Not all diabetics are created equal (in cardiovascular risk).

Current guidelines continue to recommend all persons with diabetes be considered at high risk and treated accordingly. For example, National Cholesterol Education Panel Guidelines state ‘Persons with type 2 diabetes should be managed as a coronary heart disease (CHD) risk equivalent. Treatment for low density lipoprotein (LDL) cholesterol should follow Adult Treatment Panel III recommendations for persons with established CHD’. This remains an important tenet, as up to 80% of persons with diabetes will suffer a cardiovascular event. This has led some to conclude that further risk stratification in persons with diabetes is neither warranted nor useful. However, persons with diabetes are not all at equal risk. The data from PREDICT, as well as previous coronary artery calcium (CAC) studies, demonstrate that persons with diabetes will have widely varied cardiovascular risk. While many persons with diabetes do have increased cardiovascular risk compared with the general population, there is a defined subset who have a very low cardiovascular risk and an equally large subset who are at extremely high risk. These groups cannot be discerned on the basis of HbA1c levels, duration of diabetes, age, gender or cardiovascular risk factors. Clearly, coronary calcium scanning (CCS) is not only becoming widely accepted in intermediate risk patients, such as those with several cardiovascular risk factors but data continue to accumulate indicating that presumed high-risk patients may actually benefit from further risk stratification. Increasing evidence that lower LDL-cholesterol values are warranted in persons with extremely high cardiovascular risk (with targets proposed as low as 1.8 mmol/L or 70 mg/dL). Despite this, there will be times when clinicians may need evidence to decide which patients warrant the highest doses of lipid-lowering therapy (some require dual or triple cholesterol-lowering therapies) to achieve these targets. Identifying diabetic persons with 10-, 20or 30-fold increased risk of cardiovascular events will be important to justify these therapies to both the clinician and patient. Evidence from the study by Elkeles et al. demonstrates that a simple, 5-min non-invasive, low-radiation test can potentially identify these patients at markedly increased cardiovascular risk. This study is not unique. It is one of many accumulating studies in this cohort that demonstrate this increased risk with increasing CAC scores (representing increasing atherosclerotic burden). Earlier data from Raggi et al. and Anand et al. demonstrated that persons with diabetes are up to 58 times more likely to suffer a cardiovascular event with high scores (.1000) than those without coronary calcium detected on a CT scan. This is highly consistent with larger studies including persons with and without diabetes. Large prospective studies of CCS are demonstrating marked improvement in cardiovascular risk prediction compared with Framingham or other risk stratification tools. Uniform findings between the study by Elkeles et al. (in a diabetic cohort) and Detrano et al. (from a 6814-person NIH-sponsored population-based study) support more widespread application of this test in persons at intermediate (and perhaps high) perceived cardiovascular risk. Further data are emerging from prospective cohorts in Germany and Rotterdam that further clarify the potential use of this test in more diverse, population-based groups. However, caution must be applied in the widespread application of such testing. Cost–effectiveness algorithms are just starting to be published, and the potential benefit of the increased knowledge provided by the test must be weighed against the radiation dose (albeit small for a calcium score only, unlike CT angiography, where doses can be 10-fold higher). Previous guidelines have suggested that these high-risk groups should not undergo calcium scoring; however, compelling data from these studies are likely to add to the impetus to update these recommendations. If cost–effectiveness can be established, this would make it appealing to derive LDL targets based upon something more objective, such as a measure of the atherosclerosis burden. When clinical trial data are combined, including the Treating to