Background

The treatment of unprotected left main and of coronary bifurcation still represents a challenge for interventional cardiologists due to both procedural complications and higher restenosis rates compared with non-bifurcation lesions [1,2,3,4].

The complexity of the bifurcation milieu is rooted into the unique flow patterns that characterized them, with local low and oscillatory endothelial shear stress along the lateral walls of the main vessel and of the side branch, whereas high endothelial shear stress develops at the carina. This ultimatey leads to a prothrombotic and atherogenic flow-pattern, and even after treatment with percutaneous coronary intervention (PCI), it may increase the failure rate wit need for subsequent revascularization on the target lesion (TLR) and an increased risk of stent thrombosis (ST) [4,5,6].

In the last years, BP-DES (Bioresorbable polymer drug eluting stents) have been introduced with the rationale to potentially decrease ST. Differently from durable polymer stents (DP), after the elution of the antiproliferative drug the bioreabsorbable polymer is going to dissolve leaving behind a bare metal stent, thus reducing the local inflammatory reactions and therefore the risk of thrombosis related to a permanent polymer. Clinically, this translated into lower rates of TLR for BP-DESs implanted in coronary bifurcations in the LEADER-FREE [7], although this RCT was weakened by the comparison with a first generation stent. Regarding currently implanted second generation DES, both in the EVOLVE II trial and in a recent paper of Mennuni et al. [8, 9], BP-DESs were shown to be safe and effective as compared to durable polymer, although coronary bifurcations and LM were underrepresented (respectively about 4 and 15%) [10].

In light of the intrinsic limitations of the above studies, the RAIN study (very thin stents for patients with MAIN or bifurcation in real life: the RAIN, a multicenter study) was designed to evaluate the clinical performance of ultrathin stents in everyday clinical practice. We here present an analysis of the RAIN study aimed to evaluate the safety and efficacy of BP-DES in the bifurcation setting.

Methods

The RAIN is a large multicenter retrospective observational registry (ClinicalTrials NCT03544294, retrospectively registered; see Additional file 1 for enrolling sites). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki and are consistent with ICH Good Clinical Practice as well as regulatory requirements. It was approved by an institutional review committee, and all patients provided informed consent.

Inclusion criteria

The RAIN registry included all consecutive patients from June 2015 to January 2017 undergoing complex PCI involving LM and/or bifurcation with ultrathin stents (Promus Element, Xience Alpine, Ultimaster, Synergy, and Resolute Onyx; see Additional file 1 for more details on the stents involved in this study).

Baseline and procedural data

Cardiovascular risk factors, clinical presentation, angiographic features, use of IntraVascular UltraSound (IVUS), Optical Coherence Tomography (OCT) and Fractional Flow Reserve (FFR) were recorded, along with the characteristics of the implanted stents. IVUS or OCT was used prior to stent implantation to assess the severity of the stenosis and side branch involvement, and post stent implantation to evaluate dissection and the requirement for stent optimisation. The decision to post-dilate, to perform final kissing balloon (FKB), to assess intracoronary imaging and the choice of the stenting technique (provisional versus 2-stent), was at the discretion of the treating physician.

The above data were derived from electronic patient records at each center, while follow-up data were obtained from clinical assessment, telephonic consultations or via primary care physicians and then recorded online (http://www.cardiogroup.org/RAIN/index.php?cat=home).

Endpoints

The rate of TLR was the primary endpoint, while MACE (a composite endpoint of all-cause death, myocardial infarction, TLR and stent thrombosis) and its components, along with TVR were the secondary endpoints. The analyses were performed according to PCI strategy (provisional vs two-stent).

Statistical analysis

Categorical variables are reported as count and percentages, whereas continuous variables as mean and standard deviations or interquartile range (IQR). Gaussian or not Gaussian distribution was evaluated by Kolmogorov-Smirnoff test. The t-test has been used to assess differences between parametric continuous variables, Mann-Whitney U test for non parametric variables, the chi-square test for categorical variables and Fisher’s exact test for 2 × 2 tables. The a priori statistical significance level was set at α = 0.05. To account for clustered data among centres we used a normal regression (ANOVA) approach with a fixed effect for cluster and an effect for group when data were normally distributed. For non-normal data, we used the Wilcoxon rank sum test modified to account for clustering. For propensity score, first logistic regression analysis was done for all baseline features that differed between BP-DES and DP-DES, matching was computed after division into quintiles and methods of the 1:1 nearest neighbor on the estimated propensity score [11]. Calibration was tested with Hosmer-Lermeshow, and accuracy was assessed with Area Under the Curve. Standardized differences were evaluated before and after matching to evaluate the performance of the model. All statistical analyses were performed with SPSS 21 and differences were considered significant at α = 0.05.

Results

Before propensity score with matching

At the end of the enrolling period 3001 patients had been recorded: 2120 were treated with DP-DES ad 881 with BP-DES (see Fig. 1).

Fig. 1
figure 1

design of the study

Full size image

At baseline patients in the BP-DES were more hypertensive (28.8% vs 23.2% in the DP group, p < 0.01), with a higher rate of previous MI (35.3% vs 26.6% in the DP group, p < 0.01) and they were more often admitted due to STEMI (19.1% vs 16%, p < 0.01). On the other side, patients with DP-DES were more often hyperlipidemic (58.1% vs. 54.6% in the BP group, p < 0.01; see Additional file 1: Table A).

Patients with BP-DES more frequently presented diffuse coronary disease (55.1% vs 29.9%% in the DP group, p < 0.01), and with true bifurcation involvement (i.e. Medina 1,1,1 or 0,1,1; 28.1% in the BP group vs 17.5% in the DP group, p < 0.001). They were less frequently treated with two-stent technique (76.9% vs 82.4% in the DP group, p < 0.01). Finally, there was a similar rate of total ULM disease (30% vs 26.9%, p = 0.45) (see Additional file 1: Table B).

After propensity score with matching

After multivariate adjustment, 700 patients for each group were selected. Baseline clinical features were comparable, with similar rates of presentation for STEMI (19% in the DP group vs. 18.1% in the BP group, p = 0.1, see Table 1 and Fig. 1).

Table 1 Baseline Characteristics Post PSWM
Full size table

At angiography, similar percentages of patients had a LM disease (23.1% in the DP group vs 27.1% in the BP group, p = 0.87) and true bifurcation involvement (19.7% in the DP group vs 21.9% in the BP group, p = 0.5). Provisional strategy was successfully performed in both groups in the majority of patients (82.1% in the DP group vs. 82.2% in the BP group, p = 0.68, see Table 2).

Table 2 Interventional Characteristics post PSWM
Full size table

At a median follow up of 16 (12–22) months, rates of TLR (3.7% vs 2.9%, p = 0.22) and MACE were similar (12.3% vs. 11.6%, p = 0.74), without significant differences among all the secondary endpoint (see Fig. 2).

Fig. 2
figure 2

outcomes at follow-up. Blue columns represent the percentage of events among patients receiving a DP-DES, red columns represent the percentage of events among patients receiving a BP-DES

Full size image

At sensitivity analysis for patients treated with two stents strategy, patients treated with BP-DES showed a better outcome in term of MACE (20.4% vs 10% in the DP group, p = 0.03) and TVR (12% vs 4.6% in the DP group, p = 0.05, see Fig. 3) and a trend towards TLR (3.7% vs 2.9%, p = 0.22).

Fig. 3
figure 3

outcomes at follow-up in the subgroup of patients in whom a coronary bifurcation was treated with a 2-stent technique. Blue columns represent the percentage of events among patients receiving a DP-DES, red columns represent the percentage of events among patients receiving a BP-DES

Full size image

Discussion

Our main findings may be summarised as follows:

  • Among patients treated for Left Main disease, the risk of MACE was similar for BP-DES and DP-DES.

  • Among patients treated with two stents strategy in both LM and non-LM bifurcation involvement, patients treated with BP-DES showed a better outcome in term of MACE and TVR.

  • No differences in ST are evident in BP-DES compared to DP-DES group.

To the best of our knowledge, this is the first real-world, observational registry evaluating the safety and efficacy profile of different ultrathin stents (struts thinner than 81 μm) in patients with a ULM stenosis treated with PCI using newer-generation abluminal BP- DES as compared to the DP-DES.

BP-DES have been developed to combine the best of both family of metallic stents, i.e. the efficacy of DES and the late safety associated with BMS. However, the available evidences on cardiac death, MI, or stent thrombosis are still scarce [12,13,14,15,16,17]: network meta-analyses have indicated an excess risk of BP-DES with regard to MI or stent thrombosis when compared with DP-EES (i.e. Xience, Abbott Vascular, Santa Clara, California), though their results were restricted due to heterogeneity of devices in the BP-DES group and to limited follow-up duration [12,13,14]. Of note, the meta-analyses by Kang et al. [12] and by Navarese et al. [14] included BP-BES trials using the Biosensors BioMatrix device (Biosensors International, Singapore), the meta-analysis of Bangalore et al. [13] included also trials using an other sirolimus-eluting stent (SES) with biodegradable polymer (Yukon Choice PC, Translumina, Hechingen, Germany) while the one by Cassese et al. [15] included trials on the ultrathin sirolimus-eluting Orsiro device (Biotronik, Bülach, Switzerland).

The principal finding of our current analysis is that BP-DES actually showed a similar safety and efficacy profile at 5 years compared with the DP-DES. Notably - apart from ST - event rates in our study were low keeping into account the clinical and anatomical complexity of the enrolled patients, and similar to previous studies on new-generation DES [16, 17]. These data reflect the global improvement in quality and safety outcomes of these devices mainly due to either the effect of anti-proliferative drug on restenosis and either to the thin struts technology that facilitate the endothelial coverage. However in our study the whole population analysis did not show differences between patients who received BP-DES compared to DP-DES: this might be due to the overwhelming benefit of the thin strut design over the polymer material: the difference in adverse events is in fact observed during period of time that is shorter than the complete polymers dissolution time. These findings highlight the importance of the overall DES design and biocompatibility on the clinical performance of contemporary DES [3]. The effect of stent’s struts thickness has been well established already, with thinner struts showing to produce less inflammation, vessel injury, neointimal proliferation, as well as thrombus formation when compared with thicker ones [18, 19]. Furthermore, the 5-year analysis of the COMPARE II trial also confirmed the early- and mid-term similar safety and efficacy of the BP-BES and the DP-EES, thus challenging the concept itself of the biodegradable polymer coating [20, 21]. On the other hand, Bayesian analysis in the BIOFLOW-V trial, despite limited by the analysis itself and by the particular thickness of the stent used, is encouraging for an actual role of bioresorbable polymers [22, 23]. Therefore, whether BP-DES are as safe and effective as DP-DES should be proven for each specific stent by an appropriately designed clinical trial.

The second interesting result is the trend towards better outcome in term of MACE and TVR in patients treated with BP-DES and a two stents strategy. The higher risk of subsequent events in this challenging anatomical subset is well known, and it is related either to patient either to technique drawbacks [23, 24]. Despite the low percentage of patients included in this analysis (about 20% of the patients selected by propensity score), we can speculate that in an high risk setting like a double stent bifurcation treatment in left main disease, the added value of a bioresorbable polymer in terms of vessel recovery as above explained might really make the difference in terms of outcomes. Of course, this thesis should come on top of more strong assumption about the outcomes related to the use of two stents in bifurcation, most important the technique used [2, 23].

As regard the percentage of stent thrombosis we found, we should not forget that such high numbers are actually in line with other real life reports with comparable follow-up [24]. Furthermore, our study enrolled consecutive all-comers patients, reflecting real practice outcomes that are obviously different from the safer environment of randomized trial. The percentages in the whole population analysed are actually similar to what reported in the COMPARE I trial (A Trial of Everolimus-Eluting Stents and Paclitaxel Stents for Coronary Revascularization in Daily Practice) of 1.8% [25] and of the LEADERS trial (Limus Eluted From a Durable Versus Erodable Stent Coating) of 2.6% [8]. These findings stress the importance of correct identification of critical bifurcation disease by means of fraction flow reserve or intravascular imaging [26, 27] and the value of longer double anti-platelet therapy in patients with higher thrombotic risk [28], especially when this anatomical feature comes together with higher clinical risk situation like ACS [29,30,31].

Limitations

There are several limitations to this study. First, this was not a randomized controlled study: with the use of propensity score we could not adjust for variables which were not possible to collect (e.g. the experience of the physicians performing PCI) or others among multiple factors that might be implicated in DES acute thrombogenicity and long-term vascular healing (e.g. the the polymer biocompatibility, its composition and distribution, and -in case of bioresorbable polymers- the duration of bioresorption [3, 17, 21]). It is difficult to assess the different contribution of all these confounding factors in a single study. Furthermore, the amount of patients who received a two-stent treatment for bifurcation disease is small compared to the one who received the provisional approach. Finally, although designed as an all comers study, only 23% of patients undergoing percutaneous interventions were actually enrolled in the study, so selection bias cannot be entirely ruled out.

Conclusion

The main message is that BP-DES look as safe as DP-DES even in high anatomical risk setting like LM disease. Furthermore, the provisional approach confirms itself as the safer on the long term. Finally, when a two-stent strategy is absolutely needed, the trends toward improved clinical outcomes with respect to MACE and TVR we found with BP-DES might potentially be of value to speculate about the stent choice in selected high-risk subgroups of patients at increased risk of ischemic events.