Metabolic Syndrome and Outcome Predictions: Friends or Foes?

OBJECTIVES: An analysis based on epidemiological material to show whether the term Metabolic Syndrome (MS) should be adopted when aiming at predicting coronary heart disease (CHD) and major cardiovascular disease (CVD) fatal events. MATERIAL AND METHODS: MS was defined according to the International Diabetes Federation (IDF) and risk factors were identified in the Italian Risk Factors and Life Expectancy (RIFLE) population study covering over 25,000 adult men from a pool of 19 Italian population samples. The original MS definition and the plain original units of measured risk factors were challenged in Cox proportional hazard models predicting 196 CHD and 412 major CVD fatal events in a seven-year follow-up. Parallel models were run including also total serum cholesterol as a covariate, an unfortunately excluded covariate in the MS definition. The performance of the various models was tested by the log-likelihood statistics treated with the Akaike Information Criterium (AIC). RESULTS: Models using the plain measurements of the risk factors involved were systematically and significantly outperforming any other categorized score based on the IDF-MS classification. An intermediate role was played by a model where the predictive variable was a factor score (derived from a Factor Analysis) where the MS risk factors were linearly combined. The same models also including serum cholesterol provided a significantly better prediction when compared with those without serum cholesterol, based on AIC. CONCLUSIONS: The use of a subset of classical CVD risk factors classified according to the IDF-MS criteria adds nothing better than the exclusive use of the risk factors treated by traditional procedures. The addition of serum cholesterol definitely helps in the prediction of the CHD component of major CVD events. Its omission is erroneous.