Comparison Between Bare-Metal Stents, First-Generation Drug-Eluting Stents, and Bioresorbable Vascular Scaffolds

Percutaneous coronary intervention (PCI) with stent implantation is commonly performed to restore and maintain patency in patients with obstructive coronary artery disease. Drug-eluting stent (DES) devices surpassed bare-metal stent (BMS) options due to their lower rates of target-lesion revascularization, despite similar rates of death and myocardial infarction.^1,2 However, DES implantation is associated with several limitations, such as very late stent thrombosis,^3,4 impaired coronary vasomotion,⁵ stent fracture,⁶ and interference with future PCI or computer tomography imaging. Bioresorbable vascular scaffold (BVS) stents were designed to address the limitations of DES while maintaining similar efficacy. The first commercially available BVS, Absorb (Abbott Cardiovascular), was initially received with enthusiasm, but later clouded by concerns over clinical safety with increased risk of device thrombosis in several randomized trials.^7-9 These results prompted the commercial decision to cease production of Absorb BVS in late 2017.

Many of the adverse events related to BVS have been attributed to the relatively thicker struts (150 μm).¹⁰ Hence, newer-generation BVS devices are generally designed with stronger material, thinner struts, shorter resorption profile, and lower thrombogenicity.^10-13 A drug-free BVS was also proposed in light of higher mortality with paclitaxel-coated stents used in peripheral artery disease.¹¹ The efficacy of these notions can be examined by comparing BVS with BMS devices that have thin struts (60-100 µm), or with first-generation DES devices, which have thicker struts (>100 µm) than current-generation DES options (60-90 µm). To date, data regarding these comparisons are sparse, but could be important to inform future clinical trials, especially in light of the general reservation by professional societies to use BVS stents.^13,14

This study aimed to examine the long-term clinical outcomes after implantation of BVS compared with BMS and first-generation DES in a single-center hospital registry. 

Study population and design. This retrospective study was based on a PCI registry from the cardiology divisions of Grantham Hospital and Queen Mary Hospital. The study group consisted of patients who underwent PCI with successful implantation of at least 1 everolimus-eluting Absorb BVS. The control groups were patients who underwent implantation with BMS and first-generation DES. The study was approved by the institutional review board of the University of Hong Kong/Hospital Authority Hong Kong West (reference number UW20-176). Informed consent was waived given the observational nature of the study; nonetheless, all patient records and information were anonymized and deidentified prior to analysis.

The index date was defined as the date of the PCI. Patients were excluded if they had unsuccessful stent or scaffold implantation and/or incomplete clinical and/or follow-up data. Demographic data, cardiovascular risk factors, medications, and data pertaining to the index PCI procedure were recorded at baseline. 

Definitions and outcome measures. The primary outcome measure was major adverse cardiovascular event (MACE) rate defined as a composite of all-cause mortality, non-fatal myocardial infarction (NFMI), definite or probable stent thrombosis, and ischemia-driven target-vessel revascularization (TVR), at 3 years after index PCI. Myocardial infarction, non-ST segment elevation acute coronary syndrome (NSTE-ACS), and stable coronary artery disease were defined according to the Fourth Universal Definition of Myocardial Infarction.¹⁵Definite or probable stent thrombosis was defined according to the Academic Research Consortium definition.¹⁶Hypertension was defined as resting systolic or diastolic blood pressure ≥140/90 mm Hg on 2 occasions or prescription of antihypertensive drugs. Diabetes mellitus was defined as a serum fasting glucose ≥7.0 mmol/L or prescription of antidiabetic medication. Smoking status was defined as any history of habitual smoking. Chronic kidney disease was defined as estimated glomerular filtration rate <30 mL/min/1.73 m². All data were retrieved from the electronic medical records from the territory-wide computer network of all public hospitals in Hong Kong. All outcome events were adjudicated in chart review with a standardized methodology by a physician reviewer blinded to index procedural details. Data analysis was performed by a second researcher. 

Statistical analysis. Based on prior large randomized trials comparing BMS and DES,^1,2 we estimated that MACE occurs in 20% of the patients. Power calculation showed that the study population size had a power of 0.84 to detect a 33% difference in odds of MACE. Continuous variables are expressed as mean ± standard deviation or median with interquartile range (IQR) as appropriate, and discrete variables as frequencies with percentages. Comparison of the baseline clinical characteristics was performed using Student’s t-test or one-way analysis of variance. Kaplan-Meier survival analyses with the log-rank test were carried out and Cox proportional hazards regression model was used to calculate hazard ratios (HRs) for the incidence of MACE adjusting for confounders. Multivariable logistic regression was performed to evaluate the relationship between choice of stent/scaffold (BVS, DES, and BVS) and MACE, after adjustment for statistically significant variables in univariate analysis. 

Exploratory analysis. Since stent thrombosis was a key safety concern for BVS,^17,18 we compared the composite endpoint of definite or probable stent thrombosis across choice of stent/scaffold after confounder adjustment with the same logistic regression model. Data management and statistical analyses were performed in Stata software (StataCorp/MP, version 16). A two-tailed P-value of <.05 was considered statistically significant.

A total of 1396 patients were analyzed, including 622 patients who received BMS implantation and 604 patients who received DES implantation between 2001 and 2005, and 170 patients who received BVS implantation between 2014 and 2017. In the entire study cohort, 1048 patients (75.1%) were men and the mean age was 63.7 ± 10.8 years. Patients who received BVS were younger, as well as less likely to have diabetes, hypertension, prior myocardial infarct or PCI, and acute coronary syndrome. Table 1 shows the baseline characteristics of the study population and Table 2 shows the procedural characteristics of index PCI.

Clinical outcomes at 3 years after index PCI are shown in Table 3. MACE developed in 125 patients (20.1%) in the BMS group, 81 patients (13.4%) in the DES group, and 25 patients (14.7%) in the BVS group. In crude analysis, the rates of MACE and death were significantly different across groups, while the rates of NFMI, TVR, and stent thrombosis were similar. The Kaplan-Meier estimates of MACE and individual components of MACE are shown in Figures 1-5.

Multivariable regression analysis was performed to assess the association between different stents/scaffolds and MACE. The results are shown in Table 4. In adjusted analysis, the DES group had a lower risks of MACE (adjusted odds ratio [OR], 0.58; 95% confidence interval [CI], 0.41-0.81; P<.001) compared with the BMS group. In contrast, the BVS group had similar MACE (adjusted OR, 0.91; 95% CI, 0.55-1.52; P=.72) compared with the BMS group. When compared directly with the DES group, the BVS group had similar MACE (adjusted OR, 1.45; 95% CI, 0.83-2.53; P=.19), death (adjusted OR, 1.52; 95% CI, 0.59-3.91; P=.38), NFMI (adjusted OR, 1.80; 95% CI, 0.70-4.62; P=.22), and TVR (adjusted OR, 1.33; 95% CI, 0.66-2.68; P=.42).

In the exploratory analysis, BVS was not associated with increased risk of definite or probable stent thrombosis compared with BMS (adjusted OR, 1.17; 95% CI, 0.22-6.04; P=.85) or DES (adjusted OR, 1.11; 95% CI, 0.22-5.51; P=.90).

In this retrospective study of 1396 adult patients who underwent coronary stent/scaffold implantation, first-generation DES but not BVS was associated with better outcomes at 3 years compared with BMS. Nonetheless, in direct comparison with DES, BVS was associated with similar long-term clinical outcomes. 

First-generation BVS (Absorb) consists of synthetic biodegradable polymers that dissolve approximately 3-4 years after implantation,¹⁹ with postulated advantages of restoration of vasomotor function,²⁰ less interference with future coronary computed tomography,²¹ and the appeal of no residual foreign material. Disappointingly, BVS was found to have inferior outcomes compared with second-generation DES in several randomized trials and meta-analyses, with increased risk of MACE driven by more scaffold thrombosis and target-vessel myocardial infarct.^7-9,22,23 Along with the commercial discontinuation of Absorb, 2 important questions remain unanswered: Why is BVS worse than second-generation DES, and how much worse in comparison with other approved coronary stents?

There are 2 postulated factors to explain the difference in stent performance between BVS, DES, and BMS — the thickness of the stent struts, and the presence of the antiproliferative drug coating. Many researchers attribute the worse performance of BVS to the relatively thicker struts (150 µm), a design out of necessity to preserve acceptable radial strength.¹⁰ The significance of strut thickness is illustrated in the technological advancement of DES options, where thin strut (80-100 µm) and ultra-thin strut (60-80 µm) second-generation DESs were found to have better outcomes than thick strut (>100 µm) first-generation DES options.^24,25 From our data, clinical outcomes were similar when comparing BVS and first-generation DES with comparable strut thickness. Newer-generation BVS stents are expected to feature thinner struts by employing stronger resorbable material,^10-13 and we consider it plausible that BVS may perform better when strut thickness can be reduced. There have already been promising results in early clinical outcomes of newer-generation BVS stents.^26,27  

The antiproliferative drug coating on coronary stents/scaffolds can also intervene with endothelialization, resulting in persistent exposure of foreign body, inflammation, and late stent thrombosis.²⁸ This may be a less significant issue in the context of thinner strut, second-generation DES options.^29,30 In our study, BVS was associated with outcomes similar to BMS (thin struts, <100 µm), with no signals toward worse outcomes. This supported the idea that strut thickness, rather than antiproliferative drug, should be the target for improvements in BVS design. 

One important ethical and practical aspect for further BVS research is to define the extent to which BVS has been inferior to second-generation DES. Our data showed that the performance of BVS was similar to BMS and first-generation DES with regard to outcomes at 3 years and the rate of stent thrombosis. This should reassure researchers and clinicians who are skeptical about the safety of BVS and its future use in clinical trials, since BMS and first-generation DES are still considered acceptable options for routine clinical use. Concerns with regard to whether it is ethical to randomize patients to newer-generation BVS stents could be eased.

Strengths and limitations. This study had several strengths. First, published data comparing BVS with BMS or first-generation DES are sparse. The current study identified a lower reference bound for the performance of BVS, in a wave of inferior performance compared with second-generation DES. Second, no patients were lost to follow-up due to the comprehensive documentation from the territory-wide electronic medical record system in Hong Kong. Third, our cohort represented a real-world population comprising a wide range of clinical presentations and co-morbidities. 

This study had several limitations. First, the observational nature of the study conferred risks of unmeasured confounding and bias, but the findings were extensively adjusted for many clinical and procedural variables. Second, the PCIs were performed in a different time period for the BVS group compared with the BMS or DES groups, and standards of PCI and medical therapy may have evolved over time. Certain implantation techniques, such as optimal predilation, vessel sizing, postdilation,^31-33 and use of intravascular imaging,^34,35 have been shown to improve outcomes for BVS and therefore may be more widely used in the BVS group. Nonetheless, this reflects the efficacy of the stents as used in real-world practice. 

Future directions. Ongoing research is underway to enhance the design and implantation techniques for the newer-generation BVS.^36,37 The limitations of BVS stents appear to be related to its design rather than the concept of resorption. With advancements in material science, BVS stents could be made of stronger materials that preserve radial strength even with thinner struts, shorter resorption profile, and better biocompatibility. Future clinical trials to examine the efficacy of newer-generation BVS options are much anticipated. 

In patients with BVS implantation, the risk of composite outcome of MACE (death, NFMI, TVR) at 3 years was not significantly different compared with patients with BMS or first-generation DES implantation.