Safety and Efficacy of Magmaris Bioresorbable Scaffolds Versus Contemporary Drug-Eluting Stents Including Biomatrix, Ultimaster, Orsiro, and Xience in Patients With Coronary Artery Disease: A Comprehensive Systematic Review and Extended Meta-Analysis

Suneel Arwani¹, Muhammad Ali Niaz², Tooba Idrees³, Saifullah Khan⁴, Bushra Ghaffar⁵, Joshua Odidison², Muhammad Ahmed Ashfaque², Saifullah Syed², Hira Riaz⁶

¹University of Mississippi Medical Center, Jackson, Mississippi, United States

²Royal College of Surgeons in Ireland, Dublin, Ireland

³University College of Medicine and Dentistry (UCMD), Lahore, Pakistan

⁴MedStar Washington Hospital Center, Georgetown University School of Medicine, Washington, DC, United States

⁵Quaid-e-Azam Medical College, Bahawalpur, Pakistan

⁶Kabul Medical University, Kabul, Afghanistan

* Corresponding author: Hira Riaz. Kabul Medical University, Kabul, Afghanistan.

Received: 10 June 2026; Accepted: 18 June 2026; Published: 24 June 2026

Background

Coronary artery disease (CAD) is still one of the major causes of morbidity and death globally; therefore, advancements in percutaneous coronary intervention (PCI) methods and device technology are constantly needed. By lowering restenosis and the need for repeated revascularization, drug-eluting stents (DES) have greatly enhanced clinical outcomes. However, long-term issues with permanent metallic implants-including late stent thrombosis, decreased vasomotion, and chronic inflammation—have influenced the development of next-generation stents, such as bioresorbable scaffolds (BRS) and ultrathin-strut drug-eluting stents [1,2].

The Magmaris bioresorbable magnesium scaffold has become a novel alternative among these developments, providing the mechanical support of conventional DES while removing the presence of foreign bodies over an extended period of time. Its positive safety and efficacy profile has been shown in recent single-arm studies with encouraging outcomes in vulnerable groups such as diabetic patients and those with acute coronary syndrome (ACS). [3-5]. The effectiveness of Magmaris in controlled and real-world clinical settings has been further supported by the BIOSOLVE-IV registry which demonstrated minimal rates of scaffold or stent thrombosis and target lesion failure [6]. At the same time, the Orsiro ultrathin-strut (≈60 μm struts) sirolimus-eluting stent and the Ultimaster thin-strut (≈80 μm struts) sirolimus-eluting stent—each have become more well-known because of their better durability, delivery, and healing profiles [7-9]. Furthermore, reduced restenosis, lower target lesion revascularization, and fewer adverse cardiac events (ACEs) are among the best long-term results related to these stents, as shown by clinical studies and registry data [10-14]. Ultrathin-strut DES has proven to be more effective than regular stents in head-to-head comparisons in randomized settings, especially when it comes to enhancing vascular repair and minimizing late lumen loss [1,9]. Despite these promising results, data from randomized controlled trials remain limited for more recent technologies like Magmaris. Single-arm observational studies continue to be a crucial source of empirical knowledge, providing important information about how well these devices function in various clinical and anatomical contexts [3,5]. In order to evaluate the clinical results of the Magmaris bioresorbable magnesium scaffold in relation to Biomatrix, Ultimaster, and Orsiro sirolimus-eluting stents in patients with de novo coronary artery disease, this systematic review and meta-analysis attempts to compile the most recent data. This review will offer a thorough assessment of these developing percutaneous coronary intervention techniques—with an emphasis on target lesion failure (TLF), all-cause mortality, cardiac death, myocardial infarction (MI), target-vessel myocardial infarction (TV-MI), ischaemia-driven target-lesion revascularisation (ID-TLR), and definite/probable stent/scaffold thrombosis—in order to support clinical decision-making and direct future comparative studies. However, the currently available evidence is largely limited to short- and mid-term follow-up, which remains a critical gap given that the theoretical advantages of bioresorbable scaffolds are expected to emerge over longer time horizons.

Methods

We conducted a systematic review and single-arm meta-analysis specified in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA). Our review was registered at the Open Science Framework before it was undertaken (OSF ID: 10.17605/OSF.IO/PE6SU) (Figure S50).

Search Strategy and Data Sources

A comprehensive literature search was conducted using several electronic databases from inception until March 1, 2026, including PubMed, Scopus, and ScienceDirect. Two authors independently performed the search to ensure accuracy and minimize bias. The search strategy included combinations of keywords and MeSH terms such as “drug-eluting stent,” “biodegradable scaffold,” “clinical outcomes,” “stent thrombosis,” and “myocardial infarction.” Boolean operators (AND, OR) were applied to maximize sensitivity. No restrictions were imposed on language or publication date, ensuring a comprehensive and unbiased retrieval of relevant studies. The full search strategy is detailed in Table 5.

Table 5: Search strateg.

Study Selection Criteria

Eligibility was defined using the PICOS scheme. Trials were required to enroll adults (≥18 years) with angiographically confirmed de novo coronary lesions receiving PCI due to either acute coronary syndromes (STEMI, NSTEMI, unstable angina) or stable coronary artery disease. Eligible devices included: (1) Magmaris bioresorbable magnesium scaffold (sirolimus-eluting), (2) Orsiro ultrathin-strut metallic sirolimus-eluting stent, (3) Ultimaster thin-strut metallic sirolimus-eluting stent, and (4) Biomatrix stent. At least one adjudicated clinical endpoint had to be reported at ≥6 months (preferably ≥12 months): target-lesion failure (TLF), all-cause mortality, cardiac death, myocardial infarction (MI), target-vessel MI, ischaemia-driven target-lesion revascularisation (ID-TLR), or definite/probable stent/scaffold thrombosis (ARC-2 definitions). Prospective single-arm observational studies, registries, cohorts, and single arms of randomised controlled trials were eligible. Exclusions included case reports, case series (<10 patients), retrospective chart reviews lacking prospective outcome collection, reviews, editorials, animal studies, and studies where device-specific data could not be disaggregated.

Screening and Data Extraction

All studies retrieved from the literature search were imported to EndNote X9 (Clarivate Analytics) for duplicate removal and screening. Title, abstract, and full-text screening were completed separately by two reviewers; disagreement was settled by a third senior reviewer (Cohen’s κ = 0.82). Study features, patient demographics, procedure parameters, and endpoint data were documented at 6-, 12-, 24-, and 36-month time-points using a piloted, standardised Excel form. Data were extracted in duplicate; inconsistencies were reconciled by consensus. When only graphical data were provided, PlotDigitizer (v2.6.8) was used. Study authors were contacted twice for missing or unclear data; non-responders within four weeks were excluded from pooled analyses for the relevant outcome.

Risk of Bias Assessment

Risk of bias for included studies was assessed using validated tools according to study design. Cohort studies were evaluated using the Newcastle–Ottawa Scale (NOS), which examines three domains: selection of study groups (S1–S4), comparability (C1–C2), and outcome assessment (O1–O3) (7). A maximum score of nine stars indicates the highest methodological quality. Randomized controlled trials (RCTs) were assessed using the Cochrane Risk of Bias 2 (RoB 2) tool, which evaluates bias across five domains: randomization process, deviations from intended interventions, missing outcome data, outcome measurement, and selective reporting. Each domain was rated as “low risk,” “some concerns,” or “high risk,” and an overall risk of bias judgment was assigned accordingly.

Data Synthesis and Statistical Analysis

We calculated event rates with exact 95% confidence intervals (Clopper–Pearson) for each device and each outcome. The DerSimonian–Laird random-effects model was used to make pooled estimates that accounted for expected clinical and methodological differences. Heterogeneity was quantified using I² (values >50% indicating substantial heterogeneity) and prediction intervals. Predefined subgroup analyses examined clinical presentation (stable CAD versus acute coronary syndrome), diabetes status, follow-up duration (6–12 months versus >12 months), and geographical region (Western versus Asian cohorts). Sensitivity analyses omitted studies with a high risk of bias, unpublished reports, or those involving fewer than 100 patients; influence analysis utilized leave-one-out cross-validation. Funnel plots and Egger’s test were used to check for publication bias when there were at least 10 studies for an outcome. All statistical analyses were performed using R (v4.3.2).

Outcomes

Clinical endpoints were adjudicated at ≥6 months, consistent with ARC-2 definitions.

Primary Efficacy Endpoint

Target-Lesion Failure (TLF): This composite endpoint is defined as the combination of: (1) Cardiac death—unexplained death occurring within a time window of the index procedure; (2) Target-Vessel Myocardial Infarction (TV-MI); and (3) Ischaemia-Driven Target-Lesion Revascularisation (ID-TLR).

Secondary Safety and Efficacy Endpoints

Stent/Scaffold Thrombosis (ST): Defined according to ARC-2 criteria, encompassing definite (angiographic or pathological confirmation) and probable (unexplained acute ischemic event or sudden death within the target vessel territory) events.

Myocardial Infarction (MI): Defined according to the latest universal definition, specifically including Target-Vessel MI (TV-MI), which is an MI caused by an acute ischemic event in the territory supplied by the target vessel, excluding events clearly caused by procedural complications.

Ischaemia-Driven Target-Lesion Revascularisation (ID-TLR): Any repeat revascularization (PCI or bypass surgery) performed on the target lesion due to clinical signs or symptoms of ischaemia and confirmed by angiographic evidence of restenosis.

 All-Cause Mortality: Death from any cause occurring during the specified follow-up period.

Cardiac Death: Death due to an immediate cardiac cause (e.g., MI, heart failure, lethal arrhythmia, or procedural complications) or any unexplained death.

Results

Study Selection

A total of 22,022 records were identified through database searching, with no additional records obtained from registers. After removal of 4,200 duplicate records and 322 records for other reasons, 17,500 records remained for title and abstract screening. Of these, 16,200 records were excluded based on predefined eligibility criteria.

A total of 1,300 reports were sought for full-text retrieval, of which 100 could not be retrieved. Consequently, 1,200 full-text articles were assessed for eligibility. Among these, 1,164 reports were excluded due to wrong study design (n = 520), irrelevant outcomes (n = 310), and insufficient data (n = 334). Ultimately, 36 studies met the inclusion criteria and were included in the final systematic review (Figure 1).

Figure 1: Prisma flow diagram.

Study Characteristics

The included studies evaluated four distinct stent platforms: Biomatrix, Magmaris, Orsiro, and Ultimaster.

A total of 36 studies were included, comprising a mix of randomized controlled trials (RCTs) and observational cohort studies. The included studies evaluated a range of coronary devices, predominantly focusing on magnesium-based bioresorbable scaffolds (Magmaris), biodegradable polymer drug-eluting stents (e.g., Orsiro, Ultimaster, BioMatrix), and comparative platforms including durable polymer stents and dual-therapy stents.

Sample sizes varied substantially, ranging from 30 to 3,832 participants, with several large multicenter RCTs enrolling over 2,000 patients. The studies were conducted across diverse geographic regions, including Europe, Asia, the Middle East, Africa, and multinational cohorts, reflecting a broad representation of real-world and trial populations.

Patient populations included both acute coronary syndrome (ACS) and stable coronary artery disease (CAD), with many studies enrolling mixed cohorts. ACS presentations encompassed ST-segment elevation myocardial infarction (STEMI), non-ST elevation myocardial infarction (NSTEMI), and unstable angina, while stable CAD populations included chronic coronary syndrome and silent ischemia. Follow-up durations ranged from short-term (30 days) to long-term (up to 5 years), although most studies reported outcomes at 1–2 years. Dual antiplatelet therapy (DAPT) duration varied across studies, typically ranging from 6 to 12 months, with some studies exploring shorter or extended regimens depending on clinical context and device type. Intravascular imaging was variably utilized across studies. Optical coherence tomography (OCT) and intravascular ultrasound (IVUS) were commonly employed in scaffold-based studies, particularly those involving Magmaris, whereas several RCTs relied primarily on conventional coronary angiography, with selective use of imaging modalities in a subset of patients. Overall, the included studies demonstrate considerable heterogeneity in study design, patient populations, device platforms, and adjunctive imaging and pharmacotherapy strategies. Study characteristics is given in table 1, patient characteristics is given in table 2.

Table 1: Study characteristics table.

Table 2: Patient characteristics table.

Clinical Outcomes of Biomatrix Drug-Eluting Stents

A total of 5 studies (N = 5,281 patients) were analyzed. The pooled incidence of stent thrombosis was 1% (95% CI: 0.00–0.02; I² = 85.4%; Figure S1). Cardiac death occurred in 2% of patients (95% CI: 0.01–0.04; I² = 93.9%; Figure S2), while target lesion revascularization was observed in 3% (95% CI: 0.00–0.07; I² = 98.2%; Figure S5). Target lesion failure (TLF) was reported at 7% (95% CI: 0.03–0.11; I² = 97.9%; Figure S3). Target vessel myocardial infarction was reported at 3% (95% CI: 0.01–0.04; I² = 96.4%; Figure S4).

Clinical Outcomes of Magmaris Bioresorbable Scaffolds

The analysis of Magmaris scaffolds included up to 14 studies (N = 5,199 patients). The pooled estimate for all-cause mortality was 1% (95% CI: 0.01–0.01; I² = 0.0%; Figure S7). The incidence of stent thrombosis was low at 1% (95% CI: 0.00–0.01; I² = 0.0%; Figure S11). Myocardial infarction occurred in 2% of patients (95% CI: 0.01–0.02; I² = 17.9%; Figure S9). Target lesion failure (TLF) was reported at 5% (95% CI: 0.04–0.06; I² = 66.7%; Figure S12). Target lesion revascularization (TLR) was reported at 4% (95% CI: 0.02–0.05; I² = 87.8%; Figure S18). Target vessel myocardial infarction was reported at 1% (95% CI: 0.01–0.02; I² = 0.0%; Figure S21).

Clinical Outcomes of Orsiro Drug-Eluting Stents

Outcomes for the Orsiro stent were derived from up to 13 studies (N = 5,129 patients). The pooled rate of cardiac death was 2% (95% CI: 0.01–0.04; I² = 73.3%; Figure S22). All-cause mortality was 4% (95% CI: 0.03–0.05; I² = 75.5%; Figure S23). The incidence of myocardial infarction was 1% (95% CI: 0.01–0.02; I² = 51.5%; Figure S25), and target vessel MI was 2% (95% CI: 0.01–0.03; I² = 88.1%; Figure S33). Stent thrombosis remained low at 0% (95% CI: 0.00–0.00; I² = 20.8%; Figure S27). Target lesion revascularization was observed in 2% of patients (95% CI: 0.01–0.04; I² = 83.2%; Figure S31). Target lesion failure was reported at 4% (95% CI: 0.03–0.06; I² = 83.9%; Figure S29).

Clinical Outcomes of Ultimaster Drug-Eluting Stents

The Ultimaster stent was evaluated in 10 studies comprising 10,382 patients. The pooled rate of cardiac death was 2% (95% CI: 0.01–0.03; I² = 88.0%; Figure S36), while all-cause mortality was 4% (95% CI: 0.02–0.06; I² = 95.1%; Figure S40). The rates for myocardial infarction and stent thrombosis were 2% (95% CI: 0.01–0.04; I² = 91.3%; Figure S45) and 1% (95% CI: 0.00–0.01; I² = 88.4%; Figure S49), respectively.

Clinical Outcomes of Xience Stents

For target lesion revascularization (TLR), pooled analysis of 10 studies (n = 13,642) demonstrated an event rate of 4% (95% CI: 2–6%), with substantial heterogeneity (I² = 97.4%, p < 0.0001). Individual study estimates ranged from 0% (Valgamigli 2021) to 9% (Kerkmeijer 2020).

For target lesion failure (TLF), pooled analysis of 10 studies (n = 13,780) yielded a pooled event rate of 9% (95% CI: 7–12%), again with high heterogeneity (I² = 96.3%, p < 0.0001). Event rates across individual studies ranged from 5% (Valgamigli 2021) to 16% (Kandzari 2015).

For myocardial infarction (MI), pooled analysis of 8 studies (n = 11,672) showed a pooled event rate of 4% (95% CI: 3–5%), with high heterogeneity (I² = 93.5%, p < 0.0001). Individual study event rates ranged from 1% (Valgamigli 2021) to 14% (Kandzari 2015).

Subgroup Analysis of Follow-up Duration

Subgroup analyses were conducted to assess the impact of follow-up duration on clinical outcomes.

Magmaris Scaffolds: Significant temporal differences were observed. TLF rates increased significantly from 2% (95% CI: 0.01–0.04) at 1-year follow-up to 6% (95% CI: 0.06–0.07) at 2-year follow-up (P < 0.0001; Figure S15). Similarly, TLR rates rose from 3% at 1 year to 6% at 2 years (P = 0.0010; Figure S19), indicating a time-dependent accumulation of adverse events.

Orsiro Stents: Analyses comparing follow-up durations (≤18 months vs. >18 months) showed stable long-term performance. No statistically significant differences were found between subgroups for stent thrombosis (P = 0.22), all-cause mortality (P = 0.69), myocardial infarction (P = 0.53), TLF (P = 0.63), TLR (P = 0.38), or target vessel MI (P = 0.79).

Ultimaster Stents: The subgroup analysis comparing follow-up durations (≤1 year vs. 5 years) indicated that follow-up time did not significantly influence cardiac death (P = 0.14), myocardial infarction (P = 0.26), or stent thrombosis (P = 0.26). However, a statistically significant difference was observed for all-cause mortality (P = 0.0007), with higher rates in the 5-year subgroup (8%; 95% CI: 0.06–0.10) compared to the ≤1-year subgroup (3%; 95% CI: 0.02–0.05). These findings suggest a potential late hazard period associated with scaffold resorption, during which the risk of adverse events may increase. The results are summarized in Table 3, and the subgroup analysis results are presented in Table 4.

Table 3: Summary of pooled clinical outcomes and heterogeneity statistics for studies evaluating Biomatrix, Magmaris, Orsiro, and Ultimaster drug-eluting stents.

Note: Data presents pooled estimates [95% Confidence Interval] derived from random-effects meta-analysis models. N represents the total number of patients, and I^2 indicates the percentage of heterogeneity across studies. CAD, coronary artery disease; PCI, percutaneous coronary intervention; MI, myocardial infarction; TLF, target lesion failure; TLR, target lesion revascularization; TV-MI, target vessel myocardial infarction; CI, confidence interval; I², I-squared statistic for heterogeneity; k, number of included studies; N, total number of patients.

Table 4: Subgroup analysis of clinical outcomes stratified by follow-up duration for Magmaris, Orsiro, and Ultimaster drug-eluting stents.

Note: Studies were stratified by follow-up time (e.g., < 1 year vs. > 1 year or < 18 months vs. > 18 months). The P-value for interaction indicates the statistical significance of the difference between the subgroups.TLF, target lesion failure; TLR, target lesion revascularization; MI, myocardial infarction; TV-MI, target vessel myocardial infarction; CI, confidence interval; I², I-squared statistic for heterogeneity; k, number of included studies; N, total number of patients. Cardiac death refers to death resulting from cardiovascular causes.

Heterogeneity and Publication Bias

Heterogeneity varied substantially by stent type and outcome. Magmaris and Orsiro studies generally demonstrated lower heterogeneity for mortality and thrombosis outcomes (I² < 20%), whereas Biomatrix and Ultimaster studies showed considerable heterogeneity (I² > 85%) across most endpoints. Visual inspection of funnel plots for key outcomes, including all-cause mortality and stent thrombosis, suggested generally symmetrical distributions, though some asymmetry was noted in the analysis of TLF for Magmaris stents (Figure S13), necessitating cautious interpretation of small-study effects. For the Magmaris analyses, funnel plots for all-cause mortality (Figure S6), myocardial infarction (Figure S8), and stent thrombosis (Figure S10) were broadly symmetrical, with only minor scatter among less precise studies. Funnel plots for target lesion revascularization (Figure S16) and target vessel myocardial infarction (Figure S20) similarly showed no substantial asymmetry. In contrast, within the Ultimaster dataset, the funnel plot for cardiac death (Figure S35) and the funnel plot for all-cause mortality (Figure S39) showed notable asymmetry, with smaller studies clustering on the right side of the pooled estimate, raising suspicion of publication bias or heterogeneity. The myocardial infarction plot (Figure S43) similarly displayed scattered and uneven distribution, including a large, low-precision outlier. The funnel plot for stent thrombosis (Figure S48) was also reviewed. While most outcomes showed no substantial evidence of publication bias, the asymmetry observed in Ultimaster analyses for all-cause mortality and myocardial infarction suggests potential small-study effects or selective reporting. These findings warrant cautious interpretation, particularly for efficacy comparisons involving Ultimaster.

Meta-Regression Analysis

A univariable meta-regression was conducted to examine whether follow-up duration influenced the effect sizes of key clinical outcomes for Magmaris and Ultimaster stents. For Magmaris studies, where follow-up ranged from 1.0 to 2.0 years, a positive association was observed between follow-up length and event rates; both TLR (Figure S17) and TLF (Figure S14) showed moderately positive regression slopes, indicating that longer follow-up was associated with slightly higher observed rates. For the Ultimaster stent, with follow-up durations spanning approximately 10 to 60 months, similar positive relationships were seen for all-cause mortality (Figure S41) and cardiac death (Figure S37). In contrast, the meta-regression for myocardial infarction (Figure S44) showed only a minimal, nearly flat slope, suggesting that follow-up duration did not substantially influence MI rates. Table 6 shows the results of the GRADE assessment.

Table 6: GRADE Assessment of the reported outcomes.

Risk of bias assessment

Overall, the methodological quality of included cohort studies was high. Most studies achieved the maximum NOS score of nine stars, indicating low risk of bias across selection, comparability, and outcome domains. A subset of studies scored eight stars, primarily due to minor limitations in selection procedures or incomplete control for confounding factors; however, none were deemed to have critically high risk of bias.

For randomized controlled trials, all included studies were judged to have an overall low risk of bias according to the Cochrane RoB 2 tool. Most domains, including randomization, deviations from intended interventions, and outcome measurement, were consistently rated as low risk. Some studies showed concerns related to missing outcome data or outcome reporting; however, these did not substantially affect the overall risk assessment. Collectively, the included RCTs demonstrated robust methodological quality with minimal risk of systematic bias. The results are summarised in Table 7 and 8.

Table 7: Risk of bias assessment of included cohort studies using the Newcastle–Ottawa Scale. Selection domain includes S1 (representativeness of the exposed cohort), S2 (selection of the non-exposed cohort), S3 (ascertainment of exposure), and S4 (absence of outcome at baseline). Comparability includes C1–C2 (control for confounders), and outcome includes O1 (assessment of outcome), O2 (adequacy of follow-up duration), and O3 (completeness of follow-up).

Table 8: Risk of bias assessment of randomized controlled trials using the Cochrane Risk of Bias 2 across all methodological domains.

s

Meta-Regression Analysis

To explore whether stent type was a significant explanatory variable for between-study heterogeneity across clinical outcomes, we performed inverse-variance weighted meta-regression analyses using weighted least squares (WLS) on study-level log-transformed event rates for five pre-specified endpoints: cardiac death, all-cause mortality, target lesion failure (TLF), target lesion revascularization (TLR), and target vessel myocardial infarction (TV-MI). Magmaris was used as the primary reference comparator for all outcomes except cardiac death, for which Orsiro served as the reference due to data availability. Regression coefficients, rate ratios, and model fit statistics for each outcome are presented in Tables S1A–S5B; a consolidated summary of all regression results is presented in Table S6, and between-group heterogeneity statistics are presented in Table S7. Forest plots for individual stent groups are presented in Figures S51–S55.

 Cardiac Death

The cardiac death analysis included 19 studies (N = 17,179) across three stent types — Biomatrix (4 studies), Orsiro (10 studies), and Ultimaster (5 studies) — with Orsiro serving as the reference comparator (Tables S1A–S1B; Figure S51). The pooled cardiac death rates were 1.99% (95% CI 0.94–4.17%) for Biomatrix, 2.41% (1.68–3.46%) for Orsiro, and 1.60% (0.44–5.88%) for Ultimaster. In the meta-regression model, neither Biomatrix (RR 0.802, 95% CI 0.563–1.142; p = 0.222) nor Ultimaster (RR 1.227, 95% CI 0.998–1.509; p = 0.052) differed significantly from Orsiro, although the Ultimaster comparison was borderline. The omnibus test confirmed a statistically significant overall stent-type effect (χ²(2) = 6.23, p = 0.044); however, stent type explained only 3% of between-study variance (weighted R² = 3.0%), indicating that the preponderance of heterogeneity — which was substantial across all three groups (I² = 74–97%) — was attributable to factors other than stent type.

All-Cause Mortality

The all-cause mortality analysis included 25 studies (N = 18,663) across Magmaris (10 studies), Orsiro (10 studies), and Ultimaster (5 studies), with Magmaris as the reference (Tables S2A–S2B; Figure S52). The Magmaris cohort demonstrated a pooled all-cause mortality rate of 1.25% (95% CI 0.96–1.62%) with remarkably low heterogeneity (I² = 0.0%, tau² = 0), indicating highly consistent outcomes across its study programme. By contrast, both comparator stents showed substantially higher mortality rates in the regression model: Orsiro was associated with a 4.6-fold higher all-cause mortality rate relative to Magmaris (RR 4.581, 95% CI 3.476–6.036; p < 0.001), and Ultimaster with a 3.5-fold higher rate (RR 3.458, 95% CI 2.545–4.697; p < 0.001). The omnibus test was highly significant (χ²(2) = 119.15, p < 0.0001), and stent type explained 45.0% of between-study variance — the highest model fit across all five outcomes. These findings should nonetheless be interpreted cautiously, given that the Magmaris studies predominantly comprised small single-centre trials enrolling selected patient cohorts (N = 193 per study in most cases), which may not reflect the broader, more heterogeneous populations enrolled in the multicentre Orsiro and Ultimaster programmes (I² = 93.0% and 65.9%, respectively).

Target Lesion Failure

The TLF analysis included 25 studies (N = 14,320) across Magmaris (14 studies), Biomatrix (6 studies), and Ultimaster (5 studies), with Magmaris as the reference (Tables S3A–S3B; Figure S53). Pooled TLF rates were 4.69% (3.58–6.14%) for Magmaris, 4.33% (2.89–6.48%) for Biomatrix, and 4.67% (1.91–11.44%) for Ultimaster, indicating broadly similar absolute rates. However, the regression model revealed significant between-stent differences in the weighted log-rate. Biomatrix demonstrated a significantly lower TLF rate than Magmaris (RR 0.744, 95% CI 0.612–0.905; p = 0.003), representing a relative reduction of approximately 26%. Conversely, Ultimaster was associated with a significantly higher TLF rate (RR 1.876, 95% CI 1.618–2.175; p < 0.001), though this estimate is driven predominantly by the Lee 2019 study, which contributed an outlying TLF rate of 21.7% and was the primary source of extreme heterogeneity within the Ultimaster group (I² = 98.5%). The omnibus test was highly significant (χ²(2) = 129.06, p < 0.0001), with stent type explaining 27.9% of between-study variance.

Target Lesion Revascularization

The TLR analysis included 24 studies (N = 12,782) across Magmaris (14 studies), Biomatrix (6 studies), and Ultimaster (4 studies), with Magmaris as the reference (Tables S4A–S4B; Figure S54). Pooled TLR rates were 4.18% (2.93–5.97%) for Magmaris, 2.16% (1.10–4.24%) for Biomatrix, and 3.62% (1.43–9.15%) for Ultimaster. Biomatrix demonstrated the strongest between-stent difference across all endpoints, with a significantly lower TLR rate compared to Magmaris (RR 0.431, 95% CI 0.335–0.554; p < 0.001), representing a relative reduction of approximately 57%. Ultimaster did not differ significantly from Magmaris in TLR (RR 1.092, 95% CI 0.909–1.311; p = 0.347). The omnibus test was highly significant (χ²(2) = 52.97, p < 0.0001), with stent type explaining 24.1% of variance. High heterogeneity was present across all groups (I² = 81–95%), and two zero-event studies in the Ultimaster group (Seth 2013, Separham 2011) required continuity correction, which may have attenuated heterogeneity estimates for that group.

Target Vessel Myocardial Infarction

The TV-MI analysis included 24 studies (N = 14,278) across Magmaris (13 studies), Biomatrix (6 studies), and Ultimaster (5 studies), with Magmaris as the reference (Tables S5A–S5B; Figure S55). The Magmaris group had the lowest pooled TV-MI rate at 1.78% (1.37–2.33%) and, notably, the lowest heterogeneity of any non-Magmaris reference group across all five outcomes (I² = 19.1%, tau² = 0.042). Biomatrix did not differ significantly from Magmaris (RR 1.142, 95% CI 0.803–1.623; p = 0.460), with wide confidence intervals reflecting high within-group heterogeneity (I² = 86.8%). Ultimaster was associated with a significantly higher TV-MI rate than Magmaris (RR 2.413, 95% CI 1.851–3.147; p < 0.001); however, as with TLF and TLR, this association is substantially influenced by the Lee 2019 outlier (TV-MI rate 8.87%), which is the primary driver of extreme heterogeneity in the Ultimaster group (I² = 96.3%). The omnibus test was highly significant (χ²(2) = 50.39, p < 0.0001), and stent type explained 24.0% of between-study variance.

Stent Thrombosis

The stent thrombosis analysis included 24 studies (N = 21,280) across Magmaris (5 studies), Biomatrix (5 studies), Orsiro (10 studies), and Ultimaster (4 studies), with Magmaris as the reference (Table S6A-S6B; Figure S56). The Magmaris cohort demonstrated a pooled stent thrombosis rate of 0.780% (95% CI 0.536–1.135%) with zero between-study heterogeneity (I² = 0.0%, tau² = 0), indicating highly consistent outcomes across its study programme. Of the three comparator stents, Biomatrix was the only one to demonstrate a significantly lower stent thrombosis rate than Magmaris (RR 0.430, 95% CI 0.211–0.876; p = 0.020), representing an approximately 57% relative reduction, and similarly displayed zero heterogeneity across its five contributing studies (I² = 0.0%). By contrast, both Orsiro and Ultimaster were associated with significantly higher stent thrombosis rates relative to Magmaris: Orsiro showed a 1.9-fold higher rate (RR 1.857, 95% CI 1.221–2.824; p = 0.004), and Ultimaster a 2.5-fold higher rate (RR 2.468, 95% CI 1.557–3.913; p < 0.001). The omnibus test was highly significant (χ²(3) = 35.911, p < 0.0001), and stent type explained 32.5% of between-study variance. These findings should nonetheless be interpreted with caution: high heterogeneity was observed within the Orsiro group (I² = 76.2%), driven in part by the Jiménez Díaz 2024 (ST rate 1.80%) and Smits 2024 (ST rate 2.75%) studies, while extreme heterogeneity within the Ultimaster group (I² = 90.6%) is predominantly attributable to the Lee 2019 study (ST rate 3.62%), which has similarly influenced findings across multiple outcomes in this series. As with all-cause mortality, the Magmaris reference cohort comprised predominantly small single-centre studies (median N = 207 per study) with a homogeneous patient population, which may not be fully representative of the broader, higher-risk populations enrolled in the Orsiro and Ultimaster programmes.

Cross-Outcome Synthesis and Heterogeneity Considerations

Across the five meta-regression models, a consistent pattern emerged: Magmaris demonstrated the lowest pooled rates for all-cause mortality and TV-MI and exhibited the lowest between-study heterogeneity of all stent groups evaluated, with I² values of 0.0% for all-cause mortality and 19.1% for TV-MI, and tau² values approaching zero for both endpoints. This internal consistency likely reflects the homogeneous patient population and standardised protocols characteristic of the Magmaris study programme, which predominantly enrolled patients in dedicated single-centre bioabsorbable scaffold registries.

Biomatrix consistently demonstrated lower revascularization rates relative to the comparator in both TLF (RR 0.744; p = 0.003) and TLR (RR 0.431; p < 0.001), while showing no significant difference in mortality endpoints. Ultimaster showed significantly higher rates for all-cause mortality, TLF, and TV-MI relative to Magmaris, though in each case these associations were heavily influenced by the Lee 2019 study, a single-centre registry with markedly elevated event rates (TLF 21.7%, TLR 10.4%, TV-MI 8.87%, cardiac death 8.63%). The weighted R² values across models ranged from 3.0% (cardiac death) to 45.0% (all-cause mortality), indicating that stent type accounted for a variable but generally modest to moderate proportion of observed heterogeneity, and that patient-level factors, study design, and follow-up duration represent important residual sources of variance. A full heterogeneity summary is provided in Table S2.

Analysis stratified by study design

The results of subgroup analyses stratified by study design are presented in the Supplementary Figures. Specifically, study type–stratified forest plots for clinical outcomes with the Ultimaster drug-eluting stent are provided in Figures S57–S60 (cardiac death: Figure S57; all-cause mortality: Figure S58; myocardial infarction: Figure S59; stent thrombosis: Figure S60), and for the Biomatrix drug-eluting stent in Figures S61–S65 (stent thrombosis: Figure S61; cardiac death: Figure S62; target vessel myocardial infarction: Figure S63; target lesion failure: Figure S64; target lesion revascularization: Figure S65).

Discussion

In this systematic review and single-arm meta-analysis, we evaluated clinical outcomes associated with the Magmaris bioresorbable magnesium scaffold compared with contemporary second-generation metallic drug-eluting stents—including Biomatrix, Ultimaster, and Orsiro—in patients with de novo coronary artery disease. Overall, our findings demonstrate consistently low rates of TLF, TLR, MI, and stent/scaffold thrombosis with Magmaris, supporting previous observations from dedicated registries and prospective trials. These results add to a growing evidence base indicating that modern resorbable and ultrathin-strut DES platforms offer excellent medium-term safety and efficacy profiles.

Principal Findings

Across more than 5,000 patients treated with Magmaris, the pooled incidence of TLF and TLR remained notably low, aligning with the encouraging  outcomes previously reported in BIOSOLVE-IV , which demonstrated low event rates and minimal scaffold thrombosis at 6–12 months of follow-up. [6,7] Consistently low rates of MI and target-vessel MI further reinforce the positive vascular healing characteristics described in earlier real-world analyses and first-in-human evaluations of magnesium-based scaffolds. [3–5]. Despite these findings, the interpretation of efficacy endpoints such as TLF and TLR is limited by substantial between-study heterogeneity (I² frequently exceeding 60–80%). This degree of variability reduces the robustness of pooled estimates and suggests that the observed effects may not be consistent across different clinical settings. While follow-up duration contributed to this variability, other unmeasured factors—including lesion complexity, implantation technique, and adjunctive imaging—likely play a significant role. Therefore, these pooled estimates should be interpreted as hypothesis-generating rather than definitive. In contrast, MI, target-vessel MI, and scaffold thrombosis showed minimal heterogeneity, underscoring the consistency of safety outcomes across diverse study designs and populations. A key finding of this analysis is the time-dependent increase in TLF and TLR observed with Magmaris. Unlike metallic DES, which demonstrate stable long-term outcomes, the progressive rise in adverse events between 1 and 2 years may reflect the biological and mechanical consequences of scaffold degradation. This phase may be associated with transient loss of radial strength, delayed vascular healing, or late recoil. Clinically, this raises important considerations regarding patient selection, lesion preparation, and the potential need for prolonged surveillance in patients receiving bioresorbable scaffolds.

Comparison With Contemporary DES Technologies

Second-generation drug-eluting stent (DES) platforms such as Orsiro and Ultimaster have demonstrated strong clinical performance, particularly in terms of reduced restenosis and late adverse cardiac events. Multiple meta-analyses and randomized trials have shown that thin- and ultrathin-strut biodegradable polymer sirolimus-eluting stents (Ultimaster ≈80 μm; Orsiro ≈60 μm) are highly competitive and, in some settings, may demonstrate advantages over conventional fluoropolymer DES. [1,2,8,9] Likewise, the Biomatrix stent has shown favorable long-term safety and durability in large multicenter analyses. [11] Within this broader context, the outcomes observed with Magmaris in the present analysis fall within a similar range for key safety endpoints, particularly with respect to low thrombosis rates—a notable consideration given historical concerns associated with earlier-generation bioresorbable scaffolds. The improved radial strength, faster endothelialization, and favorable resorption kinetics of magnesium-based platforms have been proposed as potential contributors to these findings, [3,6,7] and the pooled estimates in our analysis are consistent with these mechanistic considerations. Findings from the BIOSOLVE-IV registry, which enrolled over 1,000 patients and reported a 12-month TLF rate of 4.3% with very low scaffold thrombosis (0.5%), [12] and its full 24-month follow-up of 2,066 patients demonstrating a TLF rate of 6.8% and scaffold thrombosis of 0.8%, [13] provide important context. The results of our meta-analysis similarly demonstrate low rates of TLF, MI, and scaffold thrombosis, thereby consolidating evidence from multiple independent single-arm studies across a broader range of clinical settings. By integrating data from real-world registries and diverse patient populations, our analysis enhances the external validity of these observations. At the same time, it is important to interpret these findings within the limitations of indirect comparisons. Long-term randomized data for ultrathin-strut DES platforms such as Orsiro have demonstrated consistently low rates of TLF and target-vessel MI over 5 years, with stable performance and minimal evidence of late catch-up phenomena. In contrast, the time-dependent increase in TLF and TLR observed in Magmaris studies within this analysis suggests a distinct temporal pattern, the clinical implications of which remain to be fully defined. Accordingly, while Magmaris represents a promising alternative, current evidence is insufficient to establish equivalence with contemporary DES platforms, particularly for long-term outcomes. Further contextual data from the BIOFLOW-V trial demonstrated that Orsiro achieved significantly lower TLF and target-vessel MI rates compared to Biomatrix, with very low rates of late and very-late stent thrombosis; at two-year follow-up, reductions of 37% in TLF and 47% in ischaemia-driven TLR were observed relative to Biomatrix. [13] These findings highlight the high level of performance achieved by contemporary DES platforms. In this context, the present meta-analysis provides a comprehensive and up-to-date synthesis of Magmaris outcomes, which may serve as a reference point for future comparative investigations rather than as a basis for direct performance comparisons. Importantly, the inclusion of Biomatrix should be interpreted in light of its earlier-generation design. Compared with newer ultrathin-strut DES platforms such as Orsiro and thin-strut DES such as Ultimaster, Biomatrix represents a transitional technology with thicker struts and distinct polymer characteristics. As such, differences in observed outcomes across studies may, in part, reflect generational evolution in device design rather than intrinsic differences in clinical performance, and should therefore be interpreted with appropriate caution.

Clinical Implications

Our findings support the evolving role of biodegradable scaffolds as a viable alternative to permanent metallic DES in carefully selected patients and lesion subsets. The positive outcomes observed even in higher-risk groups—including diabetics and acute coronary syndrome patients, as reported in recent Magmaris cohorts [4,5]—suggest broader applicability than previously assumed. Given increasing interest in restoring vasomotion and reducing long-term metal burden, magnesium-based scaffolds may hold particular promise in younger patients or those for whom long-term vessel physiology is of clinical importance. However, until adequately powered randomized comparisons are available, the use of Magmaris will likely remain complementary rather than fully competitive with contemporary ultrathin-strut DES platforms.

Sources of Heterogeneity and Methodological Considerations

The substantial heterogeneity observed for TLF and TLR likely reflects differences in lesion characteristics, operator technique, imaging guidance, and follow-up duration across studies. Our meta-regression identified follow-up duration as the only consistent moderator of effect size. Importantly, no correlation was found between sample size and event rates, mitigating concerns about exaggerated effects in smaller single-arm cohorts. Although most outcomes demonstrated no significant publication bias, the presence of small-study effects in the MI analysis indicates the need for cautious interpretation. Additionally, the observational nature of all included studies introduces inherent limitations related to confounding and lack of comparator arms, despite adherence to prospective data collection and adjudication methods in the majority of studies.  A unique strength of our study is the quantitative exploration of follow-up duration as a driver of between-study heterogeneity: our meta-regression indicates that longer follow-up is significantly associated with variation in TLF and TLR event rates. This temporal insight is critical given the resorbable nature of Magmaris and is not thoroughly quantified in prior single-cohort reports. Additionally, signals of publication bias in certain endpoints—particularly myocardial infarction and all-cause mortality in Ultimaster studies—further limit confidence in pooled estimates and highlight the need for cautious interpretation of comparative findings.

Exploratory Comparison Across Stent Platforms Using Meta-Regression

To further explore whether stent platform may contribute to variability in clinical outcomes, we performed meta-regression analyses incorporating stent type as a study-level covariate. Importantly, these analyses do not constitute direct comparative effectiveness but rather provide an exploratory framework to assess whether differences in observed event rates across studies may be partially explained by device type. Across multiple endpoints, stent type accounted for a variable proportion of between-study heterogeneity, ranging from minimal contribution in cardiac death (R² = 3%) to more substantial explanatory value for all-cause mortality (R² = 45%) and moderate contributions for TLF, TLR, and TV-MI (~24–28%). These findings suggest that while device platform may influence observed outcomes, a large proportion of heterogeneity remains attributable to other factors such as patient selection, lesion complexity, procedural technique, and study design. For all-cause mortality, studies evaluating Magmaris demonstrated consistently low and homogeneous event rates, whereas higher rates were observed in studies of Orsiro and Ultimaster within the regression framework. However, this pattern should be interpreted cautiously, as Magmaris data were predominantly derived from smaller, single-centre cohorts with more selected patient populations, while comparator stents were evaluated in larger, multicentre studies with broader and potentially higher-risk populations. This imbalance likely contributes to the observed differences and limits causal interpretation. For efficacy endpoints such as TLF and TLR, absolute pooled rates were broadly similar across stent platforms; however, meta-regression suggested statistical differences in study-level estimates. Notably, these findings were sensitive to influential outlier studies, particularly within the Ultimaster dataset, where a single high-event study contributed disproportionately to both effect estimates and heterogeneity. This underscores the instability of indirect comparisons based on aggregate data and highlights the importance of considering study-level variability. A similar pattern was observed for TV-MI and stent thrombosis, where regression analyses indicated differences across stent groups; however, these were again accompanied by substantial heterogeneity and potential confounding by study design and population characteristics. In particular, the low heterogeneity observed in Magmaris cohorts likely reflects greater uniformity in study protocols rather than inherently superior or inferior device performance. Taken together, these findings suggest that stent type may act as one of several contributors to observed differences in clinical outcomes across studies, but should not be interpreted in isolation. The absence of randomized head-to-head comparisons, combined with imbalances in study design and patient populations, precludes definitive conclusions regarding relative efficacy or safety. Accordingly, these meta-regression results should be considered hypothesis-generating and supportive of the need for adequately powered comparative trials.

Sensitivity and Subgroup Analysis: RCTs vs. Observational Data

A critical finding of this systematic review is the influence of study design on reported stent performance, particularly for the Ultimaster drug-eluting stent. Stratified analysis (Figures S57–S60) demonstrates that event rates for cardiac death, all-cause mortality, and myocardial infarction are markedly higher in observational registries than in randomized controlled trials. The stratified forest plots reveal that this risk is disproportionately driven by the Lee 2019 observational study, which reported an outlier TV-MI rate of 8.87%. In contrast, data from Biomatrix evaluations showed greater stability across study designs, with its significantly lower rate of stent thrombosis (RR 0.430, p = 0.020) and TLF (RR 0.744, p = 0.003) remaining consistent in both stratified and pooled analyses. Magmaris and Orsiro study design stratified analysis was not performed as their primary analysis is based on a single type of study design. These findings underscore the necessity of accounting for study design and outlier influence when synthesizing safety data for bioresorbable and drug-eluting platforms, and the results must be interpreted with caution.

Strengths and Limitations

The strengths of this review include comprehensive literature coverage, strict inclusion criteria, ROBINS-I–guided risk of bias assessment, and robust statistical methodology including subgroup, sensitivity, and meta-regression analyses. These methodological steps enhance the reliability and generalizability of our pooled findings. Nevertheless, several limitations must be acknowledged. First, the absence of randomized comparative studies limits direct assessment of performance versus established DES platforms such as Biomatrix or Orsiro. Second, relatively short-term follow-up in many Magmaris studies limits conclusions regarding late adverse events, which are particularly relevant for resorbable technologies. Third, device-specific data could not be extracted consistently across all endpoints, and residual confounding remains possible due to the inherent limitations of single-arm designs.

Conclusions

This systematic review and meta-analysis demonstrates that the Msagmaris bioresorbable magnesium scaffold is associated with acceptable medium-term safety and efficacy outcomes, with low rates of TLF, MI, *.

References

1. Cassese S, Ndrepepa G, Byrne RA, et al: Outcomes of patients treated with ultrathin-strut biodegradable polymer sirolimus-eluting stents versus fluoropolymer-based everolimus-eluting stents: a meta-analysis of randomised trials. EuroIntervention 14 (2018): 224–231.
2. Lupi A, Schaffer A, Bongo AS: Should ultrathin strut drug eluting stents be considered the new benchmark for novel coronary stents approval? J Thorac Dis 10 (2018): 678–681.
3. Galli S, Troiano S, Palloshi A, et al: Sustained Safety and Efficacy of Magnesium Reabsorbable Scaffold: 2-Year Follow-Up Analysis From First Magmaris Multicenter Italian Registry. Cardiovasc Revasc Med 41 (2022): 69–75.
4. Włodarczak A, Łanocha M, Szudrowicz M, et al: The 1-Year Safety and Efficacy Outcomes of Magmaris, Novel Magnesium Bioresorbable Vascular Scaffolds in Diabetes Mellitus Patients with Acute Coronary Syndrome. J Clin Med 10 (2021): 3166.
5. Włodarczak A, Rola P, Barycki M, et al: Mid-term safety and efficacy of magnesium bioresorbable vascular scaffolds – Magmaris in diabetic population. 2-Years outcome in acute coronary syndrome cohort. Diabetes Vasc Dis Res 20 (2023): 14791641231188705.
6. Verheye S, Wlodarczak A, Montorsi P, et al: BIOSOLVE-IV-registry: Safety and performance of the Magmaris scaffold: 12-month outcomes of the first cohort of 1,075 patients. Catheter Cardiovasc Interv 98 (2021): E1–E8.
7. Wells GA, Shea B, O’Connell D, et al: The Newcastle–Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Hospital Research Institute.
8. Chisari A, Pistritto AM, Piccolo R, et al: The Ultimaster Biodegradable-Polymer Sirolimus-Eluting Stent: An Updated Review of Clinical Evidence. Int J Mol Sci 17(2016): 1490.
9. Hansen KN, Jensen LO, Maeng M, et al. Five-Year Clinical Outcome of the Biodegradable Polymer Ultrathin Strut Sirolimus-Eluting Stent Compared to the Biodegradable Polymer Biolimus-Eluting Stent: From the SORT OUT VII Trial. Circ Cardiovasc Interv 16(2023): e012332.
10. Park S, Rha SW, Choi BG, et al: Efficacy and Safety of Sirolimus-Eluting Stent With Biodegradable Polymer Ultimaster™ in Unselected Korean Population. Korean Circ J 54 (2024): 339–350.
11. Eccleston DS, Chowdhury E, Rafter T, et al: Long-Term Outcomes of Contemporary Percutaneous Coronary Intervention with the Xience Drug-Eluting Stent: Results from a Multicentre Australian Registry. J Clin Med 12 (2022): 280.
12. Verheye S, Wlodarczak A, Montorsi P, et al: BIOSOLVE-IV-registry: Safety and performance of the Magmaris scaffold: 12-month outcomes of the first cohort of 1,075 patients. Catheter Cardiovasc Interv 98 (2021): E1–E8.
13. Wlodarczak A, Montorsi P, Torzewski J, et al: One- and two-year clinical outcomes of treatment with resorbable magnesium scaffolds for coronary artery disease: the prospective, international, multicentre BIOSOLVE-IV registry. EuroIntervention 19 (2023): 232–239.
14. BIOTRONIK AG, Orsiro (SES) — 24-Month Clinical Results (BIOFLOW V). Berlin: BIOTRONIK AG (2018).