Background: Acute Kidney Injury (AKI) is associated with poor outcomes and increased costs in hospitalized patients. A recent meta analysis of controlled studies found no increased risk of AKI attributable to the use of radiocontrast media for diagnostic studies in hospitalized patients. This retrospective observational study provides a real world assessment of the risk of AKI in critically ill patients at BHSF with respect to contrast and other factors.

Methods: Data including patient demographics, physiological parameters, diagnoses and APACHE score were obtained from multiple sources in the BHSF data warehouse. Study data comprised all hospitalizations between 2010 and 2014 that included an ICU stay of at least one day. AKI was derived from creatinine records using Kidney Disease Improving Global Outcomes (KDIGO) guidelines. Multivariate logistic regression models of factors associated with worsening AKI were developed in a propensity-matched sample.

Results: Contrast was administered to 3,051 (36.3%) of the 8,394 study cases. Propensity score matching resulted in 2,749 pairs with an AKI rate of 20.4% and 6.7% hospital mortality. The relative risk of AKI attributable to contrast was 1.11 (95% CI: 1.00 to 1.24). Adjusted for demographic and diagnostic factors, the increased risk of AKI attributable to contrast was 17%, similar to the effect of age (per 15 years) and gender (21% higher in males). Additionally, a fivefold increase in risk was associated with renal comorbidity and a twofold increase with primary diagnoses of infectious disease or endocrine disorder in relation to cardiovascular disease. Patients who developed AKI were at fourfold increased risk of death regardless of the use of contrast media.

Conclusion: Since AKI is strongly predictor of mortality, even factors such as contrast that represent a modest increased risk of developing AKI should be carefully managed.

Publication Date


Presented At:

11th Annual BHSF Research Conference

Content Type


Baptist Health South Florida Affiliations:

Lisa-Mae Williams, MSN, RN

Additional Authors:

Gail Walker, PhD

Open Access

Available to all.