|
Systematic review: opioid treatment for chronic back pain: prevalence, efficacy, and association with addiction |
Martell B A, O'Connor P G, Kerns R D, Becker W C, Morales K H, Kosten T R, Fiellin D A |
|
|
CRD summary This review concluded that opioids may be efficacious for short-term pain relief amongst people with chronic back pain, but long-term efficacy beyond 16 weeks is unclear. Substance use disorders are common in patients taking opioids for back pain. The authors' conclusions did not really take the limitations of the data available into account, particularly in relation to substance use disorders.
Authors' objectives The review had three aims: to determine the prevalence of opioid treatment for patients with chronic back pain; to assess whether opioid treatment is effective in this population group; and to determine the prevalence of substance use disorders associated with opioid treatment in this population group. The second and third aims will be addressed in this abstract.
Searching MEDLINE, EMBASE and PsycINFO (all up to 2005) and the Cochrane CENTRAL Register (Issue 4, 2004) were searched for studies reported in the English language; the search terms were provided. Selected reviews and bibliographies were searched for additional studies and leading pain experts were contacted.
Study selection Study designs of evaluations included in the reviewInclusion criteria for the type of study design were not specified. Only studies on the efficacy of opioids with a quality score of at least 10 were given detailed consideration. The duration of the studies ranged from 7 days to 16 weeks.
Specific interventions included in the reviewStudies of oral, topical or transdermal opioid preparations were eligible for inclusion. The included effectiveness studies compared opioids with nonopioids or placebo, or compared different opioids. The type of opioid treatment varied between studies, as did the use of rescue medications. The average opioid dose in the placebo and nonopioid controlled trials was 73 mg/day (range: 30 to 232).
Participants included in the reviewStudies of adults (18 years and over) with chronic back pain lasting for at least 3 months, who were not obstetric patients and had no pre-existing diagnosis of opioid dependence, were eligible for inclusion. Pregnant women were excluded. The type of low back pain varied between the included efficacy studies: there was nociceptive and mechanical back pain with and without radiculopathy. The studies were of opioid-naive and opioid-experienced populations.
Outcomes assessed in the reviewOutcome inclusion criteria were not specified for the effectiveness studies, though pain was the outcome of interest. The included studies reported pain outcomes using a range of different outcome measures. Quality of life was also reported. Studies investigating the prevalence of addictive behaviours were required to assess aberrant medication-taking behaviours, any type of substance use disorder concurrent with treatment for low back pain, or the lifetime prevalence of substance abuse disorders that may have preceded treatment.
How were decisions on the relevance of primary studies made?Two reviewers assessed studies for relevance. Any disagreements were resolved through discussion and consensus.
Assessment of study quality Clinical trials and observational studies were assessed using different standardised tools. For observational studies a score of 12 or greater was considered excellent quality, while for clinical trials a score of 10 or greater was considered excellent. Two reviewers independently assessed study quality. Any differences were resolved through discussion and consensus.
Data extraction For the efficacy studies, one outcome only was extracted for each study, with preference given to the outcome measure recommended by the IMMPACT (Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials) statement. A standardised mean difference with 95% confidence interval (CI) was calculated because of the different methods of assessing pain (Hedges g for continuous pain measures and the Cox approach for binary pain outcomes). For the studies where the comparator was placebo or a nonopioid, the mean difference between treatments was calculated. For studies where the comparator was another opioid, the change from baseline was calculated. Where the two opioid groups in a study had a similar estimate, they were pooled. The authors did not state how the data were extracted.
Methods of synthesis How were the studies combined?The efficacy studies were pooled in a fixed-effect meta-analysis (or random-effects where there was evidence of statistical heterogeneity). There was a narrative synthesis of studies on the prevalence of substance misuse disorders.
How were differences between studies investigated?Studies comparing opioids with placebo or nonopioids and those comparing two or more opioids were pooled separately. Statistical heterogeneity was investigated using the I-squared statistic. Meta-regression was used to investigate the individual impact of study design, continuous or binary outcome measure, sample size, type of comparator (placebo or active), quality and study duration on the results. A sensitivity analysis was also conducted; this explored the effect of removing some specific studies.
Results of the review Fifteen studies (n=1,008) assessing the efficacy of opioids were included: 10 randomised controlled trials, 1 non-randomised trial, 2 observational studies, and 2 described as open label (although it was unclear whether they were of an observational design). Nine studies (n=1,119) on the prevalence of substance misuse disorders were included.
Treatment efficacy.
There were 6 trials available for this comparison; all were rated as excellent quality. Two were excluded from the meta-analysis as appropriate data were not available. There was a reduction in pain with opioid treatment compared with placebo or nonopioid (4 studies), though this was not statistically significant (standardised mean difference -0.199, 95% CI: -0.49, 0.11). There was no evidence of statistical heterogeneity. The subgroup analysis did not indicate any variation in the treatment effect.
There were 9 trials available where data were taken from studies comparing two or more opioids; all were rated as excellent quality. Four were excluded from the meta-analysis because appropriate data were not available. Compared with baseline, there was a reduction in pain with opioid treatment (5 studies), though this was not statistically significant (standardised mean difference -0.93, 95% CI: -1.89, -0.03). There was no evidence of statistical heterogeneity. The subgroup analysis did not indicate any variation in the treatment effect.
Prevalence of substance use disorders.
This group of studies were generally of a poor quality (mean score 11 out of 27). The prevalence of substance use disorders in chronic back pain patients receiving opioids ranged from 3 to 45% (4 studies). In the highest quality study the prevalence of such disorders was similar in patients using and not using opioids. The prevalence of aberrant medication-taking behaviours ranged from 5 to 24% (5 studies).
Authors' conclusions Opioids may be efficacious for short-term pain relief, but long-term efficacy from 16 weeks onwards is unclear. Substance use disorders are common in patients taking opioids for back pain and aberrant medication-taking behaviours occur in up to 24% of cases.
CRD commentary The review question was generally clear although inclusion criteria for the outcomes and study design were not specified. A number of relevant databases were searched, but relevant studies might have been missed as there were no specific attempts to locate unpublished studies and non-English language studies were excluded. Appropriate methods were used to minimise error and bias in the study selection and quality assessment processes; it was unclear whether similar methods were used to extract the study data. It seemed inappropriate to pool the studies though, as the authors pointed out, that there was considerable clinical heterogeneity. One of the meta-analyses was limited by the fact that it included only one treatment arm, i.e. the data was uncontrolled. The meta-analysis of opioids versus placebo was based on a fairly small number of participants (less than 500 evaluable patients). In addition, given the poor quality of the studies on the prevalence of substance use disorders, the findings of these studies should be treated with caution. The authors' conclusions did not really take the limitations of the available data into account, particularly in relation to substance use disorders.
Implications of the review for practice and research Practice: The authors stated that clinicians should reconsider treating chronic back pain with opioids and consider other treatments with similar benefit but fewer long-term adverse events.
Research: Well-conducted studies on the long-term effects of opioids, as well as an investigation of medications combined with treatments such as physical therapy and medications targeted at specific symptoms such as neuropathic pain, are required. The use of standardised pain outcome measures and a placebo that can mimic the common side-effects of the opioid being investigated would be useful. High-quality cohort studies and clinical trials assessing the long-term risk of substance abuse using validated instruments are also needed.
Funding Robert Wood Johnson Foundation; National Institute on Drug Abuse; National Institute on Alcohol Abuse and Alcoholism.
Bibliographic details Martell B A, O'Connor P G, Kerns R D, Becker W C, Morales K H, Kosten T R, Fiellin D A. Systematic review: opioid treatment for chronic back pain: prevalence, efficacy, and association with addiction. Annals of Internal Medicine 2007; 146(2): 116-127 Indexing Status Subject indexing assigned by NLM MeSH Analgesics, Opioid /therapeutic use; Chronic Disease; Clinical Trials as Topic /standards; Drug Prescriptions /statistics & Humans; Low Back Pain /drug therapy /epidemiology; Prevalence; Substance-Related Disorders /epidemiology; United States /epidemiology; numerical data AccessionNumber 12007008009 Date bibliographic record published 31/05/2007 Date abstract record published 31/05/2007 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|