Mortality in randomised controlled trials using paclitaxel-coated devices for femoropopliteal interventional procedures: an updated patient-level meta-analysis.

BACKGROUND: Numerous randomised clinical trials and real-world studies have supported the safety of paclitaxel-coated devices for the treatment of femoropopliteal occlusive disease. However, a 2018 summary-level meta-analysis suggested an increased mortality risk for paclitaxel-coated devices compared with uncoated control devices. This study presents an updated analysis of deaths using the most complete and current data available from pivotal trials of paclitaxel-coated versus control devices. METHODS: Ten trials comparing paclitaxel-coated versus control devices were included in a patient-level pooled analysis. Cox regression models were used to evaluate the effect of paclitaxel exposure on risk of death in both intention-to-treat (ITT; primary analysis) and three as-treated analysis sets accounting for treatment group crossover at the index procedure and over time. The effect of paclitaxel dose and baseline covariates were also evaluated. FINDINGS: A total of 2666 participants were included with a median follow-up of 4·9 years. No significant increase in deaths was observed for patients treated with paclitaxel-coated devices. This was true in the ITT analysis (hazard ratio [HR] 1·14, 95% CI 0·93-1·40), the as-treated analysis (HR 1·13, 95% CI 0·92-1·39), and in two crossover analyses: 1·07 (0·87-1·31) when late crossovers were censored and 1·04 (0·84-1·28) when crossovers were analysed from the date of paclitaxel exposure. There was no significant effect of paclitaxel dose on mortality risk. INTERPRETATION: This meta-analysis found no association between paclitaxel-coated device exposure and risk of death, providing reassurance to patients, physicians, and regulators on the safety of paclitaxel-coated devices. FUNDING: Becton Dickinson, Boston Scientific, Cook, Medtronic, Philips, Surmodics, and TriReme Medical.

Patient-level meta-analysis of 999 claudicants undergoing primary femoropopliteal nitinol stent implantation.

CONDENSED ABSTRACT: The factors that impact the clinical effectiveness of bare nitinol stents in claudicants with symptomatic femoropopliteal atherosclerosis are incompletely known. The authors analyzed variables that may influence stent durability and provide a benchmark for their effectiveness. Data analyzed from six studies (999 patients) included baseline noninvasive hemodynamic tests, angiographic characteristics, ultrasound defined stent patency and target lesion revascularization through 12-months. Baseline ankle-brachial index and lesion length predicted stent patency and target lesion revascularization and when combined interacted significantly to better predict outcomes. This meta-analysis provides an important comparator against which emerging therapies that treat claudicants with femoropopliteal atherosclerosis can be assessed. SUBJECT CODE: Peripheral Artery Disease BACKGROUND: The performance of bare metal nitinol stents in patients with symptomatic femoropopliteal peripheral artery disease (PAD) is not well defined. METHODS: Patient-level data from six large prospective trials sponsored by medical device manufacturers was abstracted and analyzed to identify a cohort of patients with claudication and femoropopliteal artery occlusive disease. Twelve-month binary patency and target lesion revascularization (TLR) rates were primary outcomes. Stent patency was assessed by duplex ultrasonography (DUS) and TLR was a clinically driven intervention. To characterize the effects of patient characteristics on the outcomes, meta-regression was performed via mixed effects logistic regression models with patient-level covariates. RESULTS: About 999 patients were analyzed; the mean ABI was 0.68 ± 0.18, the mean lesion length was 84 ± 53 mm, the mean lesion stenosis was 78%, and nearly two thirds of patients had mild to severe calcification. The mean Rutherford clinical category was 2.7 ± 0.6 and ranged from 2.6 to 2.8 in all studies. The 12-month patency across all studies was 69.8% and TLR rates ranged from 9.2% to 19.7%. Multivariable analysis demonstrated that baseline ABI and baseline target lesion length predicted both primary patency and TLR. Further, these two variables interacted significantly to better predict TLR outcomes when used in combination. CONCLUSION: The 12-month clinical effectiveness of bare nitinol stents to treat patients with symptomatic femoropopliteal PAD is acceptable and is impacted by clinical and lesion-specific characteristics. These data provide an important and useful benchmark to compare the clinical benefit of emerging endovascular PAD therapies. © 2017 Wiley Periodicals, Inc.