Evidence for two distinct phenotypes of chronic kidney disease in individuals with type 1 diabetes mellitus.

Abstract

In a retrospective, observational, cross-sectional, single-centre study, we assessed the prevalence and correlates of different CKD phenotypes (with and without albuminuria) in a large cohort of patients of white ethnicity with type 1 diabetes. From 2001 to 2009, 408 men and 369 women with type 1 diabetes (age 40.2±11.7years, diabetes duration 19.4±12.2years, HbA1c 7.83±1.17% [62.0±12.9mmol/mol]) were recruited consecutively. Albumin-to-creatinine ratio (ACR) and eGFR (Modification of Diet in Renal Disease) were obtained for all individuals, together with CKD stage. Diabetic retinopathy and peripheral polyneuropathy were detected in 41.5% and 8.1%, respectively, and cardiovascular disease (CVD) occurred in 8.5%. Adjudications of CKD phenotype were made by blinded investigators. Normo- (ACR <3.4), micro- (ACR 3.4-34) or macroalbuminuria (ACR ≥34mg/mmol) were present in 91.6%, 6.4% and 1.9% of individuals, respectively. eGFR categories 1 (≥90mlmin-1 [1.73m]-2), 2 (60-89mlmin-1 [1.73m]-2) and 3 (<60mlmin-1 [1.73m]-2) were present in 57.3%, 39.0% and 3.7%, respectively. The majority of participants had no CKD (89.4%), while stages 1-2 and ≥3 CKD were detected in 6.8% and 3.7%, respectively. The albuminuric (Alb+) and non-albuminuric (Alb-) phenotypes were present in 12 (41.4%) and 17 (58.6%) individuals with stage ≥3 CKD, respectively. Individuals with an ACR <3.4mg/mmol were subdivided into those with normal albuminuria (<1.1mg/mmol; 77.2%) and mildly increased albuminuria (1.1-3.4mg/mmol; 14.4%), and individuals with stage 2 CKD were subdivided into those with eGFR 75-89mlmin-1 [1.73m]-2 and 60-74mlmin-1 [1.73m]-2. ACR <3.4mg/mmol (88.7%) and even <1.1mg/mmol (70.4%) were common in individuals with eGFR 60-74mlmin-1 [1.73m]-2. The prevalence of ACR <1.1mg/mmol was lower but still significant (34.5%) in those with stage ≥3 CKD. In logistic regression analysis, stages 1-2 and ≥3 CKD were independently associated with age, HbA1c, γ-glutamyltransferase, fibrinogen, hypertension, but not with sex, BMI, smoking, HDL-cholesterol or triacylglycerol. Inclusion of advanced retinopathy removed HbA1c from the model. The CKD Alb+ phenotype correlated with diabetes duration, HbA1c, HDL-cholesterol, fibrinogen and hypertension, while the CKD Alb- phenotype was associated with age and hypertension, but not with diabetes duration, HbA1c and fibrinogen. The Alb- CKD phenotype is present in a significant proportion of individuals with type 1 diabetes supporting the hypothesis of two distinct pathways (Alb+ and Alb-) of progression towards advanced kidney disease in type 1 diabetes. These are probably distinct pathways as suggested by different sets of covariates associated with the two CKD phenotypes.