Objective: The predictive value of microalbuminuria (MA) in children with type 1 diabetes has not been defined. We describe the natural history of MA in a large cohort of children recruited at diagnosis of type 1 diabetes.
Research design and methods: Between 1985 and 1996, 514 children (279 male) who developed type 1 diabetes before the age of 16 years (91% of those eligible from a region where ascertainment of new cases is 95%) were recruited for a longitudinal study with central annual assessment of HbAlc and albumin excretion (three urine samples). Dropout rates have been < 1% per year, and 287 children have been followed for > 4.5 years.
Results: MA (defined as albumin-to-creatinine ratio > or = 3.5 and > or = 4.0 mg/mmol in boys and girls, respectively) developed in 63 (12.8%) and was persistent in 22 (4.8%) of the subjects. The cumulative probability (based on the Kaplan-Meier method) for developing MA was 40% after 11 years. HbAlc was worse in those who developed MA than in others (mean difference +/- SEM: 1.1% +/- 0.2, P < 0.001). In subjects who had been 5-11 years of age when their diabetes was diagnosed, the appearance of MA was delayed until puberty, whereas of those whose age was < 5 years at diagnosis of diabetes, 5 of 11 (45%) developed MA before puberty. The adjusted proportional probability (Cox model) of MA was greater for female subjects (200%), after pubertal onset (310%), and with greater HbAlc (36% increase for every 1% increase in HbAlc). Despite earlier differences based on age at diagnosis of diabetes (< 5, 5-11, and > 11 years), the overall cumulative risks in these groups were similar (38 vs. 29 vs. 39%, respectively) after 10 years' duration of diabetes.
Conclusions: Prepubertal duration of diabetes and prepubertal hyperglycemia contribute to the risk of postpubertal MA. The differences in rates of development of MA relating to HbAlc, sex, and age at diagnosis relative to puberty may have long-term consequences for the risk of subsequent nephropathy and for cardiovascular risk.