Sensitivity analyses, as far as I know, is performed manually in RevMan. You set the sensitivity analysis conditions manually, identify the analyzed studies to be included/excluded in every condition manually, and tick/untick the box to see the changes in the forest plot and note them manually. However, let us wait for further responses to your question if what I said may not be the case.
You need to select the studies to be included in the sensitivity analysis e.g. If you have to exclude low quality studies from the total number of studies included earlier in analysis. It depends on the purpose of sensitivity analysis, else it is the same method by just including/excluding the eligible studies for that particular outcome