Custodiol™ is an intracellular cardioplegic solution first described by Bretschneider in 1975 [4]. Initially used for organ preservation during transplantation, its application soon extended to cardiac surgery, particularly for lengthy procedures, as Custodiol™ provides myocardial protection for up to three hours with a single dose [18].

The prolonged effect of Custodiol™ compared to BCP accounts for the longer CPB times, more frequent multiple valve or aortic surgeries, and increased reoperations observed in the Custodiol™ group in our study. Additionally, these patients had a 22% higher EuroSCORE II (4.97 vs. 3.38, SMD = 0.358), indicating that surgeons preferred Custodiol™ for higher-risk patients undergoing more complex procedures. Our patient cohort’s risk profile differs significantly from other studies, as we included patients undergoing root/ascending aorta surgery and/or multiple valve surgeries.

A recent meta-analysis included 12 clinical trials [3], of which six exclusively recruited on-pump CABG patients, three focused on mitral valve repair, and three included various procedures. As compared to our study, most papers in such meta analysis excluded concomitant procedures and reoperations, and none reported aortic cross-clamping times exceeding those in our study [6,7,8, 19,20,21,22].

Few studies with an adequate sample size have compared Bretschneider-type cardioplegia and BCP in isolated valve surgery [23, 24]. One recent study compared Bretschneider cardioplegia (n = 277) with blood (Calafiore) cardioplegia (n = 194) in triple valve surgery (mitral, aortic, and tricuspid) [25]. Hence, evidence on Custodiol™ or BCP use in high-risk, technically complex surgeries, as included in our study, is limited. This is precisely the context where the impact of long-acting cardioplegia might be more pronounced.

Custodiol™ is administered in a single dose over 6–8 min, whereas BCP is given intermittently every 20 min with variable volume based on patient weight. There is no robust evidence on how this impacts CPB or cross-clamping times. For instance, Vivacqua et al. [7] found no significant differences in aortic cross-clamping times (-3 min (95% CI -11.5; 5.5)). Similarly, Mercan et al. [9] reported − 0.9 min (95% CI -11.5; 13.3). In contrast, Gaudino et al. [8] observed significant but minimal differences favoring Custodiol™ (-3 min (95% CI -5.7; -0.3)). A meta-analysis of clinical trials with 1,327 patients found no differences in aortic cross-clamping times (weighted difference 1.51 min (95% CI -1.6; 4.6)) or CPB duration (4.9 min (95% CI -5; 14.8)) [3]. In our study, after balancing differences using propensity score analyses, CPB times were identical (102.1 min, SMD < 0.001), and ischemia times were similar (79.4 min in the Custodiol™ group vs. 79.2 min in BCP; SMD 0.006). This may be due to the initial time spent administering Custodiol™ being offset by subsequent doses of blood cardioplegia, especially when administered antegradely.

In our study, we consistently found lower hospital mortality and LCOS with MCS in the Custodiol™ group across all analyses (unadjusted, PS matching, or IPTW). This finding contrasts with current knowledge. For example, an observational study of patients undergoing triple valve surgery found no differences between groups (HTK: 16.2%; BCP: 18.2%; p = 0.619) [25]. Another study with 7,263 patients undergoing aortic valve surgery (HTK n = 5,998; blood n = 1,007) also found no significant differences (HTK 4% vs. blood 4.2%, p = 0.9) [24]. The discrepancies between studies may be due to differing patient profiles (Ie: our study did not exclude patients with endocarditis, chronic kidney disease, etc.) and types of procedures: (Ie: We included mostly single valve surgery, but also double and triple valve surgery. We did not exclude thoracic aortic repair or concomitant CABG, etc.). Besides, we restricted the cross-clamping times to > 30 and < 200 min, which was not a found in previous studies.

Findings regarding other events were inconsistent. For instance, there appeared to be a lower risk of AKIN III renal failure and postoperative myocardial infarction in the Custodiol™ group, but a higher risk de novo AF (Table 2). However, a meta-analysis of clinical trials [3] by Albadrani found no significant differences in terms of LCOS (aRR = 1.11 (95% CI 0.92; 1.3)), de novo AF (aRR = 0.81 (95% CI 0.6; 1.1)), or ITn levels (-0.36 ng/mL (95% CI -1.48; 0.76)).

Patients with a small body surface area are at greater risk of hemodilution and hypothermia, leading to associated risks of arrhythmias, transfusion needs, and potentially increased mortality [26]. In our study, we evaluated the modifying effect of body surface area and found no significant interaction.

Finally, we found significant differences in long-term survival both in the unadjusted sample and in the propensity score-matched groups. The comparison of restricted mean survival times demonstrated that survival differences emerged from the first year (Custodiol™: 0.96 years vs. BCP: 0.94 years, p = 0.043) and increased until the 8-year follow-up (Custodiol™: 7.06 years vs. BCP: 6.66 years, p = 0.001), with a lower mean time to death of 2.6 years (95% CI 2.1; 3.2) for BCP in the IPTW analysis. Still, as the causes of death in the follow up have not been assessed, we cannot speculate on a biological explanation for such finding. Besides, differences in survival might be explained by the unequal early mortality or because of hidden biases which could not be addressed given the retrospective nature of this analysis.

This study has important limitations. First, it is an observational and retrospective study. Despite the large sample size and robust methods to control for potential confounders, we cannot rule out undetected biases that could affect the observed associations. Additionally, the study was conducted at a single center, limiting the external validity of its findings. We did not measure important outcomes that could have provided more interpretability to the results, such as markers of myocardial damage or events other than mortality during follow-up. Furthermore, Custodiol™ was progressively preferred over BCP across the study period, so it is likely that the observed differences might be attributed not only to the type of cardioplegia but to the improvement of surgical skills, perioperative and intraoperative care during that time.

In conclusion, in the retrospective analysis of a large series of patients undergoing heart valve or aortic surgery, we found that myocardial protection with Custodiol™ was associated with greater long-term survival, a lower risk of myocardial infarction, and acute renal failure, at the expense of higher risks of postoperative atrial fibrillation.