Objective: The prevalence of post-stroke fatigue differs widely across studies, and reasons for such divergence are unclear. We aimed to collate individual data on post-stroke fatigue from multiple studies to facilitate high-powered meta-analysis, thus increasing our understanding of this complex phenomenon.
Methods: We conducted an Individual Participant Data (IPD) meta-analysis on post-stroke fatigue and its associated factors. The starting point was our 2016 systematic review and meta-analysis of post-stroke fatigue prevalence, which included 24 studies that used the Fatigue Severity Scale (FSS). Study authors were asked to provide anonymised raw data on the following pre-identified variables: (i) FSS score, (ii) age, (iii) sex, (iv) time post-stroke, (v) depressive symptoms, (vi) stroke severity, (vii) disability, and (viii) stroke type. Linear regression analyses with FSS total score as the dependent variable, clustered by study, were conducted.
Results: We obtained data from 14 of the 24 studies, and 12 datasets were suitable for IPD meta-analysis (total n = 2102). Higher levels of fatigue were independently associated with female sex (coeff. = 2.13, 95% CI 0.44-3.82, p = 0.023), depressive symptoms (coeff. = 7.90, 95% CI 1.76-14.04, p = 0.021), longer time since stroke (coeff. = 10.38, 95% CI 4.35-16.41, p = 0.007) and greater disability (coeff. = 4.16, 95% CI 1.52-6.81, p = 0.010). While there was no linear association between fatigue and age, a cubic relationship was identified (p < 0.001), with fatigue peaks in mid-life and the oldest old.
Conclusion: Use of IPD meta-analysis gave us the power to identify novel factors associated with fatigue, such as longer time since stroke, as well as a non-linear relationship with age.
Keywords: Depression; Fatigue; Fatigue Severity Scale; Individual data; Meta-analysis; Stroke.
Copyright © 2018 Elsevier Inc. All rights reserved.