Cardioplegia Strategies in Minimally Invasive Aortic Valve Replacement: An Inverse Probability of Treatment Weighting Analysis.

Background and

Objectives

Optimal myocardial protection during minimally invasive aortic valve replacement (MIAVR) is debated. We compared four cardioplegia strategies. Materials and

Methods

Consecutive MIAVR patients (January 2010-April 2025) at a single centre were analysed retrospectively. Cardioplegia regimens were Buckberg (n = 131), Calafiore (n = 153), Custodiol HTK (n = 146) and St Thomas' (n = 113). Because substantial baseline imbalances were present in the unadjusted cohort, inverse probability of treatment weighting (IPTW) based on a multinomial propensity score was applied to achieve covariate balance between groups. IPTW was performed using a comprehensive propensity model that incorporated (1) baseline demographic and clinical characteristics, (2) anatomical factors, including bicuspid valve morphology. Procedural time variables were assessed in secondary sensitivity analyses. After IPTW application, all variables, including procedural times, achieved balance (ASMD < 0.1). Postoperative outcomes were then compared in this fully balanced pseudo-population. The Scheffé post hoc test was performed.

Results

Groups were demographically comparable except for more bicuspid valves in Buckberg. New-onset paroxysmal atrial fibrillation occurred in 31.2% (Buckberg) and 26.5% (St Thomas') versus 7.1% (Calafiore) and 2.0% (Custodiol) (p < 0.01). Respiratory insufficiency followed a similar pattern (p = 0.02). Intensive-care and hospital stay, major complications, left-ventricular ejection fraction, and 30-day mortality (0.6-3.0%) were equivalent. Bicuspid anatomy independently prolonged operative metrics but did not influence biomarkers.

Conclusions

After comprehensive inverse probability weighting that balanced groups on all baseline characteristics and anatomical factors, Calafiore and Custodiol cardioplegia strategies maintained significantly lower rates of new-onset atrial fibrillation (9.3% and 3.8% vs. 28.5% for Buckberg, p < 0.01), reduced myocardial injury biomarker release (peak CK 520 and 510 vs. 920 U/L, p < 0.01), and decreased respiratory complications (7.8% and 8.1% vs. 16.2%, p = 0.01), while mortality, stroke, and resource utilisation measures remained comparable across strategies.