Fig. 2.

Primary outcome: time to cold sore recurrence by treatment arm, and in subgroup analyses. Forest plots show the subgroup analyses of the primary outcome, time to cold sore recurrence and 95% confidence intervals. ∗p-value for subgroup x treatment arm interaction, adjusted for randomisation strata (age group, presence of comorbidity, geographical location, and study stage). In stage 1, participants randomised to the control group received no intervention; in stage 2, participants randomised to the control group received placebo saline intradermal injection. †Few participants prevented the estimation of the treatment effect in this stratum.