Abstract
OBJECTIVE: Risk estimation tools have been developed to predict coronary heart disease (CHD) in type 2 diabetes (T2D). To evaluate augmentation following the addition of lipoprotein(a) [Lp(a)] to risk calculation, we performed a pilot study.
METHODS: A total of 90 successive T2D patients were included. Details of clinical and biochemical features were obtained. Lp(a) was determined using ELISA. CHD risk estimation was performed using Framingham, QRISK-3, SCORE-2D, INTERHEART, and European Atherosclerosis Society (EAS) algorithms with and without Lp(a). Descriptive statistics are reported.
RESULTS: Mean age of patients was 55.0 ± 8 years, BP systolic/diastolic 133.7 ± 12/95.0 ± 9 mm Hg, body mass index (BMI) 26.0 ± 1.9 kg/m2, waist-hip ratio 0.96 ± 0.08, fasting glucose 198.0 ± 38 mg/dL, HbA1c 9.3 ± 1.3%, total cholesterol 197.0 ± 26 mg/dL, LDL cholesterol 114.2 ± 25 mg/dL, non-HDL cholesterol 153.8 ± 27 mg/dL, and triglycerides 197.8 ± 44 mg/dL. Lp(a) was mean 23.1 ± 9.7 mg/dL and median 22.0 (25-75 IQR 15.9-29.5) mg/dL. Mean risk scores were Framingham 11.2 ± 8.7, QRISK-3 28.6 ± 15.3, INTERHEART 21.0 ± 6.0, SCORE-2D 14.9 ± 8.3, and EAS 29.2 ± 15.2. Patients with raised Lp(a) >30 mg/dL had higher levels of total, LDL, and non-HDL cholesterol and triglycerides (p < 0.01). Spearman's correlation of Lp(a) with risk scores was Framingham 0.127, QRISK-3 0.174, INTERHEART 0.137, SCORE-2D 0.050, and EAS 0.320, while EAS-Lp(a) was 0.397. In different risk algorithms, high risk for CHD were: Framingham 14.4%, QRISK-3 64.4%, INTERHEART 45.6%, SCORE-2D 30.0%, EAS 71.1%, and EAS with Lp(a) 74.4%. Area under the curve (AUC) for Lp(a) with various scores were Framingham 0.53 (CI: 0.39-0.68; p = 0.644), QRISK-3 0.57 (CI: 0.42-0.71), INTERHEART 0.55 (CI: 0.39-0.69), SCORE-2D 0.47 (CI: 0.32-0.61), EAS 0.65 (CI: 0.50-0.79), and EAS-Lp(a) 0.68 (CI: 0.54-0.83). In addition, adding Lp(a) to the EAS risk calculator increased risk reclassification by a range of 4.6-19.3%.
CONCLUSION: Substantial variation in coronary artery disease (CAD) risk prediction using various clinical algorithms is observed in T2D. The EAS algorithm provides the most robust estimate. The addition of Lp(a) to the risk algorithms augments risk stratification significantly. The results of this pilot study need confirmation with larger prospective studies.