P1228Impact of history of vascular disease on lipoprotein(a) associations with cardiovascular outcomes: a Mendelian randomisation study in 283,658 UK Biobank participants

Abstract

Abstract Background Mendelian randomisation (MR) studies have been widely used to assess the causal relevance of risk factors for disease and to elucidate the likely effects of novel drug targets. MR studies have confirmed the causal relevance of lipoprotein(a) [Lp(a)] for risk of coronary heart disease (CHD), but the impact of prior vascular disease on these associations remains unclear, and is potentially important when testing the effects of Lp(a) lowering therapies. Purpose To examine the associations of LPA variants, as a proxy for Lp(a) levels, with risk of CHD, ischaemic stroke and peripheral vascular disease (PVD), by history of vascular disease in UK Biobank (UKB). Methods An LPA genotype score was constructed in 283,658 unrelated white British UKB participants, using 2 SNPs (rs3798220 and rs10455872) previously shown to be independently associated with, and explain ∼40% of the variance in, Lp(a) levels. The LPA score was tested for association with hospitalisation for CHD (n=23,165), presumed ischaemic stroke (n=2,579) and PVD (n=3,780) using Cox PH models. Estimated hazard ratios (per LPA variant) for cardiovascular outcomes are reported both overall and by history of vascular disease (CHD, ischaemic stroke or PVD) at baseline. Results As observed in previous studies, LPA was associated with a higher risk of CHD (Hazard ratio [HR] 1.27 per LPA variant; 95% CI: 1.24–1.31). However, the effect was substantially attenuated in participants with versus without prior vascular disease (p-interaction = 3x10–5; Figure). The effect of the LPA score on PVD was comparable to that for CHD (p-heterogeneity = 0.25), whilst the effect on ischaemic stroke was significantly weaker than that for CHD (p-heterogeneity = 6x10–3). As for CHD, the effects of the LPA score on both ischaemic stroke and PVD were weaker in those with versus those without prior vascular disease. Furthermore, the heterogeneity between the effects on cardiovascular outcomes was more extreme when stratified by history of CHD. Conclusions This MR study shows that the effects of LPA on risk of CHD, ischaemic stroke and PVD are weaker in individuals with a history of vascular disease, perhaps reflecting the effects of medications used in this setting. Therefore, MR studies in the general population may overestimate the anticipated effects of Lp(a)-lowering therapies in randomised trials, which are typically conducted in individuals with prior vascular disease. Acknowledgement/Funding British Heart Foundation