Abstract: TH-PO025
Risk Factors for AKI in High-Risk Populations: Systematic Review and Meta-Analyses
Session Information
- AKI: Epidemiology, Risk Factors, Prevention
October 25, 2018 | Location: Exhibit Hall, San Diego Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Acute Kidney Injury
- 101 AKI: Epidemiology, Risk Factors, and Prevention
Authors
- Zafar, Waleed, Geisinger, Dnville, Pennsylvania, United States
- Garcia-Arce, Andres, Geisinger, Dnville, Pennsylvania, United States
- Hu, Yirui, Geisinger, Dnville, Pennsylvania, United States
- Kwiecien, Sara J., Geisinger, Dnville, Pennsylvania, United States
- Singh, Gurmukteshwar, Geisinger Medical Center- Nephrology, Danville, Pennsylvania, United States
- Saunders, Sara, Geisinger, Dnville, Pennsylvania, United States
- Bucaloiu, Ion D., Geisinger Medical Center, Danville, Pennsylvania, United States
- Kirchner, H. Lester, Geisinger Clinic, Danville, Pennsylvania, United States
- Ho, Kevin, Geisinger Medical Center, Danville, Pennsylvania, United States
Background
Predictive models help identify patients at increased risk for Acute kidney injury (AKI) for preventive care. It is unclear how various risk factors perform in different clinical settings. We systematically reviewed predictive models for AKI in four high risk settings and performed meta-analyses of common AKI risk factors
Methods
We searched MedLine, EMBASE, CINAHL and Cochrane Library, and performed hand searches of the retrieved reference lists, from 2010 through 2017 for English language studies of prediction models for hospital-acquired AKI (KDIGO, AKIN or RIFLE criteria) in adult hospital populations, patients undergoing cardiac surgery, patients in intensive care unit (ICU) and those undergoing contrast procedures. PRISMA-P 2015 statement was used for data extraction and study appraisal. For the meta-analyses of risk factors, studies with insufficient or non-standardized data and unadjusted risk factors were excluded. Random effects meta-analyses were used to assess pooled adjusted odds ratio (OR) for diabetes (DM), congestive heart failure (CHF) and chronic kidney disease (CKD) for developing AKI
Results
74 studies testing AKI prediction models in the four clinical settings were included. Of these, 10 were in the general hospital setting, 8 were in the ICU, 27 in percutaneous coronary interventions, and 29 in cardiac surgery. There was considerable heterogeneity among studies in the choice and definition of predictors. Meta-analyses included 14 studies reporting diabetes (135955 patients), 16 studies reporting CKD (137846 patients) and 22 studies reporting CHF (180224 patients). Stratified analyses based on high-risk clinical settings were consistent with the overall results (Table 1)
Conclusion
In the combined as well as sub-populations, DM, CKD, and CHF are associated with increased risk for AKI. Analyses of differential risk levels in sub-populations was limited by heterogeneity. Using uniform definitions for AKI and its risk factors in the development of predictive models for AKI is essential for their clinical application
Meta-analyses results
Overall estimate OR (95% CI) | Percutaneous coronary intervention OR (95% CI) | Cardiac surgery OR (95% CI) | |
Diabetes | 1.72 (1.38-2.14) | 1.86 (1.47-2.34) | 1.26 (1.01-1.56) |
CKD | 1.47 (1.32-1.64) | 2.53 (1.10-5.83) | 1.31 (1.15-1.51) |
CHF | 1.78 (1.54-2.06) | 2.07 (1.62-2.65) | 1.55 (1.31-1.83) |
Funding
- Private Foundation Support