Abstract
BACKGROUND: Peritonitis in peritoneal dialysis and cirrhosis remains common and leads to morbidity. Neutrophil gelatinase-associated lipocalin (NGAL) has been evaluated as a rapid adjunctive biomarker. METHODS: Following PRISMA-DTA and PROSPERO registration (CRD420251105563), we searched MEDLINE, Embase, Cochrane Library, LILACS, Scopus, and Web of Science from inception to 31 December 2024, and ran an update on 30 June 2025 (no additional eligible studies). Diagnostic accuracy studies measuring NGAL in peritoneal/ascitic fluid against guideline reference standards were included. When 2 × 2 data were not reported, we reconstructed cell counts from published metrics using a prespecified, tolerance-bounded algorithm (two studies). Accuracy was synthesized with a bivariate random effects (Reitsma) model; 95% prediction intervals (PIs) were used to express heterogeneity; small-study effects were assessed by Deeks' test. RESULTS: Thirteen studies were included qualitatively and ten were entered into a meta-analysis (573 cases; 833 controls). The pooled sensitivity was 0.95 (95% CI, 0.90-0.97) and specificity was 0.86 (0.70-0.94); likelihood ratios were LR+ ≈7.0 and LR- 0.06. Between-study variability was concentrated on specificity: the PI for a new setting was 0.75-0.98 for sensitivity and 0.23-0.99 for specificity. Deeks' test showed evidence of small-study effects in the primary analysis; assay/platform and thresholding contributed materially to heterogeneity. CONCLUSIONS: NGAL in peritoneal/ascitic fluid demonstrates high pooled sensitivity but variable specificity across settings. Given the wide prediction intervals and the signal for small-study effects, NGAL should be interpreted as an adjunct to guideline-based criteria-not as a stand-alone rule-out test. Standardization of pre-analytics and assay-specific, locally verified thresholds, together with prospective multicenter validations and impact/economic evaluations, are needed to define its clinical role.