It has long been appreciated that diabetes is associated with a specific dyslipidemia characterized by elevated triglyceride, reduced high-density lipoprotein (HDL) cholesterol,1 and elevated levels of small, dense low-density lipoprotein (LDL), with these abnormalities in turn associated with insulin resistance,2,3 and with increased cardiovascular disease (CVD) risk both in type 24 and type 1 diabetes.5 The fibrates gemfibrozil6,7 and bezafibrate8 reduce CVD among diabetic people, but the evidence for CVD reduction with fenofibrate in the large Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) trial has been difficult to interpret, in part because of the greater use of statins among patients randomized to the placebo arm of the study.9 Although FIELD showed reductions in non-fatal myocardial infarction, total CVD, and revascularization, CVD mortality increased, a finding that raises concern in the general use of this agent among diabetic patients. Subset analysis showed greater benefit among the group with low HDL cholesterol, particularly when triglyceride levels were elevated at the same time.10 The trial also revealed evidence of reduction in albuminuria, and decreased macular edema and proliferative retinopathy, leading to a lessened requirement for laser photocoagulation;11 the two observations suggest an effect on microvascular disease, although there is evidence of reduction in the glomerular filtration rate, so the agent should be used cautiously.12
In this regard, the lipid substudy of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial was of great interest, comparing patients randomized to more and less intensive glycemic control, with both groups randomized to simvastatin versus simvastatin plus fenofibrate. Fenofibrate showed a trend to increasing HDL cholesterol, particularly during the initial year of the study, and led to a sustained reduction in triglyceride levels. Overall, there was no significant reduction in CVD, but the subset with triglyceride ≥204 mg/dL and HDL cholesterol ≤34 mg/dL did show a 29% lower CVD event rate. Furthermore, the progression of diabetic retinopathy was significantly reduced with both glycemic control (P = 0.003) and fenofibrate use (P = 0.006).13 Among those randomized to intensive glycemic control, 5.3% of those receiving a fenofibrate, but 7.1% of those receiving placebo, had progression, which was a non-significant trend; in the standard glycemic control arm, 7.6% of those receiving fenofibrate and 13.4% of those receiving placebo had progression, suggesting the benefit was seen with either of the two interventions, although not clearly demonstrating a beneficial effect of their interaction (Fig. 1). It is interesting to speculate as to whether the retinopathy effect is directly related to lipid-lowering or reflects participation of peroxisome proliferator-activated receptor-α in the process of retinal ischemia leading to vasoproliferation.
Is it, then, appropriate to recommend fibrates for CVD prevention among persons with diabetes? The intervention may lead to benefits in the subset of people with the typical diabetic dyslipidemia pattern. Should we, in addition, consider fibrates to accompany our efforts to control glycemia in the prevention of microvascular disease? This may be a reasonable recommendation for persons with or at risk of developing retinopathy, particularly when the risk of hypoglycemia limits the ability to attain excellent glycemic control. Triglyceride levels were an independent risk factor for retinopathy among type 1 diabetic patients in the EURODIAB study,14 further suggesting that fibrate treatment should be beneficial. Indeed, we should see the combined use of lipid-lowering and glycemic treatment as a specific example of the general principle that the management of diabetes today must involve appropriate multifactorial treatment to optimally reduce disease complications.15