Psychological factors associated with Long COVID: a systematic review and meta-analysis.

Background: Despite the immense impact of Long COVID on public health and those affected, its aetiology remains poorly understood. Findings suggest that psychological factors such as depression contribute to symptom persistence alongside pathophysiological mechanisms, but knowledge of their relative importance is limited. This study aimed to synthesise the current evidence on psychological factors potentially associated with Long COVID and condition-relevant outcomes like quality of life. Methods: In this systematic review and meta-analysis, MEDLINE, PsycINFO, and the Cochrane Database of Systematic Reviews were searched for peer-reviewed studies published in English from 2019 to January 2, 2024. Studies providing cross-sectional or longitudinal data on the association between at least one psychological variable and the presence of Long COVID (primary outcome) or condition-relevant secondary outcomes (symptom severity, impairment, quality of life, and healthcare utilisation) were included. Psychological constructs with at least five comparisons were pooled as odds ratio (OR) for categorical data and standardised mean difference (SMD) for continuous data in random-effects meta-analyses of cross-sectional studies with control groups. This review is registered with PROSPERO, CRD42023408320. Findings: 113 studies (n = 312,831 patients with Long COVID) provided data on at least one psychological variable, 63 in cross-sectional group comparisons, 53 in cross-sectional associations, and 18 longitudinal. Most reported findings related to depression and anxiety, and - less frequently - to physical activity, posttraumatic stress disorder, stress, and history of mental illness. Depression (OR 2.35; 95% CI, 1.49-3.70) and anxiety (OR 2.53; 95% CI, 1.76-3.61) were significantly associated with Long COVID and higher in affected patients than controls (depression: SMD 0.88; 95% CI, 0.66-1.11; anxiety: SMD 0.74; 95% CI, 0.50-0.99), while results for physical activity and stress were non-significant. In most prospective studies, the investigated psychological constructs significantly predicted Long COVID. Interpretation: Evidence suggests depression and anxiety to be co-occurring phenomena and predictive factors of Long COVID. Future studies should prospectively investigate psychological constructs such as emotion regulation or dysfunctional symptom expectations, which are well-known risk factors and therapeutic targets of persistent somatic symptoms in other medical conditions, but are so far understudied in Long COVID. Funding: None.