pathology, diagnosis, and classification of pressure ulcers: comparing clinical and imaging...

Original Research Article – Clinical Science

Pathology, diagnosis, and classification of pressure ulcers:comparing clinical and imaging techniques

JANE NIXON, PhDa; GILLIAN CRANNY, MSca; SENGA BOND, PhDb

Pressure ulcer classification systems are based on the clinical manifestations of the skin and tissue layeraffected rather than underlying pathology. The objective of this study was to compare the validity of theclinical grading of erythema (blanching and nonblanching) with a measurement of skin perfusion. Therefore,an exploratory study comparing erythema with laser Doppler imaging of the sacrum and buttock skin areaswas undertaken. Acute and major elective general, vascular, and orthopedic surgical in-patients, aged 55years or over with an expected length of hospital stay of 5 or more days were recruited. Fifty laser Dopplerimages from 37 patients were obtained and included in a discriminant analysis. Discriminant analysis sug-gested that blanching and nonblanching erythema were physiologically distinct from ‘‘normal’’ skin; clinically,these could be assessed with reasonable accuracy. Imaging also determined that high blood flow of differingintensity characterized blanching and nonblanching erythema. There was no evidence of the ‘‘no flow’’phenomenon. (WOUND REP REG 2005;13:365–372)

PU Perfusion unitPressure ulcers are described as ‘‘an area of localizeddamage to the skin and underlying tissue caused bypressure, shear, friction and/or a combination ofthese.’’1 They are complex lesions of the skin andunderlying structures and vary in size and severity oftissue layer affected.2

Classification systems have been developed tocategorize the severity of pressure ulcers and includedescriptors ranging from erythema (blanching and non-blanching) of intact skin to full-thickness tissue loss,1,3–5

but pressure ulcer classification is based upon clinicalmanifestations and tissue layer affected rather thanunderlying histopathology. Although the development

of international classification systems has led toimproved consensus,1,4 an ongoing debate remainsregarding the description, inclusion, and clinicalassessment of skin changes of intact skin (i.e., Grade/Stage 1 pressure ulcers).6–8 Specifically, the classifica-tion of alteration in darkly pigmented skin,8 blanchingerythema, nonblanching erythema, and nonblanchingerythema with other skin changes (e.g., local indura-tion, edema, pain, or warmth) as a Grade 1 pressureulcer or not is not evidence based.

Histological differences between ‘‘normal’’ skin,blanching erythema, and nonblanching erythema arereported. Witkowski and Parish2 examined biopsies ofskin and described histological changes observed fromblanching erythema through to early pressure ulcersand black eschar. Similarities and transitional phaseswere noted, with a spectrum of histological changesand overlap between normal, blanching erythema, andnonblanching erythema.

However, clinical observations of alteration todarkly pigmented skin, blanching erythema, and non-blanching erythema, with or without other skinchanges, including local induration, edema, pain, anddiscoloration, have not been assessed in relation to

From the Clinical Trials Research Unita, University ofLeeds, Leeds, and School of Population andHealth Sciencesb, University of Newcastle,Newcastle upon Tyne, United Kingdom.

Manuscript received: January 20, 2004Accepted in final form: January 20, 2005Reprint requests: Jane Nixon, PhD, Deputy Head CTRU,

Clinical Trials Research Unit, University of Leeds,17 Springfield Mount, Leeds, LS2 9NG, UnitedKingdom. Fax: þ44 113 343 1471; Email:[email protected].

Copyright # 2005 by the Wound Healing Society.ISSN: 1067-1927.

365

skin viability (i.e., skin perfusion or subsequent skin/tissue loss), and their clinical significance is not fullyunderstood.

The mechanisms leading to tissue breakdown arelargely theoretical, and reviews of the pathological lit-erature identified six possible mechanisms within threefunctional units.9–11 The functional units include thecapillaries, the interstitial space, and the cells. Themechanisms suggested include local ischemia follow-ing occlusion of capillaries;12 endothelial damage ofarterioles and the microcirculation due to the applica-tion of disruptive and shearing forces;13 external pres-sure for a prolonged period resulting in direct celldeath (dry black eschar);2 occlusion of skin bloodflow and subsequent injury due to abrupt reperfusionof the vascular bed;4 (2) disturbance to the interstitialequilibrium resulting in cell membrane rupture andcapillary bursting,14 and sustained deformation as atrigger for cell damage.11,15

New and emerging areas of interest are occlusionof skin blood flow and subsequent injury due to abruptreperfusion of the vascular bed in relation to superficialpressure ulcer development9 and sustained deforma-tion as a trigger for cell damage in relation to deeppressure ulcer development.10,11,16 The dominantpathological mechanism for superficial and deeppressure ulcers has not been determined, and it ispossible that all mechanisms play a role in theirdevelopment.

The focus of this study was upon the emergingevidence suggesting that reperfusion injury may be animportant pathological process in the development ofsuperficial pressure ulcers. It is unclear whether clini-cally defined skin changes are pathologically differentand whether blanching erythema and nonblanchingerythema with or without other skin changes (e.g., localinduration, edema, pain, and discoloration) reflectreactive hyperemia (high blood flow) or reperfusioninjury (low blood flow). Consequently, there is a needto validate clinical signs and symptoms of pressureassault of the skin against physiological measuresof skin perfusion. Thus, the purpose of this clinicalinvestigation was to explore the potential for improvingthe diagnosis of Grade 1 pressure ulcers and futureapplication of technologies to assess darkly pig-mented skin.

MATERIALS AND METHODSThe Laser Doppler Perfusion Imager (MoorInstruments Ltd, Devon, UK) provides high-resolutionimaging of Doppler flux, and for the skin, providesassessment of full dermal thickness. Low-power redlight (wavelength ¼ 6.32 nm) is directed to the skin.The Doppler-shifted light from moving blood and

nonshifted light from tissue is detected and processedto yield flux. This is an arbitrary value and not a mea-sure of absolute flow—the signal is a product of thenumber of red blood cells moving in the sample volumeand the mean velocity of the moving red blood cells.Because it is neither velocity nor flow, the term fluxhas been adopted.17

The clinical research nurse performed laserDoppler imaging. For initial setup, various aspectswere standardized, including bed height (100 cm),scan head angle (67.5 �), scan head height (125 cm), anddistance of the scan head from the skin (80 cm).Patients were placed in a lateral position, and the but-tocks and sacral areas were outlined using a laser beamarea marker facility. Minor adjustments were made tothe scan head angle so that imaging commenced at thebed sheet/skin interface. Imaging then commenced, thelaser moving in a raster motion across the skin fromthe bed sheet upward. The single-image function on theMoor Instruments LDI (Version 3.01) was used and thescan speed standardized to 4 ms/pixel.

Pilot StudyA pilot study was undertaken to assess the practicalutility of laser Doppler imaging in a clinical environ-ment, inform the main study design, and provide datato identify potentially useful summary variables forinclusion in statistical analysis and calculate samplesize.18 The laser Doppler imaging setup methodwas unchanged, and the images obtained during thepilot phase were retained for inclusion in the mainanalysis.

Study designExploratory laser Doppler imaging was undertaken onsurgical in-patients at St. James’s University Hospital.Local research ethics committee approval wasobtained, and patients were recruited if they werescheduled for elective major general or vascular sur-gery or an acute orthopedic, vascular, or general surgi-cal admission; were aged 55 years or over on the day ofsurgery; and the expected length of stay was of 5 ormore days. Elective major surgery was defined as aplanned surgical procedure with an average surgicaltime of 120 minutes or more. Patients were excludedif they were admitted under the care of general surgerysubspecialities, including liver, urology, and breast sur-gery; had dark skin pigmentation that precluded reli-able identification of blanching or nonblanchingerythema; or had skin conditions over the sacrum orbuttocks that precluded reliable identification ofblanching and nonblanching erythema.

WOUND REPAIR AND REGENERATION366 NIXON ET AL. JULY–AUGUST 2005

Skin classification scale and sample sizeThe skin classification scale was adapted from interna-tional classification scales1,4 to meet practical datacollection requirements (Table 1). Specifically, Grade0 (no skin changes) was included to clearly distinguishskin assessment of normal skin from missing data.Grade 5 (black eschar) was included as a separategrade until wound debridement enabled classificationby tissue layer. In addition, blanching erythema, non-blanching erythema, and nonblanching erythema withother skin changes were recorded and classified asGrade 1a, 1b, and 1bþ, respectively.18

To detect differences in mean blood flow (perfu-sion units) between clinical skin grades—Grade 0,Grade 1a, Grade 1b, and Grade 1bþ—a minimumsample size of 42 scans was estimated using pilotstudy data.18 This was based on an analysis of variancewith 95 percent power at the 5 percent significancelevel.

Data collectionElective general and vascular patients were recruitedpreoperatively and informed written consent obtained.Skin on the buttocks and sacrum was assessed imme-diately postoperatively and daily until discharge using acombination of the following clinical and physiologicalmeasures: clinical skin assessment (preoperatively,postoperatively, and daily until discharge), laserDoppler imaging (1/2 hour and 1 hour postoperatively),and laser Doppler imaging of Grade 1b and Grade 1bþskin areas observed during follow-up, where feasible.An expert clinical research nurse or the researcherconducted clinical skin assessment using the visualassessment of the color change following the applica-tion of light finger pressure to the skin.

Acute general, vascular, and orthopedic surgicalpatients were recruited up to 72 hours after admission.Skin on the buttocks and sacrum was assessed

clinically once each day until discharge. If a skin areawas assessed as nonblanching, where feasible, the areawas scanned using the laser Doppler imager.

Image analysisThe Moor LDI image processing software package(Version 3.01) was used to quantify various character-istics of the laser Doppler scans. The main scan areawas defined using the ‘‘Region of Interest’’ facility.Where necessary, edges were cut to remove bed sheets,areas of leg and perineum, and ‘‘interference lines’’resulting from movement during imaging. The mainscan area was then outlined using the ‘‘box’’ functionand saved, thus enabling repeated processing usingidentical image dimensions.

The images were displayed in perfusion units (PUs)using a 6-color palette on a 16-increment range set from0 to 1000 plus. PUs were used to generate all summaryvalues for the image. Perfusion units adjust for distanceand normalize for the gain. Summary values were gen-erated for unsmoothed and smoothed images.Smoothing modifies each pixel according to the eightneighboring pixels, by calculation of a weighted aver-age. Various options were explored to identify keycharacteristics or variables to summarize the imageand enable discrimination between skin grades.

Histograms showed that the distribution of pixelswas skewed in all cases, suggesting little variability inbackground values (Figure 1). Some histograms werenoted to have a second histogram peak; this was aparticular characteristic associated with Grade 1bareas (Figure 2).

The preliminary processing led to a pragmaticapproach whereby the three areas assessed clinicallywere summarized as one clinical grade (the worstgrade recorded), and image analysis was performedusing each scan as a single area, even where two dis-tinct peaks were observed. The preliminary processingalso identified a number of potentially useful summary

Table 1. Skin classification scale used for this study18

Grade Description

0 No skin changes1a Redness to skin (blanching)1b Redness to skin (nonblanching)1bþ Redness to skin (nonblanching) plus one or more:

PainIndurationHeatEdemaDiscoloration (specify)

2 Partial thickness wound involving epidermis/dermis only3 Full thickness wound involving subcutaneous tissue4 Full thickness wound through subcutaneous tissue to

muscle or bone5 Black eschar

Pixels

18000

14000

10000

6000

2000

0 200 400 600 800 1000

Flux (PU)

FIGURE 1. Histogram of pixel values showing typical distribution.

WOUND REPAIR AND REGENERATIONVOL. 13, NO. 4 NIXON ET AL. 367

variables for inclusion in statistical analysis. Theseincluded mean, minimum, and maximum, and PUvalues for unsmoothed and smoothed images and sum-mary variables including ‘‘medium’’ (proportion of pix-els with PU between 300 and 600) and ‘‘high’’(proportion of pixels with PU greater than 600).

Discriminant analysisTo delineate the factors that would predict the classifi-cation scans by skin classification group, discriminantanalysis was used with independent variables, includ-ing the mean, minimum, maximum, and summary vari-ables ‘‘medium’’ and ‘‘high’’ in PUs for unsmoothed andsmoothed scans. The dependent or criterion variablewas the clinically assessed skin grade. Discriminantanalysis generates two new variables (discriminantfunctions) that are a linear combination of the originalindependent variables, which maximize separationbetween the skin classification groups. The discrimi-nant scores for each scan were computed by applyingthe discriminant function formulae. A territorial mapwas constructed to identify the boundaries used forclassifying scans into groups based on the discriminantscores.

To assess how well the discriminant functionworked, the percentage of correct classifications wascalculated for each skin classification group. This wasdone using the classification rules created to classifythe original scans. All statistical analyses were under-taken using SPSS (Chicago, IL).

RESULTSA total of 143 patients consented to participate in thepilot study (April to July 1998) and the main study(September 1998 to May 1999), including 93 electiveand 50 acute patients, but only 37 patients had

successful laser Doppler imaging. Reasons for the lowcapture rate are detailed in Table 2.

In total, 50 laser Doppler images of sacral andbuttock areas from 37 patients were obtained andincluded in the discriminant analysis. Laser Dopplerimages, with a corresponding black-and-white imageof the area scanned, are illustrated in Figure 3. Thesample of 37 patients comprised 21 women and 16men admitted for vascular (n ¼ 19), general (n ¼ 13),and orthopedic (n ¼ 5) surgery. The majority ofpatients were elective admissions (n ¼ 32), with onlyfive acute admissions, and the mean age of the samplewas 72.4 years (range 55–88 years).

There were organizational problems associatedwith laser Doppler imaging, including the availabilityof the clinical research nurse immediately postopera-tively and the availability of staff to transfer equipmentto the ward areas, and these exacerbated difficulties inobtaining images of Grades 1b and 1bþ skin changes.Where imaging of 1b areas were planned, in somecases, the 1b area had subsequently resolved to a 1aat the time of imaging, and in one, pressure ulcer devel-opment had occurred. No images of Grade 1bþ wereobtained.

Clinical skin classification and associated variablesfor the 50 scans after smoothing are detailed inTables 3 and 4. A Kruskal–Wallis test for an overalldifference between the three groups indicates statistic-ally significant differences between Grades 0, 1a, and1b for maximum and mean PUs, suggesting differencesin blood flow between skin grades (Table 3).

Pixels

9000

7000

5000

3000

1000

0 200 400 600 800 1000

Flux (PU)

FIGURE 2. Histogram of pixel values for a Grade 1b skin areashowing a two-peak distribution.

Table 2. Reasons patients entered into study were notscanned using laser Doppler imaging

Causative Event Number of Patients

Elective patients

Loss to follow-upCanceled surgery 3Patient withdrawal 2Lost forms 2Postoperative problemsPatient refused 2Unable to turn (pain, restless,hemodynamically unstable, on trolley)

17

ICU admission 9CRN not available 14Psoriasis 1Reason not recorded 11Total Electives 61

Acute patients

Loss to follow-upPatient withdrawal 2Early discharge/death 21b/1bþ not observed 35Pressure ulcer 11b observed but reason not recorded 5Total Acutes 45Grand total 106


Discriminant analysis was repeated with differentcombinations of independent variables to identify those

variables that together best predicted the clinical clas-sification. The independent variables that, in combina-tion, correctly classified 36 of 50 (72 percent) scansincluded the summary variables smoothed ‘‘medium’’and smoothed ‘‘high.’’ The pooled within-group correla-tions between discriminating variables and discrimi-nant functions are detailed in Table 5 and illustratethat Function 1 is correlated with the larger proportion

A B

C

D

FIGURE 3. Laser Doppler imaging of skinassessed clinically as Grades 0, 1a, and 1b.(A) Grade 0, (B) Grade 1a, (C) Grade 1a,and (D) Grade 1b.

Table 3. Skin classification grade and variable parametersfor smoothed scans

PUs Grade 0 Grade 1a Grade 1b

n 16 26 8Maximum PUMinimum 264 217 836Maximum 1293 1311 2115Mean 541.2 906.5 1432.5Median 484.5 972 1421*Standard deviation 262.2 365.9 468.5

Minimum PUMinimum 5 3 1Maximum 39 43 41Mean 23.1 18.3 17.9Median 22.5 18.5 15**Standard deviation 9.8 10.0 15.0

Mean PUMinimum 73.2 61.7 79.6Maximum 168.3 273.9 339.5Mean 108.6 156.9 230.3Median 101.9 144.8 283.4***Standard deviation 34.8 73.1 104.4

*p ¼ < 0.001; **p ¼ 0.29; ***p ¼ 0.009.

Table 4. Skin classification grade and variable parameters‘‘medium’’ and ‘‘high’’ for smoothed scans

Summary Variables Grade 0 Grade 1a Grade 1b

n 16 26 8MediumMinimum 0 0 0.02Maximum 0.04 0.26 0.18Median < 0.01 0.08 0.10Standard deviation 0.01 0.09 0.07HighMinimum 0 0 < 0.01Maximum < 0.01 0.15 0.23Median 0 0.01 0.14Standard deviation < 0.01 0.04 0.09


of ‘‘high’’ pixels (high-intensity scans) and Function 2with the larger proportion of ‘‘medium’’ pixels (med-ium-intensity scans.)

The scans were then plotted using the discriminantscores (Figure 4) and classified by group (Table 6).Classification compares the clinically assessed gradeagainst the predicted grade based upon the discrimi-nant function scores, for each scan. Results suggestthat ‘‘high-intensity scans’’ (large Function 1 or largeproportion of ‘‘high’’ pixels) discriminates betweenGrade 1b and the others. Scans that are not high inten-sity (smaller Function 1) can be classified as Grade 0and 1a by examining the proportion of ‘‘medium’’ pixelsas large Function 2 (Grade 1a) and smaller Function 2(Grade 0).

DISCUSSIONThe aim of this exploratory study was to assess thevalidity of the clinical grading of erythema by compar-ing grades derived from clinical observation with mea-surement of skin blood flow using laser Dopplerimaging. Despite variability in scan quality due topatient movement during imaging and the small samplesize, significant differences in the maximum and mean

PUs between grades and good discrimination betweenclinically observed grades was found using summaryimage data in discriminant analysis. A discriminantfunction containing the variables ‘‘medium’’ and‘‘high’’ correctly classified 72 percent of scans, andthree general patterns in skin blood flow were identi-fied. These patterns were ‘‘high-intensity scans’’ (largeFunction 1) Grade 1b; ‘‘medium-intensity scans’’ (largeFunction 2) Grade 1a; and ‘‘low-intensity scans’’ (smallFunction 2) Grade 0.

In terms of misclassification by grade, differencesbetween high-intensity blanching and high-intensity non-blanching erythema would appear to be difficult to differ-entiate clinically by applying light finger pressure. Thisreflects experience in practice and evidence from thepathological examination of skin biopsies that identifiedsimilarities and transitional phases and overlap betweennormal skin and blanching and nonblanching erythema.2

Blood flow ranged from 1 PU to 2115 PU acrossskin areas, but the distribution of minimum andmean blood flows for all scans illustrates little vari-ability in ‘‘normal’’ skin blood flow, despite theabsence of any environmental controls (e.g., ambienttemperature), patient grouping (e.g., vascular) orskin temperature. Imaging of clinically assessedGrade 1a and 1b areas illustrates that high skinblood flow characterized blanching and nonblanchingerythema, suggesting that the responses observedwere not pathologically different but reflected thecapacity of the skin to increase blood flow locallyup to 10-fold compared with baseline. No evidencewas observed of the ‘‘no-flow phenomenon’’ asso-ciated with reperfusion injury.

This research has important limitations. The sam-ple size calculation undertaken assumed that a univari-ate analysis would discriminate between the threegroups, but image analysis requires consideration of anumber of variables simultaneously to identify discri-minating features. It is suggested that the sampleshould be at least five times as many subjects pergroup as the number of variables to be examined,19

therefore a sample of at least 90 skin sites needs tobe included in a definitive analysis.

Overall, practical difficulties in performing thelaser Doppler imaging led to a shortfall in sample

Table 5. Correlation of the variables ‘‘medium’’ and ‘‘high’’with discriminant Functions 1 and 2

Discriminant Functions

Variables

1 2

High 0.977* 0.213Medium 0.377 0.926*

*Largest absolute correlation between each variable and any discriminant function.

Discriminant Functions

All groups

Function 1

543210–1–2

Func

tion

2

4

3

2

1

0

–1

–2

GRADE

Group Centroids

1b

1a

0

2

1

0

FIGURE 4. Laser Doppler images plotted by discriminantfunctions.

Table 6. Clinical skin classification and predicted laserDoppler image classification

Clinical

Predicted Group Membership

Classification 0 1a 1b Total

0 14 2 0 161a 8 17 1 261b 1 2 5 8Total 23 21 6 50


size and the need to include all scans (includingthose obtained in the pilot phase) in the main analysis.In addition, the sample size was not sufficient to testthe discriminants on an independent sample ofimages.

There were also difficulties in summarizing theimages due to the number of different combinationsfor clinical skin assessment and blood flow patterns.Three skin areas were assessed clinically (sacrum, leftbuttock, and right buttock), but image analysisinvolved summarizing each scan as one single area.The image analysis compared the summary statisticsfor this combined single area to the worst of threegrades allocated clinically. Various patterns emerged;for example, in some cases a clinical assessment ofthree Grade 1a areas was recorded, but the corres-ponding scan was characterized by only one area ofhigh blood flow (Figure 3b). In other cases, two areasof high blood flow characterized the correspondingscan (Figure 3c). Future research will need todevelop reliable image-processing methods andaccurate mapping of clinical assessments to enableseparate analysis of multiple high-blood flow-areasfrom the same image.

The practical difficulties of performing laserDoppler imaging of critically ill, elderly, and immobilepatients reduced the potential sample size and resultedin an absence of Grade 1bþ skin areas for analysis.Questions remain about skin blood flow patterns asso-ciated with nonblanching erythema, where localinduration, edema, pain, and/or discoloration is alsoobserved and whether this is characterized by a highblood flow response or ischemic reperfusion injury(i.e., low blood flow).

Despite the limitations and exploratory nature ofthe study, discriminant analysis suggests that blanchingand nonblanching erythema are physiologically distinctfrom ‘‘normal’’ skin and that, clinically, these can beassessed with reasonable accuracy. Also, the good levelof agreement found between clinical observation and ameasure of skin blood flow suggests that, despite someof the difficulties encountered, this technology may beclinically useful in the assessment of patients withdarkly pigmented skin.

In conclusion, despite the limitations and explo-ratory nature of the research, laser Doppler imagingprovided a general picture of the physiological rangein blood flow values for normal skin and areas oferythema after localized pressure assault in anuncontrolled physical environment. The distributionof minimum and mean blood flows for all scansillustrates little variability in ‘‘normal’’ skin bloodflow. Imaging also determined that blanching andnonblanching erythema are characterized by highblood flow of differing intensity, suggesting that theresponses observed are not pathologically different

but reflect the capacity of the skin to increase bloodflow locally up to 10-fold compared with baseline.There was no evidence of the ‘‘no flow’’ phenom-enon, but the inability to image Grade 1bþ skinareas limited the study.

ACKNOWLEDGMENTSFunding for this research was provided by the TissueViability Society (UK) Research Training Fellowship(1997), Northern and Yorkshire Region (UK) SmallGrant (1998) and Smith and Nephew FoundationNursing Research Fellowship (UK). We thank the staffof St. James’s University Hospital, Leeds, includingward, theater, and anesthetic recovery staff, who sup-ported the research; Dr. Steven Smye (Head of MedicalPhysics) and Mr. Julian Scott (consultant vascular sur-geon), who supported the development of ideas and theuse of imaging technologies; and Jayne Upperton (clin-ical research nurse).

REFERENCES1. Pressure Ulcer Treatment Guidelines. Oxford: European Pressure

Ulcer Advisory Panel, 1999.

2. Witkowski JA, Parish LC. Histopathology of the decubitus ulcer.

J Am Acad Dermatol 1982;6:1014–21.

3. Torrance C. Pressure sores aetiology, treatment and prevention.

London: Croom-Helm, 1983.

4. Pressure ulcers in adults: prediction and prevention: quick refer-

ence guide for clinicians. Rockville, MD: Agency for Health Care

Policy and Research, U.S. Department of Health and Human

Services, 1992.

5. Reid J, Morison M. Towards a consensus. Classification of pres-

sure sores. J Wound Care 1994;3:157–60.

6. Bethell E. Controversies in classifying and assessing grade 1

pressure ulcers. J Wound Care 2003;12:33–6.

7. Bergstrom N, Braden B. Predictive validity of the Braden

Scale among black and white subjects. Nursing Res 2002;51:

398–403.

8. NPUAP. Stage 1 Assessment in darkly pigmented skin.

www.npuap.org/positn4.htm. 1998.

9. Nixon J. The pathophysiology and aetiology of pressure ulcers. In:

Morison M, editor. The prevention and treatment of pressure

ulcers. Edinburgh: Mosby, 2001:17–36.

10. Bouten CV, Oomens CW, Baaijens FP, Bader DL. The etiology of

pressure ulcers: skin deep or muscle bound? Arch Phys Med

Rehabil 2003;84:616–9.

11. Bosboom EMH. Deformation as a trigger for pressure sore

related muscle damage. PhD dissertation, Technische

Universiteit Eindhoven, 2001.

12. Kosiak M. Etiology of decubitus ulcers. Arch Phys Med Rehabil

1961;42:19–29.

13. Barton AA, Barton M. The pressure sore. The management and

prevention of pressure sores. London: Faber, 1981:26–38.

14. Krouskop TA. A synthesis of the factors which contribute to

pressure sore formation. Med Hypoth 1983;11:255–67.

15. Breuls RG, Oomens CW, Bouten CV, Bader DL, Baaijens FP.

A theoretical analysis of damage evolution in skeletal muscle


tissue with reference to pressure ulcer development. J Biomech

Eng 2003;125:902–9.

16. Bosboom EM, Bouten CV, Oomens CW, van Straaten HW,

Baaijens FP, Kuipers H. Quantification and localisation of

damage in rat muscles after controlled loading: a new approach

to the aetiology of pressure sores. Med Eng Phys 2001;23:

195–200.

17. Fronek A. Clinical value of laser doppler fluxmetering. In:

Bernstein EF, editor. Vascular diagnosis. St. Louis, MO.: Mosby,

1993:626–9.

18. Nixon J, Smye S, Scott J, Bond S. The diagnosis of early

pressure sores: report of the pilot study. J Tiss Viability

1999;9:62–6.

19. Altman DG. Practical statistics for medical research. London:

Chapman & Hall, 1991.


pathology, diagnosis, and classification of pressure ulcers: comparing clinical and imaging...

Documents