Are COX-2 NSAIDs the Most Efficacious Treatment for Chronic Nonspecific Low Back Pain?

Chronic nonspecific low back pain (CNLBP) secondary to nonspecific musculoskeletal inflammation of three months or more duration is one of the leading causes of activity limitation, impaired quality of life, and occupational truancy.¹ Pharmacologic intervention remains an essential part of the treatment paradigm (at least in the short-term), in addition to physical therapy and exercise. Though the female gender and middle to elderly-aged individuals are more commonly affected, the global mean point prevalence of activity-limiting low back pain endured more than one day was estimated to be 11.9%, and the global mean lifetime prevalence of low back pain is 38.9%,² with North American estimates as high as 85%.³

The authors of “Drug Therapy for the Treatment of Chronic Nonspecific Low Back Pain: Systematic Review and Meta-analysis,” published in Pain Physician, conducted a systematic review and meta-analysis of randomized controlled trials from 2002 to 2012 that assessed oral and injection drug therapy for chronic nonspecific low back pain in adults (18 years and older). Of the 23 trials analyzed, there was a trend of use of cyclo-oxygenase-2 (COX-2) nonsteroidal anti-inflammatory drugs (NSAIDs), tramadol, and opioids in this patient population. Antiepileptics and antidepressants were studied less commonly and mainly used in adjuvant roles.

Use of COX-2 NSAIDs resulted in the greatest margin of pain relief, with a standardized mean difference (SMD) of -12.03 (95% confidence interval [-15 to -9.06]). This was followed by opioids, with a SMD of -5.18 (95% CI of -8.3 to -2.05), and partial opioid agonists such as buprenorphine, with a SMD of -7.46 (95% CI of -11.87 to -3.04). Opioids were likewise more effective than placebo for global functional improvement.

Though statistically insignificant for pain relief, tramadol demonstrated improved functional status using the fixed-effect model of global improvement, with weighted mean difference of -0.24. Analysis of the individual opioid oxycodone showed a similar trend of statistical insignificance for pain relief, but patients who received treatment with the drug demonstrated significant global improvement. Oxymorphone was statistically significant for both pain relief and global improvement. Data from a few pooled studies demonstrated improved pain relief indices for topiramate, duloxetine, and bupropion, suggesting a limited role for antiepileptics and antidepressants for CNLBP. Interestingly, there was no statistically significant efficacy difference found between traditional NSAIDs and COX-2 NSAIDs, but COX-2 NSAIDs showed fewer side effects after statistically pooling data from two studies.

Commentary from Dr. Mitchell

This systematic review of the pharmacologic treatment of chronic low back pain was comprehensive for analysis of oral agents, identification of biases, and demonstration of study heterogeneity or homogeneity. The heterogeneity of patient populations studied is inevitable when the primary disease (chronic low back pain) is broad and inclusive of a spectrum of underlying etiologies.

Although statistical pooling was enabled by statistical models for clinical heterogeneity in this study, the qualitative criteria used were flawed in that it specified CNLBP as lower back pain with or without radiation. Radicular back pain is a feature of neurogenic back pain—a subdivision which was semantically excluded from the study analysis of nonspecific back pain. The treatment effects may be skewed by inclusion of neurogenic back pain patients in this group, and perhaps tramadol is shown to be more effective than if this subgroup were excluded, thus inflating global improvement measures.

One of the inescapable flaws of any meta-analysis is the intrusion of methodological variation. For example, what defines chronic non-specific low back pain? Low back pain may be the phenotype of myriad causes, including but not limited to: musculoskeletal strain, degenerative disc disease, disc herniation, spinal stenosis, osteoporosis, trauma, and vertebral compression fracture or combination thereof. Certain etiologies such as disc herniation and lumbar muscle strain may respond more favorably to NSAIDs by dint of their inflammatory mechanisms. Rarer outlier etiologies, such as ankylosing spondylitis, osteomyelitis, and metastatic or primary tumors, may be more responsive to local site injection or focused pharmaco-surgical interventions.

As a result of studying the entity of chronic low back pain, treatment effects may show dramatic variations across different studies. To minimize this effect in the meta-analysis, chronic nonspecific low back pain excluded causes such as: metastatic disease, inflammation, and infection, as well as exceptional cohorts of pregnant women and cases symptomatic of mechanical injury. An alternative means of eliminating the outlier effect of a heterogeneous cohort would be to prospectively compare two drugs in one particular etiology of back pain that is objectively proven, such as spinal stenosis or anterolisthesis. However, such a study would suffer from the shortcoming of reduced number of subjects and reduced power to study the effect of the therapeutic intervention, which can be more easily demonstrated in a systematic review.Gender stratification would be interesting as well, since studies have shown a greater percentage of female CNLBP sufferers.Sampling error may also be at play as some providers may favor administration of one class of drugs over another, prescribing one agent to everyone and thus improving the probability of success and under-prescribing another agent, including using it in situations in which it is doomed to fail. A side-by-side comparative study of NSAIDs versus opioids would provide useful prospective data of CNSLBP.

In addition, special considerations of combination therapy were not addressed by this analysis. Controlling for concomitant administrations of NSAIDs, opioids, tramadol, antiepileptics, and antidepressants was not addressed, and thus, the reported effect of individual agents may be confounded.

Although effective use of adjuvant epidural corticosteroids and facet-joint and trigger point injections of local anesthetics (bupivacaine, sarapin) for chronic low back pain has been demonstrated,^4,5 injection drug therapy was not mentioned in this review. Due to significant methodological heterogeneity noted at times amongst the included studies, single summary measures, such as mean prevalence of pain relief, should be interpreted with caution and with consideration of the specific characteristics of the prospective patient being treated. For example, a patient with chronic low back pain and stage III chronic renal disease may not be the optimum candidate for NSAID therapy despite their demonstration of superiority in this systematic review, and local epidural steroid injections with fewer systemic adverse effects may be the best individualized treatment option.

Overall, the study demonstrated how the prescient and short-term use of NSAIDs and opioids with pharmacological sensitivity to an individual’s comorbidities may result in pain relief, improved functioning, and reduced vocational and avocational absenteeism.

References:

1. Chung JW, Zeng Y, Wong TK. Drug Therapy for the Treatment of Chronic Nonspecific Low Back Pain: Systematic Review and Meta-analysis. Pain Physician 2013; 16:E685-E704.

2. Hoy D, Bain C, Williams G, et al. A systematic review of the global prevalence of low back pain. Arthritis Rheum 2012; 64(6):2028-37.

3. White AP, Arnold PM, Norvell DC, et al. Pharmacologic management of chronic low back pain: Synthesis of the evidence. Spine 2011; 36:S131-S143.

4. Tonkovich-Quaranta LA, Winkler SR. Use of Epidural Corticosteroids in Low Back Pain. Ann Pharmacother 2000; 34(10):1165-72.

5. Staal JB, de Bie R, de Vet HC. Injection therapy for subacute and chronic low back pain. Cochrane Database Syst Rev 2008; CD001824.

6. Bannwarth B, Kostine M, Shipley E. Nonspecific low back pain: Assessment of available medications. Joint Bone Spine 2012; 79(2):134—36.