To compare the accumulation of hospital days, a proposed proxy for overall drug safety, between antipsychotic initiators and noninitiators with Alzheimer's disease (AD).Nationwide exposure-matched cohort.Finnish community dwellers who received an incident AD diagnosis in 2005‒2011 (n = 70,718). For each antipsychotic initiator, 1 noninitiator was matched on age, sex, and time since AD diagnosis (n = 19,909 matched pairs).Accumulation of hospital days was measured during a 2-year follow-up from the national hospital discharge register. Antipsychotic use was ascertained from the National Prescription Register. Association between antipsychotic initiation and accumulation of hospital days was analyzed using negative binomial model.During the 2-year follow-up, antipsychotic initiators were hospitalized on average for 52.5 (standard deviation 97.7) days and matched noninitiators for 34.7 (standard deviation 72.4) days. Of antipsychotic initiators 23.8% and of noninitiators, 34.1% did not have any hospital days. Antipsychotic initiators had 53% more hospital days (adjusted incidence rate ratio 1.53; 95% confidence interval 1.47‒1.59) than noninitiators. Strongest associations were observed during the first 6 months. Antipsychotic initiators had more hospital days with primary diagnosis codes of dementia; mental and behavioral disorders; factors influencing health status; diseases of the respiratory, genitourinary, and circulatory system; certain infectious and parasitic diseases; and symptoms not elsewhere classified, than noninitiators.Antipsychotic initiators accumulated more hospital days than noninitiators, especially within the first 6 months after initiation. This may indicate adverse events or difficulties in treating the most severe behavioral and psychological symptoms of dementia and health problems triggering them. After initiating antipsychotics, careful and regular monitoring is needed to assess response and decrease the risk of adverse effects and events.