Ankylosing spondylitis (AS) is a chronic inflammatory disease of unknown cause and belongs to a group of diseases known as spondyloarthropathies (SpA), which includes reactive arthritis, arthritis/spondylitis in inflammatory bowel disease, psoriatic arthritis/spondylitis and undifferentiated SpA. Non-steroidal anti-inflammatory drugs (NSAIDs) have been the main treatment for AS. For those refractory or intolerant to NSAIDs, the disease modifying antirheumatic drugs (DMARDs) have been used as a second line approach. Sulfasalazine (SSZ) is the best studied DMARD in AS, but its efficacy remains unclear.
To evaluate the efficacy and toxicity of sulfasalazine for the treatment of ankylosing spondylitis.
Relevant randomised and quasi-randomised trials in any language were sought using the following sources: CENTRAL (Cochrane Central Register of Controlled Trials, Issue 2, 2003), MEDLINE (1966 to June Week 4 2003), EMBASE (1980 to 2003 Week 26), CINAHL (1982 to June Week 3 2003) and the reference section of retrieved articles.
We evaluated randomised and quasi-randomised trials examining the efficacy of sulfasalazine on ankylosing spondylitis.
Unblinded trial reports were reviewed independently by two reviewers according to the selection criteria. Disagreements on the inclusion of the studies were resolved, where necessary, by recourse to a third reviewer. The methodological quality of included trials were independently assessed by the same reviewers on randomization, concealment, blindness (participants, care providers and outcome investigators), description of withdrawals and drop-outs and intention-to-treat analysis. The same reviewers independently entered the data extracted from the included trials, using RevMan double entry facility. Results were combined using weighted mean difference or standardised mean difference for continuous data, and relative risk for dichotomous data.
Twelve studies met the inclusion criteria but only eleven were included in the data analysis. The pooled analysis showed that the difference between intervention groups was significant only in erythrocyte sedimentation rate (ESR) (WMD -4.79, 95% CI -8.80 to -0.78) mm/h) and morning stiffness VAS-100 mm (visual analogue scale 100 mm, where 0 = no stiffness and 100 = severe) (WMD -13.89, 95% CI -22.54 to -5.24), favouring SSZ over placebo. The trial with the largest sample (Clegg 1996) and that with the longest treatment duration (Kirwan 1993) had similar results. Both trials found that SSZ showed evidence of benefit in the occurrence of peripheral joint symptoms and peripheral responses in patients with peripheral arthritis. Nissila 1988 is the only trial in which SSZ showed benefit in primary outcome analyses, including back pain, chest expansion, occiput-to-wall test and patient's general well being. Compared with other trials, the patients in this trial had the shortest disease duration and the highest level of baseline ESR and contained the greatest proportion of patients with peripheral arthritis. Significantly more withdrawals for side effects (RR 1.50, 95% CI 1.04 to 2.15, NNH 23, 95% CI 10 to 288) and for any reason (RR 1.33, 95% CI 1.03 to 1.73, NNH 17, 95% CI 8 to 180) were found in SSZ compared with placebo group although severe side effects were rare (1 of the 469 patients taking SSZ).
Across all AS patients, SSZ demonstrated some benefit in reducing ESR and easing morning stiffness, but no evidence of benefit in physical function, pain, spinal mobility, enthesitis, patient and physician global assessment. Patients at early disease stage, with higher level of ESR (or active disease) and peripheral arthritis might benefit from SSZ.