Background: Serum fibroblast growth factor 23 (FGF23) is associated with acute kidney injury (AKI) and mortality in patients with critical illnesses. However, the accurate predictive performance of FGF23 on AKI remains inconclusive.
Methods: Meta-analysis was performed using data sources including PubMed, Web of Science, EMBASE, and Cochrane (until June 1, 2021). Cohort or observational studies including patients with AKI and serum FGF23 level as the index test were included. The primary outcome was the AKI detective accuracy. This study has been registered in PROSPERO (CRD42021249930).
Results: Eleven studies with 1946 patients in seven countries were included. Across all settings, the sensitivity and specificity for serum FGF23 levels to predict AKI were 82% (95% CI, 66-91%) and 77% (95% CI, 67-85%), respectively. The diagnostic odds ratio of FGF23 was 15.51 (95% CI, 4.89-49.19), with the pooled positive likelihood ratio of 3.62 (95% CI, 2.25-5.83) and a negative likelihood ratio of 0.23 (95% CI, 0.11-0.50). The area under the receiver operating characteristic curve to detect AKI was 0.86 (95% CI, 0.82-0.88). C-terminal FGF23 had a better performance than intact FGF23.
Conclusions: Plasma FGF23 is a valuable biomarker for incident AKI in critically ill patients. Comparisons of FGF23 with other biomarkers in AKI still need more studies to prove.
Keywords: Acute kidney injury; diagnosis; fibroblast growth factor 23; meta-analysis.
AJTR Copyright © 2021.