PROGNOSTIC VALUE OF URIC ACID IN RELATION TO DIFFERENT AMBULATORY BLOOD PRESSURE COMPONENTS IN THE ABP-INTERNATIONAL STUDY

Abstract

Objective: The intriguing relationship between uric acid (UA) and cardiovascular events (CVE) and mortality is under active scrutiny. Increased UA levels have been associated with elevated BP and adverse prognosis. However, the extent to which UA maintains its prognostic value over and above specific ambulatory BP components is not well defined. Design and method: We followed 5244 participants of the Ambulatory Blood Pressure International (ABP-I) study with UA determined at entry. We grouped participants by gender-specific UA quartiles (divisions: 4.6, 5.4, and 6.4 mg/dl in men; 3.4, 4.2 and 5.0 mg/dl in women). Cox models were used to estimate the magnitude and changes in UA hazard ratio (HR) for CVE and mortality in relation with different ambulatory BP (ABP) components. The Akaike and Bayes information criteria (AIC, BIC) were used to compare non-nested models. Results: The 24-hour, daytime and nighttime systolic BPs increased linearly (p < 0.001 for all) across the gender-specific UA quartiles. We found no evidence for significant relationships between increasing UA and diastolic ABPs. Participants in the top UA quartile were older (p < 0.001), more frequently diabetic (p < 0.001), and had decreased renal function (p < 0.001), higher cholesterol (p < 0.001), and BMI (p < 0.001) than the others. During 35,087 person-years of follow-up there were 423 CVE (93 fatal, 330 nonfatal) and 185 deaths. The CVE rate per 100 person-years was 0.99, 0.95, 1.11, and 1.81, and mortality was 1.01, 0.55, 0.93, and 2.01, suggesting a possible non-linear relationship across UA quartiles. In fully adjusted survival models including different ABP components, the HR for CVE and mortality for the top UA quartile always remained significant (all p < 0.05 and p < 0.02 for CVE and mortality, respectively) with no evidence for HRs heterogeneity (p = 0.965 CVE, p = 0.999 mortality). Based on AIC and BIC, CVE models including either the average 24-hour or the night systolic BP were equally informative and better than those including daytime BP. Conclusions: In ABP-I participants, UA was a powerful risk marker for subsequent cardiovascular events and mortality and added prognostic value independently of specific ABP components.