P6197The effect of lipoprotein (a) on primary prevention of cardiovascular disease and the interaction with conventional lipid markers: a sex-based sensitivity analysis from a 10-year cohort study

Abstract

Abstract Background/Introduction In the context of novel risk factors in cardiovascular disease (CVD) spectrum, lipoprotein (a) (Lp(a)) is a highly discussed biomarker with promising evidence. Purpose The association between Lp(a) and 10-year first fatal/non fatal CVD event in free of CVD males and females was evaluated. Methods A longitudinal prospective study was conducted during 2001–2012, studying 1,514 males and 1,528 females (aged >18 years old). Follow-up assessment of CVD (2011–2012) was achieved in n=2,020 participants (n=317 cases). Of them, baseline Lp(a) was measured in n=1,890 participants. The recommended threshold of 50mg/dL was used to define abnormal Lp(a) status (≥50mg/dL). Effect-size of Lp(a) on CVD was evaluated through Cox-regression analysis while its discrimination ability through C-statistics. Results Ten-year CVD event rate was 14% and 24% in participants with Lp(a)<50 mg/dL and Lp(a)≥50 mg/dL, respectively (p=0.05). In multivariate analysis those with Lp(a)≥50 mg/dL had two times higher risk to develop CVD compared with participants with normal Lp(a) (Hazard Ratio (HR)=2.18, 95% Confidence Interval (95% CI) 1.11, 4.28, p=0.04). Sex-based stratified analysis revealed that the independent Lp(a)-effect on CVD was retained only in males (HR=2.00, 95% CI 1.19, 2.56, p=0.05); while in females significance was lost when adjusting for low and high density lipoprotein (LDL-C, HDL-C), triglycerides and statins use (p for sex interaction=0.01). Sensitivity analyses revealed that Lp(a) significantly increased CVD risk only in case of abnormal HDL-C, apolipoprotein A1 and triglycerides; interestingly, the interaction between these lipid markers, sex and Lp(a) was significant (p for interaction=0.001) implying that this observation could be sex-mediated. C-indexes and correct classification rates of a standard model with three different levels of adjustment (i.e. Lp(a) or conventional lipid markers or combined lipid markers) were evaluated per sex. In females, the highest total correct classification rate was higher in model adjusted for conventional lipid markers (89.6%) with the rate corresponding to CVD cases being more than twice as high in Lp(a)-adjusted model (19.6% vs. 8.5%) and lower than in fully adjusted model (15.7%). A similar ranking was observed in case of C-indexes (0.831 vs. 0.820 vs. 0.829). Males presented the best total correct classification rate in fully adjusted model (96.5%). Case-related correct classification rate was about 3 times higher in Lp(a)-adjusted model compared with the respective rate in females (24.7% vs. 8.5%). C-index after Lp(a) adjustment in the model with conventional lipid markers increased by 0.01 (i.e. 0.772 vs. 0.784). Conclusion While ever increasing efforts have sought to elucidate Lp(a) as a therapeutic target or risk-prediction biomarker in CVD prevention clinical recommendations remain to be guided with appropriate conclusive evidence, mostly from a sex-centered standpoint. Acknowledgement/Funding The ATTICA study is supported by research grants from the Hellenic Cardiology Society [HCS2002] and the Hellenic Atherosclerosis Society [HAS2003].