In a presentation at the 2015 Vascular Interventional Advances meeting (VIVA; 2–5 November, Las Vegas, USA), Krishna Rocha-Singh, St John’s Hospital, Springfield, USA, gave an insight into the VIVA Superficial Femoral Artery Nitinol Stent Patient Level Meta-analysis, suggesting that patients with higher ankle-brachial index and shorter lesions have the most positive outcomes. He emphasised the importance of “robust and independently adjudicated” data and more standardised endpoints.
In 2007, Rocha-Singh and colleagues worked with the industry to provide and promote a performance goal as an assessor and design trial comparator for the evaluation of femoropopliteal bare nitinol stents in patients with symptomatic peripheral arterial disease. Rocha-Singh said that this was endorsed by the Food and Drug Administration (FDA) and soon became the standard comparator used by all subsequent nitinol superficial femoral artery stents trials.
The investigators wanted their performance goals to: foster collaboration with physicians, the FDA and industry; provide a “least burdensome pathway” to device approval and discourage off-label promotion; promote uniform reporting standards and endpoint assessment; and encourage industry to initiate larger, more robust registries of nitinol stenting in claudicants. All of these objectives, Rocha-Singh said, have been achieved.
The performance goals and the resulting relationship with industry provided the team with six adjudicated patient level datasets (COMPLETE SE, DURABILITY II, RESILIENT, STROLL, VIBRANT and ZILVER PTX BMS arm) of 999 trial patients. This allowed them to make comparisons across the different sets of results and analyse any correlations.
The methodology of the meta-analysis, based on the previously established performance goals, used 2.0 peak systolic velocity ratio (PSVR) primary patency as the cut-off value for five studies and 2.5 PSVR for the sixth. Intervention was considered to equal patency failure in five of the six trials. The team also looked at target lesion revascularisation, defined as clinically-driven in five studies and uncertain in one.
Patient-level data were provided for each trial—the importance of which Rocha-Singh “cannot overemphasise as they allow for very robust analysis”—and endpoints were treated as binary outcomes. Logistic regression models were used for association of baseline factors with 12-month endpoints.
The mean lesion length was 83.8mm. The analyses allowed Rocha-Singh and colleagues to show that lesion length was associated with primary patency and clinically-driven target lesion revascularisation at 12 months. After 12 months in lesion tertile 1 (lesion length <53mm), 12-month patency was 79.9% and target lesion revascularisation rate was 8.3%. In lesion tertile 2 (53–99.5mm) patency was 71.8% and revascularisation was 11%, while in the third tertile (≥99.5mm) the patency was 57.7% and target lesion revascularisation was as high as 20.1% (p<0.001). This correlation remained true when the results were put into quartiles, with the lowest quartile (<43.6mm) showing 12-month patency of 82% and revascularisation of 7.8% and the highest quartile (112.5–360mm) showing patency of 55.1% and revascularisation of 20.2% (p<0.001).
Also important was the analysis of the impact of pre-procedural ankle-brachial pressure index (ABI) scores on the 12-month outcomes of a procedure. In tertile 1 (ABI <0.62), 12-month patency was 60.4% and target lesion revascularisation was 17.8%. In tertile 2 (ABI 0.62–0.75), 12-month patency was 72.3% and revascularisation 11.5%, while in tertile 3 (ABI >0.75), patency was 76% and revascularisation 10.5% at 12 months (p
Rocha-Singh did note that “with any trial there are limitations and caveats”. In this case, he told the audience that there was selection bias as the team only studied FDA-approved bare metal stents. Furthermore, the selection of cut-offs for lesion length and ankle brachial index post-hoc could be seen as a limitation, though Rocha-Singh said that sensitivity analyses for tertiles/quartiles helped assess robustness. Full pre-specification of the published data were not possible as the trials were already well-known and published, and although there was a variation of definition across studies the large sample size helps offset variability concerns.
Rocha-Singh and colleagues believe that “Access to robust, de-identified, independently adjudicated, patient-level data from superficial femoral artery nitinol trials provides important insights into the ‘ideal patient’ cohort who may benefit from nitinol stenting through 12-month follow-up.” In addition, the team believe that additional analyses correlating stent-related surrogates (such as absolute PVS) with clinical outcomes (such as clinically-driven target lesion revascularisation) “may allow for consideration of more clinically relevant primary endpoints in superficial femoral artery stents.”