Letter by Barlis et al Regarding Article, “Two-Year Clinical Outcomes With Drug-Eluting Stents for Diabetic Patients With De Novo Coronary Lesions: Results From a Real-World Multicenter Registry
To the Editor:
The role of drug-eluting stents within the field of interventional cardiology has recently come under close scrutiny in the wake of increasing concerns about late stent thrombosis and doubts on the overall cost-effectiveness of these devices. The prevailing view among experts, however, is that this technology retains an important role, particularly in patients at high risk of restenosis such as those with diabetes mellitus, as suggested by randomized trials.1 It was thus with great interest that we read the article by Ortolani et al,2 which presents the results of a nonrandomized prospective registry on the long-term clinical outcome of diabetic patients undergoing percutaneous coronary intervention with drug-eluting or bare metal stents. The authors should be commended for this systematic collection of “real world” data and the 100% 2-year follow-up achieved. Counterintuitively, however, this study demonstrated a limited benefit from the use of drug-eluting stents in diabetic patients, a benefit only seen in non–insulin-dependent patients. Before accepting the results of this study as representative of “real-world” percutaneous coronary interventions, a number of points must be considered.
As expected in a nonrandomized study, significant differences in numerous clinical and especially angiographic variables can be found between the drug-eluting stent and bare-metal stent groups. The authors used propensity score regression adjustment rather than stratification or matching to account for these differences and present the C statistic and Hosmer-Lemeshow test as indicators of adequacy of the propensity model. Doubts, however, remain on the efficacy of this technique to balance covariates between such heterogeneous groups, which is the primary purpose of propensity score adjustment. Moreover, only the grouping variable and propensity score were included in the model, even though inclusion in the regression model of all (or important) covariates that were used to build the propensity score has been shown to improve estimation.3 Matching or stratification with demonstration of balance in this large population would have been a welcome “sensitivity analysis,” removing any doubts on the appropriateness and adequacy of the propensity score adjustment.
Lack of benefit in insulin-dependent diabetic patients is also highly counterintuitive. Insulin dependence with no indication of the level of diabetic control (eg, HB1Ac levels) is a generic classification for a prospective study. The lack of angiographic follow-up in a study addressing diabetic patients is another substantial limitation because “silent” events are frequent in this population and may lead to late clinical events beyond the 2-year follow-up period.4 Moreover, as the authors point out, the incidence of stent thrombosis was likely underestimated given that only the angiographically confirmed thrombosis rate was reported. A more contemporary and informed approach would have used the current Academic Research Consortium classification into definite, probable, or possible stent thrombosis.5 Finally, outcome differences between different types of drug-eluting stent may have been of interest.
Ortolani P, Balducelli M, Marzaroli P, Piovaccari G, Menozzi A, Guiducci V, Sangiorgio P, Tarantino F, Geraci G, Castriota F, Tondi S, Saia F, Cooke RM, Guastaroba P, Grilli R, Marzocchi A, Maresta A. Two-year clinical outcomes with drug-eluting stents for diabetic patients with de novo coronary lesions: results from a real-world multicenter registry. Circulation. 2008; 117: 923–930.
D'Agostino RB Jr. Propensity scores in cardiovascular research. Circulation. 2007; 115: 2340–2343.
Biagini E, Schinkel AF, Bax JJ, Rizzello V, van Domburg RT, Krenning BJ, Bountioukos M, Pedone C, Vourvouri EC, Rapezzi C, Branzi A, Roelandt JR, Poldermans D. Long-term outcome in patients with silent versus symptomatic ischaemia during dobutamine stress echocardiography. Heart. 2005; 91: 737–742.
Cutlip DE, Windecker S, Mehran R, Boam A, Cohen DJ, van Es GA, Steg PG, Morel MA, Mauri L, Vranckx P, McFadden E, Lansky A, Hamon M, Krucoff MW, Serruys PW. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation. 2007; 115: 2344–2351.