Cognitive impairments four months after COVID-19 hospital discharge: Pattern, severity and association with illness variables

Abstract

The ongoing Coronavirus Disease 2019 (COVID-19) pandemic has affected more than 100 million people and clinics are being established for diagnosing and treating lingering symptoms, so called long-COVID. A key concern are neurological and long-term cognitive complications. At the same time, the prevalence and nature of the cognitive sequalae of COVID-19 are unclear. The present study aimed to investigate the frequency, pattern and severity of cognitive impairments 3-4 months after COVID-19 hospital discharge, their relation to subjective cognitive complaints, quality of life and illness variables. We recruited patients at their follow-up visit at the respiratory outpatient clinic, Copenhagen University Hospital, Bispebjerg, approximately four months after hospitalisation with COVID-19. Patients underwent pulmonary, functional and cognitive assessments. Twenty-nine patients were included. The percentage of patients with clinically significant cognitive impairment ranged from 59% to 65% depending on the applied cut-off for clinical relevance of cognitive impairment, with verbal learning and executive functions being most affected. Objective cognitive impairment scaled with subjective cognitive complaints, lower work function and poorer quality of life. Cognitive impairments were associated with d-dimer levels during acute illness and residual pulmonary dysfunction. In conclusion, these findings provide new evidence for frequent cognitive sequelae of COVID-19 and indicate an association with the severity of the lung affection and potentially restricted cerebral oxygen delivery. Further, the associations with quality of life and functioning call for systematic cognitive screening of patients after recovery from severe COVID-19 illness and implementation of targeted treatments for patients with persistent cognitive impairments.

Keywords: COVID-19; Cognitive impairment; Pulmonary dysfunction; Quality of life.

Fig. 1

Flow-chart for recruitment of patients in post-COVID cognition assessments.

Fig. 2

Proportion of patients with clinically relevant global or selective cognitive impairments using different approaches for determining the clinically relevance of impairments. (A) Using the recommended cut-off for global impairment defined as scores ≥0.5 below the expected SCIP Total scores and – for selective impairments – scores ≥1 SD below the expected scores on ≥2 individual tests based on patients’ age, sex and education years yielded n = 18 patients 62%) with global impairments (dark red) and n = 1 patient (3%) with selective impairments (light red); i.e., 65% patients being cognitively impaired. (B) With a more conservative cut-off for global impairment defined as SCIP Total scores ≥1 below demographically adjusted norms and – for selective impairments – performance ≥1 SD below the demographically adjusted norms on ≥2 individual tests yielded n = 11 patients (38%) with global impairments (dark red) and n = 6 patients (21%) with selective impairments (light red); i.e., 59% patients being cognitively impaired. (C) Finally, comparisons with age- and education matched healthy controls (n = 100) and use of the cut-off for global and selective impairments as in (B), indicated that n = 11 patients (38%) had global impairments (dark red) and n = 7 (24%) had selective impairments (light red); i.e., 62% showed clinically relevant cognitive impairments. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Pattern of cognitive impairments in patients (n = 29) four months after COVID-19 in comparison (A) with normative scores adjusted for age, sex and education estimated with regression models and (B) with an age-, sex- and education-matched healthy control group (n = 100). Most pronounced impairments were seen in verbal learning (VLT-L) and executive function (TMT-B). Graphs represent the mean and error bars the standard error of the mean. * p<0.05; ** p<0.01; *** p<0.001.