Abdominal infection: 6 studies (n=312).
Pelvic infection: 17 studies (n=819).
Total: 23 studies (n=1,131).
ABDOMINAL INFECTION
Clindamycin, 600 mg, 8 hourly dosage:
mean weighted cure rate (cured patients only) 76% (95% confidence interval, CI: 59, 92%, p=0.03);
mean weighted success rate (cured plus improved patients) 90% (95% CI: 81, 99%, p=0.28).
Clindamycin, 900 mg, 8 hourly dosage:
mean weighted cure rate 91% (95% CI: 85, 96%, p=0.03);
mean weighted success rate 93% (95% CI: 88, 97%, p=0.28).
Sensitivity analysis:
Only one study (600 mg regimen) shared the same definition of cure with the review; this demonstrated a cure rate of 87% and a success rate of 93%.
PELVIC INFECTION
Clindamycin, 600 mg, 8 hourly dosage:
mean weighted cure rate 83% (95% CI: 73, 93%, p=0.51);
mean weighted success rate 87% (95% CI: 80, 94%, p=0.51).
Clindamycin, 900 mg, 8 hourly dosage:
mean weighted cure rate 89% (95% CI: 86, 93%, p=0.51);
mean weighted success rate 90% (95% CI: 86, 94%, p=0.51).
Sensitivity analysis:
All studies of the 600 mg regimen used cure definitions similar to that of the review, whereas one study of the 900 mg regimen used a dissimilar definition. When this study was excluded from the analysis, both the cure and success rates were 90%.
Homogeneity of studies:
In most cases, studies were not significantly different from one another, thus permitting the combining of results. The exceptions were studies of the 900 mg dosage in pelvic infection. The results of the sensitivity analysis suggested that differences between studies were likely to have been caused by random error rather than systematic differences.