The SPORT Study: Pro
The SPORT study is the largest-ever lumbar spine investigation funded by the U.S. National Institutes of Health (NIH). Its goals were to compare surgical and nonsurgical treatments for patients with lumbar disc herniation, spinal stenosis, or degenerative lumbar spondylolisthesis with spinal stenosis. These three diagnoses represent the vast majority of surgical disorders involving the lumbar spine in the adult population. An innovative study design consisted of concurrent enrollment of the three diagnostic groups into a surgical trial with randomized and observational cohorts with identical enrollment criteria1. In addition, informed choice through the use of pre-enrollment patient decision aids was utilized. All data for this study were gathered electronically and housed at the site of the primary investigator, Dr. James Weinstein, at Dartmouth-Hitchcock Medical Center. The study population consisted of 5580 screened patients, 4246 of whom were eligible to participate. Following review of the shared decision-making video by the eligible patients, 2505 were enrolled. The randomized group consisted of 1094 patients, 542 of whom were assigned to surgical treatment and 552, to nonoperative treatment. The observational group consisted of 1411 patients, 913 of whom were treated surgically and 498 of whom received nonoperative care. This observational group represented the patients who did not wish to be randomized. The study population was gathered from thirteen participating sites representing the East and West Coasts, Midwest, South, and Central Plains. This geography was a representative cross-section of the U.S. population. Outcomes were assessed by the primary measures of the Short Form-36 (SF-36) bodily pain and physical function scales (on which a higher score is better) and the Oswestry Disability Index (on which a lower score is better). The secondary measures were sciatica bothersomeness, patient satisfaction, and work status. Data were gathered at six weeks, three months, six months, one year, two years, and four years (data gathering is to continue for at least another five years). Utility measures consisting of the Health Utilities Index, EQ-5D (derived from the EuroQol instrument), SF-6D (derived from the SF-36), and a visual analog pain score were also recorded. Finally, cost-effectiveness data for all three groups were collected.
Patient eligibility criteria were consistent with well-established guidelines, including predominant radiculopathy or neurogenic claudication with a confirmatory imaging study, usually a magnetic resonance imaging scan. Exclusion criteria consisted of previous surgery or a progressive neurologic deficit, an age of less than eighteen years, or a nondegenerative spinal disorder. Nonoperative treatment consisted of "usual care." No specific nonoperative protocols were used. Usual care was selected because it reflects the real world situation. Surgical management consisted of open discectomy for a disc herniation, decompressive laminectomy for spinal stenosis, and decompressive laminectomy with intertransverse process fusion with use of autogenous bone graft with or without pedicle screws for degenerative spondylolisthesis.
The challenges of a prospective, randomized clinical trial in surgery are many. Recruiting a sufficient number of subjects to be statistically valid is critical. The enrollment for SPORT has far exceeded all other prospective studies on this topic. A high follow-up percentage is essential to validate the results. SPORT has maintained a 74% rate of follow-up at four years. The placebo effect of surgery has been well documented. Removing the placebo effect from patients’ assessment of their procedure cannot be done without a so-called sham surgical trial. Despite the inherent validity of a sham, the substantial ethical challenges, institutional review board approvals, blinding, and safety concerns are difficult to overcome. Other stated concerns of the SPORT trial include the lack of standardized nonoperative treatment. Patients received the customary nonoperative modalities such as anti-inflammatory medicine, physical therapy, and epidural injections. This "usual care" reflects the standard practice for nonoperative management of these disorders.
The results of the randomized portion of the SPORT trial for intervertebral disc herniation, spinal stenosis, and degenerative lumbar spondylolisthesis are presented in two different groups: the intent-to-treat group and the as-treated group. Intent-to-treat analysis assigns subjects to their original treatment group so that if they crossed over from nonoperative to operative treatment or vice versa, their results would be analyzed by their original grouping. As-treated analysis determines outcomes on the basis of the actual treatment received. Only the intervertebral disc herniation surgical group had more severe symptoms and disability than the nonoperative cohort at baseline. Yet, a high percentage of crossovers occurred for all three diagnoses both from surgery to nonoperative treatment and from nonoperative care to surgery. The high percentage of crossovers compromised the intent-to-treat analysis and, as planned for in the initial protocol, an as-treated analysis was performed for each cohort. In SPORT, it would have been impractical and unethical to force patients to accept a treatment if, in the course of the trial, they chose to change their original assignment. The as-treated analysis likely provides a more realistic representation of the results of operative compared with nonoperative care for these three diagnostic groups. In general, one might suggest that the intent-to-treat analysis underestimates the true treatment effect and the as-treated analysis overestimates the true treatment effect. The results of the SPORT trial likely are somewhere in between. It is noteworthy that the as-treated analyses were nearly identical to the as-treated observational cohort analyses.
On the basis of exact protocols for both study arms (randomized clinical trial and observational), the large volume of enrolled patients, multiple sites, multiple physicians, and standardized outcome measures, one can be assured of the generalizability of the SPORT results. Figure 1 demonstrates the primary outcome results for intervertebral disc herniation at four years for the randomized and observational cohorts. The randomized intent-to-treat results demonstrated no significant difference between surgery and nonoperative treatment at four years. The as-treated analysis, however, showed significant positive effects for surgery compared with nonoperative treatment at four years2-4. Figure 2 shows the results of the secondary outcome measures for the intervertebral disc herniation group, comparing the randomized and observational as-treated groups for surgery compared with nonoperative treatment at four years2-4. For degenerative lumbar spondylolisthesis, on the basis of the as-treated comparisons, patients treated surgically for degenerative spondylolisthesis with spinal stenosis showed substantially greater improvement with respect to pain and function than nonsurgically treated patients at four years5,6. The results for patients with lumbar degenerative spinal stenosis demonstrated that, in the combined as-treated analysis, surgery showed significantly more improvement in all primary outcomes compared with nonsurgical treatment7.
The results demonstrating the superiority of surgery at two years of follow-up for patients with a lumbar disc herniation or degenerative spondylolisthesis were maintained at the four-year follow-up period4,6, and the results for the four-year follow-up for lumbar stenosis were under review for publication. The debate for which this article was written was titled, "The SPORT Study: Did NIH Get Its $15 Million Worth?" The answer is unequivocally yes. Much has been learned from this huge undertaking, with much of the credit going to the principal investigator, Dr. James Weinstein, Director, The Dartmouth Institute for Health Policy and Clinical Practice and Professor and Chairman of Orthopaedic Surgery, and his colleagues who designed SPORT. Surgical randomized trials are challenging from several perspectives. First, blinding surgical treatment creates serious ethical concerns. Second, standardizing treatment across multiple centers is a difficult task. Third, minimizing observer bias is very difficult. This is why 90% of the orthopaedic literature represents research from nonrandomized studies8. There have been several written and verbalized concerns about SPORT. One is that usual nonoperative treatment was used rather than standardized best practice. There was no difference in therapy approach or intensity at the thirteen participating sites. Nonoperative therapy was noted to be effective in the SPORT trial. Also, there is no conclusive effective nonoperative treatment in the literature that would represent best practice. The second concern is how to account for patient expectations of surgery compared with the results of the intervention itself. An analysis by Lurie et al. regarding the interaction among expectation, preferences, and outcomes was under review at the time of writing9.
Sham surgery has been used to control this variable in both orthopaedic and nonorthopaedic surgical studies10,11. However, sham surgery, although considered, was not a viable option for this trial. The third concern raised about SPORT was the high rate of crossover among the three groups from surgery to nonoperative treatment and vice versa. This confounded statistical analysis and is why the trial results have been reported as intent-to-treat, as-treated, and combined.
In SPORT, symptom severity for patients with disc herniation influenced patient treatment choice and those choosing or crossing over to surgery had worse symptoms at baseline. This was not the case for those with stenosis. Purists argue that unless findings are based on the intent-to-treat analysis, the study loses credibility and the conclusions remain suspect. Yet, in SPORT, the as-treated results were nearly identical for both cohorts (randomized clinical trial and observational) and, given the design of SPORT and its follow-up of patients, the results appear to be the best we may ever achieve. SPORT enrolled over 2500 patients from thirteen centers in eleven states and dealt with "real world" issues. We cannot force a specific treatment for elective conditions, when patient preferences or expectations change with symptom severity and family needs. SPORT demonstrated the superiority of surgery in the conditions studied but also demonstrated nonoperative treatment to be effective or at least an option that patients may wish to choose despite their level of symptoms. Finally, the cases in the SPORT study represent very well-defined examples of these diagnoses, with history, physical examination, and confirmatory images being carefully defined, and we must not extrapolate the results of SPORT to less well-defined cases.
The pundits argue that SPORT, for all its glamour, cost, and press, has taught us nothing new. Nothing could be further from the truth. SPORT enrolled 2500 patients and maintained a follow-up rate of >74% at four years, with funding by the National Institutes of Health for five more years of continued follow-up. SPORT for the first time established the cost-effectiveness of treatments for these three conditions. These enrollment numbers, follow-up percentages, and the longitudinal analysis are powerful in themselves. The generalizability of SPORT will be hard to debate, and patients will be the largest benefactors for years to come. Physicians will have the best data available to offer patients a truly informed choice12. With the health-care debate raging, SPORT has come at just the right time and it is an exemplar of the recently published Institute of Medicine's report on Initial National Priorities for Comparative Effectiveness Research13. SPORT has demonstrated the effectiveness of both surgery and nonoperative treatment. It has shown that surgery is safe with a low complication and recurrence rate. It has demonstrated that there is no urgency to perform surgery as the results are effective early in the symptom cycle as well as later on. SPORT has demonstrated that surgery is cost-effective and improves the quality of life14,15. It has confirmed that patients with the classic diagnostic criteria for disc herniation and stenosis with milder symptoms improve with surgery, but patients with severe symptoms improve more. The power of the SPORT data and the conclusions drawn from SPORT can be used to educate patients to make better-informed decisions about their treatment12,16.
SPORT is the most comprehensive evidence-based study undertaken for the conditions of lumbar disc herniation, spinal stenosis, and degenerative lumbar spondylolisthesis. The data gathered from SPORT will generate valuable long-term treatment outcome studies. The SPORT studies are the "gold standard" against which all future evidence-based spine research will be judged.
The SPORT Study: Con
There are multiple considerations that help consumers to determine whether they "got their money's worth." These include the demand for the item, the availability of the item, and, perhaps most importantly, the constituency or viewpoint of the consumer. The suppliers (in this case the investigators) clearly see their product as something everyone needs. To them, it is undervalued and they are providing it at a bargain rate, and it should be updated frequently. The consumers come from various perspectives: the surgeon who welcomes stronger information validating the surgical procedures being done; the patient who wants to know what treatment is best; the taxpayers, employers, and insurance companies who fund medical care and want to know the balance of cost, benefit, risk, and effectiveness; and the NIH, which is ultimately accountable to the taxpayer to fulfill its mission. In order to determine whether NIH got its money's worth, one must understand the goals of NIH and evaluate critically whether these goals were met.
According to the NIH web site, the goals of the agency are as follows:
- foster fundamental creative discoveries, innovative research strategies, and their applications as a basis to advance significantly the Nation's capacity to protect and improve health;
- develop, maintain, and renew scientific human and physical resources that will ensure the Nation's capability to prevent disease;
- expand the knowledge base in medical and associated sciences in order to enhance the Nation's economic well-being and ensure a continued high return on the public investment in research; and
- exemplify and promote the highest level of scientific integrity, public accountability, and social responsibility in the conduct of science17.
Critical evaluation of the SPORT trial regarding each of these goals will answer our question. With respect to the first goal, the SPORT trial did not foster or discover a fundamental creative discovery. The trial did apply an innovative research strategy in surgical trials in the combining of randomized and observational cohorts to follow prospectively surgical and nonsurgical patients. This balances the ethics and logistics to gain the most information from a surgical trial that is not randomized and double-blinded. Will this application of an innovative strategy provide the basis to advance substantially the nation's capacity to protect and improve health? As this was a treatment trial, the protection of the nation's health will not be changed. Whether the trial provides a basis to improve health appears to depend on the specific measure evaluated, thus not clearly and substantially meeting this goal. Only one of the objectives of the first goal was met and only partially so. Grade D.
The second goal of NIH's mission is in regard to the development, maintenance, and renewal of human and physical resources that will ensure the nation's capability to prevent disease. As a treatment trial, SPORT failed to meet this goal, which is focused on prevention. SPORT did help to develop, maintain, and renew our knowledge base regarding surgical management of lumbar spinal disorders involving leg pain. The trial also was used in the training and career development of medical research personnel, including epidemiologists, physicians, and surgeons. Grade D.
The third goal of NIH is to "expand the knowledge base in medical and associated sciences in order to enhance the Nation's economic well-being and ensure a continued high return on the public investment in research." While the knowledge base related to surgical research and epidemiological research has been improved, two key questions remain. First, was the improvement in the knowledge base substantial and clinically useful? Most agree with the authors that the findings of SPORT essentially confirm those of prior studies that have shown an early benefit for surgical management that decreases in magnitude over time. The minimum clinically important difference has been reported for some measures related to spine surgery, including the SF-36, the Oswestry Disability Index, and the leg pain index18. In the SPORT trial, the primary outcome measure of the Oswestry Disability Index had a treatment effect of 3.6 at four years for lumbar disc herniation and 12.6 for degenerative spondylolisthesis, both below the minimum clinically important difference of 12.8. The treatment of leg pain is considered by most to be a key factor in surgical decision-making. In six papers by the investigators of the SPORT trial that had been published through June 2009, leg pain was a secondary measure of clinical success2-7. The leg pain index is thought to have a minimum clinically important difference of 1.6; however, depending on how it was determined, it could be as high as 2.85. As such, the only times when any of the diagnostic groups experienced a clinically important difference was at the three-month follow-up. The sciatic bothersomeness index was believed to have a minimum clinically important difference of 3 points19, resulting again in the three-month time point and perhaps the four-year time point being the only clinically important differences and those just detectable. The fact that there was no difference in return to work emphasizes that there was no direct impact on the nation's economic well-being as a result of the SPORT trial. Whether a high return on the public's investment in research is gained is a complex question but fundamentally should be reflected in clinically important and significant results that change the way medicine is practiced. Grade D.
The fourth goal of NIH is to "exemplify and promote the highest level of scientific integrity, public accountability, and social responsibility in the conduct of science." This the investigators of SPORT did very well. The trial sets a new standard for conducting surgical trials in seeking a rigorous randomized trial while accommodating the highest standard of ethics possible. The reporting of primary results and the depth of secondary information and articles published is a clear demonstration of responsible use and dissemination of the data acquired through the trial. Grade A.
On the basis of these stated goals of NIH, the overall grade for the SPORT trial is a D+. For perspective, however, it is appropriate to compare some other measures of value for $15,000,000. The SPORT trial, through May 2009, listed twenty-seven articles on its web site, indicating an expense of $555,555 per article. Alternative uses of $15,000,000 may have provided for the surgical care of 1100 patients having decompression for herniated discs. This amount may have also employed thirty orthopaedic surgeons for a year (salary, benefits, and expenses).
Summary
The SPORT study is the most comprehensive clinical study ever performed on lumbar spinal conditions. Despite its obvious limitations (high crossover rate and nonstandardized nonoperative treatment), it nevertheless provides useful information for the care of patients with common lumbar spine problems. It also provides insight into the value of prospectively obtained observational data.
The attendees at this symposium at the 2009 Annual Meeting of the AOA had a generally favorable view of SPORT. A majority were convinced by the data that the surgery was effective for lumbar disc herniation (57% indicated yes; 37%, maybe; and 7%, no), lumbar spinal stenosis (66% indicated yes; 29%, maybe; and 5%, no), and degenerative spondylolisthesis (58% indicated yes; 30%, maybe; and 12%, no). When asked whether the NIH did receive its "$15 million worth," 62% of those attending this debate agreed, 17% disagreed, and 21% were uncertain. Despite the demonstrated generalizability of the results between the randomized and observational groups in SPORT, 56% of the attendees were not convinced that observational studies were "just as meaningful" as randomized trials. Almost all present at the session (95%) agreed with the suggestion that funding agencies should continue to support clinical studies like SPORT.
Back Pain Surgery: Pro
Low back pain is a very common symptom that affects a great percentage of the population and is one of the most common reasons for visits to a health-care provider. Although there are many reasons for low back pain, the natural history is such that most episodes of low back pain are self-limiting and the natural history is favorable without the need for surgical intervention. Unfortunately, many patients have persistent low back pain, and the treatment for it, whether it be with conservative management or operative treatment, is controversial and highly debated.
Certainly, not every patient with a symptom of low back pain is a candidate for surgery. However, after a reasonable period of conservative treatment that has failed to resolve the symptoms, a carefully selected group of patients with certain diagnoses who experience the symptoms of low back pain may be amenable to surgery. Determining whether a patient is a surgical candidate relies on many factors, including defining the diagnosis, clearly defining the source of the pain (otherwise known as the pain generator), and having a reasonable surgical treatment to appropriately address the surgical pathology.
One of the first steps toward determining the appropriateness of operative treatment is to define the symptoms and corresponding pathology in order to make the diagnosis. Symptoms of low back pain may be reported as axial pain, buttock pain, or pain radiating down the proximal legs. Low back pain may be from degeneration of the disc, facets, or other supporting structures of the spine. But, it can also be referred pain, or pain resulting from a muscle strain, or symptoms of a more systemic condition such as arthritis. The primary clinical point is that grouping patients with the diagnosis, or with symptoms of low back pain, mixes patients with multiple potential pain generators. In a review of the literature, many studies examining the results of spine surgery had a heterogeneous group of patients with different primary diagnoses who may also have had other associated symptoms in addition to low back pain. While at face value, the results may show no benefits, when patients with similar diagnoses are compared within studies, the results may be more favorable for some pathologies and less favorable for others20-22.
Patients with scoliosis, spondylolisthesis, herniated discs, instability, adjacent-segment degeneration, disc degeneration, prior surgery, or a pseudarthrosis may all have symptoms of low back pain, but they may respond differently to surgical intervention21. It is critical to look at the patient population and compare the results of surgery and nonoperative treatments in studies in the literature in which the patients had the same diagnosis and reasons for the low back pain in order to appropriately examine the results of surgical treatment. Many of the reviews of the literature in the past have examined the studies treating patients with spondylosis with or without spondylolisthesis, stenosis, disc herniations, or prior surgery and concluded that there is little evidence to support the surgical treatment of these patients23,24. In the more recent literature, studies have segregated patients with similar diagnoses in order to determine the results of operative and nonoperative treatments20-22.
An examination of the evidence in support of surgical interventions for patients with low back pain shows that patients do respond with reasonable results after surgical treatment. In order to support the viewpoint that surgical treatment is indicated for patients with certain diagnoses and symptoms of low back pain, one must clearly define the diagnosis and define the source of the pain prior to recommending surgery.
When examining the evidence for the results of surgical treatment for patients with a specific diagnosis, one finds that certain conditions respond more favorably than others to operative treatment. Glassman et al. evaluated 428 patients having lumbar fusion for degenerative conditions with prospectively collected data21. They found that the health-related quality-of-life scores (HRQOL) at the time of the two-year follow-up showed the most substantial improvement with surgery for patients who had spondylolisthesis (an average 22.7-point improvement) and those with scoliosis (an average 21.2-point improvement). Patients had lesser improvements with other conditions, such as disc degeneration, postdiscectomy revision surgery, instability, stenosis, and adjacent-level degeneration, in that order.
This diagnosis-related outcome is further proven by studies published from the SPORT data. Weinstein et al. compared surgical and nonsurgical treatment for patients with degenerative spondylolisthesis5,6. They found that patients who had surgical treatment had substantially greater pain relief and improvement in function at four years compared with the patients treated nonoperatively. Patients with spondylolisthesis can have an element of instability, which can also lead to symptoms of low back pain. The literature clearly shows some support for the treatment of patients with spondylolisthesis and low back pain.
A lumbar herniated disc is actually a quite common diagnosis and can typically result in radiculopathy as well as symptoms of low back pain. Although the typical patient with a herniated disc often has radicular pain in a dermatomal distribution as the primary symptom, many of these patients have an element of low back pain. There is good evidence that patients with herniated discs have a favorable outcome after operative interventions compared with that after nonoperative treatments specifically for low back pain3,25.
Patients with spinal deformities, such as kyphosis or scoliosis in the thoracolumbar area, certainly may be prone to degeneration later in life. These degenerative changes may result in arthritis as well as some potential increased coronal and sagittal plane deformities and segmental instability. Although these patients may present with a myriad of symptoms, low back pain or pain in the thoracolumbar area can certainly be a predominant part of the overall symptomatology. The literature has shown that carefully selected patients with carefully selected procedures can have excellent outcomes21,26-29. This diagnosis is quite complicated because of the fact that there are typically multiple spinal levels involved, and there is still some debate regarding the appropriate selection of levels for spinal fusion. Often the decision is made on the basis of the sagittal and coronal plane deformities as well as the amount of degeneration at the adjacent levels both radiographically and even with provocative discography. In this population, the literature clearly has shown that there are some substantial benefits for properly selected patients who undergo surgical intervention21,28.
Degenerative disc disease at one or multiple levels is probably the most common and most controversial area regarding surgical intervention for the treatment of low back pain. Clearly this is a multifactorial diagnosis, and we do not have completely reliable tests to determine the pain generator. Discography has been used but has been questioned as a useful diagnostic tool to help to judge whether the disc is the pain generator. In addition, age-appropriate radiographic changes are certainly expected in certain groups. It is thought that the critical element for success with surgical intervention is identifying the diagnosis and the pain generator and then choosing the appropriate diagnostic tests to determine whether there is a relationship between the radiographic appearance of the discs and the symptoms of low back pain.
Gibson et al., in the Cochrane reviews, examined randomized controlled trials looking at all surgical treatments for patients with lumbar spondylosis23,24. They discussed the various studies and determined that study quality was variable, many were not appropriately randomized, and there was a lack of long-term follow-up. Unfortunately, the diagnoses in these studies included patients with degeneration, spondylolisthesis, and stenosis. They also looked at different forms of treatment including intradiscal therapy. Definite conclusions could not be formulated, possibly because of the heterogeneity of the patients and the varied surgical treatments.
Resnick et al., in a literature review of the cases of patients with degenerative disc disease, carefully graded the quality of the evidence30. When confining the patients to those with degenerative disc disease without stenosis or spondylolisthesis, they found level-I evidence to support the use of lumbar fusion for properly selected patients who had failed conservative treatment. In addition, they found only level-III evidence for cognitive and physical therapy for the treatment of patients with degenerative disc disease.
Carreon et al. also examined surgical and nonsurgical treatments for lumbar spondylosis20. They found evidence to support the role of fusion for patients with degenerative disc disease and low back pain. They also found that diagnosis was related to outcome, with spondylolisthesis and degenerative disc disease having a more favorable prognosis than chronic low back pain. Glassman et al. examined a group of 497 patients in a multicenter study of lumbar fusion for one and two-level degenerative disc disease22. They found improvement with fusion, supporting the use of surgery in these patients.
Surgery for degenerative disc disease with low back pain has been investigated in prospective randomized studies, demonstrating the efficacy of surgical treatments. One U.S. Food and Drug Administration study comparing lumbar fusion and lumbar arthroplasty with up to five years of follow-up showed that both the arthroplasty devices and the fusion controls have around 72% to 78% rates of patient satisfaction31. Another prospective study demonstrated that 84% to 91% of the patients had an improvement in low-back-pain disability scores after a lumbar arthroplasty or a lumbar fusion32.
Although low back pain is a generalized symptom that can be a part of the symptomatology for patients with multiple diagnoses and different abnormalities in the lumbar spine, the evidence in the present study shows that in the properly selected patients in whom the diagnosis is clear and the pain generator has been appropriately identified, surgical intervention is a reasonable option. The studies have clearly shown that, in carefully and properly selected patients, lumbar fusion may be a very useful surgical procedure to help with the symptoms of low back pain. This focuses the onus for the surgery on proper patient selection, which is an important part of the success of any surgical treatment.
Back Pain Surgery: Con
Low back pain is a ubiquitous malady affecting people of all ages, ethnic groups, and socioeconomic classes. The lifetime prevalence of low back pain in adults approaches 85%33. The incidence of low back pain in the United States is 15% to 20% per year and is the most common symptom necessitating an evaluation by a physician34. The cost of treating low back pain is enormous and is increasing at a staggering pace in the United States. In 2005, the United States spent $85.9 billion treating back pain, which is a 65% increase from the $52.1 billion spent in 199735.
In this "modern era" of medicine, with our arsenal of sophisticated surgical procedures, patients often believe that surgery can deliver instant relief of low back pain. Unfortunately, for the millions of Americans battling back pain, they may not be any closer to getting that quick relief. Even with the latest innovative procedures, indications for surgical treatment and successful outcomes for mechanical low back pain remain controversial and elusive. Scientific evidence continues to support the use of a nonsurgical, conservative approach to treating patients with discogenic low back pain36.
Understanding the natural history of low back pain can help to direct appropriate management and provide patient education and reassurance. Patients can present with a variety of symptoms ranging from mild discomfort to debilitating pain with lower-extremity symptoms. The intervertebral disc, facet joint, and intra-abdominal pathological conditions are all potential pain generators in axial back pain and make accurate diagnosis very challenging. Fortunately, most patients recover quickly and spontaneously. Greater than 90% of patients with acute low back pain have resolution of their symptoms within three months, and most have enough symptom relief within six weeks to return to work37. This natural history occurs regardless of whether patients seek treatment and supports conservative treatment as the initial intervention.
Recurrence is part of the natural history of low back pain, and some patients continue to have mild back pain or discomfort for greater than three months after seeking treatment38. Approximately 5% to 10% of patients develop a chronic form, in which disabling pain continues despite the passage of time, activity modification, and extensive conservative treatment22. Fortunately, chronic pain can have little impact on the well-being of a patient if he or she can function normally. Therefore, in patients with chronic pain, treatment should focus on functional capacity with return to work and usual activities as the primary goal and relief of pain as a secondary goal39,40. Operative treatment of low back pain is considered only when nonoperative measures have been exhausted and both pain and disability continue to be present.
The early pioneering work of Cloward demonstrated that lumbar fusion can provide relief from discogenic pain, and it paved the way for operative treatment of mechanical low back pain41. Over the years, multiple fusion techniques were employed to address low back pain. Posterolateral arthrodesis, posterior lumbar interbody arthrodesis, anterior lumbar interbody arthrodesis, and circumferential arthrodesis techniques were described for the treatment of degenerative disc disease34,42-45. Innovative instrumentation devices, such as pedicle screws and interbody fusion cages as well as biologic growth factors, were developed to enhance these fusion techniques46,47. Unfortunately, most of the surgical studies have been retrospective and good outcomes have been inconsistent48.
Until recently, there were no published randomized controlled trials on the effectiveness of fusion for chronic low back pain compared with conservative treatment. In 2001, the Swedish Lumbar Spine Study Group reported on their trial of lumbar fusion compared with physiotherapy for the treatment of chronic low back pain49. The results of 294 patients who had been randomized into surgical and nonsurgical treatment groups were described. Seventy-two patients had conservative treatment, and 222 had one of three different fusion techniques. At the two-year follow-up evaluation, independent assessors rated the outcome as excellent or good in 46% of the patients in the surgical group compared with 18% in the conservative treatment group. Sixty-three percent of the surgical patients also rated their results as much better or better compared with 29% in the conservative group. The authors reported that patients in the surgical group had significantly better improvement with regard to pain and disability and a higher rate of return to work. Because more patients returned to work after surgery, the investigators found that lumbar fusion was a cost-effective treatment49.
The major criticism of the Swedish study is the nature of the conservative treatment. Patients in the nonoperative group were treated with "usual" care and did not have a regimented rehabilitation program. Furthermore, since all of the study patients had had a minimum of two years of pain, patients in the control group might have perceived nonsurgical treatment as certain to fail again. This could have contributed to negative patient expectations and may explain why the results of the control group were worse than most epidemiological studies on the natural history of low back pain.
More recently, a Norwegian study randomized patients with chronic low back pain to posterolateral lumbar fusion compared with a cognitive intervention and exercise program50. This study addressed the major criticism of the usual nonoperative treatment in the Swedish study by utilizing a structured regimen of cognitive intervention and exercises in the nonoperatively treated patients. They reported the results for sixty-one patients with chronic low back pain and disc degeneration at L4-L5 or L5-S1, or both. Thirty-five patients were treated with lumbar fusion, and twenty-six patients were treated nonsurgically with the supervised cognitive intervention and exercise program. This supervised rehabilitation program taught patients how to suppress and reinforce various peripheral stimuli. Patients were also educated on the intricate interplay between the various lumbar pain-generating structures. In addition, they were encouraged to stay active and engage in ordinary activities of daily life. Exercise programs focused on developing endurance and coordination of the deep abdominal and lumbar multifidus muscles. At the end of one year, the authors found that fear-related avoidance and forward flexion were substantially improved after conservative treatment. They also found no significant difference in back pain, use of analgesics, life satisfaction, and return to work between the groups treated operatively and nonoperatively. In contrast to the Swedish trial, this study suggests that lumbar fusion and nonoperative treatment with a modern rehabilitation program produce comparable outcomes.
New procedures, surgical techniques, and devices continue to flood the market with promises of eliminating back pain, decreasing surgical morbidity, and preserving function. While there has been recent interest in intradiscal electrothermal therapy and total disc arthroplasty as alternatives to fusion, there is limited evidence to support their use in patients with low back pain.
The role of intradiscal electrothermal therapy in the treatment of discogenic back pain is controversial. The three small, randomized controlled trials lack consistent clinical outcomes and long-term follow-up. In the first study, Barendse et al. evaluated thirteen patients after only eight weeks and found only one patient with a successful outcome after intradiscal electrothermal therapy stimulation51. Freeman et al. showed no difference between patients undergoing intradiscal electrothermal therapy and a control group undergoing a sham procedure52. In a third study, Pauza et al. reported on a highly selected group of sixty-four patients and found that intradiscal electrothermal therapy resulted in a significantly greater improvement with regard to disability and pain53. The mixed results of these small limited trials suggest that the use of intradiscal electrothermal therapy for discogenic back pain remains at best controversial.
Adjacent-segment disease is a well-recognized entity in the lumbar spine that often requires revision surgery54. There is renewed interest in motion-preserving technology to retard or halt adjacent-segment disease, and a variety of implants, including posterior dynamic rods, nucleus pulposus replacements, facet replacements, and total disc arthroplasty, have been developed. Currently, only total disc arthroplasty has been approved by the U.S. Food and Drug Administration for discogenic back pain.
McAfee et al. compared the SB CHARITÉ disc arthroplasty device (DePuy Spine, Raynham, Massachusetts) and the BAK anterior interbody fusion device (Spine-Tech, Minneapolis, Minnesota) and found no significant difference in Oswestry disability scores between the artificial disc and fusion groups at two years55. Zigler et al. and Delamarter et al. each reported six-month results after arthroplasty with the ProDisc (Synthes, Paoli, Pennsylvania) compared with circumferential 360° fusion 56,57. Both found that patients who had a disc replacement had a trend toward better Oswestry disability scores; however, at six months, there was no significant difference in pain, disability, or patient satisfaction. While these initial results seem promising, without long-term data from large studies, these devices remain a controversial choice in the treatment of discogenic low back pain. Clinicians must also remain cognizant that these studies were sponsored by industry, which can lead to publication bias. Industry-funded studies are more likely to show positive results and are more frequently published than independent studies58. Although cooperation with industry is necessary, surgeons need to remain intimately involved in the design, data collection, conduct, and analysis of randomized controlled trials in order to produce quality, unbiased studies59.
Despite these equivocal results from both old and new techniques, Americans continue to spend more each year for the treatment of back pain. The total cost of treating back pain was an estimated $85.9 billion dollars in 2005, an increase of >65% since 1997. Much of the staggering rise in cost can be attributed to an increase in the number of fusion procedures. The number of spine fusions increased by 77% between 1996 and 2001 and continues to grow annually60. At the same time, the cost of spinal fusion continues to escalate. The cost of performing spinal fusions in 2004 was estimated to be $17.6 billion, more than three times the cost in 199861. Clearly, we continue to do more and spend more to treat our patients with back pain, although there is little evidence that surgery reliably decreases pain, increases function, or enables patients to return to work.
Patients who undergo surgery are also exposed to the potential morbidities of fusion surgery. Spinal fusion carries a higher rate of complications compared with decompressive spine procedures such as discectomy and laminectomy62. Blood transfusion rates are sixfold higher and postoperative mortality rates are double those of nonfusion spine procedures. Other common complications associated with spinal fusion include infection, instrument failure, nonunion, neural injury, reoperation, and spinal fluid leak. Patients can also have catastrophic complications such as pulmonary embolus, paralysis, blindness, and death. Posterior lumbar surgery itself has been shown to cause paraspinal muscle damage and pain from prolonged retractor use63.
We need to provide proven, effective treatments for low back pain, especially when our enormous and rapidly rising health-care costs are being carefully scrutinized. Scientific evidence continues to support the use of a nonsurgical, conservative approach to treating discogenic back pain. Although some patients develop a chronic disabling pain, the available literature continues to suggest that surgery should be offered cautiously. Future research should focus on identifying those who will benefit from surgery and not on what type of surgery. Perhaps a better understanding of who should undergo surgery will lead to better and consistently good surgical outcomes. Until then, low back pain is most appropriately managed with nonsurgical treatment.
Summary
Lumbar arthrodesis is a commonly performed operative procedure for the treatment of low back pain. Recently, total disc arthroplasty has gained some acceptance among surgeons and patients. However, the indications for and results of back pain surgery remain controversial and confusing. Available information suggests that meaningful functional improvement from these procedures is debatable and that the cost of such elective operations is high. Currently, lumbar disc replacement has gained minimal support from governmental and private payers. Among those attending this symposium at the 2009 Annual Meeting of the AOA, the vast majority concurred that Medicare and private insurance should not necessarily pay for disc replacement surgery. Interestingly, among this skeptical group of orthopaedic surgeons, only 23% believed that degeneration of the intervertebral disc is the major cause of low back pain. When asked the hypothetical question, "If you experienced chronic low back pain with degenerative changes at one level, what course of treatment would you opt for?" 61% responded that they would choose nonoperative treatment and 38%, that they would choose no treatment. Of more than 100 respondents, only one responded that he or she would undergo fusion and one admitted a willingness to undergo disc replacement in this hypothetical scenario.