Abstract / Summary
Background: Lipoprotein (a) [Lp(a)] is an independent atherosclerotic cardiovascular disease risk factor, with levels largely determined by variation in the LPA gene. Despite interests in therapeutic Lp(a) lowering, the phenotypic effects of very low Lp(a) are incompletely characterized, with some studies suggesting increased type 2 diabetes risk. Objectives: Estimate associations between very low Lp(a), inferred by LPA loss of function variant rs41272114, with type 2 diabetes and glycemic biomarkers. Methods: A multi-population study of European, Hispanic/Latino, and South Asian populations from five cohort studies was assembled. Effects of the null allele rs41272114 (CC, wild-type; T, null allele) on type 2 diabetes and glycemic biomarkers (fasting glucose, fasting insulin, glycosylated hemoglobin (HbA1c), triglyceride-to-high-density lipoprotein ratio, and HOMA-IR) were estimated using multivariable-adjusted and cohort-specific linear or logistic regression models that were combined using inverse variance meta-analysis. Results: Of n=519,892 participants, the T null allele was common (n=49,841 CT and n=2,039 TT) and n=38,320 participants had type 2 diabetes (prevalence range: 4.7, 29.5%). Participants with the TT genotype had a median Lp(a) ranging 1.5-1.9 nmol/L compared to median Lp(a) ranging 15.8-30.0 nmol/L in participants with the CC genotype. We did not observe increased odds of type 2 diabetes by rs412722114 genotype (OR CTvs.CC = 0.97, 95% CI: 0.94, 1.00; OR TTvs.CC = 1.01, 95% CI: 0.89, 1.15). Mean levels of glycemic biomarkers also did not differ by rs41272114 genotype. Conclusions: In a large multi-population study with high burdens of type 2 diabetes, life-long very low Lp(a) did not increase type 2 diabetes risk.