Background: Acute kidney injury (AKI) following cardiac surgery in infants is common and associated with longer mechanical ventilation times and length of stay. Prior studies have shown that a lack of responsiveness to bolus dose furosemide has been associated with increased incidence of AKI. However, these studies have excluded patients on continuous furosemide infusions who are often younger, more hemodynamically unstable, and at higher risk for AKI. Hypothesis: Decreased urine output after initiation of a furosemide infusion predicts development of AKI. Methods: A retrospective cohort study of infants (<1 year of age) was conducted who underwent cardiac surgery requiring cardiopulmonary bypass from 2020-2023 and were on a furosemide infusion post-operatively. The primary outcome was post-operative AKI as defined using KDIGO (Kidney Disease: Improving Global Outcomes) guidelines. A furosemide response score (FRS) was calculated using the total urine output divided by the total dose of furosemide administered over a given time. Univariate and multivariate analyses were performed, and receiver operating characteristic curves used to determine the discriminatory ability of the FRS with regards to the primary outcome. Results: A total of 155 infants (median age at surgery 9 days) were included. Overall, 77% of infants met definition for the primary outcome of AKI; 32% stage 1, 30% stage 2, and 15% stage 3. Lower FRS at 4, 10, and 24 hours after infusion initiation were associated with AKI development (p<0.0001). Younger age, smaller weight at surgery, lack of pre-operative feedings, pre-operative inotropic use and higher surgical complexity were also associated with a higher risk of AKI. An FRS (mL/mg/kg) cutoff of <11.3 at 4 hours, <25.5 at 10 hours, and <55.3 at 24 hours had good discriminatory ability (area under the curve 0.7-0.75) in predicting AKI stage 2 or 3 (p<0.001). Controlling for the significant univariate variables listed above, the FRS cutoffs remained an independent risk factor for the development of AKI stage 2 or 3, (p<0.01) and were significantly associated with longer mechanical ventilation days and length of stay. Conclusion: Lack of responsiveness to a furosemide infusion (as measured by urine output corrected for dose of furosemide) is associated with the development of AKI in a high-risk cohort of infants. These novel findings may be helpful to predict those at risk and allow for further investigations for modifiable risk factors.
Read full abstract