Compared with first-generation drug-eluting stents (DES) that harbored the risk of stent thrombosis, new-generation DES with durable polymers (DP-DES) have successfully lowered the risk of stent thrombosis while maintaining the lower rate of in-stent restenosis, compared with bare-metal stents (1, 2). In addition, advances in stent manufacturing technology enabled the reduction of stent strut thickness while securing sufficient radial force, and the development of a biocompatible polymer that allows stable release of anti-proliferative drugs has significantly reduced the frequency of stent failure (3). Despite these advances in technology, very late stent thrombosis and neoatherosclerosis still contribute to late fatal clinical outcomes in some subjects who successfully underwent percutaneous coronary intervention (PCI) with new-generation DES (4). As a plausible explanation, persistence of polymer in the stent platform could continuously evoke chronic inflammation, delay endothelial healing, and accelerate neoatherosclerosis (4, 5). However, if the polymer disappears by gradual biodegradation within a certain period after PCI, additional polymer-related complications may not occur. Therefore, DESs with biodegradable polymers (BP-DES) have been developed and are currently being actively utilized in contemporary clinical practice (6). Despite the theoretical superiority of BP-DES over DP-DES, prior reports, including randomized trials that enabled the use of BP-DES in clinical practice, mostly failed to demonstrate the superiority of BP-DES compared with DP-DES in short-term periods (around 1-year follow-up) with a relatively insufficient number of study participants, considering the low rates of cardiac events after implantation of new-generation DES (7–10). To date, the clinical benefits of BP-DES over DP-DES are controversial (1, 7–11). In this regard, we sought to investigate the long-term clinical impact of BP-DES compared with DP-DES utilizing the large-volume nationwide cohort that covers the entire populations who received first- and new-generation DES implantation for coronary artery disease in Korea (CONNECT DES cohort registry).

Baseline clinical characteristics and medical history of the whole cohort population before and after stabilized IPTW are presented in Table 1. After stabilized IPTW, there was no evidence of inequality in the baseline clinical characteristics and medications between the two groups (all SMD <0.1, Supplementary Figures 1, 2), except for the duration of dual antiplatelet therapy (DAPT) and year of index PCI. The incidence and relative hazards for the clinical outcomes of interest between the two groups after stabilized IPTW are presented in Table 2. At 5 years after index PCI, the incidence of all-cause death was significantly lower in patients treated with BP-DES (11.3 vs. 13.0% in those treated with DP-DES, HR 0.92, 95% CI, 0.88–0.96, p < 0.001; Figure 2A), while showing no statistically significant difference at 2 years after PCI (5.7 vs. 6.0%, respectively, HR 0.95, 95% CI, 0.89–1.01, p = 0.238). Statistical significance of reduced all-cause death was achieved in the BP-DES group at 3 years after index PCI (7.7 vs. 8.4% in DP-DES group, HR 0.93, 95% CI, 0.87–0.99, p = 0.015). Similarly, use of BP-DES was associated with a lower incidence of cardiovascular death (7.4 vs. 9.6% in those treated with DP-DES, HR 0.82, 95% CI, 0.77–0.87, p < 0.001; Figure 2B), and MI (7.4 vs. 8.7% in those treated with DP-DES, HR 0.90, 95% CI, 0.86–0.94, p = 0.006; Figure 2C) at 5 years after index PCI, while showing no statistically significant difference at 2 years after index PCI (4.6 vs. 4.9%, respectively, HR 0.93, 95% CI, 0.85–1.01, p = 0.120 for cardiovascular death; 5.0 vs. 5.1%, respectively, HR 0.98, 95% CI, 0.92–1.05, p = 0.461 for MI). The incidence and relative hazards for the outcomes of interest between the two groups before stabilized IPTW are presented in Supplementary Table 3. In a landmark analysis between 2 and 5 years after index PCI, the use of BP-DES was distinctly associated with reduced occurrence of all-cause death (5.9 vs. 7.4% in patients treated with DP-DES, HR 0.91, 95% CI, 0.86–0.96, p < 0.001; Figure 2D), cardiovascular death (3.1 vs. 4.9%, respectively, HR 0.73, 95% CI, 0.67–0.80, p < 0.001; Figure 2E) and MI (2.7 vs. 3.9%, respectively, HR 0.79, 95% CI, 0.72–0.87, p < 0.001; Figure 2F) (Supplementary Table 4). Multivariable regression analysis also revealed consistent favorable impact of BP-DES compared with DP-DES on all-cause or cardiovascular death at 5-years after PCI (Supplementary Tables 5, 6). A subgroup analysis showed that BP-DES had a consistent beneficial effect on the 5-year incidence of all-cause death (Figure 3), cardiovascular death (Figure 4), or MI (Supplementary Figure 3) across subgroups.

To the best of our knowledge, the results of our analyses were derived from a nationwide cohort with the largest study population ever published. Of note, a major strength of our study was inclusion of all patients treated with thin-strut DESs that are commonly used world-wide in daily clinical practice. Therefore, very-high-risk patients who were usually excluded in prior randomized studies were entirely included in this study. In addition, there was completeness in monitoring for demographic characteristics, comorbidities, medication history, occurrence of clinical events requiring hospitalization, and prescription status of essential cardiac medications such as anti-platelet agents during 5-year follow-up after index PCI. Through analysis of the death certificates provided by the National Statistical Office, we could discriminate cardiovascular death from all-cause death. Therefore, we could investigate the impact of remnant polymer in new-generation DES on cardiovascular death. The principal findings of our study are as follows: (1) In the early term period, the rates for the clinical outcomes of interest were not significantly different between BP- and DP-DES-treated patients; however, (2) the cumulative incidence of clinical events was gradually different between the two groups as time passed after index PCI. Favorable results for the clinical outcomes of interest were distinctly apparent at 5 years after index PCI in BP-DES-treated patients.

It has been suggested that anti-proliferative drugs per se and their polymer carriers in the coronary arterial bed continuously evoke chronic vascular inflammation (4, 5) that leads to delayed vascular healing and incomplete endothelial strut coverage and potentially contributes to clinical presentation of very late stent thrombosis and MI (17). Autopsy studies elucidated a more robust extra-cellular matrix deposition in the re-stenotic lesions of implanted DP-DES compared with bare-metal stents (18), which are considered to play a crucial role in the development of neoatherosclerosis and consequent stent failure (4).

Prior randomized clinical trials and meta-analyses comparing BP- and DP-DES have demonstrated confounding results (6–9, 11). The BIOFLOW V randomized trial showed a significant clinical benefit of biodegradable polymer sirolimus-eluting stents (n = 884 patients) over the durable polymer everolimus-eluting stents (n = 450 patients) by reducing 1-year incidence rate of target-lesion failure (11). However, the Bio-RESORT randomized trial showed no statistically significant difference of primary endpoint at 12 months between biodegradable polymer everolimus-eluting (n = 1,172 patients) and sirolimus-eluting stents (n = 1,173 patients) vs. durable polymer zotarolimus-eluting stents (n = 1,169) (9). A meta-analysis that included 19,886 patients derived from 16 randomized clinical trials found no statistically distinct benefit of BP-DES over DP-DES during a mean follow-up duration of 26 months (6). It might be inappropriate to prove the clinical benefits of BP-DES over DP-DES within 12 months after index PCI. Because DAPT for about 12 months was usually recommended in most studies, the complications or clinical presentations caused by persistent polymer in DP-DES might be masked by DAPT.

Neoatherosclerosis is one of the main mechanisms of very late stent thrombosis after DES implantation (19). Persistence of polymer of DP-DES in the coronary vascular bed may also provoke chronic inflammation characterized by infiltration of inflammatory cells such as macrophage or lymphocytes, playing a central role in the progression of neoatherosclerosis (4). And, the development of neoatherosclerosis is a time-dependent phenomenon; the frequency of neoatherosclerosis increases with the stent age. Our prior intracoronary optical coherence tomography study showed that neoatherosclerosis was observed in 45.7% of patients between 3 and 5 years, 61.5% between 5 and 7 years and 73.9% more than 7 years after DES implantation (20). Therefore, neoatherosclerosis observed at the very late period was significantly associated with very late stent thrombosis. Furthermore, since the polymer degradation are generally known to take about 6–24 months (21, 22), it could be difficult to detect a statistical difference in clinical outcomes according to the polymer properties in early period. These prior findings may provide a plausible explanation for the beneficial effects of BP-DES that appear after a certain period of time post-new generation DES implantation. Indeed, in a prior large nationwide cohort that compared the effect of BP- and DP-DES in real-world practice (total 57,487 patients; BP-DES-treated patients, 10,032 and DP-DES-treated patients, 47,455), there was no significant difference in the occurrence of all-cause mortality or myocardial infarction for 2 years after index PCI (10). Similarly, our analysis also revealed no statistically significant difference in the occurrence of the outcomes of interest at 2 years after new-generation DES implantation; however, significant differences were gradually observed at 3 years post-DES implantation, favoring BP-DES. Based on our prior optical coherence tomography analysis and this newly acquired nationwide claims data analysis, we may postulate that the adverse effects of permanent remnant polymer in the coronary artery start to appear at about 3 years post-PCI and become distinct at about 5 years post-PCI. In addition, the increase in risk of cardiovascular events due to permanent remnant polymer is statistically clear, but the degree is somewhat modest, so it may be necessary to analyze a very large number of patients in order to identify statistically clear differences. The BIOSCIENCE randomized trial reported that 5-year risk of target lesion failure was similar between biodegradable polymer sirolimus-eluting stent (n = 1,063 patients) vs. durable polymer everolimus-eluting stent (n = 1,056 patients) (23). The ISAR-TEST 4 randomized trial recently reported that the 10-year incidence of major adverse cardiac events was not statistically different between biodegradable polymer sirolimus-eluting stents (n = 1,299 patients) vs. durable polymer everolimus-eluting stents (n = 652 patients) (24). These studies did not show a clinical benefit of BP-DES over DP-DES at 5- or 10-year follow-up (23, 24). The main reason for failure to demonstrate the theoretical benefits of BP-DES over DP-DES in these clinical studies might be inappropriately small number of study populations. With the number of all-cause deaths during 5-year follow-up and a total sample size of 127,731, our study had about 82.1% power to detect an HR of 0.80 in the comparison of the BP-DES group with the DP-DES group at a two-tailed alpha level of 5.0%. Further, because in contrast to usual design of randomized controlled trials, we did not exclude the very high-risk patients, the incorporation of very high-risk patients could have attributed to demonstrating a clinical benefit of BP-DES that become distinct over time.

This study has several limitations. First, the findings from this observational study cannot be applied to establish causal relationships and persistent residual confounding factors should be considered in the interpretation of our results, although we tried to minimize the bias through IPTW. Furthermore, as the NHIS database does not contain laboratory parameters or clinical information except for the diagnostic codes, unmeasured variables could have affected the results of our analyses. Second, because this database does not include angiographic or procedural information, including the extent and complexity of coronary artery disease, the impact of high-risk procedural characteristics was not fully considered in the interpretation of our analyses. Furthermore, since the result of electrocardiography was not available in this database, the impact of clinical presentation such as ST-elevation MI or non-ST elevation MI was not considered in statistical analyses. Third, to protect patients' personal information and avoid unnecessary conflict with the device manufacturer, the NHIS database provides information of DESs after sufficient encryption work, which includes DES generation, polymer degradability, and strut thickness. Finally, the temporal difference between the groups still remained even after IPTW-adjustment. Thus, care should be taken in interpretation of our result, although we adjusted the temporal difference in all regression analyses. Therefore, further analyses such as impact of the eluted drug or stent material was not performed, which should be carried out through following research.

In this nationwide cohort of all patients treated with new-generation DES in Korea, BP-DES implantation was associated with a lower risk of all-cause death, cardiovascular death or MI. This difference appeared around 3 years after DES implantation and became more distinct with time.

The data analyzed in this study is subject to the following licenses/restrictions: the datasets generated for the analyses are not publicly available because of strict government restrictions. Requests to access these datasets should be directed to M-KH, mkhong61@yuhs.ac.

The studies involving human participants were reviewed and approved by Severance Hospital Institutional Review Board. The Ethics Committee waived the requirement of written informed consent for participation.

S-JL, D-WC, C-MN, and M-KH contributed to the conception and design and verified the data and conducted all analyses. S-JL and M-KH wrote the study protocol. D-WC and C-MN performed the programming to extract the data from the NHIS database. M-KH had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. YS, S-JH, C-MA, J-SK, B-KK, Y-GK, DC, E-CP, and YJ provided a critical review of manuscript. All authors read and approved on the final publication.

This work was supported by the Cardiovascular Research Center, Seoul, South Korea.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.