Prognosis of fatigue. A systematic review

The objective of the study was to summarize evidence on the course and prognostic factors of fatigue in primary care patients and in the community.

Two reviewers independently screened identified citations, discussed eligible studies, and assessed methodological quality of selected studies. Data concerning study population, duration of follow-up, measurement of fatigue, outcome, and prognostic factors were extracted. Studies with populations selected by a specific disease or postpartum condition were excluded.

We selected 21 articles reporting on 11 (partly) primary care cohorts and six community cohorts. Follow-up was up to 1 year in primary care and up to 4 years in the community, and in most studies that presented duration of fatigue, participants were chronically fatigued. Because of wide heterogeneity of studies, a qualitative analysis was performed. Recovery of fatigue varied widely, but no differences were found between settings. Sufficient evidence for an association with recovery was found for lower severity of fatigue, and limited evidence was found for good self-reported health, mental health, and psychological attributions. A major deficit in methodological quality of most studies was a potential bias due to low or selective response or loss to follow-up.

Most studies on fatigue included patients with long symptom duration at baseline, making it difficult to study prognosis early in the course of fatigue. To provide clear evidence on prognosis in fatigued persons, prognostic studies should use an optimal design including selection of an inception cohort with limited duration of fatigue at baseline, a sufficient sample size, and information on rates and selectivity of response and loss to follow-up.