We thank, respectively, Drs. di Mario and Dimopoulos and Drs. Parent and Rinfret for their comments on our work (1). In fact, it was conceived on the basis of experimental and clinical evidence suggesting that local rapamycin or paclitaxel could have an adverse effect on circulating monocytes (2), respectively, that coronary endothelial function in humans is compromised up- as well as downstream of the stent eluting such drugs (3- 4). Thus, the “broader context” requested by Parent and Rinfret actually incited the study. It is, indeed, the case that the restenosis severity in the stented segment was pronounced, and that the question may be raised whether our study population is representative for all patients receiving drug-eluting stents (DES). In the context of a rather prominent in-stent restenosis, the reasoning by Parent and Rinfret can even be followed to the point where they hypothesize “resistance to the antiproliferative and immunosuppressive drugs” in the DES group. However, to link the proposed opposition to antiproliferation with impaired collateral growth is, even semantically, an oxymoron. This is because all the evidence of the past 40 years has very convincingly conveyed 1 message: collateral growth is intimately tied to inflammatory (i.e., proliferative) processes (5).

It is correct that after stent implantation, there is no stimulus for collateral development. However, our study has not focused on stimuli, but on potential anti-stimuli of collateral growth. Furthermore, rapamycin and paclitaxel have been shown by independent groups to elicit an adverse effect on endothelial function both up- and downstream of the implantation site. Drs. di Mario and Dimopoulos also claim that alternative ways of data analysis would have probably uncovered their frail character. First, to apply the proposed nonparametric statistical analysis provides 38 higher collateral flow index ranks in the bare-metal stent (BMS) than in the DES group, and 21 lower ranks (p = 0.0020). Second, it was maybe not pointed out clearly enough that ours was a study with a matched design for stenosis severity at the time of follow-up. In this context, Table 3 of our article has already provided the fractional flow reserve values (0.828 ± 0.184 vs. 0.838 ± 0.177) asked for by di Mario and Dimopoulos, thus documenting that the paired stenotic lesions were well comparable hemodynamically. Third, Table 3 of our article has displayed even more information (i.e., electrocardiogram [ECG] data against the argument that the CFI difference between the groups “is unlikely to be clinically relevant”: signs of ischemia during coronary balloon occlusion on intracoronary ECG were present in 33 of 60 patients in the BMS group, but in 50 of 60 patients of the DES group). The magnitude of ECG ST-segment elevation during a temporary occlusion has been and is employed as a surrogate marker for infarct size in case of a permanent occlusion for decades. As di Mario and Dimopoulos certainly know, infarct size is gradually and not binarily influenced by the product of coronary occlusion time, myocardial area at risk, oxygen consumption at and during the time of occlusion, and the inverse of collateral supply.