Abstract

Background: Acute kidney injury (AKI) is a deadly complication of cirrhosis, and treatments are limited. Vasoconstrictors have been used to treat hepatorenal syndrome AKI (HRS-AKI), but the impact of augmenting mean arterial pressure (MAP) on all AKI types has not been studied, limiting the application of clinically-proven treatments for HRS-AKI to non-HRS-AKI. Herein, we precisely characterize the impact of MAP on AKI reversal, independent of subtype.

Methods: Among decompensated cirrhosis patients in our ongoing, prospective cohort, we identified incident AKI episodes AKI was defined as a ≥50% increase in creatinine (sCr) from an outpatient baseline, a minimum of 7 days prior, that required hospitalization and lasted >48 hours. AKI was staged according to guidelines using the peak hospitalization sCr. AKI type was defined per guidelines and independently validated by 3 investigators. AKI reversal was defined as a return in sCr to within 0.3mg/dL of baseline. To determine the impact of MAP on AKI reversal, we completed time-dependent Cox-regression models with time beginning at the time of peak sCr and ending at death, discharge, or AKI reversal. We determined the optimal MAP cut-off for AKI reversal by comparing log-rank statistics. To investigate if the initial MAP or MAP trajectories were associated with AKI reversal, we calculated time-dependent MAP intercepts and slopes over time.

Results: Among 1,626 participants (43% Female, 21% NAFL, 88% White) with a median MELDNa of 16 (14-20) followed for a median 1.5 years (0.6 – 3.5), we identified 306 episodes of AKI that lasted at least 48 hours and required hospitalization. Of these 306 episodes of AKI, 140 (46%) were HRS-AKI. As compared to those with non-HRS-AKI, those with HRS-AKI had significantly lower MAPs (73 v. 78), more severe AKI episodes (≥ Stage 2 AKI: 85% v. 72%), and less AKI reversal (16% v. 30%)(p≤0.004 for all). In time-dependent Cox models accounting for baseline sCr and AKI stage, each 5 mmHg increase in MAP was associated with 1.18x (1.07–1.31) the hazard of AKI reversal. AKI type (e.g., HRS-AKI) was not associated with AKI reversal in uni- or multivariable analyses (p>0.05 for both). There was no interaction between AKI type and MAP (p>0.05). In univariable, time-dependent Cox models, the optimal MAP cutoff most associated with AKI reversal was ≥ 84mmHg (p=0.005)(Figure 1). In time-dependent Cox models accounting for baseline sCr and AKI stage, both MAP slope (aHR 2.2 (1.3 – 3.6) per 1 mmHg increase per minute) and MAP intercept (1.01 per 1 mmHg increase (1.01 – 1.02)) were significantly associated with AKI reversal.

Conclusion: Our analysis highlights that MAP, specifically ≥84 mmHg, is associated with AKI reversal, independent of AKI type. This was true for both initial MAP (intercept) and the change in MAP over time (slope). Our data support the investigation of strategies to maintain a MAP≥84 mmHg among all cirrhosis patients with AKI, regardless of AKI type.