Low LDL Cholesterol and Increased Risk of Type 2 Diabetes

Lipid lowering therapy using HMG-CoA reductase inhibitors (statins) is associated with an approximately 9-12% increase in the risk of new-onset type 2 diabetes (T2DM). The risk of diabetes could be increased by statins directly; however, genetic approaches have also implicated low LDL cholesterol (LDL-C) concentrations as a risk factor for T2DM. Mendelian randomization studies using functional variants in both HMGCR (GeneID:3156) and PCSK9 (GeneID:255738) genes found a higher risk of T2DM in individuals with variants associated with lower LDL-C concentrations. Since PCSK9 and HMGCR are involved in lipid metabolism through distinct molecular pathways, the altered glycemic effect associated with variants in both genes is likely to be the result of their common effect on LDL-C concentrations. Despite the findings from statin clinical trials and genetic studies, there is little direct evidence implicating low LDL-C concentrations with increased risk of T2DM. Individuals who have very low LDL-C concentrations not due to lipid lowering therapy can provide insights into the relationship between low LDL-C concentrations and T2DM. Here, we used de-identified electronic health records (EHRs) to test the hypothesis that low LDL-C concentrations are associated with T2DM. ]]> The study was approved by Vanderbilt University Medical Center Institutional Review Board (IRB#150137). Patients in the Synthetic Derivative who had one or more outpatient LDL-C measurement were eligible to enter the cohort. LDL-C measurements were excluded if they were performed (1) in hospital; (2) before 5 years of age; (3) within 30 days of a serum albumin level < 3 g/dL; and (4) after the first mention of a statin in the EHR. The low LDL-C cohort was defined as individuals with median LDL-C levels ≤ 60 mg/dL and no LDL-C measurement ever ≥ 80 mg/dL; the normal LDL-C cohort was defined as those with median LDL-C levels between 90 mg/dL and 130 mg/dL and no LDL-C measurement ever ≥ 150 mg/dL or ≤ 80 mg/dL. We excluded individuals with median LDL-C concentrations between 60 and 90 mg/dL to reduce misclassification between the low and normal LDL-C cohorts. We manually reviewed the charts of 97 individuals with LDL-C concentrations less than 20 mg/dL to ensure that they were not receiving statin therapy. We removed individuals who had a median LDL-C <0 (n=8) (on review these were found to be an artifact of the Friedewald formula for calculating LDL-C). ]]>