Statistical Analysis

Characteristics for the whole cohort, as well as for the groups of patients who did and did not develop AKI were tabulated using count (%) or median (interquartile range [IQR]). Univariate analysis was conducted using Fisher's exact test for binary variables, Chi-square test for categorical variables, and the Wilcoxon rank-sum test for continuous variables with statistical significance assessed at the p  < 0.05 level.

Multivariable analysis for the primary outcome (severe AKI) was conducted using logistic regression taking into account potential confounders in the relationship between perioperative and postoperative hemodynamic risk factors and AKI. To simplify the model and use the most clinically intuitive values, we used a predetermined set of intra- and postoperative characteristics believed to be clinically relevant. For further simplification, we used postoperative mean CVP and mean MAP rather than the calculated values MPP and DPP or DBP (a component of MAP). Similarly, because donor age, recipient age, and body surface area were highly collinear (Pearson's correlation coefficient = 0.8–0.92, p  < 0.001), only recipient age was included in the model. Furthermore, because the scale and clinical significance of 1-unit changes in cardiac index, MAP, CVP, and HCT varied widely, these measures were coded as binary variables for ease of interpretation of clinical impact. Dichotomization occurred on median value for all variables except for postoperative MAP which was dichotomized at <65 mm Hg (= 25th percentile), an important cut-off identified in other research studies. ²⁶ Lastly, because NIRS data were only available for 46 patients, the impact of lowest renal NIRS was assessed in a subsample of patients with NIRS data.

Multivariable outcomes for secondary outcomes (ventilator days, post-HT ICU days, and post-HT hospital days) were conducted using a Poisson's model for skewed data. Potential risk factors for secondary outcomes included severe AKI, as well as a predetermined set of preoperative characteristics which included recipient age, race, pre-HT diagnosis, pre-HT VAD support, and allograft ischemic time. For all multivariable models, results are reported as odds ratios (OR) or incident rate ratios (IRR) with 95% confidence intervals (CI) as appropriate.

Overall posttransplant survival in patients with and without a history of AKI are depicted with Kaplan–Meier curves. Patients who did not die were censored event free at the end of the data collection period. Comparison of HT survival between those with and without AKI groups was performed using the log-rank test.

Clinical data were collected using RedCap (version: 4.5.2, Vanderbilt University), a web-based application designed to support data capture for research studies. ²⁷ Statistical analysis was conducted using Stata version 16 (College Station, Texas, United States). The protocol for this study was approved by Stanford University's Institutional Review Board Panel on Medical Human Subjects (protocol no.: 9731). Individual consent was waived. The clinical and research activities being reported are consistent with the Principles of the Declaration of Istanbul as outlined in the “Declaration of Istanbul on Organ Trafficking and Transplant Tourism.”