ORlGINAL CONTRlBUTION

Surgical vs Nonoperative Treatment for Lumbar Disk Herniation The Spine Patient Outcomes Research Trial (SPORT): A Randomized Trial

James N. Weinstein, DO, MSe

Tor D. Tosteson, ScD

Jon D. Lurie, MD, MS

Anna N. A. ScD

Brett Hanscom, MS

Jonathan S.

William A. Abdu,

Man S. MD

Scott D. Boden, MD

Richard A. MD, MPH

PhD

MS

L UMBAR DI5KECTOMY 15 THE MOST

common surgical procedure performed in the United States

for patients having back and leg

symptoms; the vast majority of the pro- cedures are elective. However, lumbal' disk herniation is often seen on imag- ing studies in the absence of symp- toms1.2 and can regress over time vvith-

out surgery.3 Up to 15- fold variation in regional diskectomyrates in the United States4 and lower rates internationaHy

mise questions regarding the appropri-

ateness of some of these surgeries.5,6

Several studies have compared sur- gical and nonoperative treatment of pa- tients wirh herniated but base-

line differences between treatment groups, small or lack of validated outcome measures these

studies limit conclu- sions

Context Lumbar diskectomy ia the most common surgical procedure performed for back and leg symptoms in US patients, but the efficacy of the procedure relative to nonoperative care remains controversiaL

Objective To assess the efficacy of surgery for lumbar intervertebral disk herniation.

Design, Setting, and Patients The Spine Patient Outcomes Research Trial, a ran- domized dinical trial enrolling patients between March 2000 and November 2004 from 13 multidisdplinary spine dinks in 11 US states. Patients were 501 surgical candidates (mean age, 42 years; 42 % women) with imaging-confirmed lumbar intervertebral disk

hemiation and persistent signs and symptoms of radiculopathy for at least 6 weeks.

Interventions Standard open diskectomy vs nonoperative treatment individualized to the patient.

Main Outcome Measures Primary outcomes were changes rrom baseline for the Medical Outcomes Study 36-item Short-Form Health Survey bodily paln and physical runction scales and the modified Oswestry Disability Index (American Academy of Or- thopaedic Surgeons MODEMS version) at 6 weeks, 3 months, 6 months, and 1 and 2

years from enrollment. Secondary outcomes included sciatica severity as measured by the Sciatica Bothersomeness Index, satisfaction with symptoms, se!f-reported improve- ment, and employment status.

Results Adherence to assigned treatment was limited: 50% of patients assigned to

surgery received surgery within 3 months of enrollment, while 30% of those assigned to nonoperative treatment received surgery in the same period. IntenHo-treat analy- ses dernonstrated substantial improvements for all primary and secondary outcomes in both treatment groups. Between-group differences in improvements were consis- tently in favor of surgery for all periods but were small and not statistically significant for the primary outcomes.

Conclusions Patients in both the surgery and the nonoperatíve treatment groups improved substantially over a 2-year period. Because of the large numbers of patients who crossed over in both directions, conclusions about the superiority or equivalence of the treatments are not warranted based on the intent-to-treat analysis.

Trial Registration ciinicaltrials.gov Identifier: NCT00000410 JAMA. 2006;296:2441-2450 www.jama.com

2000 to compare the outcomes of sur.

SURGlCAL VS NONOPERATlVE TREATMENT FOR LUMBAR DISK HERNIATlON

oVln treatment but othenvise met all the

other criteria for indusion and who agreed to undergo follow-up accord- ing to the same protocol. This article

reports intent-to-treat results through 2 years for the randomized cohort.

Flgure 1. Flow Diagram of the SPORT Randomized Controlled Trial of Dlsk Herniation: Exdusion, Enrollment, Randomization, and Follow-up

i 2720 Pallenta Screened for Efiglbilìty

Assigned 10 Reoeive Surgery

6-wk Foltow-up

Had Date Availabla

40 M1ssad Visit 2 Wlthdrew' o Dlad'

74 Underwaf\t Surgery (32%)'t

198 Had Dala Available

45 Mlssad Vislt .2 Wrthdraw o Dled

115 UndelWen! Surgery (50%)

6-mo Follow-up

200 Had Date Availabla

37 Mlssad Vialt

ß Withdrew o Dlad

132 Underwent Surgery (57%)

1-y Follow-up

202 Had Data Avaltable

29 Missed Visll 14 WithdrèW

o Ciad

Underwent Surgery (59%)

FoIlow-up

186 Had Date. Avaiiabie

32 MIssad \!ïstl

23 Withdrew o Died 4 2-'1 Fcltcw-up Not Completed

729 Candldates

Noooperatlva

Raceive

219 Had Date Avaiiabte

37 Missed Vi.!t o Wlthdrew' o Dled'

211 Had 44 Mlsaed \tls~

1 Wlthdrew o Diad

71

6-ma Fcllow-up

210 Had Data Avallabie 4 i Mlssed VlslI

, 5 Wlthdraw

I 0 Died

~ Underwent Surgary (39%)

1-y Follow-up

213 Had Dats Avaiiabla

27 Missed Vlalt

15 Wlthdrew 1 Died

i 03 Underwant Surgef/ (43%)

187 Had Data Avaiiable

31 Missed Vlsit

27 Wllhdrew

2 Dled 9

1^^ ur

für surgery;

METHODS Study Design SPORT was conducted at 13 multidis- dplinary spine practices in 11 US states

(Califomia, Georgia, Illinois, Maine, Michigan, Missouri, Nebraska, New York, New Hampshire, Ohio, Pennsyl-

vania). The human subjects commit- tee of each participating institution ap- proved a standardized protocoL All patients provided written informed con-

sent. An independent data and safety

monitoring board monitored the study

at 6-month intervals.13

Patient Population

Patients were considered for inclu- sion iE theywere 18 years and olderand diagnosed by participating physidans

during the study enrollment period as

having intervertebral disk herniation and persistent symptoms despite some

nonoperative treatment for at least 6

weeks. The content of preenrollment non operative care was not prespeci- fied in the protocol but included edu-

cation/counseling (71%), physical therapy (67%), epidural injections

(42%), chiropractìc therapy(32%), anti-

inflammatorymedications (61%), and opioid analgesics (40%).

Specific inclusion criteria at enroll-

ment were rarucular pain (below the knee for lower lumbar herniations, into the an- tedor thigh for upper lumb ar hernia- tions) and evidence oE nerve-raot irrita-

tion with a positive nerve-root tension sign (straight leg be- tween 300 and 700 or femoral

ora

sensation in

showíng disk hemiation (protrusion, ex- trusion, or sequestered fragment)!4 at a

level and side corresponding to the clini- cal symptoms. Patients with multiple hemiations were inc1uded if only one of

the hemiations was considered symp- tomatic (ie, if only one was planned to

be opera ted on). Exclusion criteria included prior lum-

bar surgery, cauda equina syndrome,

scoliosis greater than 150, segmental in- stability(> 100 angular motion or >4-mm lIanslation), vertebral fractures, spine in- feetion or tumor, inflammatory spondy- loarthropathy, pregnancy, comorbid eon- ditions conlIaindicating surgery, or inability/umvillingness to have surgery

wíthin 6 months.

Study Interventions

The surgery was a standard open dis- kectomy wíth examination of the in- volved nerve roOt.15.l6 The proeedure agreed on by all participating centers

was performed under general or loeal anesthesia, wíth patients in the prone or knee-ehest position. Surgeons were encouraged to use loupe magnifica-

tion or a microseope. Using a midline incision reflecting the paraspinous muscles, the interlaminar space was en- tered as described by Delamarter and McCuUough.15 In some cases the me- dial border of the superior facet was re- moved to provide a dear view of the in- volved nerve root. Using a small annular incision, the fragment of disk was re-

moved as described by Spengler. 16 The

canal was inspected and the foramen probed for residual disk 01' bony pa- thology. The nerve root was decom- pressed, leaving it freely mobile.

The non operative treatment group received "usual care," with the study protocol recommending that the mini- mum nonsurgical treatment include

at least active physical therapy, educa- wíth home exercise in-

nonsteroidal

SURGICAL VS NON OPERATIVE TREATMENT FOR LUMBAR DlSK HERNIATlON

Table 1. Patient Baseline Demographie Characteristics, Comorbid Condítions, Clinícal Findings, and Health Status Measures*

Patlents, No, (%)

43

93

status

Pein with reise

SF-36

Scietlca indices, moon

SURGìCAL VS

Table 2. Treatments

GliniciansiserJÍces Educatlcn/oounseling 299

52 119 36

195

MedicaUens NSAiDs GOX-2 inhibitors Gral steroids

NarceUos MlJscla relaJamts

Gther Devicas

Brece Gelset

142 13

180 102

193 101

15 147 66

172

18 (6) 9(3)

12 (4) 38(12) 25(8) 12 (4) 27 (8)

None 216 (68)

Abbreviatioos: COX.2, cyclooxygenase 2; NS?JDS, non. staroidal anti-lnnammalory drJgs; TENS. transcutane- aus elootñcaJ nerve stimulaton.

*Patients wI10 had used ciinlclans. treatments. medioa- tlens, and d6'Aoos withln 1

year foliowing enrollment Of

untlllhe time 01 SUfgery; 323 patlents eil her had no BUf-

gerv 10 Ihe (jffit y'OOl' 01 enrollmenl or had at least 1 regu- lariÝsd'eduredIDlk:w.up vlsitprior tosurgery al wrJch non- operaWe treatment Informatioo c.ould be assessed.

Study Measures The primary measures were the Medi- cal Outcomes Study 36-Item Short- Form Health Survey (SF- 36) bodily pain

and physícal functìon scales18-21 and the Americall Academy of Orthopaedìc Sur- geons MODEMS version of the Oswe- stry Disability Index (ODI)Y As specí-

fied in the trial protocol, the primary outcomes were changes from baseline in these scales at 6 weeks, 3 months, 6

mOllths, alld land 2 years from enrollment.

Secondary measures included pa- tient improvemcm, work status, and satisfaction with eunent

care. Symptom se-

DISK HERNIATION

eligibility. For recruitment and in- formed consent, evidence-based vid- eotapes described the surgícal aud non- operative treatments and the expected

benefits, risks, and uncertainties.26,27 Particípants were offered enrollment in either the randomized trial or a con- current observational cohon, the re- sults of whích are reported in a com- panion article.

Enrollment began in March 2000 and

ended in November 2004. Baseline vari- ables were collected prior to random- ization. Patients self-reported race and ethnicity using Nationallnstitutes of Health categories.

Computer-genera ted random treat-

ment assignment based on permuted blocks (randomly genera ted blocks of 6, 8, 10, and 12)28 within sites oc- curred immediately after enrollment via

an automated system at each site, en- suring proper allocation concealment. Study measures were collected at base-

Une and at regularly scheduled fol- low-upvisits. Short-term follow-upvis-

its occurred at 6 wecks and 3 months. If surgery was delayed beyond 6 weeks, additional follow-up da ta were ob- tained 6 weeks and 3 months pos top- eratively. Longer-term follow-up vis- its occurred at 6 months, 1 year from enrollment, and annually thereafter.

Statistical Analyses

We originally determined a sampIe size

of 250 patients in each treatment group to be suffícient (wÜh a 2-sided signífí-

canee level of .05 and 85% power) to

deteet a lO-point difference in the SF-36 bodily pain and physical functioning

scales or a similar effect size in the ODI. This difference corresponded to pa- tients' reports ofbeing "a lÜde better"

in the Maine Lumbar (MLSS).29 The sampIe size caIculation allowed for up to 20% data but did not aecount any levels oE

2 years. To adjust for the possible effect

of missing da ta on the study results, the analysis of mcan changes for continu- ous outcomes was performed using

maximum likelihoad estimation for longitudinal mixed-effects models un- der "missing at random" assumptions and including a term for treatment cen- ter. Comparative analyses were per- formed using the single

methods of baseline value carried for-

ward and last value earried forward, as wen as a longitudinal mixed model con- trolling for covariates associated with missed visits.30

For binary secondary outcomes, lon- gitudinal logistic regression models were fitted using generalized estimat- ing equatians31 as implemented in the

PROC GENMOD program of SAS ver- sion 9.1 (SAS Institute lne, Cary, NC). Treatment effects were estimated as dif- fe rene es in the estimated proportions in the 2 treatment groups.

P<'05 (2-sided) was used to es tab- lish statistical significanee. For the pri- mary outeomcs, 95% eonfidenee inter- vals (CIs) for mean treatment effects

were calculated at each designated time point. Global tests of the joint hypoth- esis of no treatment effect at any of the designated periods were performed us- ing Wald tests32 as implemented in SAS.

These tests aceount for the intraindi- vidual correlation due to repeated mea- surements aver time.32

N onadherence to randomly as- signed treatment may mean that the intention-to-treat analysis under- estimates the real beneHt of the treat- ment.33,34 As a preplanned sensítívÜy

analysis, we also estimated an "as- treated" longitudinal analysis based on comparisons of those actually treated

and Re- peated measures of outcomes were used as the variables, and treat- ment received was included a time-

eovariate.

included to adjust for possible con- founding.

RESUL TS

SPORT achieved full enrollment, with 501 (25%) ofl991 eligible patients en- rolled in the randomized trial. A total of 472 participants (94%) completed at

least 1 follow-up visit and were in- cluded in the analysis. Data were avail- able for between 86% and 73% of pa- tients at each of the designated follow-up times (FIGURE 1).

Patient Charaderistics

Baseline patient characteristics are

shown in T ABLE 1. Overall, the study population had a mean age of 42 years,

with majorities being male, white, em- ployed, and having attended at least

some college; 16% were receiving dis- ability compensation. All patients had

radicular leg pain, 97% in a classic der-

matomal distribution. Most of the her-

niations were at L5-S1, posterolateral,

and were extrusions by imaging crite-

ria.14 The 2 randomized groups were similar at baseline.

Nonoperative Treatments A variety of nonoperative treatments

were used during the study (TABLE 2). Most patients received educationl counseling (93%) and anti-inflamma-

tory medications (61%) (nonsteroidal

anti-inflammatory drugs, cydooxyge-

nase 2 inhibitors, or oral steroids); 46%

received opiates; more than 50% re- ceived injections (eg, epidural ste- raids); and 29% were prescribed activ- ity restriction. Fony-four per cent received active physícal therapy dur- ing the trial; however, 67% had re- ceived it prior to enrollment.

Surgical Treatment and Complications

TABLE 3 the characteristícs oE SUI-

rreatmentand The

SURGICAL vs NONOPERATIVE TREATMENT FOR LUMBAR DISK HERNIA nON

plications of childbirth 11 months af-

ter enrollment. The most common inrraoperative complication was dural

tear (4%). There were na postopera- tive complications in 95% of patients.

Reoperation occurred in 4% of pa- tients within 1 year of the initial sur-

gery; more than 50% of the reopera- tions were for recurrent herniations at the same level.

Nonadherence

Nonadherence to treatment assign- ment affected both groups, ie, some pa- tients in the surgery group chose to de- lay 01' decline surgery, and same in the nonoperative treatment gr01.1p crossed

over to receive surgery CFigure 1). The

characteristics of crassover patients that

were statistically different from pa- tients who did not cross over are shown in TABLE 4. Those more likely to cross

over to receive surgery tended to have

lower incomes, worse baseline symp- toms, more haseline disability on the

om, and were more likely to rate their

symptoms as gettíng worse at enroll-

ment than the other patients receiving

nonoperative treatment. Those more likely to cross over to receive nonop- erative care were older, had higher in-

comes, were more likely to have an 1.1p-

per lumbar disk hemiation, less likely to have a positive straight leg-raising

test result, had less pain, better physi-

cal function, less disability on the om, and were more likely to ra te their symp- toms as getting better at enrollment than the other surgery patients.

Missing Data The rates of missing data were equiva-

lent between the. groups at each time point, with no evidence of differential

dropout to assigned treat-

ment. Characteristics of with missed visits were very similar ta those

of the rest oE the cohort that

with data were

Table 3. Operative Treatments, and Events

level

L4.5 L5-S1

Operation time, mean min Blood mean Blood Length

Same da'1 1 Night 2:2 Nlghts

IntraoperatIve compiicatlonst Dural tearíspinal fluid leak Vascuiar injury Other None

9 89

145 79.1 64.7

4

65 137 37

10 1 2

230

superficial 4 9

226 Other

None Postsurgical reoperatlon,

No. (rate)~ 1 y

Additional surgery Recurrent hernlation Complicatlon or other New conditlon

2'1 Additional surgery Recurrent hemlaUon Complication 01' other

New condition

9(4) 5(2) 4(2)

o

13 (5) 8(3) 4(2)

o

*Data on surglcallevel, blood loss, loogtl1 of etay, and oorn. plloaUons were not avallable for 7 aurgioal patients. Oe. talled surglcaJ information was available for 243 of 247 patlents who had surgery,

tNone cf Ihe followlog ware reported: aspiration, opera- tion at wrang level,

tAny reported complications up to 8 weeks postoperatioo.

None af the following were reported: blood transfusion, cerebrospinal fluid leak, deep wound infectlon, nerve- root inJury, paralysis. cauda equina inJury, wound dehia- oenee, wound hemafoma.

9Rates ere Kaplan-Meier estlmate,s.

Intent-to- Treat Analyses

TABLE 5 shows estimated mean changes

from baseline and the treatment ef- feets (differences in ehanges from base-

line between treatment groups) for 3

months, 1 year, and 2 years. For each

measure and at each point, the treat-

ment effect favors surgery. The treat-

ment effects for the primary outCOIDes

were small and not statístícally

cant at any of the As shown in

FIGURE 2, treatment

SURGICAL VS NONOPERATlVE HERNIA TlON

..

Table 4. Statistlcally Significant Baseline Demdgraphic Characteristics, Comorbid Conditions, Clinical Findings, and Health Status Measures,

by Adherence With Treatment Assignment !..

Assigned to Surgery

No.(%)

Assigned to Nonoperative Treatment

No.(%)

No No

Herniatlon level:j: .45

1 were not symptoms, Range,

severe symptoms, Range tor Oswestrý Disability Index, 0-100; tor soiatica indices, 0-24, tHigher :j:Lower score

Table 5. Treatment Effects for Primar)' and Secondary Outcomes Based on Intent-to-Treat Analyses*

3 Months 1 Year 2 Years

Surgery Nonoperative (n = 198) (n = 211)

Effeot (95%01)

Surgery Nonoperative Treatment Surgery Nonoperative

(n = 202) (n = 213) Effeot (95% CI) (n = 186) (n = 187) Effect

(95% CI)

Prlmary outcomes SF-36 score, mean (SE)t

Bodlly pain

Physicai

Oswestry

signifieant. This test aeeounts for in-

traindividual eorrela tions as de- scribed in the "Methods" seetion.

For the seeondary outcome of scí-

atìca bothersameness, Table 5 and

FIGURE 3 show that there were greater

improvements in thc Scíatica Bother-

someness Index in the surgery graup

at an designated follow-up times: 3

months (treatment effect, -2.1; 95% CI,

-3.4 to -0.9),1 ycar (treatment effect,

-1.6; CI, -2.9 to -0.4), anel 2 years

(treatment effect, -1.6; 95% CI, -2.9 to

-0.3), with resülts of the global hy-

po thesis test being statistically signifi-

cant (P= .003). Patient satisfaction v"ith

symptoms and treatment showed small

effects in favor of surgery while em- plo))"ment status showed small effects

in favor of nonoperative care, but none

of these changes was statistically sig-

nificant. Self-rateel progress showed a

small statistically significant advan-

tage far surgery (P= .04).

As-treated analyses based on treat-

ment receiveel were perfonneel with ael-

justments for the time of surgery and

factors affecting treatment crossover

and missing data. These yieldeel far dif-

ferent results than the intenHo-treat analysis, with strong, statistically sig-

niHcant advantages seen for surgery at

all follow-up times through 2 years. For example, at 1 year the estimated treat-

ment eHeets for the SF-36 bodily pain

SURGICAL VS TREATMENT FüR LUMBAR DlSK HERNIATlON

anel physieal function scales, the ODI,

and the sciatica mcasures were 15.0

CI, 10.910 19.2), 11.5 Cl,

13.6 to 21.5), -15.0 (95% CI, -18.3 to

-l1.ï), and -3.2 CI, -4.3 to -2.1), respeetivdy.

Sensitivity analysis was performed

for 4 different analytic methods of

dealing with the missing data, One

method was based on simple mean ehanges far an patients with data at a

given time point with no special

adjustment for missing data. Two methods useel single imputation methoels-baseline value carried for-

ward and last value earrieel forward.J!

The latter method used the same

mixed-models approach for estimat-

ing mean changes as given in Iable 5

but also adjusted for faetors aHecting

the likelihood of missing data. Treat-

ment effect estimates at 1 year ranged

from 1.6 to 2.9 for the SF-36 bodily

pain seale, 0.74 to 1.4 for thc physical

function scale, -2.2 to -3.3 for the

001, and -1.1 to -1.6 for the scìatiea

measures. Given these ranges, there

appear to be no substantial differences

between any of these methods.

COMMENT Both operateel and nonoperateel pa-

tients with intervertebral disk hernia-

tion improved substantially over a

2-year peliocl. The intent-to-treat analy-

sis in this trial showeel no statistically

significant treatment effects for the pri-

mary outcomes; thc seconclary mea-

sures of sciatica severity and self-

reporteel progress did show statistically

significant aelvantages for surgery,

These results must be vieweel in the

eontext of the substantial rates of non-

aelherenee to assigned treatment. The

pattern of nonadherence is striking be-

cause, unlike many surgical studies,

both the surgical anel non operative

treatment groups were affecteel,35 The

most eomparable previous tria18 had

26% erossover into surgcry at 1 year,

but only 2% erossover out of surgery.

The mixing of treatments elue to cross-

over can be expected to ereate a bias to-

ward the nulp4 The large effects seen

in the as-treated analysis and the char-

acteristics of the crossover patients sug-

gest that the intent-to-treat analysis un- derestimates the true effeet of surgery.

SPORT findings are consistent with

clinical experience in that relief of leg

pain was the most strildng and consis-

tent i.mprovement with surgery. Im- portantly, all patients in this trial had

leg pain with physical examination and imaging Hndings that confirmed a disk

hemiation. There was litde evidenee of

harm from eìther treatment. No pa- tients in either group developed cauda

equina syndrome; 95% of surgical pa-

tients hael no intraoperative complica-

Figure 2. Mean Scores Over Time for SF-36 Bodlly Pain and Physical Function Scales and Oswestry Disability Index

Bodily Paln

1"9.13::3<3" t~on:n=81

Physioal Funotlon Oswestry Dìsability Index

Gi Surgery

A Nonoperative Care

for

SURGICAL VS NONOPERATIVE TREATMENT FOR LUMBAR DlSK HERNIATlON

tions. The most common complica- tion, dural tear, occurred in 4% of patients, sinÜlar to the 20Ic, to 7% noted

in the meta-analysis Hoffman et aV 2.2% seen in the MLSS, and 4% in the

recent series from Stanford.

Oue limitatiou is the potential lack of representativeness of patients agreeing

to be randomized to surgelY or nonop- erative care; however, the characterís- ties of patients agreeing to participate in SPORT were simílar to those in other studies. The mean age of 42

years was similar to the mean ages in the

MLSS,29 the series of Spangfort, 37 and the

randomized trial by Weber, and only slightly older than iliose in the recent se-

ries from Stanford (37.5 years).3Ô The

proportion of patients receiving work- ers' compensation in SPORT (16%) was

similar to the proportion in the Stan-

ford population (19%) but lower than

that in the MLSS population (35%), which specifically oversampled pa- tients receiving compensation. Base-

hne functional status was also similar,

with a mean baselíne om of 46.9 in SPORT vs 47.2 in ilie Stanford series, and

a mean baseline SF-36 physical func-

tíon score of 39 in SPORT vs 37 in the

MLSS.

The strict eligibílity criteria, how- ever, may limit the generalizability of these results. Patients unable to toler-

ate symptoms for 6 weeks and demand- ing earlier surgical intervention were

not included, nor were patients with-

out dear signs and symptoms of nt- diculopathy with confi.nnatory im ag-

ing. We can draw no conclusions regarding ilie efficacy of surgeryin iliese

other groups. However, our entry cri-

tería followed published guidelines for

patient selection for elective diskec-

tomy, and our results should apply to the majority of patients facing a surgi- cal clecision.38,39

To fuHy understand the treatment effect of surgelY compared \v"Íth non- operative treatment, it is worth noting

how each group fared. The improve- ments with surgery in SPORT were símilar to those of prior series at 1 year: for the 001, 31 points vs 34 points in the Stanford series; for the bodily pain

40 points vs 44 in the MLSS; and for sciatica bothersomeness, 10 points

vs 11 in the MLSS. Similarly, Weber8 reported 66% "good" results in the SUf-

gery group, compared with the 76% re- porting "major improvement" and 65%

satisfied with their symptoms in SPORT.

Figure 3. Measures Over Time for Scìatica Bothersomeness Index, Employment Status, Satisfaction With Symptoms, Satisfaction With (are, and Self-rated ImproVèrnent

Sciatlca Bothersomeness SaUsfaction Wlth Symptoms

~ o

()) c !!J

:::;

lj;

,[ ::i'

Global p~ ,003

6 12

Months From Baselíne

SatÎsfactlon Wlth Gare

Work Status

CJ>

C :g o S ?f!.

Months From Baseline

Self-rated Improvement

V) E .8 ~ JJ

~ S 1:J '" "" .SQ 1iJ

rf) ?f!.

24 24

Months From ßaseline

The observed improvements with nonoperative treatment in SPORT were greater than those in the MLSS, result-

ing in the small estimateel treatment effect. The nonoperative improve- ment of 35, anel 9 points in bodily

physícal function, anel sciatica

bothersomeness, respectively, were

much greater than thc improvements of 20,18, and 3 points reported in the

MLSS. The greater improvement with nonoperative treatment in SPORT may be rdated to the large proportion of pa-

tients (43%) who underwent surgery in this group.

The major limitation ofSPORT is the

degree of nonadherence with random-

ized treatment. Given this degree of

crossovcr, it is unlikely that the intent-

to-treat analysis can form thc basis of a valid estimate of the true treatment effect of surgery. The "as-treated" analy-

sis with adjustments for possible con- founders showed much larger effects in favor of surgical treatment. However, this approach does not have the strong

protection against confounding that is

afforded by randomization. We can-

not exclude the possibility that base-

line differences between the as- treated groups, or the selective choke

of some but not other patients to cross

over into surgery, may have affected

these results, even after controlling for

important covariates. Due to practical

and ethical constraints, this srudy was

not masked through the use of sham

procedures. Therefore, any improve-

ments seen with surgery may include

some degree oE "placebo effect."

Another potential limitation is that

the choice of nonoperative treatments

was at Ù1e discretion of the treating phy-

sidan anel patient. However, given the

limited evidence regarding efficacy for

most treatments for lum-

bal' disk herniation and individnal vari- a

SURGICAL VS NüNüPERATIVE TREATMENT FüR LUMBAR orSK HERNIATlüN

transcutaneous electrical nerve stimu-

lation, and braces and corsets, and lÜgher rates of epidural steroid injec-

tions anel use of narcotic analgesics.

This flexible nonoperative protocol had

the advantages of individualizatiol1 that cOl1sidered patient preferences in the

choke of nonoperative treatment and oE reflecting current practice among multidisciplinary spine practices. How-

ever, we cannot make any conclusion regarding the effect of surgery vs any specific nonoperative treatment. Simi- larly, we cannot adequately assess the

relative efficacy of any differences in surgical technique.

CONCLUSION

Patients in both the surgery and non- operative treatment groups improved

substantially over the first 2 years.

Between-group differences in improvements were consistently in favor of surgery for an outcomes and

at an time perioels but were small and

not statistically significant except for

the secondary measures oE sciatica severity and self-rated improvement.

Because of the high numbers of patients who crossed over in both

directions, conclusions about the

superiority or equivalence of the treat-

ments are not warranted based on the

intent-to-treat analysis alane.

Author Contrlbutlons: Dr Weinstein had full access

to all of the data In the study and takes responslbillty

for the integrity of the data and the accuracy of the data analvsis.

Study concept and design: Weinstein, T, Tosteson, Lurle, A. Tosteson, Acquisition of data: Weinstein, T. Tosteson, tude, Abdu, Hillbrand, Boden. Analysis and interpretation of data: Weinstein,

T, Tosteson, Lune, A. Tosteson, Hanscom, Skinner,

Abdu, Hillbrand, Deyo. Drafting of the manuscript: Weinstein, T. Tosteson,

Lurle. Critical revision of the manuscript for important in. telledual content: \Neinstein, T. A. Tosteson, Hanscom, Skinner, Abdu,

Boden, Deyo.

cal Dedsian Making, proceeds to the Department of Or. thopaedics, Dartmouth, Dr Lurie reports that he Fe.

ceives grant support Irom St Francl. Medlcai Technologles and Amerlcan Board 01 Orthopaedic Sur.

gery; has served on advlsory boards for Ortho. MacNeíl Pharmaceutlcals, the Robert Graham Center

of the American Academy of Famííy Practlce, and Cen. toeor; and as a consultant for Myexpertdoctor.com, Pa. elfic Business Group on Health, and the FOíJndation for

Inforrned Medical Decìslon Maklng, Dr A. T csteson re. ports recelvlng funding Irom St Frands Medlcal Tech. nologies. Mr Hanscom reports worklng for the Na- tional Spine Network and recelving funding from Medtronlc. Dr Boden reports seNlng as a consultantfor

Medtronlc. Dr Deyo reports recelvlng research pro.

gram benefits as a gift to the University cf Washington

Irom Synthes, a manufacturer of stJfgical appllances, from which he receives no personal benefits, No other dis.

closures were reDorted. FundingISupport: This was funded by the Na-

tionallnstitute of Arthritis and Musculoskeletal and

Skin Diseases (NIAMS) (U01-AR45444.01 A 1) and the

Office of Research on Women', Health, National in. stltutes of Health; and by the National Institute of Oc- cupatlonal Safety and Heaith, US Centers lor Disease

Controi and Preventlon. The Multldiscìplinary Cllni- cal Research Center in Musouioskeletal Diseases Is

funded by NIAMS (P60.AR048094.01A1), Dr Lurie

Is supported bya Research Career Award from NIAMS (1 K23 AR 048138-01). Role oftne Sponsors: The funding sources had no role in design or conductofthe study; the collectlon, man. agement, analysis, 0, interpretation of the data; orthe preparation, review, or approval of the manuscript.

project Coordlnator, Princlpallnvestlgators, Partici- patlng institutions, and Nurse Coordlnators (in order

by enrol1ment); Judl L. Forman, MPH (Project Coordl. nator), Wiíliam A. Abdu, MD, MS, Barbara Butler- Schmidt, RN, MSN, J. J. Hebb, RN, BSN (Dartmouth- Hitchcock Medical Center); Hany Herkowitz, MD, Gloria

Bradley, BSN, Melissa Lwie, RN (Wiliiam Beaumont Hos. pital); Todd Aibert, MD, Alan Hilibrand, MD, Caroie Si-

mon, RN, MS (Rothman institute atThomas Jefferson

Hospital); Frank Cammisa, MD, Brenda Green, RN, BSN

(Hospital for Special Surgery); Michael Longley, MD, Nancy Fullmer, RN, Ann Marie Fredericks, RN, MSN, CPNP (Nebraska Foundatlon forSpinal Research); Scott

Boden, MD, 5a1iy Lashley, BSN, MSA (Emory University-

The Emory Clinic); Lawrence Lenke, MD, Georgia Sto- bbs, RN, BA (Washington Unlversity); Sanford Emery,

MD, Chrls Furey, MD, Kathy Hlgglns, RN, PhD, CNS

(Unlversity Hospitals of Cleveiand/Case Western Re- seNe); Thomas Errìco, MD, Alex Lee, RN, BSN (Hos. pltal for Joint Dlseases); Harley Goldberg, DO, Pat Ma- lone, RN, MSN, ANP (Kaiser Permanente); Serena Hu,

MD, Pat Malone, RN, MSN, ANP (University of Call.

fornia-San Francìsco); Gunnar Andersson, MD, PhD,

Howard An, MD, Margaret Hlckey, RN, MS (Rush-

Presbyterian-St Luke's Medical Center); Robert Keller,

MD (Mai ne Spine and Rehabilitation).

Enrolllng Physicíans: Wllliam Abdu <Dartmouth- Hltchcock Medlcal Center); David Montgomery, Harry Herkowitz (Wii!iam Beaumont Hospital); Ted Conliffe,

Alan Hilibrand (Rothman Institute at Thomas Jefferson Frank Cammisa (Hos-

Univer;ity-

Foun-

SURGICAL VS NONOPERATlVE TREATMENT FOR LUMBAR DlSK HERNIATlON

nente); Lawrence Kurz (Beaumont); Joseph Dryer (Hos- pital for Joint Diseases); Sanford Emery (University Hospitals of Cleveland/Case Western Reserve); Susan

Dreyer, Howard Levy (Emory); Patrick Bowman (Ne- braska Foundation for Spinal Research); Thomas Errìça

Lee Thibodeau, MD (Maìne Spine and Jeffrey Fischgrund (Beau- mont); Mark Splaine (Darlmouth); John Bendo (Hos- pital for Joint Diseases); Taylor Smith (University of Cali- fomia-San Fmndsco); Enc Phillips (Nebraska Foundation

Research); Dilip Sengupta (Dartmouth); David (Emory); Henry Schmidek (Dartmouth); Har-

iey Go!dberg (Kaiser); Robert Rose (Dartmouth); Slg Ber-

yen (University of Californìa-San Francisco); Frank Phi!- lips, Howard An (Rush-Presbyterian-St Luke's Medical Center); Colleen Olson (Dartmouth); Anthony Margh- erita, John Metzler ('lVashingfon Universìty); Jeffrey Go!d- stein (Hospital for Joint Diseases); Phaedra Mcdonough (Dartmouth); James Farmer (Hospital for Special Sur- gerý); Marsoiais (Case Western); Gwmar Andersson (Rush-Presbyterian-St Luke's); Hilda Magnadottir, Jlm Weinstein, Jon Lurie (Darúnouth); J. X. Yoo (Case West- em); John HeUer (Emory); Jeffrey Spivak (Hospital for Joint Diseases); Roland Hazard (Dartmouth); Michael Schaufele (Emory); Jeffrey Florman (Maine Spine and Rehabilitation); Phìlip Bernini (Dartmouth); Eerie Tr- uurnees (Beaumont); I<. Daniel Riew (Washingfon Uni- verslty); Timothy Burd (Nebraska Foundation for Spl- nai Research); John Rhee (Emory); Henry Sohlman (Case Western); Richard Perry (Hospital for Joint Diseases); Ed- ward Goldberg (Rush-Presbyterian-St luke's); Christo- pher Furey (Case Western). Acknowledgment We thank Tamara S. Morgan, MA, Department of Orthopaedic Surgery, Dartmauth Medi- cal Schaol, for graphie design and produetion, jour- nal Interfacing, and shepherdingthe SPORT study from its original submission and Into the foreseeable fu- ture. Ms Margan Is funded through the department and partially by SPORT. This study Is dedlcated to the memory of Brleanna Weinstein.

REFERENCES

1. Boden SD, Davls DO, Dina TS, Patronas NJ, Wiesel

SW. Abnormal magnetlc-resonance scans of the lum- bar spine in asymptomatic subjeets: a prospeetlve Investlgation. J BOne Joint Surg Am. 1990;72:403-408. 2. Jensen MC, Brant-Zawadzki MN, Obuchowski N, Modic MT, Malkasian D, Ross JS. Magnetic reso- nance imaglng of the lumbar spine in people without back pain. N Engi J Med. 1994;331 :69-73. 3. Saai JA, Saai JS. Nonoperative treatment of herni- ated lumbar Intervertebral dise with radfculopathy. Spine. 1989;14:431-437. 4. Weinstein JN; Dartmouth Atlas Working Group. Dartmouth Atlas of Musculoskeletal Health Care. chl- oago, 111: Amerlcan Hospital Assoclation Press; 2000. 5. Deyo RA, Weinstein JN. Low back pain. N Engl J

Med. 2001 ;344:363-370.

6. Weinstein JN, Bronner KK, Morgan TS, Wenn- berg JE. Trends and geographie variations in major sur-

gery for degenerative diseases of the hip, knee, and spine. Health Aff (Millwood). 2004;(suppl Web ex- dusive):var81-89.

7. Hoffman RM, Wheeler KJ, Deyo RA. Surgery for herniated lumbar dises: a iiterature synthesis. J Gen Intern Med. 1993;8:487-496.

8. Weber H. Lumbar dise herniation: a controlled, pro- speetlve study with ten years of observation. Spine.

1983;8:131-140. 9. Buttermann GR. Treatment of lumbar dise hemia- lion: epldural steroid injeetion compared wìth discec- tomy: a prospeetive, randomized study. J Bane Joint 5urg Am. 2004;86:670-679. 10. Gibson JN, Grant IC, WaddelI G. The Cochrane review of surgery for lumbar disc prolapse and de- generative lumbar spondylosls. Spine. 1999;24:1820-

1832. 11. Gibson JN, Grant IC, WaddelI G. Surgery forium- bar disc prolapse. Cachrane Database Syst Rev. 2ooo;(3):CD001350.

12. Jordan J, Shawver Morgan Weinstein Kon- stantinou K. Herniated lumbar Clin June

2003:1203-1215. 13. B!rkmeyer NJ, Weinstein JN, Tasteson AN, et ai. Design of the Spine Patient Outcomes Research Trial (SPORT). Spine. 2002;27:1361-1372. 14. Fardon DF, Mlletle PC. Nomenclature and das- sification of lumbar disc pathology: recommenda- tions offne Combined T ask Forees of the North Amerl- can Spine Sodety, ."Imerican Sodety of Spine Radiology, and American Society of Neuroradlology. Spine. 2001; 26:E93-E113. 15. Delamarter R, McCullough J. Microdisceetomy and microsurgicallaminotomies. In: Frymoyer J, ed. The Adult Spine: Principles and Pradice. 2nd ed. Phila-

delphia, Pa: lippincotl-Raven Publlshers; 1996. 16. Spengler DM. Lumbar disceetomy: results with IIm- Ited dlsc excislon and selectlve foraminotomy. Spine. 1982;7:604-607. 17. Cummins J, lurie JD, Tosteson T, et al. Descrlp- tive epidemiology and prior healthcare utilizatlon of patlents in the Spine Patient Outcomes Research Tri- al's (SPORT) three observationai cohorts: disc hernla- tion, spinai stenosis, and degenerative spondylolistnesls. Spine. 2006;31 :806-814. 18. Ware JE Jr, Sherboume D. The MOS 36-lfem short-

form health survey (SF-36), I: conceptual framework and Item seiection. Med Care. 1992;30:473-483. 19. Ware JE Jr. SF-36 Health Survey: Manual and In- terpretation Gulde. Boston, Mass: Nimrod Press;

1993. 20. McHomey CA, Ware JE Jr, Lu JF, Sherbourne CD. The MOS 36-item Short-Form Heaith Survey (SF- 36), 111: tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care. 1994;32:40-66. 21. Stewart AL, Greenfleid S, Hays RD, et al. Func- tional status and well-being of patients with chronic conditions: results from the Medleal Outcomes Study. JAMA. 1989;262:907-913. 22. Daltroy lH, Cals-Saril WL, Kau JN, Fossel i\H, liang MH, The North Amerlcan Spine Soelety iumbar

spine outcome assessment Instrument: rellability and validity tests. Spine.

23. Deyo RA, Diehl AK. medi- cal care for low-back pain. Spine. 1986;11 :28-30. 24. Atlas SJ, Deyo RA, Patriok DL, Convery K, Keller RB, DE. The Quebec Task Force classification for disorders and the severlty, treatment, and outcomes of sciatica and lumbar spinal stenosis. Spine. 1996;21 :2885-2892. 25. Patrick DL, Deyo RA, Atlas SJ, Slnger DE, Chapin A, Keller RB. health-related of life in with Spine.

Interaetive

211. 27. Weinstein JN. Partnership: doetor and patient: ad- vocacy for Informed choice vs. informed consent. Spine. 2005;30:269-272. 28. Friedman L, Furberg C, DeMets D. The mndom- ization process. In: Fundamentals of Clinieal Trials. 3rd ed. Cambridge, Mass: Springer-Verlag; 1998:61- 81. 29. Atlas 5J, Deyo RA, Keller RB, et al. Tne Maine Lum- bar Spine Study, 11: 1-year outcomes of surgical and nOl1surglcal management of sdatlca. Spine. 1996;21 : 1777-1786. 30. Little R, Rubin D. 5tatistieal Analysis With Miss- ing Data. 2nd ed. Philadelphia, Pa: John Wiley & Sons; 2002. 31. Diggie P, Haeagery P, Uang K, ZegerS. The Analy- sis of Longitudinal Data. 2nd ed. Oxford, England: Ox- ford Unlversity Press; 2002. 32. Fitzmaurice G, Laird N, Ware J. Applied Longi- tudinal Analysis. Philadelphia, Pa: John Wiley & Sons;

2004. 33. Aitman DG, Schulz KF, Moher D, et al. The re- vlsed CONSORT statement for reporting random- ized trials: explanation and eJaboration. Ann Intern Med.2001;134:663-694.

34. Melnert Cl. C!lnical Trials: Design, Conduct, and Analysis. New Y ork, NY: Oxford Universlty Press; 1986. 35. Kuppermann M, Varner RE, Summitt RL Jr, et al.

Effeet of hystereetomy vs medical treatment on health- related quallty of Ilfe and sexual functJoning: the medi- cine or surgery (Ms) mndomlzed trial.JAMA. 2004;291: 1447-1455. 36. Carragee EJ, Han MY, Suen PW, Kim D. Clinlcal

outcomes after lumbar disceetomy for sciatica: the ef- feets offragment type and anular competence. J Bone Joint Surg Am. 2003;85:102-108. 37. Spangfort EV. The lumbar disc herniation: a com- puter-aided analysis of 2,504 operations. Acta Or- thop Scand Suppl. 1972;142:1-95. 38. Agency for Health Care Poiicy and Research. Acute Low Back Problems in Adult>. Bethesda, Md: US Dept of Health & Human Services; 1994. 39. North American Spine Sodety. Herniated disco in:

North American Spine Society Phase !II Cllnieal Guide- lines for Multidisciplinary Spine Care Speclalists. LaGrange, 111: NASS; 2000.