Abstract
This cohort study investigates whether an association exists between antipsychotic use and mortality in patients with serious mental illness who are diagnosed with COVID-19.
Schizophrenia spectrum disorders are associated with increased mortality in the setting of COVID-19 infection.1,2,3 Among several possible explanations for this increased risk is the role of adverse effects of antipsychotic medication, which has not been systematically examined in this population. Our goal was to investigate whether antecedent antipsychotic use was associated with mortality among patients with serious mental illness diagnosed with COVID-19.
Methods
We conducted a retrospective cohort study using the New York University Langone Health electronic health record system. Adults diagnosed with COVID-19 infection between March 3, 2020, and February 17, 2021, who had a preexisting diagnosis of schizophrenia, schizoaffective disorder, or bipolar disorder (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes F20, F25, F30, or F31), were included. Individuals with schizophrenia or schizoaffective disorder were assigned to the schizophrenia spectrum disorder group; those with more than 1 diagnosis were categorized hierarchically, as reported previously.1,4 The exposure of interest was antipsychotic use at COVID-19 diagnosis. Electronic health records were reviewed to verify the accuracy of prescriptions; patients with antipsychotic discontinuation or nonadherence were assigned to the unexposed group. The primary end point was death within 60 days of COVID-19 diagnosis. The following covariates were considered based on their known or hypothesized association with the outcome and potential to confound the association of interest: sociodemographic characteristics, including patient-reported race (Asian, Black, White, or other race [including Native American, multiple races, and any race not included in a list of more than 40 races], and unknown race) and ethnicity (Hispanic [Chicano/Chicana, Costa Rican, Cuban, Dominican, Guatemalan, Hispanic/Latino, Honduran, Mexican, Mexican American, Nicaraguan, Salvadoran, South American, Spaniard] or non-Hispanic), age, and insurance type; psychiatric diagnosis; medical comorbidities, including body mass index; and smoking status (Table 1). The study was approved by the institutional review board of the New York University Grossman School of Medicine, with a waiver of informed consent granted based on the determination that there was no more than minimal risk to participants. The final logistic regression model assessed for an association between antipsychotic exposure and mortality using odds ratios and 95% CIs, adjusting for covariates that differed between groups with 2-tailed testing. Covariates of borderline statistical significance (P < .10) were included in the final models because of the limited sample size. Analyses were conducted using SAS statistical software, version 9.4 (SAS Institute Inc).
Table 1. Characteristics of Study Population by Antipsychotic Use.
| Characteristic | Antipsychotic use, No. (%) (N = 464) | P valuea | |
|---|---|---|---|
| Yes | No | ||
| No. of patients | 196 (42.2) | 268 (57.8) | |
| Age, median (IQR), y | 59 (39.5-68.5) | 56 (38-66) | .10 |
| Sex | |||
| Female | 98 (50.0) | 127 (47.4) | .58 |
| Male | 98 (50.0) | 141 (52.6) | |
| Race | |||
| Asian | 7 (3.6) | 6 (2.2) | .54 |
| Black | 31 (15.8) | 51 (19.0) | |
| White | 116 (59.2) | 145 (54.1) | |
| Otherb | 28 (14.3) | 49 (18.3) | |
| Unknown | 14 (7.1) | 17 (6.3) | |
| Ethnicity | |||
| Hispanicc | 33 (16.8) | 45 (16.8) | >.99 |
| Non-Hispanic | 163 (83.2) | 223 (83.2) | |
| Smoking status | |||
| Current | 28 (14.3) | 47 (17.5) | .79 |
| Former | 50 (25.5) | 64 (23.9) | |
| Never | 106 (54.1) | 143 (53.4) | |
| Missing data | 12 (6.1) | 14 (5.2) | |
| Insurance type | |||
| Private | 38 (19.4) | 78 (29.1) | .05 |
| Public | 144 (73.5) | 171 (63.8) | |
| Self-pay | 14 (7.1) | 19 (7.1) | |
| Neighborhood poverty level, (below FPL)d | |||
| Very high (≥30%) | 11 (5.6) | 12 (4.5) | .56 |
| High (≥20%) | 28 (14.3) | 47 (17.5) | |
| Medium (≥10%) | 73 (37.2) | 87 (32.5) | |
| Low (<10%) | 83 (42.4) | 122 (45.5) | |
| Psychiatric diagnosis | |||
| Schizophrenia spectrum disordere | 91 (46.4) | 91 (34.0) | .007 |
| Bipolar disorder | 105 (53.6) | 177 (66.0) | |
| BMI, median (IQR) | 29.8 (25.8-34.0) | 28.2 (24.3-32.4) | .02 |
| Hypertension | 118 (60.2) | 142 (53) | .12 |
| Myocardial infarction | 17 (8.7) | 26 (9.7) | .71 |
| Diabetes | 72 (36.7) | 80 (29.9) | .12 |
| Chronic kidney disease | 63 (32.1) | 80 (29.9) | .60 |
| Obstructive lung disease | 72 (36.7) | 97 (36.2) | .91 |
| Inpatient visits, median (IQR), No. | 1 (0-2) | 1 (0-2) | .12 |
| Ambulatory visits, median (IQR), No. | 10 (2-28) | 10 (1-29) | .48 |
| Duration, median (IQR), y | 3.4 (1.0-4.8) | 3.5 (1.2-5.6) | .16 |
| Inpatient | 101 (51.5) | 118 (44) | .11 |
| Died | 21 (10.7) | 20 (7.5) | .22 |
Abbreviations: IQR, interquartile range; FPL, federal poverty level; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); FPL, federal poverty level; IQR, interquartile range.
Reported P values for χ2 test or Fisher exact test for categorical variables and t test for continuous variables.
Other included Native American, multiple races, and other race by patient report.
Hispanic ethnicity included Chicano/Chicana, Costa Rican, Cuban, Dominican, Guatemalan, Hispanic/Latino, Honduran, Mexican, Mexican American, Nicaraguan, Salvadoran, South American, Spaniard.
Neighborhood poverty level was missing for 1 subject in the antipsychotic treatment group.
Includes diagnoses of schizophrenia and schizoaffective disorder.
Results
A total of 464 patients (mean [SD] age, 53 [17.1] years; 239 men [51.5%] and 225 women [48.5%]) were included, of which 196 (42.2%) were treated with antipsychotic medication. Forty-one patients (8.8%) died. The 60-day case fatality rate among patients with a schizophrenia spectrum disorder (n = 182) was 13.7%, and the case fatality rate among patients with bipolar disorder (n = 282) was 5.7%. Age, body mass index, insurance type, and psychiatric diagnosis differed between groups and were included in the fully adjusted model (Table 2). Antipsychotic treatment was not significantly associated with mortality (odds ratio, 1.00; 95% CI, 0.48-2.08; P = .99). However, a diagnosis of a schizophrenia spectrum disorder was associated with a near 3-fold increased risk of mortality compared with bipolar disorder (odds ratio, 2.88; 95% CI, 1.36-6.11; P = .01).
Table 2. Multivariable-Adjusted Risk Model for 60-Day Case Fatality.
| Variable | Odds ratio (95% CI) | P value |
|---|---|---|
| Antipsychotic use (yes) | 1.00 (0.48-2.08) | .99 |
| Insurance type | ||
| Private | 1.09 (0.10-12.02) | >.99 |
| Public | 1.11 (0.13-9.58) | |
| Self-pay | 1 [Reference] | |
| Age | 1.11 (1.07-1.15) | <.001 |
| Psychiatric diagnosis | ||
| Schizophrenia spectrum disordera | 2.88 (1.36-6.11) | .006 |
| Bipolar disorder | 1 [Reference] | |
| BMI | 0.99 (0.92-1.05) | .66 |
Abbreviation: BMI, body mass index.
Includes diagnoses of schizophrenia and schizoaffective disorder.
Discussion
In this cohort study of adults with serious mental illness diagnosed with COVID-19 infection in a New York City medical system, antipsychotic treatment was not associated with an increased risk of mortality.
A growing body of evidence has suggested that people with schizophrenia spectrum disorders may have an increased risk of fatal illness after COVID-19 infection, but the mechanism is not clear. An association between antipsychotic medication and increased risk of COVID-19 mortality has been reported in population-based studies,5,6 but these studies did not take psychiatric diagnosis into account. We did not observe an association between antipsychotic use and mortality in this cohort of adults with serious mental illness. Study limitations include the inability to validate psychiatric diagnoses and capture deaths that were not documented in the electronic health record. The limited sample size precluded analysis of individual antipsychotic medications, which may differ in their associated effects. Further research is needed to understand what underlies increased mortality risk in this population to address worsening health disparities.
References
- 1.Nemani K, Li C, Olfson M, et al. Association of psychiatric disorders with mortality among patients with COVID-19. JAMA Psychiatry. 2021;78(4):380-386. doi: 10.1001/jamapsychiatry.2020.4442 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bulletin S, Tzur Bitan D, Krieger I, et al. COVID-19 prevalence and mortality among schizophrenia patients: a large-scale retrospective cohort study. Schizophr Bull. Published online February 19, 2021. doi: 10.1093/schbul/sbab012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Fond G, Pauly V, Leone M, et al. Disparities in intensive care unit admission and mortality among patients with schizophrenia and COVID-19: a national cohort study. Schizophr Bull. 2021;47(3):624-634. doi: 10.1093/schbul/sbaa158 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Walker ER, McGee RE, Druss BG. Mortality in mental disorders and global disease burden implications: a systematic review and meta-analysis. JAMA Psychiatry. 2015;72(4):334-341. doi: 10.1001/jamapsychiatry.2014.2502 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Reilev M, Kristensen KB, Pottegård A, et al. Characteristics and predictors of hospitalization and death in the first 11 122 cases with a positive RT-PCR test for SARS-CoV-2 in Denmark: a nationwide cohort. Int J Epidemiol. 2020;49(5):1468-1481. doi: 10.1093/ije/dyaa140 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Poblador-Plou B, Carmona-Pírez J, Ioakeim-Skoufa I, et al. ; EpiChron Group . Baseline chronic comorbidity and mortality in laboratory-confirmed COVID-19 cases: results from the PRECOVID study in Spain. Int J Environ Res Public Health. 2020;17(14):1-14. doi: 10.3390/ijerph17145171 [DOI] [PMC free article] [PubMed] [Google Scholar]