Study Design
Enrollment in SPORT was conducted at thirteen multidisciplinary spine practices in eleven states across the United States. The study design and methods of SPORT have been previously published22.
Patients
The SPORT protocol was approved by the human subject committees at each participating center. Criteria for inclusion in the lumbar disc herniation cohort were age greater than eighteen years, radicular pain for at least six weeks with a positive nerve root tension sign and/or neurological deficit, and a confirmatory cross-sectional imaging study demonstrating intervertebral disc herniation at a level and side corresponding to their symptoms. Exclusion criteria included cauda equina syndrome, a progressive neurological deficit, malignancy, scoliosis of >15°, herniation cephalad to L2, prior back surgery, and other established contraindications to elective surgery.
Study Interventions
Patients were offered participation in either a randomized cohort or a concurrent observational cohort. Those in the randomized cohort received computer-generated random treatment assignments. Those in the observational cohort chose their treatment with their physician. Patients in the operative treatment group were designated to undergo lumbar discectomy. Patients in the nonoperative treatment group were managed with minimum nonsurgical treatment, which included active physical therapy, education, counseling with home exercise instruction, and nonsteroidal anti-inflammatory drugs if tolerated. Additionally, patients could receive any therapies above the minimum required protocol that were deemed appropriate by their physician. Because of the extensive crossover in the randomized cohort in which some patients who had been randomized to the nonoperative treatment group instead received operative treatment and vice versa, and because of the similar baseline characteristics and outcomes when comparing the randomized and observational cohorts when analyzed by treatment, the two groups were combined in this as-treated analysis.
Study Measures
Data used in this study were obtained from patient questionnaires completed at baseline, six weeks, three months, six months, one year, two years, three years, and four years after enrollment or surgery. The primary outcome measures included the physical function and bodily pain domains of the Short Form-36 (SF-36)23 and the American Academy of Orthopaedic Surgeons MODEMS (Musculoskeletal Outcomes Data Evaluation and Management System) version of the Oswestry Disability Index24. The secondary outcome measures included patient self-reported improvement, work status, and satisfaction with current symptoms and with treatment25. Symptom severity was measured by means of the Low Back Pain Bothersomeness Index and the Sciatica Bothersomeness Index26-28. The SF-36 scales range from 1 to 100 points, with higher scores indicating less severe symptoms. The Oswestry Disability Index also ranges from 0 to 100 points, but higher scores indicate more severe symptoms. The Sciatica Bothersomeness Index ranges from 0 to 24 points, and the Low Back Pain Bothersomeness Index ranges from 0 to 6 points; in these indices, higher scores indicate more severe symptoms.
Statistical Analysis
The primary analyses compared changes in the clinical outcome measures from baseline as a function of the patient BMI at the time of enrollment in the study. BMI was calculated as the weight in kilograms divided by the height in meters squared according to National Institutes of Health (NIH) criteria29. For study purposes, the patients were divided into two groups on the basis of BMI: nonobese (BMI, <30 kg/m2) and obese (BMI, ≥30 kg/m2). The baseline characteristics of nonobese and obese patients were compared with use of a chi-square test for categorical variables and a t test for continuous variables. The primary analyses compared operative and nonoperative treatments with use of changes from baseline at each follow-up evaluation, with a mixed effects longitudinal regression model, including a random individual effect to account for the correlation between repeated measurements within individuals. A formal interaction term of the treatment and obesity was in the model for comparing the treatment effect in the subgroups. Because of the crossover, analyses were based on treatments actually received in the combined randomized and observational cohorts. In these as-treated analyses, the treatment indicator was a time-varying covariate, allowing for variable times of surgery. Times were measured from the beginning of treatment, which was the time of surgery for the operative treatment group and the time of enrollment for the nonoperative treatment group. Therefore, all changes from baseline prior to surgery were included in the estimates of the nonoperative treatment effect. After surgery, changes were assigned to the operative treatment group with follow-up evaluations measured from the date of surgery. Repeated measures of outcomes were used as the dependent variables, and treatment received was included as a time-varying covariate.
To adjust for potential confounding, baseline variables associated with missing data or treatment received were included as adjusting covariates in longitudinal regression models (marital status, smoking status, compensation, herniation location, working status, depression, other comorbidities, self-rated health trend, duration of the most recent episode, or treatment preference). In addition, data for race, center, age, sex, and baseline outcome scores (for SF-36, the Oswestry Disability Index, the Sciatica Bothersomeness Index, and the Low Back Pain Bothersomeness Index) were included in the longitudinal regression models. Secondary and binary outcomes were analyzed with use of generalized estimating equations that assumed a compound symmetry working correlation structure. The outcomes were stratified by obesity, and outcomes between these subgroups at each time point were compared with use of a multiple-degrees-of-freedom Wald test. Across the four-year duration of follow-up, overall comparisons of area-under-curve between these subgroups were made with use of a Wald test.
Computations were done with use of SAS procedures PROC MIXED for continuous data and PROC GENMOD for binary and non-normal secondary outcomes (SAS version 9.1, Windows XP Pro; SAS Institute, Cary, North Carolina). Significance was defined as p < 0.05 on the basis of a two-sided hypothesis test with no adjustments made for multiple comparisons. Such adjustments were not made because the analyses of the current study are not multiple independent comparisons from which one is looking for a single significant result, but rather consist of the longitudinal assessment over time of different dimensions of outcomes, including symptoms, function, and disability.
Source of Funding
Funds were received from the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) (U01-AR45444), the Office of Research on Women’s Health, the National Institutes of Health, the National Institute of Occupational Safety and Health, and the Centers for Disease Control and Prevention. The Multidisciplinary Clinical Research Center in Musculoskeletal Diseases is funded by NIAMS (P60-AR048094). Funds were used for statistical analysis and manuscript publication.
The results of the current study demonstrated that obese patients are more likely to undergo surgical intervention for lumbar disc herniation than nonobese patients and that obesity leads to increased operative time, blood loss, and length of hospital stay. The link between obesity and increased surgical intervention for lumbar disc herniation has been previously reported5,30. In two separate cross-sectional studies assessing >3700 patients, Böstman demonstrated a significant association between increased BMI and operative treatment of lumbar disc herniation5,30. Although the explanation of this finding remains largely unknown, the finding from the current study that obese patients have a pretreatment preference for surgery may provide the answer.
The perioperative findings in this study are at odds with the medical literature. A prospective study by Andreshak et al. comparing perioperative findings between obese and nonobese patients undergoing lumbar spine surgery demonstrated no difference in operative time, blood loss, or length of hospital stay12. The explanation for this difference in findings may be that the present study only included patients with a diagnosis of lumbar disc herniation and the study by Andreshak et al.12 evaluated patients with a variety of diagnoses. This discrepancy warrants further investigation because of the known increased risk of infection associated with these perioperative factors31,32.
We believe that our results largely supported the hypothesis that obesity is associated with worse outcomes after operative or nonoperative treatment of lumbar disc herniation. At the four-year follow-up evaluation, both operatively and nonoperatively managed obese patients had significantly less improvement in the primary outcome measures of the study. The only exception was the statistically similar improvement from baseline in the SF-36 bodily pain score in obese and nonobese patients managed nonoperatively. Disease-specific secondary outcome measures demonstrated a worse outcome for operatively managed obese patients in the Sciatica Bothersomeness Index, the Low Back Pain Bothersomeness Index, and work status. The published data on this topic are limited and serve to both support and refute the findings of the current study. In a review of nearly 1100 cases of sciatica treated nonoperatively, multivariate analysis showed that obesity was associated with adverse six-month outcomes19. More recently, Cole and Jackson evaluated the use of minimally invasive techniques to treat lumbar disc herniation in obese patients and concluded that this minimally invasive approach is the preferred technique to manage these patients because of favorable results and a trend toward reduced infectious complications33.
In the current study, recurrence of disc herniation and need for additional surgical procedures did not differ significantly between obese and nonobese patients. This finding contradicts other published literature21. In 2010, Meredith et al.21 retrospectively reviewed seventy-five patients who had undergone a microdiscectomy, eight (10.7%) of whom had experienced a recurrent herniation. Meredith et al. associated obesity with a twelve times higher likelihood of postoperative recurrent herniation and a thirty times higher likelihood of requiring a reoperation and concluded that weight loss counseling should be incorporated in preoperative discussions21. Because of the small sample size of that previous study, the findings of our study are likely more representative of the general population.
The current study advances the literature by identifying the expected results of operative and nonoperative intervention in obese patients with the specific diagnosis of symptomatic lumbar disc herniation. To our knowledge, this study has the largest series in the literature specific to lumbar disc herniation. This study also has the benefit of a prospective design, in which large samples of geographically diverse patients received treatment at multiple medical centers and were followed for several years. These attributes served to limit bias, increase the precision of our findings, and provide favorable generalizability of our results to clinical practice.
There were important methodological limitations to consider with this as-treated subgroup analysis of data. The initial SPORT study had been powered to compare operative treatment with nonoperative treatment for lumbar disc herniation and had not been powered for specific subgroup analyses. Even though the initial SPORT study22 had not been powered to detect differences in clinical outcome based on BMI, the current subgroup analysis did detect such differences with significance. Confounding baseline variables in the current study, including sex, level of education, income, the presence of a compensation claim, and medical comorbidities, differed between the obese patients and the nonobese patients. These differences may have affected the results of the present study. Other limitations included a lack of reassessment of BMI at specific postoperative time points. In other studies, obese patients had gained an average of 1.5 kg by the twenty-month average follow-up period34. This reassessment of BMI at specific postoperative time points would have helped to identify the effect of surgery on obesity and whether pain from a spine etiology limits the ability of obese patients to lose weight. Finally, the imprecision of BMI as a marker of obesity was another important limitation. Other measures, such as skin-fold thickness or body surface area, may be more specific measures of obesity to include in future studies.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. In addition, one or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.