OBJECTIVES: Antipsychotic medications are extensively used in nursing homes for management of behavioral and psychiatric disorders in the elderly. Prior research suggests that pneumonia is one of the common causes of antipsychotic-related mortality in this population. None of the studies compared typical and atypical antipsychotics with respect to pneumonia. This study examined the risk of pneumonia with use of typical versus atypical antipsychotics in dual eligible elderly nursing home residents. METHODS: The study involved a retrospective cohort design matched on propensity score using Medicare and Medicaid Analytical eXtract (MAX) data from four US states. The study population included elderly dual eligible (Medicaid and Medicare) nursing home residents (aged > 65 years) who initiated antipsychotics anytime during July 1, 2001 and December 31, 2003. The risk of pneumonia during the 6-month follow-up period was modeled using Cox proportional model and extended Cox hazard model stratified on matched pairs based on propensity scores, using atypical agents as the reference category. RESULTS: Analysis of Medicaid-Medicare data revealed that there were 49,904 antipsychotic (46,293 atypical and 3,611 typical) users in the unmatched cohort and 7,218 (3,609 atypical and 3,609 typical) users in the matched cohort. The unadjusted rate of pneumonia was 8.17% (295) for atypical users and 5.21% (188) for typical users. The results of Cox regression [average Hazards Ratio, HR, 1.24; 95% CI, 0.94 -1.64] and extended regression [<50 days: HR, 1.17; 0.83-1.66 and 50-180 days: HR, 1.36; 0.87-2.14] suggest that, there was no difference in risk of pneumonia among typical and atypical users. CONCLUSIONS: The study found no differential risk of pneumonia among typical versus atypical antipsychotic use in dual eligible nursing home residents. Given the differential risk of mortality with typical and atypical use in nursing homes, more research is needed to evaluate other contributory factors of mortality with respect to these two antipsychotic classes.