DIABETES AND CARDIOVASCULAR EVENTS: USEFULNESS AND LIMITS OF RISK FUNCTIONS

Cardiovascular diseases represent a leading cause of death and disability worldwide. Risk functions and algorithms have been constructed to estimate cardiovascular risk as the weighed sum of different variables, considered as risk factors. Basically all utilized functions consider age, gender, cigarette smoking status, some estimate of systemic blood pressure and of plasma lipid profile, and diabetes. The functions derived from the Framingham study are the most commonly utilized worldwide, even if their applicability to some geographical areas has been heavily questioned. Furthermore, almost all functions, including Framingham’s, consider diabetes only as a dichotomous (yes / no) variable, despite its heavy impact on cardiovascular risk. On the other hand, the risk engine developed by the United Kingdom Prevention of Diabetes Study (UKPDS) takes into account some variables which characterize diabetes and its severity.Evidence from an outpatient population of diabetic subjects in northern Italy has underlined the extreme variability of risk stratification when utilizing different risk functions; whereas anglo-american functions overestimate cardiovascular risk, Italian charts and functions tend to underestimate risk, particularly in women. Estimation of cardiovascular risk is largely dependent on the function adopted, and on the baseline risk of each cohort. Functions should be designed which are geographically homogeneous for a selected population, and which take into account variables that are specific for diabetes. This should hopefully allow appropriate identification of high risk subjects, and ultimately help to optimize the allocation of health care resources.