In this study, we compared two strategies of treating high-grade symptomatic aortic stenosis in combination with coronary artery disease. The interventional treatment followed a staged procedure protocol with treatment of the coronary artery disease first and the TAVI procedure following, whereas in case of surgery, a one-procedure strategy was followed.

Interestingly, despite higher initial bleeding and stroke incidences in the SAVR + CABG group, early mortality was not significantly different, although there was a trend towards lower mortality after TAVI and PCI at 30 days. However, later timepoints indicated a higher mortality in the TAVI + PCI group, which became a significant effect at three years after aortic valve replacement in the unmatched population. In the propensity score matched cohorts that were adequately balanced with regard to age, risk scores and comorbidities, this effect lost its statistical significance.

Several other studies have investigated the strategy of treating AS and CAD in elderly patients. Most recently, Khedi et al. have published the 12-month results from a multicentre, prospective, non-inferiority, randomised controlled trial comparing FFR-guided PCI plus TAVI versus SAVR plus CABG. They applied a complex primary endpoint consisting of all-cause mortality, myocardial infarction, disabling stroke, clinically driven target-vessel revascularisation, valve reintervention, and life-threatening or disabling bleeding at twelve months after treatment. The trial was positive meeting the non-inferiority criteria with signs for superiority for the FFR-PCI/TAVI group [16]. However, in addition to major technical differences (such as the randomised design and use of FFR-guided PCI) the trial also differs from our data in having a smaller patient population and a shorter follow-up period.

If only one of the two pathologies is singled out, randomised trials or their meta-analyses already indicate trends: In the SYNTAX trial, CAD was treated by either CABG or Taxus stents (first generation drug eluting stents). Though this stent platform has been outperformed meanwhile [17], its use allowed for non-inferior outcome of mortality in low- or mid-complex CAD patients (SYNTAX Score ≤ 32), whereas higher complexities (SYNTAX Score > 33), in particular three-vessel disease, displayed an increased rate of major adverse cardiac and cerebrovascular events (MACCE) after five years [18]. This finding was validated by a meta-analysis of 11 trials comprising 11.518 patients with multivessel disease, which showed a five-year all-cause mortality of 11.5% after PCI vs. 8.9% after CABG. The mortality difference was especially significant among diabetic patients (15.5% vs. 10.0%), but was not observed in non-diabetic patients or those with left main disease [19]. Moreover, a large meta-analysis of TAVI and SAVR in seven randomised controlled trials across high, intermediate and low risk levels (PARTNER 1 A, 2 A, 3, CoreValve US, NOTION, SURTAVI) provided follow-up for more than three years and revealed that an initial survival advantage of TAVI vs. SAVR in the first six months was lost and reversed to favouring SAVR over TAVI after 24 months [20].

These results suggest that a combination of both entities could replicate an early advantage of the interventional strategy, but also a long-term disadvantage, at least in certain patient groups. Several other reports have been probing the competing strategies of either complete interventional or complete surgical or hybrid treatments. In 2018, Barbanti and colleagues reported results from the Italian OBSERVANT multicentre cohort study. In 472 propensity score matched patients for either strategy, the TAVI + PCI group (with 92% of PCIs done before and 8% at the time of TAVI) did not statistically differ from the SAVR + CABG cohort, although there were 25% fewer deaths in the surgical group after three years (59 vs. 79). Notably, the observed mortality rates at the 3-year timepoint were very similar to our results with 33.5% in the interventional and 25.0% in the surgical group [21].

In 2019, Søndergaard and colleagues used the prospective SURTAVI data set of 1660 intermediate-risk AS patients and analysed 332 cases of concomitant coronary revascularisation for two years. The distribution pattern of CAD was mild (including a SYNTAX Score mean of 8.3%), with > 60% of cases presenting as 1-vessel treatment and < 10% in each group requiring 3-vessel approaches. As a result, the 169 interventionally treated patients were statistically indistinguishable from the 163 surgically treated patients with mortality rates of 16% vs. 14% at two-year follow up (p = 0.62) [22].

Following up, in 2021, Alperi and colleagues included 156 pairs of propensity score matched patients with complex coronary artery disease (SYNTAX Score of 26.3 vs. 26.9 and left main involvement of 55.8% vs. 57.1% in the interventional vs. surgical group, respectively) in their analysis. Again, no statistical difference was found in mortality at five years between interventional (38.1%) and surgical patients (32.0%), though the follow-up was unevenly distributed, with the interventional cases at five years too rare for statistical response [23].

In contrast, Baumbach and colleagues reported a prospective single centre cohort study of 626 patients with either surgical (SAVR + CABG, 464 patients), interventional (TAVI + PCI, 112 patients) or hybrid treatment (TAVI + OP/MIDCAB, 50 patients). Three-year follow-up revealed a significant survival adavantage in the surgical group (80% vs. 49%, p < 0.001). Of note, the mortality in the hybrid group resided exactly with the purely interventional group. However, the potential impact of this study’s finding was limited by age, EuroSCORE I and left ventricular function all significantly favoring the surgical group, in addition to a lack of propensity score matching and a follow-up adherence of 53% for the surgical cohort or even lower for the other groups [24]. Nevertheless, this study is particularly noteworthy as it provides the first significant evidence suggesting a potential long-term advantage of surgical treatment for AS and CAD. In our study, even with propensity score matching and a follow-up adherence of 98.1% for one year and 90.9% for three years, the surgical group was trending towards a survival advantage until the end of the three-year observation period.

Moreover, Amat-Santos and coworkers published a propensity score matched retrospective study of the Spanish TAVI registry in 2024, revealing significantly higher rates of the combined primary endpoint of stroke and mortality in the surgical group after 30 days and one year (15.8% vs. 9.9%, p = 0.033) in a population at low risk (STS-PROM of 3.0% and 3.4% after matching) [25]. Of note, the mortality at one year was similar in both groups (10.1 vs. 7.1%, p = 0.291), bearing similarities to our study in an intermediate-risk cohort.

Further research is warranted beyond the existing prospective and retrospective data to optimise treatment strategies for intermediate-risk patients presenting with AS and CAD.

There are limitations to this study. Firstly, because of its retrospective design, acquiring certain types of data, e.g. causes of death, was severely limited, resulting in the fact that cardiovascular mortality in particular could not be reported for the entire study population. A subgroup analysis on cardiovascular mortality is included in the supplementary data under 4). Secondly, data on the SYNTAX Score was missing for 16 patients in the TAVI + PCI group and 8 patients in the SAVR + CABG group, which is why this variable could not be included in the propensity score matching and remained significantly different between the two treatment groups. In addition, there are potentially important factors for clinical outcome that were not examined in this study, including complete revascularisation and prothesis degeneration. Subgroup analyses for coronary and valve reinterventions, however, can be found in the supplementary data under 5). Thirdly, the EuroSCOREs that were used as an inclusion criterion in this study are now outdated and surgical risk scores in general are not entirely suitable for TAVI patients. Lastly, interventions that date back to 2012 were included in this study and the techniques and materials used in percutaneous procedures have significantly improved since then.