© The Author(s), published by EDP Sciences, 2024

Introduction

The use of Del Nido Cardioplegia (DNC) as an alternative to hyperkalaemic crystalloid blood cardioplegia (HKB) is an accepted practice in cardiothoracic surgery [1]. The use of single-dose DNC for a cross-clamp time of less than 90 min originates from paediatric cardiac surgery [2, 3]. A single dose of DNC provides satisfactory myocardial protection for approximately 90 min [4]. The use of single-dose DNC in adult cardiac surgery with cross-clamp times of less than 90 min is now supported by randomized control trials including that published by Ad et al. [5]. The recent prospective randomized trial by Garcia-Suarez included 474 patients in different settings of adult cardiac surgery not excluding complex procedures and showed comparable outcomes between HKB and DNC [6]. The outcomes of multi dose DNC and determination of safe ischaemic time in adults and in adults with extended cross-clamp time remain unclear. To date there is no clinical guideline supporting the delivery of DNC in cases with extended bypass times.

Current animal models and trials in explanted hearts have shown superior myocardial function and troponin profiles with single dose compared to multi dose DNC regiments [7, 8]. There is also clear evidence of ischaemic changes inferenced by troponin profiles after 90 min of ischaemic conditions in these cohorts [7, 8]. The existing literature on DNC in adults with cross-clamp times over 90 min is currently limited to small patient cohorts or sub-analyses of larger studies and interchangeably defines aortic cross-clamp time and ischaemic time making direct comparisons challenging [4, 9–16]. Ross et al recently reported our unit’s initial experience of DNC however only a small subset (40 patients) had aortic cross-clamp times greater than 90 min in the DNC group [17], while Willekes et al. have reported a propensity matched study of patients with prolonged aortic cross-clamp times [18].

This study reports a review of the safety of DNC within our practice in patients with cross-clamp time exceeding 90 min. We compared those who received HKB with those who received DNC. The primary aim was to assess safety and efficacy based on post-operative Troponin T profile, with the secondary aim to compare post-operative major adverse cardiac events between the two groups.

Methods

This is a single centre, retrospective cohort study including patients who underwent cardiac or aortic surgery, without circulatory arrest, with a cross-clamp time longer than 90 min. Patients from both Flinders Medical Centre and Flinders Private Hospital were included. Patient data was prospectively collected from the Flinders Cardiac Surgery Registry and the Australian New Zealand Collaborative Perfusion Registry from June 2014 to December 2022. During this interval 845 of 5094 patients had aortic cross-clamp times greater than 90 min, of which 188 were excluded from the study (10 no cardioplegia data, 178 as outside of the study period) (Figure 1). Ethics approval for this audit was granted by the Southern Adelaide Clinical Human Research Ethics Committee and the South Australia Local Health Network Office for Research (Quality Registry ID: 2265).

Figure 1 Consort diagram. The 178 patients either had both DNC and HPK cardioplegia (55 cases) or were excluded as the surgeons performing operations did not use both cardioplegic techniques (112 cases). Abbreviations: DNC – Del Nido cardioplegia, HKB – hyperkalemic blood cardioplegia.

Patients were analysed according to their cardioplegia regimen of either HKB or DNC. DNC was introduced in November 2018. Clinical management, anaesthesia, composition and delivery methods of our DNC and HKB have been previously published [17]. Specifically for DNC, after placement of the aortic cross-clamp, cardioplegic arrest was induced with an antegrade induction dose of 1 L delivered at a flow rate of 200–300 mL/min at 6 °C, targeting aortic root pressures >100 mmHg and less than 150 mmHg. This dose was followed by subsequent 500 mL doses at 60-minute intervals delivered antegrade or retrograde as required. In cases of severe aortic regurgitation, a combination of retrograde and ostial cardioplegia was used. Hyperkalaemic blood cardioplegic arrest was induced with tepid (34 °C) hyperkalaemic blood/crystalloid cardioplegia (induction, 30 mmol/L) at induction and maintained with intermittent doses (maintenance, 15 mmol/L) every 20–30 min. Similar flows and pressures were targeted. In both groups, in addition to the timing of cardioplegia doses, the return of electrical or myocardial contractility was an indication for re-dosing.

Definitions of clinical demography and outcomes were standardised on those reported by the Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) National Database. Maximum ischemic time was defined as the maximum duration between completion of cardioplegia delivery and either the beginning of the next cardioplegia delivery or reperfusion with cross-clamp removal; our data reflects expected redosing at >90 min in DNC and >30 min in HKB. Mortality was defined as death in a hospital or within 30 days from surgery. Perioperative myocardial infarction was defined as having at least two of the following criteria: ∆Troponin T > 20 ug/L, new regional wall motion abnormalities on echocardiography, Q wave changes on electrocardiogram (ECG). A positive troponin profile was defined as having a troponin value at 72 h following surgery which is the highest troponin measured within 72 h of the index procedure. Acute Kidney Injury (AKI) was defined as postoperative creatinine greater than 150% baseline in accordance with the serum creatinine criteria of the renal Risk, Injury, Failure, Loss of renal function and End-stage renal disease (RIFLE) classification.

The registries meet the Australian Commission on Safety and Quality in Health Care National Operating Principles for Australian Clinical Quality Registries (https://www.safetyandquality.gov.au/publications-and-resources/resource-library/framework-australianclinical-quality-registries). Database managers and staff meet weekly to undertake quality assurance processes. The unit’s general anaesthetic, intraoperative monitoring, cardiopulmonary bypass (CPB), blood conservation and post-operative renal replacement protocols have been previously published [17].

Statistical analysis

Patients that received DNC were 1:1 propensity-matched without replacement with patients that received HKB, with cross-clamp-times greater than 90 min, providing 194 matched pairs (Figure 1). Preoperative risk factor variables included in propensity matching were age, sex, diabetes, insulin-dependent diabetes, chronic obstructive pulmonary disease, pulmonary hypertension, New York Heart Association classification, left ventricular dysfunction, emergency procedure, cerebrovascular disease, redo procedure, smoking history, elevated preoperative troponin, procedure type, cardiopulmonary bypass time, aortic cross-clamp time, and procedure time. A sensitivity analysis was performed on patients with cross-clamp times greater than 120 min yielding 64 matched pairs.

Stata v 15.1 (StataCorp LLC, Texas) was used for all statistical analyses. Pre-operative, intra-operative and post-operative outcomes were compared between the cohorts. Continuous variables are reported as median with interquartile range and are compared using the Wilcoxon rank-sum test. Categorical variables are reported as a number of patients and group percentage and compared using Fisher’s exact test for variables with binary measures and Pearson’s χ² test for categorical variables. A p-value of <0.05 was considered statistically significant for all analyses without adjustment for multiple comparisons.

Equivalence in outcome for cardioplegia type was evaluated on the incidence of myocardial infarction, positive troponin profile, and continuous postoperative troponin values at 6, 12 and 72 h, peak, and the 72 h area under the curve, calculated based on the method described by Lakens using the Two One-Sided Tests (TOST) procedure [4]. Continuous troponin values underwent log transformation to approximate normality. The TOST procedure utilised the Fishers exact Z-test for proportions. With upper and lower equivalence boundaries defined as Cohen’s d ± 0.3, to detect a type I error rate of 0.05 we had a power of 95% for myocardial infarction and 80% for peak postoperative troponin value between primary analysis groups.

Results

Patient characteristics

Five thousand and ninety-four adult cardiac cases were performed between 12.6.2014 and 31.12.2022, with 845 having an AXC (aortic cross-clamp) time of greater than 90 min. Six hundred and fifty-seven patients were eligible for inclusion in propensity analyses (see consort diagram in Figure 1), resulting in 194 patients in each group for AXC time > 90 min, and 64 patients in each group for the > 120 min AXC time, with both cohorts demonstrating similar pre-operative characteristics respectively (Tables 1 and 2).

Primary analysis

Seventy percent of cases in both groups were valve replacement or valve/coronary grafting surgery. There was no significant difference in total AXC time between the two groups. DNC was delivered at a colder temperature, requiring less volume and fewer doses of cardioplegia. Patients receiving HKB were more likely to receive combined antegrade and retrograde cardioplegia delivery. The maximum period of ischemia was significantly greater for DNC compared with HKB (98 (90, 109) vs. 32 (28, 36) p < 0.001) and the spontaneous return of rhythm was more likely with the DNC compared to HKB (89% vs. 59%, p < 0.001) (Table 3).

Clinical outcomes were the same between DNC and HKB for rates of post-operative intra-aortic balloon pump (IABP), myocardial infarction, acute kidney injury, stroke, return to theatre, or mortality (Table 4). The DNC had a higher rate of return to theatre for bleeding (6% vs. 2%).

The median troponin value at time points 6, 12, and 72 h postoperatively, the maximum postoperative troponin, area under the curve nor the number of patients with positive troponin profile showed any difference between the two groups (Table 5 and Figure 2), similarly, there was no difference in time to peak troponin (Table 6). Equivalence testing found DNC to be equivalent to HKC for all troponin measures other than peak troponin (Table 5) (p = 0.101).

Figure 2 Box and whisker plot for Troponin profile based on cardioplegia strategy for greater than 90-minute ischaemic time. Solid middle bar is the median, top and bottom of box the 75th and 25th percentile, with upper and lower adjacent values.

Sensitivity analysis

Similar findings were found for the AXC time > 120-minute subgroup with no differences in preoperative characteristics and clinical outcomes (Tables 2, 7, and 8). Troponin profiling did demonstrate differences, with the 72 h, maximum postoperative troponin value and the area under the release curve being higher in the DNC group (p < 0.05, Table 9, Figure 3). The median troponin profile at 6 and 12 h, and the number of patients with positive troponin profile between cohorts were not different (Table 6). Equivalence testing found DNC to be not equivalent in troponin measures other than for positive troponin profile (Table 9).

Figure 3 Troponin profile based on cardioplegia strategy for greater than 120-minute ischaemic time. Solid middle bar is the median, top and bottom of box the 75th and 25th percentile, with upper and lower adjacent values.

Discussion

Del Nido cardioplegia has a well-established safety profile in myocardial ischaemic times of up to 90 min [6], with current literature providing little consensus on management protocols for extended AXC times and optimal reporting of clinical endpoints [13, 19, 20]. Clinical advantages of DNC compared to HKB are thought to be mediated by lidocaine’s inhibition of cardiomyocyte sodium channels, prevention of hypertonic myocardial oedema moderated by mannitol and competitive inhibition of calcium influx by magnesium [7, 21]. Comfortable dosing intervals, and advantages over glycaemic control and reperfusion arrhythmias make it a popular alternative for myocardial protection [6]. The continued reporting of clinical experiences is mandated to build an evidence base upon which practice may evolve.

Our results showed patients receiving DNC had significantly longer ischaemic time compared to the matched HKB group, while demonstrating increased rates of return of spontaneous activity, with no significant differences in clinical outcomes (Table 4). While there was overall equivalence in myocardial injury as inference by post-operative troponin T release (Tables 5 and 6) the optimal timing for DNC re-dosing and clinical endpoints for equivalence remain unclear. Sensitivity analysis of patients with cross-clamp times of greater than 120 min similarly demonstrated no significant differences in clinical outcomes (Table 8), however troponin results were equivocal (Tables 6 and 9).

Troponin profile

The post-operative release of high-sensitivity troponin T is one measure reflecting the efficacy of myocardial protection and was chosen for this study due to its lower false positive rate compared to other biomarkers such as CKMB. Recent meta-analyses reporting troponin release in the setting of DNC are of limited value and provide little guidance on strategies for cases with extended AXC times [22, 23]. In reporting the troponin T profile for DNC out to 72 h we found equivalence in the median timepoint values and area under the curve between propensity-matched groups in our primary analysis. This is in keeping with the initial single dose experience from our unit suggesting a robustness in the safety profile of our cardioplegia protocol despite DNC having threefold longer ischemic times (98 min vs. 32 min) [17] and the recent report by Willekes et al. [18] where they reported similar findings for patients with extended AXC times. The sensitivity analysis highlights the variation in redosing within this early experience with 25% of patients in the sub-study receiving a single dose of DNC. In this cohort although troponin T profile values within 72 h were not equivalent, equivalence was observed in the incidence of positive troponin T at 72 h. With further ongoing evaluation, the benefit of DNC may be ascertained.

The literature supporting multi-dose DNC is evolving, with reports in early studies of multi-dose DNC showing no difference in post-operative troponin T profile [10, 13], while Willekes [18] showed lower release of Troponin T their functional assessments showed no differences between DNC and HKB. Other studies comparing DNC and HKB have also shown no difference however are limited by design bias [4, 24]. Existing literature reporting lower troponin profiles with DNC has also reflected a concurrent reduction in AXC time [13, 19]. By comparison our study included as a variable in our propensity matching AXC time and this may explain troponin equivalence rather than reduction, even though DNC patients experienced longer ischaemic times. Our early (6 and 12 h) troponin T measurements are consistent with previously reported experiences of non-inferiority for DNC [5, 25]. Ad et al., found lower and earlier peak troponin with DNC; in contrast, Garcia-Suarez et al incorporated more diverse and complex procedures observing an earlier peak (<12 h) with DNC [5, 6]. In contrast to ours, their dosing strategy was a single 1000 mL induction dose mixed with autologous blood (4:1 crystalloid: blood) followed by 500 mL redosing for ischaemic periods >90 min or in patients with spontaneous activity [6].

Clinical outcomes

Clinical outcomes of DNC in cases with prolonged AXC times and utilising multi-dose strategies varies in the literature [9, 10, 19, 20, 22, 26]. Our study demonstrated no difference in postoperative major adverse events including transfusion, IABP use, maximum inotropic duration, myocardial infarction, acute kidney injury, stroke or mortality. There was no difference in minimum haemoglobin suggesting there is no significant haemodilution with DNC. This was replicated in our sub-group analysis which demonstrated equivalence in outcomes (Table 8), in keeping with a large cohort study sub-analysis done by Koda et al and the recently published prospective RCT by Garcia-Suarez et al. however we did find a higher incidence of stroke with DNC [6, 9]. In contrast, other cohort studies have demonstrated a higher rate of IABP, stroke, and inotropic support and higher peak postoperative creatinine levels in multi-dose DNC [13, 26]. In this report, it is significant that the median number of cardioplegia doses in the 90 min ACX time primary analysis was one therefore inference on multi-dosing may only be based on the sub-analysis.

In the primary analysis the preference for mode of cardioplegia delivery varied among our groups with DNC administered predominantly antegrade (75%) while HKB cardioplegia was delivered by combination antegrade and retrograde delivery in 64% of patients. This variation likely reflects the patient population chosen (those with prolonged cross-clamp times) and the delivery strategy for multiple dose HKB frequently utilising a combined antegrade and retrograde delivery strategy in these patients. The significance of the mode of re-delivery is unclear and given the high variability in existing literature, highlights the need for randomized controlled trials and standardized guidelines given variable clinical implementation. Expectedly the number of cardioplegia doses was 6 (HKB) compared to 1 (DNC) in the primary analysis and 7 (HKB) versus 2 (DNC) in the sensitivity analysis. The maximum ischaemic time was 32 and 98 min in the HKB and DNC groups reflecting our institutions redosing at approximately 30 and 90 min respectively.

The higher rate of return of spontaneous rhythm and lower need for defibrillation on removal of AXC in DNC is well reported [4, 6, 27–29]. Our results showed the DNC cohort was more likely to return to spontaneous rhythm with multi-dose DNC regiments consistent in primary and sub-analyses (Tables 3 and 7) and is consistent with findings our previously published experience [17].

Quality improvement

There is no clear dosing regimen for extended cross-clamp time and DNC use. Current protocols are based on experience, vary widely and make inter study comparisons challenging. The dosing regimens in the literature range from an initial dose of DNC of 1000–1200 mL with an additional maintenance dose of 300–1000 mL every 60 min after 90 min of cross-clamp time [4–7, 9–11, 17].

Following our initial evaluation of DNC use, adjustments to practice were made and commonly a DNC dosing protocol with initiation with 1000 mL DNC induction dose followed by a further 500 mL DNC at 90 min ischemic time was utilised. Following evaluation of our current data, departmental morbidity and mortality reviews and recent publications discussing redosing and timing of DNC for cases with extended AXC [19, 20], we have developed our current multi-dose DNC dosing protocol.

The current dosing protocol stipulates that when the total ischemic time is expected to be less than 90 min a single 1000 mL induction DNC dose is used. If the ischemic time is expected to exceed 90 min a 1000 mL DNC induction dose is given followed by maintenance doses of 500 mL DNC administered at approximately 60 min intervals thereafter.

Limitations

This is a retrospective observational study with inherent limitations associated with the study design. Our report represents an evolving clinical practice reinforced by a culture of measurement and review, visible by multiple dosing protocol adjustments over time, with the most recent iteration of the delivery protocol being adopted in 2023. The choice of variables for the propensity matching is an inherent limitation of the statistical methodology, and additionally, we have not performed sensitivity analyses on the route of cardioplegia administration or surgeon performing the procedure. The variation introduced by being unable to propensity match surgeons is an inherent limitation of the retrospective study design. It is important to note the median number of cardioplegia doses in the 90 min ACX time primary analysis was one, therefore inference on multi-dosing is based on the sub-analysis alone. With regards to clinical outcomes one of the major concerns regarding multidose DNC is lidocaine toxicity and its hypothesized results on myocardial contractility and arrhythmia. Arrhythmia and a quantitative reproducible measure of cardiac output such as inotropic requirements or cardiac index post-operatively were not included. Similarly, rates of return to theatre were higher in the DNC cohort however inferences on the manner and cause of this were not included. A further limitation is that we have not reported a 24-hour Troponin value as one of the two hospitals did not collect this time point, and our troponin profiles are not indexed to renal function nor type of procedure which can introduce confounders to their interpretation relative to these variables. Finally, a significant consideration is the learning curve for the operative team in adopting a new cardioplegic strategy, this will inadvertently introduce patient selection bias and additional confounders about timing of redosing, choice of cardioplegic agent, mode, and rate of delivery; which cannot be accounted for in a retrospective study design.

Conclusion

This is the largest quantitative cohort study on DNC use in patients with prolonged AXC time in Australia or New Zealand and adds significantly to the limited international reports. Del Nido cardioplegia in more complex and longer procedures has demonstrated equivalent myocardial protection compared with multiple doses of HKB cardioplegia. This study is a quality assurance measure driven by clinical experience and our departmental morbidity and mortality review processes. Randomised, multi-centre trials are needed to develop an evidence-based protocol for multi-dose DNC.

Funding

No financial funding was required for this project.

Conflicts of interest

The authors declare no conflict of interest.

Data availability statement

Data use agreements restrict the distribution of raw study-related data files. Requests for data will be reviewed by the ANZCPR Steering Committee.

Authors contribution statement

Malgorzata (Maggie) Szpytma: Drafting of manuscript, data interpretation, input on drafts, approval of final manuscript.

Damian Gimpel: Project conception, input on drafts, approval of final manuscript.

Jordan Ross: Input on drafts, approval of final manuscript.

Richard F. Newland: Project conception, design and performance of data analysis, input on drafts, approval of final manuscript.

Gareth Crouch: Contributed data, input on drafts, approval of final manuscript.

Gregory D. Rice: Contribution and interpretation of data, input on drafts, approval of final manuscript.

Jayme S. Bennetts: Contribution and interpretation of data, input on drafts, approval of final manuscript.

Robert A. Baker: Project conception and supervision, design and contributed to data analysis, drafting of manuscript, input on revisions, approval of final manuscript.

Ethics approval

Ethics approval for this audit was granted by the Southern Adelaide Clinical Human Research Ethics Committee and the South Australia Local Health Network Office for Research (Quality Registry ID: 2265).

References

All Tables

All Figures

	Figure 1 Consort diagram. The 178 patients either had both DNC and HPK cardioplegia (55 cases) or were excluded as the surgeons performing operations did not use both cardioplegic techniques (112 cases). Abbreviations: DNC – Del Nido cardioplegia, HKB – hyperkalemic blood cardioplegia.
In the text

	Figure 2 Box and whisker plot for Troponin profile based on cardioplegia strategy for greater than 90-minute ischaemic time. Solid middle bar is the median, top and bottom of box the 75th and 25th percentile, with upper and lower adjacent values.
In the text

	Figure 3 Troponin profile based on cardioplegia strategy for greater than 120-minute ischaemic time. Solid middle bar is the median, top and bottom of box the 75th and 25th percentile, with upper and lower adjacent values.
In the text