Background: We included both parallel and cross-over studies in our recent systematic review on pharmacological and pacemaker treatment for neurally mediated reflex syncope, the most common cause of transient loss of consciousness. When calculating treatment effects, we wanted to take into account that patients included in a parallel randomized controlled trial receive either intervention or control treatment, while patients included in a randomized cross-over trial receive both kinds of treatment. In Review Manager 5 (RevMan) there are no tools available to take into account differences between randomized controlled trials and randomized cross-over trials. Methods: Using Microsoft Excel, we combined the results of parallel and cross-over studies for both continuous and binary outcomes in two steps: we first calculated treatment effects and respective variances for parallel randomized controlled trials and randomized cross-over studies separately and then calculated combined treatment effects using weights determined by the variances of these two types of studies. Next, we imported the log (odds ratios) and respective standard errors in RevMan using the generic inverse variance method to compose our final review. In a previous concept version of our review in RevMan we had calculated treatment effects irrespective of the design of included studies for all available outcome measures in studies comparing beta-blocker and placebo treatment. We used these calculations as a reference. Results: Treatment effects of placebo-controlled beta-blocker studies were highly different when taking differences between parallel randomized controlled trials and randomized cross-over trials into account. After calculation of combined treatment effects taking into account differences in study design, we observed no consistent significant differences in studies investigating placebo-controlled pharmacological and/or pacemaker treatment for neurally mediated reflex syncope. Conclusions: Treatment effects can be very different when taking the design of included studies into account.