SERUM URIC ACID, INDEPENDENTLY OF ARTERIAL HYPERTENSION, PREDICTS FATAL HEART FAILURE IN A LARGE COHORT OF MEN AND WOMEN. SEARCH FOR A CUT-OFF VALUE

Abstract

Objective: Searching for a prognostic cut-off value of serum uric acid (SUA) in predicting incident heart failure (HF) in a large regional-based Italian cohort of men and women in the frame of the URRAH study (URic Acid Right for heArt Health). Design and method: The ongoing large database URRAH collects data from studies and cohorts from hypertension centres and epidemiological laboratories including subjects with at least 1 measure of SUA and a follow-up of ∼20 years. Incident HF was defined in 23,475 subjects on the basis of ICD10 codes and double-checked with general practitioners and hospital files. Multivariate dichotomic Cox regression models having fatal and morbid HF as dependent variables, adjusted for arterial hypertension (AH), age, sex, diabetes, hematocrit, LDL-cholesterol, smoking and chronic renal disease were used to search for an association between SUA as a continuous variable and HF. Two prognostic cut-off values (one for fatal and one for morbid HF), identified by means of receiver operating curves (ROC) and able to discriminate between subjects doomed to develop the event, were then used as independent predictors to divide people into those <cut-off and >cut-off in further multivariate Cox models adjusted for the confounders. Results: In Cox analysis, SUA as a continuous variable was a significant predictor of fatal [odds ratio, OR 1.101 (1.011–1.200), p = 0.028] and morbid [OR 1.179 (1–031–1.349), p = 0.016] incident HF, independently of AH. Diuretics reduced the risk of fatal HF only [OR 0.448 (0.259–0.774), p = 0.005]. ROC showed that >5.70 mg/dl (CI 4.40 to 6.77, sensitivity 42.67, specificity 71.15, p < 0.0001) was the univariate prognostic cut-off value for fatal HF, and >4.20 mg/dl (CI 3.36 to 6.05, sensitivity 75.06, specificity 31.14, p < 0.003) for non-fatal HF. Nevertheless, only the cut-off for fatal HF was accepted as multivariate predictor in Cox analysis, the hazard ratios being 1.645 (1.284 - 2.109, p < 0.0001), with diuretic acting as a protective factor (p = 0.007), while that for morbid HF was rejected (p = 0.4). Conclusions: In conclusion, a clear prognostic cut-off value of SUA for fatal HF exists (>5.70 mg/dl) also after adjustment for confounders including AH, while SUA did not predict morbid HF.