Is All Coronary Heart Disease Prevention in Type 2 Diabetes Mellitus Secondary Prevention?

DIABETES IS associated with a 2- to 4-fold increase in coronary heart disease (CHD) (1–3). In the overall population, clinically established CHD is associated with a 3- to 7-fold increase in CHD mortality (4). Plasma cholesterol levels are a strong predictor of risk in patients with clinical CHD (4) and diabetes (3). The much higher risk of CHD in patients with clinical CHD and diabetes has led both the National Cholesterol Education Program (NCEP) (5) and the American Diabetes Association (ADA) (6) to recommend lower goals for low-density lipoprotein cholesterol (LDL-C) in these risk groups. The possible role of lipid lowering to reduce CHD in diabetic patients has been enhanced since the publication of data on the efficacy of lipid lowering with simvastatin (7, 8) and pravastatin (9) in diabetic subgroups with preexisting CHD. In the Scandinavian Simvastatin Survival Study (n = 202) (4S) (8), diabetic patients received more benefit from lipid lowering than did nondiabetic patients (55% vs. 32% reduction, respectively) in lowering major CHD (fatal + nonfatal CHD); in the Cholesterol and Recurrent Events (CARE) study (n = 586) (9), similar benefits were seen in diabetic and nondiabetic subgroups [27% vs. 25% reduction in CHD events (major CHD + revascularizations)]. Unfortunately, no current data exist on possible effectiveness of statins in diabetic patients without CHD. The Helsinki Heart Study (10) gemfibrozil reduced CHD in diabetic patients by 60%, although the result was not significant because of the small number of subjects (n = 135) and the lower event rate in this primary prevention cohort. Diabetic patients with CHD have a worse prognosis than nondiabetic patients with CHD (11–13). Because of high case fatality rates in diabetic patients with preexisting CHD (11–13) and previous data on the effectiveness of LDL-C lowering in diabetic subjects (8, 9), it has been suggested that new statin vs. placebo-controlled trials in the prevention of CHD should not be “initiated” in diabetic patients with preexisting CHD (14). The 1-yr case fatality rate for first myocardiol infarction (MI) (from the onset of symptoms, thus including prehospitalization mortality) in the FINMONICA (13) was 45% in diabetic men and 39% in diabetic women. These 1-yr case fatality rates were significantly higher than the rates in nondiabetic patients (38% and 25%, respectively). Of the diabetic patients who died, 50% of men and 25% of women died before hospitalization. These patients, by definition, could not benefit from secondary prevention strategies, indicating that aggressive management of cardiovascular risk factors in diabetic subjects (especially diabetic men) should precede the onset of clinical CHD. The NCEP (5) suggests an initiation level for pharmacological therapy in subjects with clinical CHD more than 130 mg/dL and a treatment goal of less than 100 mg/dL. The recent ADA recommendation (6) modifies the above recommendation by suggesting the initiation level be 100 mg/dL or greater, whereas the goal remains less than 100 mg/dL. The lower initiation level was, in part, based on the greater risk of recurrent CHD in diabetic than in nondiabetic patients with clinical CHD. It may be questioned why the goal for diabetic subjects with prior CHD should not be much lower than 100 mg/dL (i.e. LDL <70 mg/dL); however, in the absence of clinical trial data in either diabetic or nondiabetic patients of better outcomes for much lower LDL cutpoints than with an LDL-C of 100 mg/dL, this was not indicated. The current NCEP guidelines (5) for diabetic subjects without clinical CHD varies depending on the presence of at least two other cardiovascular risk factors. Thus, the LDL-C level for initiation of therapy varies from 160–190 mg/dL, and the LDL-C goal varies from 130–160 mg/dL. The new ADA guidelines (6) are considerably more aggressive. Considering that diabetic women are relatively (although perhaps not absolutely) at greater risk of CHD than diabetic men, the ADA considered all diabetic patients as having “2 risk factors.” In addition, the ADA recommends that the LDL-C level for initiation of therapy and the goal of therapy be set at 130 mg/dL. A footnote to one of the tables in the ADA report suggested that the goal for LDL-C might be less than 100 mg/dL if there were additional CHD risk factors (smoking, family history of CHD, low high-density lipoprotein cholesterol, hypertension, or microalbuminuria). It is expected that many, if most, diabetic patients have additional risk factors for CHD. One possibility in assessing whether diabetic patients, irrespective of the presence of CHD, should have as aggressive lipid lowering as patients with clinically established CHD is to examine the risk of CHD and cardiovascular disease events in diabetic subjects with and without prior CHD (relative to nondiabetic subjects with and without prior CHD). In previous reports, the excess of CHD risk in patients with prior MI (3- to 7-fold) (4) is higher than the excess risk of CHD in diabetic patients (2- to 4-fold) (1–3), but comparisons are difficult across different populations. Furthermore, patients with diabetes are overrepresented in patients with prior MI (1–3), and diabetic patients with MI have a worse prognosis than nondiabetic patients with CHD (11–13). Thus, the risk of recurrent CHD in the overall population might be overestimated by the inclusion of diabetic patients in previous studies. We have recently examined this issue in a 7-yr follow-up in 1373 nondiabetic subjects and 859 diabetic subjects from the East West study, a population-based study of diabetes in Finland (15). The 7-yr incidence of MI in nondiabetic patients with and without MI at baseline was 18.8% and 3.5%, respectively (P < 0.001), whereas the 7-yr incidence of MI in diabetic patients with and without MI at baseline was 45.0% and 20.2%, respectively (P< 0.001). The hazard ratio for CHD mortality for diabetic patients without prior MI compared with nondiabetic patients with prior MI was not significant (hazard ratio, 1.4; 95% confidence interval, 0.7, 2.6) after adjustment for age and gender, suggesting similar prognosis in the two groups. To further assess this issue (16), we compared the intima-media wall thickness (IMT) in the common carotid artery (CCA) and internal carotid artery (ICA) in 43 diabetic patients with clinical CHD, 446 diabetic patients without clinical CHD, 47 nondiabetic subjects with clinical CHD, and 975 nondiabetic subjects without clinical CHD in the Insulin Resistance Atherosclerosis Study (IRAS). Both diabetes and CHD were associated with increased atherosclerosis in the CCA. Likewise, diabetes was significantly associated with increased atherosclerosis in the ICA; however, CHD was not associated with ICA IMT. As expected, diabetic patients with coronary artery disease had the greatest IMT CCA (0.948 mm), whereas nondiabetic patients without coronary artery disease had the least atherosclerosis (CCA, 0.792 mm). Subjects with diabetes but without CHD had slightly greater IMT (CCA, 0.868 mm) than nondiabetic subjects with CHD (CCA, 0.861 mm), although these differences were not statistically significant. These two preliminary reports suggest that diabetic patients without preexisting vascular disease have similar risk of CHD as nondiabetic subjects with vascular disease. If diabetic patients were treated to the same goal as nondiabetic patients with prior CHD, as defined by NCEP (5), the initiation level for pharmacological treatment of LDL-C would be more than 130 mg/dL and the LDL cholesterol goal would be less than 100 mg/dL; in this case, most diabetic patients would be eligible for pharmacologic lipid lowering therapy. This could imply a large increase in pharmacological therapy and, thus, could produce an objectionable increase in health care expenditures. No cost benefit data are available for lipid lowering in diabetic patients. Recent data based on 4S study (i.e. secondary prevention) suggest a 34% reduction in hospitalization (17), which would markedly reduce the cost of HMG-CoA reductase inhibitor therapy. In a further analyses of the 4S data (18), the investigators suggested that for some patients, pharmacologic therapy might be cost-saving, once indirect costs associated with morbidity of CHD were taken into account. Because the risk of CHD in diabetic patients without prior CHD is similar to that of nondiabetic subjects with prior CHD, it is possible that lipid lowering might be cost effective in diabetic patients even without prior CHD. However, cost-benefit studies should be done specifically for diabetic patients. Diabetic subjects without prior CHD have a similar degree of atherosclerosis and rate of CHD as nondiabetic patients with prior CHD. Because LDL-C lowering with HMG-CoA reductase inhibitors seems to be at least as effective in diabetic patients as in nondiabetic patients, a strong case can be made that diabetic patients without vascular disease should be treated similarly to nondiabetic patients with vascular disease, with respect to aggressive treatment of lipid therapy. Definitive data, however, need to be collected from clinical trials in diabetic patients without vascular disease. “Clinical Perspectives” are an occasional feature of The Journal of Clinical Endocrinology & Metabolism. They present the opposing views of invited contributors on a topic. All reprints must include the complete Clinical Perspective, so that each section can be read in context.