2.1 Study Design

We conducted a prospective, protocol-driven, longitudinal, observational study of patients who underwent angioplasty with the IN.PACT Admiral paclitaxel DCB (Medtronic, Santa Rosa, California, USA) in France in 2018. Data were captured in the French National Health Data System (Système National des Données de Santé [SNDS]), which is the French nationwide claims database containing healthcare reimbursement data from patients affiliated with one of the compulsory health insurance providers representing more than 99% of French residents, including more than 10 years of follow-up. The use of deidentified SNDS data for this analysis was approved by the Ethics and Scientific Committee for Research, Studies and Evaluations in the Field of Health and Commission Nationale de l'Informatique et des Libertés.

SNDS contains data from three main sources: Système National d'Information Inter-Régimes de l'Assurance Maladie, which consists of all outpatient reimbursed health expenditures (Données de Consommation Inter-Régime [DCIR]), medical and administrative data from hospital stays (Programme de Médicalisation des Systèmes d'Information [PMSI]), and cause of death. Data from DCIR and PMSI have been linked for each patient to allow for follow-up across different settings of care, including the following: outpatient practices and hospital admissions related to medicine, surgery, and obstetrics; follow-up and rehabilitation care; and hospitalization at home. Each patient's healthcare use is tracked from birth or immigration in France to death or emigration, even if the person is not working, changes occupation, or retires, and irrespective of socioeconomic status.

SNDS contains information on beneficiaries' age, sex, region of residence, death date, complementary universal health coverage status, long-term diseases, and all outpatient healthcare consumption including all reimbursed prescription drugs identified by their Anatomical Therapeutic Chemical code, the date of delivery, quantity, and brand name. Medical procedures performed on an outpatient basis or in a healthcare institution are identified by the classification commune des actes médicaux (common classification of medical procedures), laboratory procedures are identified by the nomenclature des actes de biologie médicale (clinical pathology test nomenclature), and paramedical or medical visits are identified by a prestation code specific to the DCIR (PRS_NAT_REF). Codes of interest are detailed in Supporting Information S1: Table 1.

Through the PMSI, the SNDS also includes medical summaries of all hospitalizations from all private or public hospitals, including the date of stay, medical procedures, innovative and costly medications and medical devices covered by an add-on list (“Liste en sus”), the primary diagnosis (main reason for admission), related diagnoses (specifies the disease context of the primary diagnosis), and diagnoses related to other comorbidities, all encoded according to the International Classification of Disease, 10th edition (ICD-10). For certain medical devices (notably those covered by the “Liste en sus”), the type and brand of devices used are also tracked and reported.

For this analysis, we utilized SNDS data from the DCIR and the PMSI hospitalization data; specifically, data from hospitalizations in the medicine, surgery, and obstetrics sectors.

2.2 Study Population

This study comprised all adult patients with at least one reimbursement in SNDS for an IN.PACT Admiral DCB between January 1, 2018, and December 31, 2018. The study index date was the first hospital entry date for an angioplasty performed during the inclusion period. Data extraction began on January 1, 2015 and ended on December 31, 2021 to allow for (1) a 3-year historical period before the index procedure to identify patient comorbidities and any history of revascularization, and (2) a 3-year follow-up period from the date of angioplasty.

Patients included in the analysis had treatment for lower limb PAD. Patients who underwent angioplasty within 30 days before or 30 days after the index procedure were excluded from the analysis, as these were considered likely to be scheduled interventions (i.e., bilateral procedures) rather than interventions for restenosis. Because SNDS does not capture details of specific lesions, it was not possible to accurately identify the lesion characteristics.

2.3 Study Outcomes

The primary outcomes of this analysis were all-cause mortality, major amputation of a lower limb (including target and nontarget limbs and vessels), and rate of reintervention (any infrainguinal reintervention or new intervention of the target lesions, nontarget lesions, target limbs, or contralateral limbs performed through endovascular or surgical procedures) within 1 year after the index procedure.

Outcomes after the index procedure are described for the total study population and according to three clinical subgroups of interest: CLTI versus claudication, diabetes versus no diabetes, and history of revascularization versus no history of revascularization. Since there is no diagnostic code for CLTI, patients were defined as having CLTI if at least one diagnostic code associated with the presence of trophic disorders (ICD-10 code L97–lower limb ulcer, not elsewhere classified) was entered during the index stay, or if the index stay lasted more than two nights. By default, patients who did not meet any of the above criteria were considered to have claudication. Presence of diabetes was defined according to the Caisse Nationale de l'Assurance Maladie des Travailleurs Salariés algorithm, which identifies diabetes via hospitalizations for diabetes or for one of its complications, diabetes registered as a long-term disease, or at least three prescriptions of oral or injectable antidiabetic drugs (on three different dates) in the year preceding the index procedure [15]. The history of revascularization on the lower limb was defined as a lower limb intervention (regardless of laterality of the lesion) recorded in the database between 30 days and 2 years before the index procedure.

The SNDS database tracks any new procedure performed after the index procedure; however, the reason for and location of the new procedure (i.e., new lesion, reintervention, or bilateral intervention) are not captured. Similarly, deployment of stents during the index procedure is captured in the database with a procedure code for angioplasty with stenting, but the type and location of the stent may not be identified. Patients with one or more procedures at the infrainguinal level and one or more stents, without procedures identified at the iliac level, were considered to have had stent use during the index procedure, as well as patients for whom multi-leveled procedures and multiple stents were identified, without evidence of the location of stenting (Supporting Information S1: Figure 1).

2.4 Statistical Analysis

Event rates for primary outcomes (all-cause mortality, major amputation rate, and revascularization rate at 1 year) were estimated for all patients who had angioplasty with the IN.PACT Admiral DCB in 2018 using Kaplan−Meier analyses. Event rates at 1 year and 3 years after the index procedure were also described within the predefined subgroups (CLTI vs. claudication, diabetes vs. no diabetes, and history vs. no history of revascularization) and compared using the log-rank test. Kaplan−Meier curves describing freedom from all-cause mortality, major amputation, and revascularization over the 3-year follow-up were also plotted with their confidence intervals (CIs). Patients were censored on the date of death, the date of last reimbursement for care (in case of no benefit for a period of at least 10 months), or 3 years after inclusion, whichever occurred first. Statistical analyses were carried out by IQVIA (La Défense Cedex France) using SAS version 9.3 (SAS Institute, North Carolina, USA).

4.1 Limb Salvage

Our 3-year major amputation rate of 4.6% in routine clinical practice is low given the unselected population, high proportion of patients with CLTI, and the fact that this figure may include amputation of the nontarget limb. Three-year target limb amputation rates in the IN.PACT studies were very low (IN.PACT Global, 1.0%; IN.PACT SFA, 0%; and IN.PACT SFA Japan, 0%) [18-20]. Together, these findings demonstrate successful limb salvage with this DCB.

4.2 Mortality

A recently updated meta-analysis of mortality rates following the use of paclitaxel-coated interventional devices for the treatment of lower extremity PAD supports the safe use of paclitaxel DCBs in randomized trial settings [10]. The current large, national registry further supports the use of paclitaxel-coated devices in real-world settings. The 1-year (7.5%) mortality rates in our study compare favorably with rates in other real-world evaluations of patients with PAD, including the COPART Registry in France (21.4%) [24] and a Canadian analysis (8.2% annual mortality) [25], as well as in real-world studies of paclitaxel-coated device use in Sweden (10.2%) [26] and the United States (8.6%) [27]. Three-year mortality rates vary widely across real-world PAD studies (9.7%−47.0%) and the rate of 19.7% in the current study falls in the middle of that range, consistent with half of the previous studies (Figure 4) [18, 28-34].

4.3 High-Risk Subgroups

CLTI is the most severe manifestation of PAD and presents with ischemic pain at rest, tissue loss, or both. In post hoc analysis of the 5-year follow-up in the IN.PACT Global study, participants with CLTI had significantly higher cumulative incidence of all-cause mortality (37.4% vs. 17.4%; p < 0.001) and major target limb amputation (6.8% vs. 1.1%; p < 0.001) compared to participants with intermittent claudication [35]. Consistent with these findings, we observed significantly higher rates of all-cause mortality (30.6% vs. 13.9%, p < 0.001), major amputation (9.5% vs. 2.3%, p < 0.001), and total reintervention (50.8% vs. 39.8%, p < 0.001) in patients with CLTI compared to patients with intermittent claudication at 3 years. Our findings for patients with CLTI are also consistent with rates of all-cause mortality (33.0% for surgery and 37.6% for endovascular therapy) and above-the-knee amputation (10.4% and 14.9%, respectively) at median 2.7 years follow-up in the BEST-CLI randomized trial of patients with CLTI and infrainguinal PAD with a single segment of the great saphenous vein that could be used for surgery [36]. However, higher rates of all-cause mortality (53% for vein bypass and 45% for endovascular treatment) and major amputation (20% and 18%, respectively) at 2-year follow-up have been reported in the BASIL-2 randomized, multicenter trial in patients with CLTI [37].

The known increased mortality risk for PAD patients with comorbid diabetes mellitus [38] was not observed in the 3-year analysis of IN.PACT Global [18] but was evident in the post hoc analysis of 5-year follow-up, in which only all-cause mortality was higher in patients with diabetes mellitus compared to those without diabetes (23.8% vs. 16.6%; p < 0.001) [35]. In our real-world study, not only was the 3-year rate of all-cause mortality (25.5% vs. 16.1%, p < 0.001) significantly higher among patients with diabetes than those without diabetes, but the same was true for major amputation (7.9% vs. 2.5%, p < 0.001), and total reintervention (48.1% vs. 40.6%, p < 0.001).

Patients with PAD who have undergone a prior revascularization procedure carry a particularly high risk for additional adverse limb events. In the IN.PACT global study, a history of prior revascularization was present in 52% of patients and was a significant multivariable predictor of major adverse events [3]. Previous ipsilateral revascularization was also identified as a predictor of clinically driven target lesion revascularization among women in IN.PACT SFA randomized trial [39]. Corroborating these results in a real-world French population, our study observed that a history of revascularization was associated with poorer outcomes, including all-cause mortality (22.3% vs. 18.3%, p = 0.005), major amputation (7.3% vs. 3.2%, p < 0.001), and total reintervention (51.7% vs. 38.9%, p < 0.001), compared to those without prior revascularization.

4.4 Study Limitations

This was a study of real-world experience with a DCB in femoropopliteal artery disease using data from the French national healthcare claims database. It was a single-arm retrospective analysis with no comparator group and no adjudication of outcomes. The SNDS database does not contain detailed lesion characteristics or reintervention details, such as vessel location or side; thus, major amputations and interventions could be in any vessel or the contralateral limb, resulting in overestimations compared to other prospective registries. These different definitions of major amputation and reintervention used in our study preclude direct comparison of these rates with other studies of DCBs. The lack of a diagnostic code for CLTI and the necessary use of a surrogate measure was also a limitation. Nonetheless, this longitudinal, prospective national study with near-total coverage in France is a useful tool for confirming or identifying potential differences in outcomes between clinical trials and routine practice.