Diabetic Nephropathy in 27,805 Children, Adolescents, and Adults With Type 1 Diabetes: Effect of diabetes duration, A1C, hypertension, dyslipidemia, diabetes onset, and sex

The German Diabetes Documentation System (DPV) is a prospective, nationwide survey of demographic, anthropometric, and diabetes-related characteristics of patients with type 1 diabetes (19,20). Local data-control authorities approved data collection and anonymous analysis for study purposes; participating centers are listed in the online appendix (available at http://dx.doi.org/10.2337/dc07-0282). Data acquisition for the present analysis was done until February 2007 and included a total number of 27,805 children, adolescents, and adults with type 1 diabetes consecutively registered at 262 centers for pediatric or adult diabetes care. The following independent risk factors for development of nephropathy were analyzed by logistic regression and stepwise selection of parameters: diabetes duration, age at diagnosis, sex, blood pressure, A1C, dyslipidemia, cumulative A1C, and smoking.

Screening for microalbuminuria was performed by the following methods: 1) measurement of the urine albumin-to-creatinine ratio (UAC) in a random spot collection, 2) 24-h collection with creatinine, and 3) timed (e.g., overnight) collection. Microalbuminuria or macroalbuminuria was defined as at least two increased urine albumin tests during the follow-up period. Within the most recent year, albuminuria was determined by the UAC ratio in 12.1%, AER (24 h) in 31.7%, and albumin concentration from a timed overnight collection in 56.2% of subjects. Thresholds were AER ≥20 μg/min or a UAC ≥2.5 mg/mmol, according to guidelines of American Diabetes Association (21). Macroalbuminuria was defined as AER ≥200 μg/min or a UAC ≥35 mg/mmol. Albumin and creatinine were measured by the center-specific laboratory methods that had to meet German internal and external quality requirements for laboratory analysis. Patients with ESRD (requiring dialysis/transplantation) were included in the Kaplan-Meier survival analysis to adequately reflect prevalence of severe nephropathy but were excluded from regression analysis for macroalbuminuria-associated covariates.

Glycemic control was assessed as mean A1C during the follow-up period. Single-center A1C methods were standardized according to the DCCT reference range of 4.05–6.05% (22).

Systolic and diastolic blood pressure were measured according to current guidelines. Age-specific normal values were obtained from the Task Force on Blood Pressure Control in Children and Adolescents (23). Hypertension was defined as the median value >95th percentile of at least three independent measurements. Hypertension was classified as a dichotomous variable in macroalbuminuric patients but as a continuous variable (systolic and diastolic blood pressure) in the large cohorts of normoalbuminuric and microalbuminuric patients.

Dyslipidemia was diagnosed if at least one lipid parameter was permanently increased. Cutoffs were for total cholesterol >200 mg/dl, LDL cholesterol >160 mg/dl, and triglycerides >150 mg/dl. In logistic regression analysis for microalbuminuria, triglycerides, and HDL and LDL cholesterol were treated as independent, continuous variables, whereas for macroalbuminuria, dyslipidemia was treated as a dichotomous variable.

Subjects’ smoking habits were asked at each single visit. If at least one cigarette per day was reported, patients were classified as smokers.

Patients with dyslipidemia and systolic or diastolic hypertension were treated with lipid-lowering or antihypertensive drugs. Lipid-lowering drugs used were statins, fibrates, or cholesterol absorption inhibitors such as ezetimibe. Treatment with antihypertensive drugs were classified and documented as ACE inhibitors, angiotensine receptor blockers, β-blockers, diuretics, and calcium channel blockers. Pharmacotherapy with these respective drugs was documented in a qualitative fashion.

The SAS 9.1 statistical software package was used for data evaluation and statistical analysis. In detail, the relative contribution of covariates to risk for nephropathy was analyzed by logistic regression and stepwise selection of parameters, Kaplan-Meier method for survival analysis, and the log-rank test for trend. Differences between groups of normoalbuminuria, microalbuminuria, and overt nephropathy were analyzed by the nonparametric Mann-Whitney U test (Table 1). The odds ratio (95% Wald confidence limits) are reported for logistic regression analysis, and data were corrected for use of antihypertensive and lipid-lowering drugs. Data are presented as means ± SE, where appropriate.

The study cohort was analyzed in February 2007 from 262 diabetes centers in Germany and Austria, with a total number of 49,027 patients with type 1 diabetes being continuously followed in the DPV. At the time of analysis, 27,805 patients had at least two documented urine tests with clear results of normal or albuminuric reading in at least two follow-up visits and gave consent for scientific data evaluation. We assume that the cohort analyzed is representative for all patients followed in the DPV, as A1C (mean; included 7.99 vs. all 7.92%), male sex (52.5 vs. 51.9%), age (21.1 vs. 21.4 years), diabetes duration (8.3 vs. 7.8 years), and age at diabetes onset (12.9 vs. 13.6 years) were largely comparable. The following were the characteristics of the analyzed cohort: follow-up time (median [interquartile range]) 2.5 years (0.5–5.3), follow-up visits 12.0 (3.0–26.0), A1C determinations 9.0 (9–21), and blood pressure measurements 12 (3.0–26.0). Age at the most recent visit (Fig. 1A) and age at diagnose (Fig. 1B) are predominantly in the pediatric and adolescent age. At their most recent visit, normal urine tests were found in 26,605 patients (95.6%) and at least two abnormal urine tests were found in 1,200 patients (4.3%). Among the latter group, 281 of 1,200 (23.4%) had at least two urine tests in the defined range of macroalbuminuria (AER >200 μg/min, UAC >35 mg/mmol/l) or ESRD, with ongoing dialysis or after renal transplantation. In detail, 52 patients were only macroalbuminuric and 229 proceeded toward ESRD.

Comparing patients with normoalbuminuria and micro- or macroalbuminuria, the variables A1C, systolic and diastolic blood pressure, tobacco consumption, and serum lipids, including cholesterol, LDL cholesterol, and triglycerides, were found to be higher in patients with increased albumin excretion (Table 1). But these changes were parallel with increasing age, changed insulin treatment regimens, and had longer diabetes duration, and therefore these findings were only noticed in a descriptive fashion. Treatment with antihypertensive drugs increased with severity of nephropathy, but it still was unacceptably low in patients with micro- or macroalbuminuria (Table 1).

Kaplan-Meier analysis revealed that after median diabetes duration of 40 years, 25.4% (95% CI 22.3–28.3) of patients had microalbuminuria and 9.4% (8.3–11.4) had macroalbuminuria or ESRD (Fig. 2). To describe the long-term influence of A1C on nephropathy, patients were stratified into two groups: one with A1C levels above and one with A1C levels below 7.5%. The group with with A1C levels >7.5% developed microalbuminuria (log-rank test, χ² = 51.6, P < 0.0001) and macroalbuminuria (log-rank test, χ² = 8.2, P < 0.0042) significantly earlier than those with levels <7.5% (Figs. 3A and B). Differences between groups became evident after diabetes duration of ∼10 years for microalbuminuria and 20 years for macroalbuminuria.

Logistic regression analysis with stepwise selection of parameters identified diabetes duration, dyslipidemia, and mean A1C as significant risk factors for development of microalbuminuria. When blood pressure and lipid levels were analyzed as continuous variables, and not as dichotomous cutoff levels, systolic and diastolic blood pressure, but also triglyceride and LDL cholesterol concentration, became independent and significant risk factors for the development of microalbuminuria. In the relatively small number of patients with macroalbuminuria, diabetes duration, A1C, male sex, and dyslipidemia were associated with the development of overt nephropathy. Also, young age at diagnosis reduced the risk of microalbuminuria, when adjusted for diabetes duration and other independent covariates (Table 2). Smoking and sex were included in a stepwise selection of potential risk factors, but only male sex was associated with the development of macroalbuminuria.

We analyzed the prevalence of incipient and overt nephropathy in a large cohort of 27,805 children, adolescents, and adults with type 1 diabetes. These data give a representative and up-to-date profile of diabetic nephropathy and associated risk factors in Germany and Austria. Detailed analysis of risk factors is of particular importance, as these could be influenced and controlled during diabetes follow-up.

Within 40 years of diabetes duration, calculated prevalence was 25.4% for persistent microalbuminuria but <10% for macroalbuminuria or ESRD. This is a lower rate of nephropathy than reported earlier from cohorts followed in Denmark, England, and western Australia, with similar age of patients (11,17,24). The most prominent difference in our study was a lower A1C level compared with these studies. But finally, many factors, including cohort characteristics and study design, could explain deviating cumulative incidence of nephropathy.

In our study, microalbuminuria was associated with higher A1C, serum lipids, in-detail LDL cholesterol and triglycerides, and systolic and diastolic blood pressure as well as with older age at diagnosis. Macroalbuminuria, documented in a much smaller number of patients, was linked to A1C, dyslipidemia, and male sex.

Increased A1C as a marker of chronic hyperglycemia is the most established and unquestioned risk factor for diabetic kidney disease in adult- and pediatric-onset type 1 diabetes (11,13,14,17). The DCCT and the follow-up Epidemiology of Diabetes Interventions and Complications study clearly demonstrate that previous intensive treatment of diabetes with near-normal glycemia has an extended benefit in delaying the development and progression of diabetic nephropathy (9,10). By Kaplan-Meier analysis, we found that a cumulative A1C of >7.5% significantly raised the probability of micro- and macroalbuminuria. This effect became evident after diabetes duration of >10 years for microalbuminuria and >20 years for macroalbuminuria. Therefore, efforts to normalize metabolic control should be started right from diabetes onset, although consequences on renal function might not be seen before adult age.

We found both systolic and diastolic blood pressure to be independently associated with microalbuminuria. Hypertension, per se, was not an independent risk factor for nephropathy and was lower in patients with macroalbuminuria or dialysis. This fact might reflect strict antihypertensive treatment regimens in patients with advanced renal disease. But whether or not blood pressure is the cause or result of nephropathy, hypertension should strictly be treated also in pediatric and adolescent patients (25).

Dyslipidemia, as the strongest single factor for elevated LDL cholesterol, is associated with progression of diabetic kidney disease (12,26). We found that both LDL cholesterol and triglycerides are independently associated with microalbuminuria. Furthermore, association between serum advanced glycation end products and increased serum cholesterol imply involvement of lipids in the formation of advanced glycation end products and, by this way, in renal disease (27).

The rate of smokers among patients with type 1 diabetes is alarmingly high, in detail 10.5% of adolescents and 34.8% of young adults in Germany and Austria (28); however, it has recently been controversially discussed whether smoking is a risk factor for development or progression of diabetic kidney disease (13,29,30). We did not find that smoking increased the likelihood of nephropathy, but programs to keep children and adolescents away from starting smoking should be an important factor of continuous diabetes care, first of all to prevent macrovascular disease (15,16).

Sex influence on nephropathy has been found to be age dependent. In adolescents, female sex increases the risk of microalbuminuria (14,17), while in adults and in terms of advanced nephropathy, men are at higher risk to develop renal disease, if data were adjusted to social class, A1C, smoking status, and blood pressure (11,16). We found no link between sex and microalbuminuria but higher risk for macroalbuminuria in male subjects.

Several studies have indicated that prepubertal duration of diabetes delays the onset of diabetic nephropathy. In our study, microalbuminuria was delayed by very early onset of diabetes and confirmed these previous prospective studies (14,31,32). We do not suggest that poor metabolic control in prepubertal children does not add to the risk of microvascular complications, but there is evidence that it does so at a lower rate (33).

In conclusion, diabetes duration, long-term metabolic control (A1C), dyslipidemia, and male sex have been identified as independent risk factors for the development of nephropathy in this large cohort. We conclude from our data that diabetes care must continue to focus on long-term metabolic control as well as on reduction of additional risk factors like dyslipidemia and blood pressure.

Financial support was given by the German Ministry of Health, the German Diabetes Association, Bundesärztekammer, Nationales Aktionsforum Diabetes Mellitus, the Dr-Bürger-Büsing-Foundation, and Novo Nordisk Germany.

We kindly acknowledge each of the 262 participating diabetes centers in Germany and Austria. These centers are listed in the online appendix.