Clinical and vessel characteristics associated with hard outcomes after PCI and their combined prognostic implications

Abstract

Abstract Background While clinical outcomes after percutaneous coronary intervention (PCI) have improved in recent decades, cardiac death or myocardial infarction (MI) still occurs in patients undergoing contemporary PCI. An integrative risk model for post-PCI hard clinical endpoints incorporating clinical risk factors, angiographic and physiological vessel characteristics has not been established. Purpose To identify adverse clinical and vessel characteristics related to hard outcomes after PCI and to investigate their individual and combined prognostic implications. Methods From an individual patient data meta-analysis of 17 cohorts of patients who underwent post-PCI fractional flow reserve (FFR) measurement after drug eluting stent (DES) implantation, 2,081 patients with available 10 clinical characteristics (age, sex, clinical diagnosis at index procedure, diabetes, hypertension, dyslipidemia, smoking history, chronic kidney disease [CKD], prior myocardial infarction, left ventricular ejection fraction [LVEF]) and 8 vessel characteristics (target vessel, reference vessel diameter, % diameter stenosis, minimum lumen diameter [MLD], post-PCI % diameter stenosis, post-PCI MLD, total stent length in the target vessel, and post-PCI FFR) were analyzed. The least absolute shrinkage and selection operator (LASSO) model was applied for feature selection. The primary outcome was cardiac death or target vessel-MI at 2-years. Results The mean age of patients was 64.2±10.2 years, and the mean angiographic % diameter stenosis was 63.9±14.3%. Among 10 clinical and 8 vessel features, 4 adverse clinical characteristics (ACCs; age ≥65 years, diabetes, chronic kidney disease, and left ventricular ejection fraction <50%) and 2 adverse vessel characteristics (AVCs; post-PCI FFR ≤0.80 and total stent length ≥54 mm) were selected by the LASSO model. All ACCs and AVCs were independent predictors for the primary outcome (HR 3.33 [95% CI 1.34 – 8.25], P =0.009 for age ≥65 years; HR 2.62 [1.22 – 5.65], P =0.014 for diabetes; HR 5.95 [2.52 – 14.1], P <0.001 for CKD; HR 2.91 [1.17 – 7.20], P =0.021 for LVEF <50%; HR 3.79 [1.73 – 8.28], P <0.001 for post-PCI FFR ≤0.80; HR 2.42 [1.08 – 5.39], P =0.031 for total stent length ≥54 mm). The number of AVCs had additive predictability for the primary endpoint to that of ACCs (AUC 0.72 vs 0.78, P =0.03), and vice versa (AUC 0.68 vs 0.78, P =0.03). The cumulative event rate increased in the order of none, either, and both of ACCs ≥2 and AVCs ≥1 (0.3%, 2.4%, and 5.3%, P-for-trend <0.01) (Figure). Conclusions In patients undergoing DES implantation, adverse clinical and vessel characteristics were associated with the risk of cardiac death or target vessel-MI. Since these characteristics showed independent and additive prognostic value, their integrative assessment can optimize post-PCI risk stratification.