Estimating the Risk of Radiocontrast-Associated Nephropathy

Abstract

Estimates of the incidence of radiocontrast-associated nephropathy vary widely and suffer from misclassification of the cause of AKI and confounding. Using the Nationwide Inpatient Sample, we created multiple estimates of the risk of radiocontrast-associated nephropathy among adult patients hospitalized in the United States in 2009. First, we stratified patients according to the presence or absence of 12 relatively common diagnoses associated with AKI and evaluated the rate of AKI between strata. Next, we created a logistic regression model, controlling for comorbidity and acuity of illness, to estimate the risk of AKI associated with radiocontrast administration within each stratum. Finally, we performed an analysis stratified by the degree of preexisting comorbidity. In general, patients who received radiocontrast did not develop AKI at a clinically significant higher rate. Adjusted only for the complex survey design, patients to whom radiocontrast was and was not administered developed AKI at rates of 5.5% and 5.6%, respectively. After controlling for comorbidity and acuity of illness, radiocontrast administration associated with an odds ratio for AKI of 0.93 (95% confidence interval, 0.88 to 0.97). In conclusion, the risk of radiocontrast-associated nephropathy may be overstated in the literature and overestimated by clinicians. More accurate AKI risk estimates may improve clinical decision-making when attempting to balance the potential benefits of radiocontrast-enhanced imaging and the risk of AKI.

Keywords: acute renal failure; nephrotoxicity; outcomes.

Figure 1.

Assembly of the cohort of included hospitalizations from the total sample.

Figure 2.

The relationship between contrast administration, demographic data, and clinical characteristics for each level of co-morbidity burden. Demographic and clinical characteristics of the study sample, stratified by comorbidity score.

Figure 3.

The variable relationship between contrast administration and AKI across the examined disease states. Risk of AKI, entire sample and diagnosis-defined strata.