LAUSR

Type of funding sources: None. An earlier published prospective cohort study evaluating the impact of malnutrition on cardiac surgery outcomes showed serum albumin ≤42 g/dL to be associated with a poor prognosis [1]. However, the association with kidney outcomes was not studied. The adverse effect of preoperative hypoalbuminemia on postoperative kidney injury and its cut-off value in on-pump cardiac surgery patients is still controversial [2]. A retrospective matched case-control study based on the secondary analysis of the open-access dataset [3] was conducted. The research hypothesis was that preoperative hypoalbuminemia is independently associated with a higher risk of acute kidney injury (AKI) requiring postoperative new-onset dialysis. We defined cases as patients treated with new-onset dialysis for AKI and controls as those who did not. Hypoalbuminemia was defined as serum albumin of ≤42 g/L based on findings from the original study. The dataset counted 1178 adults from multiple centers who underwent elective cardiopulmonary bypass (CPB) cardiac surgery in 2011. We excluded patients with missing data and compared key demographic and medical history characteristics. Following the hypothesis, subjects were matched on the type of surgery they received (coronary, valvular, or combined) in a multiple conditional logistic regression model that included the hypoalbuminemia status, diabetes history, CPB time, and logistic EuroScore. Additionally, univariate regression with and without conditional matching was performed. In total, 831 (39 cases and 792 controls) subjects were included. Subjects shared similar demographics and medical history. Expectedly, cases had prolonged CPB and aortic cross-clamp time. The median CPB time in cases and controls was 120 and 79 min, respectively (p=0.001). The median aortic cross-clamp time was 79 in cases and 54 in controls (p=0.001). Hypoalbuminemia was observed in 18 cases (46%) in cases compared to 241 (30%) in controls (p=0.038). Multiple conditional regression demonstrated that the risk of postoperative AKI requiring dialysis was not associated with preoperative hypoalbuminemia (OR 1.43, [0.70, 2.89]). Similarly, univariate conditional regression showed no association (OR 1.84 [0.95, 3.53]). Only univariate unmatched regression resulted in statistical significance(OR 1.96 [1.03, 3.74]). The higher logistic EuroScore (OR 1.11, [1.06, 1.15]) and the longer cardiopulmonary bypass time (OR 1.01, [1.01, 1.02]), but not the history of diabetes (OR 0.61, [0.19, 1.88]), increased the odds of renal replacement therapy requirement. In our study, preoperative hypoalbuminemia was not independently associated with a higher risk of postoperative AKI requiring new-onset dialysis after elective on-pump cardiac surgery. Patients who required dialysis had lower preoperative albumin levels, but the association between hypoalbuminemia ≤ 42 g/L and kidney outcomes could be confounded.