Drug-Eluting Stent Solutions: Comparing Randomized Clinical Trials and Registries

What are some major differences in patient follow-up between a randomized clinical trial and a registry? In recent drug-eluting stent (DES) randomized trials, we’ve come to appreciate how difficult it is to get our arms around the issue of clinical-driven ischemic target lesion or target vessel revascularization (TVR). Follow-up which can be angiographic or clinical may help us with this issue. TVR rates in RCTs with angiographic follow-up will always be higher than in RCTs that have only clinical follow-up. In angiographic follow-up, the patient undergoes re-catheterization usually within nine months following the initial procedure after which the cine is sent to a core lab to be compared with the original and post-implant cines. A clinical follow-up does not include the same degree of rigor, as it can either be in the form of the patient presenting at the clinic or a phone conversation follow-up, but with no diagnostic procedure involved. With angiographic follow-up as commonly dictated by protocol in an RCT some patients arrive at the cath lab with minimal or no symptoms, yet will still have an 80-90% lesion. As registries usually include broader patient criteria and presumably more complex lesions (mimicking their occurrence in clinical practice), we would expect more patient events which should lead to increased ischemic TVR. Yet registry data uniformly reports markedly lower TVR rates, even though the patients are more complex and more likely to have restenosis, a result of primarily clinical follow-up, not angiographic as in RCTs. As a physician, what do you look for in randomized clinical trials versus registries? As previously stated, I look to RCTs to answer focused, critical questions that are statistically powered to show the superiority or non-inferiority of one treatment over another. For registry data, I want to see that the number of complex patients, such as saphenous vein graft (SVG) lesions, AMIs, diabetics (especially insulin-dependent), and bifurcations fit with what I see clinically in my practice. Secondly, in RCTs we have to be wary of subset analysis that is not pre-defined. This is often provocative and hypothesis-generating, but we have to realize that the analysis is not powered to be definitive and should not be interpreted as such. Also in registries, the same issue of subset review is at play. Remember, no core lab exists and no one reviews the cines, so operator-reported percentages of stenosis and lesion length are data points that are best left out of any analysis of registry data. In the DES world, randomized trials have compared DES to bare metal stents (BMS). As the BMS control group diminishes, what do you see in the future for randomized trials? If the comparison becomes DES-to-DES, what would you see as the primary endpoint? We can no longer ethically randomize patients to BMS in a DES world. Thus, the primary endpoint will be the comparison of major adverse cardiac events (MACE) carefully monitored and adjudicated with respect to 30-day death, MI, sub-acute thrombosis (SAT), ischemic TVR and 1-year TVR. I think we’ll also demand 2-4 year follow-up with DES to be sure we don’t see the same five-year late catch-up in TVR rates that we had with brachytherapy. Do you have any observations of the ARRIVE I and II registries that you could share? Dr. John Lasala, the principal investigator for ARRIVE I, reported the 30-day data for TCT. We were happy with the real world reported acute and sub-acute stent thrombosis rate of 1.3% and with the quality of the data submitted by the enrolling sites. Importantly, in both registries, an effort was made to be sure we captured consecutive patients. I think we were successful on this point. Are there certain terms that translate from trials to registries? MACE is a composite measurement designed to capture all major cardiac events in an RCT, whereas cardiac adverse event is a measurement in registries reported by the site without the benefit of core lab verification or the inclusion of all events. Cardiac adverse event rates are often a composite of cardiac death, MI and reported re-interventions of the target vessel. It’s important the two are not used interchangeably unless pre-specified definitions are exactly alike. For instance, whether a revascularization event is ischemic-driven is something much easier to define in an RCT than in a registry. Also, it’s important to remember that differences in definitions for clinical events like sub-acute thrombosis do exist. For instance, in TAXUS IV, stent thrombosis was defined as angiographically-proven thrombosis or unexplained cardiac death, which occurs within 30 days of stent implant, the same definition we use in ARRIVE. However, other registries and even some clinical trials have defined SAT as only angiographically proven, which will definitely give a lower number. Just remember to compare apples to apples! Sponsored by Boston Scientific Corporation. ‘MACE’ is a composite measurement designed to capture all major cardiac events in an RCT, whereas ‘cardiac adverse event’ is a measurement in registries reported by the site without the benefit of core lab verification or the inclusion of all events.

NULL