Cognitive domains affected <scp>post‐COVID</scp>‐19; a systematic review and <scp>meta‐analysis</scp>
Jack B. Fanshawe, Brendan Sargent, James Badenoch, Aman Saini, Cameron Watson, Aleksandra Pokrovskaya, Daruj Aniwattanapong, Isabella Conti, Charles Nye, Ella Burchill, Zain Hussain, Khanafi Said, Elinda Kuhoga, Kukatharmini Tharmaratnam, Sophie Pendered, Bernard Mbwele, Maxime Taquet, Greta K. Wood, Jonathan Rogers, Adam Hampshire, Alan Carson, Anthony S. David, Benedict Michael, Timothy R. Nicholson, Stella‐Maria Paddick, Charles E. Leek
Abstract
BACKGROUND AND PURPOSE: This review aims to characterize the pattern of post-COVID-19 cognitive impairment, allowing better prediction of impact on daily function to inform clinical management and rehabilitation. METHODS: A systematic review and meta-analysis of neurocognitive sequelae following COVID-19 was conducted, following PRISMA-S guidelines. Studies were included if they reported domain-specific cognitive assessment in patients with COVID-19 at >4 weeks post-infection. Studies were deemed high-quality if they had >40 participants, utilized healthy controls, had low attrition rates and mitigated for confounders. RESULTS: Five of the seven primary Diagnostic and Statistical Manual of Mental Disorders (DSM-5) cognitive domains were assessed by enough high-quality studies to facilitate meta-analysis. Medium effect sizes indicating impairment in patients post-COVID-19 versus controls were seen across executive function (standardised mean difference (SMD) -0.45), learning and memory (SMD -0.55), complex attention (SMD -0.54) and language (SMD -0.54), with perceptual motor function appearing to be impacted to a greater degree (SMD -0.70). A narrative synthesis of the 56 low-quality studies also suggested no obvious pattern of impairment. CONCLUSIONS: This review found moderate impairments across multiple domains of cognition in patients post-COVID-19, with no specific pattern. The reported literature was significantly heterogeneous, with a wide variety of cognitive tasks, small sample sizes and disparate initial disease severities limiting interpretability. The finding of consistent impairment across a range of cognitive tasks suggests broad, as opposed to domain-specific, brain dysfunction. Future studies should utilize a harmonized test battery to facilitate inter-study comparisons, whilst also accounting for the interactions between COVID-19, neurological sequelae and mental health, the interplay between which might explain cognitive impairment.