A New Approach to Assessing the Statin Therapy and Diabetes Link

By Nisha Acharya, MD; James F. Bena, MS; Pooja Manroa, MD; Shristi Kunwar, MD; Subramanian Kannan, MBBS, MD; and Marwan Hamaty, MD, MBA

Statins have been linked to an increased risk of type 2 diabetes mellitus (T2DM) in a number of post hoc and meta-analyses, while others analyses have reported reduced progression to T2DM in association with statin therapy. In these studies, glycemia at baseline typically was not well-defined. It mostly relied on fasting glucose, which may miss a large number of patients with early diabetes. In addition, physician reporting had been among the most often used methods of identifying diabetes incidence.

Our retrospective study within Cleveland Clinic’s Endocrinology & Metabolism Institute took a different approach to identifying patients with no known T2DM at baseline by using the 2-hour oral glucose tolerance test (OGTT) rather than fasting glucose. Patients categorized as having no known diabetes either had a normal OGTT or prediabetes. We also accounted for glucose levels at one hour during OGTT, which is a strong predictor for developing T2DM, especially when it exceeds 155 mg/dL. We also took additional steps vs. previous studies to verify diabetes status at follow up, as described in the information that follows.

Using these criteria, our analysis found no increased risk of T2DM associated with statin use.

Defining progression to T2DM

A retrospective search of electronic medical records between 2000 and 2010 identified 2,370 patients older than 18 years who had no diabetes at baseline based on OGTT and who had further follow up within Cleveland Clinic. Information on statin use was collected at baseline OGTT and during follow up, with the follow-up period extending through December 2012.

Incident diabetes was defined per laboratory findings based on American Diabetes Association criteria, use of diabetes medications and ICD-9 codes compatible with T2DM. All reported cases of T2DM were supplemented by manual chart review if the diagnosis was based on either medication use or ICD-9 codes. Accordingly, patients who had been prescribed diabetes medications at follow up for reasons other than T2DM (836 patients) were excluded. Other excluded patients included those who had positive antibodies to glutamic acid decarboxylase, insulin or islet cells, or patients who had pancreatic surgery (11 patients total).

Time to development of T2DM was evaluated using Kaplan-Meier estimates, and risk was measured according to Cox proportional hazards models. All models were adjusted for glycemia, and the first set of models was adjusted for clinical risk factors. The second set of models was adjusted for the use of other medications.

No increased risk with statin use

The final analyses included 1,523 patients, of which 62 percent were women and 76 percent were white. The mean age was 54.6 years, and the mean BMI was 31 (SD 7.6) kg/m2. Median follow up was 40 months, with some patients developing T2DM in less than one month. In the cohort, 779 patients (51 percent) never used statins, 90 (6 percent) were or had been on a statin prior to baseline with no record of subsequent use after OGTT, and 190 (13 percent) had been prescribed statin therapy only at follow up (after OGTT). The remaining 464 patients (30 percent) were taking a statin both before and after baseline. Atorvastatin and simvastatin accounted for the majority of statin therapy (47 percent and 42 percent, respectively).

Diabetes was diagnosed in 274 patients (18 percent). After adjusting for glycemia and clinical risk factors, there was no increased risk for developing T2DM among patients who used statins at any time compared with patients who never used statins (HR [hazard ratio] 0.86; 95 percent CI [confidence interval], 0.66–1.12). Results were consistent when the statistical models adjusted for glycemia and medication therapies.

In a subgroup analysis based on statin use (after OGTT only, prior to and after OGTT, or only prior to OGTT), patients who used statins only prior to OGTT, an increased risk for developing diabetes was observed (23 out of 90 patients, or 25 percent). As with the other subgroups in the study, the analysis included an adjustment for glycemia and other clinical risk factors (HR 2.87; 95 percent CI, 1.76–4.68).

Future research needed

While our analysis found no evidence for a higher risk of developing T2DM with the general use of statins, we did not study the potential effects of any specific statins. Additional research could include a subanalysis of individual statins.

An intriguing observation in our study was the increased risk of developing T2DM in association with the interruption of statin use, although the number of patients in this subgroup was relatively small. It’s possible that the higher risk in this group could be due to differences in lifestyle management. Future research could examine this possibility as well as other potential explanations, such as mechanisms that could lead to both the discontinuation of statins (i.e., side effects) and the development of diabetes.

Marwan Hamaty, MD, is an endocrinologist in the Department of Endocrinology, Diabetes & Metabolism Institute. Drs. Acharya and Kannan were endocrine fellows in the department when the research was conducted. Drs. Manroa and Kunwar are internal medicine physicians in Cleveland Clinic’s Medicine Institute. Mr. Bena is a lead biostatistician in the Biostatistics section of Quantitative Health Sciences at Cleveland Clinic.