The percutaneous coronary intervention prior to transcatheter aortic valve implantation (ACTIVATION) trial: study protocol for a randomized controlled trial

Among the possible advantages of revascularization prior to TAVI may be a protective effect against the ischemic burden of the procedure, including as it does periods of hypotension. The absence of contractile reserve is associated with increased mortality after sAVR [12], and significant stenoses not intervened upon could contribute to this. Surgical revascularization for multi-vessel coronary artery disease has been found to be an independent factor predictive of improvement of left ventricular ejection fraction (LVEF) after sAVR [13], and similar benefits for revascularization by PCI may exist. Improving coronary flow in symptomatic patients with significant flow-limiting stenoses may maximize this beyond the valvular intervention. Wave intensity analysis of coronary flow has demonstrated marked reduction in the diastolic suction wave in aortic stenosis (the dominant wave in coronary perfusion [14]), which significant stenosis may impair further [15].

The risks of PCI are well described: death, myocardial infarction, coronary artery bypass grafting (CABG), stroke, vascular access complications, renal insufficiency, allergy and stroke/TIA [16, 17]. PCI is also associated with an incidence of stent thrombosis of up to 1% with significant mortality [18, 19]. The presence of severe aortic stenosis could have a detrimental effect upon the ability to withstand these. Indeed, the hypotension experienced by patients during the TAVI procedure (especially during rapid right ventricular pacing) may actually increase the risk of stent thrombosis. There has been increasing recognition recently of the adverse outcomes associated with major bleeding post-PCI [20, 21]. Current opinion is to continue aspirin for life and clopidogrel for one month after bare metal stenting and 12 months after insertion of a drug-eluting stent. This would possibly impact upon bleeding complications in transfemoral, subclavian and transaortic TAVI and would certainly be of concern in patients undergoing TAVI via the transapical approach, given the direct myocardial access required. Angiography in the 24 hours, or even the 5 days, prior to sAVR has been shown to be associated with acute kidney injury post-operatively [22]. These risks may also exist if PCI and TAVI are inadequately separated or combined in a hybrid procedure. Finally, patients would require two admissions with attendant implications to cost and risk. This latter consideration might be offset against the possible reduction in length of stay in the post-TAVI period.

The data on coronary artery disease in TAVI and PCI in this context is certainly provocative, but there are several major limitations, given that the studies are all nonrandomized, registry-type with relatively small sample sizes.

The lack of data on pre-procedural revascularization has led to some variability between centers on the level of coronary interventions undertaken prior to implantation. A retrospective analysis by Masson et al. stratified TAVI patients according to myocardium at risk due to coronary artery disease, finding no difference in 30-day or 1-year mortality between the groups [4]. However the small group sizes involved would have required a mortality difference between even the two most extreme groups at 30 days of greater than one third to gain significance, limiting our use of this data. The access approach used was not included in their risk model, an important omission given the known differences in risk profiles between, for example, TF-TAVI and TA-TAVI cohorts. In the German TAVI registry, CAD was defined as previous revascularization (either PCI or CABG) OR a stenosis of ≥50% and was noted in 62% of 1,382 patients. These patients had a greater in-hospital mortality (10.0 versus 5.5%, P <0.01), required more frequent cardio-pulmonary resuscitation (7.8 versus 3.5%, P <0.01) and suffered greater 30-day mortality (log rank P = 0.041). But there were significant differences in the group demographics: more frequent peripheral vascular disease, lower left ventricular (LV) ejection fraction and higher logistic EuroScores in patients with CAD [23]. The Italian CoreValve registry noted that in patients with critical ostial disease, myocardial infarction (MI) within 12 months of TAVI was greater in those patients who were not revascularized prior to TAVI [24]. A recent meta-analysis of the effects of CAD upon outcome in the midterm showed no effect; but again, of seven included studies, five defined CAD as previous revascularization, and the remaining two used 50% stenosis severity [25].

The presence of pre-existing treated coronary artery disease has been identified as an adverse risk factor for procedural and long term outcome [5]. With regards to the effects of pre-TAVI PCI, Masson and colleagues performed a sub-analysis of just 15 patients in whom PCI was performed at the discretion of the individual cardiologist: the resultant selection bias in turn clouds the interpretation of any findings. The mortality in the overall group one-year after TAVI was reported as 77.9% and was found to be 80% in those with pre-emptive PCI - though no comparison between the two cohorts was performed [4]. Abdel Wahab et al. reported that 55 patients who underwent PCI prior to TAVI showed no difference in adverse events at either 30 days (2% versus 6%; P = 0.27) or 6 months (9% versus 14%; P = 0.42) when compared to a group undergoing isolated TAVI [26]. The feasibility of PCI for CAD in patients with severe AS has been demonstrated by Goel et al. in a cohort of 258 patients over a 10-year period with a favorable comparison with propensity matched patients without the aortic disease [27].

There has been a rapid, global expansion in the use of TAVI to treat aortic stenosis in patients who are not candidates for sAVR. The efficacy of this technique has been successfully demonstrated in randomized, controlled trials [28, 29] and in large registries [30, 31], and there is now a need to improve patient safety and outcome, including how to manage patients with concomitant CAD. Given the paucity of data on the issue and variation in practice and recommendations we feel a randomized controlled trial (RCT) is essential for safe evidence-based practice. ACTIVATION is the first randomized controlled trial of elective PCI prior to TAVI. Its findings will help define the optimum revascularization strategy in this procedure and help create evidence-based guidelines on this controversial issue.

ACTIVATION is a prospective, randomized, open-label trial that aims to test the hypothesis that revascularization of significant coronary artery disease by PCI prior to TAVI (Test arm) will result in a rate of mortality and re-hospitalization at 12 months that is non-inferior to TAVI without such revascularization (control). Because the control patients in this study are receiving a CE/FDA approved valve replacement, a trial arm comparison would not be likely to demonstrate superiority in either direction.

The trial has an intentional non-inferiority design. The safety of TAVI with revascularization performed as deemed necessary by the Heart Team has been demonstrated in previous trials and registries [29, 30, 32]. From the available data detailed in the statistical design section, our hypothesis is that the no PCI strategy will have comparable effectiveness to the PCI (the active control) strategy, hopefully helping the cardiology community decide whether the extra procedure of pre-TAVI PCI can be safely avoided. From an ethical standpoint, a sham procedure in the no PCI arm could not be countenanced. Given the complexities of the TAVI procedure and the recommended Heart Team approach, we could not blind operators as to whether patients had undergone revascularization or not. To minimize bias towards the null, we have robust inclusion/exclusion criteria and clear end-point definitions and boundaries. Any deviation from the protocol in delivery of strategy will be carefully recorded, including any concomitant therapies.

Patient eligibility will be confirmed by the study investigators according to the inclusion and exclusion criteria in Table 1.

The primary endpoint is a composite of mortality and rehospitalization at 12 months since randomization. Mortality is defined as death due to any cause, the exact nature and date of which will be recorded. All deaths will be considered cardiovascular-related unless there is documentation to the contrary.

These consist of:

The definitions are in accordance with the Valve Academic Research Consortium (VARC) guidelines and its update and are specified in Table 2[33].

Ethics review was approved by United Kingdom National Research Ethics Committee (London Dulwich) (11/LO/1596).

Randomization will take place once the patient has provided written, informed consent. This will be performed via an online portal. Randomization of the treatment assignment is based on the block method using randomly varying block sizes and will be stratified by center. Once the patient is randomized s/he will be monitored for the duration of the follow-up irrespective of subsequent clinical management. The intended timeframe for PCI should be stated at randomization. After PCI the patient should receive dual anti-platelet therapy for 30 days - one of which can then stop if the patient receives a bare-metal stent (BMS). All PCI patients should undergo TAVI within the 2 weeks after this 30-day post-PCI time point. Any clinical events which occur in the test arm during this period will be attributed to the percutaneous revascularization on an intention-to-treat-basis. The flow diagram in Figure 1 illustrates the recruitment and study process.

ACTIVATION will be conducted at up to 20 hospitals in Europe providing TAVI using the Edwards system. Trial administration and data management will be carried out by an independent clinical trial organization, the European Cardiovascular Research Center, Massy, France. The trial is registered at http://​www.​controlled-trials.​com (ISRCTN75836930).

For the primary endpoint analysis, cumulative event rate will be calculated using the Kaplan-Meier estimates, and the Greenwood standard errors for these estimates. A 95% one-sided upper confidence limit will be computed for the difference in the primary composite outcome of mortality of rehospitalization at 12 months between arms (Test - Control). The Test arm (PCI) will be judged not inferior to the Control arm (no PCI) if the upper confidence limit is less than 7.5% - the predetermined non-inferiority margin. This was chosen based upon expert opinion and precedent (PARTNER cohort A [29]). In addition to formal analysis of non-inferiority endpoints, the Kaplan-Meier curves will be presented and compared by log rank test by treatment group and hazard ratio and its 95% confidence interval will be calculated using Cox regression model. Also, the combination of the all cause mortality and time to first recurrent hospitalization will be analyzed using the unmatched win ratio approach [35].

For secondary time-to-event outcomes, event rates at 30 and 365 days will be calculated using the Kaplan-Meier method and compared by log rank test. Their hazard ratios will be estimated using Cox regression model. Outcomes with repeated measurements will be analyzed using generalized linear mixed models. Continuous variables will be summarized using number of observations, median (interquartile range) or mean (SD) depending on variable distributions, whereas categorical variables will be summarized by the number and percentage of events. Chi-square tests and Mann-Whitney methods will also be used for comparative purposes. All analyses will be performed both on intention to treat and as treated populations.

The trial is now actively recruiting at centers in the United Kingdom, France and Germany.

We are optimistic that the trial has achievable enrollment targets and will answer an important question in this nascent field.