SPINE Volume 26, Number 10, pp 1179–1187
©2001, Lippincott Williams & Wilkins, Inc.
Surgical and Nonsurgical Management of Sciatica
Secondary to a Lumbar Disc Herniation
Five-Year Outcomes From the Maine Lumbar Spine Study
Steven J. Atlas, MD, MPH,* Robert B. Keller, MD,† YuChiao Chang, PhD,*
Richard A. Deyo, MD, MPH,‡ and Daniel E. Singer, MD*
Study Design. A prospective cohort study.
Objective. To assess 5-year outcomes for patients with
sciatica caused by a lumbar disc herniation treated surgi-
cally or nonsurgically.
Summary of Background Data. There is limited knowl-
edge about long-term treatment outcomes of sciatica
caused by a lumbar disc herniation, particularly the rela-
tive benefits of surgical and conservative therapy in con-
temporary clinical practice.
Methods. Eligible, consenting patients recruited from
the practices of orthopedic surgeons, neurosurgeons, and
occupational medicine physicians throughout Maine had
baseline interviews with mailed follow-up questionnaires
at 3, 6, and 12 months and annually thereafter. Clinical
data were obtained at baseline from a physician question-
naire. Outcomes included patient-reported symptoms of
leg and back pain, functional status, satisfaction, and em-
ployment and compensation status.
Results. Of 507 patients initially enrolled, 5-year out-
comes were available for 402 (79.3%) patients: 220 (80%)
treated surgically and 182 (78.4%) treated nonsurgically.
Surgically treated patients had worse baseline symptoms
and functional status than those initially treated nonsur-
gically. By 5 years 19% of surgical patients had under-
gone at least one additional lumbar spine operation, and
16% of nonsurgical patients had opted for at least one
lumbar spine operation. Overall, patients treated initially
with surgery reported better outcomes. At the 5-year fol-
low-up, 70% of patients initially treated surgically re-
ported improvement in their predominant symptom (back
or leg pain) versus 56% of those initially treated nonsur-
gically (P ⬍ 0.001). Similarly, a larger proportion of sur-
gical patients reported satisfaction with their current sta-
tus (63% vs. 46%, P ⬍ 0.001). These differences persisted
after adjustment for other determinants of outcome. The
relative advantage of surgery was greatest early in fol-
low-up and narrowed over 5 years. There was no differ-
ence in the proportion of patients receiving disability
compensation at the 5-year follow-up. The least symp-
tomatic patients at baseline did well regardless of initial
treatment, although function improved more in the sur-
gical group.
From the *General Medicine Division and the Medical Practices Eval-
uation Center, Medical Services, Massachusetts General Hospital,
Harvard Medical School, Boston, Massachusetts; the †Maine Medical
Assessment Foundation, Augusta, Maine; and the ‡Center for Cost and
Outcomes Research and the Departments of Medicine and Health Ser-
vices, University of Washington, Seattle, Washington.
Supported by grants from the Agency for Health Care Policy and Re-
search (HS-06344, HS-08194, and HS-09804).
Acknowledgment date: April 18, 2000.
Acceptance date: August 21, 2000.
Device status category: 1.
Conflict of interest category: 14.
We appreciate the assistance of Monica McLain, MA, with data anal-
ysis and Valerie Soucie with patient follow-up.
Conclusions. For patients with moderate or severe sci-
atica, surgical treatment was associated with greater im-
provement than nonsurgical treatment at 5 years. How-
ever, patients treated surgically were as likely to be
receiving disability compensation, and the relative benefit
of surgery decreased over time. [Key words: sciatica, her-
niated lumbar disc, prospective cohort study, outcome
research, lumbar disc surgery, natural history] Spine
2001;26:1179 –1187
Sciatica3,22 caused by a lumbar herniated disc is the most
common cause of radicular leg pain in adult working
populations.7 Although most patients improve over sev-
eral weeks, surgical treatment is frequently considered
for patients with symptoms that are persistent or severe.1
More than 250,000 elective lumbar spine operations are
performed each year in the United States, with discec-
tomy being the most common procedure.24 Despite the
availability of clinical guidelines, wide variations exist in
r a t e s o f s u r g e r y a c r o s s d i ff e r e n t g e o g r a p h i c
regions.5,11,24,25
The long-term benefit of surgical versus nonsurgical
treatment for patients with sciatica caused by a herniated
lumbar disc has been assessed in only one randomized
clinical trial.27 In this trial surgery was superior to con-
servative treatment at the 1-year follow-up and nonsig-
nificantly better at 4 years. By 10 years the outcomes of
the two treatments were similar. This study randomized
only 126 inpatients at a single referral center in Norway
and was begun almost 30 years earlier. The goal of the
current study was to assess whether these outcomes ap-
ply to contemporary clinical practice in the United
States.12
Materials and Methods
Details about the study design and methods, and 1-year out-
comes have been previously published.4,12 The Maine Lumbar
Spine Study is a prospective observational study of patients
presenting with sciatica to orthopedic surgeons, neurosur-
geons, and occupational medicine physicians in community-
based practices throughout the state of Maine. Treatment, ei-
ther surgical or nonsurgical care, was not prescribed by the
study but was determined in a routine clinical manner by the
patient and the physician. We have previously reported that for
patients choosing surgery 95.9% underwent open discectomy.
Operative findings were consistent with a extruded disc frag-
ment or herniation in 91.2%.4 For patients electing nonsurgi-
cal treatment, back exercises, physical therapy, bed rest, spinal
manipulation, narcotic analgesics, and epidural steroids were
most frequently used.4
1179
1180 Spine • Volume 26 • Number 10 • 2001
Study Population. The diagnosis of sciatica caused by a lum-
bar disc herniation was based on the physician’s assessment.
Specific radiographic findings were not required for study en-
try. To restrict the study to patients for whom surgery would be
elective and acceptably safe, patients were excluded for the
following reasons: prior lumbar spine surgery, cauda equina
syndrome, developmental spine deformities, vertebral frac-
tures, spine infection or tumor, inflammatory spondylopathy,
pregnancy, or severe comorbid conditions. Five hundred seven
patients with sciatica were enrolled from 1990 to 1992, with
enrollment stratified to obtain roughly equal numbers of sur-
gical and nonsurgical patients. Patients initially choosing non-
surgical treatment but who underwent surgery before the
3-month evaluation (n ⫽ 38, 14%) were considered in all anal-
yses to have been surgically treated. These crossover patients
had baseline characteristics and clinical findings similar to
other surgical patients, and 5-year outcomes by treatment
group were not altered by how these patients were classified.
Because patients referred to the study represented only a sam-
ple of those eligible, a survey of randomly selected surgical
patients who were not enrolled found similar baseline features
and 2-year outcomes compared with enrolled patients.4,12
Study Protocol. For eligible consenting patients, baseline in-
terviews were conducted in person by trained research assis-
tants. Follow-up was obtained by mailed questionnaires at 3, 6,
12, 24, 36, 48, and 60 months. Physicians completed a baseline
questionnaire including history, physical and neurologic find-
ings, diagnostic procedure results, and planned treatment.
Baseline imaging studies (CT, MRI or myelogram) were or-
dered as directed by the treating physicians. These were avail-
able for independent review in 55% of patients.4 The Human
Subjects Review Committee of the University of Washington,
Seattle, and the Maine Medical Center Institutional Review
Board, Portland, Maine, approved the study protocol.
Baseline and Outcome Assessment. Patient data collected
at baseline included demographic information, employment
and disability status, comorbid conditions, past spine history,
physical examination and imaging findings, symptoms, and
functional status. Symptoms in the past week, including the
frequency (from 0 “not at all” to 6 “always”) and bothersome-
ness (from 0 “not bothersome” to 6 “extremely bothersome”)
of low back pain, leg pain, leg or foot weakness, leg numbness,
and pain in the back or leg while sitting, were assessed at base-
line and follow-up. Sciatica frequency and bothersome indexes,
each with scores ranging from 0 to 24, were created by sum-
ming the four leg-related questions.19 For these scales higher
scores indicated more severe symptoms.
Back-specific functional status was measured using the
modified Roland disability scale.19,21 Modified Roland scores
range from 0 to 23, reflecting a sum of items checked as being
present.19,21 Generic health status was assessed with the Med-
ical Outcomes Study Short Form 36-item questionnaire (SF-
36).26 The SF-36 describes 8 domains of health with each
scored from 0 (poor health) to 100 (optimal health). Higher
scores reflect worse function for the Roland scale but better
function for the SF-36.
Change in symptoms and functional status was assessed by
subtracting results at 5-year follow-up from those at baseline.
Patients were also asked to describe the improvement in low
back and leg pain at each follow-up visit relative to baseline.
Responses ranged from “much worse” to “completely gone”
on a 7-point scale. The primary symptom outcome was im-
provement in the patient’s predominant symptom, either back
or leg pain, as indicated at baseline. The outcome was catego-
rized as improved if the response was “better,” “much better,”
or “completely gone,” the same if the response was “about the
same” or “a little better,” and worse if the response was “a
little worse” or “much worse.”
Patient satisfaction was assessed with three questions: rating
of overall quality of life, satisfaction with the current state, and
for surgically treated patients, whether they would still choose
a back operation. Patients were considered to be satisfied with
their current state if they replied they were “delighted,”
“pleased,” or “mostly satisfied” on a 7-point scale. Disability
and work status were assessed at each follow-up.2 Additional
lumbar spine surgery during the follow-up period (reopera-
tions for the surgically treated cohort and first operations for
the nonsurgically treated cohort) was assessed for all patients
using physician office records, a state hospital discharge data-
base, and patient responses to follow-up surveys.
Statistical Analyses. Statistical analyses were performed us-
ing a commercial software package (Statistical Analysis Sys-
tem, SAS Institute, Cary, NC). We compared change in symp-
toms and functional status at 5 years between surgical and
nonsurgical groups using linear regression models controlling
for baseline scores. Ratings of current findings compared with
baseline used categorical responses, so distributions were di-
rectly compared between treatment groups using ␹2 tests or
Fisher’s exact tests. To adjust for baseline differences in patient
characteristics between the two treatment groups, logistic re-
gression models were used to estimate the marginal effect of
treatment option on the change over 5 years. Analyses stratified
by the baseline Sciatica Bothersome Index score were also used
to compare treatments in more homogeneous populations.4
To examine the pattern of change over time, we performed
repeated-measures analysis to include data from all follow-ups
(n ⫽ 482, 95.1%). Mixed-effects models and logistic regression
models with Generalized Estimating Equations were used to
model the correlation structure of the repeated measures within
each patient.14 The treatment ⫻ time interaction was used to
test how treatment effect differed over time in these models.
Results
Baseline Characteristics
Patients in both treatment groups were generally young,
healthy, and predominantly male (Table 1). Most were
not college graduates and had been employed in the past
4 weeks. For the majority, symptoms had been present
less than 6 months. Patients treated surgically were less
likely to have had a past episode of back pain or sciatica,
or be receiving Workers’ Compensation. However, sur-
gically treated patients were more likely to have abnor-
mal physical findings, moderate or severe findings on
imaging studies, severe back and leg pain, and greater
back-related and overall disability than patients treated
nonsurgically (Tables 1 and 2). Despite worse back-
related symptoms and findings, surgically treated pa-
tients had better SF-36 general health perceptions (Table
1). Although the clinical features of the treatment groups
were different, there was considerable overlap.
 Five-Year Outcomes of Sciatica • Atlas et al 1181

Table 1. Patient Characteristics and Features of Back Disorder at Baseline Evaluation

Treatment Group

Surgical Nonsurgical
Characteristic [n (%)]* (n ⫽ 220) (n ⫽ 182) P Value†

Age (yr) [mean (SD)] 43.2 (12.2) 42.6 (11.9) 0.62

Gender (male) 137 (62.3) 108 (59.3) 0.61
Education (college graduate) 75 (34.1) 47 (25.8) 0.08
Comorbid illnesses (yes)‡ 62 (28.2) 41 (22.5) 0.21
Employed in past 4 weeks 123 (55.9) 91 (50.0) 0.27
Receiving or applying for Workers’ Compensation 58 (26.5) 90 (50.0) ⬍0.001
Past episodes of back pain (none) 86 (39.1) 93 (51.1) 0.02
Positive straight leg raise test 160 (73.1) 88 (49.2) ⬍0.001
Abnormal examination findings [mean (SD)]§ 1.4 (1.0) 1.1 (0.9) ⬍0.001
Radiographic image reviewed (n)¶ 120 74
Moderate or severe findings# 99 (82.5) 44 (59.5) ⬍0.001
Quebec classification (category 4 or 6)** 103 (47.2) 54 (30.2) ⬍0.001
Length of current episode (⬍6 mo) 150 (68.2) 129 (71.3) 0.52
Unilateral leg pain (yes) 185 (84.1) 129 (72.1) 0.005
SF-36 score (0–100) [mean (SD)]††
Physical function 32.4 (26.9) 45.7 (26.6) ⬍0.001
Bodily pain 19.4 (15.8) 35.5 (23.8) ⬍0.001
Role emotional 49.5 (43.7) 55.2 (44.0) 0.20
Mental health 62.1 (19.1) 65.6 (20.7) 0.08
General health 77.9 (18.7) 72.2 (19.5) 0.003
* The data are expressed as the number of patients with the percentage in parentheses, except as noted. Denominators differ slightly among variables because
not all patients answered each question on the survey.
† P values compare surgical and nonsurgical treatment groups using Fisher’s exact test or t test.
‡ Any self-reported chronic pulmonary disease, heart disease, stroke, cancer, or diabetes.
§ The mean number of positive physical examination findings for a patient including unilateral strength, sensation, or reflex abnormality (range 0 –3 findings).
¶ Any computerized tomography, magnetic resonance imaging, or myelogram available for independent review.
# Global rating from normal to severe by study neuroradiologist blinded to treatment group and clinical information.
** The Quebec classification system defines 11 categories including sciatica with distal extremity radiation and neurologic findings (category 4), or the above with
radiographic findings of nerve root compression (category 6).3,22
†† Higher scores indicate better function.

Change in Symptoms and Functional Status Over
5 Years
Five-year outcomes were available for 402 of 507
(79.3%) patients, 80.6% of those treated surgically and
77.8% of those treated nonsurgically. Patients rated the
frequency and bothersomeness of low back pain and leg
symptoms in the past week at baseline and at the 5-year
follow-up (Table 2). For both low back pain and leg
pain, surgically treated patients rated the frequency and
severity of symptoms as worse at baseline evaluation and
better at the 5-year follow-up compared with patients
treated nonsurgically. The frequency and bothersome-
ness of sciatica symptoms were also worse at baseline
and better at follow-up for surgical patients. For each
symptom question the change in symptoms at 5 years
was significantly better for surgically treated patients (all
P ⬍ 0.001). Similarly, surgically treated patients re-
ported significantly greater improvement in back-specific

Table 2. Change in Symptoms and Functional Status From Baseline to 5-year Follow-up According to Initial Treatment
Surgical Treatment* Nonsurgical Treatment*

Variable [mean (SD)] Baseline‡ 5-year‡ Change§ Baseline‡ 5-year‡ Change§ P Value¶

Low back pain in the past week

Frequency score 4.4 (1.9) 1.9 (1.7) ⫺2.4 (2.2) 3.7 (2.2) 2.5 (1.8) ⫺1.2 (2.2) ⬍0.001
Bothersome score 4.2 (2.0) 2.0 (1.8) ⫺2.2 (2.3) 3.5 (2.1) 2.5 (1.9) ⫺1.0 (2.2) ⬍0.001
Leg pain in the past week
Frequency score 5.0 (1.4) 1.5 (1.7) ⫺3.5 (2.1) 3.5 (2.1) 2.0 (1.9) ⫺1.5 (2.2) ⬍0.001
Bothersome score 5.2 (1.4) 1.5 (1.8) ⫺3.7 (2.3) 3.6 (2.2) 1.9 (1.8) ⫺1.7 (2.2) ⬍0.001
Sciatica index†
Frequency 17.8 (5.2) 6.3 (6.1) ⫺11.4 (7.3) 12.3 (6.6) 7.5 (6.4) ⫺4.6 (7.1) ⬍0.001
Bothersome 17.6 (5.1) 5.8 (6.2) ⫺11.7 (7.2) 11.9 (6.3) 7.2 (6.5) ⫺4.6 (6.9) ⬍0.001
Modified Roland scale 17.6 (4.2) 6.2 (6.7) ⫺11.4 (7.0) 13.3 (6.0) 7.6 (6.8) ⫺5.8 (6.8) ⬍0.001
* The number of patients reporting each variable ranged from 212 to 218 for surgical patients and from 178 to 180 for nonsurgical patients.
† The frequency and bothersome indices are the sum of four questions (each response scored from 0 to 6): leg pain, leg or foot weakness, leg numbness, and
pain in the back or leg while sitting.
‡ Higher mean scores at baseline and 5 years indicate worse symptoms or function for all variables. Low back and leg pain scores range from 0 to 6. Sciatica index
scores range from 0 to 24, and modified Roland scale scores range from 0 to 23.
§ Change is calculated as score at 5 years minus score at baseline. Negative values indicate improvement for all variables.
¶ P values compare the change between surgical and nonsurgical treatment groups using multiple linear regression models that control for baseline score.
1182 Spine • Volume 26 • Number 10 • 2001

Table 3. Patient-Reported Improvement in Symptoms and Satisfaction and Disability Status at 5-Year Follow-Up
Treatment Group

5-Year Outcomes (%) Surgical Nonsurgical P Value*

Low back pain compared to baseline† (n ⫽ 212) (n ⫽ 179) 0.009

Improved 69.8 54.8
Same 19.8 30.7
Worse 10.4 14.5
Leg pain compared to baseline† (n ⫽ 206) (n ⫽ 170) 0.008
Improved 71.4 59.4
Same 15.5 28.8
Worse 13.1 11.8
Predominant symptom compared to baseline†‡ (n ⫽ 206) (n ⫽ 173) 0.005
Improved 70.4 56.1
Completely gone 27.7 11.6
Much better 29.1 26.0
Better 13.6 18.5
Same 16.5 30.0
Worse 13.1 13.9
Quality of life (at least moderately improved)§ 79.4 56.3 0.001
Satisfied with current state (yes)¶ 63.0 45.9 0.001
If surgery, still choose back operation (yes)** 82.1 —
If receiving Workers’ Compensation at entry, current status (n ⫽ 58) (n ⫽ 90)
Receiving any disability compensation 20.7 20.2 1.00
Working 82.8 70.0 0.12
If employed at entry, current status (n ⫽ 123) (n ⫽ 91)
Receiving any disability compensation 4.9 6.7 0.76
Working 91.1 84.6 0.20
* P values assessed using Fisher’s exact or ␹2 tests for categorical variables.
† Symptom severity was reported to be improved if the response was “better” to “completely gone,” the same if the response was “about the same” or “a little
better,” and worse if the response was “a little worse” or “much worse.”
‡ The predominant symptom, either back or leg pain, as rated by the patient at baseline.
§ N ⫽ 218 for surgical and N ⫽ 174 for nonsurgical cohorts.
¶ N ⫽ 219 for surgical and N ⫽ 181 for nonsurgical cohorts.
** Not asked in the 5-year questionnaire. Includes patient responses from the most recent 3- or 4-year follow-up (n ⫽ 190).

functional status using the modified Roland scale. Gen- tures, patients treated surgically were still more likely to
eral health perceptions from the SF-36 worsened slightly report improvement in the predominant pain symptom
over 5 years to a comparable degree in both treatment (odds ratio [OR] ⫽ 2.1; 95% confidence interval [CI],
groups (data not shown). 1.3–3.6). Similarly, patients treated surgically were more
likely to report satisfaction with their current state at 5
Patients’ Global Evaluation at 5 Years
years (OR ⫽ 3.0; 95% CI, 1.8 –5.2). In these models not
When asked to contrast their status at 5 years with that
receiving Workers’ Compensation at study entry, less
at baseline, a higher percentage of patients treated surgi-
severe back pain and sciatica, and better baseline scores
cally reported improved low back pain and leg pain (Ta-
on the SF-36 role emotional subscale were also associ-
ble 3). The predominant pain symptom, either low back
ated with greater satisfaction.
or leg pain based on the patient’s report at entry, was
Although patients treated surgically had significantly
improved in 70% of surgical patients compared with
greater improvement in symptoms, functional status,
56% of nonsurgical patients (P ⫽ 0.005).
and satisfaction than patients treated nonsurgically,
Patient satisfaction at the 5-year follow-up was also
work disability outcomes were similar (Table 3). Among
significantly better among those treated surgically. Satis-
patients receiving Workers’ Compensation at study en-
faction with the patient’s current state was reported by
try, 21% of surgically treated and 20% of nonsurgically
63% of surgically treated patients versus 46% of non-
treated patients were receiving some form of disability
surgically treated patients (P ⬍ 0.001). Among surgical
compensation at 5 years (P ⫽ 0.94). Regardless of Work-
patients, most (82%) stated that if they could go back in
ers’ Compensation or working status at entry, the per-
time, they would still definitely or probably choose to
centage of surgically treated patients working at the
have their back operation.
5-year follow-up was slightly but not statistically sig-
Surgical treatment remained an independent predictor
nificantly higher than for nonsurgically treated pa-
of symptom improvement and satisfaction in logistic re-
tients (P ⬎ 0.1).
gression models. Baseline factors associated with im-
provement in the predominant pain symptom included Lumbar Spine Surgery After Initial Treatment
younger age, not receiving Workers’ Compensation, Fifty-three of 273 surgically treated patients (19.4%)
shorter duration of symptoms, and less severe SF-36 had at least one reoperation over 5 years (median time to
bodily pain score. Even after controlling for these fea- reoperation, 19 months). Forty-five of these patients

Figure 1. Five-year outcomes of
additional operation for patients
initially treated surgically or
nonsurgically. Percent report-
ing improvement in the predom-
inant symptom and satisfaction
with the current state for pa-
tients initially treated surgically
(44 having reoperation vs. 175
who did not) or nonsurgically
(32 having subsequent opera-
tion vs. 149 who did not).
(84.9%) completed the 5-year follow-up. Their baseline
findings were similar to surgical patients not undergoing
a reoperation (data not shown). However, 5-year pre-
dominant pain symptom and satisfaction outcomes were
significantly worse for these patients compared with
those who had only one operation (Figure 1).
Among patients initially treated nonsurgically, 38 of
234 (16.2%) underwent a lumbar spine operation be-
tween 3 and 60 months of follow-up (median time to
operation, 10 months). Of these 38 patients, 32 (84.2%)
completed the 5-year follow-up, and their baseline find-
ings were comparable with the nonsurgical patients who
did not have surgery (data not shown). Five-year out-
comes were slightly but nonsignificantly worse for non-
surgical patients who had a subsequent operation com-
pared with those who did not (Figure 1).
Outcomes Among Patients Least Symptomatic
at Baseline
Patients who were least symptomatic at entry to the
study appeared to benefit less from surgery because the
Table 4. Five-Year Outcomes of Least Symptomatic Patients at Entry*
5-Year Outcomes
Predominant symptom (% better)‡
Sciatica Frequency Index [mean change (SD)]§
Roland scale [mean change (SD)]§ ⫺10.8 (7.4)
Quality of life (% at least moderately improved)
Satisfied with current state (% yes)
* Least symptomatic (mild) patients defined as having scores at entry in the lowest quartile (0 –11) of the Sciatic Bothersome Index.4
† P values calculated using Fisher’s exact test or t tests.
‡ Predominant symptom at entry, either leg or back pain, reported as improved (“better,” “much better,” or “completely gone”).
§ Difference between 5-year follow-up and baseline. A negative number implies improved functional status.
Five-Year Outcomes of Sciatica • Atlas et al 1183
outcomes of those treated nonsurgically in this group
were generally good (Table 4). For example, the relative
odds of reporting satisfaction with current state at 5
years was 1.3 (surgical vs. nonsurgical treatment) among
patients in the least symptomatic tertile, compared with
3.7 for patients in the highest two tertiles of the baseline
Symptom Bothersome Index (P ⫽ 0.02, interaction be-
tween treatment group and baseline severity in logistic
regression models). This reflected the fact that 66% of
nonsurgically treated patients were satisfied in the least
symptomatic group compared with only 30% in the
moderate and severe baseline symptom categories. The
improvement attributable to surgery was also smaller
among the least symptomatic patients for pain improve-
ment, quality of life, change in sciatica frequency, and
the Roland functional status (although not all interac-
tions were statistically significant).
Outcomes of Those Not Completing 5-Year Follow-Up One
hundred five patients did not complete the 5-year follow-
up: 53 surgically and 52 nonsurgically treated. Such pa-
Treatment Group
Surgical Nonsurgical
(n ⫽ 24) (n ⫽ 80) P Value†
77.2 69.3 0.60
⫺4.2 (6.3) ⫺2.1 (5.6) 0.19
⫺4.3 (7.0) ⬍0.001
87.5 74.3 0.26
70.3 65.8 0.65
1184 Spine • Volume 26 • Number 10 • 2001

Figure 2. Time course of symp-

tom and functional status out-
comes. Assessed at initial eval-
uation and at 3-, 6-, 12-, 24-, 36-,
48-, and 60-month follow-up for
nonsurgical (-F-) and surgical
(-䉫-) treatment. Values are
mean ⫾ 2 SE. A, Sciatica fre-
quency index. B, Modified Ro-
land score. P ⬍ 0.05 for inter-
action between time and group.

tients were younger (39.5 vs. 42.9 years), more likely to
be male (73.3% vs. 61%), receiving Workers’ Compen-
sation (48.6% vs. 37.1%), and less likely to have com-
pleted college (13.3% vs. 30.3%) (all P ⬍ 0.05). Eighty
of these 105 patients (76.2%) returned at least one prior
follow-up survey. Although the last outcomes reported
by these patients were worse than those completing the
5-year follow-up, they were similarly worse for both sur-
gical and nonsurgical patients. If one includes outcomes
from the last available follow-up survey (3– 60 months)
for the surgical (n ⫽ 265, 97.1%) or nonsurgical (n ⫽
217, 92.7%) cohort, the comparative results and conclu-
sions are unchanged. For example, the predominant
symptom was improved in 67.7% of surgically treated
and 54.0% of nonsurgically treated patients (P ⫽ 0.003).
Time Course of Outcomes
We have previously shown improvement from baseline
to 3 months for all patients, although surgically treated
patients had significantly greater change.4 From the
3– 60-month follow-up, there was narrowing of the rel-
ative benefit of surgical treatment on the frequency of
sciatica symptoms (Figure 2A) and the Roland back-
specific functional status score (Figure 2B). Most of this
narrowing occurred over the first 2 years and subse-
quently remained essentially unchanged (Figure 2). Yet,

Figure 3. Time course of satis-
faction outcome. Assessed at
3-, 6-, 12-, 24-, 36-, 48-, and 60-
month follow-up for nonsurgical
(-F-) and surgical (-䉫-) treat-
ment. Percent satisfied with
current state ⫾ 2 SE.
despite narrowing of symptoms and functional status
outcomes over time, 5-year results remained significantly
better for patients surgically treated (Tables 2 and 3).
The percentage of patients reporting satisfaction with
their current state at each follow-up improved for both
treatment groups over time to a similar extent (P ⫽ 0.24
for interaction between time and group, Figure 3).
Discussion
Our study prospectively followed patients with sciatica
caused by a herniated lumbar disc initially treated either
surgically or nonsurgically over 5 years. Patients treated
surgically in this study had more severe baseline symp-
toms and worse functional status than those treated non-
surgically but had better 5-year outcomes. Although sur-
gically treated patients reported better symptoms,
functional status, and satisfaction, the relative advantage
of surgery decreased over time. Disability and work sta-
tus outcomes were similar regardless of initial treatment.
Our results are similar to other studies that have ex-
amined long-term outcomes of patients with sciatica be-
cause of a herniated lumbar disc.10 In the Weber study,27
good results were reported at 4 years in 70.2% of pa-
tients initially randomized to surgery and 51.5% of those
initially receiving conservative treatment, although this
difference did not reach statistical significance. These re-
sults are almost identical to our finding of improvement
in the predominant pain symptom; our results achieved
statistical significance because of larger numbers. Other
prospective, randomized studies of patients with sciatica
have predominantly compared short-term outcomes of
standard discectomy, chemonucleolysis, or conservative
care. In the one study that provided long-term follow-up,
outcomes after 10 years were at least moderately im-
proved in 77% of patients randomized to chymopapain
Five-Year Outcomes of Sciatica • Atlas et al 1185
(20% had subsequent surgery) and 38% of patients ran-
domized to conservative care (although 47% had subse-
quent surgery).8 Few other studies have prospectively
evaluated long-term (at least 3 years) outcomes of sur-
gery for herniated lumbar disc, and we are aware of only
one that included nonsurgically treated patients. 17
Among uncontrolled studies of surgically treated pa-
tients, positive results have been reported in 50 – 87%
over a 3–10-year follow-up.13,17,28
In our study the 5-year follow-up revealed a narrow-
ing of the difference in outcomes among treatment
groups, although surgically treated patients continued to
report significantly greater improvement at 5 years. The
narrowing of outcomes occurred primarily over the first
2 years and was related to continued improvement in
nonsurgically treated patients rather than deteriorating
outcomes in surgical patients. Further narrowing of these
results over time seems unlikely given the stable out-
comes we observed over the last 3 years of follow-up.
Although surgical treatment was associated with bet-
ter symptoms, functional status, and satisfaction, it had
no significant effect on disability or work outcomes at 5
years. Disability and work status are likely related to
many factors in addition to medical treatment. These
factors may include accommodations in the workplace,
job characteristics (tasks, autonomy, satisfaction, etc.),
other sources of income, and local economic fac-
tors.9,15,18 In contrast, symptoms and daily functioning
may be more directly affected by medical treatment. Our
results suggest the importance of measuring patient-
oriented outcomes in addition to work and disability
status and reporting these outcomes separately.
Almost 20% of patients in the surgical cohort had a
second spinal operation over 5 years. This rate is consid-
erably higher than the 10% cited by Hoffman et al in
1186 Spine • Volume 26 • Number 10 • 2001
their review of lumbar disc surgery,10 but it is compara-
ble to a recent report of reoperation rates in Washington
state.16 Outcomes of our surgical patients undergoing a
reoperation were worse compared with those who did
not have additional surgery. Reported outcomes of re-
peat lumbar disc surgery are variable and may relate to
the etiology of the recurrent symptoms.20
Among nonsurgically treated patients, 16% had spine
surgery between 3 and 60 months. Their outcomes were
similar to other nonsurgically treated patients and infe-
rior to surgically treated patients. However, it is unclear
whether initial surgical treatment would have changed
these results.
Our study has several strengths. It is a prospective
study providing long-term follow-up of contemporary
comparison groups with sciatica. We collected detailed
data, using validated measures of symptoms and func-
tion. Our cohort was assembled from community-based
clinical practice, making it likely that our results are
broadly generalizable.
The study’s major limitation is its observational, non-
randomized design. Because treatment was not assigned
in a random fashion, we cannot be certain that differ-
ences in outcomes between treatment groups were exclu-
sively because of surgery. Patients treated surgically were
less likely to be receiving Workers’ Compensation and to
have had a past spine episode, and more likely to have
abnormal physical and imaging findings, severe pain,
and greater back-related disability than patients treated
nonsurgically. However, surgical treatment remained a
significant predictor of improved symptoms and satisfac-
tion even after controlling for baseline differences. An-
other potential limitation is that the diagnosis was based on
physician assessment and did not require definitive imaging
or electrophysiologic validation. However, our goal was to
focus on outcomes in usual clinical care and not to alter
physician practice and decision making. Finally, although
follow-up rates were high (80%), our results may overstate
outcomes of both surgery and conservative care to a similar
extent because dropouts appeared to have worse outcomes
than those remaining in the study.
Our results do not imply that surgery is the preferred
treatment for all patients with sciatica caused by a her-
niated lumbar disc. Patients with mild symptoms did
well regardless of treatment. For those with moderate or
severe symptoms, surgery may hasten recovery and result
in better outcomes compared with nonsurgical treat-
ment. However, optimal surgical outcomes can take
3–12 months to achieve, up to 20% may undergo addi-
tional surgery, and work and disability outcomes are
similar to those treated nonsurgically. Patients reluctant
to undergo surgery may elect conservative care, knowing
that their symptoms will likely improve, albeit more
slowly, and they may remain more symptomatic than
those surgically treated.
Patients need to better understand the risks and ben-
efits of alternative treatments for a lumbar disc hernia-
tion. Wide variation in rates of surgery across geographic
regions suggests different thresholds for recommending
an operation.5,11,24,25 We have previously demonstrated
worse outcomes in higher surgical rate areas in Maine
compared with the lowest rate area.11 As new data be-
come available regarding outcomes and predictors, un-
biased educational programs may assist patients and
their physicians to make the most appropriate treatment
decision by providing them with more detailed and indi-
vidualized information.6,23
Key Points
● Outcomes of 402 patients with sciatica caused
by a lumbar disc herniation treated surgically or
nonsurgically were evaluated over a 5-year period.
● Surgically treated patients had more severe
symptoms and worse functional status at baseline
and better outcomes at 5-year follow-up compared
with nonsurgically treated patients.
● Surgical treatment remained a significant deter-
minant of 5-year symptom improvement and satis-
faction even after adjusting for other independent
predictors.
● The relative benefit of surgery declined over time
but remained superior to nonsurgical treatment.
● Work and disability outcomes as well as out-
comes of those least symptomatic at baseline were
similar among those treated surgically or
nonsurgically.
References
1. Andersson GBJ, Brown MD, Dvorak J, et al. Consensus summary on the
diagnosis and treatment of lumbar disc herniation. Spine 1996;24(suppl):
75– 8.
2. Atlas SJ, Chang Y, Kammann E, Keller RB, Deyo RA, Singer DE. Long-
term disability and return to work among patients who have a herniated
lumbar disc: the effect of disability compensation. J Bone Joint Surg Am
2000;82:4 –15.
3. Atlas SJ, Deyo RA, Patrick DL, Convery K, Keller RB, Singer DE. The
Quebec Task Force Classification for spinal disorders and the severity, treat-
ment, and outcomes of sciatica and lumbar spinal stenosis. Spine 1996;21:
2885–92.
4. Atlas SJ, Deyo RA, Keller RB, et al. The Maine Lumbar Spine Study: II.
1-year outcomes of surgical and nonsurgical management of lumbar spinal
stenosis. Spine 1996;21:1787–95.
5. Cherkin DC, Deyo RA, Loeser JD, Bush T, Waddell G. An international
comparison of back surgery rates. Spine 1994;19:1201– 6.
6. Deyo RA, Cherkin DC, Weinstein J, et al. Involving patients in clinical
decisions: impact of an interactive video program on outcomes and use of
back surgery. Presented at the annual scientific meeting of the International
Society for the Study of the Lumbar Spine, Kona, Hawaii, June 1999.
7. Frymoyer JW. Back pain and sciatica. N Engl J Med 1988;318:291–300.
8. Gogan WJ, Fraser RD. Chymopapain: a 10-year, double-blind study. Spine
1992;17:388 –94.
9. Hadler NM. Back pain in the workplace: what you lift or how you lift
matters far less than whether you lift or when. Spine 1997;22:935– 40.
10. Hoffman RM, Wheeler KJ, Deyo RA. Surgery for herniated lumbar disc: a
review. J Gen Intern Med 1993;8:487–96.
11. Keller RB, Atlas SJ, Soule DN, et al. Relationship between rates and out-
comes of operative treatment for lumbar disc herniation and spinal stenosis.
J Bone Joint Surg Am 1999;81:752– 62.
12. Keller RB, Atlas SJ, Singer DE, et al. The Maine Lumbar Spine Study: I.
Background and Concepts. Spine 1996;21:1769 –76.
13. Lewis PJ, Weir BKA, Broad RW, Grace MG. Long-term prospective study of
lumbosacral discectomy. J Neurosurg 1987;67:49 –53.

14. Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS System for Mixed
Models. Cary, NC: SAS Institute Inc., 1996.
15. Magora A. Investigation of the relation between low back pain and occupa-
tion: V. Psychological aspects. Scand J Rehabil Med 1973;5:191– 6.
16. Malter AD, McNeney B, Loeser JD, Deyo RA. Five-year reoperation rates
after different types of lumbar spine surgery. Spine 1998;23:814 –20.
17. Nykvist F, Hurme M, Alaranta H, Einola S. A prospective 5-year follow-up
study of 276 patients hospitalized because of suspected lumbar disc hernia-
tion. Int Disabil Stud 1989;11:61–7.
18. Papageorgiou AC, Macfarlane GJ, Thomas E, Croft PR, Jayson MI, Silman
AJ. Psychosocial factors in the workplace: do they predict new episodes of
low back pain? Spine 1997;22:1137– 42.
19. Patrick DL, Deyo RA, Atlas SJ, Singer DE, Chapin A, Keller RB. Assessing
health-related quality of life in patients with sciatica. Spine 1995;20:1899 –
909.
20. Postacchini F. Management of herniation of the lumbar disc. J Bone Joint
Surg Br 1999;81:567–76.
21. Roland M, Morris R. A study of the natural history of back pain: 1. Devel-
opment of a reliable and sensitive measure of disability in low-back pain.
Spine 1983;8:141– 4.
22. Spitzer WO. Scientific approach to the assessment and management of ac-
tivity-related spinal disorders: a monograph for clinicians. Report of the
Quebec Task Force on Spinal Disorders. Spine 1987;12(suppl 7):1–59.
23. Spunt BS, Deyo RA, Taylor VM, Leek KM, Goldberg HI, Mulley AG. An
interactive videodisc program for low back pain patients. Health Educ Res
1996;11:535– 41.
Point of View
Jon D. Lurie, MD, MS
SPORT, The Spine Center
Dartmouth Hitchcock Medical Center
Lebanon, New Hampshire
The report by Atlas and colleagues on the 5-year out-
comes from the Maine Lumbar Spine Study adds signif-
icantly to our understanding of the outcomes of surgical
and nonsurgical care for patients with sciatica. They find
early symptom improvement with surgery at 1 year
(71% improved with surgery vs. 43% for nonsurgical
patients) and a persistent but smaller relative advantage
of surgery at 5 years due to continued improvement in
the nonsurgical group (70% improved with surgery vs.
56% for nonsurgical patients). These finding are remark-
ably similar to the results of Weber’s earlier randomized
trial— 65% good results with surgery versus 36% for
nonsurgical patients at 1 year, and 70% good results
with surgery versus 52% for nonsurgical patients at 4
years.3
The major limitations of the current study are well
delineated by the authors in their discussion. There were
substantial baseline differences between the two groups,
and though attempts were made to control for these dif-
ferences in the analysis, the potential for residual con-
founding remains. In addition, the lack of definitive im-
aging in all cases may have resulted in some patients
without clear disc herniations being included in the non-
surgical group. These difficulties are inherent in the ob-
servation design of the study and have been addressed in
the newly begun Spine Patient Outcomes Research Trial
Five-Year Outcomes of Sciatica • Atlas et al 1187
24. Taylor VM, Deyo RA, Cherkin DC, Kreuter W. Low back pain hospitaliza-
tion: recent United States trends and regional variations. Spine 1994;19:
1207–13.
25. Volinn E, Mayer J, Diehr P, Van Koevering D, Connell FA, Loeser JD. Small
area analysis of surgery for low back pain. Spine 1992;17:575– 81.
26. Ware JE, Sherbourne CD. The MOS 36-item short-form survey (SF-36): 1.
Conceptual framework and item selection. Med Care 1992;30:473– 83.
27. Weber H. Lumbar disc herniation: a controlled, prospective study with ten
years of observation. Spine 1983;8:131– 40.
28. White AH, von Rogov P, Zucherman J, Heiden D. Lumbar laminectomy for
herniated disc: a prospective controlled comparison with internal fixation
fusion. Spine 1987;12:305–7.
Address reprint requests to
Steven J. Atlas, MD
Medical Practices Evaluation Center
Massachusetts General Hospital
50 Staniford Street
Boston, MA 02114
E-mail: satlas@partners.org
(SPORT), a multicenter, randomized trial of surgery com-
pared to nonsurgical treatment for patients with lumbar
intervertebral disc herniations, spinal stenosis, or degener-
ative spondylolisthesis confirmed by MRI or CT.1
The prospect of data from a large randomized trial is
exciting, but our patients need our help now. While we
still cannot give them clear answers, we can help them
make reasonable choices. Data from both randomized
and observation studies suggest a symptomatic benefit at
least in the first 3 months to 5 years with surgery in
patients with persistent sciatica. The outcomes without
surgery are generally quite favorable however, and, as
Atlas and colleagues show, patients with more mild
symptoms have such a good prognosis it is difficult to
improve on with surgery. By unbiasedly presenting the
current best evidence regarding available options and
their expected outcomes,2 we can help our patients to
make more informed and hopefully wiser choices.
References
1. Spine Patient Outcomes Research Trial, NIH #U01-AR45444 – 01A1; JN
Weinstein, Principal Investigator.
2. Deyo RA, Cherkin DC, Weinstein J, et al. Involving patients in clinical deci-
sions: Impact of an interactive video program on use of back surgery. Med
Care 2000;38:959 – 69.
3. Weber H. Lumbar disc herniation. A controlled, prospective study with 10
years of observation. Spine 1983;8:131– 40.