Influence of Gender and Repeated Urine Sampling on the Association of Albuminuria with Coronary Events

Abstract

Background: The association between albuminuria and coronary heart disease (CHD) is well-known, but uncertainties related to day-to-day variability and effect modification of gender complicate the risk assessment process. This study evaluates the associations of CHD with albuminuria level in men and women based on the number of urine samples. Methods: Nine thousand one hundred and fifty-eight adults provided 3 urine samples and were followed for 14 years in the population-based HUNT-2 cohort study. The association of myocardial infarction or coronary death with different albumin-creatinine ratio (ACR) cut-offs, based on gender and number of positive ACRs, were estimated by hazard ratios (HRs) and adjusted for by Framingham variables. Results: Associations between ACR and CHD were similar in men and women. For example, HRs for moderately increased (3.0≤ ACR ≤30.0 mg/mmol) vs. normal albuminuria (ACR <1.0 mg/mmol) were 1.40 (95% CI 1.27-2.03) and 1.61 (95% CI 1.15-1.71) respectively, (p_sex-equality = 0.3). However, median intra-individual day-to-day ACR coefficient of variation was 22.4% in women vs. 17.5% in men (p < 0.001). Two or 3 positive ACRs were required to establish a significant association with CHD at levels below 4.0 mg/mmol in women, while one positive ACR implied a significant association at all levels in men. Based on receiver-operating-characteristics curves, the Youden index suggested possible equal cut-offs for women (1.12 mg/mmol) and men (0.88 mg/mmol), p = 0.06. Conclusions: There were no significant gender differences in the association between albuminuria and coronary events. However, women had increased intra-individual albuminuria variability compared to men, necessitating several positive urine samples if mildly increased albuminuria is used in coronary risk evaluation.