Lipoprotein (a), coronary artery disease risk, and MACE in individuals without standard modifiable risk factors in the UK Biobank

Abstract Background A significant proportion of patients with coronary artery disease (CAD) lack standard modifiable risk factors (SMuRFs; smoking, hypercholesterolaemia, diabetes and hypertension), for example with recent analyses demonstrating that around 25% of patients with ST-elevation myocardial infarction are SMuRF-less(1). With continued focus on primary prevention in cardiovascular disease, the SMuRF-less subset is likely to increase over time. A body of evidence has implicated lipoprotein (a) in CAD, but no such studies have investigated this association in a purely SMuRF-less population. Purpose To investigate the role of lipoprotein (a) in predicting CAD risk and major adverse cardiovascular events (MACE) in a large, SMuRF-less cohort. 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. SMuRFs were defined as 1) smoking (current or historical), 2) hypercholesterolaemia – lipid lowering medication on enrolment or total cholesterol >5.5mmol/L or LDL-C >3.5mmol/L, 3) previous diagnosis of diabetes mellitus, or HbA1c >48mmol/mol, 4) previous diagnosis of hypertension. MACE was defined as a composite of coronary artery disease – previous myocardial infarction, percutaneous coronary intervention or coronary artery bypass graft surgery –, stroke, or all-cause mortality. Results Of the total UK Biobank cohort, 46,139 participants were identified as SMuRF-less together with a valid lipoprotein (a) level. Median follow-up duration was 12yrs. In univariate analyses (Table 1), there was no significant difference in lipoprotein (a) concentrations between individuals with and without MACE (16.9 vs. 17.1nmol/L respectively, p = 0.913). The overall event rate was low, with just 3% MACE (1381 events) overall. Older age, male gender, apolipoprotein A, apolipoprotein B, HbA1c level, HDL cholesterol level and body mass index were all associated with MACE. In regard to all-cause mortality alone, there was no association with lipoprotein (a) level, although the other aforementioned parameters were all significantly associated with this outcome. Similarly, lipoprotein (a) was not a significant predictor of stroke. However, there was a large difference in lipoprotein (a) levels between patients with and without CAD (27.2 vs. 17.1nmol/L respectively, p = 0.004). In Kaplan-Meier analysis, lipoprotein (a) level above the 77th percentile (60.3nmol/L) was significantly associated with CAD (p=0.032). In Cox regression models, despite the low event rate, lipoprotein (a) was a significant predictor of CAD (HR 1.00-1.01, p=0.01), and importantly the only potentially modifiable parameter (N = 39,433, 111 events, Table 2). Conclusion Lipoprotein (a) independently predicts CAD in those without SMuRFs, highlighting the importance of checking this parameter in such patients, particularly given the novel antisense technologies in development which may lower lipoprotein (a)(2).Univariate analysesCox regression for CAD