In the last several years, FDA has used meta-analyses to address some important drug safety problems including antidepressants and suicidality, rosiglitazone and cardiovascular outcomes, and anti-epileptic drugs and suicidality. General meta-analysis principles and methodology are well established. However, in drug safety problems, special considerations arise. These considerations include, for example, the identification of events and the choice of appropriate methodology when events are sparse. Using recent FDA examples, this talk will both review the general meta-analysis principles and raise the special considerations of planning and executing a meta-analysis for drug safety problems.