In our results, we found that concomitant PTA at the time of PCI was associated with an increased risk of all-cause death, ESRD, amputation and PTA after discharge, whereas it was not associated with increased risks of CV deaths, repeat coronary revascularization, MI, stroke or bleeding events. Our results also showed that experiencing amputation and PTAs after discharge strongly mediated the association between concomitant PTAs and the risk of all-cause death.
Symptomatic LEADs are frequently found in patients with significant CAD (1). The presence of LEAD in patients with acute coronary syndrome indicates widespread atherosclerosis and is often associated with higher CV mortality and morbidity (14, 15). Therefore, revascularization for symptomatic LEADs is often required in patients undergoing PCI, and concomitant PTA is frequently performed at the time of PCI as a default strategy worldwide. Bartus et al. reported a case series of 66 patients undergoing PTA and PCI during a single hospital stay and claimed that concomitant PTA was safe and should be encouraged for patients’ better quality of life (3). A recent study by Koren et al. also reported in a small retrospective cohort that concurrent PTA had a similar 1-year MACE and mortality as staged PTA in patients undergoing PCI, advocating concurrent PTA (16). However, concomitant PTA at the time of PCI still concerns various aspects, including procedure complexity, high contrast media exposure, repeated use of parenteral anticoagulants, multiple puncture sites and higher rates of ISR and acute thrombosis in PTA compared to PCI. PTA is also known to be followed by extensive systemic inflammation, which might promote coronary artery ISR (5, 7). Therefore, the safety and efficacy of concomitant PTAs should be investigated thoroughly. Our results may provide some confidence in the safety of concomitant PTAs regarding CV outcomes and bleeding but may imply potential hazards of concomitant PTAs regarding all-cause deaths and unfavorable renal and limb outcomes, which demands further investigation. Our results are particularly more concerning because the rates of revascularization and amputation have been reported as < 3% per year in patients with stable LEAD in a recent randomized controlled trial (17) and in an observational study (18), which are similar to those in the PCI-only group in our results (Fig. 2).
CV outcomes, including MACE, MI and repeat coronary revascularization, would be driven by coronary ISRs, which are uncommon in the second-generation DES era. In fact, the incidences of MI (concomitant PTA vs. PCI only, 4.7% vs. 5.8%) and repeat coronary revascularization (17.6% vs. 17.1%) during 5 years of observation were quite low in the both groups in our results. The similar risks of bleeding events and stroke between the groups are expected because parenteral anticoagulants would not affect the bleeding risk after discharge, and the same dual-antiplatelet agent therapy regimens were used in the two groups (1, 8). In contrast, the greater exposure to radiocontrast media may have contributed to the higher incidences of ESRD in the PTA + PCI group (19).
Because patients underwent PTA between 2014 and 2015 in the current study, the majority of PTAs were performed using plain balloon angioplasty and BMSs, which exhibit ISR rates as high as 30–50% in a year (20, 21). These high ISR rates after PTA may have resulted in the higher incidences of PTA after discharge and amputation in the PTA + PCI group. The development of CLI has already been recognized as a predictor of death (22). Repeat hospitalization and procedures for limb revascularization and wound care may have to be followed by the risk of procedural complications, malnutrition, infection, bleeding and thrombotic events, thus eventually leading to death. Our results also showed that experiencing PTAs after discharge or amputation significantly mediated the association between concomitant PTAs and a higher risk of all-cause death. Although DES has been preferred to BMS in recent PTAs, the ISR rate of DES was still not remarkably different from that of BMS after PTA for femoropopliteal lesions (23). Therefore, the association between concomitant PTA and unfavorable limb outcomes may not have been different, although the study population included more patients undergoing PTA with DES.
The associations between concomitant PTA and unfavorable limb outcomes were stronger in the subgroups of females, patients with CKD and those with CCIs < 7; this may also imply the role of ISR in worsening the outcomes of patients undergoing concomitant PTA. Patients with CKD typically have more severe LEAD and thus would have experienced amputation or PTA after discharge more frequently. However, it is difficult to apply this reasoning to the results in females and patients with fewer comorbidities. A high ISR rate rather than a heavy LEAD burden may explain the results more easily, given that females could develop ISR more easily because of their smaller vessel diameter (24) and that frequent deaths may have interfered with the observation of amputation/PTA after discharge as a competing event in the patients with a high CCI.
The E-values for the LLCI of the HRs for all-cause deaths and ESRD were only 1.17 and 1.21 in the PSM cohort, which indicates that potential unmeasured confounders including fragility might explain the association of the concomitant PTA with these outcomes. In contrast, the E-values for the LLCI of the HR for amputation and PTA after discharge were substantially high (3.33 and 7.75), and it is unlikely for any potential unmeasured factors to cancel the associations in the PSM cohort. Furthermore, to minimize the concern that patients in the PTA + PCI group might have had more severe LEADs than those in the PCI only group, we excluded patients with any conditions potentially related to ALI, CLI and limb amputation and included patients only with stable LEAD who had not undergone any endovascular procedures for the 3 preceding years. In fact, a previous study showed that in patients with stable LEAD, an ankle-brachial index < 0.5 increased the risk of amputation and limb revascularization, but when age, sex, CLI, CKD and diabetes were adjusted, the HRs of having an ankle-brachial index < 0.5 were only 1.96 and 2.69 for amputation and limb revascularization, respectively, which are smaller than the E-values observed in our results (25).
This study has several limitations. First, since this study is a retrospective observational study, interpretation of the associations shown in the results as causality should be cautious. Although we adjusted confounders using multiple statistical methods, including PSM, and displayed the strength of the associations using the E-values, unmeasured confounders may still significantly impact the associations shown in the results. In particular, the severity of symptoms could not be investigated but may be significantly different between the two groups. Moreover, the E-values for the LLCI of the HR for all-cause deaths and ESRD were small; thus, the associations may be explained by unmeasured confounders such as fragility or atherosclerosis burdens. Second, although this study was conducted using a nationwide insurance claim database, the number of patients included in the PTA + PCI group was only 279, because the enrollment period was only 2 years, only 30% of the eligible patients were randomly selected for the study population because of the KNIHSS policy, and those with ALI or CLI were excluded. Larger-scale investigations with longer enrollment periods are desired to fully reflect contemporary clinical practice. Third, our study was focused on investigating the clinical outcomes related to concomitant PTA in patients undergoing PCI. Therefore, our results did not include any data producing insights into the mechanisms behind the associations of concomitant PTAs with worse renal and limb outcomes and deaths. Although our results showed that the increased risk of all-cause death was significantly mediated by worse limb outcomes, we could only speculate that worse limb outcomes may have resulted from ISR, without any related data analysis results. Finally, we did not investigate the quality of life, which is the most important reason for any patients with stable LEAD to undergo concomitant PTA and may have compensated for the unfavorable clinical outcomes.
In conclusions, concomitant PTA at the time of PCI was not associated with an increased risk of CV events or death, whereas it was associated with increased risks of renal impairment, repeat PTA after discharge and amputation. Experiencing PTAs after discharge and amputation may lead to an increased risk of death in patients undergoing concomitant PTAs. Although no mechanisms underlying these results were revealed in our study, the negative impacts of concomitant PTA on renal and limb outcomes suggest that concomitant PTA should be more cautiously selected as a treatment strategy for stable LEAD in patients undergoing PCI. Further investigations are required, including large-scale randomized controlled trials comparing clinical outcomes, including renal and limb outcomes, between concomitant PTA and delayed PTA in patients undergoing PCI.