Early acute kidney injury after pulmonary endarterectomy: incidence, associated perioperative factors, and early outcome in a single-centre retrospective cohort

Pulmonary endarterectomy (PEA) is the standard treatment for operable chronic thromboembolic pulmonary hypertension, but factors associated with acute kidney injury (AKI) remain uncertain. We examined perioperative factors associated with AKI within 48 h and evaluated an exploratory early postoperative risk-identification model. We screened 126 consecutive adults with operable chronic thromboembolic pulmonary hypertension who underwent PEA at a tertiary referral centre between January 2018 and December 2020. Eight met predefined exclusion criteria and one of 118 eligible patients had no assessable renal criterion. AKI was defined by the 2012 Kidney Disease: Improving Global Outcomes serum creatinine and urine-output criteria. Eleven patients whose status could not be determined, because one criterion was negative and the other unavailable, were excluded, leaving 106 for analysis. Multiple imputation assessed missing-covariate sensitivity, and the combined model underwent bootstrap validation, calibration and decision-curve analysis. AKI occurred in 42 of 106 adults (39.6%): 35 stage 1, four stage 2 and three stage 3; alternative handling of the indeterminate cases gave incidences between 35.9% and 45.3%. When extracorporeal circulation (ECC), deep hypothermic circulatory arrest and aortic cross-clamp durations were analysed together, only ECC duration remained independently associated with AKI (adjusted odds ratio 1.032 per minute, 95% confidence interval 1.012–1.052; p = 0.001). Age was the only independently associated preoperative variable (1.034 per year, 1.004–1.065; p = 0.027). An integrated model combining age (1.035 per year, 1.007–1.063; p = 0.014) and ECC duration (1.025 per minute, 1.009–1.041; p = 0.003) gave apparent and optimism-corrected areas under the receiver operating characteristic curve of 0.718 and 0.706, with a calibration slope of 0.97. Albumin at intensive care unit admission was inversely associated with AKI in univariable analysis but was not independently associated in the integrated model ( p = 0.129) and improved neither fit nor discrimination. Haemodynamic improvement was similar in both groups, but 30-day mortality was higher among patients with AKI (14.3% vs. 0%; p = 0.003). Among adults undergoing PEA for operable chronic thromboembolic pulmonary hypertension, longer ECC duration and older age were associated with AKI within 48 h. Similar haemodynamic recovery did not exclude renal injury. The model is exploratory and requires external validation. ClinicalTrials.gov, NCT07132619. Retrospectively registered 13 August 2025.