Sample size
An estimate of sample size was based on the assumption that the frequency of caesarean section for genital herpes would be 25% in the placebo group and that this could be reduced to 5% in the treatment group. With a 90% probability of detecting this difference significant at the 5% level, the trial would require 120 women to participate. All data were entered into an electronic database. The results presented are an intention-to-treat analysis using the SPSS statistical computer package, with the measure of effect expressed as an odds ratio (OR) with 95% confidence intervals (95% CI) for the incidence of the outcome measures in acyclovir recipients compared with placebo recipients. The adjusted odds ratios control for the effect of centre on those data which are pooled from both centres.