Calcified Tissue International, vol. 84, no. 3, pp. 186–194,2009.

[50] Z. A. Cohen, V. C. Mow, J. H. Henry, W. N. Levine, andG. A. Ateshian, “Templates of the cartilage layers of thepatellofemoral joint and their use in the assessment ofosteoarthritic cartilage damage,” Osteoarthritis and Cartilage,vol. 11, no. 8, pp. 569–579, 2003.

[51] J.-P. Raynauld, J. Martel-Pelletier, F. Abram et al., “Analysisof the precision and sensitivity to change of differentapproaches to assess cartilage loss by quantitative MRI in alongitudinal multicentre clinical trial in patients with kneeosteoarthritis,” Arthritis Research and Therapy, vol. 10, no. 6,article no. R129, 2008.

[52] D. J. Hunter, J. Niu, Y. Zhang et al., “Change in cartilagemorphometry: a sample of the progression cohort of theOsteoarthritis Initiative,” Annals of the Rheumatic Diseases,vol. 68, no. 3, pp. 349–356, 2009.

[53] W. Wirth, M.-P. Hellio Le Graverand, B. T. Wyman etal., “Regional analysis of femorotibial cartilage loss in asubsample from the Osteoarthritis Initiative progressionsubcohort,” Osteoarthritis and Cartilage, vol. 17, no. 3, pp.291–297, 2009.

[54] R. J. Buck, B. T. Wyman, M.-P. H. L. Graverand, W.Wirth, and F. Eckstein, “An efficient subset of morphologicalmeasures for articular cartilage in the healthy and diseasedhuman knee,” Magnetic Resonance in Medicine, vol. 63, no. 3,pp. 680–690, 2010.

[55] M.-P. Hellio Le Graverand, R. J. Buck, B. T. Wymanet al., “Subregional femorotibial cartilage morphology inwomen—comparison between healthy controls and partici-pants with different grades of radiographic knee osteoarthri-tis,” Osteoarthritis and Cartilage, vol. 17, no. 9, pp. 1177–1185, 2009.

[56] M. Hudelmaier, C. Glaser, J. Hohe et al., “Age-related changesin the morphology and deformational behavior of knee jointcartilage,” Arthritis and Rheumatism, vol. 44, no. 11, pp.2556–2561, 2001.

16 Arthritis

[57] F. Eckstein, M. Reiser, K.-H. Englmeier, and R. Putz, “In vivomorphometry and functional analysis of human articularcartilage with quantitative magnetic resonance imaging—from image to data, from data to theory,” Anatomy andEmbryology, vol. 203, no. 3, pp. 147–173, 2001.

[58] H. Z. Tameem, L. E. Selva, and U. S. Sinha, “Morphologicalatlases of knee cartilage: Shape indices to analyze cartilagedegradation in osteoarthritic and non-osteoarthritic pop-ulation,” in Proceedings of the 29th Annual InternationalConference of IEEE Engineering in Medicine and BiologySociety (EMBC ’07), pp. 1310–1313, August 2007.

[59] K. A. Beattie, J. Duryea, M. Pui et al., “Minimum joint spacewidth and tibial cartilage morphology in the knees of healthyindividuals: a cross-sectional study,” BMC MusculoskeletalDisorders, vol. 9, article no. 119, 2008.

[60] S. Koo, G. E. Gold, and T. P. Andriacchi, “Considerations inmeasuring cartilage thickness using MRI: factors influencingreproducibility and accuracy,” Osteoarthritis and Cartilage,vol. 13, no. 9, pp. 782–789, 2005.

[61] J.-P. Pelletier, J.-P. Raynauld, M.-J. Berthiaume et al., “Riskfactors associated with the loss of cartilage volume onweight-bearing areas in knee osteoarthritis patients assessedby quantitative magnetic resonance imaging: a longitudinalstudy,” Arthritis Research and Therapy, vol. 9, no. 4, articleno. R74, 2007.

[62] F. Eckstein, M. Yang, A. Guermazi et al., “Referencevalues and Z-scores for subregional femorotibial cartilagethickness—results from a large population-based sample(Framingham) and comparison with the non-exposedOsteoarthritis Initiative reference cohort,” Osteoarthritis andCartilage, vol. 18, no. 10, pp. 1275–1283, 2010.

[63] F. Eckstein, M. Hudelmaier, and R. Putz, “The effects ofexercise on human articular cartilage,” Journal of Anatomy,vol. 208, no. 4, pp. 491–512, 2006.

[64] G. E. Gold, D. Burstein, B. Dardzinski, P. Lang, F. Boada,and T. Mosher, “MRI of articular cartilage in OA: novelpulse sequences and compositional/functional markers,”Osteoarthritis and Cartilage, vol. 14, no. 1, pp. 76–86, 2006.

[65] C. G. Peterfy, G. Gold, F. Eckstein, F. Cicuttini, B. Dardzinski,and R. Stevens, “MRI protocols for whole-organ assessmentof the knee in osteoarthritis,” Osteoarthritis and Cartilage,vol. 14, no. 1, pp. 95–111, 2006.

[66] P. A. Hardy, M. P. Recht, D. Piraino, and D. Thomasson,“Optimization of a Dual Echo in the Steady State (DESS)free-precession sequence for imaging cartilage,” Journal ofMagnetic Resonance Imaging, vol. 6, no. 2, pp. 329–335, 1996.

[67] F. Eckstein, M. Kunz, M. Hudelmaier et al., “Impact ofcoil design on the contrast-to-noise ratio, precision, andconsistency of quantitative cartilage morphometry at 3 Tesla:a pilot study for the osteoarthritis initiative,” MagneticResonance in Medicine, vol. 57, no. 2, pp. 448–454, 2007.

[68] F. Eckstein, M. Kunz, M. Schutzer et al., “Two year lon-gitudinal change and test-retest-precision of knee cartilagemorphology in a pilot study for the osteoarthritis initiative,”Osteoarthritis and Cartilage, vol. 15, no. 11, pp. 1326–1332,2007.

[69] G. H. Welsch, T. C. Mamisch, T. Hughes, S. Domayer,S. Marlovits, and S. Trattnig, “Advanced morphologicaland biochemical magnetic resonance imaging of cartilagerepair procedures in the knee joint at 3 tesla,” Seminars inMusculoskeletal Radiology, vol. 12, no. 3, pp. 196–211, 2008.

[70] C. G. Peterfy, E. Schneider, and M. Nevitt, “The osteoarthritisinitiative: report on the design rationale for the magnetic

resonance imaging protocol for the knee,” Osteoarthritis andCartilage, vol. 16, no. 12, pp. 1433–1441, 2008.

[71] S. Maschek, W. Wirth, B. Wyman et al., “Differences inquantitative MR imaging of cartilage morphology betweensagittal versus coronal acquisitions of the femorotibial joint,”Osteoarthritis and Cartilage, vol. 17, supplement 1, p. S 216,2009, abstract.

[72] D. Inglis, M. Pui, G. Ioannidis et al., “Accuracy and test-retest precision of quantitative cartilage morphology ona 1.0 T peripheral magnetic resonance imaging system,”Osteoarthritis and Cartilage, vol. 15, no. 1, pp. 110–115, 2007.

[73] F. Eckstein, H. C. Charles, R. J. Buck et al., “Accuracy andprecision of quantitative assessment of cartilage morphologyby magnetic resonance imaging at 3.0T,” Arthritis andRheumatism, vol. 52, no. 10, pp. 3132–3136, 2005.

[74] S. R. Morgan, J. C. Waterton, R. A. Maciewicz et al.,“Magnetic resonance imaging measurement of knee cartilagevolume in a multicentre study,” Rheumatology, vol. 43, no. 1,pp. 19–21, 2004.

[75] P. R. Kornaat, S. Koo, T. P. Andriacchi, J. L. Bloem, and G. E.Gold, “Comparison of quantitative cartilage measurementsacquired on two 3.0T MRI systems from different manufac-turers,” Journal of Magnetic Resonance Imaging, vol. 23, no. 5,pp. 770–773, 2006.

[76] M. Hudelmaier, W. Horger, C. Pfau et al., “Cross-calibrationof magnetic resonance sequences and scanners for quan-tifying articular cartilage morphology,” Osteoarthritis andCartilage, vol. 11, supplement A, p. S72, 2003, abstract.

[77] E. Schneider, M. NessAiver, D. White et al., “The osteoarthri-tis initiative (OAI) magnetic resonance imaging qualityassurance methods and results,” Osteoarthritis and Cartilage,vol. 16, no. 9, pp. 994–1004, 2008.

[78] F. Eckstein, R. J. Buck, B. T. Wyman et al., “Quantitativeimaging of cartilage morphology at 3.0 Tesla in the presenceof gadopentate dimeglumine (Gd-DTPA),” Magnetic Reso-nance in Medicine, vol. 58, no. 2, pp. 402–406, 2007.

[79] F. Eckstein, B. T. Wyman, R. J. Buck et al., “Longitudinalquantitative MR imaging of cartilage morphology in thepresence of gadopentetate dimeglumine (gd-dtpa),” MagneticResonance in Medicine, vol. 61, no. 4, pp. 975–980, 2009.

[80] S. J. Gandy, P. A. Dieppe, M. C. Keen, R. A. Maciewicz, I.Watt, and J. C. Waterton, “No loss of cartilage volume overthree years in patients with knee osteoarthritis as assessed bymagnetic resonance imaging,” Osteoarthritis and Cartilage,vol. 10, no. 12, pp. 929–937, 2002.

[81] A. E. Wluka, S. Stuckey, J. Snaddon, and F. M. Cicuttini,“The determinants of change in tibial cartilage volume inosteoarthritic knees,” Arthritis and Rheumatism, vol. 46, no.8, pp. 2065–2072, 2002.

[82] F. M. Cicuttini, A. E. Wluka, Y. Wang, and S. L. Stuckey,“Longitudinal study of changes in tibial and femoral cartilagein knee osteoarthritis,” Arthritis and Rheumatism, vol. 50, no.1, pp. 94–97, 2004.

[83] J.-P. Raynauld, J. Martel-Pelletier, M.-J. Berthiaume et al.,“Quantitative magnetic resonance imaging evaluation ofknee osteoarthritis progression over two years and cor-relation with clinical symptoms and radiologic changes,”Arthritis and Rheumatism, vol. 50, no. 2, pp. 476–487, 2004.

[84] M.-J. Berthiaume, J.-P. Raynauld, J. Martel-Pelletier et al.,“Meniscal tear and extrusion are strongly associated withprogression of symptomatic knee osteoarthritis as assessedby quantitative magnetic resonance imaging,” Annals of theRheumatic Diseases, vol. 64, no. 4, pp. 556–563, 2005.

Arthritis 17

[85] J.-P. Raynauld, J. Martel-Pelletier, M. -J. Berthiaume et al.,“Long term evaluation of disease progression through thequantitative magnetic resonance imaging of symptomaticknee osteoarthritis patients: correlation with clinical symp-toms and radiographic changes,” Arthritis Research andTherapy, vol. 8, no. 1, article R21, 2005.

[86] A. E. Wluka, A. Forbes, Y. Wang, F. Hanna, G. Jones, andF. M. Cicuttini, “Knee cartilage loss in symptomatic kneeosteoarthritis over 4.5 years,” Arthritis Research and Therapy,vol. 8, no. 4, article no. R90, 2006.

[87] O. Bruyere, H. Genant, M. Kothari et al., “Longitudinalstudy of magnetic resonance imaging and standard X-raysto assess disease progression in osteoarthritis,” Osteoarthritisand Cartilage, vol. 15, no. 1, pp. 98–103, 2007.

[88] D. J. Hunter, M. A. Bowes, C. B. Eaton et al., “Can cartilageloss be detected in knee osteoarthritis (OA) patients with 3–6 months’ observation using advanced image analysis of 3TMRI?” Osteoarthritis and Cartilage, vol. 18, no. 5, pp. 677–683, 2010.

[89] J.-P. Raynauld, C. Kauffmann, G. Beaudoin et al., “Reliabilityof a quantification imaging system using magnetic resonanceimages to measure cartilage thickness and volume in humannormal and osteoarthritic knees,” Osteoarthritis and Carti-lage, vol. 11, no. 5, pp. 351–360, 2003.

[90] F. Cicuttini, A. Wluka, Y. Wang, and S. Stuckey, “Thedeterminants of change in patella cartilage volume in oste-oarthritic knees,” Journal of Rheumatology, vol. 29, no. 12, pp.2615–2619, 2002.

[91] R. B. Frobell, M.-P. Le Graverand, R. Buck et al., “The acutelyACL injured knee assessed by MRI: changes in joint fluid,bone marrow lesions, and cartilage during the first year,”Osteoarthritis and Cartilage, vol. 17, no. 2, pp. 161–167, 2009.

[92] P. J. Watson, T. A. Carpenter, L. D. Hall, and J. A. Tyler,“Cartilage swelling and loss in a spontaneous model ofosteoarthritis visualized by magnetic resonance imaging,”Osteoarthritis and Cartilage, vol. 4, no. 3, pp. 197–207, 1996.

[93] E. Calvo, I. Palacios, E. Delgado et al., “High-resolution MRIdetects cartilage swelling at the early stages of experimentalosteoarthritis,” Osteoarthritis and Cartilage, vol. 9, no. 5, pp.463–472, 2001.

[94] E. Calvo, I. Palacios, E. Delgado et al., “Histopathological cor-relation of cartilage swelling detected by magnetic resonanceimaging in early experimental osteoarthritis,” Osteoarthritisand Cartilage, vol. 12, no. 11, pp. 878–886, 2004.

[95] E. Vignon, M. Arlot, and D. Hartmann, “Hypertrophic repairof articular cartilage in experimental osteoarthrosis,” Annalsof the Rheumatic Diseases, vol. 42, no. 1, pp. 82–88, 1983.

[96] M. E. Adams and K. D. Brandt, “Hypertrophic repair ofcanine articular cartilage in osteoarthritis after anteriorcruciate ligament transection,” Journal of Rheumatology, vol.18, no. 3, pp. 428–435, 1991.

[97] R. B. Frobell, M. C. Nevitt, M. Hudelmaier et al., “Femorotib-ial subchondral bone area and regional cartilage thickness—across-sectional description in healthy reference cases and var-ious radiographic stages of osteoarthritis in 1003 knees fromthe Osteoarthritis Initiative,” Arthritis Care Res (Hoboken),vol. 62, no. 11, pp. 1612–1623, 2010.

[98] R. J. Buck, B. T. Wyman, M.-P. Hellio Le Graverand, M.Hudelmaier, W. Wirth, and F. Eckstein, “Osteoarthritis maynot be a one-way-road of cartilage loss—comparison ofspatial patterns of cartilage change between osteoarthriticand healthy knees,” Osteoarthritis and Cartilage, vol. 18, no.3, pp. 329–335, 2010.

[99] F. Eckstein, W. Wirth, D. J. Hunter et al., “Magnitude andregional distribution of cartilage loss associated with gradesof joint space narrowing in radiographic osteoarthritis—datafrom the Osteoarthritis Initiative (OAI),” Osteoarthritis andCartilage, vol. 18, no. 6, pp. 760–768, 2010.

[100] F. Eckstein, S. Muller, S. C. Faber, K.-H. Englmeier, M.Reiser, and R. Putz, “Side differences of knee joint cartilagevolume, thickness, and surface area, and correlation withlower limb dominance—an MRI-based study,” Osteoarthritisand Cartilage, vol. 10, no. 12, pp. 914–921, 2002.

[101] R. D. Altman and G. E. Gold, “Atlas of individual radio-graphic features in osteoarthritis, revised,” Osteoarthritis andCartilage, vol. 15, no. 1, pp. 1–56, 2007.

[102] R. D. Altman, M. Hochberg, W. A. Murphy Jr., F. Wolfe, andM. Lequesne, “Atlas of individual radiographic features inosteoarthritis,” Osteoarthritis and Cartilage, vol. 3, pp. 3–70,1995.

[103] W. Wirth, O. Benichou, C. K. Kwoh et al., “Spatial patterns ofcartilage loss in the medial femoral condyle in osteoarthriticknees: data from the osteoarthritis initiative,” MagneticResonance in Medicine, vol. 63, no. 3, pp. 574–581, 2010.

[104] R. J. Buck, B. T. Wyman, M.-P.H. Le Graverand, M.Hudelmaier, W. Wirth, and F. Eckstein, “Does the use ofordered values of subregional change in cartilage thicknessimprove the detection of disease progression in longitudinalstudies of osteoarthritis?” Arthritis Care and Research, vol. 61,no. 7, pp. 917–924, 2009.

[105] W. Wirth, R. Buck, M. P. Hellio Le Graverand-Gastineau etal., “Ordered values of cartilge loss in knee subregions moreefficiently differentiate progression at different radiographicstages of OA: data from the OA Initiative,” Osteoarthritis andCartilage, vol. 17, supplement 1, p. S203, 2009, abstract.

[106] F. Cicuttini, J. Hankin, G. Jones, and A. Wluka, “Comparisonof conventional standing knee radiographs and magneticresonance imaging in assessing progression of tibiofemoraljoint osteoarthritis,” Osteoarthritis and Cartilage, vol. 13, no.8, pp. 722–727, 2005.

[107] J.-P. Raynauld, J. Martel-Pelletier, P. Bias et al., “Pro-tective effects of licofelone, a 5-lipoxygenase and cyclo-oxygenase inhibitor, versus naproxen on cartilage loss in kneeosteoarthritis: a first multicentre clinical trial using quantita-tive MRI,” Annals of the Rheumatic Diseases, vol. 68, no. 6, pp.938–947, 2009.

[108] C. Peterfy, J. Li, S. Zaim et al., “Comparison of fixed-flexionpositioning with fluoroscopic semi-flexed positioning forquantifying radiographic joint-space width in the knee: test-retest reproducibility,” Skeletal Radiology, vol. 32, no. 3, pp.128–132, 2003.

[109] S. Botha-Scheepers, M. Kloppenburg, H. M. Kroon et al.,“Fixed-flexion knee radiography: the sensitivity to detectknee joint space narrowing in osteoarthritis,” Osteoarthritisand Cartilage, vol. 15, no. 3, pp. 350–353, 2007.

[110] H. C. Charles, V. B. Kraus, M. Ainslie, and M.-P. HellioLe Graverand-Gastineau, “Optimization of the fixed-flexionknee radiograph,” Osteoarthritis and Cartilage, vol. 15, no. 11,pp. 1221–1224, 2007.

[111] M. C. Nevitt, C. Peterfy, A. Guermazi et al., “Longitudi-nal performance evaluation and validation of fixed-flexionradiography of the knee for detection of joint space loss,”Arthritis and Rheumatism, vol. 56, no. 5, pp. 1512–1520,2007.

[112] M.-P. Hellio Le Graverand, E. P. Vignon, K. D. Brandt et al.,“Head-to-head comparison of the Lyon Schuss and fixed

18 Arthritis

flexion radiographic techniques. Long-term reproducibilityin normal knees and sensitivity to change in osteoarthriticknees,” Annals of the Rheumatic Diseases, vol. 67, no. 11, pp.1562–1566, 2008.

[113] J. G. Adams, T. McAlindon, M. Dimasi, J. Carey, and S.Eustace, “Contribution of meniscal extrusion and cartilageloss to joint space narrowing in osteoarthritis,” ClinicalRadiology, vol. 54, no. 8, pp. 502–506, 1999.

[114] D. R. Gale, C. E. Chaisson, S. M. S. Totterman, R. K. Schwartz,M. E. Gale, and D. Felson, “Meniscal subluxation: associationwith osteoarthritis and joint space narrowing,” Osteoarthritisand Cartilage, vol. 7, no. 6, pp. 526–532, 1999.

[115] D. J. Hunter, Y. Q. Zhang, X. Tu et al., “Change in jointspace width: hyaline articular cartilage loss or alteration inmeniscus?” Arthritis and Rheumatism, vol. 54, no. 8, pp.2488–2495, 2006.

[116] J. Duryea, G. Neumann, J. Niu et al., “Comparison ofradiographic joint space width with magnetic resonanceimaging cartilage morphometry: analysis of longitudinaldata from the Osteoarthritis Initiative,” Arthritis Care andResearch, vol. 62, no. 7, pp. 932–937, 2010.

[117] J. Duryea, J. Li, C. G. Peterfy, C. Gordon, and H. K. Genant,“Trainable rule-based algorithm for the measurement ofjoint space width in digital radiographic images of the knee,”Medical Physics, vol. 27, no. 3, pp. 580–591, 2000.

[118] G. Neumann, D. Hunter, M. Nevitt et al., “Location specificradiographic joint space width for osteoarthritis progres-sion,” Osteoarthritis and Cartilage, vol. 17, no. 6, pp. 761–765,2009.

[119] J. N. Belo, M. Y. Berger, M. Reijman, B. W. Koes, and S.M. A. Bierma-Zeinstra, “Prognostic factors of progressionof osteoarthritis of the knee: a systematic review of observa-tional studies,” Arthritis Care and Research, vol. 57, no. 1, pp.13–26, 2007.

[120] F. Eckstein, W. Wirth, M. I. Hudelmaier et al., “Relationshipof compartment-specific structural knee status at baselinewith change in cartilage morphology: a prospective obser-vational study using data from the osteoarthritis initiative,”Arthritis Research and Therapy, vol. 11, no. 3, article no. R90,2009.

[121] M. L. Davies-Tuck, J. Martel-Pelletier, A. E. Wluka et al.,“Meniscal tear and increased tibial plateau bone area inhealthy post-menopausal women,” Osteoarthritis and Carti-lage, vol. 16, no. 2, pp. 268–271, 2008.

[122] A. Chang, K. Moisio, J. S. Chmiel et al., “Subregional effectsof meniscal tears on cartilage loss over 2 years in kneeosteoarthritis,” Annals of the Rheumatic Diseases, 2010.

[123] M. D. Crema, A. Guermazi, L. Li et al., “The association ofprevalent medial meniscal pathology with cartilage loss inthe medial tibiofemoral compartment over a 2-year period,”Osteoarthritis and Cartilage, vol. 18, no. 3, pp. 336–343, 2010.

[124] K. A. Jung, S. C. Lee, S. H. Hwang et al., “High frequency ofmeniscal hypertrophy in persons with advanced varus kneeosteoarthritis,” Rheumatology International, vol. 30, no. 10,pp. 1325–1333, 2010.

[125] W. Wirth, R. B. Frobell, R. B. Souza et al., “A three-dimensional quantitative method to measure meniscusshape, position, and signal intensity using MR images: apilot study and preliminary results in knee osteoarthritis,”Magnetic Resonance in Medicine, vol. 63, no. 5, pp. 1162–1171, 2010.

[126] J.-P. Raynauld, J. Martel-Pelletier, M.-J. Berthiaume et al.,“Correlation between bone lesion changes and cartilage

volume loss in patients with osteoarthritis of the knee asassessed by quantitative magnetic resonance imaging over a24-month period,” Annals of the Rheumatic Diseases, vol. 67,no. 5, pp. 683–688, 2008.

[127] F. Cicuttini, A. Wluka, J. Hankin, and Y. Wang, “Lon-gitudinal study of the relationship between knee angleand tibiofemoral cartilage volume in subjects with kneeosteoarthritis,” Rheumatology, vol. 43, no. 3, pp. 321–324,2004.

[128] R. Von Eisenhart-Rothe, H. Graichen, M. Hudelmaier, T.Vogl, L. Sharma, and F. Eckstein, “Femorotibial and patellarcartilage loss in patients prior to total knee arthroplasty,heterogeneity, and correlation with alignment of the knee,”Annals of the Rheumatic Diseases, vol. 65, no. 1, pp. 69–73,2006.

[129] A. J. Teichtahl, M. L. Davies-Tuck, A. E. Wluka, G. Jones,and F. M. Cicuttini, “Change in knee angle influences the rateof medial tibial cartilage volume loss in knee osteoarthritis,”Osteoarthritis and Cartilage, vol. 17, no. 1, pp. 8–11, 2009.

[130] A. J. Teichtahl, A. E. Wluka, and F. M. Cicuttini, “Frontalplane knee alignment is associated with a longitudinalreduction in patella cartilage volume in people with kneeosteoarthritis,” Osteoarthritis and Cartilage, vol. 16, no. 7, pp.851–854, 2008.

[131] G. Zhai, C. Ding, F. Cicuttini, and G. Jones, “A longitudinalstudy of the association between knee alignment and changein cartilage volume and chondral defects in a largely non-osteoarthritic population,” Journal of Rheumatology, vol. 34,no. 1, pp. 181–186, 2007.

[132] B. Vanwanseele, F. Eckstein, R. M. Smith et al., “Therelationship between knee adduction moment and cartilageand meniscus morphology in women with osteoarthritis,”Osteoarthritis and Cartilage, vol. 18, no. 7, pp. 894–901, 2010.

[133] F. M. Cicuttini, A. Wluka, M. Bailey et al., “Factors affectingknee cartilage volume in healthy men,” Rheumatology, vol.42, no. 2, pp. 258–262, 2003.

[134] A. J. Teichtahl, A. E. Wluka, Y. Wang et al., “Obesity and adi-posity are associated with the rate of patella cartilage volumeloss over 2 years in adults without knee osteoarthritis,” Annalsof the Rheumatic Diseases, vol. 68, no. 6, pp. 909–913, 2009.

[135] A. E. Wluka, F. Hanna, M. Davies-Tuck et al., “Bone marrowlesions predict increase in knee cartilage defects and loss ofcartilage volume in middle-aged women without knee painover 2 years,” Annals of the Rheumatic Diseases, vol. 68, no. 6,pp. 850–855, 2009.

[136] A. E. Wluka, Y. Wang, M. Davies-Tuck, D. R. English, G.G. Giles, and F. M. Cicuttini, “Bone marrow lesions predictprogression of cartilage defects and loss of cartilage volumein healthy middle-aged adults without knee pain over 2 yrs,”Rheumatology, vol. 47, no. 9, pp. 1392–1396, 2008.

[137] C. Ding, F. Cicuttini, F. Scott, C. Boon, and G. Jones,“Association of prevalent and incident knee cartilage defectswith loss of tibial and patellar cartilage: a longitudinal study,”Arthritis and Rheumatism, vol. 52, no. 12, pp. 3918–3927,2005.

[138] A. E. Wluka, C. Ding, G. Jones, and F. M. Cicuttini, “The clin-ical correlates of articular cartilage defects in symptomaticknee osteoarthritis: a prospective study,” Rheumatology, vol.44, no. 10, pp. 1311–1316, 2005.

[139] F. Cicuttini, C. Ding, A. Wluka, S. Davis, P. R. Ebeling, andG. Jones, “Association of cartilage defects with loss of kneecartilage in healthy, middle-age adults: a prospective study,”Arthritis and Rheumatism, vol. 52, no. 7, pp. 2033–2039,2005.

Arthritis 19

[140] C. Ding, J. Martel-Pelletier, J.-P. Pelletier et al., “Two-year prospective longitudinal study exploring the factorsassociated with change in femoral cartilage volume in acohort largely without knee radiographic osteoarthritis,”Osteoarthritis and Cartilage, vol. 16, no. 4, pp. 443–449, 2008.

[141] C. Ding, F. Cicuttini, and G. Jones, “Tibial subchondralbone size and knee cartilage defects: relevance to kneeosteoarthritis,” Osteoarthritis and Cartilage, vol. 15, no. 5, pp.479–486, 2007.

[142] R. B. Frobell, W. Wirth, M. Nevitt et al., “Presence, location,type and size of denuded areas of subchondral bone in theknee as a function of radiographic stage of OA—data fromthe OA initiative,” Osteoarthritis and Cartilage, vol. 18, no. 5,pp. 668–676, 2010.

[143] D. J. Hunter, L. Li, Y. Q. Zhang et al., “Region of interest anal-ysis: by selecting regions with denuded areas can we detectgreater amounts of change?” Osteoarthritis and Cartilage, vol.18, no. 2, pp. 175–183, 2010.

[144] E. B. Dam, I. Byrjalsen, M. A. Karsdal, P. Qvist, and C.Christiansen, “Increased urinary excretion of C-telopeptidesof type II collagen (CTX-II) predicts cartilage loss over 21months by MRI,” Osteoarthritis and Cartilage, vol. 17, no. 3,pp. 384–389, 2009.

[145] O. Bruyere, J. Collette, M. Kothari et al., “Osteoarthritis,magnetic resonance imaging, and biochemical markers: aone year prospective study,” Annals of the Rheumatic Diseases,vol. 65, no. 8, pp. 1050–1054, 2006.

[146] J. P. Pelletier, J. P. Raynauld, J. Caron et al., “Decrease inserum level of matrix metalloproteinases is predictive of thedisease-modifying effect of osteoarthritis drugs assessed byquantitative MRI in patients with knee osteoarthritis,” Annalsof the Rheumatic Diseases, vol. 69, no. 12, pp. 2095–2101,2010.

[147] F. Eckstein, M. P. Hellio-Le Graverand, C. Charles etal., “Which baseline factors predict cartilge thinning andthickening in OA knees—results from the A9001140 study,”Osteoarthritis and Cartilage, vol. 17, supplement 1, p. S215,2009, Abstract No 407.

[148] C. Ding, F. Cicuttini, L. Blizzard, and G. Jones, “Smokinginteracts with family history with regard to change in kneecartilage volume and cartilage defect development,” Arthritisand Rheumatism, vol. 56, no. 5, pp. 1521–1528, 2007.

[149] M. L. Davies-Tuck, A. E. Wluka, A. Forbes et al., “Smokingis associated with increased cartilage loss and persistenceof bone marrow lesions over 2 years in community-basedindividuals,” Rheumatology, vol. 48, no. 10, pp. 1227–1231,2009.

[150] F. M. Cicuttini, G. Jones, A. Forbes, and A. E. Wluka, “Rateof cartilage loss at two years predicts subsequent total kneearthroplasty: a prospective study,” Annals of the RheumaticDiseases, vol. 63, no. 9, pp. 1124–1127, 2004.

[151] S. K. Tanamas, A. E. Wluka, J. -P. Pelletier et al., “Theassociation between subchondral bone cysts and tibial car-tilage volume and risk of joint replacement in people withknee osteoarthritis: a longitudinal study,” Arthritis Research& Therapy, vol. 12, article R58, 2010.

[152] G. A. Hawker, L. Stewart, M. R. French et al., “Understandingthe pain experience in hip and knee osteoarthritis—anOARSI/OMERACT initiative,” Osteoarthritis and Cartilage,vol. 16, no. 4, pp. 415–422, 2008.

Hindawi Publishing CorporationArthritisVolume 2011, Article ID 854651, 6 pagesdoi:10.1155/2011/854651

Clinical Study

Patella Eversion Reduces Early Knee Range of Motionand Muscle Torque Recovery after Total Knee Arthroplasty:Comparison between Minimally Invasive Total Knee Arthroplastyand Conventional Total Knee Arthroplasty

Tokifumi Majima,1 Osamu Nishiike,2 Naohiro Sawaguchi,2 Kouichi Susuda,3

and Akio Minami2

1 Department of Joint Replacement and Tissue Engineering, Graduate School of Medicine, Hokkaido University, N-15, W-7, Kita-Ku,Sapporo 060-8638, Japan

2 Department of Orthopaedic Surgery, Graduate School of Medicine, Hokkaido University, Sapporo 060-8638, Japan3 Shin-Sapporo Orthopaedic Hospital, Sapporo 004-0022, Japan

Correspondence should be addressed to Tokifumi Majima, tkmajima@med.hokudai.ac.jp

Received 9 August 2010; Revised 26 October 2010; Accepted 30 November 2010

Academic Editor: Changhai Ding

Copyright © 2011 Tokifumi Majima et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

We hypothesized that patella eversion during total knee arthroplasty (TKA) reduces early return of active knee extension andflexion, quadriceps muscle strength, and postoperative pain. In 100 conventional TKA knees and 100 minimally invasive TKA(MIS TKA) knees, we compared knee range of motion (ROM), postoperative pain, and quadriceps muscle strength at 1 day, 4days, 1 week, 2 weeks, 3 weeks, 4 weeks, 12 weeks, 1 year, and 5 years after surgery. The differences of surgical approach betweenMIS TKA and conventional TKA of this study are length of skin incision with subcutaneal flap and patella eversion. In MISTKA, skin incision is shorter than conventional TKA. Furthermore, patella is not everted in MIS TKA procedure. There were nosignificant differences in preoperative factors. Postoperative improvement of ROM, postoperative muscle strength recovery, andpostoperative improvement of visual analog scale were faster in patients with MIS TKA when compared to that in patients withconventional TKA. On the other hand, no significant difference was observed in complication, 5-year clinical results of subjectiveknee function score, and the postoperative component angle and lower leg alignment. These results indicate that patella eversionmay affect muscle strength recovery and postoperative pain.

1. Introduction

Total knee arthroplasty (TKA) is an established treatmentfor advanced arthritis of the knee [1, 2]. At the time ofsurgery, medial parapatellar approach, midvastus approach,or subvastus approach is used. The most popular approach ofTKA is a medial parapatellar approach that splits quadricepstendon with eversion of the patella [3]. Having operatedwithout disruption of the extensor mechanism subvastusapproach theoretically preserves the quadriceps musclesfunction. It has been reported that subvastus approach inTKA could regain quadriceps muscle strength faster than

the medial parapatellar approach [4–6]. On the other hand,it is reported that performing a knee arthroplasty througha subvastus approach is difficult in patients with limitedpreoperative knee flexion or in obese patients [7]. Therefore,some surgeons use the midvastus muscle-splitting approach.With the use of this approach, complete eversion and lateraldisplacement of the patella are not inhibited, allowing foradequate exposure to the joint [8]. In these three approaches,surgeons need an approximately 20 cm length incision ofthe skin and patella eversion. That may affect length ofhospitalization, medical cost, postoperative pain, and lengthof rehabilitation.

2 Arthritis

Minimally invasive surgical techniques have been used inmany types of surgical procedures, both arthroscopic surgeryand open surgery. Minimally invasive surgery (MIS) asapplied to TKA has been developed to improve pain, periodof time consumed for functional recovery, blood loss, andhospitalization. Previous reports indicated the advantages ofMIS TKA according to 2-year short-term clinical results [9–17]. In the previous study, however, no mid-term more than5-year clinical results have been reported. Furthermore, noreport explained the mechanism of quick pain relief andmuscle strength recovery in MIS TKA.

In this paper, we hypothesized that patella eversion willreduce early return of active knee extension and flexion,quadriceps strength, and postoperative pain. To test thishypothesis, we compared knee range of motion (ROM),postoperative pain, and quadriceps muscle strength afterTKA in conventional TKA and MIS TKA. The purpose ofthis study was to evaluate the knee ROM, postoperative pain,and quadriceps muscle strength within one year after surgery.Furthermore, mid-term at least 5-year clinical results of theMIS TKA and conventional TKA were compared.

2. Materials and Methods

Between April 2004 and August 2005, a prospective compara-tive study was conducted in two hospitals. 200 primary TKAsof 180 patients with medial compartmental osteoarthriticknee were enrolled. Conventional TKA was performed in100 knees at one hospital. MIS-TKA was performed in 100knees with age and gender matched to conventional TKAgroup at the other hospital. All of the patients were evaluated.Follow-up period ranged from 60 to 76 months (average:66 months). Age at surgery ranged from 60 to 82 years old(average: 70.3 years old). Average preoperative Hospital forSpecial Surgery scoring system (HSS score) [18] was 46.7 ±8.5 points in the MIS group and 48.6 ± 7.6 points in theconventional group. No patients received previous surgeryon the affected knees. No patients had flexion contracturemore than 30 degrees. General anesthesia with epiduraltubing was made by the anesthetist. All of the surgery wasperformed by a single surgeon with subvastus approachusing Scorpio NRG PS (Stryker, Mahwah, NJ). The patellawas resurfaced in all patients. All of the components werefixed with bone cement.

For the MIS group, after the tourniquet was inflated,a curved medial skin incision extending from the superiorpole of the patella to the top of the tibial tubercle was made.Medial arthrotomy was made from the superior pole of thepatella to just above the insertion of pes anserinus, andthen subperiosteal medial soft tissue release was performedas standard procedure. The patella was displaced laterallywithout eversion, and first resection of the patella wasperformed to increase working space. The tibial resectionwas then performed from the anteromedial side with anextramedullary guide. The distal femur was also resectedusing an intramedullary guide. The rotation of femoralcomponent was set parallel to the tibial cut surface underindividual 40 N load of medial and lateral space using

Figure 1: Subvastus arthrotomy with implants is shown frommedial side. The vastus medialis oblique muscle and its aponeurosisremain intact and fully attached to the patella.

ligament tension with knee flexed at 90 degrees (EndoPlus,Marl, Germany). Further resection of the patella was doneto match the thickness of patellar components. Figure 1shows subvastus arthrotomy with implant in the MIS TKAgroup. In the standard group, the surgery was done througha midline incision of about 13 to 18 cm and a medialparapatellar arthrotomy; the patella was everted laterallywhen the knee was flexed. Tourniquet time was measured inall patients. A suction drain and epidural tube were left for 24hours in all patients. Postoperative pain management was thesame for both groups, which included continuous epiduralinjection of 2.5 ml/hr of bupivacaine hydrochloride hydrate(Marcain, AstraZeneca, Osaka, Japan) and nonsteroidal anti-inflammatory drugs for 2 weeks. Both groups went throughthe same clinical pathway with continuous passive motionand bedside physiotherapy started on the second day aftersurgery.

Total blood loss and possible time of straight leg raisewere measured. Visual analogue scale (VAS) score of painwas scheduled to evaluate at day 1, day 4, 1 week, 2 weeks,3 weeks, and 4 weeks after surgery in every patient. Whenresponding to a VAS item, patients specify their level of painto a statement by indicating a position along a continuoushorizontal line (100 mm in length) between two endpoints.The VAS score is determined by measuring in millimeters.In VAS score, 0 mm means no pain, and 100 mm meansvery severe pain. Passive knee ROM was measured usinga goniometer. Knee extension muscle torque was recordedat 1 week, 2 weeks, 3 weeks, 4 weeks, 3 months, 12months, and the final visit. Peak isokinetic torque of theipsilateral and contralateral quadriceps muscle was measuredat 60 degrees/second of angular velocity using the Cybex IIdynamometer (Lumex Inc, Ronkonkoma, NY). This musclecontraction involved knee extension. The torque measuredin the untreated knee at the preoperative examination wasdefined as the reference value in each patient, and torque

Arthritis 3

Table 1: Patient demographics.

Conventional TKA MIS-TKA P value

Age 69.2± 8.1 71.8± 7.7 N.S

Sex, F/M 82/18 79/21 N.S

Height (cm) 154.1± 6.0 151.5± 5.2 N.S

Weight (kg) 60.8± 6.1 62.2± 5.8 N.S

Preoperative kneeextension (degree)

−10± 8.5 −8.5± 10.1 N.S

Preoperative kneeflexion (degree)

115± 10.5 120± 11.4 N.S

Preoperative mechanicalaxis in one leg standingradiograph (degree)

190.8± 8.9 188.6± 10.6 N.S

Percentage ofpreoperative peakisokinetic torque of thequadriceps muscle (%)∗

82± 14 85± 11 N.S

Preoperative visualanalogue scale of pain(mm)

75± 11 72± 13 N.S

Preoperative knee score 48.6± 7.6 46.7± 8.5 N.S∗

preoperative peak value in uninvolved knee was defined as a reference value(100%).

measured at each postoperative period was represented asa ratio (percentage) of the reference value. With regardto radiology, anteroposterior (AP), lateral, and long-legweight-bearing views were obtained preoperatively and at thefinal followup. All radiographs were assessed independentlyby two observers except for evaluation by surgeons withregard to implant position according to the Knee SocietyTKA roentgenographic evaluation form [19] and mechanicalaxis of lower leg allowing for an unbiased radiographicassessment.

Statistically, the 2 groups were compared using Studentis t-test for continuous data. For muscle strength, change inROM, and change in VAS a two-way analysis of variance witha Bonferroni correction for multiple comparisons was usedto analyze the change in time and difference between groups.A P value less than .05 was considered significant.

3. Results

The detailed preoperative comparison between the 2 groupsis shown in Table 1. There were no significant differences inpreoperative factors.

There was no significant difference in the meantourniquet time (MIS group: 98.9 ± 18.2 minutes;conventional group: 88.6 ± 14.6 minutes). On the otherhand, tourniquet time of the early 50 cases was significantlylonger than that in the latest 50 cases in MIS-TKA (P < .004)(Figure 2). The mean total blood loss also showed nodifference between the MIS group (548 ± 348 ml) and theconventional group (631 ± 335 ml). No lateral retinacularrelease was required in both of the MIS group and theconventional group. Active straight-leg raise was achievedquicker (P < .05) in the MIS group (1.02 ± 0.14 days)

Table 2: Clinical results.

MIS-TKA Conventional TKA P value

Mean tourniquet time(min)

98.9± 18.2 88.6± 14.6 N.S

Total blood loss (ml) 548± 348 631± 335 N.S

Possible SLR (day) 1.02± 0.14 2.13± 2.30 P < .05

HSS knee score 89.0± 7.5 87.3± 7.1 N.S

Knee extension (degree) −2.0± 3.1 −3.0± 4.1 N.S

Knee flexion (degree) 130± 8.1 126± 6.5 N.S

SLR: straight leg raise.

R2 = 0.2154

0

20

40

60

80

100

120

140

160

0 20 40 60 80 100

Tim

e(m

in)

Case number

Figure 2: Tourniquet time of 100 cases in MIS-TKA group isshown. Learning curve was observed.

than in the conventional group (2.13 ± 2.30 days). Therewere no intraoperative complications in both of the groups.Postoperatively, 5 cases in the MIS group and 2 cases in theconventional group had minor wound complications. Therewere no deep or superficial infections in either group. Therewas no symptomatic deep venous thrombosis in all cases.There were no perioperative fatal complications in eithergroup. HSS knee score improved to 89.0 ± 7.5 in the MISgroup and 87.3± 7.1 in the conventional group (Table 2).

Average postoperative pain between day 4 and 3 weeksafter surgery, as recorded on the VAS, was significantly lowerin the MIS group (Figure 3). Until 4 weeks after surgery,average knee ROM was larger in the MIS group than in theconventional group (Figure 4). The changes of quadricepsmuscle strength in both groups are shown in Figure 5. Peakisometric torque of the quadriceps muscle in the MIS grouprecovered more quickly than the conventional group. Threemonths after surgery, there were no significant differences inknee ROM and quadriceps muscle strength.

There were no significant differences in radiologicalmeasurements. At the final followup, mechanical axis was180.4 ± 2.8◦ in the MIS group and 180.9 ± 2.5◦ in theconventional group. Cases of overall lower leg alignmentof neutral ±3◦ were 79% and 81% in the MIS group andthe conventional group, respectively. Angles of componentalpha, beta, gamma, and delta are 94.7 ± 3.6, 88.9 ± 2.0,90.6 ± 2.7, and 90.3± 2.6◦ in the MIS group and 96.5± 1.9,

4 Arthritis

0102030405060708090

100

Pre op

MIS groupConventional group

∗∗: P < .01

∗: P < .05∗∗

∗∗

∗∗

∗

( mm

)

Time after surgery

Val

ue

ofvi

s ual

anal

o gu

esc

ale

p/o 1day

p/o 4days

p/o 1week

p/o 2weeks

p/o 3weeks

p/o 4weeks

Figure 3: Change of the visual analogue scale (mm) of pain in bothgroups until 4 weeks after surgery.

89.7 ± 2.3, 92.3 ± 2.1, and 91.1 ± 2.3◦ in the conventionalgroup, respectively.

4. Discussion

In this paper, we hypothesized that patella eversion willdeteriorate early return of knee range of extension andflexion, quadriceps strength, and postoperative pain. Thispaper showed that postoperative improvement of ROM,postoperative muscle strength recovery, and postoperativeimprovement of pain were greater in patients with MISTKA when compared to patients with conventional TKA.On the other hand, no significant difference was observedin complication, 5-year mid-term results of knee functionscore, or the postoperative component angle and lower legalignment. The strength of the study lies in that we firstlyexplained the mechanism of pain relief and muscle strengthrecovery in MIS TKA.

The differences of surgical approach between MIS TKAand conventional TKA of this study are length of skinincision with subcutaneal flap and patella eversion. It hasbeen reported that knee pain with muscle contraction playeda small role in the reduction of muscle activation [20].Concerning the peak torque of quadriceps muscle, tractionforce to muscle by patellar eversion during surgery mayaffect the recovery of muscle strength. As for postoperativepain, differences in visual analogue scale of pain exist afterwound healing. These results indicate that patella eversionmay affect postoperative muscle torque recovery and pain.However, there is possibility that the reduced length of theincision might have contributed to the results. To assess thispossibility, we are planning to compare the result of TKA inthe same skin incision with patellar eversion and that withoutpatellar eversion for future research.

In the present study, the risks and benefits of MISTKA and conventional TKA were also compared. Theresults showed that MIS-TKA is better in postoperative earlyrecovery of muscle strength, while concerning the compo-nent placement and lower leg alignment, no deterioration

0

20

40

60

80

100

Pre op

MIS groupConventional group

∗∗: P < .01

∗: P < .05

∗∗∗

∗ ∗

∗∗

Time after surgery

120

140

Flex

ion

and

exte

nsi

onan

gle

(deg

rees

)

−201

week2

weeks3

weeks4

weeks weeks12 12

monthsFinal

Figure 4: Change in the knee range of motion in both groups until4 weeks after surgery. Higher range data set represents knee flexionangle. Lower range data set represents knee extension angle.

Pre op

MIS groupConventional group

∗∗: P < .01

∗: P < .05

∗∗∗∗

∗∗ ∗

Time after surgery

1week

2weeks

3weeks

4weeks weeks

12 12months

Final

Rel

ativ

eva

lue

ofp

eak

torq

ue

ofqu

adri

ceps

mu

scle

(%)

0

20

40

60

80

100

120

Figure 5: Change in percentage of peak isokinetic torque of thequadriceps muscle in the MIS group and the conventional groupis shown. Preoperative peak value in untreated knee was defined asa reference value (100%).

occurred. These results are consistent with previous reports[9–17, 21, 22]. However, the MIS-TKA through a limitedfield has difficulties. The higher rate of complications such asmalalignments, fixation errors, soft tissue damage, increasedoperating time, and blood loss were also noted [10–13, 15,23, 24]. Knowing these complications at the time of MISsurgery, we meticulously checked the position of instrumentand cut surface to avoid outliers. This effort may result in nocase of outliers in alignment, bleeding, and wound healing.

Benefits from an earlier recovery with MIS TKA lead toearlier discharge, reduced physiotherapy, and reduced dosageof pills for pain control. These benefits are convenient forpatients. Furthermore, the benefits may reduce medical cost.On the other hand, there were no differences in functionand pain 12 weeks after surgery. Three months after TKA,functional recovery and wound healing involving muscle,

Arthritis 5

ligament, and joint capsule may reach plateau whether MISsurgery was performed or not.

Operating through the subvastus approach in MIS TKAwas technically demanding for access and overall visibility.Consequently, the tourniquet time was increased by anaverage of 10 minutes in the initial 50 cases of MIS TKA.On the other hand, this did not influence the incidenceof deep venous thrombosis. Despite reduced access insurgical approach, component geometry and leg alignmentdistributed in acceptable range. We, as a surgical team, havegone through a learning curve and have not experiencedintraoperative complications.

Poor visualization of the knee joint may leads to wrongbone cutting that lead to poor clinical results [25, 26].Concerning implications of the results, surgeons may extendskin incision without patella eversion when exposure of theknee joint is not enough to visualize the area of bone cuttingand soft tissue release.

A limitation of this study was that the patients werenot assigned randomly at the same hospital. This fact hasa possibility of confounding variables. Another limitationwas that 5-year mid-term clinical data with MIS TKA canprovide limited information concerning their efficacy andsafety. Long-term follow-up studies are necessary.

In conclusion, MIS TKA is better in early functionalrecovery of the knee compared to conventional TKA. Thiseffectiveness may owe to the approach without patellaeversion.

References

[1] C. S. Ranawat, W. F. Flynn, S. Saddler, K. K. Hansraj, andM. J. Maynard, “Long-term results of the total condylarknee arthroplasty: a 15-year survivorship study,” ClinicalOrthopaedics and Related Research, no. 286, pp. 94–102,1993.

[2] M. A. Kelly and H. D. Clarke, “Long-term results of pos-terior cruciate-substituting total knee arthroplasty,” ClinicalOrthopaedics and Related Research, no. 404, pp. 51–57, 2002.

[3] J. Insall, “A midline approach to the knee,” Journal of Bone andJoint Surgery A, vol. 53, no. 8, pp. 1584–1586, 1971.

[4] A. Berth, D. Urbach, W. Neumann, and F. Awiszus, “Strengthand voluntary activation of quadriceps femoris muscle in totalknee arthroplasty with midvastus and subvastus approaches,”Journal of Arthroplasty, vol. 22, no. 1, pp. 83–88, 2007.

[5] G. S. Roysam and M. J. Oakley, “Subvastus approach for totalknee arthroplasty: a prospective, randomized, and observer-blinded trial,” Journal of Arthroplasty, vol. 16, no. 4, pp. 454–457, 2001.

[6] E. Cila, V. Guzel, M. Ozalay et al., “Subvastus versus medialparapatellar approach in total knee arthroplasty,” Archives ofOrthopaedic and Trauma Surgery, vol. 122, no. 2, pp. 65–68,2002.

[7] S. Knezevich, G. A. Engh, F. E. Preidis et al., “Compari-son of subvastus quadriceps-sparing and standard anteriorquadriceps-splitting approaches in total and unicompartmen-tal knee arthroplasty,” Orthopaedic Transactions, vol. 16, article615, 1992.

[8] G. A. Engh, B. T. Holt, and N. L. Parks, “A midvastus muscle-splitting approach for total knee arthroplasty,” Journal ofArthroplasty, vol. 12, no. 3, pp. 322–331, 1997.

[9] M. Tenholder, H. D. Clarke, and G. R. Scuderi, “Minimal-incision total knee arthroplasty: the early clinical experience,”Clinical Orthopaedics and Related Research, no. 440, pp. 67–76,2005.

[10] T. O. Boerger, P. Aglietti, N. Mondanelli, and L. Sensi,“Mini-subvastus versus medial parapatellar approach in totalknee arthroplasty,” Clinical Orthopaedics and Related Research,no. 440, pp. 82–87, 2005.

[11] D. F. Dalury and D. A. Dennis, “Mini-incision total kneearthroplasty can increase risk of component malalignment,”Clinical Orthopaedics and Related Research, no. 440, pp. 77–81,2005.

[12] R. S. Laskin, “Minimally invasive total knee replacement usinga mini-mid vastus incision technique and results,” SurgicalTechnology International, vol. 13, pp. 231–238, 2004.

[13] A. J. Tria Jr. and T. M. Coon, “Minimal incision totalknee arthroplasty early experience,” Clinical Orthopaedics andRelated Research, no. 416, pp. 185–190, 2003.

[14] P. M. Bonutti, D. J. Neal, and M. A. Kester, “Minimal incisiontotal knee arthroplasty using the suspended leg technique,”Orthopedics, vol. 26, no. 9, pp. 899–903, 2003.

[15] S. B. Haas, S. Cook, and B. Beksac, “Minimally invasive totalknee replacement through a mini midvastus approach: a com-parative study,” Clinical Orthopaedics and Related Research,no. 428, pp. 68–73, 2004.

[16] R. A. Berger, S. Sanders, T. Gerlinger, C. Della Valle, J. J. Jacobs,and A. G. Rosenberg, “Outpatient total knee arthroplastywith a minimally invasive technique,” Journal of Arthroplasty,vol. 20, no. 7, supplement 3, pp. 33–38, 2005.

[17] M. W. Pagnano and R. M. Meneghini, “Minimally inva-sive total knee arthroplasty with an optimized subvastusapproach,” Journal of Arthroplasty, vol. 21, no. 4, supplement1, pp. 22–26, 2006.

[18] C. S. Ranawat, J. Insall, and J. Shine, “Duo condylar kneearthroplasty. Hospital for Special Surgery design,” ClinicalOrthopaedics and Related Research, vol. 120, pp. 76–82,1976.

[19] F. C. Ewald, “The Knee Society total knee arthroplastyroentgenographic evaluation and scoring system,” Clini-cal Orthopaedics and Related Research, no. 248, pp. 9–12,1989.

[20] R. L. Mizner, S. C. Petterson, J. E. Stevens, K. Vandenborne,and L. Snyder-Mackler, “Early quadriceps strength loss aftertotal knee arthroplasty: the contributions of muscle atrophyand failure of voluntary muscle activation,” Journal of Boneand Joint Surgery A, vol. 87, no. 5, pp. 1047–1053, 2005.

[21] P. M. Bonutti, M. A. Mont, and M. A. Kester, “Minimallyinvasive total knee arthroplasty: a 10-feature evolutionaryapproach,” Orthopedic Clinics of North America, vol. 35, no. 2,pp. 217–226, 2004.

[22] A. J. Tria Jr., “Minimally invasive total knee arthroplasty: theimportance of instrumentation,” Orthopedic Clinics of NorthAmerica, vol. 35, no. 2, pp. 227–234, 2004.

[23] R. S. Laskin, B. Beksac, A. Phongjunakorn et al., “Minimallyinvasive total knee replacement through a mini-midvastusincision: an outcome study,” Clinical Orthopaedics and RelatedResearch, no. 428, pp. 74–81, 2004.

[24] A. F. Chen, R. K. Alan, D. E. Redziniak, and A. J. Tria, “Quadri-ceps sparing total knee replacement. The initial experience

6 Arthritis

with results at two to four years,” Journal of Bone and JointSurgery B, vol. 88, no. 11, pp. 1448–1453, 2006.

[25] P. Aglietti and R. Buzzi, “Posteriorly stabilised total-condylarknee replacement. Three to eight years’ follow-up of 85 knees,”Journal of Bone and Joint Surgery B, vol. 70, no. 2, pp. 211–216,1988.

[26] R. S. Jeffery, R. W. Morris, and R. A. Denham, “Coronalalignment after total knee replacement,” Journal of Bone andJoint Surgery B, vol. 73, no. 5, pp. 709–714, 1991.

Hindawi Publishing CorporationArthritisVolume 2011, Article ID 683970, 16 pagesdoi:10.1155/2011/683970

Review Article

Developmental Mechanisms in Articular Cartilage Degradationin Osteoarthritis

Elena V. Tchetina

Institute of Rheumatology, Russian Academy of Medical Sciences, Kashirskoye Shosse 34A, Moscow 115522, Russia

Correspondence should be addressed to Elena V. Tchetina, etchetina@mail.ru

Received 5 August 2010; Accepted 9 December 2010

Academic Editor: Henning Bliddal

Copyright © 2011 Elena V. Tchetina. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Osteoarthritis is the most common arthritic condition, which involves progressive degeneration of articular cartilage. The mostrecent accomplishments have significantly advanced our understanding on the mechanisms of the disease development andprogression. The most intriguing is the growing evidence indicating that extracellular matrix destruction in osteoarthritic articularcartilage resembles that in the hypertrophic zone of fetal growth plate during endochondral ossification. This suggests commonregulatory mechanisms of matrix degradation in OA and in the development and can provide new approaches for the treatmentof the disease by targeting reparation of chondrocyte phenotype.

1. Introduction

Osteoarthritis (OA) is the most common joint disease, whichis associated with a risk of mobility disability. It affectsapproximately 12% of the aging Western population, while aquarter of people aged over 55 have an episode of persistentknee pain [1]. The pathology of OA involves the wholejoint and is associated with focal and progressive hyalinearticular cartilage loss, concomitant sclerotic changes in thesubchondral bone, and the development of osteophytes. Softtissue structures in and around the joint including synovium,ligaments, and muscles are also involved [2].

OA affects predominantly articular cartilage, whichdegrades by gradual loss of its extracellular matrix (ECM)composed mainly of aggrecan and type II collagen. Loss oflarge proteoglycan aggrecan decreases cartilage compressivestiffness and precedes the damage to collagen fibrillarnetwork, which is responsible for tensile properties of thetissue [3]. Aggrecan degradation is associated with upreg-ulation of aggrecanases a disintegrin and metalloproteasewith thrombospondin motifs (ADAMTS-) 4 and 5 as wellas matrix metalloproteinases (MMPs) [4]. The excessivecleavage of type II collagen in OA is assumed to becaused by the upregulation of the synthesis and activities ofcollagenases [5–7], in particular MMP-13 [8–10]. Presently,

it is believed that articular cartilage destruction in OA resultsfrom excessive loading, age-related changes, and metabolicimbalance in the tissue [11–13].

OA also exhibits features of a systemic disease as it hasbeen shown to involve vascular pathology [14, 15] as wellas T-cell immune response [16, 17] associated with upreg-ulation of cytokines such as interleukin (IL-) β and tumornecrosis factor (TNF)α [3, 18], which aggravate cartilageresorption [19]. As the mechanism of OA development is notcompletely understood, the disease manifestations, whichare associated with cartilage resorption and inflammation,suggest a treatment involving inhibition of proinflammatorycytokines or MMP activity to prevent matrix destruction.However, it does not result in disease modification andproduces severe side effects [20, 21].

Articular cartilage degeneration in OA is also associatedwith changes in chondrocyte phenotype [13, 22, 23]. Specif-ically, these changes resemble those observed during chon-drocyte differentiation in endochondral ossification and arecharacterized by cell cloning, expression of differentiation-related genes such as parathyroid hormone-related peptide(PTHrP) [24], type X collagen [25–27], annexins andalkaline phosphatase (ALP) [28, 29], osteocalcin [30], matrixcalcification [31, 32], as well as apoptotic cell death ofterminally differentiated chondrocytes [33, 34]. All these

2 Arthritis

cellular changes including increased cleavage of type IIcollagen by MMP-13 are also associated with chondrocytehypertrophy observed in the growth plate [35]. This suggeststhat, as articular cartilage shares a common embryologicalorigin with the epiphyseal growth plate [36], destructionof cartilage matrix in OA may involve some of the samecellular and regulatory mechanisms that govern normalchondrocyte terminal differentiation and ECM resorption inskeletal growth and repair [22].

The aim of this paper is to summarize current evi-dence supporting the involvement of molecular mechanismsobserved in the course of chondrocyte progression throughthe growth plate in cartilage matrix destruction in OA.

2. Zonal Gene Expression in EpiphysealGrowth Plate

A central process in endochondral bone formation is aprogressive differentiation of proliferating matrix assemblingchondrocytes to growth-arrested hypertrophic cells. Thisinvolves remodeling and mineralization of the cartilagematrix and leads eventually to its subsequent replacement bybone.

Primary mammalian growth plate physis is structurallyorganized and can be divided into zones, namely, the resting,proliferative, and hypertrophic. Resting zone chondrocytesshow very limited cell division evidenced by low proliferatingcell nuclear antigen (PCNA) expression [37]. They elaboratean extensive extracellular matrix, which is composed pre-dominantly of type II collagen and proteoglycan aggrecan;however, it also contains other collagen types VI, IX, XI,link protein, and small leucine-rich proteoglycans (SLRPs)such as decorin and fibromodulin [37]. Expression of severalregulatory growth factors, such as bone morphogenetic pro-teins (BMPs-) 3, 5, 7, fibroblast growth factor (FGF-) 2, andtransforming growth factor (TGF)β1–3 has been detected inthis zone as well [37–43].

In contrast to resting zone, proliferative zone chondro-cytes actively divide, which is evidenced by the expression ofcyclins [35, 44] and the presence of PCNA positive cells [45].They produce long columns of flattened cells and expresshyaline ECM similar to resting zone chondrocytes. The spacefor the cells newly formed in the course of cell division isgenerated by the matrix-degrading activity of collagenasesMMP-13, MT1-MMP [46, 47], and other MMPs such asMMP-3 [48]. These cells also express proliferation-specificgrowth factors, namely, TGFβ1–3, FGF-2 [35, 43, 49, 50],PTHrP, insulin growth factor (IGF-) I and II [35, 51–53],a cell death inhibitor that regulates apoptosis Bcl-2 (B-celllymphoma-2) [54], and transcription factor Sox9 (SRY-typehigh-mobility-group box transcription factor 9) [35, 55].Although PTHrP [56], TGFβ2, and FGF-2 [57, 58] have beenreported to stimulate MMP-13 expression in rodents, in theearly proliferative zone of the growth plate, their expressiondoes not induce significant matrix loss probably due to thelack of gelatinase (MMP-2 and -9) expression [35, 59].

Cessation of cell division in the growth plate is associatedwith upregulation of cell cycle inhibitors p18, p19, and p21

[60], growth arrest and DNA damage-inducible (GADD)45beta gene [61, 62], as well as apoptosis inhibitors Bcl2 andBag1 (Bcl2-associated athanogene 1), a Bcl2-binding proteincapable of enhancing Bcl2 activity [42, 63], and a marker ofapoptosis caspase 3 [42]. At this point, chondrocytes partiallyresorb their extracellular matrix, enlarge, round up, andfinally mature into hypertrophic cells, which express type Xcollagen (COL10A1), a marker of chondrocyte hypertrophy.Alkaline phosphatase shows the most pronounced expressionalso in hypertrophic chondrocytes [64, 65]. This phenotypicmodification in growth plate chondrocytes is associated withdramatic alteration in regulatory gene expression, namely,upregulation of growth factors such as TGFβ1 and -3 [35,50], BMP-2, -4, -6, and -7 [39, 40, 66–68], connective tissuegrowth factor (CTGF) [69], vascular endothelial growthfactor (VEGF) [59, 70], and Indian hedgehog (Ihh) [35,71, 72]. Inflammation-related cytokine IL-1 expression alsohas been observed only in the hypertrophic chondrocytes[73].

These regulatory growth factors are expressed in asso-ciation with runt-related transcription factor (RUNX)2,which is essential both for osteoblast differentiation [74] andchondrocyte maturation during endochondral ossification[75–78], and is capable of inducing MMP-13 expression[79, 80].

Expression of these growth and transcription factors isalso associated with upregulation of matrix proteins, suchas collagen type II (COL2A1) concomitantly with theirdegrading enzymes MMP-13 and gelatinases MMP-2 and -9 [35, 67]. At this time, overt type II collagen degradationoccurs [46] indicating that genes for both matrix synthesisand degradation are coregulated. However aggrecan remainsretained in the tissue at that time [3].

In the lower hypertrophic zone, mineralization (orcalcification) of residual matrix remaining after its resorp-tion is initiated focally [3]. This involves deposition ofhydroxyapatite mineral [81]. Mineralization process in thelower hypertrophic zone of the growth plate is associatedwith expression of osteocalcin, which is a marker of matureosteoblasts and is involved in chondrocyte mineralizationand Ca+2 homeostasis [28]. Upregulation of ankylosis pro-tein (Ank), which is responsible for transport of intracellularinorganic pyrophosphate to the extracellular milieu, hasbeen also observed in this zone [82]. Mineralization islikely regulated by annexins II, V, and VI, which are highlyexpressed in the hypertrophic and terminally differentiatedmineralizing growth plate chondrocytes and form calciumchannels enabling formation of first mineral phase [83, 84].For example, annexin V has been shown to be capable ofupregulating annexins II, VI, osteocalcin, Runx2, and ALPas well as stimulating apoptotic activity in the lowest part ofthe growth plate [83, 85]. In contrast, TGFβ2, which is alsoexpressed by lower hypertrophic chondrocytes [35], is mostprobably involved in osteoblast formation [86].

Therefore, chondrocyte maturation in the growth plate isassociated with expression of stage-specific set of regulatorygrowth and transcription factors producing changes incellular phenotype and synthesis of stage-specific extracel-lular matrix, which eventually degrades in the hypertrophic

Arthritis 3

zone. All these cellular activities require careful and specificcoordination.

3. Regulation of Growth Plate ChondrocyteDifferentiation

Chondrocyte differentiation is initiated in the center of thecartilaginous bone rudiment and is thought to be inducedby hypoxia and/or nutrient deficiency [87]. The pace ofchondrocyte differentiation is regulated by various agentsincluding paracrine and autocrine growth factors and hor-mones [3, 88]. They are responsible for specific regulatorymolecule expression by chondrocytes in the course of theirprogression through the growth plate.

Growth factors secreted by fetal chondrocytes are incharge of mutually exclusive processes of chondrocyte pro-liferation and terminal differentiation. Thus, proliferation-related growth factors such as basic fibroblast growth factorand parathyroid hormone-related peptide stimulate restingchondrocytes to proliferate and suppress terminal differen-tiation of hypertrophic chondrocytes [89–96]. In addition,PTHrP, in combination with Indian hedgehog, regulateschondrocyte differentiation through the establishment ofa negative feedback mechanism, whereby Ihh and PTHrPcan together suppress hypertrophy [97–100]. Alternatively,interactions of Ihh with syndecan 3, which serves as a growthfactor coreceptor, are important for restricting mitoticactivity to the proliferative zone of mammalian growth plate[101].

Transforming growth factor betas are multifunctionalmolecules regulating cellular proliferation, differentiation,and extracellular matrix function [75, 102]. TGFβ trans-ported from apoptotic chondrocytes to the region of celldivision would be expected to stimulate matrix production,delay hypertrophic differentiation, and thus maintain growthplate width [103, 104]. TGFβ1 [105–107] and TGFβ2[108] each are able to suppress chondrocyte hypertrophyby coordinate inhibition of collagenase expression. Thisis partially associated with upregulation of PTHrP geneexpression that exerts both PTHrP-dependent and PTHrP-independent effects on endochondral bone formation [105,109, 110]. TGFβ2 in synergy with FGF-2 has been alsoshown to suppress chondrocyte maturation and hypertrophy[111, 112].

BMP signaling is also essential for chondrocyte progres-sion through the growth plate [66]. Zone specific expressionof various BMPs suggests their involvement in chondrocytephenotypic changes in the course of both proliferationand hypertrophy. Thus, BMP-2 and -6 have been shownto promote chondrocyte hypertrophy by upregulation ofIhh and type X collagen expression and downregulationof FGF signaling involving Runx2 [113–118]. At the sametime, BMP-2 and -9 augmented mitogenic effect of IGF-1,while BMP-5 increased cell proliferation and cartilage matrixsynthesis [119, 120].

IGF-1, a structural and functional analog of insulin, pro-motes chondrocyte proliferation and differentiation whileit inhibits apoptosis [89, 93]. It is also an important

regulator of PTHrP-Ihh feedback loop. The lack of IGFresults in downregulation of Ihh expression and upregulationof PTHrP [51]. IGF-1 favors chondrocyte hypertrophicdevelopment as it induced type X collagen and alkalinephosphatase in avian sternal chondrocytes [108, 112]. Inaddition, insulin and IGF-1 [121] both are strong stimulatorsof aggrecan and type II collagen synthesis [122].

Furthermore, chondrocyte differentiation in the growthplate is regulated by various transcription factors [123].Transcription factors Sox9 and -4 have been shown todetermine the rate of chondrocyte differentiation intohypertrophy and the expression of chondrocyte-specificmatrix molecules including Col2A1, Col9A2, Col11A1, andaggrecan [124–130]. They are also required to preventconversion of proliferating chondrocytes into hypertrophicchondrocytes [55]. Transcription factors Runx1-3 are themost important as they play a crucial role both in chon-drocyte maturation and had been shown to induce MMP-13 expression [77, 80, 125, 131]. Recently, the involvementof several other transcription factors such as Shox/Shox2,Dlx5, and MEF2C has been shown to control skeletal growththat suggests their potential contribution in ectopic chondro-cyte hypertrophy development [132, 133]. Wnt/beta-cateninsignaling can also mediate chondrocyte hypertrophy as it iscapable of upregulating type X collagen, Runx2, and alkalinephosphatase expression while inhibiting Sox9 and type IIcollagen expression [92].

Prostaglandin E2 (PGE2), a potent lipid molecule thatregulates a broad range of physiologic reactions, can inhibitgrowth plate chondrocyte differentiation by downregulationof differentiation-related genes COL10A1, VEGF, MMP-13, and alkaline phosphatase expression as well as theirenzyme activity [134, 135]. At the same time, low con-centrations of this prostaglandin are capable of increasingproliferation of growth plate chondrocytes [136, 137]. Incontrast, chemokine stromal sell-derived factor 1, annexinV, and Ank have been shown to stimulate hypertrophy,mineralization, and apoptosis, when they are overexpressedin nonmineralizing growth plate chondrocytes [82, 85, 138,139].

Extracellular matrix proteins produced by chondrocyteshave also exhibited a capacity to regulate growth platechondrocyte hypertrophy. Thus, type II collagen, aggrecan,and matrilin-3 are likely to inhibit hypertrophy as thesematrix component deficiency produced premature mat-uration in mutant chondrocytes [140–142]. Furthermorea functional link between chondrocyte hypertrophy andextracellular matrix degradation is also supported by the factthat downregulation of chondrocyte hypertrophy evidencedby suppression of type X collagen, Runx2 and MMP-13expression is associated with inhibition of collagen cleavageactivity in cultured hypertrophic growth plate chondrocytestreated with MMP-13 inhibitor [8, 143, 144]. This indicatesa functional link between chondrocyte hypertrophy andextracellular matrix degradation.

It is necessary to note that variable effects of regulatorymolecules are carefully coordinated to provide accuracy inthe process of endochondral ossification. Thus, it has beendemonstrated that growth plate chondrocyte progression

4 Arthritis

to hypertrophy is a subject to negative control that canbe arrested at various checkpoints [112]. Accordingly, anearly proliferative phenotype in avian fetal chondrocyteshas been reassumed by treatment with TGFβ2, FGF-2, andinsulin in combination, while differential Ihh expression wasresponsible for acquisition of the late proliferative phenotypein hypertrophic cells [112]. In another study, the releaseof terminally differentiated hypertrophic chondrocytes fromtheir environment also resulted in downregulation of type Xcollagen synthesis, activation of proliferation, and reinitia-tion of aggrecan synthesis [145].

Therefore, chondrocyte differentiation is carefully regu-lated in the course of endochondral ossification. Eventually,epiphyseal chondrocytes give rise to articular cartilage, whosestructural components and regulatory networks at leastpartially resemble that in the growth plate.

4. Zonal Gene Expression in HealthyArticular Cartilage

Healthy articular cartilage is characterized by a very lowexpression of collagens type II, VI, IX, and XI [146] andrelatively high turnover rate for aggrecan [147]. It is alsocharacterized by expression of matrix turnover genes suchas MMP-3 [148], occasionally detected MMP-1, -8, -13[149], and growth factors TGFβ1 [150] and PTHrP [24,151]. Antiangiogenic factor chondromodulin-1 [152, 153],p16INK4α, and Gadd45α/β genes, the latter is associatedwith environmental and intrinsic stress [154, 155], areexpressed in all the cartilage zones. At the same, time noexpression of type I and X collagens [156, 157], a completelack of expression of TGFβ2, IGF-1, Ihh [158, 159], annexinVIII [160], and osteocalcin [30] was observed in healthycartilage.

Articular cartilage can be divided into superficial, mid-,and deep zones; the latter is followed by the calcified cartilageproviding junction of the cartilage to the subchondral bone[157]. These zones differ in expression of specific matrixmolecules, their modifying enzymes, and regulatory growthfactors, which are responsible for articular cartilage integrityand function. Although normal articular chondrocytes areless metabolically active than the growth plate chondrocytes,some similarity in gene expression pattern in the individualcartilage zones has been noted.

Superficial zone of healthy articular cartilage containsflattened chondrocytes surrounded by specialized extracel-lular matrix rich in thin collagen fibrils [161] and smallleucine-rich proteoglycans-decorin and biglycan [162]. Italso contains the lowest amount of predominant cartilageproteoglycan aggrecan compared to other zones of articularcartilage. This zone is rich in regulatory molecules suchas TGFβ1 and -3 and BMP 1–6 [37, 40]. Proliferativepotential of these cells is indicated by the expression ofcyclin D2; however, it may be suppressed by cell divisioninhibitors such as growth arrest specific protein (Gas)-1 andGadd45α, which are also expressed in this cartilage zone[37]. This is supported by the lack of superficial chondrocyteproliferative activity determined by PCNA staining [163].

MMP-3 expression was observed in this cartilage zone moreoften than MMP-1, -8, and -13, however these proteinases donot produce any matrix degradation and are likely involvedin matrix turnover [149, 164]. Expression of antiapoptoticBcl2 and Bag1 genes was detected predominantly in this zonein old mice, while it was observed throughout the articularcartilage in the young animals [63].

Mid-zone chondrocytes are round in shape, surroundedby ECM composed of thick collagen fibrils and rich inaggrecan. Chondrocytes in this zone do not show anyproliferative activity determined by PCNA staining similarto superficial zone cells [163]. However, these cells are likelyto possess a potential for proliferation, as FGF-2, capableof inducing proliferation in normal articular chondrocytesin culture [165], has been detected in the mid-zone ofmouse articular cartilage [49]. BMP1–7 expression was alsoobserved in the mid-zone of normal articular cartilage[40].

Deep zone chondrocytes are grouped in clusters andresemble hypertrophic chondrocytes of the growth plate [3].In this zone, cartilage matrix has the highest content ofaggrecan [166], the lowest amounts of small leucine-richproteoglycans [162], and the largest diameter of collagenfibrils. Similar to hypertrophic zone of the growth plate,BMP1–7 [40], Ihh expression [167], and the highest amountof annexin VI-positive cells were observed in the deepzone of human articular cartilage [163]. The lowest partof the deep zone, which is partly calcified, expressed amarker of chondrocyte hypertrophy type X collagen andis rich in alkaline phosphatase. MMP-13 expression [149]and negligible activity of chondrocyte apoptosis was alsosometimes observed here [168, 169].

However, in spite of low activity of cellular and matrixturnover, healthy articular cartilage possesses a strongmetabolic potential, whose activation is observed duringdevelopment of pathological condition such as OA.

5. Early Development of Osteoarthritis

Early OA changes in articular cartilage are associated withsignificant metabolic activation of articular chondrocytes.This involves sequential and zonal upregulation of chondro-cyte differentiation-related genes as well as an increase inthe activity of the same MMPs, which are responsible formatrix degradation in the hypertrophic zone of the growthplate. Spatially, these genes are upregulated in the mid- andsuperficial zones of articular cartilage, where lately the firstsigns of cartilage destruction occur.

Mild OA changes (Mankin 1–4) are characterized by theloss of proteoglycans in the surface area [163, 170, 171].Although these changes were not accompanied by significantstructural disturbances in the tissue, they were associatedwith increased type II collagen and aggrecan synthesis,upregulation of chondrocyte proliferation evidenced byincreased PCNA and Ki67 staining and MMP-13 expression[172–175]. This was followed by the cellular changes similarto those observed in hypertrophic zone of the growth plateas indicated by type X collagen production, collagenase

Arthritis 5

and alkaline phosphatase staining, and increased type IIcollagen cleavage activity in the mid-zone [172, 176]. Later,chondrocyte activation extends to the superficial zone, whereit is accompanied by chondrocyte apoptosis evidenced bythe presence of the cells carrying DNA nicks [172]. IL-1βexpression is also upregulated both in the superficial anddeep cartilage zones in early OA [177]. Similar to biphasicMMP-13 expression in the growth plate, upregulation of thiscollagenase in the articular cartilage was initially precededand later accompanied by type X collagen and alkalinephosphatase expression [172, 176].

An early OA articular cartilage degeneration is observedfocally. Spatial distribution of chondrocyte differentiation-related gene expression in the areas adjacent to and remotefrom the early lesion also resembles that in the growthplate and is associated with increased collagenase cleavage oftype II collagen. Thus, collagenases MMP-1, MMP-14 (MT1-MMP), and aggrecanase ADAMTS-5 (but not ADAMTS-4), cytokines IL-1α/β and TNF-α, chondrocyte terminaldifferentiation-related genes COL10A1, MMP-13, MMP-9,Ihh, and caspase 3 were often upregulated in the vicinityof the lesion. Growth factors associated with growth platechondrocyte proliferation, namely, FGF-2, PTHrP, and TGFβ1/2, as well as the matrix molecules COL2A1 and aggrecan,were expressed adjacent to and remote from the lesion[22]. In addition, a distinct spatial reorganization in humansuperficial chondrocytes in remote area from early OAlesions has been recently reported [178].

However, of all genes, only caspase 3 and ADAMTS-5 expression was exclusively seen in association with earlylesions. Elevation of collagenase activity was associated witha frequent elevation of expression of COL10A1, caspase 3, IL-1α/β, MMP-1, and ADAMTS-5 and a decreased expression ofSox-9, TGF-β1, TGF-β2, TNF-α, and aggrecan [22].

Moderate OA changes in the articular cartilage (Mankin6–9), which are characterized by the lack of fibrillations,some loss of superficial zone, and some clustering of cells[163, 171], are associated with the increase of PCNA stainingin the superficial zone and annexin VI and VIII antigenupregulation in the mid- and deep zones [160, 163, 179].

Therefore, articular chondrocyte activation in early OA,which is the most pronounced in the superficial and mid-zones, resembles that observed during chondrocyte matura-tion in the growth plate.

6. Gene Expression in LateOsteoarthritic Cartilage

Severe OA (Mankin ≥10) is characterized by extensivefissuring and fibrillation, clustering of chondrocytes, and lossof cartilage [163]. Cartilage zonal organization is disturbed.The superficial zone degradation produces rough fibrillatedsurface, fissures, and cracks extending to the calcified zone.This is accompanied by severe proteoglycan loss followedby degradation of type II collagen [164, 180]. Collagendegradation occurs around chondrocytes. At this time,upregulation of MMPs-13, -2, -11, ADAMTS as well asexpression of collagens type I, II, III, VI, and X were

observed near the articular surface [148, 181–183] and wasaccompanied by strong expression of IL-1β and TNFα [149].At the same time, collagen replenishment is limited as Col2AN-propeptide, a marker of collagen synthesis, was detectedonly in the deep zone close to subchondral bone [184].As it was stated above, all these gene activities have beenalso observed in the hypertrophic zone of the fetal growthplate.

Chondrocyte terminal differentiation-related geneexpression is also observed in cell clusters located aroundfissures [185, 186]. In these clusters, both collagen type IIand X synthesis [187] as well as TGFβ3 and its receptorregulator Smad-2P expression were observed [188].

PCNA and syndecan-3, the markers of early fetalchondrocyte differentiation, as well as annexin VI andalkaline phosphatase, which are involved in terminal stageof differentiation, all are upregulated near articular surfacein human OA articular cartilage [30, 163]. At the same time,annexin VIII and osteocalcin, which were never detected innormal articular cartilage, were observed in the mid- anddeep zones in late OA cartilage [30, 160]. An increase inBMP-2 expression [188] was associated with upregulationof tumor suppressor p53 expression [189] and cyclin-dependent kinase inhibitor p16INK4a upregulation in all thecartilage zones [155] indicating inhibition of proliferativepotential in late OA chondrocytes. However, repression ofantiproliferative factor Tob1 has been also reported in the latestage of knee OA cartilage [183].

The most severely damaged rodent knee OA articu-lar cartilage has shown significantly reduced expressionof proliferation-related growth factors and their signalingmolecules such as PTHrP, TGFβ3 and Smad-2P, TGFβ1and its receptor II [188, 191]. However, in human hip OA,both downregulation and upregulation of TGFβ1–3 isoformexpression compared to healthy cartilage have been reported[43, 192], while one study failed to detect any upregulationof chondrocyte differentiation and hypertrophy markersassociated with late OA [193].

Antiangiogenic factor chondromodulin-1 downregula-tion concomitant to VEGF upregulation indicating increasedvascular invasion into cartilage in advanced OA has been alsoobserved [194, 195]. This was accompanied by upregulationof chondrocyte apoptosis, a marker of the final step ofchondrocyte differentiation, which was more pronounced inOA cartilage compared to normal specimens [189].

Overall degrading activity prevailed over synthesis asserum levels of Col2A N-propeptide were lower than that ofcollagen degradation products in late OA patients comparedto controls indicating the uncoupling of collagen synthesisand degradation in OA [190]. Moreover, serum increasein both Col2A N-propeptide and collagen degradationfragments was often indicative on the most aggressive diseaseprogression [196].

Thus, the similarity in the gene expression profiles asso-ciated with matrix destruction in OA articular cartilage andin the hypertrophic zone of the growth plate observed in themajority of studies suggests an acquisition of hypertrophicphenotype traits by OAarticular chondrocytes.

6 Arthritis

7. Inhibition of Articular ChondrocyteHypertrophy Suppresses OA CartilageDegeneration

The similarity of ECM degradation in OA to that in thehypertrophic zone of primary growth plate involves upregu-lation of type II collagen cleavage by collagenase and expres-sion of regulatory differentiation-related growth factors andmatrix proteins, which are associated with chondrocytehypertrophy [6, 22, 25, 28]. Therefore, the above observationthat hypertrophic changes in the growth plate chondrocytesare reversible [112] suggests a possibility for OA articularchondrocytes to regain healthier phenotype when they aretreated by the agents inhibiting fetal hypertrophy.

In fact, the same growth factors, namely, TGFβ2, FGF-2, and insulin, which were previously used individually or incombination to suppress hypertrophy in growth plate chon-drocytes [112], have been shown to be capable of arrestingtype II collagen cleavage, chondrocyte differentiation-relatedgene, and proinflammatory cytokine expression in humanOA articular cartilage explants [197]. Another combinationof TGFβ1 and IGF1 inhibited collagen degradation, a markerof extracellular matrix destruction, which was inducedby oncostatin M and TNFα in bovine articular cartilage[198]. It has been also shown that TGFβ inhibition ofchondrocyte differentiation is likely mediated by Smad2/3pathway through modulation of Runx2 function [75]. Inanother study, a major proliferation-related growth factorPTHrP downregulated terminal differentiation-related genesin cultured mineralizing articular chondrocytes from thedeep zone as well as in chondrogenic articular cartilage con-structs [199, 200]. It is worth to note here that growth factors,which were capable of inhibiting collagen degradation inOA articular cartilage, are predominantly expressed in theproliferative zone of the growth plate and are required forchondrocyte proliferation in the development.

Similar effect on suppression of collagen cleavage in asso-ciation with inhibition of chondrocyte hypertrophy-relatedgenes and proinflammatory cytokines IL-1β and TNFαhas been observed on treatment of OA explants with lowconcentrations of PGE2 [201]. Although PGE2 is expressedin all the growth plate zones, it is primarily required forfetal chondrocyte proliferation [137] and is capable alsoof inhibiting their terminal differentiation [134, 135] andexpression of proinflammatory mediators [202]. At the sametime, PGE2 at higher concentrations has been shown to exertstimulating effects on cartilage degradation [203].

Alternatively, downregulation of the genes, which areexpressed in the hypertrophic zone of the growth plate andassociated with chondrocyte terminal differentiation suchas Hedgehog signaling, TGFβ1/BMP signaling, transform-ing growth factor-beta-activated kinase (TAK) 1, cyclin-dependent kinase inhibitor p16INK4a, ADAMTS5, RUNX2,and caspases, resulted in abrogation of matrix degeneration,less type X collagen production, and MMP-13 expression[155, 204–210]. Therefore, upregulation of the genes asso-ciated with growth plate chondrocyte proliferation or down-regulation of hypertrophy-related genes favors acquisition ofhealthier phenotype in OA articular chondrocytes.

However, while direct inhibition of cartilage degradationby the agents capable of regulating chondrocyte differentia-tion is an attractive means to counteract articular cartilagedegeneration in OA, it has several limitations. Thus, fol-lowing the inhibition of cartilage degradation by individualgrowth factors (GFs) reparation of articular cartilage in OAin vivo may require a combination of GF [18, 211, 212]. Forexample, being the most efficient in suppressing OA articularcartilage destruction [197], TGFβ2 alone may not be capableof restoring the anabolic functions of healthy articularcartilage since it has been reported to downregulate typeII collagen and aggrecan synthesis [108, 213]. In contrast,in responsive individuals, insulin may facilitate tissue repairas it is a principal anabolic agent in the articular cartilage[214]. FGF-2 can also promote cartilage repair [212] byitself or inducing local TGFβ or its own expression [215].In addition, combinations of these and other growth factorshave been shown to produce synergistic effect in maintainingsynthesis of matrix molecules in articular and growth platechondrocytes [108, 216, 217].

Another concern on GF application is related to theirpossible catabolic effects. Although no evidence has beenobtained that TGFβ2 can act catabolically in human OAarticular cartilage [197], destructive potential of this growthfactor at high concentration was observed in normal articularcartilage in vivo after its intraarticular injections, whichproduced joint swelling, fibroblastic proliferation of synovialmembrane, and profound loss of articular cartilage pro-teoglycan in rabbit joints [213]. Therefore, the delivery ofexact therapeutic amount of the growth factor to the site ofarticular cartilage destruction may be important. This hasbeen demonstrated in a recent study, where deleterious effectof TGFβ1 capable of inducing synovial fibrosis has beencounteracted by combined overexpression of TGFβ1 and itsinhibitor Smad7 [188]. This resulted both in prevention ofproteoglycan (PG) loss and in increase in PG content inmouse OA cartilage.

8. Modeling of OA-Related Changes in HealthyArticular Cartilage Is Associated withChondrocyte Hypertrophy Development

In healthy adult articular cartilage, chondrocyte differen-tiation does not occur in the noncalcified cartilage [218].However, when maturational arrest is abolished, chondro-cyte differentiation-related genes, which are barely expressedin healthy articular cartilage, become upregulated followedby hypertrophic changes in the cells and extracellular matrix.If this notion is true, stimulation of degradation in healthyarticular cartilage should be accompanied by chondrocytehypertrophy development. The relieve of transcriptionalrepression can be attained by cartilage treatment withazacytidine C (Aza-C), which replaces cytidine bases ingenomic DNA during replication and disturbs methylationpattern of cytidines (CpG islands) in target gene promoters.This was associated with upregulation of PTHrP, governingchondrocyte proliferation in the growth plate, as well aschondrocyte hypertrophy-related collagen type X, Ihh, and

Arthritis 7

alkaline phosphatase gene expression, and the increase inchondrocyte cell size in healthy articular cartilage [219–221].

On the other hand, alterations associated with chondro-cyte hypertrophy in the growth plate are always accompa-nied by overt extracellular matrix resorption producing itsdegradation fragments [3]. Therefore, it is not surprisingthat collagen and/or fibronectin degradation peptides, whichcan be also released on mechanical destruction of articularcartilage caused by trauma or joint overload in case ofanterior cruciate ligament transection, have been shown tobe capable of inducing articular cartilage degradation byupregulation of collagenase and MMPs activity [222–224].Besides, these peptides upregulated chondrocyte prolifer-ation, production of type X collagen and apoptotic cellson the surface of articular cartilage explants [172, 224].Collagen fragments may also account for OA-like changesinduced in healthy cartilage by overexpression of MMP-13, which were associated with chondrocyte hypertrophyin mouse articular cartilage [47]. In addition, other matrixdisturbances such as lack of matrilin-3 by a correspondinggene knockout produced premature chondrocyte maturationto hypertrophy and formed predisposition to develop severeOA in mice [141].

Functional disturbances in the regulatory genes involvedin chondrocyte differentiation can also produce OA-relatedchanges in healthy articular cartilage resembling chondro-cyte maturation in the growth plate. Thus, deficiency inTGFβ signaling, which is essential for articular cartilagemaintenance and had been induced either by overexpressionof functionless TGFβ type II receptor [225] or by deletion ofSmad3 signaling [104, 226], caused accelerated chondrocytedifferentiation associated with type X collagen expressionand OA-like changes in articular cartilage. It has been alsoshown that the absence of signaling through Fgfr (fibroblastgrowth factor receptor) 3 in the joints of Fgfr3(−/−)mice produced premature cartilage degeneration and earlyarthritis [227]. In contrast, TGFalpha signaling suggestscatabolic potential of this growth factor as it has been shownto stimulate articular chondrocyte proliferation, formationof cell clusters followed by expression of matrix-degradingenzymes MMP-13, cathepsin C and downregulation of Sox9,as well as collagen and aggrecan expression in rat articularosteochondral explants [228].

Being an important factor of OA articular cartilagepathology proinflammatory cytokines such as TNFα and IL-1β have been shown to mediate articular cartilage degrada-tion by upregulation of matrix-degrading MMPs [229]. It hasbeen observed recently that increased expression of proin-flammatory agents such as TNFα, chemokines IL-8, growth-related oncogene α (GROα), or the multiligand receptorfor advanced glycation end products (RAGE) induced alsochondrocyte hypertrophy evidenced by collagen type Xexpression [230, 231]. This suggests a link between inflam-mation and altered differentiation in articular chondrocytes[231]. Interestingly, the impairment TGFβ signaling by IL-1β was mediated by downregulation of TGFβRII [232]. Theloss of function of this receptor has previously been linked tochondrocyte hypertrophy induction and OA development inanimal studies [225].

Therefore, OA-like alterations in healthy articular car-tilage induced by the mediators, which are upregulated inthe hypertrophic zone during endochondral ossification,are accompanied by ECM degradation and associated witharticular chondrocyte hypertrophy.

9. Conclusions

The data presented here shows a significant progress inour understanding of molecular mechanisms of articularcartilage degradation in OA. They involve at least in partsimilar machinery of extracellular matrix resorption inthe hypertrophic zone of the growth plate and in OAarticular cartilage in the course of its degeneration. Theobservation that profound cellular phenotypic changes inarticular chondrocytes occur prior the overt cartilage matrixdegradation monitored histologically suggests that articularchondrocyte phenotype modifications can be recognizedvery early in the disease at gene expression level favoringtimely disease recognition, which could help its prevention.This implies also innovative opportunities in suppression ofcartilage matrix degradation targeting inhibition of chon-drocyte hypertrophy and suggests new targets for therapeuticintervention. For this purpose, further studies are requiredin search of new agents generating programmable articularchondrocyte phenotype modification.

Abbreviations

OA: OsteoarthritisECM: Extracellular matrixMMP: MetalloproteinasesIL: InterleukinADAMTS: A disintegrin and metalloprotease with

thrombospondin motifsPTHrP: Parathyroid hormone-related peptideTNF: Tumor necrosis factorALP: Alkaline phosphatasePCNA: Proliferating cell nuclear antigenSLRPs: Small leucine-rich proteoglycansBMP: Bone morphogenetic proteinFGF: Fibroblast growth factorTGF: Transforming growth factorTGFR: Transforming growth factor receptorIGF: Insulin growth factorBcl-2: B-cell lymphoma 2Sox9: SRY-type high-mobility-group box

transcription factor-9GADD45beta: Growth arrest and DNA damage-inducible

45betaBag1: Bcl2-associated athanogene 1, a

Bcl2-binding protein capable of enhancingBcl2 activity

COL10A1: Type X collagenBMP: Bone morphogenetic proteinsCTGF: Connective tissue growth factorVEGF: Vascular endothelial growth factorIhh: Indian hedgehogRUNX2: Runt-related transcription factor

8 Arthritis

COL2A1: Collagen type IIAnk: Ankylosis proteinPGE2: Prostaglandin E2Gas1: Growth arrest specific protein1Aza-C: Azacytidine CFgfr3: Fibroblast growth factor receptor 3RAGE: Multiligand receptor for advanced glycation

end productsGRO: Growth-related oncogenePG: ProteoglycanTAK: Transforming growth factor-beta-activated

kinaseGF: Growth factor.

Acknowledgment

The author is supported by the Russian Foundation for BasicResearch (Project no. 09-04-01158a).

References

[1] D. J. Hunter and D. T. Felson, “Osteoarthritis,” BritishMedical Journal, vol. 332, no. 7542, pp. 639–642, 2006.

[2] S. B. Abramson and M. Attur, “Developments in thescientific understanding of osteoarthritis,” Arthritis Research& Therapy, vol. 11, no. 3, article 227, 2009.

[3] A. R. Poole, “Cartilage in health and disease,” in Arthritisand Allied Conditions: A Textbook of Rheumatology, W.Koopman, Ed., pp. 223–269, Lippincott, Williams & Wilkins,Philadelphia, Pa, USA, 15th edition, 2005.

[4] H. Nagase and M. Kashiwagi, “Aggrecanases and cartilagematrix degradation,” Arthritis Research & Therapy, vol. 5, no.2, pp. 94–103, 2003.

[5] A. R. Poole, F. Nelson, L. Dahlberg et al., “Proteolysisof the collagen fibril in osteoarthritis,” Biochemical SocietySymposium, no. 70, pp. 115–123, 2003.

[6] A. R. Poole, F. Guilak, and S. B. Abramson, “Etiopathogenesisof osteoarthritis,” in Osteoarthritis: Diagnosis and Medi-cal/Surgical Management, R. W. Moskowitz, R. D. Altman, M.C. Hochberg, J. A. Buckwalter, and V. M. Goldberg, Eds., pp.27–49, Williams & Wilkins, Lippincott, Pa, USA, 4th edition,2007.

[7] A. R. Poole, M. Kobayashi, T. Yasuda et al., “Type IIcollagen degradation and its regulation in articular cartilagein osteoarthritis,” Annals of the Rheumatic Diseases, vol. 61,no. 2, pp. ii78–ii81, 2002.

[8] C. W. Wu, E. V. Tchetina, F. Mwale et al., “Proteolysisinvolving matrix metalloproteinase 13 (collagenase-3) isrequired for chondrocyte differentiation that is associatedwith matrix mineralization,” Journal of Bone and MineralResearch, vol. 17, no. 4, pp. 639–651, 2002.

[9] R. C. Billinghurst, L. Dahlberg, M. Ionescu et al., “Enhancedcleavage of type II collagen by collagenases in osteoarthriticarticular cartilage,” Journal of Clinical Investigation, vol. 99,no. 7, pp. 1534–1545, 1997.

[10] L. Dahlberg, R. C. Billinghurst, G. Webb et al., “Selectiveenhancement of collagenase-mediated cleavage of residenttype II collagen in cultured osteoarthritic cartilage and arrestwith a synthetic inhibitor that spares collagenase 1 (matrixmetalloproteinase 1),” Arthritis and Rheumatism, vol. 43, no.3, pp. 673–682, 2000.

[11] T. Aigner, J. Haag, J. Martin, and J. Buckwalter, “Osteoarthri-tis: aging of matrix and cells—going for a remedy,” CurrentDrug Targets, vol. 8, no. 2, pp. 325–331, 2007.

[12] T. Aigner and N. Gerwin, “Growth plate cartilage as devel-opmental model in osteoarthritis research potentials andlimitations,” Current Drug Targets, vol. 8, no. 2, pp. 377–385,2007.

[13] P. M. van der Kraan, E. N. Blaney Davidson, and W. B.van den Berg, “A role for age-related changes in TGF-beta signaling in aberrant chondrocyte differentiation andosteoarthritis,” Arthritis Research & Therapy, vol. 12, no. 1,pp. 201–214, 2010.

[14] D. M. Findlay, “Vascular pathology and osteoarthritis,”Rheumatology, vol. 46, no. 12, pp. 1763–1768, 2007.

[15] P. Ghosh and P. A. Cheras, “Vascular mechanisms inosteoarthritis,” Best Practice and Research: Clinical Rheuma-tology, vol. 15, no. 5, pp. 693–709, 2001.

[16] H. de Jong, S. E. Berlo, P. Hombrink et al., “Cartilageproteoglycan aggrecan epitopes induce proinflammatoryautoreactive T-cell responses in rheumatoid arthritis andosteoarthritis,” Annals of the Rheumatic Diseases, vol. 69, no.1, pp. 255–262, 2010.

[17] L. I. Sakkas and C. D. Platsoucas, “The role of T cells inthe pathogenesis of osteoarthritis,” Arthritis and Rheumatism,vol. 56, no. 2, pp. 409–424, 2007.

[18] C. J. Malemud, “Anticytokine therapy for osteoarthritis:evidence to date,” Drugs and Aging, vol. 27, no. 2, pp. 95–115,2010.

[19] M. Kobayashi, G. R. Squires, A. Mousa et al., “Roleof interleukin-1 and tumor necrosis factor α in matrixdegradation of human osteoarthritic cartilage,” Arthritis andRheumatism, vol. 52, no. 1, pp. 128–135, 2005.

[20] C. B. Little and A. J. Fosang, “Is cartilage matrix breakdownan appropriate therapeutic target in osteoarthritis—insightsfrom studies of aggrecan and collagen proteolysis?” CurrentDrug Targets, vol. 11, no. 5, pp. 561–575, 2010.

[21] R. F. Loeser, “Molecular mechanisms of cartilage destructionin osteoarthritis,” Journal of Musculoskeletal Neuronal Inter-actions, vol. 8, no. 4, pp. 303–306, 2008.

[22] E. V. Tchetina, G. Squires, and A. R. Poole, “Increasedtype II collagen degradation and very early focal cartilagedegeneration is associated with upregulation of chondrocytedifferentiation related genes in early human articular carti-lage lesions,” Journal of Rheumatology, vol. 32, no. 5, pp. 876–886, 2005.

[23] R. Yagi, D. McBurney, D. Laverty, S. Weiner, and W. E. Hor-ton Jr., “Intrajoint comparisons of gene expression patternsin human osteoarthritis suggest a change in chondrocytephenotype,” Journal of Orthopaedic Research, vol. 23, no. 5,pp. 1128–1138, 2005.

[24] R. Terkeltaub, M. Lotz, K. Johnson et al., “Parathyroidhormone-related protein is abundant in osteoarthritic car-tilage, and the parathyroid hormone-related protein 1-173 isoform is selectively induced by transforming growthfactor β in articular chondrocytes and suppresses generationof extracellular inorganic pyrophosphate,” Arthritis andRheumatism, vol. 41, no. 12, pp. 2152–2164, 1998.

[25] K. von der Mark, T. Kirsch, A. Nerlich et al., “Type X collagensynthesis in human osteoarthritic cartilage: indication ofchondrocyte hypertrophy,” Arthritis and Rheumatism, vol. 35,no. 7, pp. 806–811, 1992.

[26] I. Girkontaite, S. Frischholz, P. Lammi et al., “Immunolo-calization of type X collagen in normal fetal and adult

Arthritis 9

osteoarthritic cartilage with monoclonal antibodies,” MatrixBiology, vol. 15, no. 4, pp. 231–238, 1996.

[27] G. D. Walker, M. Fischer, J. Gannon, R. C. Thompson, and T.R. Oegema, “Expression of type-X collagen in osteoarthritis,”Journal of Orthopaedic Research, vol. 13, no. 1, pp. 4–12, 1995.

[28] T. Kirsch, B. Swoboda, and H. D. Nah, “Activation ofannexin II and V expression, terminal differentiation, min-eralization and apoptosis in human osteoarthritic cartilage,”Osteoarthritis and Cartilage, vol. 8, no. 4, pp. 294–302, 2000.

[29] D. Pfander, T. Cramer, E. Schipani, and R. S. Johnson,“HIF-1α controls extracellular matrix synthesis by epiphysealchondrocytes,” Journal of Cell Science, vol. 116, no. 9, pp.1819–1826, 2003.

[30] O. Pullig, G. Weseloh, D. L. Ronneberger, S. M. Kakonen,and B. Swoboda, “Chondrocyte differentiation in humanosteoarthritis: expression of osteocalcin in normal andosteoarthritic cartilage and bone,” Calcified Tissue Interna-tional, vol. 67, no. 3, pp. 230–240, 2000.

[31] G. A. Karpouzas and R. A. Terkeltaub, “New developments inthe pathogenesis of articular cartilage calcification,” CurrentRheumatology Reports, vol. 1, no. 2, pp. 121–127, 1999.

[32] K. Johnson and R. Terkeltaub, “Upregulated ank expressionin osteoarthritis can promote both chondrocyte MMP-13expression and calcification via chondrocyte extracellular PPiexcess,” Osteoarthritis and Cartilage, vol. 12, no. 4, pp. 321–335, 2004.

[33] F. J. Blanco, R. Guitian, E. Vazquez-Martul, F. J. de Toro,and F. Galdo, “Osteoarthritis chondrocytes die by apoptosis:a possible pathway for osteoarthritis pathology,” Arthritis andRheumatism, vol. 41, no. 2, pp. 284–289, 1998.

[34] C. M. Robertson, A. T. Pennock, F. L. Harwood, A. C.Pomerleau, R. T. Allen, and D. Amiel, “Characterizationof pro-apoptotic and matrix-degradative gene expressionfollowing induction of osteoarthritis in mature and agedrabbits,” Osteoarthritis and Cartilage, vol. 14, no. 5, pp. 471–476, 2006.

[35] E. Tchetina, F. Mwale, and A. R. Poole, “Distinct phases ofcoordinated early and late gene expression in growth platechondrocytes in relationship to cell proliferation, matrixassembly, remodeling, and cell differentiation,” Journal ofBone and Mineral Research, vol. 18, no. 5, pp. 844–851, 2003.

[36] A. V. Ham and D. H. Cormack, Ham’s Histology, Lippincott,Philadelphia, Pa, USA, 1987.

[37] S. Yamane, E. Cheng, Z. You, and A. H. Reddi, “Geneexpression profiling of mouse articular and growth platecartilage,” Tissue Engineering, vol. 13, no. 9, pp. 2163–2173,2007.

[38] P. Krejci, D. Krakow, P. B. Mekikian, and W. R. Wilcox,“Fibroblast growth factors 1, 2, 17, and 19 are the predom-inant FGF ligands expressed in human fetal growth platecartilage,” Pediatric Research, vol. 61, no. 3, pp. 267–272,2007.

[39] O. Nilsson, E. A. Parker, A. Hegde, M. Chau, K. M. Barnes,and J. Baron, “Gradients in bone morphogenetic protein-related gene expression across the growth plate,” Journal ofEndocrinology, vol. 193, no. 1, pp. 75–84, 2007.

[40] H. C. Anderson, P. T. Hodges, X. M. Aguilera, L. Missana,and P. E. Moylan, “Bone morphogenetic protein (BMP)localization in developing human and rat growth plate,metaphysis, epiphysis, and articular cartilage,” Journal ofHistochemistry and Cytochemistry, vol. 48, no. 11, pp. 1493–1502, 2000.

[41] S. Matsunaga, T. Yamamoto, and K. Fukumura, “Temporaland spatial expressions of transforming growth factor-βs

and their receptors in epiphyseal growth plate,” InternationalJournal of Oncology, vol. 14, no. 6, pp. 1063–1067, 1999.

[42] T. A. Damron, S. Mathur, J. A. Horton et al., “Temporalchanges in PTHrP, Bcl-2, Bax, caspase, TGF-β, and FGF-2 expression following growth plate irradiation with orwithout radioprotectant,” Journal of Histochemistry andCytochemistry, vol. 52, no. 2, pp. 157–167, 2004.

[43] M. P. Verdier, S. Seite, K. Guntzer, J. P. Pujol, and K.Boumediene, “Immunohistochemical analysis of transform-ing growth factor beta isoforms and their receptors in humancartilage from normal and osteoarthritic femoral heads,”Rheumatology International, vol. 25, no. 2, pp. 118–124, 2005.

[44] F. Beier, Z. Ali, D. Mok et al., “TGFβ and PTHrP controlchondrocyte proliferation by activating cyclin D1 expres-sion,” Molecular Biology of the Cell, vol. 12, no. 12, pp. 3852–3863, 2001.

[45] H. I. Roach, G. Mehta, R. O. C. Oreffo, N. M. P. Clarke, andC. Cooper, “Temporal analysis of rat growth plates: cessationof growth with age despite presence of a physis,” Journal ofHistochemistry and Cytochemistry, vol. 51, no. 3, pp. 373–383,2003.

[46] F. Mwale, C. Billinghurst, W. Wu et al., “Selective assemblyand remodelling of collagens II and IX associated withexpression of the chondrocyte hypertrophic phenotype,”Developmental Dynamics, vol. 218, no. 4, pp. 648–662, 2000.

[47] L. A. Neuhold, L. Killar, W. Zhao et al., “Postnatal expres-sion in hyaline cartilage of constitutively active humancollagenase-3 (MMP-13) induces osteoarthritis in mice,”Journal of Clinical Investigation, vol. 107, no. 1, pp. 35–44,2001.

[48] A. L. Armstrong, H. J. Barrach, and M. G. Ehrlich, “Identi-fication of the metalloproteinase stromelysin in the physis,”Journal of Orthopaedic Research, vol. 20, no. 2, pp. 289–294,2002.

[49] F. H. Wezeman and M. R. Bollnow, “Immunohistochemicallocalization of fibroblast growth factor-2 in normal andbrachymorphic mouse tibial growth plate and articularcartilage,” Histochemical Journal, vol. 29, no. 6, pp. 505–514,1997.

[50] A. Horner, P. Kemp, C. Summers et al., “Expression anddistribution of transforming growth factor-β isoforms andtheir signaling receptors in growing human bone,” Bone, vol.23, no. 2, pp. 95–102, 1998.

[51] Y. Wang, S. Nishida, T. Sakata et al., “Insulin-like growthfactor-I is essential for embryonic bone development,”Endocrinology, vol. 147, no. 10, pp. 4753–4761, 2006.

[52] P. de los Rios and D. J. Hill, “Cellular localization andexpression of insulin-like growth factors (IGFS) and IGFbinding proteins within the epiphyseal growth plate ofthe ovine fetus: possible functional implications,” CanadianJournal of Physiology and Pharmacology, vol. 77, no. 4, pp.235–249, 1999.

[53] D. M. Shinar, N. Endo, D. Halperin, G. A. Rodan, andM. Weinreb, “Differential expression of insulin-like growthfactor-I (IGF-I) and IGF-II messenger ribonucleic acid ingrowing rat bone,” Endocrinology, vol. 132, no. 3, pp. 1158–1167, 1993.

[54] Q. Ma, X. Li, D. Vale-Cruz, M. L. Brown, F. Beier, andP. LuValle, “Activating transcription factor 2 controls Bcl-2promoter activity in growth plate chondrocytes,” Journal ofCellular Biochemistry, vol. 101, no. 2, pp. 477–487, 2007.

[55] H. Akiyama, M. C. Chaboissier, J. F. Martin, A. Schedl,and B. de Crombrugghe, “The transcription factor Sox9has essential roles in successive steps of the chondrocyte

10 Arthritis

differentiation pathway and is required for expression of Sox5and Sox6,” Genes and Development, vol. 16, no. 21, pp. 2813–2828, 2002.

[56] D. W. Burton, M. Foster, K. A. Johnson, M. Hiramoto, L.J. Deftos, and R. Terkeltaub, “Chondrocyte calcium-sensingreceptor expression is up-regulated in early guinea pig kneeosteoarthritis and modulates PTHrP, MMP-13, and TIMP-3expression,” Osteoarthritis Cartilage, vol. 13, no. 5, pp. 395–404, 2005.

[57] J. A. Urıa, M. Balbın, J. M. Lopez et al., “Collagenase-3 (MMP-13) expression in chondrosarcoma cells and itsregulation by basic fibroblast growth factor,” AmericanJournal of Pathology, vol. 153, no. 1, pp. 91–101, 1998.

[58] P. Borden, D. Solymar, A. Sucharczuk, B. Lindman, P.Cannon, and R. A. Heller, “Cytokine control of interstitialcollagenase and collagenase-3 gene expression in humanchondrocytes,” Journal of Biological Chemistry, vol. 271, no.38, pp. 23577–23581, 1996.

[59] G. Haeusler, I. Walter, M. Helmreich, and M. Egerbacher,“Localization of matrix metalloproteinases, (MMPs) theirtissue inhibitors, and vascular endothelial growth factor(VEGF) in growth plates of children and adolescents indi-cates a role for MMPs in human postnatal growth andskeletal maturation,” Calcified Tissue International, vol. 76,no. 5, pp. 326–335, 2005.

[60] C. Li, L. Chen, T. Iwata, M. Kitagawa, X. Y. Fu, and C. X.Deng, “A Lys644Glu substitution in fibroblast growth factorreceptor 3 (FGFR3) causes dwarfism in mice by activationof STATs and ink4 cell cycle inhibitors,” Human MolecularGenetics, vol. 8, no. 1, pp. 35–44, 1999.

[61] K. Tsuchimochi, M. Otero, C. L. Dragomir et al., “GADD45βenhances Col10a1 transcription via the MTK1/MKK3/6/p38axis and activation of C/EBPβ-TAD4 in terminally differenti-ating chondrocytes,” Journal of Biological Chemistry, vol. 285,no. 11, pp. 8395–8407, 2010.

[62] K. Ijiri, L. F. Zerbini, H. Peng et al., “A novel role forGADD45β as a mediator of MMP-13 gene expression duringchondrocyte terminal differentiation,” Journal of BiologicalChemistry, vol. 280, no. 46, pp. 38544–38555, 2005.

[63] M. D. Kinkel, R. Yagi, D. McBurney, A. Nugent, and W.E. Horton Jr., “Age-related expression patterns of Bag-1 and Bcl-2 in growth plate and articular chondrocytes,”Anatomical Record Part A, vol. 279, no. 2, pp. 720–728, 2004.

[64] J. P. Tuckermann, K. Pittois, N. C. Partridge, J. Merregaert,and P. Angel, “Collagenase-3 (MMP-13) and integral mem-brane protein 2a (Itm2a) are marker genes of chondro-genic/osteoblastic cells in bone formation: sequential tem-poral, and spatial expression of Itm2a, alkaline phosphatase,MMP-13, and osteocalcin in the mouse,” Journal of Bone andMineral Research, vol. 15, no. 7, pp. 1257–1265, 2000.

[65] F. M. D. Henson, M. E. Davies, J. N. Skepper, and L.B. Jeffcott, “Localisation of alkaline phosphatase in equinegrowth cartilage,” Journal of Anatomy, vol. 187, no. 1, pp.151–159, 1995.

[66] R. Pogue and K. Lyons, “BMP signaling in the cartilagegrowth plate,” Current Topics in Developmental Biology, vol.76, pp. 1–48, 2006.

[67] Y. Wang, F. Middleton, J. A. Horton, L. Reichel, C. E. Farnum,and T. A. Damron, “Microarray analysis of proliferative andhypertrophic growth plate zones identifies differentiationmarkers and signal pathways,” Bone, vol. 35, no. 6, pp. 1273–1293, 2004.

[68] T. Sakou, T. Onishi, T. Yamamoto, T. Nagamine, T. K.Sampath, and P. Ten Dijke, “Localization of Smads, the

TGF-β family intracellular signaling components duringendochondral ossification,” Journal of Bone and MineralResearch, vol. 14, no. 7, pp. 1145–1152, 1999.

[69] T. Fukunaga, T. Yamashiro, S. Oya, N. Takeshita, M. Taki-gawa, and T. Takano-Yamamoto, “Connective tissue growthfactor mRNA expression pattern in cartilages is associatedwith their type I collagen expression,” Bone, vol. 33, no. 6,pp. 911–918, 2003.

[70] J. Alvarez, L. Costales, R. Serra, M. Balbın, and J. M.Lopez, “Expression patterns of matrix metalloproteinasesand vascular endothelial growth factor during epiphysealossification,” Journal of Bone and Mineral Research, vol. 20,no. 6, pp. 1011–1021, 2005.

[71] H. E. MacLean and H. M. Kronenberg, “Localization ofIndian hedgehog and PTH/PTHrP receptor expression inrelation to chondrocyte proliferation during mouse bonedevelopment,” Development Growth and Differentiation, vol.47, no. 2, pp. 59–63, 2005.

[72] J. M. Kindblom, O. Nilsson, T. Hurme, C. Ohlsson, and L.Savendahl, “Expression and localization of Indian hedgehog(Ihh) and parathyroid hormone related protein (PTHrP)in the human growth plate during pubertal development,”Journal of Endocrinology, vol. 174, no. 2, pp. R1–R6, 2002.

[73] F. Yamashita, K. Sakakida, K. Kusuzaki, H. Takeshita, andA. Kuzuhara, “Immunohistochemical localization of inter-leukin 1 in human growth cartilage,” Nippon SeikeigekaGakkai Zasshi, vol. 63, no. 5, pp. 562–568, 1989.

[74] P. Ducy, R. Zhang, V. Geoffroy, A. L. Ridall, and G. Karsenty,“Osf2/Cbfa1: a transcriptional activator of osteoblast differ-entiation,” Cell, vol. 89, no. 5, pp. 747–754, 1997.

[75] P. M. van der Kraan, E. N. Blaney Davidson, A. Blom, andW. B. van den Berg, “TGF-beta signaling in chondrocyteterminal differentiation and osteoarthritis. Modulation andintegration of signaling pathways through receptor-Smads,”Osteoarthritis and Cartilage, vol. 17, no. 12, pp. 1539–1545,2009.

[76] T. Komori, “A fundamental transcription factor for bone andcartilage,” Biochemical and Biophysical Research Communica-tions, vol. 276, no. 3, pp. 813–816, 2000.

[77] S. Takeda, J. P. Bonnamy, M. J. Owen, P. Ducy, and G.Karsenty, “Continuous expression of Cbfa1 in nonhyper-trophic chondrocytes uncovers its ability to induce hyper-trophic chondrocyte differentiation and partially rescuesCbfa1-deficient mice,” Genes and Development, vol. 15, no.4, pp. 467–481, 2001.

[78] C. Ueta, M. Iwamoto, N. Kanatani et al., “Skeletal malfor-mations caused by overexpression of Cbfa1 or its dominantnegative form in chondrocytes,” Journal of Cell Biology, vol.153, no. 1, pp. 87–99, 2001.

[79] D. Porte, J. Tuckermann, M. Becker et al., “Both AP-1 andCbfa1-like factors are required for the induction of interstitialcollagenase by parathyroid hormone,” Oncogene, vol. 18, no.3, pp. 667–678, 1999.

[80] M. J. G. Jimenez, M. Balbın, J. M. Lopez, J. Alvarez, T.Komori, and C. Lopez-Otın, “Collagenase 3 is a target ofCbfa1, a transcription factor of the runt gene family involvedin bone formation,” Molecular and Cellular Biology, vol. 19,no. 6, pp. 4431–4442, 1999.

[81] F. Mwale, E. Tchetina, C. W. Wu, and A. R. Poole, “Theassembly and remodeling of the extracellular matrix inthe growth plate in relationship to mineral deposition andcellular hypertrophy: an in situ study of collagens II and IXand proteoglycan,” Journal of Bone and Mineral Research, vol.17, no. 2, pp. 275–283, 2002.

Arthritis 11

[82] W. Wang, J. Xu, B. Du, and T. Kirsch, “Role of the progressiveankylosis gene (ank) in cartilage mineralization,” Molecularand Cellular Biology, vol. 25, no. 1, pp. 312–323, 2005.

[83] W. Wang, J. Xu, and T. Kirsch, “Annexin-mediated Ca2+

influx regulates growth plate chondrocyte maturation andapoptosis,” Journal of Biological Chemistry, vol. 278, no. 6, pp.3762–3769, 2003.

[84] T. Kirsch, “Annexins—their role in cartilage mineralization,”Frontiers in Bioscience, vol. 10, pp. 576–581, 2005.

[85] W. Wang, J. Xu, and T. Kirsch, “Annexin V and terminaldifferentiation of growth plate chondrocytes,” ExperimentalCell Research, vol. 305, no. 1, pp. 156–165, 2005.

[86] R. K. Gill, R. T. Turner, T. J. Wronski, and N. H. Bell,“Orchiectomy markedly reduces the concentration of thethree isoforms of transforming growth factor β in rat bone,and reduction is prevented by testosterone,” Endocrinology,vol. 139, no. 2, pp. 546–550, 1998.

[87] I. M. Shapiro, C. S. Adams, T. Freeman, and V. Srinivas,“Fate of the hypertrophic chondrocyte: microenvironmentalperspectives on apoptosis and survival in the epiphysealgrowth plate,” Birth Defects Research Part C, vol. 75, no. 4,pp. 330–339, 2005.

[88] R. C. Olney, J. Wang, J. E. Sylvester, and E. B. Mougey,“Growth factor regulation of human growth plate chon-drocyte proliferation in vitro,” Biochemical and BiophysicalResearch Communications, vol. 317, no. 4, pp. 1171–1182,2004.

[89] M. R. Hutchison, M. H. Bassett, and P. C. White, “Insulin-likegrowth factor-I and fibroblast growth factor, but not growthhormone, affect growth plate chondrocyte proliferation,”Endocrinology, vol. 148, no. 7, pp. 3122–3130, 2007.

[90] H. M. Kronenberg, “PTHrP and skeletal development,”Annals of the New York Academy of Sciences, vol. 1068, no. 1,pp. 1–13, 2006.

[91] T. F. Li, Y. Dong, A. M. Ionescu et al., “Parathyroid hormone-related peptide (PTHrP) inhibits Runx2 expression throughthe PKA signaling pathway,” Experimental Cell Research, vol.299, no. 1, pp. 128–136, 2004.

[92] Y. F. Dong, Y. Soung do, E. M. Schwarz, R. J. O’Keefe,and H. Drissi, “Wnt induction of chondrocyte hypertrophythrough the Runx2 transcription factor,” Journal of CellularPhysiology, vol. 208, no. 1, pp. 77–86, 2006.

[93] D. Kiepe, S. Ciarmatori, A. Haarmann, and B. Tonshoff,“Differential expression of IGF system components in pro-liferating vs. differentiating growth plate chondrocytes: thefunctional role of IGFBP-5,” American Journal of Physiology,vol. 290, no. 2, pp. E363–E371, 2006.

[94] A. C. Karaplis, A. Luz, J. Glowacki et al., “Lethal skeletal dys-plasia from targeted disruption of the parathyroid hormone-related peptide gene,” Genes and Development, vol. 8, no. 3,pp. 277–289, 1994.

[95] N. Amizuka, H. Warshawsky, J. E. Henderson, D. Goltzman,and A. C. Karaplis, “Parathyroid hormone-related peptide-depleted mice show abnormal epiphyseal cartilage develop-ment and altered endochondral bone formation,” Journal ofCell Biology, vol. 126, no. 6, pp. 1611–1623, 1994.

[96] E. C. Weir, W. M. Philbrick, M. Amling, L. A. Neff,R. Baron, and A. E. Broadus, “Targeted overexpressionof parathyroid hormone-related peptide in chondrocytescauses chondrodysplasia and delayed endochondral boneformation,” Proceedings of the National Academy of Sciences ofthe United States of America, vol. 93, no. 19, pp. 10240–10245,1996.

[97] B. Lanske, A. C. Karaplis, K. Lee et al., “PTH/PTHrP receptorin early development and Indian hedgehog-regulated bonegrowth,” Science, vol. 273, no. 5275, pp. 663–666, 1996.

[98] A. Vortkamp, K. Lee, B. Lanske, G. V. Segre, H. M.Kronenberg, and C. J. Tabin, “Regulation of rate of cartilagedifferentiation by Indian hedgehog and PTH-related pro-tein,” Science, vol. 273, no. 5275, pp. 613–622, 1996.

[99] E. Yoshida, M. Noshiro, T. Kawamoto, S. Tsutsumi, Y.Kuruta, and Y. Kato, “Direct inhibition of Indian hedgehogexpression by parathyroid hormone (PTH)/PTH-relatedpeptide and up-regulation by retinoic acid in growth platechondrocyte cultures,” Experimental Cell Research, vol. 265,no. 1, pp. 64–72, 2001.

[100] B. St-Jacques, M. Hammerschmidt, and A. P. McMahon,“Indian hedgehog signaling regulates proliferation anddifferentiation of chondrocytes and is essential for boneformation,” Genes and Development, vol. 13, no. 16, pp.2072–2086, 1999.

[101] M. Pacifici, T. Shimo, C. Gentili et al., “Syndecan-3: a cell-surface heparan sulfate proteoglycan important for chondro-cyte proliferation and function during limb skeletogenesis,”Journal of Bone and Mineral Metabolism, vol. 23, no. 3, pp.191–199, 2005.

[102] A. B. Roberts and M. B. Sporn, “Physiological actions andclinical applications of transforming growth factor-β (TGF-β),” Growth Factors, vol. 8, no. 1, pp. 1–9, 1993.

[103] G. Gibson, “Active role of chondrocyte apoptosis in endo-chondral ossification,” Microscopy Research and Technique,vol. 43, no. 2, pp. 191–204, 1998.

[104] X. Yang, L. Chen, X. Xu, C. Li, C. Huang, and C. X. Deng,“TGF-β/Smad3 signals repress chondrocyte hypertrophicdifferentiation and are required for maintaining articularcartilage,” Journal of Cell Biology, vol. 153, no. 1, pp. 35–46,2001.

[105] R. Serra, A. Karaplis, and P. Sohn, “Parathyroid hormone-related peptide (PTHrP)-dependent and -independenteffects of transforming growth factor β (TGF-β)onendochondral bone formation,” Journal of Cell Biology,vol. 145, no. 4, pp. 783–794, 1999.

[106] C. M. Ferguson, E. M. Schwarz, P. R. Reynolds, J. E.Puzas, R. N. Rosier, and R. J. O’Keefe, “Smad2 and 3mediate transforming growth factor-β1-induced inhibitionof chondrocyte maturation,” Endocrinology, vol. 141, no. 12,pp. 4728–4735, 2000.

[107] R. T. Ballock, A. Heydemann, L. M. Wakefield, K. C.Flanders, A. B. Roberts, and M. B. Sporn, “TGF-β1 preventshypertrophy of epiphyseal chondrocytes: regulation of geneexpression for cartilage matrix proteins and metallopro-teases,” Developmental Biology, vol. 158, no. 2, pp. 414–429,1993.

[108] K. Bohme, K. H. Winterhhlter, and P. Bruckner, “Terminaldifferentiation of chondrocytes in culture is a spontaneousprocess and is arrested by transforming growth factor-β2 andbasic fibroblast growth factor in synergy,” Experimental CellResearch, vol. 216, no. 1, pp. 191–198, 1995.

[109] R. A. Terkeltaub, K. Johnson, D. Rohnow, R. Goomer, D.Burton, and L. J. Deftos, “Bone morphogenetic proteins andbFGF exert opposing regulatory effects on PTHrP expressionand inorganic pyrophosphate elaboration in immortalizedmurine endochondral hypertrophic chondrocytes (MCTcells),” Journal of Bone and Mineral Research, vol. 13, no. 6,pp. 931–941, 1998.

12 Arthritis

[110] D. B. Pateder, R. N. Rosier, E. M. Schwarz et al., “PTHrPexpression in chondrocytes, regulation by TGF-β, and inter-actions between epiphyseal and growth plate chondrocytes,”Experimental Cell Research, vol. 256, no. 2, pp. 555–562, 2000.

[111] H. Nagai and M. Aoki, “Inhibition of growth plate angiogen-esis and endochondral ossification with diminished expres-sion of MMP-13 in hypertrophic chondrocytes in FGF-2-treated rats,” Journal of Bone and Mineral Metabolism, vol. 20,no. 3, pp. 142–147, 2002.

[112] V. Szuts, U. Mollers, K. Bittner et al., “Terminal differentia-tion of chondrocytes is arrested at distinct stages identified bytheir expression repertoire of marker genes,” Matrix Biology,vol. 17, no. 6, pp. 435–448, 1998.

[113] Q. Wang, X. Wei, T. Zhu et al., “Bone morphogenetic protein2 activates Smad6 gene transcription through bone-specifictranscription factor Runx2,” Journal of Biological Chemistry,vol. 282, no. 14, pp. 10742–10748, 2007.

[114] T. Kobayashi, K. M. Lyons, A. P. McMahon, and H. M. Kro-nenberg, “BMP signaling stimulates cellular differentiationat multiple steps during cartilage development,” Proceedingsof the National Academy of Sciences of the United States ofAmerica, vol. 102, no. 50, pp. 18023–18027, 2005.

[115] B. S. Yoon, R. Pogue, D. A. Ovchinnikov et al., “BMPsregulate multiple aspects of growth-plate chondrogenesisthrough opposing actions on FGF pathways,” Development,vol. 133, no. 23, pp. 4667–4678, 2006.

[116] M. C. Stewart, R. M. Kadlcek, P. D. Robbins, J. N. Macleod,and R. T. Ballock, “Expression and activity of the CDKinhibitor p57Kip2 in chondrocytes undergoing hypertrophicdifferentiation,” Journal of Bone and Mineral Research, vol. 19,no. 1, pp. 123–132, 2004.

[117] D. Zhang, E. M. Schwarz, R. N. Rosier, M. J. Zuscik, J. E.Puzas, and R. J. O’Keefe, “ALK2 functions as a BMP typeI receptor and induces Indian hedgehog in chondrocytesduring skeletal development,” Journal of Bone and MineralResearch, vol. 18, no. 9, pp. 1593–1604, 2003.

[118] C. D. Grimsrud, P. R. Romano, M. D’Souza et al., “BMP-6 is an autocrine stimulator of chondrocyte differentiation,”Journal of Bone and Mineral Research, vol. 14, no. 4, pp. 475–482, 1999.

[119] T. Takahashi, E. A. Morris, and S. B. Trippel, “Bonemorphogenetic protein-2 and -9 regulate the interaction ofinsulin-like growth factor-I with growth plate chondrocytes,”International Journal of Molecular Medicine, vol. 20, no. 1, pp.53–57, 2007.

[120] G. Mailhot, M. Yang, A. Mason-Savas, C. A. Mackay,I. Leav, and P. R. Odgren, “BMP-5 expression increasesduring chondrocyte differentiation in vivo and in vitro andpromotes proliferation and cartilage matrix synthesis inprimary chondrocyte cultures,” Journal of Cellular Physiology,vol. 214, no. 1, pp. 56–64, 2008.

[121] Z. Laron, “Insulin-like growth factor 1 (IGF-1): a growthhormone,” Molecular Pathology, vol. 54, no. 5, pp. 311–316,2001.

[122] W. Hui, A. D. Rowan, and T. Cawston, “Insulin-like growthfactor 1 blocks collagen release and down regulates matrixmetalloproteinase-1, -3, -8, and -13 mRNA expression inbovine nasal cartilage stimulated with oncostatin M incombination with interleukin 1α,” Annals of the RheumaticDiseases, vol. 60, no. 3, pp. 254–261, 2001.

[123] V. Lefebvre and P. Smits, “Transcriptional control of chon-drocyte fate and differentiation,” Birth Defects Research PartC, vol. 75, no. 3, pp. 200–212,2005.

[124] T. Hattori, C. Muller, S. Gebhard et al., “SOX9 is a majornegative regulator of cartilage vascularization, bone marrowformation and endochondral ossification,” Development, vol.137, no. 6, pp. 901–911, 2010.

[125] M. Wuelling and A. Vortkamp, “Transcriptional networkscontrolling chondrocyte proliferation and differentiationduring endochondral ossification,” Pediatric Nephrology, vol.25, no. 4, pp. 625–631, 2010.

[126] Q. Zhao, H. Eberspaecher, V. Lefebvre, and B. de Crom-brugghe, “Parallel expression of Sox9 and Col2a1 in cellsundergoing chondrogenesis,” Developmental Dynamics, vol.209, no. 4, pp. 377–386, 1997.

[127] B. de Crombrugghe, V. Lefebvre, R. R. Behringer, W. Bi, S.Murakami, and W. Huang, “Transcriptional mechanisms ofchondrocyte differentiation,” Matrix Biology, vol. 19, no. 5,pp. 389–394, 2000.

[128] L. J. Ng, S. Wheatley, G. E. O. Muscat et al., “SOX9binds DNA, activates transcription, and coexpresses withtype II collagen during chondrogenesis in the mouse,”Developmental Biology, vol. 183, no. 1, pp. 108–121, 1997.

[129] W. Bi, W. Huang, D. J. Whitworth et al., “Haploinsufficiencyof Sox9 results in defective cartilage primordia and pre-mature skeletal mineralization,” Proceedings of the NationalAcademy of Sciences of the United States of America, vol. 98,no. 12, pp. 6698–6703, 2001.

[130] S. Reppe, E. Rian, R. Jemtland, O. K. Olstad, V. T. Gautvik,and K. M. Gautvik, “Sox-4 messenger RNA is expressed inthe embryonic growth plate and regulated via the parathyroidhormone/parathyroid hormone-related protein receptor inosteoblast-like cells,” Journal of Bone and Mineral Research,vol. 15, no. 12, pp. 2402–2412, 2000.

[131] Y. Soung do, Y. Dong, Y. Wang et al., “Runx3/AML2/Cbfa3regulates early and late chondrocyte differentiation,” Journalof Bone and Mineral Research, vol. 22, no. 8, pp. 1260–1270,2007.

[132] L. A. Solomon, N. G. Berube, and F. Beier, “Transcrip-tional regulators of chondrocyte hypertrophy,” Birth DefectsResearch Part C, vol. 84, no. 2, pp. 123–130, 2008.

[133] H. Drissi, M. Zuscik, R. Rosier, and R. O’Keefe, “Tran-scriptional regulation of chondrocyte maturation: potentialinvolvement of transcription factors in OA pathogenesis,”Molecular Aspects of Medicine, vol. 26, no. 3, pp. 169–179,2005.

[134] T. F. Li, M. J. Zuscik, A. M. Ionescu et al., “PGE2 inhibitschondrocyte differentiation through PKA and PKC signal-ing,” Experimental Cell Research, vol. 300, no. 1, pp. 159–169,2004.

[135] X. Zhang, N. Ziran, J. J. Goater et al., “Primary murinelimb bud mesenchymal cells in long-term culture completechondrocyte differentiation: TGF-β delays hypertrophy andPGE2 inhibits terminal differentiation,” Bone, vol. 34, no. 5,pp. 809–817, 2004.

[136] Z. Schwartz, R. M. Gilley, V. L. Sylvia, D. D. Dean, and B.D. Boyan, “The effect of prostaglandin E2 on costochondralchondrocyte differentiation is mediated by cyclic adenosine3’,5’-monophosphate and protein kinase C,” Endocrinology,vol. 139, no. 4, pp. 1825–1834, 1998.

[137] C. Brochhausen, P. Neuland, C. J. Kirkpatrick, R. M.Nusing, and G. Klaus, “Cyclooxygenases and prostaglandinE2 receptors in growth plate chondrocytes in vitro and insitu—prostaglandin E2 dependent proliferation of growthplate chondrocytes,” Arthritis Research & Therapy, vol. 8, no.3, article R78, 2006.

Arthritis 13

[138] L. Wei, K. Kanbe, M. Lee et al., “Stimulation of chondrocytehypertrophy by chemokine stromal cell-derived factor 1 inthe chondro-osseous junction during endochondral boneformation,” Developmental Biology, vol. 341, no. 1, pp. 236–245, 2010.

[139] W. Wang and T. Kirsch, “Annexin V/β5 integrin interactionsregulate apoptosis of growth plate chondrocytes,” Journal ofBiological Chemistry, vol. 281, no. 41, pp. 30848–30856, 2006.

[140] M. S. Domowicz, M. Cortes, J. G. Henry, and N. B. Schwartz,“Aggrecan modulation of growth plate morphogenesis,”Developmental Biology, vol. 329, no. 2, pp. 242–257, 2009.

[141] L. van der Weyden, L. Wei, J. Luo et al., “Functional knockoutof the matrilin-3 gene causes premature chondrocyte mat-uration to hypertrophy and increases bone mineral densityand osteoarthritis,” American Journal of Pathology, vol. 169,no. 2, pp. 515–527, 2006.

[142] O. Barbieri, S. Astigiano, M. Morini et al., “Depletion ofcartilage collagen fibrils in mice carrying a dominant negativeCol2a1 transgene affects chondrocyte differentiation,” Amer-ican Journal of Physiology, vol. 285, no. 6, pp. C1504–C1512,2003.

[143] R. M. Borzi., E. Olivotto, S. Pagani et al., “Matrix metal-loproteinase 13 loss associated with impaired extracellularmatrix remodeling disrupts chondrocyte differentiation byconcerted effects on multiple regulatory factors,” Arthritisand Rheumatism, vol. 62, no. 8, pp. 2370–2381, 2010.

[144] C. J. Malemud, “Matrix metalloproteinases: role in skeletaldevelopment and growth plate disorders,” Frontiers in Bio-science, vol. 11, no. 2, pp. 1702–1715, 2006.

[145] Q. Chen, D. M. Johnson, D. R. Haudenschild, and P. F.Goetinck, “Progression and recapitulation of the chondro-cyte differentiation program: cartilage matrix protein is amarker for cartilage maturation,” Developmental Biology, vol.172, no. 1, pp. 293–306, 1995.

[146] J. A. Buckwalter and H. J. Mankin, “Articular cartilage repairand transplantation,” Arthritis and Rheumatism, vol. 41, no.8, pp. 1331–1342, 1998.

[147] T. E. Hardingham, A. J. Fosang, and J. Dudhia, “Aggrecan, thechondroitin sulfate/keratan sulfate proteoglycan from carti-lage,” in Articular Cartilage and Osteoarthritis, K. Kuettner etal., Ed., Raven Press, New York, NY, USA, 1992.

[148] T. Aigner, A. Zien, D. Hanisch, and R. Zimmer, “Geneexpression in chondrocytes assessed with use of microarrays,”Journal of Bone and Joint Surgery. American, vol. 85, no. 1, pp.117–123, 2003.

[149] L. C. Tetlow, D. J. Adlam, and D. E. Woolley, “Matrix met-alloproteinase and proinflammatory cytokine production bychondrocytes of human osteoarthritic cartilage: associationswith degenerative changes,” Arthritis and Rheumatism, vol.44, no. 3, pp. 585–594, 2001.

[150] W. B. van den Berg, “Growth factors in experimentalosteoarthritis: transforming growth factor β pathogenic?”Journal of Rheumatology, vol. 22, no. 43, pp. 143–145, 1995.

[151] X. Chen, C. M. Macica, A. Nasiri, and A. E. Broadus,“Regulation of articular chondrocyte proliferation and dif-ferentiation by Indian hedgehog and parathyroid hormone-related protein in mice,” Arthritis and Rheumatism, vol. 58,no. 12, pp. 3788–3797, 2008.

[152] Y. Hiraki, H. Inoue, K. I. Iyama et al., “Identificationof chondromodulin I as a novel endothelial cell growthinhibitor: purification and its localization in the avascularzone of epiphyseal cartilage,” Journal of Biological Chemistry,vol. 272, no. 51, pp. 32419–32426, 1997.

[153] H. Kitahara, T. Hayami, K. Tokunaga et al., “Chondro-modulin-I expression in rat articular cartilage,” Archives ofHistology and Cytology, vol. 66, no. 3, pp. 221–228, 2003.

[154] K. Ijiri, L. F. Zerbini, H. Peng et al., “Differential expression ofGADD45β in normal and osteoarthritic cartilage: potentialrole in homeostasis of articular chondrocytes,” Arthritis andRheumatism, vol. 58, no. 7, pp. 2075–2087, 2008.

[155] H. W. Zhou, S. Q. Lou, and K. Zhang, “Recovery of functionin osteoarthritic chondrocytes induced by p16INK4a-specificsiRNA in vitro,” Rheumatology, vol. 43, pp. 555–568, 2004.

[156] T. Aigner and L. McKenna, “Molecular pathology and patho-biology of osteoarthritic cartilage,” Cellular and MolecularLife Sciences, vol. 59, no. 1, pp. 5–18, 2002.

[157] A. R. Poole, “What type of cartilage repair are we attemptingto attain?” Journal of Bone and Joint Surgery. American, vol.85, no. 1, pp. 40–44, 2003.

[158] A. Scharstuhl, H. L. Glansbeek, H. M. van Beuningen,E. L. Vitters, P. M. van der Kraan, and W. B. van denBerg, “Inhibition of endogenous TGF-β during experimentalosteoarthritis prevents osteophyte formation and impairscartilage repair,” Journal of Immunology, vol. 169, no. 1, pp.507–514, 2002.

[159] P. K. Bos, G. J. V. M. van Osch, D. A. Frenz, J. A.N. Verhaar, and H. L. Verwoerd-Verhoef, “Growth factorexpression in cartilage wound healing: temporal and spatialimmunolocalization in a rabbit auricular cartilage woundmodel,” Osteoarthritis and Cartilage, vol. 9, no. 4, pp. 382–389, 2001.

[160] A. H. White, R. E. B. Watson, B. Newman, A. J. Freemont,and G. A. Wallis, “Annexin VIII is differentially expressedby chondrocytes in the mammalian growth plate duringendochondral ossification and in osteoarthritic cartilage,”Journal of Bone and Mineral Research, vol. 17, no. 10, pp.1851–1858, 2002.

[161] G. E. Kempson, H. Muir, C. Pollard, and M. Tuke, “Thetensile properties of the cartilage of human femoral condylesrelated to the content of collagen and glycosaminoglycans,”Biochimica et Biophysica Acta, vol. 297, no. 2, pp. 456–472,1973.

[162] A. R. Poole, L. C. Rosenberg, A. Reiner, M. Ionescu,E. Bogoch, and P. J. Roughley, “Contents and distribu-tions of the proteoglycans decorin and biglycan in normaland osteoarthritic human articular cartilage,” Journal ofOrthopaedic Research, vol. 14, no. 5, pp. 681–689, 1996.

[163] D. Pfander, B. Swoboda, and T. Kirsch, “Expression of earlyand late differentiation markers (proliferating cell nuclearantigen, syndecan-3, annexin VI, and alkaline phosphatase)by human osteoarthritic chondrocytes,” American Journal ofPathology, vol. 159, no. 5, pp. 1777–1783, 2001.

[164] W. Wu, R. C. Billinghurst, I. Pidoux et al., “Sites ofcollagenase cleavage and denaturation of type II collagenin aging and osteoarthritic articular cartilage and theirrelationship to the distribution of matrix metalloproteinase 1and matrix metalloproteinase 13,” Arthritis and Rheumatism,vol. 46, no. 8, pp. 2087–2094, 2002.

[165] J. Quintavalla, C. Kumar, S. Daouti, E. Slosberg, and S. Uziel-Fusi, “Chondrocyte cluster formation in agarose cultures as afunctional assay to identify genes expressed in osteoarthritis,”Journal of Cellular Physiology, vol. 204, no. 2, pp. 560–566,2005.

[166] A. Maroudas, M. T. Bayliss, and M. F. Venn, “Further studieson the composition of human femoral head cartilage,” Annalsof the Rheumatic Diseases, vol. 39, no. 5, pp. 514–523, 1980.

14 Arthritis

[167] S. A. Semevolos, M. L. Strassheim, J. L. Haupt, and A. J.Nixon, “Expression patterns of hedgehog signaling peptidesin naturally acquired equine osteochondrosis,” Journal ofOrthopaedic Research, vol. 23, no. 5, pp. 1152–1159, 2005.

[168] T. Aigner, M. Hemmel, D. Neureiter et al., “Apoptoticcell death is not a widespread phenomenon in normalaging and osteoarthritic human articular knee cartilage: astudy of proliferation, programmed cell death (apoptosis),and viability of chondrocytes in normal and osteoarthritichuman knee cartilage,” Arthritis and Rheumatism, vol. 44, no.6, pp. 1304–1312, 2001.

[169] K. Kuhn, D. D. D’Lima, S. Hashimoto, and M. Lotz, “Celldeath in cartilage,” Osteoarthritis and Cartilage, vol. 12, no. 1,pp. 1–16, 2004.

[170] H. Dumond, N. Presle, P. Pottie et al., “Site specific changesin gene expression and cartilage metabolism during earlyexperimental osteoarthritis,” Osteoarthritis and Cartilage,vol. 12, no. 4, pp. 284–295, 2004.

[171] J. C. Fernandes, J. Martel-Pelletier, V. Lascau-Coman etal., “Collagenase-1 and collagenase-3 synthesis in normaland early experimental osteoarthritic canine cartilage: animmunohistochemical study,” Journal of Rheumatology, vol.25, no. 8, pp. 1585–1594, 1998.

[172] E. V. Tchetina, M. Kobayashi, T. Yasuda, T. Meijers, I.Pidoux, and A. R. Poole, “Chondrocyte hypertrophy can beinduced by a cryptic sequence of type II collagen and isaccompanied by the induction of MMP-13 and collagenaseactivity: implications for development and arthritis,” MatrixBiology, vol. 26, no. 4, pp. 247–258, 2007.

[173] H. Lorenz, W. Wenz, M. Ivancic, E. Steck, and W. Richter,“Early and stable upregulation of collagen type II, collagentype I and YKL40 expression levels in cartilage duringearly experimental osteoarthritis occurs independent of jointlocation and histological grading,” Arthritis Research &Therapy, vol. 7, no. 1, pp. R156–R165, 2005.

[174] J. R. Matyas, P. F. Ehlers, D. Huang, and M. E. Adams, “Theearly molecular natural history of experimental osteoarthri-tis: I. Progressive discoordinate expression of aggrecan andtype II procollagen messenger rna in the articular cartilage ofadult animals,” Arthritis and Rheumatism, vol. 42, no. 5, pp.993–1002, 1999.

[175] J. R. Matyas, D. Huang, M. Chung, and M. E. Adams,“Regional quantification of cartilage type II collagen andaggrecan messenger RNA in joints with early experimentalosteoarthritis,” Arthritis and Rheumatism, vol. 46, no. 6, pp.1536–1543, 2002.

[176] C. T. G. Appleton, D. D. McErlain, V. Pitelka et al., “Forcedmobilization accelerates pathogenesis: characterization ofa preclinical surgical model of osteoarthritis,” ArthritisResearch & Therapy, vol. 9, article R13, 2007.

[177] F. Moldovan, J. Pelletier, F. C. Jolicoeur, J. M. Cloutier, andJ. Martel-Pelletier, “Diacerhein and rhein reduce the ICE-induced IL-1β and IL-18 activation in human osteoarthriticcartilage,” Osteoarthritis and Cartilage, vol. 8, no. 3, pp. 186–196, 2000.

[178] B. Rolauffs, J. M. Williams, M. Aurich, A. J. Grodzinsky,K. E. Kuettner, and A. A. Cole, “Proliferative remodeling ofthe spatial organization of human superficial chondrocytesdistant from focal early osteoarthritis,” Arthritis and Rheuma-tism, vol. 62, no. 2, pp. 489–498, 2010.

[179] N. Miosge, M. Hartmann, C. Maelicke, and R. Herken,“Expression of collagen type I and type II in consecutivestages of human osteoarthritis,” Histochemistry and CellBiology, vol. 122, no. 3, pp. 229–236, 2004.

[180] A. P. Hollander, I. Pidoux, A. Reiner, C. Rorabeck, R. Bourne,and A. R. Poole, “Damage to type II collagen in aging andosteoarthritis starts at the articular surface, originates aroundchondrocytes, and extends into the cartilage with progressivedegeneration,” Journal of Clinical Investigation, vol. 96, no. 6,pp. 2859–2869, 1995.

[181] A. R. Poole, M. Alini, and A. P. Hollander, “Cellular biologyof cartilage degradation,” in Mechanisms and Models inRheumatoid Arthritis, pp. 163–203, Academic Press, London,UK, 1995.

[182] T. Aigner, A. Zien, A. Gehrsitz, P. M. Gebhard, and L.McKenna, “Anabolic and catabolic gene expression patternanalysis in normal versus osteoarthritic cartilage using com-plementary DNA-array technology,” Arthritis and Rheuma-tism, vol. 44, no. 12, pp. 2777–2789, 2001.

[183] M. Gebauer, J. Saas, J. Haag et al., “Repression of anti-proliferative factor Tob1 in osteoarthritic cartilage,” ArthritisResearch & Therapy, vol. 7, no. 2, pp. R274–R284, 2005.

[184] A. C. Bay-Jensen, T. L. Andersen, N. Charni-Ben Tabassiet al., “Biochemical markers of type II collagen breakdownand synthesis are positioned at specific sites in humanosteoarthritic knee cartilage,” Osteoarthritis and Cartilage,vol. 16, no. 5, pp. 615–623, 2008.

[185] D. Pfander, R. Rahmanzadeh, and E. E. Scheller, “Presenceand distribution of collagen II, collagen I, fibronectin,and tenascin in rabbit normal and osteoarthritic cartilage,”Journal of Rheumatology, vol. 26, no. 2, pp. 386–394, 1999.

[186] K. Veje, J. L. Hyllested-Winge, and K. Ostergaard, “Topo-graphic and zonal distribution of tenascin in human articularcartilage from femoral heads: normal versus mild and severeosteoarthritis,” Osteoarthritis and Cartilage, vol. 11, no. 3, pp.217–227, 2003.

[187] H. Lorenz and W. Richter, “Osteoarthritis: cellular andmolecular changes in degenerating cartilage,” Progress inHistochemistry and Cytochemistry, vol. 40, no. 3, pp. 135–163,2006.

[188] E. N. Blaney Davidson, E. L. Vitters, P. M. van der Kraan,and W. B. van den Berg, “Expression of transforming growthfactor-β (TGFβ) and the TGFβ signalling molecule SMAD-2P in spontaneous and instability-induced osteoarthritis:role in cartilage degradation, chondrogenesis and osteophyteformation,” Annals of the Rheumatic Diseases, vol. 65, no. 11,pp. 1414–1421, 2006.

[189] N. Yatsugi, T. Tsukazaki, M. Osaki, T. Koji, S. Yamashita, andH. Shindo, “Apoptosis of articular chondrocytes in rheuma-toid arthritis and osteoarthritis: correlation of apoptosis withdegree of cartilage destruction and expression of apoptosis-related proteins of p53 and c-myc,” Journal of OrthopaedicScience, vol. 5, no. 2, pp. 150–156, 2000.

[190] P. Garnero, X. Ayral, J. C. Rousseau et al., “Uncoupling oftype II collagen synthesis and degradation predicts progres-sion of joint damage in patients with knee osteoarthritis,”Arthritis and Rheumatism, vol. 46, no. 10, pp. 2613–2624,2002.

[191] E. Gomez-Barrena, O. Sanchez-Pernaute, R. Largo, E. Calvo,P. Esbrit, and G. Herrero-Beaumont, “Sequential changes ofparathyroid hormone related protein (PTHrP) in articularcartilage during progression of inflammatory and degener-ative arthritis,” Annals of the Rheumatic Diseases, vol. 63, no.8, pp. 917–922, 2004.

[192] M. Pombo-Suarez, M. T. Castano-Oreja, M. Calaza, J.Gomez-Reino, and A. Gonzalez, “Differential upregulation ofthe three transforming growth factor beta isoforms in human

Arthritis 15

osteoarthritic cartilage,” Annals of the Rheumatic Diseases,vol. 68, no. 4, pp. 568–571, 2009.

[193] C. J. Brew, P. D. Clegg, R. P. Boot-Handford, J. G. Andrew,and T. Hardingham, “Gene expression in human chondro-cytes in late osteoarthritis is changed in both fibrillated andintact cartilage without evidence of generalised chondrocytehypertrophy,” Annals of the Rheumatic Diseases, vol. 69, no. 1,pp. 234–240, 2010.

[194] T. Hayami, M. Pickarski, Y. Zhuo, G. A. Wesolowski, G.A. Rodan, and LE. T. Duong, “Characterization of articularcartilage and subchondral bone changes in the rat anteriorcruciate ligament transection and meniscectomized modelsof osteoarthritis,” Bone, vol. 38, no. 2, pp. 234–243, 2006.

[195] T. Pufe, W. Petersen, B. Tillmann, and R. Mentlein, “Thesplice variants VEGF121 and VEGF189 of the angiogenicpeptide vascular endothelial growth factor are expressed inosteoarthritic cartilage,” Arthritis and Rheumatism, vol. 44,no. 5, pp. 1082–1088, 2001.

[196] M. Sharif, J. Kirwan, N. Charni, L. J. Sandell, C. Whittles,and P. Garnero, “A 5-yr longitudinal study of type IIAcollagen synthesis and total type II collagen degradation inpatients with knee osteoarthritis—association with diseaseprogression,” Rheumatology, vol. 46, no. 6, pp. 938–943, 2007.

[197] E. V. Tchetina, J. Antoniou, M. Tanzer, D. J. Zukor, andA. R. Poole, “Transforming growth factor-β2 suppressescollagen cleavage in cultured human osteoarthritic cartilage,reduces expression of genes associated with chondrocytehypertrophy and degradation, and increases prostaglandin E2

production,” American Journal of Pathology, vol. 168, no. 1,pp. 131–140, 2006.

[198] W. Hui, T. Cawston, and A. D. Rowan, “Transforming growthfactor β1 and insulin-like growth factor 1 block collagendegradation induced by oncostatin M in combination withtumour necrosis factor α from bovine cartilage,” Annals ofthe Rheumatic Diseases, vol. 62, no. 2, pp. 172–174, 2003.

[199] J. Jiang, N. L. Leong, J. C. Mung, C. Hidaka, and H.H. Lu, “Interaction between zonal populations of articularchondrocytes suppresses chondrocyte mineralization andthis process is mediated by PTHrP,” Osteoarthritis andCartilage, vol. 16, no. 1, pp. 70–82, 2008.

[200] W. Kafienah, S. Mistry, S. C. Dickinson, T. J. Sims, I. Lear-month, and A. P. Hollander, “Three-dimensional cartilagetissue engineering using adult stem cells from osteoarthritispatients,” Arthritis and Rheumatism, vol. 56, no. 1, pp. 177–187, 2007.

[201] E. V. Tchetina, J. A. Di Battista, D. J. Zukor, J. Antoniou,and A. R. Poole, “Prostaglandin PGE2 at very low con-centrations suppresses collagen cleavage in cultured humanosteoarthritic articular cartilage: this involves a decreasein expression of proinflammatory genes, collagenases andCOL10A1, a gene linked to chondrocyte hypertrophy,”Arthritis Research & Therapy, vol. 9, no. 4, p. R75, 2007.

[202] J. Akaogi, T. Nozaki, M. Satoh, and H. Yamada, “Role of PGE2

and EP receptors in the pathogenesis of rheumatoid arthritisand as a novel therapeutic strategy,” Endocrine, Metabolic &Immune Disorders-Drug Targets, vol. 6, no. 4, pp. 383–394,2006.

[203] M. Attur, H. E. Al-Mussawir, J. Patel et al., “Prostaglandin E2exerts catabolic effects in osteoarthritis cartilage: evidence forsignaling via the EP4 receptor,” Journal of Immunology, vol.181, no. 7, pp. 5082–5088, 2008.

[204] D. D’Lima, J. Hermida, S. Hashimoto, C. Colwell, and M.Lotz, “Caspase inhibitors reduce severity of cartilage lesions

in experimental osteoarthritis,” Arthritis and Rheumatism,vol. 54, no. 6, pp. 1814–1821, 2006.

[205] A. Scharstuhl, E. L. Vitters, P. M. van der Kraan, andW. B. van den Berg, “Reduction of osteophyte formationand synovial thickening by adenoviral overexpression oftransforming growth factor β/bone morphogenetic proteininhibitors during experimental osteoarthritis,” Arthritis andRheumatism, vol. 48, no. 12, pp. 3442–3451, 2003.

[206] A. C. Lin, B. L. Seeto, J. M. Bartoszko et al., “Modulatinghedgehog signaling can attenuate the severity of osteoarthri-tis,” Nature Medicine, vol. 15, no. 12, pp. 1421–1425, 2009.

[207] E. Araldi and E. Schipani, “MicroRNA-140 and the silencingof osteoarthritis,” Genes and Development, vol. 24, no. 11, pp.1075–1080, 2010.

[208] S. S. Glasson, R. Askew, B. Sheppard et al., “Deletion of activeADAMTS5 prevents cartilage degradation in a murine modelof osteoarthritis,” Nature, vol. 434, no. 7033, pp. 644–648,2005.

[209] S. Kamekura, Y. Kawasaki, K. Hoshi et al., “Contributionof runt-related transcription factor 2 to the pathogenesisof osteoarthritis in mice after induction of knee jointinstability,” Arthritis and Rheumatism, vol. 54, no. 8, pp.2462–2470, 2006.

[210] A. R. Klatt, G. Klinger, O. Neumuller et al., “TAK1 downreg-ulation reduces IL-1β induced expression of MMP13, MMP1and TNF-alpha,” Biomedicine and Pharmacotherapy, vol. 60,no. 2, pp. 55–61, 2006.

[211] J. Weisser, B. Rahfoth, A. Timmermann, T. Aigner, R.Brauer, and K. von der Mark, “Role of growth factors inrabbit articular cartilage repair by chondrocytes in agarose,”Osteoarthritis and Cartilage, vol. 9, pp. S48–S54, 2001.

[212] T. Yamamoto, S. Wakitani, K. Imoto et al., “Fibroblast growthfactor-2 promotes the repair of partial thickness defects ofarticular cartilage in immature rabbits but not in maturerabbits,” Osteoarthritis and Cartilage, vol. 12, no. 8, pp. 636–641, 2004.

[213] P. R. Elford, M. Graeber, H. Ohtsu et al., “Induction ofswelling, synovial hyperplasia and cartilage proteoglycan lossupon intra-articular injection of transforming growth factorβ-2 in the rabbit,” Cytokine, vol. 4, no. 3, pp. 232–238, 1992.

[214] S. B. Trippel, “Growth factor inhibition: potential role in theetiopathogenesis of osteoarthritis,” Clinical Orthopaedics andRelated Research, no. 427, pp. S47–S52, 2004.

[215] J. I. Shida, S. Jingushi, T. Izumi, T. Ikenoue, and Y.Iwamoto, “Basic fibroblast growth factor regulates expressionof growth factors in rat epiphyseal chondrocytes,” Journal ofOrthopaedic Research, vol. 19, no. 2, pp. 259–264, 2001.

[216] P. C. Yaeger, T. L. Masi, J. L. Buck de Ortiz, F. Binette, R. Tubo,and J. M. McPherson, “Synergistic action of transforminggrowth factor-β and insulin-like growth factor-I inducesexpression of type II collagen and aggrecan genes in adulthuman articular chondrocytes,” Experimental Cell Research,vol. 237, no. 2, pp. 318–325, 1997.

[217] A. Barbero, S. Grogan, D. Schafer, M. Heberer, P. Mainil-Varlet, and I. Martin, “Age related changes in human articularchondrocyte yield, proliferation and post-expansion chon-drogenic capacity,” Osteoarthritis and Cartilage, vol. 12, no.6, pp. 476–484, 2004.

[218] A. V. Babarina, U. Mollers, K. Bittner, P. Vischer, and P.Bruckner, “Role of the subchondral vascular system in endo-chondral ossification: endothelial cell-derived proteinasesderepress late cartilage differentiation in vitro,” MatrixBiology, vol. 20, no. 3, pp. 205–213, 2001.

16 Arthritis

[219] J. O. P. Cheung, M. C. Hillarby, S. Ayad et al., “A novelcell culture model of chondrocyte differentiation duringmammalian endochondral ossification,” Journal of Bone andMineral Research, vol. 16, no. 2, pp. 309–318, 2001.

[220] M. J. Zuscik, J. F. Baden, Q. Wu et al., “5-azacytidinealters TGF-β and BMP signaling and induces maturation inarticular chondrocytes,” Journal of Cellular Biochemistry, vol.92, no. 2, pp. 316–331, 2004.

[221] M. L. Ho, JE. K. Chang, S. C. Wu et al., “A novel terminaldifferentiation model of human articular chondrocytes inthree-dimensional cultures mimicking chondrocytic changesin osteoarthritis,” Cell Biology International, vol. 30, no. 3, pp.288–294, 2006.

[222] T. Yasuda, E. Tchetina, K. Ohsawa et al., “Peptides of typeII collagen can induce the cleavage of type II collagen andaggrecan in articular cartilage,” Matrix Biology, vol. 25, no. 7,pp. 419–429, 2006.

[223] T. Yasuda and A. R. Poole, “A fibronectin fragment inducestype II collagen degradation by collagenase through aninterleukin-1-mediated pathway,” Arthritis and Rheumatism,vol. 46, no. 1, pp. 138–148, 2002.

[224] G. A. Homandberg, “Cartilage damage by matrix degrada-tion products: fibronectin fragments,” Clinical Orthopaedicsand Related Research, no. 391, pp. S100–S107, 2001.

[225] R. Serra, M. Johnson, E. H. Filvaroff et al., “Expression of atruncated, kinase-defective TGF-β type II receptor in mouseskeletal tissue promotes terminal chondrocyte differentiationand osteoarthritis,” Journal of Cell Biology, vol. 139, no. 2, pp.541–552, 1997.

[226] T. F. Li, M. Darowish, M. J. Zuscik et al., “Smad3-deficientchondrocytes have enhanced BMP signaling and accelerateddifferentiation,” Journal of Bone and Mineral Research, vol. 21,no. 1, pp. 4–16, 2006.

[227] G. Valverde-Franco, J. S. Binette, . et al., “Defects in articularcartilage metabolism and early arthritis in fibroblast growthfactor receptor 3 deficient mice,” Human Molecular Genetics,vol. 15, no. 11, pp. 1783–1792, 2006.

[228] C. T. G. Appleton, S. E. Usmani, S. M. Bernier, T. Aigner,and F. Beier, “Transforming growth factor α suppression ofarticular chondrocyte phenotype and Sox9 expression in arat model of osteoarthritis,” Arthritis and Rheumatism, vol.56, no. 11, pp. 3693–3705, 2007.

[229] M. B. Goldring and K. B. Marcu, “Cartilage homeostasis inhealth and rheumatic diseases,” Arthritis Research & Therapy,vol. 11, no. 3, article 224, 2009.

[230] D. L. Cecil, K. Johnson, J. Rediske, M. Lotz, A. M. Schmidt,and R. Terkeltaub, “Inflammation-induced chondrocytehypertrophy is driven by receptor for advanced glycation endproducts,” Journal of Immunology, vol. 175, no. 12, pp. 8296–8302, 2005.

[231] D. Merz, R. Liu, K. Johnson, and R. Terkeltaub, “IL-8/CXCl8and growth-related oncogene α/CXCL1 induce chondrocytehypertrophic differentiation,” Journal of Immunology, vol.171, no. 8, pp. 4406–4415, 2003.

[232] C. Bauge, F. Legendre, S. Leclercq et al., “Interleukin-1βimpairment of transforming growth factor β1 signaling bydown-regulation of transforming growth factor β receptortype II and up-regulation of Smad7 in human articularchondrocytes,” Arthritis and Rheumatism, vol. 56, no. 9, pp.3020–3032, 2007.

Hindawi Publishing CorporationArthritisVolume 2011, Article ID 591253, 6 pagesdoi:10.1155/2011/591253

Research Article

The Self-Administered Patient Satisfaction Scale forPrimary Hip and Knee Arthroplasty

N. Mahomed,1 Rajiv Gandhi,1 Lawrence Daltroy,2 and J. N. Katz3

1 Department of Surgery, Toronto Western Hospital, Toronto, ON, Canada2 Harvard Medical School, Boston, MA, USA3 Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA

Correspondence should be addressed to N. Mahomed, nizar.mahomed@uhn.on.ca

Received 2 July 2010; Revised 10 November 2010; Accepted 2 December 2010

Academic Editor: Changhai Ding

Copyright © 2011 N. Mahomed et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Introduction. The objective of this study was to develop a short self-report questionnaire for evaluating patient satisfaction with theoutcome of hip and knee replacement surgery. Methods. This scale consists of four items focusing on satisfaction with the extent ofpain relief, improvement in ability to perform home or yard work, ability to perform recreational activities, and overall satisfactionwith joint replacement. This instrument does not measure satisfaction with process of care. The responses are scored on a Likertscale, with the total score ranging from 25 to 100 per question. The instrument was tested on 1700 patients undergoing primarytotal hip and total knee replacement surgery, evaluated preoperatively, at 12 weeks, and one year postoperatively. Psychometrictesting included internal consistency, measured with Cronbach’s alpha, and convergent validity, measured by correlation withchanges in measures of health status between the preoperative, 12-week, and one-year evaluations. Results. The internal consistency(reliability) of the scale, measured by the Cronbach’s alpha, ranged from 0.86 to 0.92. The scale demonstrated substantial ceilingeffects at 1 year. The scale scores correlated modestly with the absolute SF-36 PCS and WOMAC scores (ρ = 0.56–0.63) and alsowith the WOMAC change scores (ρ = 0.38–0.46) at both 12-week and 1-year followups. Conclusions. This instrument is valid andreliable for measuring patient satisfaction following primary hip and knee arthroplasty and could be further evaluated for use withother musculoskeletal interventions.

1. Introduction

Total knee (TKA) and total hip arthroplasty (THA) areeffective procedures for relieving pain and restoring function.Increasingly, patient satisfaction is being recognized as animportant measure of health care quality [1].

Several studies have shown a discrepancy betweensurgeon and patient assessment of medical and surgicaloutcomes, particularly in assessing pain and function [2–5]. Surgeons generally focus on range of motion, alignment,and stability—objective measures—while patients are moreconcerned with overall functionality of the joint—subjectivemeasures [6, 7]. Moreover, the outcome of satisfaction likelystretches beyond improved mobility and pain relief, butrather encompasses other factors such as fulfillment of pre-operative expectations [8, 9]. A complete evaluation of totaljoint replacement (TJR) outcomes should therefore include

clinical measures, generic and disease-specific functionalhealth outcomes scales, and patient satisfaction.

There are a number of published scales on patientsatisfaction, but these focus on issues of process of care, suchas the office environment, the patient-physician interaction,or general satisfaction with care [10–12]. Although manyauthors have reported on patient satisfaction after TJR,we are unaware of any validated self-report scale thatevaluates patient satisfaction with the results of muscu-loskeletal treatment outcomes. Knowledge of modifiablefactors predicting patient dissatisfaction following surgery,such as expectations or comorbidity, would be valuable forphysicians and patients in order to improve satisfaction withsurgical outcomes.

The goal of this study was to develop and performpsychometric testing (internal consistency and convergentvalidity) for a short self-report questionnaire evaluating

2 Arthritis

patient satisfaction with the outcome of hip and knee arthro-plasty. The Self-Administered Patient Satisfaction Scale(SAPS) was designed to be used in conjunction with otherclinical measures and functional health status instrumentsto evaluate the results of hip and knee arthroplasty. Wehypothesized that individuals who experienced less pain andbetter functioning and greater improvements in these areaswould likewise show greater satisfaction with their surgery[8, 13].

2. Methods

2.1. Scale Description. Four items, as determined byliterature review and an expert consensus panel—rheumatologists, an orthopaedic surgeon, and a behavioralscientist—were chosen reflecting various facets of patientfunctioning most affected by TJR. The items includepatients’ overall satisfaction with surgery, the extent of painrelief, the ability to perform home or yard work, and theability to perform recreational activities. Items are scored ona 4-point Likert scale with response categories consisting ofvery satisfied (100 points), somewhat satisfied (75 points),somewhat dissatisfied (50 points), and very dissatisfied (25points). The scale score is the unweighted mean of the scoresfrom the individual items, ranging from 25 to 100 per item(with 100 being most satisfied).

Face validity was assessed by having the scale reviewed bya panel of independent experts in the field of rheumatologyand orthopedics.

2.2. Instrument Testing. As part of our registry, patients arerecruited from a single Canadian academic institution, theToronto Western Hospital, prior to undergoing primary TKAand THA. For this study, we included patients aged 18years or above with a diagnosis of primary or secondaryosteoarthritis (OA), or inflammatory arthritis. All patientsgave informed consent to participate in this study. All datawere collected by an independent assessor not involved in thepatient’s medical care. The study protocol was approved bythe local ethics committee.

2.3. Data Collection. Baseline demographic data of age,gender, body mass index (BMI), and ethnicity were collectedby patient self-report at the time of surgery. Ethnicity wasrecorded under the categories of White/European, Black,Asian, or Aboriginal as defined by the United States Census[14]. Asian refers to individuals who classified themselves asSouth Asian (India, Pakistan, Bangladesh and Sri Lanka) orEast Asian (China, Japan, Taiwan, Korea).

Functional status was assessed at baseline, 12-weekand 1-year followup using the Western Ontario McMas-ter University Osteoarthritis Index (WOMAC) [15], andthe Medical Outcomes Short Form 36 (SF 36) PhysicalComponent Score (PCS) [16, 17]. For ease of presentation,the WOMAC scores were rescaled from 0–100 with highscores representing better function and pain relief. TheSF 36 is the most common generic health scale used ina joint arthroplasty population [18]. The WOMAC index

has become the standard scale adopted by the AmericanAcademy of Orthopaedic Surgeons and the Council OfMusculoskeletal Societies for the assessment of joint (hip andknee) functional outcomes. Patient satisfaction was assessedat the same followup points using the SAPS.

2.3.1. Scale Characteristics

Reliability. The distributions of scale scores were examinedwith tests of normality including skewness and kurtosis andgraphically with histograms. We tested for floor and ceilingeffects by calculating the percentage of respondents scoringat the lowest and the highest scale levels, respectively. Theinternal consistency (reliability) of the scale was evaluatedusing the Cronbach’s alpha coefficient. The Cronbach’s alphacalculates how well a set of variables correlate with oneanother and with the aggregate scale score [19]. It is used toassess the capacity of the scale to measure a unidimensionalconcept. Values of 0.7 or greater are considered acceptable,values greater than 0.8 are considered good, and greater than0.9 are considered excellent [19].

Convergent Validity. We compared the results of the satisfac-tion scale against the absolute total WOMAC score and SF 36PCS as well as against the change in total WOMAC score atboth 12-week and 1-year followup. The change score repre-sents the improvement in pain and function from surgeryeach reported patient calculated as the difference betweenthe followup score and the baseline score. The WOMACchange score has been shown by others to be related topatient-reported satisfaction with TJR [13]. Spearman’s rankcorrelation coefficient was used as not all data was normallydistributed.

Separate statistical analyses were conducted for hip andknee patients. All analyses were completed with SPSS version13.0 (SPSS, Chicago, IL, USA). All reported p values are 2-tailed with an alpha of 0.05.

3. Results

The validity and internal consistency (reliability) of the scalewere assessed on complete data from 843 hip arthroplastypatients and 857 knee arthroplasty patients. This representsa response rate of 1700/2000 (85%). The demographic data,functional outcomes, and satisfaction scores of these groupsare presented in Table 1. The mean age of the group was65.2 years (range 19–88, SD 11.3) while 44.4% were males.The knee arthroplasty patients, on average, were a fewyears older than the hip arthroplasty patients with a greaterpercentage of females. The demographic data of respondersand nonresponders was not clinically or statistically different(data not shown).

3.1. Response Patterns. The majority of responses were in thevery satisfied or somewhat satisfied categories. More patientswere satisfied with pain relief compared to improvement intheir ability to do work or recreational activities [18, 20].All scale scores demonstrated a negative skew for both hip

Arthritis 3

and knee patient populations, indicating that respondents’scores were clustered toward the positive end of the healthspectrum.

Table 2 presents the distribution of responses for eachitem in the scale. Substantial ceiling effects (62% and52%) were found for the 1-year hip and knee scale scores,respectively. Modest ceiling effects (39% and 30%) werefound for the 12 week hip and knee scale scores, respectively.There was minimal evidence of floor effects for both hip andknee patients as less than 3% scored at the lowest end of thescale.

3.2. Scale Characteristics: Reliability. The 12 week and 1-year cronbach’s alpha (reliability) coefficients for satisfactionscores are given in Table 3.

3.3. Scale Characteristics: Convergent Validity

3.3.1. Hips. The 12 week Spearman’s correlation coefficientfor total WOMAC scores and satisfaction scores was 0.57,P < .001. At 1-year followup, the correlation coefficient was0.60, P < .0001. For the outcome of the total WOMACchange score, the Spearman’s correlation coefficients at 12-week and 1-year followup were 0.43, P < .001 and 0.44,P < .001, respectively.

The 12 week Spearman’s correlation coefficient betweenthe SF36 PCS and satisfaction scores was 0.55, P < .001. At 1-year followup, the correlation coefficient was 0.55, P < .0001.For the outcome of the PCS change score, the Spearman’scorrelation coefficients at 12-week and 1-year followup were0.46, P < .001 and 0.46, P < .001, respectively.

3.3.2. Knees. The 12 week Spearman’s correlation coefficientfor total WOMAC scores and satisfaction scores was 0.61,P < .001. At 1-year followup, the correlation coefficient was0.64, P < .0001. For the outcome of the total WOMACchange score, the Spearman’s correlation coefficients at 12-week and 1-year followup were 0.41, P < .001 and 0.38,P < .001, respectively.

The 12 week Spearman’s correlation coefficient betweenthe SF36 PCS scores and satisfaction scores was 0.53, P <.001. At 1-year followup, the correlation coefficient was 0.57,P < .0001. For the outcome of the PCS change score, theSpearman’s correlation coefficients at 12-week and 1-yearfollowup were 0.38, P < .001 and 0.41, P < .001, respectively.

Figures 1 and 2 demonstrate the relationship betweenthe quartiles of satisfaction scores and mean total WOMACscores for hip and knee patients at 1-year, respectively. Forthe hip patients, the mean WOMAC scores for quartiles 2 to 4are similar due to the ceiling effect as satisfaction scores werehigh across those quartiles. The knee patients demonstratedmore of a graded response between quartiles of satisfactionand mean WOMAC scores indicating less of a ceiling effect.

4. Discussion

The assessment of TJR outcomes has evolved from focusingprimarily on clinical outcomes, to patient report measures

0

20

40

60

80

100

120

Q1 Q2 Q3 Q4

Satisfaction quartiles

Tota

lWO

MA

Csc

ores

Figure 1: Mean 1-year total WOMAC scores (with SD) comparedacross quartiles of 1-year satisfaction scores for hip patients.

0

20

40

60

80

100

120

Q1 Q2 Q3 Q4

Satisfaction quartiles

Tota

lWO

MA

Csc

ores

Figure 2: Mean 1-year total WOMAC scores (with SD) comparedacross quartiles of 1-year satisfaction scores for knee patients.

and patient satisfaction. It has been estimated that between9 and 30% of patients may be dissatisfied following surgeryand thus a greater understanding of the determinants ofpatient satisfaction may help to improve subjective outcomes[8, 21–23]. This short four-item satisfaction scale is areliable (high internal consistency) and valid instrumentfor measuring satisfaction with the outcome of TJR. Thedistribution of responses for each item revealed that patientswere more satisfied with improvement in pain than function.This correlates with clinical experience and the publishedliterature, where pain relief is more reliably achieved thanimprovement in function with TJR [18]. One group showedthat 73% of patients were very satisfied with pain relief, butonly 50% were very satisfied with their ability to performleisure activities following TKA [20]. Other potential rea-sons for patient dissatisfaction with surgery may be poormental health, unfulfilled expectations, the patient-surgeonrelationship, or length of the incision [8, 9, 22, 24–27].

Further, we found that the total satisfaction scores weregreater at 12-week and 1-year following THA as comparedto TKA, which is consistent with the literature that showsgreater satisfaction and health improvement following hip

4 Arthritis

Table 1: Demographic data, satisfaction scores, and total WOMAC scores for hip and knee replacement patients used in scale assessment.

Hips Knees

n= 843 n = 857

Mean age (SD) 64.8 (12.6) 67.3 (9.9)

% Males 46% 36.8%

Mean BMI kg/m2 (SD) 27.4 (5.5) 30.4 (6.5)

% White 76% 71%

Mean/median 12 week

Satisfaction 90.6/100 83.4/87.5

score (range,SD) (25–100,13.8) (25–100,17.2)

Mean/median 12 week

total WOMAC 76.8/81 68.9/70

Score (range,SD) (20–100,17.2) (4–100,18.2)

Mean/median 12 week

SF36 PCS 40.1/39.8 36.8/35.8

scores (range,SD) (11.5–65.1,10.1) (13.6–61.6,9.2)

Mean/median 1 year

Satisfaction 92.4/87.5 84.2/100

score (range,SD) (25–100,13.5) (25–100,19.0)

Mean/median 1-year total

WOMAC 82.4/88 72.7/76

scores (range,SD) (15–100,16.9) (8–100,18.9)

Mean/median 1-year SF36

PCS scores (range,SD) 44.2/44.6 (16.6–67.6,11.0) 34.4/38.7 (14.7,68.5,10.5)

Table 2: Percentage distribution for responses for each item.

Hips Knees

12 wks (%) 1 year (%) 12 wks (%) 1 year (%)

How satisfied are you with the results of your surgery ?

Very satisfied 83.4 83.0 64.1 65.8

Somewhat satisfied 13.4 13.6 28.7 22.2

Somewhat dissatisfied 1.8 2.1 5.1 8.1

Very dissatisfied 1.4 1.3 2.1 3.9

How satisfied are you with the results of your surgery for improvingyour pain?

Very satisfied 84.4 86.9 63.4 67.6

Somewhat satisfied 12.3 10.7 27.7 21.9

Somewhat dissatisfied 2.3 1.6 5.8 7.3

Very dissatisfied 1.0 0.8 3.2 3.2

How satisfied are you with the results of surgery for improving yourability to do home or yard work?

Very satisfied 61.0 71.7 40.0 50.2

Somewhat satisfied 30.3 20.1 45.7 33.9

Somewhat dissatisfied 5.0 5.9 9.8 10.7

Very dissatisfied 3.7 2.3 4.4 5.2

How satisfied are you with the results of surgery for improving yourability to do recreational activities?

Very satisfied 53.6 65.8 34.0 42.8

Somewhat satisfied 34.7 24.7 46.9 36.2

Somewhat dissatisfied 7.1 6.9 12.8 13.1

Very dissatisfied 4.6 2.6 6.2 7.9

Arthritis 5

Table 3: Cronbach’s alpha coefficients for internal consistency(reliability) of the satisfaction scale.

Hips Knees

12-week satisfaction 0.86 0.91

1-year Satisfaction 0.91 0.92

arthroplasty [28, 29]. Hip arthroplasty consistently demon-strates superior functional outcomes as compared to theknee, potentially because the ball and socket design of a hipjoint is easier to replicate with metallic implants as comparedto a more complicated hinge joint such as the knee.

The scale has a broad distribution of scores with sub-stantial ceiling effects at 1-year. Similar floor/ceiling effectsare seen in widely used scale such as the SF 36 [30–32] andin other satisfaction scales [33]. The instrument should takeapproximately 2 minutes to complete. It could therefore beadded to other outcome instruments to more completelyevaluate the results of TJR without significantly increasingrespondent burden. This scale differs from others in theliterature in that it only focuses on satisfaction with theoutcome of an intervention rather than the process of care[10–12].

We found a slightly better correlation for the absolutescores of the WOMAC index as compared to the WOMACchange scores at both 12-week and 1-year followup. This sug-gests that satisfaction following TJR may be more predictedby the final status reached rather than the relative benefitgained from the surgery. Moreover, it suggests that althoughpatient satisfaction is related to improvement in pain andfunction, these domains are not directly correlated. Thusthe satisfaction scale measures a different but interrelateddomain.

Limitations of this study include the use of the Likertresponse scale that yields ordinal rather than interval data.Caution with parametric analysis is thus required. However,the Cronbach’s alpha coefficient indicated good reliabilityfor summing the individual items into a summary score.Second, we did not assess test-retest reliability due to thenature of the existing study designs of the two cohorts used.Third, the scale was not validated on a cohort of revisionarthroplasty patients, and this would be required prior to itsuse in this population. Fourth, although the scale asks aboutgeneral satisfaction “with results of surgery”, it would requirevalidation on other musculoskeletal surgical interventionsbefore its use could be generalized.

In conclusion, this satisfaction scale may be used inconjunction with other outcome instruments to morecomprehensively evaluate the results of primary hip andknee replacement surgery. It has been shown that patientsatisfaction following surgery does not always correlatewith surgeon assessments [3, 6] and this scale providesa simple instrument to explore the complex relationshipsbetween patient baseline pain, functioning, expectations ofsurgery, and satisfaction with outcome. Identification of anymodifiable risk factor for patient dissatisfaction with surgerypresents an opportunity of improving patients perceivedoutcomes.

References

[1] C. K. Ross, C. A. Steward, and J. M. Sinacore, “A comparativestudy of seven measures of patient satisfaction,” Medical Care,vol. 33, no. 4, pp. 392–406, 1995.

[2] P. N. Baker, J. H. van der Meulen, J. Lewsey, and P. J.Gregg, “The role of pain and function in determining patientsatisfaction after total knee replacement. Data from thenational joint registry for England and Wales,” Journal of Boneand Joint Surgery B, vol. 89, no. 7, pp. 893–900, 2007.

[3] A. J. Janse, R. J. B. J. Gemke, C. S. P. M. Uiterwaal, I. VanDer Tweel, J. L. L. Kimpen, and G. Sinnema, “Quality oflife: patients and doctors don’t always agree: a meta-analysis,”Journal of Clinical Epidemiology, vol. 57, no. 7, pp. 653–661,2004.

[4] P. M. Rothwell, Z. McDowell, C. K. Wong, and P. J. Dorman,“Doctors and patients don’t agree: cross sectional studyof patients’ and doctors’ perceptions and assessments ofdisability in multiple sclerosis,” British Medical Journal, vol.314, no. 7094, pp. 1580–1583, 1997.

[5] R. B. G. Brokelman, C. J. M. van Loon, and W. J. Rijnberg,“Patient versus surgeon satisfaction after total hip arthro-plasty,” Journal of Bone and Joint Surgery B, vol. 85, no. 4, pp.495–498, 2003.

[6] P. H. J. Bullens, C. J. M. van Loon, M. C. de Waal Malefijt,R. F. J. M. Laan, and R. P. H. Veth, “Patient satisfaction aftertotal knee arthroplasty: a comparison between subjective andobjective outcome assessments,” Journal of Arthroplasty, vol.16, no. 6, pp. 740–747, 2001.

[7] P. C. Amadio, “Outcomes measurements. More questions;some answers,” Journal of Bone and Joint Surgery A, vol. 75,no. 11, pp. 1583–1584, 1993.

[8] P. C. Noble, M. A. Conditt, K. F. Cook, and K. B. Mathis,“The John Insall Award: patient expectations affect satisfactionwith total knee arthroplasty,” Clinical Orthopaedics and RelatedResearch, no. 452, pp. 35–43, 2006.

[9] C. A. Mancuso, E. A. Salvati, N. A. Johanson, M. G. E.Peterson, and M. E. Charlson, “Patients’ expectations and sat-isfaction with total hip arthroplasty,” Journal of Arthroplasty,vol. 12, no. 4, pp. 387–396, 1997.

[10] D. H. Solomon, D. W. Bates, J. Horsky, E. Burdick, J. L.Schaffer, and J. N. Katz, “Development and validation of apatient satisfaction scale for musculoskeletal care,” ArthritisCare and Research, vol. 12, no. 2, pp. 96–100, 1999.

[11] T. J. Sole and P. E. Lipsky, “Satisfaction of patients attendingan arthritis clinic in a county teaching hospital,” Arthritis Careand Research, vol. 10, no. 3, pp. 169–176, 1997.

[12] S. Grogan, M. Conner, P. Norman, D. Willits, and I. Porter,“Validation of a questionnaire measuring patient satisfactionwith general practitioner services,” Quality in Health Care, vol.9, no. 4, pp. 210–215, 2000.

[13] R. B. Bourne, “Measuring tools for functional outcomes intotal knee arthroplasty,” Clinical Orthopaedics and RelatedResearch, vol. 466, no. 11, pp. 2634–2638, 2008.

[14] “2000 Census of Population, Public Law 94-171 RedistrictingData File: Race,” U.S. Census Bureau.

[15] N. Bellamy, W. W. Buchanan, C. H. Goldsmith, J. Campbell,and L. W. Stitt, “Validation study of WOMAC: a health statusinstrument for measuring clinically important patient relevantoutcomes to antirheumatic drug therapy in patients withosteoarthritis of the hip or knee,” Journal of Rheumatology, vol.15, no. 12, pp. 1833–1840, 1988.

[16] C. A. McHorney, J. E. Ware Jr., and A. E. Raczek, “The MOS36-Item Short-Form Health Survey (SF-36): II. Psychometric

6 Arthritis

and clinical tests of validity in measuring physical and mentalhealth constructs,” Medical Care, vol. 31, no. 3, pp. 247–263,1993.

[17] C. A. McHorney, J. E. Ware Jr., J. F. Lu, and C. D. Sherbourne,“The MOS 36-item Short-Form Health Survey (SF-36): III.Tests of data quality, scaling assumptions, and reliability acrossdiverse patient groups,” Medical Care, vol. 32, no. 1, pp. 40–66,1994.

[18] O. Ethgen, O. Bruyere, F. Richy, C. Dardennes, and J. Y.Reginster, “Health-related quality of life in total hip and totalknee arthroplasty: a qualitative and systematic review of theliterature,” Journal of Bone and Joint Surgery A, vol. 86, no. 5,pp. 963–974, 2004.

[19] B. Dawson and R. G. Trapp, “Correlation and regression,” inBasic and Clinical Biostatistics, Lange, 4th edition, 1994.

[20] R. J. Wright, C. B. Sledge, R. Poss, F. C. Ewald, M. E. Walsh, andE. A. Lingard, “Patient-reported outcome and survivorshipafter kinemax total knee arthroplasty,” Journal of Bone andJoint Surgery A, vol. 86, no. 11, pp. 2464–2470, 2004.

[21] L. D. Dorr, D. Thomas, W. T. Long, P. B. Polatin, and L. E.Sirianni, “Psychologic reasons for patients preferring mini-mally invasive total hip arthroplasty,” Clinical Orthopaedicsand Related Research, no. 458, pp. 94–100, 2007.

[22] O. Robertsson, M. Dunbar, T. Pehrsson, K. Knutson, and L.Lidgren, “Patient satisfaction after knee arthroplasty: a reporton 27,372 knees operated on between 1981 and 1995 inSweden,” Acta Orthopaedica Scandinavica, vol. 71, no. 3, pp.262–267, 2000.

[23] R. Dickstein, Y. Heffes, E. I. Shabtai, and E. Markowitz, “Totalknee arthroplasty in the elderly: patients’ self-appraisal 6 and12 months postoperatively,” Gerontology, vol. 44, no. 4, pp.204–210, 1998.

[24] N. N. Mahomed, M. H. Liang, E. F. Cook et al., “Theimportance of patient expectations in predicting functionaloutcomes after total joint arthroplasty,” Journal of Rheumatol-ogy, vol. 29, no. 6, pp. 1273–1279, 2002.

[25] J. G. Anderson, R. L. Wixson, D. Tsai, S. D. Stulberg, and R. W.Chang, “Functional outcome and patient satisfaction in totalknee patients over the age of 75,” Journal of Arthroplasty, vol.11, no. 7, pp. 831–840, 1996.

[26] L. D. Dorr, D. Thomas, W. T. Long, P. B. Polatin, and L. E.Sirianni, “Psychologic reasons for patients preferring mini-mally invasive total hip arthroplasty,” Clinical Orthopaedicsand Related Research, no. 458, pp. 94–100, 2007.

[27] R. Gandhi, J. R. Davey, and N. N. Mahomed, “Predictingpatient dissatisfaction following joint replacement surgery,”Journal of Rheumatology, vol. 35, no. 12, pp. 2415–2418, 2008.

[28] C. A. Jones, D. C. Voaklander, D. W. C. Johnston, and M.E. Suarez-Almazor, “Health related quality of life outcomesafter total hip and knee arthroplasties in a community basedpopulation,” Journal of Rheumatology, vol. 27, no. 7, pp. 1745–1752, 2000.

[29] P. Rissanen, S. Aro, P. Slatis, H. Sintonen, and P. Paavolainen,“Health and quality of life before and after hip or kneearthroplasty,” Journal of Arthroplasty, vol. 10, no. 2, pp. 169–176, 1995.

[30] P. G. O’Mahony, H. Rodgers, R. G. Thomson, R. Dobson, andO. F. W. James, “Is the SF-36 suitable for assessing health statusof older stroke patients,” Age and Ageing, vol. 27, no. 1, pp. 19–22, 1998.

[31] J. E. Brazier, R. Harper, J. Munro, S. J. Walters, and M. L.Snaith, “Generic and condition-specific outcome measures forpeople with osteoarthritis of the knee,” Rheumatology, vol. 38,no. 9, pp. 870–877, 1999.

[32] N. K. Aaronson, M. Muller, P. D. A. Cohen et al., “Translation,validation, and norming of the Dutch language version ofthe SF-36 Health Survey in community and chronic diseasepopulations,” Journal of Clinical Epidemiology, vol. 51, no. 11,pp. 1055–1068, 1998.

[33] D. B. DiBenedetti, K. Gondek, P. P. Sagnier et al., “Thetreatment satisfaction scale: a multidimensional instrumentfor the assessment of treatment satisfaction for erectiledysfunction patients and their partners,” European Urology,vol. 48, no. 3, pp. 503–511, 2005.

Hindawi Publishing CorporationArthritisVolume 2011, Article ID 580632, 6 pagesdoi:10.1155/2011/580632

Research Article

Serum Leptin Concentration Positively Correlates withBody Weight and Total Fat Mass in Postmenopausal JapaneseWomen with Osteoarthritis of the Knee

Jun Iwamoto,1 Tsuyoshi Takeda,1 Yoshihiro Sato,2 and Hideo Matsumoto1

1 Institute for Integrated Sports Medicine, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan2 Department of Neurology, Mitate Hospital, Tagawa, Fukuoka 826-0041, Japan

Correspondence should be addressed to Jun Iwamoto, jiwamoto@sc.itc.keio.ac.jp

Received 11 August 2010; Accepted 5 January 2011

Academic Editor: Marco Amedeo Cimmino

Copyright © 2011 Jun Iwamoto et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The objective of the present study was to identify factors correlated with the serum leptin concentration in women with knee OA.Fifty postmenopausal Japanese women with knee OA (age: 50–88 years) were recruited in our outpatient clinic. Plain radiographsof the knee were taken, and urine and blood samples were collected. Dual-energy X-ray absorptiometry (DXA) scanning wasperformed for the whole body and lumbar spine, and factors correlated with the serum leptin concentration were identified. Asimple linear regression analysis showed that body weight, body mass index, whole-body bone mineral density (BMD), total fatmass, and total fat percentage, but not age, height, lumbar spine BMD, lean body mass, serum and urinary bone turnover markers,or the radiographic grade of knee OA, were significantly correlated with the serum leptin concentration. A multiple regressionanalysis showed that among these factors, only body weight and total fat mass exhibited a significant positive correlation with theserum leptin concentration. These results suggest that the serum leptin concentration might be related to increases in body weightand total fat mass, but not to BMD or bone turnover markers, in postmenopausal women with OA.

1. Introduction

Leptin, the product of the ob gene, is a peptide hormonesecreted primarily by the adipocytes and plays an importantrole in the regulation of body weight by centrally inhibitingfood intake and stimulating energy expenditure [1]. Leptinenters the circulation and crosses the blood-brain barrierto reach its primary target, receptors in the hypothalamus.However, a clinical study has shown that the serum leptinconcentration is elevated in obese persons and is correlatedwith the percentage of body fat [2]. Thus, obesity is morelikely to be caused by central mechanisms regulating foodintake and energy expenditure than by defective signalingoriginating in adipocytes and affecting these central mech-anisms [2].

Leptin regulates bone metabolism [3]. Leptin decreasesbone formation and increases bone resorption via thesympathetic nervous system [4, 5]. Conversely, circulatingleptin also regulates bone metabolism directly by binding to

leptin receptors on bone marrow stromal cells, osteoblasts,and osteoclasts, functioning to increase osteoblast activityand decrease osteoclast activity [6]. Clinically, the resultsof cross-sectional and longitudinal cohort studies examin-ing relationships among the serum leptin concentration,bone mineral density (BMD), and/or bone formation andresorption markers remain contradictory [7–9]. A recentprospective study demonstrated that the initial serum lep-tin concentration, together with specific-body compositionparameters, determined the loss in total and femoral neckBMD values in physically active older women [10].

Osteoarthritis (OA) of the knee is the most commontype of arthritis and is the major cause of chronic muscu-loskeletal pain and mobility disability in elderly populations,representing a significant burden on health care provision.Well-established risk factors for OA include aging, obesity,and female sex [11]. The serum leptin concentration isknown to be two to threefold higher in women than inmen, independent of adiposity [12]. It has been hypothesized

2 Arthritis

that leptin might be a systemic or local factor mediatingthe metabolic link between obesity and OA and partiallyaccounting for the gender disparity of this disease [13].

Accordingly, leptin was speculated to mediate themetabolic link connecting fat mass, knee OA, and BMDin women. However, very few studies have addressed thisissue in women with knee OA. The objective of the presentstudy was to identify possible correlations between theserum leptin concentration and various factors includingage, height, body weight, body mass index (BMI), BMD, leanbody mass, total fat mass, total fat percentage, bone turnovermarkers, and the radiographic grade of OA in women withknee OA.

2. Subjects and Methods

2.1. Subjects. Postmenopausal Japanese women with OA ofthe knee who visited our outpatient clinic (orthopaedics andsports medicine clinic) in 2002 or 2003 were recruited. Allthe patients had consulted experts in knee OA treatment atour clinic because of knee pain. The diagnosis of OA of theknee was made based on the clinical symptoms, the resultsof a physical examination, and X-ray findings of the knee.All the patients had mild, moderate, or severe OA of the knee(grades 1–4) according to the Kellgren and Lawrence methodof grading (grade 0, normal; grade 1, possible osteophytesonly; grade 2, definite osteophytes and possible joint-spacenarrowing; grade 3, moderate osteophytes and/or definitejoint-space narrowing; and grade 4, large osteophytes, severejoint-space narrowing, and/or bony sclerosis) [14]. Becausepharmacological treatments such as oral nonsteroidal anti-inflammatory drugs (NSAIDs) and intra-articular injectionsof hyaluronates are effective for symptom relief [15, 16], mostof the patients were receiving weekly to monthly intraarticu-lar injections of hyaluronate sodium and/or oral NSAIDs.

After one-month of the cessation of intraarticular injec-tions of hyaluronate sodium, dual-energy X-ray absorptiom-etry (DXA) scanning was performed on the whole bodyand the whole lumbar spine. Urine and serum samples werecollected from all the patients between 9:00 am and 11:00 am,and the serum leptin and urinary and serum bone turnovermarkers were measured. Factors correlated with the serumleptin level were then determined. Informed consent wasobtained from each patient.

2.2. Measurements of Serum Leptin and Serum and UrinaryBone Turnover Markers. The urinary levels of pyridino-line and deoxypyridinoline were measured using high-performance liquid chromatography (HPLC) (normal range:17.7–41.9 pmol/μmol Cr and 2.8–7.6 nmol/mmol Cr, resp.)[17], and the urinary levels of cross-linked N-terminaltelopeptide of type I collagen (NTX) were measured usingan enzyme-linked immunosorbent assay (ELISA) (normalrange: 9.3–54.3 nM BCE/mM Cr) [17]. The serum levelsof bone-specific alkaline phosphatase (BAP) were measuredusing a chemiluminescent enzyme immunoassay (CLEIA)(normal range: 7.9–20.9 U/L) [17]. The serum levels ofosteocalcin (OC) were measured using an immunora-diometric assay (IRMA) (normal range: 3.1–12.7 ng/mL).

The serum levels of leptin were measured using a radioim-munoassay (RIA) (normal range: 2.5–21.8 ng/mL).

2.3. DXA Scanning. First, DXA scanning was performed onthe whole body in the supine position using a Norland XR-36instrument (Norland, Fort Atkison, WI, USA). The BMD ofthe whole body, lean body mass, total fat mass, and total fatpercentage were measured. Second, DXA scanning was alsoperformed on the lumbar spine in the supine position, andthe BMD of the lumbar spine (L2–L4) in the anteroposteriorview was measured.

According to the Japanese diagnostic criteria for thediagnosis of osteoporosis [18, 19], patients with a lumbarspine or hip BMD <70% of the young adult mean (YAM)or of 70–80% of the YAM and a history of osteoporoticfractures are diagnosed as having “osteoporosis”. Patientswith a lumbar spine BMD, but not a whole-body BMD,�80% of the YAM and between 70%–80% of the YAMwithout any history of osteoporotic fractures are diagnosedas being “normal” and having “osteopenia”, respectively. Dataregarding the YAM for whole-body BMD is not available.

2.4. Statistical Analysis. A simple linear regression analysiswas used to examine possible correlations between theserum leptin level and age; body weight; height; BMI;lumbar spine BMD; whole-body BMD; lean body mass;total fat mass; total fat percentage; serum and urinary boneturnover markers including serum BAP and osteocalcinand urinary pyridinoline, deoxypyridinoline, and NTX; andthe radiographic grade of knee OA. A multiple regressionanalysis was used to determine factors correlated with theserum leptin level among the factors that were significantlycorrelated with the serum leptin level in a simple linearregression analysis. All statistical analyses were performedusing the Stat View-J5.0 program on a Windows computer. Asignificance level of P < .05 was used for all the comparisons.

3. Results

3.1. Characteristics of Study Subjects. Table 1 shows thecharacteristics of the study subjects. The mean age was66.4 years, and the mean BMI was 24.1 kg/m2 (normalrange: 18.5–25.0 kg/cm2), corresponding to a “normal high”level for Japanese populations. The mean lumbar spineBMD was 0.864 g/cm2, which was 83.1% of the youngadult mean (diagnosed as “normal BMD”). The meanradiographic grade of OA was 2.47. The mean serum leptinlevel was 9.7 ng/mL (within normal range). Although themean urinary pyridinoline and deoxypyridinoline and theserum OC levels were within the normal ranges, the serumBAP and urinary NTX levels were higher than the normalranges [17].

3.2. Correlations of Serum Leptin Level with Various Factors.A simple linear regression analysis showed that five factors(body weight, BMI, whole-body BMD, total fat mass, andtotal fat percentage) but not age, height, lumbar spine BMD,

Arthritis 3

Table 1: Characteristics of study subjects.

Mean ± SD Range

Age (years) 66.4± 8.2 50–88

Height (m) 1.56± 0.05 1.47–1.65

Body weight (kg) 58.4± 8.2 44–80

Body mass index(kg/m2)

24.1± 3.4 16.9–33.3

Lumbar spine BMD(g/cm2)

0.864± 0.204 0.524–1.316

Whole body BMD(g/cm2)

0.866± 0.108 0.647–1.109

Lean body mass (g) 34229± 7548 20720–64203

Total fat mass (g) 24734± 6135 14302–41201

Total fat percent (%) 40.8± 5.3 29.7–51.5

Serum BAP (U/L) 28.1± 9.8 11.7–49.8

Serum OC (ng/mL) 5.7± 2.2 1.7–11.9

Urinarypyridinoline(pmol/μmol Cr)

32.5± 9.3 19.3–63.6

Urinarydeoxypyridinoline(nmol/mmol Cr)

6.5± 2.0 2.8–10.1

Urinary NTX (nmolBCE/mmol Cr)

57.9± 21.4 8.2–95.7

Serum leptin(ng/mL)

9.7± 4.7 1.8–21.1

Radiographic grade 2.47± 1.00 1–4

BMD: bone mineral density, BAP: bone-specific alkaline phosphatase, OC:osteocalcin, NTX: cross-linked N-terminal telopeptides of type I collagen,BCE: bone collagen equivalent, Cr: creatinine.Normal ranges of urinary pyridinoline, deoxypyridinoline, and NTX were17.7–41.9 pmol/μmol Cr and 2.8–7.6 nmol/mmol Cr, and 9.3–54.3 nMBCE/mM Cr, respectively. Normal ranges of serum BAP, OC, and leptin were7.9–20.9 U/L, 3.1–12.7 ng/mL, and 2.5–21.8 ng/mL, respectively.

lean body mass, serum and urinary bone turnover markers,or the radiographic grade of knee OA were significantlycorrelated with the serum leptin level (Table 2). A multipleregression analysis showed that among these five factors,body weight and total fat mass were significantly correlatedwith the serum leptin level (r2 = 0.859, Table 3). Figure 1shows the correlations between the serum leptin level andthe body weight and total fat mass according to simple linearregression analyses.

4. Discussion

The present study confirmed that the serum leptin concen-tration was significantly correlated with the body weightand total fat mass in postmenopausal women with kneeOA. These positive findings are consistent with the well-known fact that the serum leptin concentration is closelycorrelated with fat mass and decreases after weight loss[20]. OA is well known to be strongly correlated with ahigh BMI [20]. Gonzalez-Gay et al. [21] found a positivecorrelation between BMI and serum leptin in patients with

Table 2: Correlations of serum leptin level with various factors bysimple regression analysis.

Correlation coefficient P value

Age (years) −0.240 NS

Height (m) 0.162 NS

Body weight (kg) 0.792 <.0001

Body mass index(kg/m2)

0.693 <.0001

Lumbar spine BMD(g/cm2)

0.255 NS

Whole body BMD(g/cm2)

0.440 .0014

Lean body mass (g) 0.189 NS

Total fat mass (g) 0.706 <.0001

Total fat percent (%) 0.511 <.0001

Serum BAP (U/L) −0.085 NS

Serum bone Glaprotein (ng/mL)

−0.267 NS

Urinary pyridinoline(pmol/μmol Cr)

0.138 NS

Urinarydeoxypyridinoline(nmol/mmol Cr)

0.086 NS

Urinary NTX (nmolBCE/mmol Cr)

−0.151 NS

Radiographic grade −0.110 NS

A simple linear regression analysis was used to examine correlations ofserum leptin level with various factors. BMD: bone mineral density, BAP:bone-specific alkaline phosphatase, NTX: cross-linked N-terminal telopep-tides of type I collagen, BCE: bone collagen equivalent, Cr: creatinine, NS:not significant.

severe rheumatoid arthritis receiving tumor necrosis factor-α (TNF-α) antagonist therapy. In the present study, a simplelinear regression analysis showed a similar correlation inpatients with knee OA.

The relationship between obesity and OA is an importantpublic health issue. An experimental study showed thatextreme obesity arising from impaired leptin signalinginduced alterations in subchondral bone morphology with-out increasing the incidence of knee OA, suggesting thatbody fat, in and of itself, might not be a risk factor forjoint degeneration [22]. Obesity alone does not cause kneeOA, and leptin might be involved in OA because withoutleptin, obesity itself does not predispose an individual to OA[22]. Leptin receptors have been found in articular cartilage,further implying that leptin synthesis and secretion mightplay a role in OA [13]. Leptin might play a catabolic rolein cartilage metabolism and may be a disadvantage factorinvolved in the pathological process of OA [23]. Thus, thehigh concentration of serum leptin in obese individualsmight be associated with an increased risk of knee OA.As a result of the effects of sex hormones, the serumleptin concentration is higher in women than in men, evenafter adjustments for BMI, and this difference might berelevant to the influence of gender on the development and

4 Arthritis

Table 3: Correlations of serum leptin level with five factors by multiple regression analysis.

Regression coefficient Standard error P value

Body weight (kg) 0.331 0.123 .0098

Body mass index (kg/m2) −0.119 0.283 NS

Whole body BMD (g/cm2) 1.776 4.926 NS

Total fat mass (g) 3.145E − 4 1.943E − 4 .0275

Total fat percent (%) −0.087 0.178 NS

A multiple regression analysis was used to examine correlations of serum leptin level with five factors that had a significant correlation by a simple regressionanalysis. BMD: bone mineral density, NS: not significant.

40

45

50

55

60

65

70

75

80

85

0 2.5 5 7.5 10 12.5 15 17.5 20 22.5

Leptin (ng/mL)

Bod

yw

eigh

t(k

g)

Y = 45.23 + 1.349∗X ;R2 = 0.596

(a)

0 2.5 5 7.5 10 12.5 15 17.5 20 22.5

Leptin (ng/mL)

10

15

20

25

30

35

40

45×103

Tota

lmas

s(g

)

Y = 15699.583 + 928.319∗X ;R2 = 0.498

(b)

Figure 1: Correlations between the serum leptin level and body weight and total fat mass. A simple linear regression analysis was used toexamine correlations of serum leptin level with body weight and total fat mass. Body weight and total fat mass had a significant positivecorrelation with serum leptin level (both P < .0001).

frequency of OA [24]. However, the present study did notshow any significant correlation between the serum leptinconcentration and the radiographic grade of knee OA. Thus,the elevated serum leptin concentrations might not haveresulted in an increased risk of knee OA in our subjects.

Studies have shown that patients with knee OA mighthave a higher bone mineral content (BMC) and bone size,although the relation between knee OA and BMD remainsto be established [25–34]. However, Weiss et al. [7] suggestthat obese people might have a relatively low level of boneloss and a decreased risk of osteoporosis. The bone-sparingeffects of body weight might possibly be caused by increasedloading on the skeleton, the increased production of estrogenin adipose tissue, the increased bone formation caused by theanabolic effects of high levels of insulin, and/or high levelsof bone-related hormones such as leptin [7]. The presentstudy showed that postmenopausal women with knee OA(mean age: 66.4 years) had a normal lumbar spine BMDand a “normal high” BMI, despite the fact that women olderthan 65 years have an increased risk of osteoporosis [35].The normal lumbar spine BMD might partly be attributableto the “normal high” BMI and the subsequent increasesin loading to the skeleton. The significant correlationbetween the serum leptin concentration and whole-bodyBMD observed in the simple regression analysis disappeared

in a multiple regression analysis, suggesting an interactionbetween body weight or total fat mass and whole-bodyBMD. The urinary NTX and serum BAP levels in ourstudy subjects suggested that bone turnover was mildlyincreased, compared with in healthy premenopausal women[17], as a result of menopause, similar to previously reportedpostmenopausal women with osteoporosis [36]. The serumleptin concentration was not significantly correlated with theBMD and bone turnover markers, probably because it wasnot high enough to influence bone formation and resorptionand, subsequently, the BMD.

Joint space narrowing, sclerosis of the subchondral bone,and the presence of osteophytes are typical structural featuresof knee OA. Thus, both articular cartilage and subchondralbone are considered to be involved in the pathogenesis ofknee OA. However, recent evidence has shown that increasedlocal bone turnover, decreased BMC and stiffness, andtrabecular bone loss have been observed in the subchondralbone structure of knee OA [37–39]; therefore, subchondralbone abnormalities might be a major factor in diseaseprogression. The Framingham Study showed that a highBMD and BMD gain decreased the risk of radiographicknee OA progression [40]. Thus, the subchondral bone massand structure and the whole-body BMD, if it influences thesubchondral bone, might be associated with the progression

Arthritis 5

of knee OA. In the present study, however, no significantassociation was found between the radiographic grade ofknee OA and the lumbar spine and whole-body BMD (datanot shown), probably because of the small sample sizes in thesubanalyses.

Weight loss is correlated with a decrease in the progres-sion of OA, and the serum leptin concentration decreasesafter weight loss [20]. Thus, studying the effect of inter-ventions aimed at weight loss on the progression of kneeOA, serum leptin concentration, and BMD in patients withknee OA would be interesting. Adiponectin, an adipose-modulated biochemical signal, might play an important rolein the maintenance of the total BMC and regional BMDin physically active older women [10]. Thus, examining thecorrelations among serum adiponectin concentration, theradiographic grade of knee OA, BMD, body size, and fatmass in women would also be interesting. Further researchis needed to clarify these issues.

In conclusion, the present study evaluated the serumleptin concentration, the radiographic grade of knee OA,BMD, body size, and fat mass in postmenopausal womenwith knee OA. The serum leptin concentration and thelumbar spine BMD were normal. However, the BMI was“normal high,” probably because of a mild impairment inthe central mechanisms of leptin signaling. The present studyrevealed that body weight and total fat mass, but not theBMD or the radiographic grade of OA, exhibited a significantpositive correlation with the serum leptin concentration inpostmenopausal women with knee OA.

References

[1] Y. Zhang, R. Proenca, M. Maffei, M. Barone, L. Leopold, and J.M. Friedman, “Positional cloning of the mouse obese gene andits human homologue,” Nature, vol. 372, no. 6505, pp. 425–432, 1994.

[2] R. V. Considine, M. K. Sinha, M. L. Heiman et al., “Serumimmunoreactive-leptin concentrations in normal-weight andobese humans,” The New England Journal of Medicine, vol. 334,no. 5, pp. 292–295, 1996.

[3] V. Cirmanova, M. Bayer, L. Starka, and K. Zajıckova, “Theeffect of leptin on bone—an evolving concept of action,”Physiological Research, vol. 57, no. 1, supplement, pp. S143–S151, 2008.

[4] S. Takeda, F. Elefteriou, R. Levasseur et al., “Leptin regulatesbone formation via the sympathetic nervous system,” Cell, vol.111, no. 3, pp. 305–317, 2002.

[5] F. Elefteriou, J. D. Ahn, S. Takeda et al., “Leptin regulationof bone resorption by the sympathetic nervous system andCART,” Nature, vol. 434, no. 7032, pp. 514–520, 2005.

[6] F. Elefteriou, S. Takeda, K. Ebihara et al., “Serum leptin level isa regulator of bone mass,” Proceedings of the National Academyof Sciences of the United States of America, vol. 101, no. 9, pp.3258–3263, 2004.

[7] L. A. Weiss, E. Barrett-Connor, D. Von Muhlen, and P. Clark,“Leptin predicts BMD and bone resorption in older womenbut not older men: the Rancho Bernardo Study,” Journal ofBone and Mineral Research, vol. 21, no. 5, pp. 758–764, 2006.

[8] M. Yamauchi, T. Sugimoto, T. Yamaguchi et al., “Plasmaleptin concentrations are associated with bone mineral density

and the presence of vertebral fractures in postmenopausalwomen,” Clinical Endocrinology, vol. 55, no. 3, pp. 341–347,2001.

[9] M. Sato, N. Takeda, H. Sarui et al., “Association betweenserum leptin concentrations and bone mineral density, andbiochemical markers of bone turnover in adult men,” Journalof Clinical Endocrinology and Metabolism, vol. 86, no. 11, pp.5273–5276, 2001.

[10] J. Jurimae, T. Kums, and T. Jurimae, “Adipocytokine andghrelin levels in relation to bone mineral density in physicallyactive older women: longitudinal associations,” EuropeanJournal of Endocrinology, vol. 160, no. 3, pp. 381–385, 2009.

[11] M. Sowers, “Epidemiology of risk factors for osteoarthritis:systemic factors,” Current Opinion in Rheumatology, vol. 13,no. 5, pp. 447–451, 2001.

[12] T. Thomas and B. Burguera, “Is leptin the link between fat andbone mass?” Journal of Bone and Mineral Research, vol. 17, no.9, pp. 1563–1569, 2002.

[13] A. J. Teichtahl, A. E. Wluka, J. Proietto, and F. M. Cicuttini,“Obesity and the female sex, risk factors for knee osteoarthritisthat may be attributable to systemic or local leptin biosynthesisand its cellular effects,” Medical Hypotheses, vol. 65, no. 2, pp.312–315, 2005.

[14] J. H. Kellgren and J. S. Lawrence, “Radiological assessment ofosteo-arthrosis,” Annals of the Rheumatic Diseases, vol. 16, no.4, pp. 494–502, 1957.

[15] D. T. Felson, “Osteoarthritis of the knee,” The New EnglandJournal of Medicine, vol. 354, no. 8, pp. 841–848, 2006.

[16] N. Bellamy, J. Campbell, V. Robinson, T. Gee, R. Bourne,and G. Wells, “Viscosupplementation for the treatment ofosteoarthritis of the knee,” Cochrane Database of SystematicReviews, no. 2, Article ID CD005321, 2005.

[17] Y. Nishizawa, T. Nakamura, H. Ohta et al., “Guidelines for theuse of biochemical markers of bone turnover in osteoporosis(2004),” Journal of Bone and Mineral Metabolism, vol. 23, no.2, pp. 97–104, 2005.

[18] H. Orimo, Y. Sugioka, M. Fukunaga et al., “Diagnosticcriteria of primary osteoporosis,” Journal of Bone and MineralMetabolism, vol. 16, no. 3, pp. 139–150, 1998.

[19] H. Orimo, Y. Hayashi, M. Fukunaga et al., “Diagnostic criteriafor primary osteoporosis: year 2000 revision,” Journal of Boneand Mineral Metabolism, vol. 19, no. 6, pp. 331–337, 2001.

[20] L. J. Sandell, “Obesity and osteoarthritis: is leptin the link?”Arthritis and Rheumatism, vol. 60, no. 10, pp. 2858–2860,2009.

[21] M. A. Gonzalez-Gay, M. T. Garcia-Unzueta, A. Berja et al.,“Anti-TNF-α therapy does not modulate leptin in patientswith severe rheumatoid arthritis,” Clinical and ExperimentalRheumatology, vol. 27, no. 2, pp. 222–228, 2009.

[22] T. M. Griffin, J. L. Huebner, V. B. Kraus, and F. Guilak,“Extreme obesity due to impaired leptin signaling in mice doesnot cause knee osteoarthritis,” Arthritis and Rheumatism, vol.60, no. 10, pp. 2935–2944, 2009.

[23] J. P. Bao, W. P. Chen, J. Feng, P. F. Hu, Z. L. Shi, and L. D. Wu,“Leptin plays a catabolic role on articular cartilage,” MolecularBiology Reports, vol. 37, no. 7, pp. 3265–3272, 2009.

[24] O. Gualillo, “Further evidence for leptin involvement incartilage homeostases,” Osteoarthritis and Cartilage, vol. 15,no. 8, pp. 857–860, 2007.

[25] J. Iwamoto, T. Takeda, and S. Ichimura, “Forearm bonemineral density in postmenopausal women with osteoarthritisof the knee,” Journal of Orthopaedic Science, vol. 7, no. 1, pp.19–25, 2002.

6 Arthritis

[26] K. Naitou, K. Kushida, M. Takahashi, T. Ohishi, and T. Inoue,“Bone mineral density and bone turnover in patients withknee osteoarthritis compared with generalized osteoarthritis,”Calcified Tissue International, vol. 66, no. 5, pp. 325–329, 2000.

[27] M. Sowers, L. Lachance, D. Jamadar et al., “The associationsof bone mineral density and bone turnover markers withosteoarthritis of the hand and knee in pre- and peri-menopausal women,” Arthritis and Rheumatism, vol. 42, no.3, pp. 483–489, 1999.

[28] M. T. Hannan, J. J. Anderson, Y. Zhang, D. Levy, and D.T. Felson, “Bone mineral density and knee osteoarthritis inelderly men and women: the Framingham study,” Arthritis andRheumatism, vol. 36, no. 12, pp. 1671–1680, 1993.

[29] D. J. Hart, C. Cronin, M. Daniels, T. Worthy, D. V. Doyle, andT. D. Spector, “The relationship of bone density and fractureto incident and progressive radiographic osteoarthritis of theknee: the Chingford study,” Arthritis and Rheumatism, vol. 46,no. 1, pp. 92–99, 2002.

[30] A. P. Bergink, A. G. Uitterlinden, J. P. T. M. Van Leeuwen, A.Hofman, J. A. N. Verhaar, and H. A. P. Pols, “Bone mineraldensity and vertebral fracture history are associated withincident and progressive radiographic knee osteoarthritis inelderly men and women: the Rotterdam study,” Bone, vol. 37,no. 4, pp. 446–456, 2005.

[31] H. Atalar, B. Yanik, B. Ozcakar, E. Atalar, and A. Koktener,“Bone mineral density is not related to severity of osteoarthri-tis in the knee in postmenopausal women,” RheumatologyInternational, vol. 28, no. 3, pp. 233–236, 2008.

[32] D. L. Schneider, E. Barrett-Connor, J. Deborah, and M.M. Weisman, “Bone mineral density and clinical handosteoarthritis in elderly men and women: the e RanchoBernardo Study,” Journal of Rheumatology, vol. 29, no. 7, pp.1467–1472, 2002.

[33] M. C. Hochberg, M. Lethbridge-Cejku, W. W. Scott, R.Reichle, C. C. Plato, and J. D. Tobin, “Upper extremity bonemass and osteoarthritis of the knees: data from the Baltimorelongitudinal study of aging,” Journal of Bone and MineralResearch, vol. 10, no. 3, pp. 432–438, 1995.

[34] M. Abdin-Mohamed, K. Jameson, E. M. Dennison, C. Cooper,and N. K. Arden, “Volumetric bone mineral density of thetibia is not increased in subjects with radiographic kneeosteoarthritis,” Osteoarthritis and Cartilage, vol. 17, no. 2, pp.174–177, 2009.

[35] S. Fujiwara, “Effectiveness of screening for osteoporosis bybone density measurement for the prevention of fractures: areview of the evidence,” Nippon Eiseigaku Zasshi, vol. 58, no.3, pp. 338–346, 2003 (Japanese).

[36] T. Matsumoto, H. Hagino, M. Shiraki et al., “Effect ofdaily oral minodronate on vertebral fractures in Japanesepostmenopausal women with established osteoporosis: a ran-domized placebo-controlled double-blind study,” OsteoporosisInternational, vol. 20, no. 8, pp. 1429–1437, 2009.

[37] T. D. Spector, “Bisphosphonates: potential therapeutic agentsfor disease modification in osteoarthritis,” Aging, vol. 15, no.5, pp. 413–418, 2003.

[38] P. Bettica, G. Cline, D. J. Hart, J. Meyer, and T. D. Spector,“Evidence for increased bone resorption in patients withprogressive knee osteoarthritis: longitudinal results from theChingford study,” Arthritis and Rheumatism, vol. 46, no. 12,pp. 3178–3184, 2002.

[39] B. Li and R. M. Aspden, “Mechanical and material propertiesof the subchondral bone plate from the femoral head of

patients with osteoarthritis or osteoporosis,” Annals of theRheumatic Diseases, vol. 56, no. 4, pp. 247–254, 1997.

[40] Y. Zhang, M. T. Hannan, C. E. Chaisson et al., “Bone mineraldensity and risk of incident and progressive radiographic kneeosteoarthritis in women: the Framingham Study,” Journal ofRheumatology, vol. 27, no. 4, pp. 1032–1037, 2000.

Hindawi Publishing CorporationArthritisVolume 2011, Article ID 454873, 9 pagesdoi:10.1155/2011/454873

Review Article

Current Surgical Treatment of Knee Osteoarthritis

Karolin Ronn, Nikolaus Reischl, Emanuel Gautier, and Matthias Jacobi

Department of Orthopaedic Surgery, Hopital Cantonal Fribourg, 1708 Fribourg, Switzerland

Correspondence should be addressed to Matthias Jacobi, ortho@mjacobi.ch

Received 14 August 2010; Revised 4 January 2011; Accepted 28 February 2011

Academic Editor: Annamaria Iagnocco

Copyright © 2011 Karolin Ronn et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Osteoathritis (OA) of the knee is common, and the chances of suffering from OA increase with age. Its treatment should beinitially nonoperative—and requires both pharmacological and nonpharmacological treatment modalities. If conservative therapyfails, surgery should be considered. Surgical treatments for knee OA include arthroscopy, cartilage repair, osteotomy, and kneearthroplasty. Determining which of these procedures is most appropriate depends on several factors, including the location, stageof OA, comorbidities on the one side and patients suffering on the other side. Arthroscopic lavage and debridement is often carriedout, but does not alter disease progression. If OA is limited to one compartment, unicompartmental knee arthroplasty or unloadingosteotomy can be considered. They are recommended in young and active patients in regard to the risks and limited durability oftotal knee replacement. Total arthroplasty of the knee is a common and safe method in the elderly patients with advanced kneeOA. This paper summarizes current surgical treatment strategies for knee OA, with a focus on the latest developments, indicationsand level of evidence.

1. Introduction

Osteoarthritis (OA) of the knee is the commonest joint dis-order in the elderly, with a prevalence of about 30% inadults aged >60 years [1]. About half of these subjects willshow symptoms such as joint pain, stiffness, effusion andlimitation of joint function. With our aging population, theprevalence of OA in the “developed” world is expected toincrease. It is anticipated that OA will become the fourthleading cause of disability in the coming decades [2].

The etiology of knee OA is multifactorial and includesgeneralized constitutional factors (e.g., aging, sex, obesity,heredity, and reproductive variables), local adverse mechan-ical factors (e.g., joint trauma, occupational and recreationalabuse, alignment, and postmeniscectomy), and geographicfactors. There is a significant genetic component to theprevalence of knee OA, with heritability estimates from twinstudies of 0.39–0.65 independent of known environmentalor demographic confounders [3]. Genetic variations lead tochondrocyte alterations resulting in osteoarthritis [4, 5].

Diagnostic criteria for OA of the knee include patient his-tory, physical examination, and radiologic and laboratoryfindings [6]. However, the standard radiograph alone allowsin most patients definitive diagnosis of knee OA. Other

radiological modalities such as computer tomography, ultra-sound imaging, MRI and bone scan can provide alternativeor supplementary information [7].

The OA Research Society International (OARSI) has pub-lished global, evidence-based, consensus recommendationsfor the treatment of OA of the hip and knee [8–10]. Of the51 modalities of treatment addressed in the OARSI recom-mendations, 35 have been systematically reviewed includinga wide range of nonsurgical methods (e.g., physiotherapy,bracing, education, weight reduction, viscosupplementa-tion, corticoid injections, analgesia, other anti-inflammatorytreatments, etc.). Initial treatment of knee OA should beconservative. Only if symptoms persist after the appropriateuse of nonsurgical treatment, surgery should be considered.Surgical treatment options are arthroscopic debridement,cartilage repair surgery, osteotomy with axis-correction, andunicompartmental or total knee arthroplasty (TKA). We willfocus on the latest.

Surgical indication and choice of treatment is based onsymptoms (e.g., pain and knee function), OA stage, andpatient-related factors such as age, level of physical activity,and patient’s comorbidities. Radiological evidence of OAalone (joint space narrowing, osteophytes, etc.) does notjustify surgical intervention, which is indicated only in

2 Arthritis

combination with relevant symptoms. Finally, it is thepatient’s degree of suffering, in correlation to radiologicalevidence of OA, which determines the time point of surgery.It is important that indication with OA, surgery is alwaysa relative indication. Only in case of progressive kneeinstability associated to OA surgical treatment (total kneearthroplasty) should not be unnecessary delayed. The choiceof surgical treatment, however, underlies in general practicepersonal, regional, and industry-influenced preferences asindications for different surgical and nonsurgical treatmentmodalities interfere with each other.

The present paper will discuss accepted surgical treat-ment options in knee OA. We focus on the latest develop-ments, indications, and the chosen treatment’s efficiency.

2. Surgical OA Treatment

2.1. Arthroscopic Lavage and Debridement. Arthroscopictechniques include lavage and debridement of the knee (e.g.,shaving of rough cartilage or smoothening of the degen-erated meniscus). In theory, arthroscopy for OA shouldrelieve symptoms by removing the debris and inflammatorycytokines that cause synovitis [11, 12]. Debridement canremove torn meniscal fragments and loose cartilage flaps.However the role of arthroscopy in treating knee OA iscontroversial [8–10]. Although widely used, there is a lackof evidence showing it to have a significant benefit. Acontrolled trial study by Moseley et al. [13] showed thatthere was no benefit comparing arthroscopic lavage anddebridement with shame surgery. In 2007, Siparsky et al.carried out an evidence-based review of the literature onthe arthroscopic treatment of knee OA and found limitedsupport for its use [14]. Dervin et al. [15] showed theimportance of patient selection before knee arthroscopy.Patients with evident lesions of the meniscus or cartilageflaps may benefit from surgery. Another study confirms that,in well-selected middle-aged patients with knee arthritis,arthroscopic debridement may be valuable for providingtransient relief of symptoms [16]. Patients with less extensivearthritis as seen by radiography, less severe involvementof articular cartilage, and a younger age at the time ofsurgery have higher probability of improvement [17]. Ashort duration of pain and mechanical symptoms and mild-to-moderate radiographic stages of arthritis correlate witha better result [14, 18]. However, two recent Cochranereviews [18, 19] of arthroscopic lavage and debridement forknee OA identified only three well-designed studies [10, 13,16] and concluded from these that the procedure has nobenefit for OA arising from mechanical or inflammatorycauses. On the basis of available evidence, arthroscopiclavage seems to provide only short-term benefit to selectedpatients with mild radiographic OA and effusion. Arthro-scopic debridement should not be used as routine treatmentfor knee OA, although patients with symptomatic menis-cal tears and loose bodies with locking symptoms couldbenefit.

Quantification of the benefits has been limited by meth-odological problems and limited analyses in many studies

[20]. It is an outpatient procedure with less serious potentialcomplications than other surgical treatments for OA. Thepostoperative course is predictable, and the risk of compli-cations is acceptably small for most patients. It does notpreclude later definitive surgery, and so patient and surgeonmay feel it is “worth a try.” Nevertheless, it cannot alter theprogression of OA; it may only be a helpful instrument toreduce pain in well-selected patients.

2.2. Cartilage Repair Techniques. Damaged articular cartilagehas only limited or no healing capacity [21]. Repair ofthe cartilage surface has therefore been proposed. Howevercartilage repair is only indicated for focal cartilage defects,which can been seen as a precursor of OA. If the defectis to extended cartilage, repair is no longer indicated. Thedifferent techniques can be divided in bone marrow stim-ulating techniques like abrasion, drilling, or microfractureas well as in replacement techniques like mosaicplasty orosteochondral allograft transplantation and in grafting andcombined techniques like periost flap transplantation, andautologous chondrocyte implantation (ACI), autologousmatrix induced chondrogenesis (AMIC).

2.2.1. Bone Marrow Stimulating Techniques. Penetration ofthe subchondral lamina has been shown to promote cartilagerepair tissue; indeed, pluripotent stem cells arising from thesubchondral bone marrow may promote chondrogenesis inthe defect area. This technique enhances chondral resurfac-ing and takes advantage of the healing potential of the body.Pridie was the first to describe a technique whereby he useda drill to penetrate the often sclerotic subchondral lamina[22]. In former times, this was provided by an arthrotomyof the joint. Nowadays, it is usually done employing themicrofracture technique described by Steadman et al. [23–26]. Using an awl, holes which penetrate 2–4 mm into thesubchondral lamina are made at a distance of 3-4 mm fromeach other. This is relatively simple and can be done arthro-scopically. The low-cost and simplicity of this technique havepermitted its wide use. The disadvantages of the techniqueinclude limited hyaline repair tissue, variable repair cartilagevolume, and possible functional deterioration [27].

2.2.2. Osteochondral Transplantation Techniques. Recon-struction of a cartilaginous surface or of osteocartilaginousdefects can be done by transplantation of osteochondralgrafts. The graft can be autologous or allogenic. Autologoustransfer is termed “mosaicplasty” or the osteochondralautologous transfer system (OATS). These terms are usedsynonymously. It is done by taking one or several cylindrical“plugs” from the peripheries of the femoral condyles at thelevel of the patellofemoral joint, and the plugs are transferredto the defect with a special cutting devise [28–35]. Theprocedure can be open (for large defects) or arthroscopic (forsmall defects) [36]. The advantages of this technique are theuse of a bone-cartilage graft consisting of hyaline cartilage,replacing also the often affected underlying bone. Minorintegration, limited graft availability and technical difficultiesare the disadvantages of the procedure.

Arthritis 3

Figure 1: Schematic drawing of autologous cartilage implantation(ACI). The procedure consists of the following steps: (1) cartilageharvest generally performed during arthroscopic surgery, (2) cellculture with expansion of cells in monolayer flasks, and (3)reimplantation of the cells by injecting them underneath a suturedcollagen membrane.

2.2.3. Autologous Chondrocyte Implantation (ACI). In 1994,Brittberg presented the ACI technique whereby cultivatedand proliferated autologous chondrocytes are re-implantedunderneath a periosteal flap [37]. Chondrocytes are har-vested in a first procedure in which a small cartilage probeis taken arthroscopically. The cartilage is then digested andthe harvested cells expanded during 3-4 weeks in monolayerculture before implantation (Figure 1). Nowadays, theperiost membrane is replaced by a collagen membrane, andcell culture is improved by applying growth factors or byculturing cells in a three-dimensional collagen scaffold whichcan be directly implanted [38]. The disadvantages of thistechnique are the two-stage procedure and the costs of thecell culture.

Main indications for cartilage repair techniques are lim-ited size cartilage lesions especially in younger patients. Ifcartilage damage tends towards an osteoarthritic lesion, car-tilage repair procedures are not indicated. Exclusive cartilagerepair will not be successful if axial malalignment, ligamen-tous instability, or patella maltracking is the underlying causeor is associated with the cartilage lesion. Once more one ofthe key elements of successful surgery is correct indication.Diagnosis is facilitated by improved MRI techniques [39, 40].Nevertheless, many isolated cartilage lesions are recognizedonly during arthroscopy [41]. These incidental findings(which are found during arthroscopy or based on MRI)make choosing the correct treatment quite difficult. If bonenecrosis is present, debridement and bone grafting must beconsidered as a concomitant procedure. The use of ACI andother chondral resurfacing techniques is becoming increas-ingly widespread. The prevalence of symptoms after cartilagerepair procedures has been shown to decrease. Randomized

controlled trials have been done comparing ACI, microfrac-ture, and mosaicplasty [42–44]. Nevertheless, evidence ofa significant difference between ACI and other interven-tions is lacking [45]. Additional good-quality randomizedcontrolled trials with long-term functional outcomes arerequired.

2.3. Osteotomies around the Knee. Osteotomies around theknee are an accepted method for the treatment of unicom-partmental OA with associated varus or valgus deformity.Osteotomies have been carried out since the nineteenthcentury [46]. Although osteotomies were done regularly inthe first half of the twentieth century, the real breakthroughcame only with the publications of Jackson, Waugh, Gariepy,Coventry, and others in the late 1950s and 1960s [47–50]. Osteotomy became a standard treatment option forunicompartmental OA of the knee. The classic osteotomyof Coventry was a closed-wedge valgization including afibula osteotomy and was carried out proximal to the tibialtuberosity [50]. This was the most widely used technique fora long time. In the 1980s and 1990s, osteotomy around theknee lost importance due to the success of knee arthroplasty.Compared with arthroplasty, osteotomy was considered ademanding procedure with an unpredictable outcome andwas associated with significant complications. During the lastdecade, the development of new plates (particularly plateswith angular stability) and the tendency for open-wedgeosteotomy without bone graft interposition and absence ofrisk of damage to the peroneal nerve have led to a revival ofosteotomy around the knee, particularly for younger patients[51–53].

Osteotomies around the knee alter the weightbearing axisof the lower extremity [54]. The aim is to unload the dam-aged compartment and to transfer the weight load from theaffected areas by slightly overcorrecting into a valgus or varusaxis to reduce pain, slow the degenerative process, and delayjoint replacement [50, 55, 56].

Fundamental for a satisfactory postoperative outcomeis appropriate patient selection, including evaluation of allthree knee compartments. The classic inclusion criterionis OA of one compartment in combination with varus orvalgus alignment. The femoropatellar compartment shouldnot be affected by OA. Good mobility of the knee is aprerequisite, as well as ligament stability. Instability is not anabsolute contraindication because cruciate ligaments can bereconstructed together with correction of the axis [57, 58].Age is a significant factor to consider. Age >60–65 years is arelative contraindication, whereas biologic age and activitymust also be considered. Obesity and chondrocalcinosisare not strict contraindications, but the success rate andprognosis are compromised. Before osteotomy, it is ideal toconfirm clinical and radiographic findings by arthroscopyof the knee to ensure that the unaffected compartment ishealthy. This can be done in the same procedure.

Different techniques are used to correct load axis in uni-compartmental knee OA. This includes proximal tibial headosteotomies and supracondylar femoral osteotomies. Bothcan be done in an additive (open-wedge) or subtractive(closed-wedge) technique, and are regarded as established

4 Arthritis

Figure 2: Unloading osteotomy: exemplary a valgisation open-wedge high tibial osteotomy in unicompartmental OA of the medialknee compartment. The corrected position is stabilized by a platewith angular locked screws.

procedures for the treatment of varus and valgus OA (Fig-ure 2). Valgisation osteotomies are commonly done on theproximal tibia, whereas varisation osteotomies are done onthe femoral side. If the deformity is not located near the jointbut in the diaphysis of the long bones, the correction shouldbe at the site of the deformity [57].

The classic lateral closing-wedge procedure requires afibular osteotomy. This is associated with the risk of damageto the peroneal nerve reported to be up to 11% [52]. Thejoint line has a tendency to end up in an oblique position,which may complicate subsequent placement of the tibialcomponent of a total knee replacement.

Medial opening-wedge techniques of the tibial head areless demanding, more precise, and faster. This is an advan-tage, particularly in combined interventions with cruciateligament reconstruction. Only one saw cut is required, andcorrections in the frontal plane can be combined withadjustments in the sagittal plane. New plates with angularstability have been developed during recent years. Theincreased stability of these plates offers high stability, makingbone grafting dispensable [51–53]. The risk to the peronealnerve is negligible.

Most long-term studies for the closing-wedge valgisationtechnique have shown good results [59]. Good results arereported during the first years of followup, with deteriorationover the time. Insall et al. showed that at the two-yearfollowup, 97% of patients report good results, whereas afterfive years patient satisfaction decreases to 85%, which dips to63% after 9 years [60]. Only one Japanese study showed veryhigh survival (90%) after 15 years [61]. Long-term followupfor the open-wedge techniques using modern implants withlocking screws is not available. Nevertheless, the availablemid-term results are promising, and it may become the newstandard procedure for valgisation of varus OA.

Osteotomies around the knee are an effective procedurein young and active patients with early OA of one com-partment with associated varus or valgus axis. Appropri-ate patient selection, good preoperative planning, accuratesurgical technique, and correct postoperative managementcan minimize the complication rate and lead to satisfactoryoutcome.

Although unloading osteotomies are an accepted and safetreatment modality, no studies have been undertaken tocompare it with placebo or conservative treatment alone.However, it has shown to be efficient in reducing pain andimproving function [8–10]. Further comparative trials arenecessary to define its indication in relation to unicompart-mental or total knee arthroplasties.

2.4. Joint Arthroplasty. Joint arthroplasty is a well-accepted,safe, and cost-effective method for treatment of advancedknee OA. Owing to its irreversible nature, joint arthroplastyis recommended only in patients for whom other treatmentmodalities have failed or are contraindicated. Durabilityof prosthetic components is limited to about 15–20 yearsbut survival of unicompartmental arthroplasties is generallyinferior. Therefore arthroplasties should be avoided inpatients younger than 60 years whenever possible. If OAis limited to one compartment, unicompartmental kneearthroplasty (UKA) or unloading osteotomy can be consid-ered, otherwise TKA with or without patellar resurfacing isindicated.

2.4.1. Unicompartmental Knee Arthroplasty (UKA). Sinceone of the first followup studies reported in the 1970s byMarmor, UKA has received increased interest [62]. UKA isindicated in cases where OA involves only one of thethree compartments of the knee: the medial tibiofemoral,lateral tibiofemoral or patellofemoral compartment. Thecommonest UKA replaces the contact surfaces of the medialtibiofemoral compartment with two metallic prostheticdevices and inserts a polyethylene inlay between them(Figure 3). For successful medial UKA, the initial conditionsmust provide a well-preserved lateral compartment withrespect to meniscus and cartilage [63]. The implant is unre-strained in the sagittal plane, so the stability of the prosthesisdepends on intact cruciate ligaments [64]. Considerablemalalignment of the limb is a contraindication. Overcor-rection to the contralateral compartment must be avoidedbecause it may result in progression of OA and persistingsymptoms [65]. Equally, undercorrection is associated withincreased likelihood of revision and clinical failure of theUKA [66].

One advantage of UKA includes a less invasive surgicaltechnique [67]. In particular, the patella is not evertedand the extensor mechanism is not damaged, permittinga much more rapid recovery and earlier discharge fromhospital. It also provides preservation of bone stock, morenormal knee kinematics, and greater physiological function[68].

The use of modern implants and surgical techniques hasimproved the outcome and survival associated with medial

Arthritis 5

Figure 3: Treatment of an isolated medial compartment OA byunicompartmental arthroplasty.

Figure 4: Treatment of advanced knee OA by total knee arthro-plasty (example without patella resurfacing).

UKA [67]. The 10-year survival for medial unicompartmen-tal knee arthroplasty (UKA) is highly variable and rangesfrom 80.2 to 98% [69, 70]. Anyhow UKA has signicantlypoorer long-term survival than TKA [69]. The target groupof UKA differs from that of TKA. UKA is usually done inyounger patients with less severe disease, who have betterultimate function, but who wear-out their joints morerapidly.

Outcomes for the treatment of lateral unicompartmentalknee OA are rarely reported [71]. These results are lesspredictable than those of medial unicompartmental OA,despite recent improvements in implant design. The femoralcondyle undergoes a greater translation than the medialcondyle during flexion, which may result in instability anddislocation of the tibial insert in mobile-bearing prosthesis[72]. The kinematics of the lateral compartment suggests thata fixed-bearing component may offer a better solution [73].

Isolated femoropatellar OA occurs in 10% of patientswith knee OA. Underlying disorders often include prior trau-ma to the patella, patellar maltracking, trochlear dysplasia,

and degeneration secondary to deep bending and overuse.Few patients undergo isolated patellofemoral replacement,although this number is increasing [74, 75]. Failure ofisolated femoropatellar arthroplasty is more common thanwith femorotibial replacements, and the reasons are still notclearly defined. TKA should be considered also for isolatedfemoropatellar OA, particularly in older patients.

2.4.2. Total Knee Arthroplasty (TKA) (Figure 4). In advancedknee OA, with more than one compartment involved andfailure of conservative treatments, TKA has been shown tobe a highly effective treatment that results in substantialimprovement in patient functioning and health-related qual-ity of life [76]. Until now it has been the first-line procedurefor end-stage knee OA. The long-term results of TKA havebeen well documented with survival rates of up to 98% at 15years. [77]. Results in younger patients are mostly reportedto be inferior with 76% survival rates at 10 years [78].

Although TKA is effective for end-stage arthritis of theknee, postoperative pain occurs or persists in one out of eightpatients despite an absence of clinical or radiological abnor-malities [79]. The main complications are femoropatellarproblems, loosening of components, infections and stiffnessof the knee. There is a correlation between existing comor-bidities of patients and the range of motion and conditionof the knee postoperatively [80]. Nevertheless, there havealso been substantial refinements of understanding in thetreatment of complications. The importance of patient-related factors to outcome of TKA is shown, and thesefactors should influence preoperative counseling of patientsawaiting TKA.

One of the central problems in persistent postoperativepain is the femoropatellar joint. However, a general benefitfrom patella resurfacing has never been proven, and theindications of patella resurfacing are not clearly defined [81,82]. Complications involving the extensor mechanism andthe femoropatellar joint remain the primary noninfectiousindications for revision TKA [83].

Motivated by the sometimes unsatisfactory results, ef-forts have been undertaken during recent years to improvethe outcome of total knee replacement. These strategiesinclude minimally invasive surgery (MIS), intraoperativecontrol with computer-navigated surgery (CAS) or betterinstrumentation, improvements in the biomechanic andanatomic design of the implants, and improvements in thefixation of implants.

Minimal Invasive Surgery (MIS). Most knee arthroplastiesare implanted through a parapatellar medial arthrotomywith splitting of the quadriceps tendon and the retinac-ulum/capsule beside the patella and patellar tendon. Thepatella is usually everted. The so-called “mini-invasivesurgery” avoids splitting of the quadriceps tendon. Accessis made possible through a mid-vastus approach (splittingof the vastus) or a subvastus approach. Eversion of thepatella is avoided. Skin incision is shortened to a minimum.This strategy is thought to have faster recovery times,shorter stays in hospital, fewer problems with patella baja,

6 Arthritis

and improved short-term functional outcomes [83, 84].Critics have raised questions about malalignment of theleg, malpositioning of the implants, and the length of thelearning curve for the procedure [85, 86]. Recent RCTshave failed to show a relevant advantage of this technique[87, 88].

Biomechanic and Anatomic Improvements in Implant Design.TKA copies the physiologic biomechanics of the knee jointpoorly. The course of motion is defined in the physiologicknee mainly by the cruciate ligaments and in the TKA bythe polyethylene inlay. Different types of inlays are avail-able, rotating, fixed-bearing and posterior-stabilized inlays,among others [89]. All of them fail to imitate the originalknee motion with rolling back of the femoral condyles onthe tibial plateau. Clinical results of different types of inlaysare very similar [90]. A newer inlay design, imitating the twocruciate ligaments, is available, but independent long-termfollowup is lacking [91].

Anatomic implant design has been improved by thefollowing points. First, anatomic studies have revealedthat the distal femur is more variable than intended byimplant designs. In particular, a difference between maleand females can be demonstrated. Implants with a newrelationship between the frontal to anteroposterior diameterand adapted Q-angles have been designed. This leads toan expanded implant assortment, but clinical benefit hasnot been documented by a RCT. A second improvement inimplant design is a more anatomically designed trochlea,supporting patella tracking. Third, implants are availablewhich should favor higher flexion of knee prosthesis up to155◦ due to higher posterior condyle offset affecting a higherposterior femoral translation and range of flexion [92]. Theexpected difference between standard knee prosthesis andhigh flexion prosthesis has not been observed in RCTs [93].

Implant Fixation. Cemented fixation of total knee replace-ment is a standard procedure with good long-term dura-bility. The main advantage of noncemented fixation is theshorter operating time. Whereas clinical outcome shows nosignificant difference between cemented and noncementedfixation, a recent study found a statistically significant benefittowards improved survival of the cemented compared withnoncemented components, with followup ranging from 2years to 11 years [94]. Another advantage of cementedfixation is that it is less technically challenging becausebone cuts do not require a perfect fit with the prosthesisand cement can fill the defects [95]. It is less costly andprevents early migration [96] which may potentially lead tolate clinical failures. Cement may also potentially create aneffective barrier to polyethylene debris generated from thearticular surface, thereby preventing osteolysis and implantloosening [97].

Intraoperative Control. A new technology introduced intoTKA is computer-assisted surgery (CAS). Computer navi-gation improves the precision of postoperative alignmentafter TKA [98]. Despite this effect, patients who underwent

navigated TKA did not exhibit improved clinical resultsat two years when compared with patients who had beenmanaged with conventional TKA. Studies do not revealearly benefit of navigated TKA, and long-term studies areneeded to reveal improvements in survival derived fromthe improvement in limb alignment [99]. Disadvantages arethe longer operating time, a learning curve of about 25–30operations, and the costs of the new technology.

Another new technique relies on patient-specific cuttingblocks, which are designed by using the patients MRI orcomputer tomography as template. These individual cuttingblocks allow a precise bone resection adapted to the uniqueshapes and angles of the joint. Surgery is facilitated, bloodloss may be reduced, and the duration of operation isshortened. Disadvantages of this new technology are theadditional costs for the cutting block and the fact that thetechnique relies purely on bony landmarks without payingattention to the ligament balance.

Due to the development of surgical techniques and im-proved implant technology, the outcome and function ofTKA have improved. For successful outcome, good align-ment of the tibial and femoral components (as well as correctpatella tracking) is essential, leading to lower wear of theprosthesis [100]. TKA has become a successful treatmentfor advanced and symptomatic knee OA, particularly inelderly patients. Many new developments and designs havebeen presented and brought to the “medical market” duringrecent years. They are scientifically interesting and must befollowed carefully. Nevertheless, most fall short of provingclinical improvement in the available followup period.Unfortunately, manufacturers regularly misapply these tech-nologies for advertisement purposes, even though evidenceof improvement regarding residual pain level, durability ofthe arthroplasty, and knee function is not present.

References

[1] D. T. Felson, A. Naimark, J. Anderson, L. Kazis, W. Castelli,and R. F. Meenan, “The prevalence of knee osteoarthritis inthe elderly: the Framingham Osteoarthritis Study,” Arthritis& Rheumatism, vol. 30, pp. 914–918, 1987.

[2] World Health Organiazation, The World Health Report 2002:Reducing Risks, Promoting Healthy Life, WHO, Geneva,Switzerland, 2002.

[3] T. D. Spector, F. Cicuttini, J. Baker, J. Loughlin, and D.Hart, “Genetic influences on osteoarthritis in women: a twinstudy,” British Medical Journal, vol. 312, no. 7036, pp. 940–944, 1996.

[4] P. M. van der Kraan, E. N. Blaney Davidson, A. Blom, andW. B. van den Berg, “TGF-beta signaling in chondrocyteterminal differentiation and osteoarthritis. Modulation andintegration of signaling pathways through receptor-Smads,”Osteoarthritis and Cartilage, vol. 17, no. 12, pp. 1539–1545,2009.

[5] A. M. Valdes, T. D. Spector, A. Tamm et al., “Geneticvariation in the SMAD3 gene is associated with hip and kneeosteoarthritis,” Arthritis and Rheumatism, vol. 62, no. 8, pp.2347–2352, 2010.

[6] R. Altman, E. Asch, D. Bloch et al., “Development of cri-teria for the classification and reporting of osteoarthritis:

Arthritis 7

classification of osteoarthritis of the knee,” Arthritis andRheumatism, vol. 29, no. 8, pp. 1039–1052, 1986.

[7] A. Iagnocco, G. Meenagh, L. Riente et al., “Ultrasound imag-ing for the rheumatologist XXIX. Sonographic assessmentof the knee in patients with osteoarthritis,” Clinical andExperimental Rheumatology, vol. 28, no. 5, pp. 643–646, 2010.

[8] W. Zhang, R. W. Moskowitz, G. Nuki et al., “OARSIrecommendations for the management of hip and kneeosteoarthritis—part I: critical appraisal of existing treatmentguidelines and systematic review of current research evi-dence,” Osteoarthritis and Cartilage, vol. 15, no. 9, pp. 981–1000, 2007.

[9] W. Zhang, R. W. Moskowitz, G. Nuki et al., “OARSI rec-ommendations for the management of hip and kneeosteoarthritis—part II: OARSI evidence-based, expert con-sensus guidelines,” Osteoarthritis and Cartilage, vol. 16, no. 2,pp. 137–162, 2008.

[10] W. Zhang, G. Nuki, R. W. Moskowitz et al., “OARSI rec-ommendations for the management of hip and kneeosteoarthritis—part III: changes in evidence following sys-tematic cumulative update of research published throughJanuary 2009,” Osteoarthritis and Cartilage, vol. 18, no. 4, pp.476–499, 2010.

[11] R. W. Chang, J. Falconer, S. D. Stulberg, W. J. Arnold, L. M.Manheim, and A. R. Dyer, “A randomized, controlled trialof arthroscopic surgery versus closed- needle joint lavagefor patients with osteoarthritis of the knee,” Arthritis andRheumatism, vol. 36, no. 3, pp. 289–296, 1993.

[12] D. J. Ogilvie-Harris and D. P. Fitsialos, “Arthroscopic man-agement of the degenerative knee,” Arthroscopy, vol. 7, no. 2,pp. 151–157, 1991.

[13] J. B. Moseley, K. O’Malley, N. J. Petersen et al., “Continuingmedical education: a controlled trial of arthroscopic surgeryfor osteoarthritis of the knee,” The New England Journal ofMedicine, vol. 347, no. 2, pp. 81–88, 2002.

[14] P. Siparsky, M. Ryzewicz, B. Peterson, and R. Bartz, “Arthro-scopic treatment of osteoarthritis of the knee: are thereany evidence-based indications?” Clinical Orthopaedics andRelated Research, vol. 455, pp. 107–112, 2007.

[15] G. F. Dervin, I. G. Stiell, K. Rody, and J. Grabowski, “Effectof arthroscopic debridement for osteoarthritis of the kneeon health-related quality of life,” Journal of Bone and JointSurgery A, vol. 85, no. 1, pp. 10–19, 2003.

[16] M. J. S. Hubbard, “Articular debridement versus washoutfor degeneration of the medial femoral condyle: a five-yearstudy,” Journal of Bone and Joint Surgery B, vol. 78, no. 2, pp.217–219, 1996.

[17] R. W. Jackson and D. W. Rouse, “The results of partialarthroscopic meniscectomy in patients over 40 years of age,”Journal of Bone and Joint Surgery B, vol. 64, no. 4, pp. 481–485, 1982.

[18] J. A. Rand, “Role of arthroscopy in osteoarthritis of the knee,”Arthroscopy, vol. 7, no. 4, pp. 358–363, 1991.

[19] W. Laupattarakasem, M. Laopaiboon, P. Laupattarakasem,and C. Sumananont, “Arthroscopic debridement for kneeosteoarthritis,” Cochrane Database of Systematic Reviews, no.1, Article ID CD005118, 2008.

[20] S. Reichenbach, A. W. Rutjes, E. Nuesch, S. Trelle, and P.Juni, “Joint lavage for osteoarthritis of the knee,” CochraneDatabase of Systematic Reviews, vol. 5, Article ID CD007320,2010.

[21] W. Widuchowski, P. Lukasik, G. Kwiatkowski et al., “Isolatedfull thickness chondral injuries. Prevalance and outcome oftreatment. A retrospective study of 5233 knee arthroscopies,”

Acta Chirurgiae Orthopaedicae et Traumatologiae Cechoslo-vaca, vol. 75, no. 5, pp. 382–386, 2008.

[22] K. H. Pridie, “A method of resurfacing osteoarthritic kneejoints,” Journal of Bone and Joint Surgery, vol. 41, pp. 618–619, 1959.

[23] J. R. Steadman, K. K. Briggs, J. J. Rodrigo, M. S. Kocher,T. J. Gill, and W. G. Rodkey, “Outcomes of microfracturefor traumatic chondral defects of the knee: average 11-yearfollow-up,” Arthroscopy, vol. 19, no. 5, pp. 477–484, 2003.

[24] J. R. Steadman, W. G. Rodkey, and K. K. Briggs, “Microfrac-ture to treat full-thickness chondral defects: surgical tech-nique, rehabilitation, and outcomes,” The Journal of KneeSurgery, vol. 15, no. 3, pp. 170–176, 2002.

[25] J. R. Steadman, W. G. Rodkey, K. K. Briggs, and J. J. Rodrigo,“The microfracture technic in the management of completecartilage defects in the knee joint,” Orthopaedics, vol. 28, pp.26–32, 1999.

[26] J. R. Steadman, W. G. Rodkey, and J. J. Rodrigo, “Microfrac-ture: surgical technique and rehabilitation to treat chondraldefects,” Clinical Orthopaedics and Related Research, supple-ment 391, pp. S362–S369, 2001.

[27] K. Mithoefer, T. Mcadams, R. J. Williams, P. C. Kreuz, andB. R. Mandelbaum, “Clinical efficacy of the microfracturetechnique for articular cartilage repair in the knee: anevidence-based systematic analysis,” American Journal ofSports Medicine, vol. 37, no. 10, pp. 2053–2063, 2009.

[28] L. Hangody, P. Feczko, L. Bartha, G. Bodo, and G. Kish,“Mosaicplasty for the treatment of articular defects of theknee and ankle,” Clinical Orthopaedics and Related Research,supplement 391, pp. S328–S336, 2001.

[29] L. Hangody and P. Fules, “Autologous osteochondral mosaic-plasty for the treatment of full-thickness defects of weight-bearing joints: ten years of experimental and clinical expe-rience,” Journal of Bone and Joint Surgery A, vol. 85, no. 1,supplement 2, pp. 25–32, 2003.

[30] L. Hangody and Z. Karpati, “New possibilities in themanagement of severe circumscribed cartilage damage inthe knee,” Magyar Traumatologia, Ortopedia, Kezsebeszet,Plasztikai Sebeszet, vol. 37, no. 3, pp. 237–243, 1994.

[31] L. Hangody, G. Kish, Z. Karpati, I. Udvarhelyi, I. Szigeti, andM. Bely, “Mosaicplasty for the treatment of articular cartilagedefects: application in clinical practice,” Orthopedics, vol. 21,no. 7, pp. 751–756, 1998.

[32] L. Hangody, G. K. Rathonyi, Z. Duska, G. Vasarhelyi, P. Fules,and L. Modis, “Autologous osteochondral mosaicplasty,”Journal of Bone and Joint Surgery A, vol. 86, supplement 1,pp. 65–72, 2004.

[33] L. Hangody, G. Vasarhelyi, L. R. Hangody et al., “Autologousosteochondral grafting-Technique and long-term results,”Injury, vol. 39, supplement 1, pp. 32–39, 2008.

[34] E. Gautier, D. Kolker, and R. P. Jakob, “Treatment of cartilagedefects of the talus by autologous osteochondral grafts,”Journal of Bone and Joint Surgery B, vol. 84, no. 2, pp. 237–244, 2002.

[35] R. P. Jakob, P. Mainil-Varlet, and E. Gautier, “Isolated articu-lar cartilage lesion: repair or regeneration,” Osteoarthritis andCartilage, vol. 9, supplement A, pp. S3–S5, 2001.

[36] L. Hangody, G. Kish, Z. Karpati, I. Szerb, and I. Udvarhelyi,“Arthroscopic autogenous osteochondral mosaicplasty forthe treatment of femoral condylar articular defects: a prelim-inary report,” Knee Surgery, Sports Traumatology, Arthroscopy,vol. 5, no. 4, pp. 262–267, 1997.

[37] M. Brittberg, A. Lindahl, A. Nilsson, C. Ohlsson, O. Isaks-son, and L. Peterson, “Treatment of deep cartilage defects in

8 Arthritis

the knee with autologous chondrocyte transplantation,” TheNew England Journal of Medicine, vol. 331, no. 14, pp. 889–895, 1994.

[38] D. B. F. Saris, J. Vanlauwe, J. Victor et al., “Characterizedchondrocyte implantation results in better structural repairwhen treating symptomatic cartilage defects of the knee in arandomized controlled trial versus microfracture,” AmericanJournal of Sports Medicine, vol. 36, no. 2, pp. 235–246, 2008.

[39] H. G. Potter and L. R. Chong, “Magnetic resonance imagingassessment of chondral lesions and repair,” Journal of Boneand Joint Surgery A, vol. 91, supplement 1, pp. 126–131, 2009.

[40] H. G. Potter and L. F. Foo, “Magnetic resonance imagingof articular cartilage: trauma, degeneration, and repair,”American Journal of Sports Medicine, vol. 34, no. 4, pp. 661–677, 2006.

[41] K. Hjelle, E. Solheim, T. Strand, R. Muri, and M. Brittberg,“Articular cartilage defects in 1,000 knee arthroscopies,”Arthroscopy, vol. 18, no. 7, pp. 730–734, 2002.

[42] R. Gudas, E. Stankevicius, E. Monastyreckiene, D. Pranys,and R. J. Kalesinskas, “Osteochondral autologous transplan-tation versus microfracture for the treatment of articularcartilage defects in the knee joint in athletes,” Knee Surgery,Sports Traumatology, Arthroscopy, vol. 14, no. 9, pp. 834–842,2006.

[43] G. Knutsen, J. O. Drogset, L. Engebretsen et al., “A random-ized trial comparing autologous chondrocyte implantationwith microfracture: findings at five years,” Journal of Bone andJoint Surgery A, vol. 89, no. 10, pp. 2105–2112, 2007.

[44] G. Knutsen, V. Isaksen, O. Johansen et al., “Autologouschondrocyte implantation compared with microfracture inthe knee: a randomized trial,” Journal of Bone and JointSurgery A, vol. 86, no. 3, pp. 455–464, 2004.

[45] J. Wasiak, C. Clar, and E. Villanueva, “Autologous cartilageimplantation for full thickness articular cartilage defects ofthe knee,” Cochrane Database of Systematic Reviews, vol. 3,Article ID CD003323, 2006.

[46] P. Lobenhoffer, J. van Heerwaarden, A. Staubli, and R. P.Jakob, Osteotomie Around the Knee, Thieme, 2008.

[47] J. P. Jackson, “Osteotomy for osteoarthritis of the knee.Proceedings of the Sheffi eld Regional Orthopaedic Club,”The Journal of Bone and Joint Surgery, vol. 40, no. 4, p. 826,1958.

[48] J. P. Jackson, W. Waugh, and J. P. Green, “Tibial osteotomy forosteoarthritis of the knee,” Journal of Bone and Joint SurgeryB, vol. 43, no. 4, pp. 746–751, 1961.

[49] R. Gariepy, “Genu varum treated by high tibial osteotomy.In proceedings of the joint meeting of orthopaedic associa-tions,” The Journal of Bone and Joint Surgery, vol. 46, no. 4,pp. 783–784, 1964.

[50] M. B. Coventry, “Osteotomy of the upper portion of the tibiafor degenerative arthritis of the knee. A preliminary report,”The Journal of Bone and Joint Surgery, vol. 47, pp. 984–990,1965.

[51] P. Lobenhoffer and J. D. Agneskirchner, “Improvements insurgical technique of valgus high tibial osteotomy,” KneeSurgery, Sports Traumatology, Arthroscopy, vol. 11, no. 3, pp.132–138, 2003.

[52] A. E. Staubli, C. De Simoni, R. Babst, and P. Lobenhoffer,“TomoFix: a new LCP-concept for open wedge osteotomy ofthe medial proximal tibia—early results in 92 cases,” Injury,vol. 34, supplement 2, pp. SB55–SB62, 2003.

[53] M. Wagner, A. Frenk, and R. Frigg, “New concepts forbone fracture treatment and the Locking Compression Plate,”Surgical Technology International, vol. 12, pp. 271–277, 2004.

[54] P. G. Maquet, Biomechanics of the Knee: With Applications ofthe Pathogenesis and the Surgical Treatment of Osteoarthritis,Springer, New York, NY, USA, 2nd edition, 1984.

[55] P. Maquet, “Valgus osteotomy for osteoarthritis of the knee,”Clinical Orthopaedics and Related Research, vol. 120, pp. 143–148, 1976.

[56] Y. Fujisawa, K. Masuhara, and S. Shiomi, “The effect ofhigh tibial osteotomy on osteoarthritis of the knee. Anarthroscopic study of 54 knee joints,” Orthopedic Clinics ofNorth America, vol. 10, no. 3, pp. 585–608, 1979.

[57] J. D. Aqueskirchner, A. Bernau, A. C. Burkart, and A.B. Imhoff, “Knee instability and varus malangulation—simultaneous cruciate ligament reconstruction andosteotomy (indication,planning and operative technique,results),” Zeitschrift fur Orthopadie und Ihre Grenzgebiete,vol. 140, no. 2, pp. 185–193, 2002.

[58] D. Paley and J. Pfeil, “Principles of deformity correctionaround the knee,” Orthopade, vol. 29, no. 1, pp. 18–38, 2000.

[59] M. Bonnin and P. Chambat, “Current status of valgus angle,tibial head closing wedge osteotomy in medial gonarthrosis,”Orthopade, vol. 33, no. 2, pp. 135–142, 2004.

[60] J. N. Insall, D. M. Joseph, and C. Msika, “High tibialosteotomy for varus gonarthrosis: a long-term follow-upstudy,” Journal of Bone and Joint Surgery A, vol. 66, no. 7, pp.1040–1048, 1984.

[61] K. Yasuda, T. Majima, T. Tsuchida, and K. Kaneda, “A ten-to 15-year follow-up observation of high tibial osteotomy inmedial compartment osteoarthrosis,” Clinical Orthopaedicsand Related Research, no. 282, pp. 186–195, 1992.

[62] L. Marmor, “Marmor modular knee in unicompartmentaldisease. Minimum four-year follow-up,” Journal of Bone andJoint Surgery A, vol. 61, no. 3, pp. 347–353, 1979.

[63] D. W. Murray, “Unicompartmental knee replacement: nowor never?” Orthopedics, vol. 23, no. 9, pp. 979–980, 2000.

[64] J. T. Moller, R. E. Weeth, J. O. Keller, and S. Nielsen,“Unicompartmental arthroplasty of the knee. Cadaver studyof the importance of the anterior cruciate ligament,” ActaOrthopaedica Scandinavica, vol. 56, no. 2, pp. 120–123, 1985.

[65] E. M. H. Obeid, M. A. Adams, and J. H. Newman,“Mechanical properties of articular cartilage in knees withunicompartmental osteoarthritis,” Journal of Bone and JointSurgery B, vol. 76, no. 2, pp. 315–319, 1994.

[66] S. R. Ridgeway, J. P. McAuley, D. J. Ammeen, and G. A.Engh, “The effect of alignment of the knee on the outcomeof unicompartmental knee replacement,” Journal of Bone andJoint Surgery B, vol. 84, no. 3, pp. 351–355, 2002.

[67] T. Borus and T. Thornhill, “Unicompartmental knee arthro-plasty,” Journal of the American Academy of OrthopaedicSurgeons, vol. 16, no. 1, pp. 9–18, 2008.

[68] N. M. Boehler, “Unicompartmental knee arthroplasty,” inKnee Arthroplasty, T. Sculco and E. Martucci, Eds., pp. 113–119, Springer, New York, NY, USA, 2001.

[69] E. Koskinen, P. Paavolainen, A. Eskelinen, P. Pulkkinen,and V. Remes, “Unicondylar knee replacement for primaryosteoarthritis: a prospective follow-up study of 1,819 patientsfrom the Finnish Arthroplasty Register,” Acta Orthopaedica,vol. 78, no. 1, pp. 128–135, 2007.

[70] U. C. G. Svard and A. J. Price, “Oxford medial unicom-partmental knee arthroplasty. A survival analysis of anindependent series,” Journal of Bone and Joint Surgery B, vol.83, no. 2, pp. 191–194, 2001.

[71] A. P. Sah and R. D. Scott, “Lateral unicompartmental kneearthroplasty through a medial approach: study with an

Arthritis 9

average five-year follow-up,” Journal of Bone and Joint SurgeryA, vol. 89, no. 9, pp. 1948–1954, 2007.

[72] T. V. Gunther, D. W. Murray, R. Miller et al., “Lateralunicompartmental arthroplasty with the Oxford meniscalknee,” Knee, vol. 3, no. 1-2, pp. 33–39, 1996.

[73] T. Ashraf, J. H. Newman, R. L. Evans, and C. E. Ackroyd,“Lateral unicompartmental knee replacement,” Journal ofBone and Joint Surgery B, vol. 84, no. 8, pp. 1126–1130, 2002.

[74] C. E. Ackroyd, J. H. Newman, R. Evans, J. D. J. Edridge, andC. C. Joslin, “The avon patellofemoral arthroplasty: five-yearsurvivorship and functional results,” Journal of Bone and JointSurgery B, vol. 89, no. 3, pp. 310–315, 2007.

[75] P. Cartier, J. L. Sanouiller, and A. Khefacha, “Long-termresults with the first patellofemoral prosthesis,” ClinicalOrthopaedics and Related Research, no. 436, pp. 47–54, 2005.

[76] R. L. Buly and T. P. Sculco, “Recent advances in total kneereplacement surgery,” Current Opinion in Rheumatology, vol.7, no. 2, pp. 107–113, 1995.

[77] E. M. Keating, J. B. Meding, P. M. Faris, and M. A. Ritter,“Long-term followup of nonmodular total knee replace-ments,” Clinical Orthopaedics and Related Research, vol. 404,pp. 34–39, 2002.

[78] J. A. Rand and D. M. Ilstrup, “Survivorship analysis of totalknee arthroplasty,” Journal of Bone and Joint Surgery A, vol.73, no. 3, pp. 397–409, 1991.

[79] H. Lundblad, A. Kreicbergs, and K. A. Jansson, “Prediction ofpersistent pain after total knee replacement for osteoarthri-tis,” Journal of Bone and Joint Surgery B, vol. 90, no. 2, pp.166–171, 2008.

[80] M. A. Ritter, J. D. Lutgring, K. E. Davis, M. E. Berend,J. L. Pierson, and R. M. Meneghini, “The role of flexioncontracture on outcomes in primary total knee arthroplasty,”Journal of Arthroplasty, vol. 22, no. 8, pp. 1092–1096, 2007.

[81] A. J. Smith, D. J. Wood, and M. G. Li, “Total knee replace-ment with and without patellar resurfacing: a prospective,randomised trial using the Profix total knee system,” Journalof Bone and Joint Surgery B, vol. 90, no. 1, pp. 43–49, 2008.

[82] R. S. Burnett, J. L. Boone, K. P. McCarthy, S. Rosenzweig,and R. L. Barrack, “A prospective randomized linical trial ofpatellar resurfacing and nonresurfacing in bilateral total kneearthroplasty,” Clinical Orthopaedics and Related Research, vol.464, pp. 65–72, 2007.

[83] J. H. Lonner and P. A. Lotke, “Aseptic complications aftertotal knee arthroplasty,” The Journal of the American Academyof Orthopaedic Surgeons, vol. 7, no. 5, pp. 311–324, 1999.

[84] F. Walter, M. B. Haynes, and D. C. Markel, “A randomizedprospective study evaluating the effect of patellar eversionon the early functional outcomes in primary total kneearthroplasty,” Journal of Arthroplasty, vol. 22, no. 4, pp. 509–514, 2007.

[85] A. Berth, D. Urbach, W. Neumann, and F. Awiszus, “Strengthand voluntary activation of quadriceps femoris musclein total knee arthroplasty with midvastus and subvastusapproaches,” Journal of Arthroplasty, vol. 22, no. 1, pp. 83–88,2007.

[86] J. King, D. L. Stamper, D. C. Schaad, and S. S. Leopold,“Minimally invasive total knee arthroplasty compared withtraditional total knee arthroplasty: assessment of the learningcurve and the postoperative recuperative period,” Journal ofBone and Joint Surgery A, vol. 89, no. 7, pp. 1497–1503, 2007.

[87] T. Karachalios, D. Giotikas, N. Roidis, L. Poultsides, K.Bargiotas, and K. N. Malizos, “Total knee replacementperformed with either a mini-midvastus or a standard

approach: a prospective randomised clinical and radiologicaltrial,” Journal of Bone and Joint Surgery B, vol. 90, no. 5, pp.584–591, 2008.

[88] S. S. Leopold, “Minimally invasive total knee arthroplastyfor osteoarthritis,” The New England Journal of Medicine, vol.360, no. 17, pp. 1749–1758, 2009.

[89] W. C. H. Jacobs, D. J. Clement, and A. B. Wymenga,“Retention versus removal of the posterior cruciate ligamentin total knee replacement: a systematic literature reviewwithin the Cochrane framework,” Acta Orthopaedica, vol. 76,no. 6, pp. 757–768, 2005.

[90] R. Chaudhary, L. A. Beaupre, and D. W. C. Johnston, “Kneerange of motion during the first two years after use ofposterior cruciate-stabilizing or posterior cruciate-retainingtotal knee prostheses: a randomized clinical trial,” Journalof Bone and Joint Surgery A, vol. 90, no. 12, pp. 2579–2586,2008.

[91] J. E. Arbuthnot and R. B. Brink, “Assessment of the antero-posterior and rotational stability of the anterior cruciateligament analogue in a guided motion bi-cruciate stabilizedtotal knee arthroplasty,” Journal of Medical Engineering andTechnology, vol. 33, no. 8, pp. 610–615, 2009.

[92] J. Bellemans, S. Banks, J. Victor, H. Vandenneucker, andA. Moemans, “Fluoroscopic analysis of the kinematics ofdeep flexion in total knee arthroplasty. Influence of posteriorcondylar offset,” Journal of Bone and Joint Surgery, vol. 84, no.1, pp. 50–53, 2002.

[93] Y. H. Kim, Y. Choi, and J. S. Kim, “Range of motion ofstandard and high-flexion posterior cruciate-retaining totalknee prostheses: a prospective randomized study,” Journal ofBone and Joint Surgery A, vol. 91, no. 8, pp. 1874–1881, 2009.

[94] R. Gandhi, D. Tsvetkov, J. R. Davey, and N. N. Mahomed,“Survival and clinical function of cemented and uncementedprostheses in total knee replacement: a meta-analysis,”Journal of Bone and Joint Surgery B, vol. 91, no. 7, pp. 889–895, 2009.

[95] A. V. Lombardi, C. C. Berasi, and K. R. Berend, “Evolutionof tibial fixation in total knee arthroplasty,” Journal ofArthroplasty, vol. 22, supplement 4, pp. 25–29, 2007.

[96] K. G. Nilsson, J. Karrholm, L. Carlsson, and T. Dalen,“Hydroxyapatite coating versus cemented fixation of thetibial component in total knee arthroplasty: prospectiverandomized comparison of hydroxyapatite- coated andcemented tibial components with 5-year follow-up usingradiostereometry,” Journal of Arthroplasty, vol. 14, no. 1, pp.9–20, 1999.

[97] M. A. R. Freeman and R. Tennant, “The scientific basis ofcement versus cementless fixation,” Clinical Orthopaedics andRelated Research, no. 276, pp. 19–25, 1992.

[98] K. Bauwens, G. Matthes, M. Wich et al., “Navigated totalknee replacement: a meta-analysis,” Journal of Bone and JointSurgery A, vol. 89, no. 2, pp. 261–269, 2007.

[99] J. M. Spencer, S. K. Chauhan, K. Sloan, A. Taylor, and R. J.Beaver, “Computer navigation versus conventional total kneereplacement: no difference in functional results at two years,”Journal of Bone and Joint Surgery, vol. 89, pp. 477–480, 2007.

[100] M. B. Collier, C. A. Engh, J. P. McAuley, and G. A. Engh,“Factors associated with the loss of thickness of polyethylenetibial bearings after knee arthroplasty,” Journal of Bone andJoint Surgery A, vol. 89, no. 6, pp. 1306–1314, 2007.