Comparative efficacy of open endarterectomy versus endovascular treatments in common femoral artery stenocclusive disease: A Bayesian hierarchical meta‐analysis

Abstract

Background

The management of Common Femoral Artery (CFA) stenocclusive disease presents unique challenges, with debates persisting regarding the optimal treatment strategy between Open Common Femoral Endarterectomy (CFE) and Endovascular Therapy (EVT). This study aimed to compare the efficacy of CFE versus EVT, as well as simple versus advanced EVT (aEVT), using a Bayesian hierarchical framework.

Methods

A systematic review was conducted according to PRISMA guidelines by querying PubMed, ScienceDirect, Scopus, and the Cochrane Library for articles published through June 2024. The study included randomized controlled trials and retrospective observational studies with a minimum of 40 patients per cohort. Data were synthesized using a Bayesian random-intercept logistic model to estimate pooled Odds Ratios (OR) for primary patency, freedom from reintervention, and limb salvage.

Results

The analysis included 10 comparative studies encompassing 1452 patients. Open CFE demonstrated superior outcomes compared to EVT, with a significantly reduced odds of patency loss (OR 0.480; 95% CrI: 0.367–0.626) and lower reintervention rates (OR 0.534; 95% CrI: 0.391–0.718). Limb salvage rates also favored CFE (OR 0.548; 95% CrI: 0.402–0.737). In subgroup analyses, advanced EVT (involving atherectomy) improved primary patency compared to simple EVT (OR 0.520; 95% CrI: 0.306–0.883) but did not statistically significantly reduce the need for reintervention (OR 0.910; 95% CrI: 0.427–1.894).

Conclusion

Open CFE remains the gold standard for CFA stenocclusive disease, offering superior long-term durability and limb salvage compared to endovascular interventions. While advanced endovascular techniques improve patency rates compared to simple balloon angioplasty, they fail to significantly reduce reintervention rates, likely due to the biomechanical stress of the femoral region. EVT should be considered a pragmatic option reserved primarily for high-surgical-risk patients.

Keywords

Common femoral artery endarterectomy endovascular therapy Bayesian meta-analysis primary patency limb salvage reintervention

Introduction

Peripheral artery disease (PAD) is a major global health concern, with atherosclerotic involvement leading to stenosis and occlusions that can compromise limb perfusion and function¹. The common femoral artery (CFA), in particular, plays a critical role in lower extremity blood supply and is frequently afflicted by atherosclerotic lesions due to its anatomical location and exposure to mechanical stress². The management of CFA stenocclusive disease poses unique challenges: while open common femoral endarterectomy (CFE) is recognized for its long‐term durability with reported 5‐year primary patency rates exceeding 90%^3,4 its invasiveness is associated with higher perioperative morbidity, including wound complications and extended recovery⁵. Conversely, endovascular therapy (EVT) offers a less invasive alternative with shorter hospital stays; however, EVT is often linked with increased rates of restenosis and the need for reintervention, particularly due to the biomechanical stresses imposed on stents and balloons in the groin region^6–8.

Prior systematic reviews and meta‐analyses have yielded mixed results due to heterogeneity in study design, patient populations, and evolving device technology^5,6. Recent randomized trials^7,8 and observational studies^9–11 have provided further insight, yet debates remain regarding the optimal treatment strategy. In addition, the emergence of “advanced” endovascular techniques combining atherectomy with adjunctive stenting or drug‐coated balloon therapy has further complicated the therapeutic landscape^12–14. The current clinical guidelines of the European Society of Cardiology (ESC)³ and the National Institute for Health and Care Excellence (NICE)⁴ continue to emphasize individualized treatment decisions based on disease severity, anatomical complexity, and patient comorbidities¹⁵.

The advent of Bayesian hierarchical meta‐analysis has provided a robust framework for synthesizing heterogeneous data while appropriately accounting for between‐study variability¹⁶. By incorporating prior distributions and utilizing Markov chain Monte Carlo (MCMC) sampling, Bayesian methods offer a flexible approach to estimate pooled effect sizes and assess model adequacy through predictive checks and information criteria (e.g., WAIC)¹⁷. In this study, we integrate data from 10 comparative studies and implement a Bayesian random‐intercept logistic model to compare open CFE with EVT, as well as to compare EVT versus advanced EVT (aEVT), across three critical clinical endpoints: primary patency, freedom from reintervention, and limb salvage. Moreover, we conduct a sub‐analysis contrasting simple (balloon ± drug‐coated) and advanced endovascular approaches.

We further report the outcomes from our new analysis pipeline implemented in R using packages such as rjags, bridgesampling, and runjags which provides detailed diagnostic outputs (e.g., trace plots, Gelman–Rubin statistics, and effective sample sizes) and sensitivity analyses. Our aim is to generate robust, reproducible estimates that inform evidence‐based clinical decision‐making in the management of CFA stenocclusive disease.

Methods

Study selection and data extraction

Literature search and study eligibility

We conducted this meta‐analysis in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines. Four electronic databases PubMed, ScienceDirect, Scopus, and the Cochrane Library were queried through June 2024. Search terms included “common femoral artery,” “endarterectomy,” “atherectomy,” “balloon angioplasty,” “drug‐coated balloon,” “stent,” “limb salvage,” “primary patency,” and “reintervention,” combined using Boolean operators (e.g., (common femoral) AND (endarterectomy OR atherectomy OR stent OR balloon)). The initial search retrieved 1143 articles (75 in PubMed, 74 in ScienceDirect, 910 in Scopus, and 84 in the Cochrane Library). After de‐duplication, 1025 unique records were retained for abstract/title screening.

Two reviewers independently (N.G. and T.P.) evaluated abstracts and titles. We excluded non‐English conference abstracts, studies lacking relevant comparative data, cohorts with fewer than 20 patients per study arm, and investigations focused on non‐CFA sites. This screening yielded 129 full‐text articles, from which an additional set of exclusions was applied, ultimately leaving 10 eligible studies. Inclusion criteria were as follows:

(1) Population: Adult patients (≥18 years) with CFA stenocclusive disease necessitating revascularization.

(2) Comparative Design: Each study compared either (a) open endarterectomy versus any endovascular (EV) approach, or (b) simple versus advanced endovascular interventions.

(3) Outcomes: Binomial event data (events/sample) for at least one of the following endpoints: primary patency, freedom from reintervention, or limb salvage.

(4) Minimum Sample Size: At least 40 patients.

(5) Study Type: Randomized controlled trials or retrospective/observational investigations.

A PRISMA flow diagram (S.1) details the selection process, from initial identification to final inclusion.

Study registration

This systematic review and meta-analysis were prospectively registered with PROSPERO registration number CRD42025649928. The protocol is publicly available through the PROSPERO database, ensuring transparency and reproducibility of the study methodology.

Data extraction and study characteristics

Two independent reviewers extracted data from the included 10 studies, reconciling any discrepancies through discussion. For each study, we recorded publication year and country, study design (randomized vs observational), sample sizes per group, mean age, sex distribution, and comorbidities (e.g., diabetes, coronary artery disease, and hypertension), lesion classifications (Rutherford categories, if reported), follow‐up duration, and methods of outcome assessment (Table 1).

Table 1.

Baseline characteristics, demographics, and lesion classifications of the included studies.

Study	First author	Publication year	Country	Study design	Sample size	Age mean	Males	Diabetes	Hypertension	Smoking	Hyperlipidemia	Coronary artery disease	Chronic kidney disease	Lesion type	Rutherford category
1	Nakama et al.	2022	Japan	Multicenter retrospective observational study	694	74.5	526	419	N/A	499	N/A	305	398	1 = 82, 2 = 380, 3 = 232	Rutherford 2 = 197, Rutherford 3 = 331, Rutherford 4 = 57, Rutherford 5 = 109,
2	Linni et al.	2014	Austria	Single-center randomized controlled trial	80	72	52	24	65	40	50	35	N/A	N/A	Rutherford 2 = 13, Rutherford 3 = 35, Rutherford 4 = 11, Rutherford 5 = 18, Rutherford 6 = 3
3	Guo et al.	2018	China	Single-center retrospective observational	90	68.35	48	50	25	41	31	18	6	N/A	N/A
4	Kapetanios et al.	2024	Germany	Two-center retrospective observational	42	70.48	19	13	36	31	25	7	9	1 = 15, 2 = 16, 3 =11	Rutherford 2 = 12, Rutherford 3 = 17, Rutherford 4 = 6, Rutherford 5 = 6, Rutherford 6 = 1
5	Elsherif et al.	2018	Ireland	Single-center retrospective observational	40	69.66	33	20	36	31	32	18	15	N/A	N/A
6	Stavroulakis et al.	2017	Germany	Single-center retrospective observational	47	71	26	19	45	N/A	33	22	1	N/A	Rutherford 2 = 2, Rutherford 3 = 27, Rutherford 4 = 7, Rutherford 5 = 7, Rutherford 6 = 4
7	Imran et al.	2018	USA	Single-center retrospective observational	70	70.9	46	28	66	57	66	42	30	1 = 10, 2 = 14, 3 =46	Rutherford 2,3 = 48, Rutherford 4 = 16, Rutherford 5,6 = 6
8	Kuo et al.	2019	Taiwan	Single-center retrospective observational	100	71.8	75	68	53	51	42	46	34	N/A	Rutherford 3 = 10, Rutherford 4 = 35, Rutherford 5 = 38, Rutherford 6 = 17
9	Mehta et al.	2016	USA	Multicenter retrospective observational	167	68	101	72	144	42	47	103	4	N/A	Rutherford 3 = 91, Rutherford 4 = 31, Rutherford 5 = 24, Rutherford 6 = 11
10	Gouëfic et al.	2017	France	Multicenter randomized controlled trial	117	68	99	42	89	54	87	55	14	1 = 15, 2 = 34, 3 = 68	Rutherford 2 = 3, Rutherford 3 = 98, Rutherford 4 = 12, Rutherford 5 = 3,

We classified open CFE arms as those explicitly labeled “endarterectomy” without mention of “atherectomy.” Interventions mentioning balloon angioplasty, stenting, or atherectomy were deemed endovascular (EV). For the sub‐analysis, we distinguished “advanced” EV arms (directional/orbital atherectomy ± stent/balloon) from “simple” EV arms (balloon ± drug‐coated balloon [DCB]) based on the text descriptions.

Outcome definitions

We compiled binomial event data (number of events and total participants) for three key outcomes:

(1) Primary Patency: Maintained patency of the treated artery, typically defined as ≤50% restenosis or no occlusion by duplex ultrasound or angiography at ∼12–48 months of follow‐up.

(2) Freedom from Reintervention: The proportion of patients who did not undergo any further endovascular or surgical procedure on the target lesion during the study period.

(3) Limb Salvage: The absence of major (above‐ankle) amputation throughout follow‐up.

Bayesian hierarchical model specification

For each outcome, we assumed that the observed number of events yiyi followed a binomial distribution:

yi \sim Binomial (ni, pi), yi \sim Binomial (ni, pi),

where nini is the sample size and pipi is the event probability for observation ii. To model the effect of treatment, we used a logistic regression framework with study-level random intercepts:

\ logit (pi) = α j [i] + β \times treatmenti, \ logit (pi) = α j [i] + β \times treatmenti,

where αj[i]αj[i] represents the baseline risk specific to study jj corresponding to observation ii, and ββ is the global log-odds ratio comparing the two treatment modalities. A negative value of ββ implies a benefit for the reference treatment (open CFE).

Random intercepts for each study were modeled as

α j \sim N (μ α, τ α), α j \sim N (μ α, τ α),

with τα=σα−2τα=σα−2. We assigned the following weakly informative priors:

• β∼N(0,0.52)β∼N(0,0.52)

• μα∼N(0,0.12)μα∼N(0,0.12)

• σα∼Uniform(0,5)σα∼Uniform(0,5)

For the subgroup analysis comparing simple versus advanced endovascular therapies, the same model was employed with the treatment indicator defined by the presence of atherectomy in the intervention description.

Model implementation and diagnostics

The model was implemented using the rjags package in R (version 4.4.2). Two parallel MCMC chains were run for 20,000 iterations each, discarding the initial 6000 iterations as burn-in and applying a thinning factor of 4 to mitigate autocorrelation. Convergence was assessed using trace plots, Gelman–Rubin diagnostic statistics (target R^≈1.00 R^≈1.00), and effective sample size (ESS) calculations. Posterior summaries (median, standard deviation, and 95% credible intervals) were calculated for all parameters; the posterior samples for ββ were exponentiated to yield odds ratios (ORs).

Model adequacy was evaluated via posterior predictive checks. We generated replicated datasets from the median posterior parameter estimates and compared the mean event rates in the observed and replicated data. In addition, the widely applicable information criterion (WAIC) was computed for each model to facilitate model comparison and penalize overfitting. Sensitivity analyses were performed by altering the prior distributions (e.g., using β∼N(0,12)β∼N(0,12) instead of N (0,0.52)N(0,0.52)).

Analysis pipeline

We developed a custom R function, BayesCFAStudy, to streamline the entire analysis process. This function loads the primary patency, freedom from reintervention, and limb salvage datasets; constructs subgroup data frames using text-mining techniques to distinguish between interventions (based on keywords “endarterectomy” and “atherectomy”); and runs the Bayesian models via rjags. The pipeline outputs detailed diagnostic plots including trace plots, autocorrelation plots, and forest plots of the baseline odds (exponentiated random intercepts) and computes WAIC values for each model. Additional functions perform sensitivity analyses and posterior predictive checks.

Risk-of-bias assessment

Risk of bias was evaluated separately for randomized and observational studies. For randomized controlled trials, we applied the revised Cochrane risk‐of‐bias tool (RoB 2). Domain‐level judgments were used to derive an overall risk‐of‐bias rating for each trial. Detailed RoB 2 assessments for each randomized study are provided in (Supplemental Table S2).

For observational studies, quality was appraised using the Newcastle–Ottawa Scale (NOS), which examines selection, comparability, and outcome ascertainment. Studies were awarded stars based on predefined criteria, with higher scores indicating a lower risk of bias. Full NOS ratings and justifications for each observational study are available in (Supplemental Table S3).

Results

Study characteristics

A total of 10 studies, representing 1452 patients, met the inclusion criteria. The studies were conducted in diverse regions including Japan, Austria, China, Germany, the USA, Ireland, Taiwan, and France. Key characteristics of these studies, including study design, sample size, mean age, and details of the interventions (open CFE vs EVT and endovascular subtypes) are summarized in Table 1.^{7,8,11,18–24}.

Primary patency

Data for primary patency were available from six studies comparing open CFE with EVT. The Bayesian model converged successfully (all Gelman–Rubin R^R^ values were ∼1.00), and the effective sample sizes were robust (all exceeding 6000). The posterior median for the treatment effect (ββ) was estimated to be −0.734 (95% CrI: −1.002 to −0.468). Exponentiating this estimate produced an OR of 0.480 (95% CrI: 0.367–0.626), which implies that open CFE reduces the odds of patency loss by approximately 52% compared with EVT^{7,8,18,20,23,24} (Figures 1 and 2).

Figure 1.

Bayesian diagnostic plots (trace plot, posterior density, and autocorrelation) for the primary patency analysis (endarterectomy vs endovascular therapy)

Figure 2.

Forest plot for baseline odds for primary patency (endarterectomy vs endovascular therapy

Freedom from reintervention

For freedom from reintervention, data were pooled from four studies. The posterior median ββ was −0.627 (95% CrI: −0.938 to −0.331), corresponding to an OR of 0.534 (95% CrI: 0.391–0.718)^7,20,23,24 .These findings indicate that patients undergoing open CFE are less likely to require repeat interventions compared to those treated with EVT (Figures 3 and 5).

Figure 3.

Bayesian diagnostic plots (trace plot, posterior density, and autocorrelation) for the freedom from reintervention analysis (endarterectomy vs endovascular therapy)

Figure 4.

Forest plot for baseline odds for freedom from reintervention (angioplasty vs advanced endovascular therapy)

Limb salvage

Seven studies contributed to the limb salvage outcome analysis. The overall posterior median ββ was −0.602 (95% CrI: −0.912 to −0.306), translating to an OR of 0.548 (95% CrI: 0.402–0.737)^8,20,23,24. This suggests that open CFE provides a 45% reduction in the odds of major amputation relative to EVT. Although seven studies reported limb salvage outcomes, three of them either had zero events in both treatment arms or provided insufficient detail for inclusion in the pooled analysis (Figures 4 and 6).

Figure 5.

Bayesian diagnostic plots (trace plot, posterior density, and autocorrelation) for the limb salvage analysis (endarterectomy vs endovascular therapy

Figure 6.

Forest plot for baseline odds for limb salvage (endarterectomy vs endovascular therapy)

Subgroup analysis: Simple versus advanced endovascular techniques

In the subgroup analysis comparing endovascular strategies, studies that contrasted simple balloon angioplasty (with or without drug-coated balloons) with advanced techniques (involving atherectomy ± stenting) were considered. For primary patency within this subgroup, the posterior median ββ was −0.655 (95% CrI: −1.184 to −0.125), yielding an OR of 0.520 (95% CrI: 0.306–0.883), favoring advanced endovascular approaches. However, for freedom from reintervention, the median ββ was −0.094 (95% CrI: −0.851 to 0.639), corresponding to an OR of 0.910 (95% CrI: 0.427–1.894); this result was not statistically significant^11,19,21,22. No valid two-arm comparisons could be performed for limb salvage in this subgroup (Figures 7–10).

Figure 7.

Bayesian diagnostic plots (trace plot, posterior density, and autocorrelation) for the subgroup primary patency analysis (angioplasty vs advanced endovascular therapy

Figure 8.

Forest plot for baseline odds (exp(alpha)) for subgroup primary patency (angioplasty vs advanced endovascular therapy)

Figure 9.

Forest plot for baseline odds for freedom from reintervention (endarterectomy vs endovascular therapy)

Figure 10.

Bayesian diagnostic plots (trace plot, posterior density, and autocorrelation) for the subgroup freedom from reintervention analysis (angioplasty vs advanced endovascular therapy)

Posterior predictive checks and sensitivity analyses

Posterior predictive checks for the primary patency model indicated an observed mean event rate of 0.390, compared with a replicated mean of 0.206, suggesting moderate calibration. Sensitivity analyses—using alternative priors (e.g., β∼N (0,12)β∼N (0,12), μα∼N(0,0.52)μα∼N(0,0.52), and σα∼Uniform(0,10)σα∼Uniform(0,10))—produced a median ββ of −0.645, a median μαμα of −0.238, and a median σασα of 2.196. These results confirm that the data dominate the posterior estimates.

Model comparison

WAIC was computed for each analysis as a measure of model fit while penalizing complexity. For the primary patency analysis (open CFE vs EVT), WAIC was 9324.210 (SE: 7958.619). The freedom from reintervention model had a WAIC of 2013.729 (SE: 908.718), and the limb salvage model’s WAIC was 23836.518 (SE: 14603.399). In the subgroup analyses comparing simple and advanced EVT, WAIC values were 10045.259 (SE: 8577.137) for primary patency and 22042.021 (SE: 18199.415) for freedom from reintervention. Although the standard errors were large, likely due to variability and small sample sizes, the overall direction and magnitude of the effect estimates were consistent.

Discussion

This meta-analysis represents a significant methodological advancement over prior frequentist approaches, for example, Jia et al. and Boufi et al., by employing Bayesian hierarchical modeling to account for between study variability^5,6 Traditional fixed-effect models often underestimated heterogeneity, leading to inflated confidence in pooled estimates Zhang et al.¹⁶. Our approach, using weakly informative priors (β ∼ N (0, 0.5²)) and study-level random intercepts (σα = 2.196), aligns with Spiegelhalter et al. framework, which emphasizes adaptability in multi-level data synthesis¹⁷. For instance, Wang et al.²⁵ adopted similar Bayesian methods but omitted critical subgroup analyses of endovascular subtypes a gap our study addressed by stratifying simple (balloon ± DCB) versus advanced EVT (atherectomy ± stenting), paralleling device-specific evaluations by Stavroulakis et al.¹⁹.

Convergence metrics (Gelman–Rubin R̂ ≈ 1.00, ESS >6000) surpassed benchmarks from Zhang et al. ensuring robust parameter estimation^16. Posterior predictive checks revealed discrepancies between observed (0.390) and replicated (0.206) patency rates, highlighting unmodeled confounders such as lesion calcification Boufi et al. and stent design heterogeneity factors Siracuse et al. linked to EVT outcome variability^5,13. Sensitivity analyses with alternative priors (β ∼ N (0, 1²)) confirmed result stability (Δβ < 10%), reinforcing Spiegelhalter et al.’s (2002) assertions about Bayesian robustness to prior selection.

Risk of bias and data limitations

Our exclusion of studies with <20 patients/arm versus Kapetanios et al. 2024 prioritized precision but excluded niche cohorts, a tradeoff evident in posterior predictive checks²³ Risk-of-bias assessments using Cochrane RoB 2 and NOS mirrored Vascular Quality Initiative protocols, yet operator experience disparities Siracuse et al. remained unquantified due to inconsistent reporting¹³ These limitations reflect broader challenges in vascular meta-analyses, where heterogeneous lesion classifications Rutherford et al. and follow-up protocols complicate synthesis². For example, Halpin et al. emphasized the need for standardized reporting of lesion calcification and stent types variables only partially captured in our analysis¹².

Clinical outcomes

Open CFE demonstrated a 52% reduction in patency loss odds (OR 0.48, 95% CrI: 0.367–0.626), aligning with Ballotta et al. 96% 7-year patency and Wieker et al. 90.2% 3-year patency^9,10. Mechanistically, CFE’s plaque excision eliminates restenosis nidi, whereas EVT’s transient lumen gain leaves residual calcification prone to recoil particularly under femoral biomechanical stress Bonvini et al.²⁶. EVT’s inferior performance (1-year patency 70–75% for stents vs CFE’s >90%) mirrors registry data from Malgor, Soga and Hong et al., who noted higher restenosis in stent-covered profunda origins^27–29.

Advanced EVT’s modest patency advantage (OR 0.52 vs simple EVT) aligns with Guo et al., where directional atherectomy + DCB achieved 85% 1-year patency. However, Nakama et al. CAULIFLOWER trial highlighted that even drug-eluting stents (88% 1-year patency) lag behind CFE’s long-term durability, underscoring the femoral region’s hostile biomechanics^21,24. These findings align with Kuma et al., who reported 5-year patency rates of 89% for CFE versus 65% for EVT in their single-center series³⁰.

Open CFE’s 46% lower reintervention odds (OR 0.53, 95% CrI: 0.391–0.718) reflect Malgor et al. findings of 5–10% 5-year reoperation rates²⁸. By contrast, EVT’s 15–30% reintervention rates Mehta et al. and Banerjee et al. stem from stent fractures and in-stent restenosis phenomena exacerbated by femoral flexion Kuo et al.^8,11,31. The XLPAD registry Banerjee et al. reported 38–45% cost inflation from repeated EVT, whereas CFE’s upfront invasiveness offers cost-effectiveness for patients with >2-year life expectancy³¹.

Subgroup analyses revealed nonsignificant reintervention differences between simple and advanced EVT (OR 0.91), contradicting Guo et al. but aligning with Kapetanios et al.^21,23. This paradox suggests advanced devices improve acute luminal gain but fail to address femoral stress a limitation highlighted by Hong et al., where 32% of stents required reintervention within 18 months²⁹. For example, Mehta et al. reported 24% reinterventions at 1 year for atherectomy + stenting, while Gouëffic et al. found 22% reinterventions for bare-metal stents at 12 months^11,18.

Open CFE’s 45% limb salvage advantage (OR 0.55, 95% CrI: 0.402–0.737) correlates with Nishibe et al.¹⁴ 92% salvage rates in CLTI. By preserving profunda femoris flow a critical collateral pathway CFE mitigates the “last straw” phenomenon when SFA disease progresses Rutherford et al.² EVT’s inferior performance 6% major amputation versus CFE’s 3% in Kuo et al.⁸ underscores its limitations in bifurcation disease. Hybrid approaches (e.g., iliac stenting + CFE), as in Piazza et al.,³² achieved 91–95% 3-year patency, rivaling open bypass outcomes but requiring surgical femoral repair.

Subgroup analyses and technological evolution

The subgroup analysis favoring advanced EVT for patency (OR 0.52) but not reintervention mirrors Stavroulakis et al.¹⁹ where directional atherectomy + DCB yielded 80% 1-year patency versus 68% for DCB alone. However, Kapetanios et al.²³ reported 24% 2-year reinterventions for advanced EVT versus 10% for CFE, emphasizing femoral biomechanical stress. Emerging technologies like DES Nakama et al.²⁴ may narrow this gap, but long-term data remain sparse. For instance, the CAULIFLOWER trial’s 88% 1-year DES patency versus CFE’s 82% suggests promise, yet 3-year data are pending. Additionally, a separate study by Bonvini et al.³³ focused on angioplasty with provisional stenting in isolated CFA lesions, reporting favorable mid‐term outcomes and relatively low complication rates thus reinforcing the feasibility of an endovascular‐first approach for selected patients.

ESC/NICE guidelines appropriately recommend CFE for low-risk patients but underestimate EVT’s niche in high-risk cohorts. Azéma et al.³⁴ achieved 89% technical success with EVT in octogenarians, avoiding general anesthesia critical for frail patients Boufi et al.⁵ Conversely, CFE’s 30-day morbidity (2–4%) justifies its use in operable patients despite higher upfront costs Banerjee et al.,³¹ Our Bayesian models, weighted by study design rigor, reinforce these nuanced recommendations. For example, Halpin et al.¹² noted that EVT’s perioperative mortality (1–2%) makes it preferable for patients with severe CAD or COPD.

Limitations and methodological reflections

Posterior predictive check discrepancies highlight unaccounted variables like lesion calcification Boufi et al.,⁵ and operator experience Siracuse et al.¹³ Individual participant data (IPD) meta-analyses, as advocated by Zhang et al., could adjust for these factors¹⁶. Standardized reporting of lesion characteristics (TASC II, calcification) and procedural details (stent/balloon specifications) remains imperative. For example, only 3/10 studies reported calcification scores a critical predictor of EVT failure Boufi et al.⁵.

While CFE’s durability suggests cost-effectiveness, formal analyses are lacking. Banerjee et al. (2019) calculated CFE becomes cost-effective at 3 years for patients with >50% 5-year survival. Hybrid approaches (e.g., iliac stenting + CFE), as in Piazza et al. merit further study for multi-level disease. For instance, hybrid FEIA achieved 95% 3-year patency in TASC D lesions, comparable to open aortobifemoral bypass³².

Future directions and clinical implications

Next-generation devices (e.g., calcification-resistant stents, DCB + atherectomy) may improve EVT outcomes. The CAULIFLOWER trial Nakama et al.,²⁴ reported 88% 1-year DES patency, but long-term data are needed. Stavroulakis et al.¹⁹ demonstrated that directional atherectomy + DCB reduces restenosis by 40% compared to balloon alone, yet femoral biomechanics remain a barrier.

Large RCTs comparing advanced EVT versus CFE are urgently needed. Small trials Linni et al. Gouëffic et al.^7,18 lack power for long-term endpoints. Registries like the Vascular Quality Initiative must standardize data collection to enable robust analyses. For example, Siracuse et al.¹³ highlighted operator experience as a key variable yet only 20% of studies reported this metric.

Banerjee et al. estimated that CFE becomes cost-effective at 3 years for patients with >50% 5-year survival. Future studies should integrate QALYs and reintervention costs, as EVT’s cumulative expenses often exceed CFE’s upfront costs³¹.

Conclusion

This Bayesian meta-analysis solidifies open CFE’s status as the gold standard for CFA disease, reducing patency loss by 52% and reinterventions by 46% versus EVT. Methodologically, hierarchical modeling advanced synthesis of heterogeneous data, though limitations in primary study reporting persist. While advanced EVT narrows the patency gap, its failure to reduce reinterventions underscores femoral biomechanical challenges. For high-risk patients, EVT remains pragmatic, but CFE’s durability justifies its use in operable candidates. Future research must prioritize RCTs comparing next-gen EVT devices with CFE, coupled with cost-utility analyses to guide value-based care.

Supplemental material

Supplemental Material - Comparative efficacy of open endarterectomy versus endovascular treatments in common femoral artery stenocclusive disease: A Bayesian hierarchical meta‐analysis

Supplemental Material for Comparative efficacy of open endarterectomy versus endovascular treatments in common femoral artery stenocclusive disease: A Bayesian hierarchical meta‐analysis by Nikolaos-Nektarios Giannakopoulos, Sofia Tzamtzidou, Dimitra Manou, Konstantinos Nakas, Konstantinos Seretis, Paraskevi Tsiantoula, Konstantinos Roditis, Vasileios Papaioannou, and Theofanis Papas in Vascular.

Supplemental material

Supplemental Material - Comparative efficacy of open endarterectomy versus endovascular treatments in common femoral artery stenocclusive disease: A Bayesian hierarchical meta‐analysis

Footnotes

Acknowledgments

The authors would like to thank all colleagues and staff at the Korgialeneio-Benakeio Hellenic Red Cross Hospital for their support.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Nikolaos-Nektarios Giannakopoulos

Sofia Tzamtzidou

Dimitra Manou

Konstantinos Nakas

Konstantinos Seretis

Paraskevi Tsiantoula

Konstantinos Roditis

Vasileios Papaioannou

Theofanis Papas

Data Availability Statement

All data utilized in this meta-analysis are sourced from publicly available, previously published studies. Detailed datasets generated during this analysis are available from the corresponding author upon reasonable request.*

Supplemental material

Supplemental material for this article is available online.

References

Fowkes

Rudan

, et al. Comparison of global estimates of prevalence and risk factors for peripheral artery disease in 2000 and 2010: a systematic review and analysis. Lancet 2013; 382(9901): 1329–1340. https://doi.org/10.1016/S0140-6736(13)61249-0, Epub 2013 Aug 1. PMID: 23915883.

Rutherford

Baker

Ernst

, et al. Recommended standards for reports dealing with lower extremity ischemia: revised version. J Vasc Surg 1997; 26(3): 517–538. https://doi.org/10.1016/s0741-5214(97)70045-4, Erratum in: J Vasc Surg 2001 Apr;33(4):805. PMID: 9308598.

Aboyans

Ricco

Bartelink

MEL

, et al. 2017 ESC Guidelines on the Diagnosis and Treatment of Peripheral Arterial Diseases, in collaboration with the European Society for Vascular Surgery (ESVS): document covering atherosclerotic disease of extracranial carotid and vertebral, mesenteric, renal, upper and lower extremity arteriesEndorsed by: the European Stroke Organization (ESO)The Task Force for the Diagnosis and Treatment of Peripheral Arterial Diseases of the European Society of Cardiology (ESC) and of the European Society for Vascular Surgery (ESVS). Eur Heart J 2018; 39(9): 763–816. https://doi.org/10.1093/eurheartj/ehx095. PMID: 28886620.

National Institute for Health and Care Excellence (NICE) . Peripheral arterial disease: diagnosis and management (NICE guideline NG19). 2020.

Boufi

Ejargue

Gaye

, et al. Systematic review and meta-analysis of endovascular versus open repair for common femoral artery atherosclerosis treatment. J Vasc Surg 2021; 73(4): 1445–1455. https://doi.org/10.1016/j.jvs.2020.10.026, Epub 2020 Oct 22. PMID: 33098944.

Jia

Sun

Patel

, et al. Systematic review of endovascular intervention and surgery for common femoral artery atherosclerotic disease. Br J Surg 2019; 106(1): 13–22. https://doi.org/10.1002/bjs.11026. PMID: 30582635.

Linni

Ugurluoglu

Hitzl

, et al. Bioabsorbable stent implantation vs. common femoral artery endarterectomy: early results of a randomized trial. J Endovasc Ther 2014; 21(4): 493–502. https://doi.org/10.1583/14-4699R.1, PMID: 25101576.

Kuo

Chen

Huang

, et al. Outcome of drug-eluting balloon angioplasty versus endarterectomy in common femoral artery occlusive disease. J Vasc Surg 2019; 69(1): 141–147. https://doi.org/10.1016/j.jvs.2018.05.225, Epub 2018 Aug 10. PMID: 30104098.

Wieker

Schönefeld

Osada

, et al. Results of common femoral artery thromboendarterectomy evaluation of a traditional surgical management in the endovascular era. J Vasc Surg 2016; 64(4): 995–1001. https://doi.org/10.1016/j.jvs.2016.04.036, PMID: 27666444.

10.

Ballotta

Gruppo

Mazzalai

, et al. Common femoral artery endarterectomy for occlusive disease: an 8-year single-center prospective study. Surgery 2010; 147(2): 268–274. https://doi.org/10.1016/j.surg.2009.08.004, Epub 2009 Oct 13. PMID: 19828166.

11.

Mehta

Zhou

Paty

, et al. Percutaneous common femoral artery interventions using angioplasty, atherectomy, and stenting. J Vasc Surg 2016; 64(2): 369–379. https://doi.org/10.1016/j.jvs.2016.03.418, PMID: 27763265.

12.

Halpin

Erben

Jayasuriya

, et al. Management of isolated atherosclerotic stenosis of the common femoral artery: a review of the literature. Vasc Endovasc Surg 2017; 51(4): 220–227. https://doi.org/10.1177/1538574417702773, Epub 2017 Apr 4. PMID: 28376704.

13.

Siracuse

Van Orden

Kalish

, et al. Endovascular treatment of the common femoral artery in the Vascular Quality Initiative. J Vasc Surg 2017; 65(4): 1039–1046. https://doi.org/10.1016/j.jvs.2016.10.078, Epub 2016 Dec 29. PMID: 28041804.

14.

Nishibe

Maruno

Iwahori

, et al. The role of common femoral artery endarterectomy in the endovascular era. Ann Vasc Surg 2015; 29(8): 1501–1507. https://doi.org/10.1016/j.avsg.2015.05.005, Epub 2015 Jul 4. PMID: 26148640.

15.

Kang

Patel

Conrad

, et al. Common femoral artery occlusive disease: contemporary results following surgical endarterectomy. J Vasc Surg 2008; 48(4): 872–877. https://doi.org/10.1016/j.jvs.2008.05.025, Epub 2008 Jul 17. PMID: 18639427.

16.

Zhang

Nie

, et al. Bayesian hierarchical methods for meta-analysis combining randomized-controlled and single-arm studies. Stat Methods Med Res 2019; 28(5): 1293–1310. https://doi.org/10.1177/0962280218754928, Epub 2018 Feb 13. PMID: 29433407; PMCID: PMC6719559.

17.

Spiegelhalter

Best

Carlin

, et al. Bayesian measures of model complexity and fit. J Roy Stat Soc B 2002; 64: 583–639. https://doi.org/10.1111/1467-9868.00353

18.

Gouëffic

Della Schiava

Thaveau

, et al. Stenting or surgery for de novo common femoral artery stenosis. JACC Cardiovasc Interv 2017; 10(13): 1344–1354. https://doi.org/10.1016/j.jcin.2017.03.046, PMID: 28683941.

19.

Stavroulakis

Schwindt

Torsello

, et al. Directional atherectomy with antirestenotic therapy vs drug-coated balloon angioplasty alone for common femoral artery atherosclerotic disease. J Endovasc Ther 2018; 25(1): 92–99. https://doi.org/10.1177/1526602817748319, Epub 2017 Dec 18. PMID: 29251204.

20.

Elsherif

Tawfick

Elsharkawi

, et al. Common femoral artery endarterectomy in the age of endovascular therapy. Vascular 2018; 26(6): 581–590. https://doi.org/10.1177/1708538118772682, Epub 2018 May 14. PMID: 29759039.

21.

Guo

Tong

, et al. Directional atherectomy is associated with better long-term efficiency compared with angioplasty for common femoral artery occlusive disease in rutherford 2-4 patients. Ann Vasc Surg 2018; 51: 65–71. https://doi.org/10.1016/j.avsg.2017.12.004, Epub 2018 Mar 1. PMID: 29501593.

22.

Imran

Hyder

Soukas

. Efficacy and safety of adjunctive drug-coated balloon therapy in endovascular treatment of common femoral artery disease. Cardiovasc Revascularization Med 2019; 20(3): 210–214. https://doi.org/10.1016/j.carrev.2018.06.018, Epub 2018 Jun 25. PMID: 29970328.

23.

Kapetanios

Torsello

Tsilimparis

, et al. Directional atherectomy with antirestenotic therapy versus open repair in patients with restenotic disease after surgical revascularization of the common femoral artery. J Cardiovasc Surg 2024; 65(4): 339–345. https://doi.org/10.23736/S0021-9509.24.13126-6, Epub 2024 Jul 15. PMID: 39008277.

24.

Nakama

Takahara

Iwata

, et al. CAULIFLOWER study investigators. 1-Year outcomes of thromboendarterectomy vs endovascular therapy for common femoral artery lesions: CAULIFLOWER study results. JACC Cardiovasc Interv 2022; 15(14): 1453–1463. https://doi.org/10.1016/j.jcin.2022.03.010, Epub 2022 Jun 29. PMID: 35863795.

25.

Wang

Murray

Xiao

, et al. Bayesian hierarchical models incorporating study-level covariates for multivariate meta-analysis of diagnostic tests without a gold standard with application to COVID-19. Stat Med 2023; 42(28): 5085–5099. https://doi.org/10.1002/sim.9902, Epub 2023 Sep 19. PMID: 37724773.

26.

Bonvini

Rastan

Sixt

, et al. Endovascular treatment of common femoral artery disease: medium-term outcomes of 360 consecutive procedures. J Am Coll Cardiol 2011; 58(8): 792–798. https://doi.org/10.1016/j.jacc.2011.01.070, PMID: 21835313.

27.

Soga

Tomoi

Sato

, et al. Clinical outcome after endovascular treatment for isolated common femoral and popliteal artery disease. Cardiovasc Interv Ther 2013; 28(3): 250–257. https://doi.org/10.1007/s12928-013-0164-1, Epub 2013 Jan 30. PMID: 23361951.

28.

Malgor

Ricotta

2nd Bower

, et al. Common femoral artery endarterectomy for lower-extremity ischemia: evaluating the need for additional distal limb revascularization. Ann Vasc Surg 2012; 26(7): 946–956. https://doi.org/10.1016/j.avsg.2012.02.014, PMID: 22944568.

29.

Hong

Suh

, et al. Outcomes of stents covering the deep femoral artery origin. EuroIntervention 2014; 10(5): 632–639. https://doi.org/10.4244/EIJV10I5A109, PMID: 25256202.

30.

Kuma

Tanaka

Ohmine

, et al. Clinical outcome of surgical endarterectomy for common femoral artery occlusive disease. Circ J 2016; 80(4): 964–969. https://doi.org/10.1253/circj.CJ-15-1177, Epub 2016 Feb 19. PMID: 26902450.

31.

Banerjee

Jeon-Slaughter

Armstrong

, et al. Clinical outcomes and cost comparisons of stent and non-stent interventions in infrainguinal peripheral artery disease: insights from the excellence in peripheral artery disease (XLPAD) registry. J Invasive Cardiol 2019; 31(1): 1–9. https://doi.org/10.25270/jic/18.00285, PMID: 30611122; PMCID: PMC6428413.

32.

Piazza

Ricotta

2nd Bower

, et al. Iliac artery stenting combined with open femoral endarterectomy is as effective as open surgical reconstruction for severe iliac and common femoral occlusive disease. J Vasc Surg 2011; 54(2): 402–411. https://doi.org/10.1016/j.jvs.2011.01.027, Epub 2011 Apr 30. PMID: 21531527.

33.

Bonvini

Rastan

Sixt

, et al. Angioplasty and provisional stent treatment of common femoral artery lesions. J Vasc Intervent Radiol 2013; 24(2): 175–183. https://doi.org/10.1016/j.jvir.2012.10.020, Epub 2013 Jan 28. PMID: 23369554.

34.

Azéma

Davaine

Guyomarch

, et al. Endovascular repair of common femoral artery and concomitant arterial lesions. Eur J Vasc Endovasc Surg 2011; 41(6): 787–793. https://doi.org/10.1016/j.ejvs.2011.02.025, Epub 2011 Mar 24. PMID: 21439857.