Background: The question of what potential benefits and harms are associated with certain dietary regimes used in rheumatoid arthritis is an important one for many patients and health care providers.
Objectives: To assess the effectiveness and safety of dietary interventions in the treatment of rheumatoid arthritis.
Search strategy: We searched the Cochrane Central Register of Controlled Trials (CENTRAL)(The Cochrane Library, issue 1 2008), MEDLINE, EMBASE, AMED, CINAHL and reference lists of relevant articles (up to January 2008), and contacted authors of included articles.
Selection criteria: Randomised controlled trials (RCTs) or controlled clinical trials (CCTs) where the effectiveness of dietary manipulation was evaluated. Dietary supplement studies (including fish oil supplements) were not included.
Data collection and analysis: Two authors independently selected trials for inclusion, assessed the internal validity of included trials and extracted data. Investigators were contacted to obtain missing information.
Main results: Fourteen RCTs and one CCT, with a total of 837 patients, were included. Due to heterogeneity of interventions and outcomes, baseline imbalance and inadequate data reporting, no overall effects were calculated. A single trial with a moderate risk of bias found that fasting, followed by 13 months on a vegetarian diet, may reduce pain (mean difference (MD) on a 0 to 10 scale -1.89, 95% confidence interval (CI) -3.62 to -0.16), but not physical function or morning stiffness immediately after intervention. Another single trial with a moderate risk of bias found that a 12-week Cretan Mediterranean diet may reduce pain (MD on a 0 to 100 scale -14.00, 95% CI -23.6 to -4.37), but not physical function or morning stiffness immediately after intervention. Two trials compared a 4-week elemental diet with an ordinary diet and reported no significant differences in pain, function or stiffness. Due to inadequate data reporting, the effects of vegan and elimination diets are uncertain. When comparing any dietary manipulation with an ordinary diet we found a significantly higher total drop-out of 10% (risk difference (RD) 0.10, 95% CI 0.02 to 0.18), higher treatment-related drop-out of 5% (RD 0.05, 95% CI -0.03 to 0.14) and a significantly higher weight loss (weighted mean difference -3.23, 95% CI -4.79 to -1.67 kg) in the diet groups compared to the control groups.
Authors' conclusions: The effects of dietary manipulation, including vegetarian, Mediterranean, elemental and elimination diets, on rheumatoid arthritis are still uncertain due to the included studies being small, single trials with moderate to high risk of bias. Higher drop-out rates and weight loss in the groups with dietary manipulation indicate that potential adverse effects should not be ignored.