Perioperative Hemodynamic Disturbance as a Mediator Between Preoperative Pulmonary Hypertension and Acute Kidney Injury After Cardiopulmonary Bypass-assisted Cardiac Surgery: A Retrospective Cohort Study

Technical Summary: Perioperative Hemodynamic Disturbance as a Mediator Between Preoperative Pulmonary Hypertension and Acute Kidney Injury

Problem Statement
Patients with preoperative pulmonary hypertension (PH) undergoing cardiopulmonary bypass (CPB)-assisted cardiac surgery face a significantly elevated risk of developing acute kidney injury (AKI). While the association between PH and postoperative AKI is established, the specific modifiable mediators within the perioperative phase remain insufficiently characterized. Existing literature often focuses on the direct association or general hemodynamic instability, lacking a quantified understanding of how specific intraoperative hypotension (IOH) and the intensity of vasoactive-inotropic support contribute to this risk. This study aims to determine whether and to what extent IOH and vasoactive-inotropic support act as mediators in the causal pathway between preoperative PH and postoperative AKI.

Methodology
This retrospective cohort study analyzed 2,242 adult patients who underwent elective cardiac surgery with CPB at Nanjing First Hospital between June 2018 and June 2023.

• Exposure: Preoperative PH was defined as a pulmonary artery systolic pressure (PASP) ≥ 50 mmHg, analyzed both as a continuous variable and a binary variable.
• Mediators:
  • Intraoperative Hypotension (IOH): Mean arterial pressure (MAP) data were extracted at 5-minute intervals. Four thresholds (<65, <60, <55, and <50 mmHg) were evaluated. Two metrics were calculated for each: cumulative duration below threshold (TIME) and area under the threshold (AUT).
  • Vasoactive-Inotropic Score (VIS): Calculated intraoperatively and at 24 hours postoperatively using a weighted formula based on the doses of dopamine, dobutamine, epinephrine, milrinone, vasopressin, norepinephrine, and levosimendan.
• Outcome: Postoperative AKI, defined by KDIGO criteria (≥0.3 mg/dL increase in serum creatinine within 48 hours or ≥50% increase within 7 days).
• Statistical Analysis:
  • Multivariable logistic regression identified confounders (including BMI, smoking, hypertension, preoperative creatinine, neutrophil count, NLR, platelet count, CPB duration, age, and sex).
  • Mediation Analysis: Conducted using nonparametric bootstrap resampling (n=1,000) and the Baron-Kenny stepwise approach to estimate the proportion of the total effect mediated by hemodynamic parameters.
  • Sensitivity Analysis: Continuous mediators were dichotomized at optimal ROC-derived cutoffs to verify result stability.

Key Results

• Cohort Characteristics: 33.5% (751/2,242) of patients developed postoperative AKI. The AKI group had significantly higher BMI, preoperative PASP, IOH severity, and perioperative VIS compared to the non-AKI group.
• Mediation Effects:
  • Postoperative VIS: This was the strongest mediator. It accounted for 23.9% of the total effect when PH was treated as a continuous variable and 27.6% when treated as a binary variable (both P < 0.01).
  • Intraoperative IOH (MAP_TIME <65 mmHg): This was the second most significant mediator, accounting for 5.0% (continuous PH) and 5.3% (binary PH) of the effect (both P < 0.001).
  • Intraoperative VIS: Showed no statistically significant mediation effect (1.9%; P=0.120).
• Thresholds: Sensitivity analyses identified optimal cutoffs of 164 minutes for intraoperative MAP_TIME_65 and a VIS of 3.4 for postoperative VIS. Dichotomizing variables at these thresholds confirmed the robustness of the mediation effects.
• Concordance: Results from the bootstrap method were consistent with the Baron-Kenny causal stepwise regression estimates.

Significance and Claims
The paper claims that preoperative PH is an independent risk factor for postoperative AKI, but this risk is partially modifiable through perioperative hemodynamic management.

• Mechanism: The study posits that the increased AKI risk in PH patients is not solely due to the pathology of PH itself (e.g., right ventricular dysfunction, venous congestion) but is significantly driven by the patient's inability to tolerate hemodynamic fluctuations. Specifically, PH patients are more vulnerable to sustained IOH and require higher levels of postoperative pharmacologic support, both of which independently increase AKI risk.
• Clinical Implication: The authors advocate for a "bundled, phase-specific hemodynamic management strategy."
  • Intraoperatively: Real-time monitoring is essential to minimize exposure to MAP <65 mmHg, as prolonged hypotension acts as a mediator.
  • Postoperatively: The VIS within the first 24 hours should be incorporated into early risk assessment. High VIS indicates a cumulative burden of hemodynamic disturbance that contributes significantly to AKI risk.
• Limitations: The authors acknowledge the study is a single-center retrospective analysis, subject to unmeasured confounding. Additionally, blood pressure was recorded at 5-minute intervals rather than continuously, potentially missing transient fluctuations, though the use of TIME and AUT metrics aimed to mitigate this.

The study concludes that managing AKI risk in PH patients requires shifting from isolated blood pressure or medication targets to an integrated strategy that addresses the entire perioperative hemodynamic continuum, from minimizing hypotension during surgery to optimizing circulatory support in the ICU.