Aim: In current device trials, the values of angiographic late lumen loss (LLL) have become extremely low and the relationship between LLL and clinical endpoints has not been recently re-evaluated. The impact of LLL on target lesion revascularization (TLR) in a patient- and study-level analysis of contemporary coronary devices was investigated.
Methods and results: We performed a patient-level meta-analysis of seven randomized controlled trials including 2426 patients treated with first- and second-generation drug-eluting stents (DES) and a study-level meta-analysis of 40 studies including 19 199 patients treated with CE-marked DES. In the patient-level analysis, the probability regression curve showed an exponential relationship between in-stent LLL and 2-year incidence of TLR. The optimal cut-off value of LLL based on Youden's index for 2-year TLR event was 0.50 mm. In the Cox proportional hazard model, LLL >0.50 mm was independently associated with an increased incidence of TLR up to 4 years after angiographic follow-up {adjusted hazard ratio (HR) 6.62 [95% confidence interval (95% CI) 4.67-9.39], P < 0.001}. In the meta-regression analysis of the DES studies, pooled mean value of LLL was as low as 0.23 mm (95% CI 0.20-0.26), and there was a moderate correlation between the 1- and 5-year incidence of TLR and the percentage of the lesions with LLL >0.50 mm (R2 = 0.44, P < 0.001 at 1 year, R2 = 0.40, P < 0.001 at 5 years).
Conclusion: An angiographic LLL ≤0.50 mm was not predictive of the incidence of TLR whereas a LLL >0.50 mm was. Low LLL in contemporary device trials may not be a sufficiently discriminating parameter for the comparative evaluation of devices.