key: cord-0897347-p0g9g3rn
authors: Boland, Xavier; Dratcu, Luiz
title: Antipsychotics and COVID-19: the debate goes on
date: 2021-11-18
journal: Lancet Psychiatry
DOI: 10.1016/s2215-0366(21)00396-5
sha: de0cf094f0625971c0d71b82dfedd685939492dd
doc_id: 897347
cord_uid: p0g9g3rn

nan

www.thelancet.com/psychiatry Vol 8 December 2021

Antipsychotics and COVID-19: the debate goes on

In a meta-analysis on the asso ciation of mental illness, its treatment, and COVID-19 risk, Benedetta Vai and colleagues reported a strong association between exposure to antipsychotics and COVID-19 mortality (odds ratio 3·71 [1·74-7·91]). 1 This has serious implications for the care of patients with severe mental illness receiving or about to receive antipsychotics during the pandemic. It also raises questions as to whether the use of electronic health records might have limited the scope of the analysis. Firstly, in all three studies reviewed in the analysis in which antipsychotic use was an exposure variable, patients on antipsychotics were included independently of diagnoses. Vai and colleagues acknowledged that unaccounted-for confounding factors might then have affected their analysis, as patient groups with disparate clinical profiles were included in single cohorts. For example, off-label prescription of antipsychotics in older patients with dementia or delirium is commonplace and a marker of frailty, itself a major risk factor for poor outcomes in COVID-19. 2 Yet two of the studies reported a median age of 84·2 years and 82 years, respectively, among patients who died of COVID-19. Although Vai and colleagues did conduct a sensitivity analysis adjus t ing for minimum age of the recruited cohorts, the outcomes they encountered are unlikely to represent outcomes in most patients with severe mental illness, the majority of whom are of working age and physically able.

Secondly, all three studies considered antipsychotics as a single homogenous pharmacological group. However, individual drugs can differ widely in their pharmacological profile. This is reflected in the differences in all-cause mortality rates in patients receiving different classes of antipsychotics, regardless of COVID-19 status. 3 Moreover, there is emerging evidence that the anti-inflammatory properties of some antipsychotics could in fact protect from severe COVID-19 by hindering the cytokine storms implicated in its pathogenesis. 4 Lastly, none of the studies tested adherence to treatment. However, non-adherence to antipsychotics is common and a major contributor to treatment failure. The risk of complicated COVID-19 associated with untreated psychosis might far outweigh any risk associated with the proper use of antipsychotics that are recommended for its treatment. There is evidence that, when adherence to antipsychotic treatment is ensured, patients with severe mental illness could actually have better COVID-19 outcomes than the general population. 5 Vai and colleagues highlight risks associated with the use of antipsychotic drugs in the context of the COVID-19 pandemic. However, unless specific criteria are used, scrutiny of data from electronic health records might not always reflect the complex needs of patients with severe mental illness. Further research is needed to better inform clinicians in adjusting psychotropic drug prescribing during these unprecedented times.

We declare no competing interests. 

Lower risk of SARS-CoV2 infection in individuals with severe mental disorders on antipsychotic treatment: a retrospective epidemiological study in a representative Spanish population

We thank Xavier Boland and Luiz Dratcu for their comments on our meta-analysis. 1 They have accurately identified an essential gap in the current understanding of COVID-19 impact on patients with mental disorders. Most of the available evidence in our metaanalysis relied on electronic health records, which indeed did not allow calculation of the risk associated with antipsychotics without confounding by indication. Nevertheless, risk estimates associated with antipsychotics remained significant in pooled odds ratios for adjusted estimates (2·43 [95% CI 1·81-3·25], I²=61·35%), accounting for age in all three studies and comorbidities in two of the three studies. 1 To differentiate the risks associated with specific drugs and the condition they are meant to treat, we have previously recommended reporting of medication-stratified risk estimates for each psychopharmacological drug class in patients with mental disorders. 2 We are glad to see emerging evidence answering our plea. In a recently published study by Nemani and colleagues, 3 anti psychotic treatment was not associated with increased mortality among a US cohort of 464 patients with severe mental illness, 196 of whom were treated with antipsychotic medication. However, this does not preclude potential risks associated with specific antipsychotics, 4 nor does it exclude the possibility of unaccounted-for confounding. Considering current treatment guidelines for severe mental illness, in particular among patients with