Chronic low back pain (LBP) is a common disorder. It afflicts 70% to 85% of the people in North America at some point in their lives. In the United States, back pain is the most common reason for persons under 45 years of age to limit activities, including work-related activities.1 For this reason, in part, LBP has become a major socioeconomic problem. Direct and indirect economic losses have been estimated to be nearly $90 billion annually.2
Medication is the most frequently used intervention for chronic LBP. The popularity of pharmacologic management may be related to many factors, including high patient volume and turnover in busy primary care practices, pervasive availability and marketing of pain management clinics, and raised expectations for physicians to continuously suppress pain. There is an increasing recognition that many patients are made to feel entitled to a “pain free” life. One study reported that primary care physicians prescribed one or more medications for 80% of their patients with chronic LBP and two or more medications for 35% of their patients with chronic LBP.3
The most commonly prescribed medications include nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, and antidepressants. NSAIDs are the most frequently prescribed medications worldwide and are frequently recommended as an option in chronic LBP treatment. Many other types of medications are used, however, including Tylenol, skeletal muscle relaxants, benzodiazepines, systemic corticosteroids, and antiepileptics.
While chronic LBP is common, expensive, and most frequently treated with medications, the literature is relatively devoid of high quality medical evidence upon which strong treatment recommendations can be made. This is not related to a lack of data, however; there is an abundance of information available on this common treatment method for patients with this common disorder. There are many randomized controlled trials (RCTs) available. Furthermore, several reviews regarding the most frequently prescribed medications (NSAIDs, opioids, antidepressants) have been published in recent years. These reports have not been encouraging.
Roelofs et al4 published a Cochrane review comparing NSAIDs with placebo or other medications for the treatment of acute and chronic LBP. They reviewed 65 trials representing 11,237 patients. They classified 28 of these studies as high quality. These authors concluded that NSAIDs were slightly effective for short-term relief of acute and chronic LBP and that NSAIDs were not more effective than other drugs including acetaminophen, narcotics, and muscle relaxants. A Cochrane review by Deshpande et al5 considered the efficacy of opioid medications for treatment of chronic LBP. This review included four trials. Three of the trials compared Tramadol to placebo, representing 908 patients. The fourth trial compared morphine (or morphine derivative) to NSAIDs, representing 36 patients. These Cochrane review authors reported that those receiving Tramadol had more pain relief and less difficulty with daily activities as compared with placebo, and that those receiving morphine (or derivative) had little or no difference in pain relief as compared with NSAID. They also reported that weak opioids reduce pain but had minimal effect on function.5 The third recent Cochrane review considered antidepressants. In this review, Urquhart evaluated 10 RCTs to evaluate if antidepressants are more effective than placebo for treatment of acute and chronic LBP. The pooled analysis showed no difference in pain relief between antidepressant and placebo. The qualitative analysis found conflicting evidence on the effect of antidepressants on pain intensity and no clear evidence that antidepressants reduce depression in patients with chronic LBP.6
In general, the results of treatment for chronic LBP in RCTs are less than encouraging.7 Poor treatment results, however, may be related to incorrectly classifying chronic LBP as a homogeneous entity when in fact it is heterogeneous.8–10 Chronic LBP may be best characterized as a symptom reflective of many heterogeneous disorders, each with a different cause. As such, each patient with chronic LBP may respond more favorably or less favorably to a given treatment, since any given treatment may or may not be particularly appropriate for the root cause of each individual's pain.
The heterogeneity of patients with chronic LBP has an important effect on the results of RCTs evaluating this group. Results from RCTs represent average effects (population means) and, while estimates of the average treatment effect are useful, some individuals will respond more positively or more negatively than the reported average. Such variation in results is termed heterogeneity of treatment effects (HTE).11 When the same treatment results in different outcomes in different patients, HTE is present. One way to identify HTE is to analyze the effect of treatment in subgroups of patients with certain baseline characteristics.
Subgroup analyses may be prone to spurious results, however. This is due to the problem of multiple testing.12 Many caution against subgroup analyses, especially post hoc comparisons.13 Nevertheless, identification of subgroup effects in clinical trials can generate important hypotheses about potential factors that modify treatment effects. Recommended reporting of subgroups analysis includes statistical tests of interaction or heterogeneity and a description of whether the subgroup analysis was prespecified or post hoc.13–15 When attempting to evaluate possible HTE for a systematic review, care must be taken in selecting the appropriate study designs and in ensuring the results are presented in such a way that the possibility of HTE can be assessed. Ideally, one identifies randomized trials that compare the two treatments of interest (e.g., opioids vs. placebo), and report treatment effects by subgroup (e.g., smokers and nonsmokers) in addition to the overall treatment effect. It is not uncommon to find that there is little to no treatment effect in all patients; however, with enough study power and careful planning, subgroup differences may exist. Given that only one treatment is evaluated in a case series, this design does not address the question of whether treatment differences vary according to differing subgroup characteristics.15–18
The goal of this investigation was to help determine the safety and effectiveness of medications commonly prescribed in the treatment of chronic LBP and to determine if specific subgroups respond more favorably than the mean population to these medications. As such, we developed the following clinical questions:
- What is the relative effectiveness and safety of opioids in the treatment of chronic LBP and does it differ by subpopulation?
- What is the relative effectiveness and safety of NSAIDs for the treatment of chronic LBP and does it differ by subpopulation?
- What is the relative effectiveness and safety of antidepressants for the treatment of chronic LBP and does it differ by subpopulation?
MATERIALS AND METHODS
Electronic Literature Database
A systematic search was conducted in MEDLINE and the Cochrane Collaboration Library for literature published from January 1978 through December 2010. We limited our results to Cochrane reviews, RCTs, studies with human subjects, and to articles published in the English language. All previously published systematic reviews were also checked to ensure inclusion of previously summarized RCTs regarding (1) opioids, (2) NSAIDs, and (3) antidepressants for the treatment of chronic LBP.6,19,20 While we are aware that there are many other medications commonly prescribed, including Tylenol, these three drug classes were ultimately selected for analysis, since there is not significant published data on other medications to accommodate systemic analysis. In addition, we recognize that it is important to differentiate between the short-term and long-term risks and benefits of these medications. In this inquiry regarding chronic LBP, the long-term safety and efficacy of the treatment is of particular interest. As such, we focused our analysis not only on patients with chronic LBP, but also on medication treatments ongoing for more that 12 weeks.
To summarize the overall treatment effect and safety for each of the three drug classes (opioids, NSAIDs, antidepressants), we reviewed previously published Cochrane reviews. To evaluate whether the effects of treatment varied by specific subgroups, we sought RCTs evaluating one of the three drug classes versus an alternative management for chronic LBP. More specifically, we approached the literature to identify the following: (1) RCTs designed specifically for evaluating this comparison stratifying the random assignment on one or more subgroups, (2) RCTs designed specifically for this comparison that included a subgroup analysis stratifying on one or more subgroups, (3) RCTs that made the comparison among patients within a specific subgroup (e.g., neurological deficit only or no depression) to compare with other RCTs that were conducted among patients without characteristics of this subgroup. We excluded studies that did not report medication effectiveness and/or safety (i.e., one medication vs. another medication) separately for the subgroups being compared unless they performed a statistical test for determining if the subgroup modified the effectiveness and/or safety (i.e., test for interaction). For example, if the authors reported a multivariate regression that included a subgroup variable (e.g., age or sex) and the treatment variable (e.g., drug therapy vs. alternative care), without an interaction term, the study was excluded. Articles were also excluded if they included the following patient groups: pediatric patients (<18 years of age), patients with radiculopathy, prior surgical intervention, primary diagnoses of cancer, deformity, instability, infection, and/or trauma contributing to the LBP condition. Other exclusions included non-Cochrane reviews, editorials, case reports, and non–English-written studies, and studies without subgroup analyses (Figure 1).
For the Cochrane reviews, we extracted data only relevant to chronic LBP as other conditions including acute LBP were often summarized in these reviews. For the subgroup analyses, each retrieved citation was reviewed by two independently working reviewers (D.C.N., E.E.). Some articles were excluded on the basis of information provided by the title or abstract if they clearly were not a randomized trial or were evaluating a patient population that did not meet our inclusion criteria. Otherwise, all full text reports were reviewed by the two reviewers regardless of whether the abstract reported a subgroup analysis. Any disagreement between them was resolved by consensus. From the included articles, the following data were extracted for both treatment groups if the data were available: outcome, risk factor or subpopulation, rates of outcome (where appropriate), pre- and posttreatment and change scores (where appropriate), effect estimates (e.g., odds ratio, relative risk, treatment effect), and associated P values. Tests for interaction of treatment effects were included when reported by the author.
For assessing the level of evidence for the overall treatment effect of pharmacologic interventions, we rated each Cochrane review using both a quantitative and qualitative analysis. For assessing the level of evidence for subgroup effects, level of evidence ratings were assigned to each article independently by two reviewers using criteria set by The Journal of Bone and Joint Surgery, American Volume (J Bone Joint Surg Am)21 for therapeutic studies and modified to delineate criteria associated with methodological quality and described elsewhere.22
We summarized the findings from the three Cochrane reviews. We included the study level summary findings from the meta-analyses provided by the authors including standardized mean differences (SMDs) for continuous outcome measures and risk differences (RDs) for binary outcomes. We constructed forest plots using these effect sizes and confidence intervals to provide a visual demonstration of treatment effects. For the subgroup analyses, data between studies were not pooled because of heterogeneity of subject populations, outcome measures, lack of raw data (in some instances), and differing effect estimates. In rare cases, we were able to report or calculate effect sizes from subgroup analyses.
Overall Strength of Body of Literature
To determine the overall strength of the evidence we used a modification of the GRADE criteria. Initial strength was established on the basis of study design and was subsequently upgraded or downgraded on the basis of the results. Level of evidence ratings were assigned to each article independently by two reviewers using criteria set by The Journal of Bone and Joint Surgery, American Volume (J Bone Joint Surg Am)21 for therapeutic studies and modified to delineate criteria associated with risk of bias and methodological quality described elsewhere.22 The initial strength of the overall body of evidence was considered high if the majority of the studies were level I or II and low if the majority of the studies were level III or IV. We downgraded the body of evidence one or two levels based on the following criteria: (1) inconsistency of results, (2) indirectness of evidence, (3) imprecision of the effect estimates (e.g., wide confidence intervals) (4) if the authors did not state a priori their plan to perform subgroup analyses and if there was no test for interaction. We upgraded the body of evidence one or two levels based on the following criteria: (1) large magnitude of effect or (2) dose-response gradient. The overall strength of the body of literature was expressed in terms of the impact that further research may have on the results. An overall strength of high means we have high confidence that the evidence reflects the true effect. Further research is very unlikely to change our confidence in the estimate of effect. An overall strength of moderate is interpreted as moderate confidence that the evidence reflects the true effect. Further research may change our confidence in the estimate of effect and may change the estimate. A grade of low means we have low confidence that the evidence reflects the true effect. In this case, further research would be likely to change the confidence in the estimate of effect and likely to change the estimate. A grade of insufficient suggests that evidence is unavailable or does not permit a conclusion.
Three Cochrane reviews were identified that addressed the efficacy and safety of opioids, NSAIDs, and antidepressants for the treatment of LBP. Two of the reviews (regarding opioids and antidepressants)5,6 included studies with populations comprising chronic LBP patients only (one of the 14 study populations was unspecified). In the third review (regarding NSAIDs),4 the majority of studies were conducted in patients with acute LBP; only four studies comparing NSAIDs to placebo were in chronic low back patients with results reported separately and were included in the final analysis (Figure 2).
For the subpopulations, our search strategy identified 97 total citations. Of these, 34 were excluded by abstract and 63 full text articles (all RCTs) were retrieved to determine if they met the criteria. Of these 63, all were RCTs, but only 12 reported treatment effects separately by subgroup (Figure 3). This highlights the deficiencies in the literature that describes comparative pharmacology effects by different risk factors in patients with chronic LBP. It also serves to provide hypotheses regarding the possibility of HTE by subpopulations. Data across studies could not be pooled due to the differences in subgroups, treatment comparisons, and outcomes recorded. Therefore, individual study findings are presented.
What is the relative effectiveness and safety of opioids in the treatment of chronic LBP and are there differences by subpopulation?
A meta-analysis of three trials investigated the efficacy of opioids (tramadol) versus placebo in a total of 908 patients with chronic LBP with and without radiating pain5 (Table 1). Studies were excluded if patients' LBP was due to nondegenerative pathologies such as infection, neoplasm, and compression fractures. In the pooled analysis, compared with those on placebo, patients who received tramadol reported better pain relief with a mean 10.8-point difference in the 100-mm visual analog scale (VAS). They also reported improved function, with a mean 1-point difference as measured by the Roland Disability Questionnaire. SMDs for opioids versus placebo were −0.71 (95% CI: −0.84 to −0.57) for pain and −0.17 (95% CI: −0.30 to −0.04) for disability (Table 2). This difference favoring opioids was statistically significant (Figure 4A). Only one study compared opioids with other analgesics and the authors reported better pain relief in both morphine groups (set-dose and titrated-dose) compared with the naproxen group and a greater effect in the titrated-dose versus the set-dose morphine group (Table 2). The Cochrane review's calculations, however, failed to show a statistically significant difference in pain relief between the titrated-dose group and the control group (SMD, −0.58; 95% CI, −1.42 to 0.26), probably due to the small sample size (Figure 4B). No significant difference in disability was found when the morphine groups were compared with the naproxen group (SMD, −0.06; 95% CI, −0.88 to 0.76) (Figure 4B).
Headache and nausea were the only side effects reported among all three studies comparing opioids with placebo and were more prevalent in the opioid group (risk difference [RD] = 9%; 95% CI, 6%–12% and RD = 3%; 95% CI, 0%–6%, respectively). In two studies (654 patients), somnolence, constipation, dry mouth, and dizziness were more common after treatment with opioids with RDs ranging from 7% to 9% (95% CI, 4%–13%) (Table 2). Finally, one study (336 patients) reported a significantly greater incidence of pruritus, vomiting, anorexia, and increased sweating in the opioid versus placebo group with RDs ranging from 4% to 6% (95% CI, 1%–10%). For the comparison between opioids and other analgesics, the most frequent side effects were dry mouth, drowsiness, headaches, constipation, and nausea. No intergroup differences were reported.
Three studies were identified that examined the efficacy of opioids for the treatment of chronic LBP in various subpopulations. Of these, only one trial reported significant treatment effects by subgroup. This trial is summarized below. The remaining two studies23,24 that reported no significant treatment effects by subgroup are briefly outlined in Table 1 (see Table 1, Supplemental Digital Content 1, http://links.lww.com/BRS/A542). Details of these study populations and results can be found in the Web appendices (see Supplemental Digital Content 1, http://links.lww.com/BRS/A542).
In the one trial that reported a significant treatment effect (TE) by a subgroup, the change in Bowel Function Index score and the number of complete spontaneous bowel movements was examined. A subset of 59 patients with high Bowel Function Index scores (≥50) at baseline was treated with oxycodone PR (n = 30) or oxycodone PR/naloxone PR (n = 29).25 Compared with the group receiving oxycodone PR, those on oxycodone PR/naloxone PR had a significantly greater improvement in bowel function (TE based on Bowel Function Index score: −11.8; mean difference in number of complete spontaneous bowel movements: 2.0), Table 2 in the Web appendices (see Table 2, Supplemental Digital Content 1, http://links.lww.com/BRS/A542).
What is the relative effectiveness and safety of NSAIDS for the treatment of chronic LBP and are there differences by subpopulation?
A meta-analysis of four trials compared the efficacy of NSAIDs4 with the efficacy of placebo in 1020 patients with chronic LBP with or without radiating pain (Table 1). Studies were excluded if patients' LBP was due to specific pathologies such as infection, neoplasm, metastasis, rheumatoid arthritis, or fractures. Pooled results of all four studies indicated a statistically significant decrease in pain intensity (100 mm VAS) after treatment with NSAIDs compared with placebo (weighted mean difference −12.4; 95% CI −15.5 to −9.3) (Table 3 and Figure 5).
Side effects were reported in the same four studies in 1034 patients. In the pooled results, significantly more side effects were reported in the NSAID group compared with the placebo group (risk ratio 1.24; 95% CI, 1.07–1.43) (Table 3). Specific side effects were not listed.
Four studies were identified that reported the efficacy of NSAID treatment for chronic LBP in subpopulations. Of these, only two trials reported significant treatment effects by subgroup and are summarized below; the remaining two studies26,27 that reported no significant treatment effects by subgroup are briefly outlined in Table 1 in the Web appendices (see Table 1, Supplemental Digital Content 1, http://links.lww.com/BRS/A542). Details of these study populations and results can also be found in the Web appendices (see Supplemental Digital Content 1, http://links.lww.com/BRS/A542).
Among the studies reporting significant treatment effect (TE) by subgroups, one study24 reported that anxiety was a significant factor effecting pain improvement. Anxiety was measured with the Hospital Anxiety and Depression Scale and effectiveness of treatment was measured by change in Low Back Intensity Scale scores. As compared with placebo, patients receiving 25 mg rofecoxib had decreased effectiveness of treatment with increasing anxiety (see Table 3, Supplemental Digital Content 1, http://links.lww.com/BRS/A542). In patients with no anxiety, the TE between 25 mg rofecoxib and placebo was −19.5, compared to those with mild anxiety, −9.9, and those with moderate/severe anxiety, 1.8; P = 0.013. A similar trend was not evident in differences between the 50-mg rofecoxib and placebo treatment groups or in the subgroup analysis based on the Hospital Anxiety and Depression Scale depression interpretation, as well as other variables investigated.
The second study looked at the effect of pain characteristics on pain scores in patients prescribed various types of NSAID treatments (aspirin, dextropropoxyphene and paracetamol, indomethacin, mefenamic acid, paracetamol, and phenylbutazone).28 Only two factors showed significant interactions with treatment. These were (1) whether pain was brought on by sneezing and (2) whether pain radiated from the primary focus (see Table 4, Supplemental Digital Content 1, http://links.lww.com/BRS/A542). In patients who had pain brought on by sneezing, treatment with aspirin, phenylbutazone, or dextropropoxyphone plus paracetamol, resulted in significantly better pain relief than mefenamic acid (1.06, 1.65, and 1.79 vs. 2.47, respectively, P < 0.05). The same was true when aspirin was compared with indomethacin (1.06 vs. 2.28, P < 0.05). When comparing patients with radiating (with or without sneezing) versus nonradiating pain, mefenamic acid was significantly more effective in reducing pain than phenylbutazone and indomethacin in patients with nonradiating pain (0.98 vs. 2.47 and 2.75, respectively, P < 0.05).
What is the relative effectiveness and safety of antidepressants for the treatment of chronic LBP and are there differences by subpopulation?
A meta-analysis of pain outcomes was possible in 6 of the 10 trials comparing antidepressants with placebo.6 A total of 376 patients with nonspecific chronic LBP with and without radiating pain were analyzed (Table 1). Studies were excluded if patients' LBP was due to specific pathologies such as infection, neoplasm, metastasis, osteoporosis, rheumatoid arthritis, or fractures. No significant difference was seen in pain relief (VAS) between the two study groups (SMD = −0.04; 95% CI, −0.25 to 0.17) (Table 4 and Figure 6). Functional outcome (Oswestry Disability Index, Sickness Impact Profile) and depression (Montgomery Asberg Depression Rating Scale, Beck Depression Inventory) were reported by only two small studies (132 patients) and the pooled analysis showed no significant difference in improvement of either outcome between those on antidepressants and those taking placebo (SMD = −0.06; 95% CI, −0.40 to 0.29 and MD = 0.06; 95% CI, −0.029 to 0.40, respectively) (Table 4 and Figure 6). A separate analysis was performed to evaluate the effect of different types of antidepressants on pain intensity. The pooled analysis revealed no significant difference in pain relief between either tricyclic antidepressants compared with placebo (three trials, 148 patients) or selective serotonin-reuptake inhibitors (SSRI) compared with placebo (three trials, 199 patients) (SMD = −0.10; 95% CI, −0.51 to 0.31 and SMD = 0.11; 95% CI, −0.17 to 0.39, respectively) (Table 4).
No data on safety were reported in these trials. The adverse effects of antidepressants have been well established in other studies, however, and are reviewed below in the discussion.
Five studies were identified that reported the efficacy of antidepressants for the treatment of chronic LBP in subpopulations. No significant treatment effects by subgroup were reported by any of the trials (see Table 1, Supplemental Digital Content 1, http://links.lww.com/BRS/A542). One of the studies, which compared Tofranil with placebo in groups with and without previous history of LBP, did report appreciable, though not statistically significant numerical difference between the subgroups.29 In particular, antidepressants were favored over placebo for decreasing stiffness in patients with no history of LBP, and for decreasing psychosocial symptoms of anxiety, depression, and other traits (Middlesex Hospital Questionnaire), in patients with a history of LBP (see Table 10, Supplemental Digital Content 1, http://links.lww.com/BRS/A542).
The overall strength of the evidence evaluating the efficacy of NSAIDs, opioids, and antidepressants is “high,” that is, we have high confidence that the evidence reflects the true effect and that further research is very unlikely to change our confidence in the estimate of effect (Table 5). There is little research published evaluating subgroup effects among these three drug classes. Therefore, the evidence is “insufficient” for whether specific patients respond more favorably than others to specific pharmacology management strategies. That is, evidence either is unavailable or does not permit a conclusion; however, some hypotheses can be generated and considered for future research planning.
The purpose of this systematic review was to determine the relative effectiveness and safety of opioids, NSAIDs, and antidepressants in the treatment of chronic LBP and to determine if the treatment effects differed by subpopulation. We evaluated the relative effectiveness by summarizing recently published Cochrane reviews on each of these drug classes. To examine the effects in subpopulations, we used a systematic approach that would allow us to evaluate study outcomes based on the heterogeneity of treatment effects. We reviewed the full texts of 63 RCTs and identified 12 studies that reported subgroup effects by treatment spread across the three drug classes. This demonstrates a significant deficiency in the literature. Since chronic LBP is likely to be a heterogeneous disorder, related to a heterogeneous set of causes, it is likely that certain subgroups will respond differently than others to a particular treatment. Since there is a significant deficit in the published literature, with little to no informationavailable on subgroups within the chronic LBP population, it is unfortunately difficult to determine if certain patient groups respond better to pharmacological treatment than others. Such knowledge, if it existed, would likely assist the provider in treatment decision making.
We recognize that prescription of multiple medications or multiple therapies may occur simultaneously. Concomitant therapies are likely to be the norm. For example, a patient with chronic LBP may receive two or more medication therapies, as well as exercise therapies at the same time. There may be an additive effect to such a strategy. This additive effect is discussed in the introduction article in this focus issue. In our analysis of these medications, we were not able to determine the additive effect of multiple or combination pharmacologic therapies.
We considered three classes of medications in this analysis. We recognize, however, that many other types of medications commonly prescribed. Tylenol is used quite frequently, for example, as are muscle relaxants, anticonvulsants, and others. The three drug classes analyzed (NSAIDs, opioids, and antidepressants) were selected for analysis since the bulk of the available published data is focused on these medications.
In this analysis, we strived to recognize the critical importance of differentiating between the short-term and long-term risks and benefits of these medications. In this inquiry regarding chronic LBP, the long-term safety and efficacy of the pharmacologic treatment is of particular interest. As such, we focused our analysis not only on patients with chronic LBP, but also on medication treatments ongoing for more that 12 weeks. Our conclusions and recommendations also are a reflection of this recognition; the long-term efficacy and the long-term risk profile of habituating medications are considerably different than the short-term efficacy and risks. These well recognized long-term risks include death, and are and well reported.30
From our analysis of the Cochrane reviews, we did observe that opioids are more effective than placebo with respect to pain and disability improvement. The effect size was much greater for pain than disability. When compared with NSAIDs, however, in a small trial, there was no significant difference between the two with respect to pain and disability. The rate of side effects related to opioids is significantly greater than placebo, however, ranging from differences of 2% to 9%. The side effects are well known, and can be associated with severe adverse events.31
As for differing effects by subpopulations, only one RCT reported subgroup differences for opioids. Among patients with a history of bowel dysfunction, a combination of oxycodone and naloxone led to significantly better improvement in bowel function compared with oxycodone only. Two other studies did not find treatment effect differences by age, sex, or race.
From the Cochrane review analysis, NSAIDs were found to be significantly more effective than placebo with respect to pain improvement. When compared with opioids in a small trial, however, there was no significant difference between the two with respect to pain and disability. With this comparison indicating similar effectiveness between NSAIDs and opioids, prescribing decision making may then be made on comparisons of other factors. These other factors, such as safety profile, cost, and long-term effectiveness (i.e., risk of habituation) all favor NSAIDs over opioids. NSAIDs are of course not without risk; patients were at a 24% greater risk of side effects when taking NSAIDs instead of placebo. The side effects were not summarized, but the gastrointestinal risks are well established.32
As for differing effects by subpopulations, patients with no or mild anxiety had a greater reduction in LBP intensity scores with NSAIDs (rofecoxib) compared with placebo. On the contrary, patients with moderate to severe anxiety had a greater reduction with placebo than NSAIDs. This is concordant with our hypothesis that in this heterogeneous population, certain subgroups will respond more favorably than others to a particular treatment. Specifically in this case, patients that experience chronic LBP as a manifestation of anxiety, or have LBP exacerbated by anxiety, would not be expected to find relief with an anti-inflammatory medication.
In patients who had pain brought on by sneezing, treatment with aspirin and other specific NSAIDs resulted in significantly better pain relief. Aspirin was also more effective than indomethacin in this subgroup. When comparing patients with radiating (with or without sneezing) versus nonradiating pain, mefenamic acid was significantly more effective in reducing pain than phenylbutazone and indomethacin in patients with nonradiating pain.
From the Cochrane review analysis, antidepressants were not found to be more effective than placebo with respect to pain, functional status, or depression. There were no differences found between differing types of antidepressants. While an evaluation of side effects was not reported, significant side effects are well known and are reported elsewhere.33 From the subgroup analyses, patients with a previous history of LBP had improved psychosocial outcomes (including anxiety and depression) with antidepressants compared with placebo but less improvement in stiffness.
Based on our review, both opioids and NSAIDs appear to have similar efficacy with regard to pain and disability in patients with chronic LBP. Based on the serious side effect profile of opioids, it is recommended that NSAIDs be considered the first line pharmacological treatment for chronic LBP. We recommend that opioids be avoided if possible, particularly in the long-term treatment of this chronic condition, but may be considered as a second line treatment. Antidepressants are no more effective than placebo, however, and as such, we do not recommend that they be used for the treatment of chronic LBP.
- For chronic LBP, opioids are more effective than placebo with respect to effects on pain and disability with a much greater effect on pain than disability.
- When compared to NSAIDs, opioids did not confer a significantly greater benefit with regard to effects on pain and disability.
- Antidepressants are not an effective treatment for chronic LBP.
- Based on the significant rate of side effects with opioids and the lack of convincing superiority over NSAIDs, opioids are not recommended as first line treatment for chronic LBP.
Supplemental digital content is available for this article. Direct URL citations appearing in the printed text are provided in the HTML and PDF versions of this article on the journal's Web site (www.spinejournal.com).
The authors thank Ms. Nancy Holmes, RN, for her administrative assistance.
1. Andersson GB. Epidemiological features of chronic low-back pain. Lancet 1999;354:581–5.
2. Dagenais S, Caro J, Haldeman S. A systematic review
of low back pain
cost of illness studies in the United States and internationally. Spine J 2008;8:8–20.
3. Cherkin DC, Wheeler KJ, Barlow W, et al. Medication use for low back pain
in primary care. Spine 1976;23:607–14.
4. Roelofs PD, Deyo RA, Koes BW, et al. Non-steroidal anti-inflammatory drugs
for low back pain
. Cochrane Database Syst Rev 2008;CD000396.
5. Deshpande A, Furlan A, Mailis-Gagnon A, et al. Opioids for chronic low-back pain. Cochrane Database Syst Rev 2007;CD004959.
6. Urquhart DM, Hoving JL, Assendelft WW, et al. Antidepressants
for non-specific low back pain
. Cochrane Database Syst Rev 2008;CD001703.
7. Mirza SK, Deyo RA. Systematic review
of randomized trials
comparing lumbar fusion surgery to nonoperative care for treatment of chronic back pain. Spine 2007;32:816–23.
8. Coste J, Paolaggi JB, Spira A. Classification of nonspecific low back pain
. II. Clinical diversity of organic forms. Spine (Phila Pa 1976) 1992;17:1038–42.
9. Delitto A, Erhard RE, Bowling RW. A treatment-based classification approach to low back syndrome: identifying and staging patients for conservative treatment. Phys Ther 1995;75:470–85; discussion 485–9.
10. Hall H, McIntosh G, Boyle C. Effectiveness of a low back pain
classification system. Spine J 2009;9:648–57.
11. Kravitz RL, Duan N, Braslow J. Evidence-based medicine, heterogeneity of treatment effects, and the trouble with averages. Milbank Q 2004;82:661–87.
12. Brookes ST, Whitley E, Peters TJ, et al. Subgroup analyses in randomised controlled trials: quantifying the risks of false-positives and false-negatives. Health Technol Assess 2001;5:1–56.
13. Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 Explanation and Elaboration: Updated guidelines for reporting parallel group randomised trials. J Clin Epidemiol 2010;63:e1–37.
14. Gabler NB, Duan N, Liao D, et al. Dealing with heterogeneity of treatment effects: is the literature up to the challenge? Trials 2009;10:43.
15. Wang R, Lagakos SW, Ware JH, et al. Statistics in medicine–reporting of subgroup analyses in clinical trials. N Engl J Med 2007;357:2189–94.
16. Brookes ST, Whitely E, Egger M, et al. Subgroup analyses in randomized trials
: risks of subgroup-specific analyses; power and sample size for the interaction test. J Clin Epidemiol 2004;57:229–36.
17. Lagakos SW. The challenge of subgroup analyses—reporting without distorting. N Engl J Med 2006;354:1667–9.
18. Rothwell PM. Treating individuals 2. Subgroup analysis in randomised controlled trials: importance, indications, and interpretation. Lancet 2005;365:176–86.
19. Salerno SM, Browning R, Jackson JL. The effect of antidepressant treatment on chronic back pain: a meta-analysis. Arch Intern Med 2002;162:19–24.
20. Schnitzer TJ, Ferraro A, Hunsche E, et al. A comprehensive review of clinical trials on the efficacy and safety of drugs
for the treatment of low back pain
. J Pain Symptom Manag 2004;28:72–95.
21. Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to the journal. J Bone Joint Surg Am 2003;85-A:1–3.
22. Norvell DC, Dettori JR, Fehlings MG, et al. Methodology for the systematic reviews on an evidence based approach for the management of chronic LBP. Spine 2011;36:S10–S18.
23. Hale ME, Ahdieh H, Ma T, et al. Efficacy and safety of OPANA ER (oxymorphone extended release) for relief of moderate to severe chronic low back pain
in opioid-experienced patients: a 12-week, randomized, double-blind, placebo-controlled study. J Pain 2007;8:175–84.
24. Katz N, Rauck R, Ahdieh H, et al. A 12-week, randomized, placebo-controlled trial assessing the safety and efficacy of oxymorphone extended release for opioid-naive patients with chronic low back pain
. Curr Med Res Opin 2007;23:117–28.
25. Vondrackova D, Leyendecker P, Meissner W, et al. Analgesic efficacy and safety of oxycodone in combination with naloxone as prolonged release tablets in patients with moderate to severe chronic pain. J Pain 2008;9:1144–54.
26. Aoki T, Kuroki Y, Kageyama T, et al. Multicentre double-blind comparison of piroxicam and indomethacin in the treatment of lumbar diseases. Eur J Rheumatol Inflamm 1983;6:247–52.
27. Zerbini C, Ozturk ZE, Grifka J, et al. Efficacy of etoricoxib 60 mg/day and diclofenac 150 mg/day in reduction of pain and disability in patients with chronic low back pain
: results of a 4-week, multinational, randomized, double-blind study. Curr Med Res Opin 2005;21:2037–49.
28. Evans DP, Burke MS, Newcombe RG. Medicines of choice in low back pain
. Curr Med Res Opin 1980;6:540–7.
29. Jenkins DG, Ebbutt AF, Evans CD. Tofranil in the treatment of low back pain
. J Int Med Res 1976;4:28–40.
30. Bohnert AS, Valenstein M, Bair MJ, et al. Association between opioid prescribing patterns and opioid overdose-related deaths. JAMA 2011;305:1315–21.
31. Benyamin R, Trescot AM, Datta S, et al. Opioid complications and side effects. Pain Phys 2008;11:S105–20.
32. Rainsford KD. Profile and mechanisms of gastrointestinal and other side effects of nonsteroidal anti-inflammatory drugs
(NSAIDs). Am J Med 1999;107:27S–35S; discussion 35S–6S.
33. Huszonek JJ, Dewan MJ, Koss M, et al. Antidepressant side effects and physician prescribing patterns. Ann Clin Psychiatry 1993;5:7–11.
antidepressants; drugs; heterogeneity of treatment effect; low back pain; nonsteroidal anti-inflammatories; opiates; randomized trials; systematic review
Supplemental Digital Content
© 2011 Lippincott Williams & Wilkins, Inc.