BACKGROUND: LBP is a major health problem that causes significant medical expense and much absenteeism and disability. The disease is usually self-limited, but the pain can be severe. NSAIDs are widely prescribed for LBP and are recommended in several back pain management guidelines. Their effectiveness has not been proven beyond a doubt, so the authors of this systematic review attempted to synthesize the world literature on the subject.
POPULATION STUDIED: Fifty-one studies enrolled subjects between the ages of 18 and 65 years who had LBP with or without sciatica. Subjects with specific LBP caused by infection, neoplasm, osteoporosis, fractures, or rheumatoid arthritis were excluded. Acute ( 12 weeks) and chronic (>12 weeks) LBP patients were included.
STUDY DESIGN AND VALIDITY: This was a systematic review in which the authors evaluated randomized double-blinded controlled trials. The studies were identified by searching MEDLINE (1966 to 1998), EMBASE (1988 to 1998), and the Cochrane Controlled Trials Register (issue 3, 1998). References in identified studies were also screened. The quality assessment was based on the presence or absence of 11 criteria, including randomization, allocation concealment, blinding, intention-to-treat analysis, and follow-up. Quantitative analysis was appropriately limited to clinically homogeneous studies. Separate analyses were performed for the primary outcome measures of pain intensity, overall improvement, functional status, and return to work. Qualitative analysis was performed if the studies were clinically heterogeneous (ie, if they were too different to combine) or if the data required for statistical pooling were lacking.
OUTCOMES MEASURED: The primary outcome measures in hierarchical order were: pain intensity using a visual analog scale or a numerical rating scale, a global measure such as overall improvement or proportion of patients recovered, back pain specific functional status, and return to work.
RESULTS: In acute LBP there was clear statistical evidence of slight global short-term improvement with NSAIDs compared with placebo, without any statistically significant difference in side effects. There was not enough information to determine the effectiveness of NSAIDs in chronic LBP. NSAIDs were slightly more effective than acetaminophen for both acute and chronic LBP. NSAIDs were no better than muscle relaxants, narcotics, physiotherapy, or spinal manipulation for acute LBP; they were, however, somewhat better than bed rest. There was no difference between types of NSAIDs. There was no advantage in adding muscle relaxants to NSAIDs for acute LBP, and the addition of B vitamins to NSAIDs was supported by very limited evidence.
On the basis of the patient-oriented outcomes from this review, it is reasonable to treat acute or chronic LBP with NSAIDs. All NSAIDs are equally effective and have minimal side effects, so generic ibuprofen is probably the best choice (fewer serious side effects and lower cost). Acetaminophen is a reasonable, though slightly less effective, alternative.