Abstract
Cachexia is associated with lower overall survival (OS) in patients with cancer; however, the relationship between the two is reported to differ according to the definitive criteria for diagnosing cachexia. We aimed to investigate 1) the difference in the prevalence of cachexia in patients with cancer and 2) the association between cachexia and OS, depending on the definitive criteria for diagnosing cachexia in patients with cancer. We searched PubMed and Web of Science from their inception until July 31, 2023, to identify eligible studies. We conducted a systematic review of the prevalence of cachexia in patients with cancer and performed a meta-analysis to investigate its relationship with OS. A total of 125 articles comprising 137,960 patients were included in the systematic review, and 26 articles consisting of 11,118 patients underwent meta-analysis. The overall prevalence of cachexia in patients with cancer was 33.0% (95% confidence interval [CI]: 32.8, 33.3); however, it varied according to the definitive criteria for diagnosing cachexia (13.9%-56.5%). According to the Fearon 2011 criteria, the prevalence of cachexia was associated with a high hazard ratio (HR) for OS compared with that of noncachexia [HR: 1.58 (95% CI: 1.45, 1.73)]; according to the other criteria, the HR was 2.78 (95% CI: 1.88, 4.11), indicating significant subgroup differences (P = 0.006). The dose-response curve indicated that the HR for OS plateaued at a cachexia prevalence range of 40%-50% (l-shaped relationship). The prevalence of cachexia in patients with cancer may vary depending on the definitive criteria used to diagnose cachexia. The HR for OS was higher for low cachexia prevalence. The definitive criteria should be carefully considered when assessing cachexia in patients with cancer. This trial was registered at the PROSPERO as CRD42023435474.