Comparison of urinary albumin-creatinine ratio and albumin excretion rate in the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications study

Abstract

Background and objectives: The objective of this study was to compare random urine albumin-creatinine ratio (ACR) with timed urine albumin excretion rate (AER) in patients with type 1 diabetes.

Design, setting, participants, & measurements: A total of 1186 participants in the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Study provided spot urine specimens concurrent with 4-hour timed urine collections. ACR and AER were compared using Bland-Altman plots, cross-classification of albuminuria status and its change over time, and within-person variability.

Results: Despite moderate correlation (r=0.62), ACR levels (mg/g) were lower than AER levels (mg/24 hr). This difference was greatest for men. Gender-specific estimated AER (eAER) values were empirically derived from ACR. Comparing the eAER with measured AER, agreement of prevalent microalbuminuria and macroalbuminuria classification was fair to moderate, and classification of change in albuminuria status over time was different. Intraclass correlations were 0.697 for ACR and 0.803 for AER. Effects of DCCT intensive versus conventional diabetes therapy on urine albumin excretion or classification of albuminuria were similar using the eAER versus measured AER, as were the effects of the previous glycosylated hemoglobin.

Conclusions: Systematic differences exist between urine ACR and AER, related to gender and other determinants of muscle mass. Use of ACR (or eAER) versus AER yields differences in classification of prevalent albuminuria states and changes in albuminuria states over time. These findings support the use of consistent ascertainment methods over time and further efforts to standardize and optimally interpret measurement of urine albumin excretion.

Trial registration: ClinicalTrials.gov NCT00360815 NCT00360893.

Figures

Figure 1.

Scatterplots of the AER from the timed collection versus the ACR. (A) Scatterplot of log(AER) against log(ACR), with marginal box plots and histograms for each variable. Horizontal lines at AER = 30 and 300 mg/24 h are added for reference. (B) Bland-Altman plot of the log(AER) − log(ACR) on the vertical scale and the geometric mean of the AER and ACR on the horizontal scale, using a log scale, with horizontal reference lines for a geometric mean of 0, 30, and 300. Three smoothed nonparametric regression lines are superimposed: One for all participants, one for male participants, and one for female participants.