PROGNOSTIC CUTOFF OF TRIGLYCERIDES PREDICTING CARDIOVASCULAR OUTCOME IN A LARGE REGIONAL-BASED ITALIAN DATABASE

Objective: Searching for a prognostic cut-off value of triglycerides (TG) in predicting cardiovascular (CV) outcome in a large regional-based Italian database of men and women. Design and method: A nationwide multicentre database collects data on subjects aged 18 to 95 years recruited on a regional community basis from all the territory of Italy with at least 1 measure of TG and a mean follow-up of 110.1 ± 64.3 months. A total of 18,784 subjects were included in the analysis. CV outcome was defined on the basis of ICD10 codes and double-checked with general practitioners and hospital files. Multivariate dichotomic Cox regression models having CV outcome as dependent variable, adjusted for age, sex, serum uric acid, body mass index, total and HDL-cholesterol, arterial hypertension, diabetes, chronic renal disease, smoking habit, use of antihypertensive and lipid lowering drugs were preliminarily used to search for an association between TG log-transformed as a continuous variable and CV event. The conventional cut-off value of TG (150 mg/d) and the prognostic cut-off value of TG identified by means of receiver operating characteristics (ROC) curves, both able to discriminate between subjects developing a CV event, were used as independent predictors in further multivariate Cox models adjusted for the confounders listed above. Results: During 172,329 person-years of follow-up, 1910 participants experienced CV events (11.1 per 1000 age-adjusted person-years). In Cox analysis, TG as a continuous variable was a significant predictor of CV event [odds ratio, OR, 1.189 (1.060-1.334), p = 0.003]. ROC curve showed that > 105 mg/dl (95%CI 75.8-129.3, sensitivity 61.7, specificity 48.8, p < 0.0001) was the prognostic cut-off value for CV event. The conventional cut-off (150 mg/dl) and that identified by ROC curve (105 mg/dl) were both accepted as multivariate predictors in separate Cox analyses, the hazard ratios being 1.157 (95%CI 1.037-1.290, p = 0.009) and 1.137 (95%CI 1.023-1.263, p = 0.016), respectively. Nevertheless, we observed that cut-offs of TG between 89 and 105 mg/dl, when played in the multivariate Cox analyses, seems to be protective as well. Conclusions: In conclusion, lower than expected prognostic cut-off value of TG for CV event do exist.