Abstract: FR-PO0385
Predictive Accuracy of Neutrophil Gelatinase-Associated Lipocalin (NGAL), Kidney Injury Molecular-1 (KIM-1), Cystatin C, and TIMP-2 IGFBP7 for Cardiopulmonary Bypass-Associated AKI: A Systematic Review and Meta-Analysis
Session Information
- AKI: Biomarkers, Diagnostics, and Risk Prediction
October 23, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Acute Kidney Injury
- 102 AKI: Clinical, Outcomes, and Trials
Authors
- Bhardwaj, Shlok, All India Institute of Medical Sciences Bathinda, Bathinda, PB, India
- Bansal, Vilohit, All India Institute of Medical Sciences Bathinda, Bathinda, PB, India
- Singh, Harjot, All India Institute of Medical Sciences Bathinda, Bathinda, PB, India
- Nohria, Sanyam, All India Institute of Medical Sciences Bathinda, Bathinda, PB, India
- Japnoor singh, Fnu, All India Institute of Medical Sciences Bathinda, Bathinda, PB, India
Background
AKI is a common complication after surgery using cardiopulmonary bypass (CPB). Several biomarkers are used to predict AKI; but their accuracy at different timepoints is uncertain. We aim to test predictive accuracy of various biomarkers and whether diabetes prevalence acts as a modifier.
Methods
This study was registered with PROSPERO (CRD420261376312). MEDLINE, Embase, Scopus were searched for eligible studies. Adult CPB-surgery population (>20); measurement of NGAL, KIM-1, Cystatin C, TIMP-2 IGFBP7; KDIGO equivalent AKI diagnostic criteria; and reporting of AUROC/2×2 contingency data were primary eligibility criteria. QUADAS-2 was used for assessing risk of bias. Random-effects meta-analysis of AUROCs of biomarkers at different time ranges was performed, along with meta-regression to test if diabetes prevalence affected performance.
Results
Total of 1298 unique studies were identified; 1137 being excluded at title/abstract screening, and 125 being excluded at full text screening leaving 36 included studies. Pooled AUROC for all biomarkers assessed are attached. Plasma NGAL showed the highest pooled AUROC, but this must be interpreted with caution as there were only 3 effects from 2 studies. TIMP-2 IGFBP7 showed modest accuracy despite being well studied. Urinary NGAL performed best in 0-4 hours post CPB (0.76, k=7). Its accuracy declined thereafter in 4-12 hours post CPB (0.63, k=3), which is consistent with known biokinetics. Diabetes prevalence did not modify performance for urinary NGAL (β= -0.003, p=0.82, k=6) or urinary TIMP-2 IGFBP7 (β= +0.007, p=0.68, k=9). Pre-planned sensitivity analysis was unfeasible as most contributing studies showed high risk of bias.
Conclusion
Plasma NGAL and Cystatin C show promising predictive value, albeit evidence size is low. Urinary NGAL shows best performance within 0-4h post CPB. Diabetes prevalence did not modify biomarker performance, but further investigation is warranted. Confidence in evidence presented remains constrained due to limited high-quality studies reporting full diagnostic data.
| Biomarker | k | AUROC (95% CI) | I^2 | Q p |
| Plasma NGAL | 3 | 0.81 (0.75-0.86) | 0% | 0.75 |
| Serum Cystatin C | 4 | 0.74 (0.70-0.78) | 0% | 0.44 |
| Urinary NGAL | 13 | 0.71 (0.65-0.76) | 49% | 0.03 |
| Urinary TIMP-2 IGFBP7 | 10 | 0.68 (0.62-0.74) | 78% | <0.001 |
| Urinary KIM-1 | 5 | 0.63 (0.53-0.71) | 18% | 0.33 |