Objectives: Older adults are particularly vulnerable to the negative consequences of antipsychotic exposure and are disproportionally affected by higher mortality from coronavirus disease 2019 (COVID-19). Our goal was to determine whether concurrent antipsychotic medication use was associated with increased COVID-19 mortality in older patients with preexisting behavioral health problems. We also report on findings from post-COVID follow-ups.
Design: Retrospective observational study.
Participants: Outpatients at a geriatric psychiatric clinic in New York City.
Measurements: Demographic and clinical data including medication, diagnosis and Clinical Global Impression Severity (CGI-S) scales on outpatients who had COVID-19 between February 28th and October 1st 2020 were extracted from the electronic health records (EHR) from the hospital.
Results: A total of 56 patients were diagnosed with COVID-19 (mean age 76 years; median age 75 years) and 13 (23.2%) died. We found an increased mortality risk for patients who were prescribed at least one antipsychotic medication at the time of COVID-19 infection (Fisher's exact test P = 0.009, OR = 11.1, 95% confidence interval: 1.4-96.0). This result remains significant after adjusting for age, gender, housing context and dementia (Logistic regression P = 0.035, Beta = 2.4). Furthermore, we found that most patients who survived COVID-19 (88.4%) recovered to pre-COVID baseline in terms of psychiatric symptoms. Comparison of pre- and post-COVID assessments of CGI-S for 33 patients who recovered from COVID-19 were not significantly different.
Conclusion: We observed a higher COVID-19 mortality associated with concurrent antipsychotics use in older patients receiving behavioral health services. The majority of patients in our geriatric clinic who recovered from COVID-19 appeared to return to their pre-COVID psychiatric function. More precise estimates of the risk associated with antipsychotic treatment in older patients with COVID-19 and other underlying factors will come from larger datasets and meta-analyses.