Subjects
Subjects were from a previously described cohort of patients with type 2 diabetes and nephropathy, who were followed for at least 3 years at the Steno Diabetes Center (n=227) [14]. In the present observational follow-up study, all of the above patients who developed NRA (n=79) were included, and followed from the onset of NRA either until death, or until January 2005.
Type 2 diabetes was diagnosed in patients treated by diet alone or by diet combined with oral hypoglycaemic agents; in insulin-treated patients with a BMI above normal (≥25 kg/m2 in women, ≥27 kg/m2 in men) at the time of diagnosis and in whom diabetes onset was after the age of 40 years; or in insulin-treated patients with a normal BMI if a glucagon-stimulated C-peptide value was ≥0.60 pmol/l.
Diabetic nephropathy was diagnosed clinically in patients with: (1) albuminuria >300 mg/24 h in at least two out of three consecutive samples; (2) diabetic retinopathy; and (3) no clinical or laboratory evidence of non-diabetic renal disease [2]. In all patients with abscence of diabetic retinopathy, the diagnosis of diabetic nephropathy was established by a renal biopsy showing diabetic glomerulosclerosis.
NRA was defined as persisting albuminuria >2500 mg/24 h (which corresponds to total proteinuria >3500 mg/24 h) in at least two out of three consecutive 24-h urine collections. Remission of NRA was defined as a reduction in albuminuria from NRA to <600 mg/24 h, sustained for at least 1 year during follow-up. These definitions of NRA and remission of NRA are in accordance with previous published studies [11, 12].
During the observation period, patients were seen at the Steno Diabetes Center for routine visits three or four times a year. Body weight, postprandial blood glucose, HbA1c, blood pressure and 24-h albuminuria were determined at each visit; GFR, serum creatinine and haemoglobin were measured once a year; and cholesterol was measured at least every second year. No sodium or protein restriction was applied during the study. During follow-up, the treatment goal for HbA1c was 7.5% as recommended by the Danish Medical Association. Arterial hypertension was diagnosed and treated according to the World Health Organization’s criteria (≥160/95 mmHg) until 1995 and thereafter according to the American Diabetes Association’s criteria (≥140/90 mmHg) [15]. Until 2001, no specific class of antihypertensive agents was recommended. The most commonly prescribed agents included diuretics, ACE inhibitors (ACE-I), calcium-channel blockers and beta-blockers. After 2001, agents blocking the renin–angiotensin–aldosterone system (RAAS) were recommended as the initial drug of choice [9, 10, 16]. At the Steno Diabetes Center, low-dose aspirin and statins were not used systematically until 2002 in patients with type 2 diabetes with macroalbuminuria.
Materials and methods
ESRD was defined as serum creatinine >500 μmol/l, the need for dialysis or renal transplantation. In all patients included in the study, we obtained information on survival status/date of death in January 2005 from the Danish death registry. Death certificates were obtained to establish the cause of death. In patients who died with ESRD, the cause of death was coded as ESRD irrespective of the cause of death on the death certificate.
Diabetic retinopathy was assessed by direct ophthalmoscopy until 1988, thereafter by fundus photography, both after pupillary dilatation. The degree of retinopathy was classified as nil, simplex or proliferative. Diabetic maculopathy was graded as simplex retinopathy.
GFR was measured after a single intravenous injection of 51Cr–EDTA (3.7 MBq) at 08.30 h by following the plasma clearance of the tracer for 4 h [17]. The results were standardised for 1.73 m2 body surface area. The surface area calculated at the initial GFR measurement was used to standardise all subsequent GFR measurements for a body surface area of 1.73 m2.
HbA1c, serum creatinine, cholesterol and haemoglobin were measured using standard laboratory techniques with unchanged reference intervals over the study period, as described in detail previously [14].
Albuminuria was quantified by radioimmunoassay from 1983 to 1990 (sensitivity 0.5 mg/l, coefficient of variation 9%), [18] and from 1990 to 1997 by enzyme immunoassay (sensitivity 1.1 mg/l, coefficient of variation 8%) [19]. A close correlation between radioimmunoassay and enzyme immunoassay (r=0.99) was documented before changing the methods. From 1997, the DAKO Turbidimetric method was used to measure urinary albumin excretion rate. This method is closely correlated with enzyme immunoassay (r=0.99) and has a coefficient of variation of 5%.
Blood pressure was measured in the sitting position after approximately 5–10-min rest using a standard mercury sphygmomanometer and an appropriate cuff size.
Smoking history was assessed from the patient’s records. Patients were classified as smokers if they smoked more than one cigarette per day. The study was approved by the local ethics committee and was conducted in accordance with the Helsinki declaration.
Statistical analysis
Results are expressed as mean (±SE) unless otherwise stated. Albuminuria was logarithmically transformed before analysis because of its skewed distribution and is given as geometric mean with 95% confidence intervals. Baseline values of clinical variables were calculated as mean values during the first year after onset of NRA, and clinical values at end of follow-up were calculated as mean values during the last year of observation. Comparisons were performed with Student’s t-test and Fisher’s exact test. A Cox proportional hazard regression analysis with remission as endpoint was performed to evaluate baseline predictors of subsequent remission. In this analysis, time from NRA was used as timescale. Age, sex, and baseline albuminuria, HbA1c and systolic blood pressure were included as covariates. Logistic regression with backwards selection was performed to evaluate factors that were associated with remission of NRA during follow-up. To evaluate baseline predictors and the impact of remission on ESRD and death, two other Cox proportional hazard regression analyses were performed: one with the composite endpoint of ESRD or death, the other with only death as endpoint. In both analyses, the time since NRA was used as the timescale. Based on a previous study of risk factors for progression of renal disease and mortality in type 2 diabetes at the Steno Diabetes Center [14], we decided a priori to include the following fixed baseline covariates in both models: age, sex, HbA1c and systolic blood pressure. Remission of NRA was entered as a time-dependent covariate with value 0 up to the time of remission and after that a value of 1. For patients with onset of NRA within the first 3 years after onset of nephropathy, delayed entry was used, because the inclusion criterion required at least 3 years of follow-up with nephropathy. Linear regression analysis by least-squares method was used to determine the GFR for each patient. Statistical analyses were performed using SPPS 12.0 (SPSS, Chicago, IL, USA) and using the freely available software R 2.0.1 (http://www.r-project.org). A p-value of less than 0.05 was considered statistically significant.