Four relevant RCTs were identified (n=665, range 124 to 193). All studies were double-blind, used validated methods for detecting C. difficile and had similar predefinitions of diarrhoea (although none described validated methods of measuring stool consistency). Three studies had sufficient follow-up and three studies performed an intention-to-treat analysis. All four RCTs reported withdrawal rates, but none compared these rates in the treatment and placebo groups.
Prevention of C. difficile recurrence (two RCTs): One RCT reported that S. boulardii significantly reduced recurrence of infection (RR 0.19 and OR 0.44, 95% CI 0.21 to 0.94) and one RCT found no significant effect.
One study performed subgroup analyses for the different antibiotics used, none of which was significant. But there was a trend towards reduction in C. difficile infection relapse in the recurrent treatment group of patients who received high-dose vancomycin in addition to S. boulardii (RR 0.33 and OR 0.20, 95% CI 0.04 to 1.01). The other study performed a subgroup analysis for patients with a history of recurrent C. difficile infection and found that S. boulardii significantly reduced the recurrence of infection (RR 0.30 and OR 0.29, 95% CI 0.10 to 0.84).
Adverse events: One RCT reported that there were no significant differences between intervention and control groups for overall adverse events and those related specifically to S. boulardii treatment. The other RCT reported greater frequency of thirst and constipation in the S. boulardii group compared to controls (significance not reported).
Prevention of C. difficile infection (two RCTs): Both studies performed subgroup analyses for patients with C. difficile infection if they had both clinical symptoms of diarrhoea and positive C. difficile toxin assays. Neither study found S. boulardii had a significant effect on reducing antibiotic-associated diarrhoea caused by C. difficile infection.
Adverse events: One RCT reported no side effects in either the placebo or treatment group. The other RCT found a statistically significantly higher level of intestinal gas in the S. boulardii group (7.4%) than in the placebo group (0%) among the 96% of patients who completed the adverse reaction forms.