Repeat revascularisation following PCI in the UK Biobank prospective cohort study: incidence, independent predictors and prognostic impact

Abstract

Abstract Introduction Data from the SYNTAX and EXCEL trials have previously demonstrated increased morbidity and mortality following repeat revascularisation (RR). However real-world RR rates following index PCI are poorly defined, particularly in the 'late' context of restenosis of bystander disease and de novo lesions. Purpose To identify independent predictors of RR and evaluate its prognostic impact in an unselected cohort of patients following first PCI. Methods The UK Biobank is a large prospective multicentre study which recruited over 500,000 participants in the UK aged 37–73 years at entry between 2006–2010. Data are derived from an array of physical measurements, self-reported measures and biological samples, together with longitudinal linkage to hospital inpatient records and death registries. Procedural codes (OPCS-4) were used to identify patients who had undergone first PCI from 2006 onwards. RR was defined as PCI or CABG occurring at least 9 months after the index PCI, in order to preclude instances of staged revascularisation and in-stent restenosis. Data were censored at January 2021 or date of death as appropriate. Non-parametric tests are used throughout. Results A total of 12,853 participants underwent a first PCI during the study period, with 1,394 (10.8%) requiring RR over a median follow-up of 6.5 years. Of these, 1188 underwent PCI whilst 206 were elected to CABG. The average age was 64 years in the RR cohort and 1117 (80%) were male. Univariate analyses confirmed a number of established associations with RR, including for example diabetes mellitus and hypertension. Median lipoprotein(a) concentrations were 27.6 vs. 30.3nmol/L in the RR group, p=0.066. Cox regression analyses incorporating 21 biometric and clinical parameters demonstrated that lipoprotein(a) was an independent predictor of time to RR, independent of other lipids (Table 1). The strength of this association persisted when the number of covariates was reduced to include only other lipids and classical cardiovascular risk parameters. In Kaplan-Meier analyses, RR was associated with lower all-cause mortality (Figure 1). Further multivariate time-to-event analyses confirmed this association (HR 0.59, 95% CI 0.47–0.75, p<0.001). In separate models examining PCI and CABG as distinct modalities of RR, both were independently associated with prognostic benefit (HR 0.61, p<0.001 and HR 0.48, p<0.05 respectively). Conclusions Lipoprotein(a) is identified as a major independent driver of RR. In contrast to recent trial data focussing on left main stem and multivessel disease, unselected patients undergoing RR 9 or more months following index PCI attain prognostic benefit as assessed in multivariate analyses. The findings support an invasive approach, whether with PCI or CABG, in progressive coronary artery disease following initial PCI. Funding Acknowledgement Type of funding sources: Other. Main funding source(s): 1. The Herefordshire Heart Fund, UK. 2. Postgraduate Medical Bursary, Wye Valley NHS Trust, UK. Table 1Figure 1