- Email: [email protected]
Progression of diabetic nephropathy
Kidney International, Vol. 59 (2001), pp. 702–709
Progression of diabetic nephropathy PETER HOVIND, PETER ROSSING, LISE TARNOW, ULLA M. SMIDT, and HANS-HENRIK PARVING Steno Diabetes Center, Gentofte, Denmark
Progression of diabetic nephropathy. Background. Diabetic nephropathy is a major cause of renal failure. The decline in glomerular filtration rate (GFR) is highly variable, ranging from 2 to 20, with a median of 12 mL/min/ year. The risk factors of losing filtration power (progression promoters) have not been clearly identified. Furthermore, information on optimal arterial blood pressure, glycemic control, and cholesterol levels are lacking. Methods. We measured GFR with 51Cr-EDTA plasma clearance technique, blood pressure, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol every year for seven years (range 3 to 14 years) in 301 consecutive type 1 diabetic patients with diabetic nephropathy recruited consecutively during 1983 through 1997. Diabetic nephropathy was diagnosed clinically if the following criteria were fulfilled: persistent albuminuria ⬎200 g/min, presence of diabetic retinopathy, and no evidence of other kidney or renal tract disease. In total, 271 patients received antihypertensive treatment at the end of the observation period. Results. Mean arterial blood pressure was 102 ⫾ 0.4 (SE) mm Hg. The average decline in GFR was 4.0 ⫾ 0.2 mL/min/year and even lower (1.9 ⫾ 0.5 mL/min/year) in the 30 persistently normotensive patients, none of whom had ever received antihypertensive treatment (P ⬍ 0.01). A multiple linear regression analysis revealed a significant positive correlation between the decline in GFR and mean arterial blood pressure, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol during follow-up (R2adj ⫽ 0.29, P ⱕ 0.001). No threshold level for blood pressure, glycosylated hemoglobin A1c, or serum cholesterol was demonstrated. A two-hit model with mean arterial blood pressure and glycosylated hemoglobin A1c below and above the median values (102 mm Hg and 9.2%, respectively) revealed a rate of decline in GFR of only 1.5 mL/min/year in the lowest stratum compared with 6.1 mL/min/year in the highest stratum (P ⬍ 0.001). Conclusions. The prognosis of diabetic nephropathy has improved during the past decades, predominantly because of effective antihypertensive treatment. Genuine normotensive patients have a slow progression of nephropathy. Several modifiable variables have been identified as progression promoters.
Diabetic nephropathy is a clinical syndrome characterized by persistent albuminuria, arterial blood pressure elevation, a relentless decline in glomerular filtration rate (GFR), and a high risk of cardiovascular morbidity and mortality [1]. This major life-threatening complication develops in approximately 35% of type 1 diabetic patients [2, 3]. Diabetic nephropathy is a leading cause of end-stage renal disease. However, the decline in GFR is highly variable, ranging from 2 to 20 mL/min/year [4–6]. The risk factors for losing filtration power, that is, the so-called progression promoters, have not been studied extensively. Hypertension and proteinuria contribute to the progressive loss of renal function, while other progression promoters, for example, glycemic control and lipids, are still debatable as reviewed by Rossing [7]. The identification of progression promoters is important for the creation of new powerful treatment modalities impeding the development of end-stage renal disease. The aim of our prospective observational study, started in 1983, was to evaluate the impact of several putative progression promoters in a consecutive cohort of 301 type 1 diabetic patients with diabetic nephropathy, who had serial measurements of glomerular filtration performed and at least three years of follow-up. In addition, we strived to obtain information on optimal arterial blood pressure, glycemic control, and cholesterol levels in relationship to loss of kidney function. METHODS Patients Since 1983, at Steno Diabetes Center, all type 1 diabetic patients with nephropathy have had their kidney function monitored with one yearly determination of GFR. The data are kept in a diabetic nephropathy registry, from which we included all type 1 diabetic patients who had a minimum of three years of follow-up (N ⫽ 301). The recruitment period for our prospective cohort study ended in 1997. Diabetic nephropathy was diagnosed clinically if the following criteria were fulfilled: persistent albuminuria ⬎200 g/min in at least two out of three consecutive 24-hour urine collections, presence of diabetic retinopathy, and the absence of any clinical or
Key words: renal failure, proteinuria, glycemic control, type 1 diabetes mellitus, antihypertensive treatment, blood pressure control, glomerular filtration rate, diabetic nephropathy. Received for publication May 11, 2000 and in revised form August 29, 2000 Accepted for publication September 1, 2000
2001 by the International Society of Nephrology
702
Hovind et al: Progression of diabetic nephropathy
laboratory evidence of other kidney or renal tract disease [8]. All patients took at least two daily injections of insulin and had a diabetic diet containing 45 to 55% carbohydrates, 30 to 35% fat, and 15 to 20% protein. No sodium or protein restrictions were applied during the study. A total of 271 patients were treated with antihypertensive medication during the whole or the major part of the follow-up period, 179 patients predominantly with angiotensin-converting enzyme (ACE) inhibitors. Patients were classified as taking a class of antihypertensive agent if the drug was prescribed for more than 50% of their follow-up time. Seventeen percent of the patients received monotherapy. Forty-seven percent received two agents. Thirty percent received three agents, and 6% were treated with four or more antihypertensive drugs. A stepped-care approach for antihypertensive treatment was applied. Before 1991, it included selective  blockers, diuretics, and vasodilators. After 1991, it included ACE inhibitors, diuretics, calcium channel blockers (mainly dihydropyridines), and  blockers. During the whole follow-up period, we strived to keep our patients’ blood pressure below 140/90 mm Hg. Thirty patients remained normotensive without antihypertensive treatment before or during the observation period, that is, genuine normotensive patients. The local ethical committee approved the experimental design, and all patients gave their informed consent. Procedures All investigations were made on the same day between 8 a.m. and 1 p.m. The patients had their normal breakfast and usual medication at least one hour before the procedure, during which the patients rested in a supine position. During the investigation period, the patients drank approximately 150 to 200 mL of tap water per hour. The GFR measurement was performed 3 to 24 times (median 8) in each patient during a follow-up period of 3 to 14 years (median 7). GFR was measured after a single intravenous injection of 3.7 MBq 51Cr-EDTA by determination of the radioactivity in venous blood samples taken 180, 200, 220, and 240 minutes after the injection [9]. The mean variability in GFR of each patient from day to day was 4%. Results are standardized for 1.73 m2 body surface, using the patient’s surface area at the start of the study throughout. Albuminuria was measured in timed urine collections, obtained during the four-hour clearance period. Before 1984, the urinary albumin concentration was measured by radial immunodiffusion [10], from 1984 to 1990 by radioimmunoassay [11], and from 1990 it was determined by enzyme immunoassay [12]. The assays had a sensitivity of 2, 0.5, and 1.1 mg/L and a coefficient of variation of 8, 9, and 8%, respectively. A very close correlation between radial immunodiffusion and radioimmunoassay (r ⫽ 0.98) [13] and radioimmunoassay and enzyme im-
703
munoassay (r ⫽ 0.99) [12] was documented before changing the methods. The method used for measurement of glycosylated hemoglobin A1c from venous blood samples has also changed over the years: From 1983 to 1988, high-performance liquid chromatography (HPLC) ion exchange [14] and isoelectric focusing [15] were used. These methods have a normal range of 4.1 to 6.1% and 4.1 to 6.4%. Thereafter, glycosylated hemoglobin A1c was measured with HPLC (Bio-Rad Diamat, Richmond, CA, USA) with a normal range of 4.1 to 6.4%. The correlations between this method and the two previous methods were r ⫽ 0.983 (N ⫽ 194) and r ⫽ 0.931 (N ⫽ 119), respectively. Finally, from 1994, an HPLC-based method was used (Bio-Rad Variant). The normal range remained unchanged (4.1 to 6.4%), and the coefficient of correlation between the latter and present method was r ⫽ 0.993 (N ⫽ 161). Serum cholesterol was measured with standard laboratory techniques, and the method was unchanged during the study. Arterial blood pressure was measured at each visit with a standard mercury sphygmomanometer and appropriate cuff size. The measurements were performed twice, on the right arm, after at least 10 minutes of rest in the supine position and were averaged. Diastolic blood pressure was recorded at the disappearance of Korotkoff sounds (phase V). Arterial hypertension was diagnosed according to the World Health Organization’s criteria (ⱖ160/95 mm Hg) until 1995 and thereafter according to the American Diabetes Associations criteria ⱖ140/90 mm Hg [16]. All patients visited the outpatient clinic every three to four months during the study. Blood glucose concentration, glycosylated hemoglobin A1c, albuminuria, blood pressure, and body weight were monitored, and the insulin dose and antihypertensive treatment were adjusted. Retinopathy was assessed after pupillary dilation by ophthalmoscopy and from 1991 by fundus photography and graded: nil, simplex, or proliferative diabetic retinopathy. Statistical analyses Results are expressed as means and SE, means and SD being used for descriptive information. Albuminuria is given as median (range) and logarithmically transformed before analysis because of the positively skewed distribution. In each patient, all measurements performed during the entire follow-up period were used to calculate the mean values. A linear regression analysis, least-square method, was used to determine the rate of decline in GFR for each patient. In normally distributed variables, a comparison between groups was performed by an unpaired Student t test. In non-normally distributed continuous variables, a Mann–Whitney test was used for comparison between groups. A multiple linear regression analysis with backward selection was performed with decline in GFR as
704
Hovind et al: Progression of diabetic nephropathy
Table 1. Characteristics of type 1 diabetic patients suffering from diabetic nephropathy at baseline, and clinical and laboratory data during the follow-up period
Baseline Sex M/F Age yearsb Age at onset yearsb Duration of diabetes yearsb Height cmb Retinopathy simplex/proliferative Smoking yes/no Systolic blood pressure mm Hgb Diastolic blood pressure mm Hgb Mean arterial blood pressure mm Hgb Albuminuria lg/minc Glycosylated hemoglobin A1c %b Glomerular filtration rate mL/min/1.73 m2 During follow-up periodd Rate of decline in glomerular filtration rate mL/min/yeare Systolic blood pressure mm Hge Diastolic blood pressure mm Hge Mean arterial blood pressure mm Hge Albuminuria lg/minc Glycosylated hemoglobin A1c %e Serum cholesterol mmol/Le Observation time yearsc
Total N ⫽ 301
No antihypertensive therapy N ⫽ 30
Antihypertensive therapya N ⫽ 271
192/109 36 ⫾ 11 14 ⫾ 9 22 ⫾ 8 172 ⫾ 8 99/202 164/137 140 ⫾ 19 85 ⫾ 9 103 ⫾ 11 629 (45–6177) 9.3 ⫾ 1.4 89 ⫾ 28
16/14 35 ⫾ 10 13 ⫾ 9 22 ⫾ 7 173 ⫾ 8 15/15 20/10 129 ⫾ 14 79 ⫾ 9 96 ⫾ 9 368 (204–1694) 8.9 ⫾ 1.2 106 ⫾ 20
176/95 36 ⫾ 11 14 ⫾ 9 22 ⫾ 8 172 ⫾ 9 84/187h 144/127 141 ⫾ 19f 86 ⫾ 9f 104 ⫾ 10f 691f (45–6177) 9.4 ⫾ 1.5 87 ⫾ 28f
4.0 140 82 102 557 9.3 5.7 6.7
1.9 131 77 95 384 8.9 5.0 7.0
4.3 (0.3)g 141 (0.9)f 83 (0.4)f 102 (0.5)f 585 (13–6939)h 9.3 (0.1)h 5.8 (0.1)f 6.6 (3.0–14.0)
(0.2) (0.8) (0.5) (0.4) (13–6939) (0.1) (0.1) (3.0–14.0)
(0.5) (1.9) (1.0) (1.1) (124–2358) (0.2) (0.1) (3.4–13.3)
Some patients with previously persistent albuminuria receiving antihypertensive treatment had baseline albuminuria below 200 g/min. a These patients received antihypertensive treatment during the whole or the major part of the follow-up period Data are: b means ⫾ SD, c median (range), e means (SE) d In each patient, all measurements performed during the entire follow-up period were used to calculate mean/median values f P ⬍ 0.001 as compared with patients receiving no antihypertensive therapy g P ⱕ 0.01 as compared with patients receiving no antihypertensive therapy h P ⬍ 0.05 as compared with patients receiving no antihypertensive therapy
dependent variable, including all variables with P ⬍ 0.10 in univariate analyses. A two-factor analysis of variance was used to evaluate the two hit models. The calculations were performed with a commercially available program, SPSS 8.0 (SPSS Inc., Chicago, IL, USA). To evaluate whether there were nonlinear effects of the variables on the decline in GFR, we extended the linear regression model to a model with one change point for each variable at a time. By varying the change point, we could estimate the position of the change point (threshold) and test whether inclusion of a change point improved the model [17]. These calculations were performed by the freely available software R (http://www.ci.tuwien. ac.at/R). A P value ⬍0.05 (two-sided) was considered statistically significant. RESULTS The characteristics of the patients at baseline are shown in Table 1. The demographic and clinical data of the patients in the genuine normotensive group and in the group receiving antihypertensive treatment were alike. However, the latter group had higher arterial blood pressure and albuminuria and lower GFR (P ⬍ 0.001). During the investigation period of seven (range
3 to 14) years, the mean rate of decline in GFR was 4.0 ⫾ 0.2 mL/min/year. The 30 genuine normotensive patients had a significantly lower rate of decline in glomerular filtration as compared with the hypertensive patients (1.9 ⫾ 0.5 vs. 4.3 ⫾ 0.3 mL/min/year, respectively, P ⬍ 0.01). However, the mean arterial blood pressure was also significantly lower in the normotensive group compared with the patients receiving antihypertensive medication (95 ⫾ 1 vs. 102 ⫾ 0.5 mm Hg, P ⬍ 0.001). In addition, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol were lower in the normotensive patients (P ⬍ 0.05). When comparing the 30 genuine normotensive patients with previously hypertensive patients, who on antihypertensive treatment obtained the same level of blood pressure (N ⫽ 102), there was no difference in the decline in kidney function (1.9 ⫾ 0.5 vs. 2.5 ⫾ 0.3 mL/min/year, respectively, NS). The rate of decline in GFR in patients predominantly treated with ACE inhibitors versus patients mainly treated with other blood pressure-lowering agents did not differ significantly (4.5 ⫾ 0.3 vs. 3.8 ⫾ 0.4 mL/min/year, respectively, NS). Furthermore, there was no difference in mean arterial blood pressure values between patients predominantly treated with ACE inhibitors as compared with patients mainly treated with other blood pressure-lowering agents (103 ⫾ 0.6 vs. 102 ⫾ 0.7 mm Hg, NS). No statistical
705
Hovind et al: Progression of diabetic nephropathy Table 2. Multiple linear regression analysis of progression promoters in type 1 diabetic patients suffering from diabetic nephropathy Variable
Slope (95% CI)
P value
Dependent variable Rate of decline in GFR mL/min/year Independent variables Mean arterial blood pressure per 10 mm Hg Albuminuria log10 Hemoglobin A1c % Serum cholesterol mmol/L
1.18 2.04 0.73 0.66
⬍ 0.001 ⬍ 0.001 ⬍ 0.001 0.001
(0.59–1.77) (1.07–3.01) (0.36–1.10) (0.27–1.05)
R2 adj.: 0.29 Not included in the model are age, sex, height, grade of retinopathy, smoking, and class of antihypertensive treatment.
significant differences in any of the other progression promoters were demonstrated between these two groups (data not shown). A univariate analysis of demographic, clinical, and laboratory parameters revealed a significant positive correlation between the decline in GFR (dependent variable) and systolic (r ⫽ 0.27, P ⬍ 0.001), diastolic (r ⫽ 0.37, P ⬍ 0.001) and mean arterial blood pressure (r ⫽ 0.39, P ⬍ 0.001), albuminuria (r ⫽ 0.41, P ⬍ 0.001), glycosylated hemoglobin A1c (r ⫽ 0.30, P ⬍ 0.001), and serum cholesterol (r ⫽ 0.32, P ⬍ 0.001) during the study. Men had a significantly higher rate of decline in GFR than women (4.4 ⫾ 0.3 vs. 3.4 ⫾ 0.4 mL/min/year, P ⬍ 0.04). After adjustment for other progression promoters, the rate of decline in GFR was similar in men (4.1 ⫾ 0.3 mL/min/year) and in women (4.0 ⫾ 0.4 mL/min/year, NS). Smoking, degree of retinopathy, height, age, class of antihypertensive treatment, and baseline GFR had no predictive value. When these variables were examined in combination by multiple linear regression analysis, the significant predictors of decline in GFR were high values of mean arterial blood pressure, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol (r ⫽ 0.54, r2adj ⫽ 0.29, P ⱕ 0.001; Table 2). The relationship between the rate of decline in GFR and each of the four previously mentioned progression promoters separated in quintiles is shown in Figure 1. Except for albuminuria (threshold 436 g/min, P ⬍ 0.01), none of the other potentially modifiable progression promoters showed a significant threshold, suggesting that the best fit of the data is linear or curve linear. A two-hit model with mean arterial blood pressure (previously demonstrated to be a modifiable progression promoter) combined with each of the three previouslymentioned potentially modifiable risk factors for losing filtration power is shown in Figure 2. Of the 271 patients treated with antihypertensive agents, 139 patients (51%; 95% CI, 45 to 57) achieved a systolic blood pressure below 140 mm Hg during follow-up. In 234 patients (86%; 82 to 90%), diastolic blood pressure was below 90 mm Hg, whereas 128 patients (47%; 41 to 53%) achieved a blood pressure below 140
mm Hg systolic and 90 mm Hg diastolic. Correspondingly, 202 patients (75%; 69 to 80%) had a mean arterial blood pressure below 107 mm Hg during follow-up. DISCUSSION Our long-term prospective observational study in 301 type 1 diabetic patients suffering from diabetic nephropathy revealed that the prognosis of diabetic nephropathy has improved during the past decades. Genuine normotensive patients have a slow rate of decline in GFR. Furthermore, several modifiable variables—that is, arterial blood pressure, albuminuria, glycemic control, and serum cholesterol—act as progression promoters. Finally, except for albuminuria, no statistical threshold effect on the decline in GFR of any of the identified progression promoters was demonstrated. To minimize selection bias and maximize generalizability, all type 1 diabetic patients with diabetic nephropathy according to the well-defined criteria [8] who had a minimum of three years of follow-up on their kidney function were enrolled in our prospective cohort study. To obtain a valid determination of the rate of decline in GFR the following requirements should be fulfilled: The applied GFR method should have good accuracy and precision, and repeated measurements of GFR (⬃every 6 to 12 months) should be performed. The observation period should be extended to at least two years [18]. These requirements were fulfilled in our study. A major strength in prospective cohort studies (such as ours) is a long-term follow-up period of an unselected, well-defined group of patients, which maximize the generalizability of the findings. The limitations are that the identified progression promoters can only be documented as modifiable risk factors in randomized clinical intervention trials, and interpretation of the impact of different treatment modalities is limited, since different categories of patients can receive different therapy. The natural history of diabetic nephropathy is characterized by arterial blood pressure elevation and a relentless decline in GFR of approximately 12 mL/min/year [4–6]. Data from studies evaluating the rate of decline
706
Hovind et al: Progression of diabetic nephropathy
Fig. 1. Impact of mean arterial blood pressure, glycosylated hemoglobin A1c, albuminuria, and serum cholesterol on the decline in GFR (P ⬍ 0.001 for each factor).
in GFR are presented in Table 3. While the original studies did not include patients on antihypertensive treatment, such treatment was applied in 90% of our patients, and we demonstrated a much slower rate of decline in GFR (4 mL/min/year). A retrospective analysis of 158 type 1 diabetic patients with nephropathy has revealed results nearly identical to the present findings [19]. Several prospective antihypertensive treatment trials have documented a beneficial effect on albuminuria, rate of decline in GFR, progression to end-stage renal disease, and survival in type 1 diabetic patients suffering from diabetic nephropathy [8, 20–25]. Thus, arterial blood pressure seems to have a rather complex relationship with diabetic nephropathy: nephropathy raising blood pressure and blood pressure accelerating the course of nephropathy. This acceleration may be at least partly due to impaired/abolished autoregulation of GFR [26, 27], leading to enhanced transmission of systemic blood pressure to
the glomerular capillaries and therefore to glomerular capillary hypertension. Studies in experimental diabetic animals [28, 29] and more recently in humans [30] have documented intrarenal hypertension, and have stressed the importance of this hemodynamic phenomenon in the initiation and progression of diabetic and nondiabetic glomerulopathies [31]. Impaired autoregulation of glomerular pressure and systemic blood pressure elevation induce distention– contraction of glomeruli. These alterations are associated with mesangial cell stretch, that in turn stimulates the synthesis and deposition of extracellular matrix material leading to mesangial expansion and glomerulosclerosis [32, 33]. In diabetic glomeruli, the mechanical stretch effect on extracellular matrix material synthesis is markedly aggravated by the presence of hyperglycemia [33]. Proteinuria is usually considered a marker of the extent of glomerular damage, but studies in various experi-
Hovind et al: Progression of diabetic nephropathy
Fig. 2. The decline in GFR divided according to median of mean arterial blood pressure combined with median albuminuria, glycosylated hemoglobin A1c, and serum cholesterol. Within each combination of risk factors, a comparison of the four different strata revealed a significant difference (P ⬍ 0.001).
mental animal models suggest that proteinuria per se may contribute to glomerular and tubulointerstitial lesions [34]. Our study suggests that albuminuria is an independent risk factor for progression of diabetic nephropathy, a finding that confirms and extends previous results in normotensive and hypertensive type 1 and type 2 diabetic patients with nephropathy [19, 35–39]. Similarly, proteinuria is a progression promoter in nondiabetic nephropathies [40, 41]. Furthermore, intervention that has ameliorated the progression of diabetic nephropathy has always been associated with a reduction in proteinuria as reviewed by Parving [42]. Finally, it should be recalled that initial reduction in albuminuria is the best clinical guideline predicting a beneficial effect of antihyperten-
707
sive treatment on the long-term decline in kidney function in diabetic and nondiabetic nephropathies [43–45]. Development of overt diabetic nephropathy has for many years been regarded as “a point of no return” in relationship to glycemic control as reviewed by Rossing [7]. However, this statement is based on studies using inappropriate methods for monitoring kidney function (serum creatinine), long-term glycemic control (blood glucose), and often, very few patients were studied. Applying better research techniques and in some studies a larger number of type 1 diabetic patients, a significant impact of glycemic control on progression of nephropathy has been found [19, 37, 46], as also documented in our study. Furthermore, pancreas transplantation can reverse the lesions of diabetic glomerulopathy, but reversal requires more than five years of normoglycemia [47]. Our study demonstrated that elevated serum cholesterol acts as an independent progression promoter in diabetic nephropathy. This finding supports the concept advocated by Moorhead et al [48] that hyperlipidemia promotes progression in chronic renal diseases once an initial event has damaged the glomerular capillary wall, thereby allowing passage of lipids and lipoproteins into the mesangium. Several short-term studies dealing with a limited number of diabetic patients have failed to demonstrate a beneficial effect of lipid lowering treatment on the surrogate endpoint: albuminuria [7]. Long-term trials applying a principal endpoint such as the decline in GFR are still lacking. A renoprotective effect of ACE inhibitors above and beyond the effect of blood pressure lowering has been demonstrated in some clinical trials [23, 24], while others have failed to demonstrate an additional non-hemodynamic effect of ACE inhibition [abstract; Tarnow et al, Diabetologia 42(Suppl 1):A275, 1999] [25]. In our longterm prospective observational study, we could not demonstrate a renoprotective effect of treatment with ACE inhibitors in comparison with other antihypertensive treatment modalities. However, the present study was not designed to evaluate this concept since ACE inhibitors were not available until 1989. In addition, as our study did not use a randomized design, different patient categories could in fact receive different antihypertensive treatment modalities. Previous studies have documented a relationship between the degree of glomerular lesions and GFR and progression of diabetic nephropathy in type 1 and type 2 diabetic patients [49–51]. Furthermore, a progressive loss of intrinsic ultrafiltration capacity has been shown to be the predominant cause of the declining GFR in type 2 diabetic patients with nephropathy [39]. The combined contribution from the previously identified renal structural lesions and the present demonstrated progression promoters explains the vast majority of progression in diabetic nephropathy. Unfortunately, the role of genetic
708
Hovind et al: Progression of diabetic nephropathy Table 3. Observational studies and clinical trials on progression of diabetic nephropathy in type 1 diabetic patients with and without antihypertensive treatment (AHT)
Authors Mogensen et al [4] Parving et al [5] Viberti et al [6] Bjo¨rck et al [23] Lewis et al [24] Elving et al [25] Tarnow et ale Mulec et al [19] Present study
Study design
N
Follow-up years
Observational, retrospective Observational, prospective Observational, retrospective Randomized trial Randomized trial Randomized trial Randomized trial Observational, retrospective Observational, prospective
10 14 13 40 409 29 48 158 301
2.8 2.2 2.5 2.2 2.7 2.0 4.0 8.0 6.7
⌬GFR mL/min/year
Mean arterial blood pressure mm Hg
No AHT
NonACE-I
ACE-I
No AHT
NonACE-I
ACE-I
10.9 9.0 14.4 — — — — — 1.9
— — — 5.6 10.8c 3.7 5.5
— — — 2.0 8.0c 4.9 6.5
121a 111 109b — — —
— — — 103 100 101d 103
— — — 102 96 100d 100
3.8f 3.8
4.5
— 95
102f 102
103
Glomerular filtration rate (GFR) determination was based on renal or plasma clearance of filtration markers. Blood pressure measurements were based on mean of values during follow-up. a MABP at end of the study b MABP at entry to the study c Creatinine clearance, GFR calculated from percentage decrease per year using the formula: y ⫽ y0 ⫻ exp(⫺kt), decline in GFR per year from baseline to mean follow up of 2.7 years d MABP during last 6 months of follow up e Abstract [Tarnow et al., Diabetologia 42(Suppl 1): pA275, 1999] f Antihypertensive treatment varied throughout the follow-up period
factors cannot be evaluated in our study, since systematical collection of DNA first began in 1993. The available data suggest that multifactorial intervention targeting blood pressure, albuminuria, glycemic control, and hyperlipidemia will diminish progression of nephropathy and may even induce regression of diabetic nephropathy, that is, a loss in GFR equal to the natural aging process (⬃1 mL/min/year). Recently, the success of such a combined approach in delaying progression of diabetic microangiopathy has been demonstrated in type 2 diabetic patients with microalbuminuria [52]. It should be mentioned that there appear to be no substantial differences between patients with type 2 diabetes and those with type 1 diabetes with respect to initiation, progression, and treatment of diabetic nephropathy [53]. In summary, the prognosis of diabetic nephropathy has improved, probably because of effective antihypertensive treatment. Genuine normotensive patients have a very slow progression of nephropathy. Several modifiable factors, that is, arterial blood pressure, albuminuria, glycemic control, and lipids, have been identified as progression promoters. Except for albuminuria, no thresholds for the remaining risk factors were demonstrated. ACKNOWLEDGMENTS The Paul and Erna Sehested Hansen Foundation, The Per S. Henriksen Foundation, and the Danish Diabetes Association are gratefully thanked for financial support. We acknowledge the assistance of Berit R. Jensen, Birgitte V. Hansen, and Inge-Lise Rossing. Reprint requests to Peter Hovind, M.D., Steno Diabetes Center, Niels Steensens Vej 2 DK-2820 Gentofte, Denmark. E-mail: [email protected]
REFERENCES 1. Parving H, Østerby R, Ritz E: Diabetic nephropathy, in The Kidney, edited by Brenner BM, Levine S, Philadelphia, W.B. Saunders, 2000, p 1731 2. Andersen AR, Christiansen JS, Andersen JK, et al: Diabetic nephropathy in type 1 (insulin-dependent) diabetes: An epidemiological study. Diabetologia 25:496–501, 1983 3. Rossing P, Rossing K, Jacobsen P, et al: Unchanged incidence of diabetic nephropathy in IDDM patients. Diabetes 44:739–743, 1995 4. Mogensen CE: Progression of nephropathy in long-term diabetics with proteinuria and effect of initial antihypertensive treatment. Scand J Clin Lab Invest 36:383–388, 1976 5. Parving H-H, Smidt UM, Friisberg B, et al: A prospective study of glomerular filtration rate and arterial blood pressure in insulindependent diabetics with diabetic nephropathy. Diabetologia 20: 457–461, 1981 6. Viberti GC, Bilous RW, Mackintosh D, et al: Monitoring glomerular function in diabetic nephropathy. Am J Med 74:256–264, 1983 7. Rossing P: Promotion, prediction, and prevention of progression in diabetic nephropathy. Diabet Med 15:900–919, 1998 8. Parving H-H, Andersen AR, Smidt UM, et al: Early aggressive antihypertensive treatment reduces rate of decline in kidney function in diabetic nephropathy. Lancet 1:1175–1179, 1983 9. Bro¨chner-Mortensen J: A simple method for the determination of glomerular filtration rate. Scand J Clin Lab Invest 30:271–274, 1972 10. Mancini G, Carbonara AO, Heremans JF: Immunochemical quantitation of antigens by single radial immunodiffusion. Immunochemistry 2:235–254, 1965 11. Miles DW, Mogensen CE, Gundersen HJG: Radioimmunoassay for urinary albumin using a single antibody. Scand J Clin Lab Invest 26:5–11, 1970 12. Feldt-Rasmussen B, Dinesen B, Deckert M: Enzyme immunoassay: An improved determination of urinary albumin in diabetics with incipient nephropathy. Scand J Clin Lab Invest 45:539–544, 1985 13. Rossing P, Hommel E, Smidt UM, et al: Impact of arterial blood pressure and albuminuria on the progression of diabetic nephropathy in IDDM patients. Diabetes 42:715–719, 1993 14. Svendsen PA, Christiansen JS, Søegaard U, et al: Rapid changes in chromatographically determined haemoglobin A1c induced by short-term changes in glucose concentration. Diabetologia 19:130– 136, 1980
Hovind et al: Progression of diabetic nephropathy 15. Mortensen HB: Quantitative determination of hemoglobin A1c by thin layer isoelectric focusing. J Chromatogr 182:325–333, 1980 16. American Diabetes Association: Treatment of hypertension in diabetes. Diabetes Care 16:1394–1401, 1993 17. Zoccali C, Postorino M, Martorano C, et al: The “breakpoint” test, a new statistical method for studying progression of chronic renal failure. Nephrol Dial Transplant 4:101–104, 1989 18. Levey AS, Gassman J, Hall PM, et al: Assessing the progression of renal disease in clinical studies: Effects of duration of followup and regression to the mean. J Am Soc Nephrol 1:1087–1094, 1991 19. Mulec H, Blohme´ G, Gra¨nde B, Bjo¨rck S: The effect of metabolic control on rate of decline in renal function in insulin-dependent diabetes mellitus with overt diabetic nephropathy. Nephrol Dial Transplant 13:651–655, 1998 20. Mogensen CE: Long-term antihypertensive treatment inhibiting progression of diabetic nephropathy. Br Med J 285:685–688, 1982 21. Parving H-H, Andersen AR, Smidt UM, et al: Effect of antihypertensive treatment on kidney function in diabetic nephropathy. Br Med J 294:1443–1447, 1987 22. Parving H-H, Hommel E: Prognosis in diabetic nephropathy. Br Med J 299:230–233, 1989 23. Bjo¨rck S, Mulec H, Johnsen SA, et al: Renal protective effect of enalapril in diabetic nephropathy. Br Med J 304:339–343, 1992 24. Lewis E, Hunsicker L, Bain R, et al: The effect of angiotensinconverting-enzyme inhibition on diabetic nephropathy. N Engl J Med 329:1456–1462, 1993 25. Elving LD, Wetzels JFM, Van Lier HJJ, et al: Captopril and atenolol are equally effective in retarding progression of diabetic nephropathy. Diabetologia 37:604–609, 1994 26. Parving H-H, Kastrup J, Smidt UM, et al: Impaired autoregulation of glomerular filtration rate in type 1 (insulin-dependent) diabetic patients with nephropathy. Diabetologia 27:547–552, 1984 27. Christensen PK, Hansen HP, Parving H-H: Impaired autoregulation of GFR in hypertensive non-insulin dependent diabetic patients. Kidney Int 52:1369–1374, 1997 28. Hostetter TH, Troy JL, Brenner BM: Glomerular hemodynamics in experimental diabetes mellitus. Kidney Int 19:410–415, 1981 29. Jensen PK, Steven K, Blæhr H, et al: The effects of indomethacin on glomerular hemodynamics in experimental diabetes. Kidney Int 29:490–495, 1986 30. Imanishi M, Yoshioka K, Konishi Y, et al: Glomerular hypertension as one cause of albuminuria in type II diabetic patients. Diabetologia 42:999–1005, 1999 31. Hostetter TH, Rennke HG, Brenner BM: The case for intrarenal hypertension in the initiation and progression of diabetic and other glomerulopathies. Am J Med 72:375–380, 1982 32. Riser BL, Cortes P, Zhao X, et al: Intraglomerular pressure and mesangial stretching stimulate extracellular matrix formation in the rat. J Clin Invest 90:1932–1943, 1992 33. Cortes P, Riser BL, Yee J, et al: Mechanical strain of glomerular mesangial cells in the pathogenesis of glomerulosclerosis: Clinical implications. Nephrol Dial Transplant 14:1351–1354, 1999 34. Remuzzi G, Bertani T: Is glomerulosclerosis a consequence of altered glomerular permeability to macromolecules? Kidney Int 38:384–394, 1990 35. Elving LD, Wetzels JFM, De Nobel E, et al: Erythrocyte sodiumlithium countertransport is not different in type 1 (insulin-depen-
36. 37. 38. 39. 40. 41.
42. 43.
44. 45.
46. 47. 48. 49. 50. 51. 52.
53.
709
dent) diabetic patients with and without diabetic nephropathy. Diabetologia 34:126–128, 1991 Breyer JA, Bain P, Evans JK, et al: Predictors of the progression of renal insufficiency in patients with insulin-dependent diabetes and overt diabetic nephropathy. Kidney Int 50:1651–1658, 1996 Parving H-H, Smidt UM, Hommel E, et al: Effective antihypertensive treatment postpones renal insufficiency in diabetic nephropathy. Am J Kidney Dis 22:188–195, 1993 Gall M-A, Nielsen FS, Smidt UM, et al: The course of kidney function in type 2 (non-insulin-dependent) diabetic patients with diabetic nephropathy. Diabetologia 36:1071–1078, 1993 Nelson RG, Bennett PH, Beck GJ, et al: Development and progression of renal disease in Pima Indians with non-insulin-dependent diabetes mellitus. N Engl J Med 335:1636–1642, 1996 Peterson JC, Adler S, Burkart JM, et al: Blood pressure control, proteinuria, and the progression of renal disease: The modification of diet in renal disease study. Ann Intern Med 123:754–762, 1995 The Gisen Group: Randomised placebo-controlled trial of effect of ramipril on decline in glomerular filtration rate and risk of terminal renal failure in proteinuric, non-diabetic nephropathy. Lancet 349:1857–1863, 1997 Parving H-H: Renoprotection in diabetes: Genetic and nongenetic risk factors and treatment. Diabetologia 41:745–759, 1998 Rossing P, Hommel E, Smidt UM, et al: Reduction in albuminuria predicts a beneficial effect on diminishing the progression of human diabetic nephropathy during antihypertensive treatment. Diabetologia 37:511–516, 1994 Rossing P, Hommel E, Smidt UM, et al: Reduction in albuminuria predicts diminished progression in diabetic nephropathy. Kidney Int 45(Suppl 45):S145–S149, 1994 Apperloo AJ, De Zeeuw D, De Jong PE: Short-term antiproteinuric response to antihypertensive treatment predicts long-term GFR decline in patients with non-diabetic renal disease. Kidney Int 45(Suppl 45):S174–S178, 1994 Nyberg G, Blohme´ G, Norde´n G: Impact of metabolic control in progression of clinical diabetic nephropathy. Diabetologia 30:82–86, 1987 Fioretto P, Steffes MW, Sutherland DER, et al: Reversal of lesions of diabetic nephropathy after pancreas transplantation. N Engl J Med 339:69–75, 1998 Moorhead JF, El-Nahas M, Chan MK, et al: Lipid nephrotoxicity in chronic progressive glomerular and tubulo-interstitial disease. Lancet 2:1309–1311, 1982 Mauer SM, Steffes MW, Ellis EN, et al: Structural-functional relationships in diabetic nephropathy. J Clin Invest 74:1143–1155, 1984 Østerby R, Parving H-H, Hommel E, et al: Glomerular structure and function in diabetic nephropathy -early to advanced stages. Diabetes 39:1057–1063, 1990 Østerby R, Gall M-A, Schmitz A, et al: Glomerular structure and function in proteinuric type 2 (non-insulin-dependent) diabetic patients. Diabetologia 36:1064–1070, 1993 Gæde P, Vedel P, Parving H-H, et al: Intensified multifactorial intervention in patients with type 2 diabetes mellitus and microalbuminuria: The Steno type 2 randomised study. Lancet 353:617– 622, 1999 Parving H-H: Initiation and progression of diabetic nephropathy. N Engl J Med 335:1682–1683, 1996
Progression of diabetic nephropathy PETER HOVIND, PETER ROSSING, LISE TARNOW, ULLA M. SMIDT, and HANS-HENRIK PARVING Steno Diabetes Center, Gentofte, Denmark
Progression of diabetic nephropathy. Background. Diabetic nephropathy is a major cause of renal failure. The decline in glomerular filtration rate (GFR) is highly variable, ranging from 2 to 20, with a median of 12 mL/min/ year. The risk factors of losing filtration power (progression promoters) have not been clearly identified. Furthermore, information on optimal arterial blood pressure, glycemic control, and cholesterol levels are lacking. Methods. We measured GFR with 51Cr-EDTA plasma clearance technique, blood pressure, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol every year for seven years (range 3 to 14 years) in 301 consecutive type 1 diabetic patients with diabetic nephropathy recruited consecutively during 1983 through 1997. Diabetic nephropathy was diagnosed clinically if the following criteria were fulfilled: persistent albuminuria ⬎200 g/min, presence of diabetic retinopathy, and no evidence of other kidney or renal tract disease. In total, 271 patients received antihypertensive treatment at the end of the observation period. Results. Mean arterial blood pressure was 102 ⫾ 0.4 (SE) mm Hg. The average decline in GFR was 4.0 ⫾ 0.2 mL/min/year and even lower (1.9 ⫾ 0.5 mL/min/year) in the 30 persistently normotensive patients, none of whom had ever received antihypertensive treatment (P ⬍ 0.01). A multiple linear regression analysis revealed a significant positive correlation between the decline in GFR and mean arterial blood pressure, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol during follow-up (R2adj ⫽ 0.29, P ⱕ 0.001). No threshold level for blood pressure, glycosylated hemoglobin A1c, or serum cholesterol was demonstrated. A two-hit model with mean arterial blood pressure and glycosylated hemoglobin A1c below and above the median values (102 mm Hg and 9.2%, respectively) revealed a rate of decline in GFR of only 1.5 mL/min/year in the lowest stratum compared with 6.1 mL/min/year in the highest stratum (P ⬍ 0.001). Conclusions. The prognosis of diabetic nephropathy has improved during the past decades, predominantly because of effective antihypertensive treatment. Genuine normotensive patients have a slow progression of nephropathy. Several modifiable variables have been identified as progression promoters.
Diabetic nephropathy is a clinical syndrome characterized by persistent albuminuria, arterial blood pressure elevation, a relentless decline in glomerular filtration rate (GFR), and a high risk of cardiovascular morbidity and mortality [1]. This major life-threatening complication develops in approximately 35% of type 1 diabetic patients [2, 3]. Diabetic nephropathy is a leading cause of end-stage renal disease. However, the decline in GFR is highly variable, ranging from 2 to 20 mL/min/year [4–6]. The risk factors for losing filtration power, that is, the so-called progression promoters, have not been studied extensively. Hypertension and proteinuria contribute to the progressive loss of renal function, while other progression promoters, for example, glycemic control and lipids, are still debatable as reviewed by Rossing [7]. The identification of progression promoters is important for the creation of new powerful treatment modalities impeding the development of end-stage renal disease. The aim of our prospective observational study, started in 1983, was to evaluate the impact of several putative progression promoters in a consecutive cohort of 301 type 1 diabetic patients with diabetic nephropathy, who had serial measurements of glomerular filtration performed and at least three years of follow-up. In addition, we strived to obtain information on optimal arterial blood pressure, glycemic control, and cholesterol levels in relationship to loss of kidney function. METHODS Patients Since 1983, at Steno Diabetes Center, all type 1 diabetic patients with nephropathy have had their kidney function monitored with one yearly determination of GFR. The data are kept in a diabetic nephropathy registry, from which we included all type 1 diabetic patients who had a minimum of three years of follow-up (N ⫽ 301). The recruitment period for our prospective cohort study ended in 1997. Diabetic nephropathy was diagnosed clinically if the following criteria were fulfilled: persistent albuminuria ⬎200 g/min in at least two out of three consecutive 24-hour urine collections, presence of diabetic retinopathy, and the absence of any clinical or
Key words: renal failure, proteinuria, glycemic control, type 1 diabetes mellitus, antihypertensive treatment, blood pressure control, glomerular filtration rate, diabetic nephropathy. Received for publication May 11, 2000 and in revised form August 29, 2000 Accepted for publication September 1, 2000
2001 by the International Society of Nephrology
702
Hovind et al: Progression of diabetic nephropathy
laboratory evidence of other kidney or renal tract disease [8]. All patients took at least two daily injections of insulin and had a diabetic diet containing 45 to 55% carbohydrates, 30 to 35% fat, and 15 to 20% protein. No sodium or protein restrictions were applied during the study. A total of 271 patients were treated with antihypertensive medication during the whole or the major part of the follow-up period, 179 patients predominantly with angiotensin-converting enzyme (ACE) inhibitors. Patients were classified as taking a class of antihypertensive agent if the drug was prescribed for more than 50% of their follow-up time. Seventeen percent of the patients received monotherapy. Forty-seven percent received two agents. Thirty percent received three agents, and 6% were treated with four or more antihypertensive drugs. A stepped-care approach for antihypertensive treatment was applied. Before 1991, it included selective  blockers, diuretics, and vasodilators. After 1991, it included ACE inhibitors, diuretics, calcium channel blockers (mainly dihydropyridines), and  blockers. During the whole follow-up period, we strived to keep our patients’ blood pressure below 140/90 mm Hg. Thirty patients remained normotensive without antihypertensive treatment before or during the observation period, that is, genuine normotensive patients. The local ethical committee approved the experimental design, and all patients gave their informed consent. Procedures All investigations were made on the same day between 8 a.m. and 1 p.m. The patients had their normal breakfast and usual medication at least one hour before the procedure, during which the patients rested in a supine position. During the investigation period, the patients drank approximately 150 to 200 mL of tap water per hour. The GFR measurement was performed 3 to 24 times (median 8) in each patient during a follow-up period of 3 to 14 years (median 7). GFR was measured after a single intravenous injection of 3.7 MBq 51Cr-EDTA by determination of the radioactivity in venous blood samples taken 180, 200, 220, and 240 minutes after the injection [9]. The mean variability in GFR of each patient from day to day was 4%. Results are standardized for 1.73 m2 body surface, using the patient’s surface area at the start of the study throughout. Albuminuria was measured in timed urine collections, obtained during the four-hour clearance period. Before 1984, the urinary albumin concentration was measured by radial immunodiffusion [10], from 1984 to 1990 by radioimmunoassay [11], and from 1990 it was determined by enzyme immunoassay [12]. The assays had a sensitivity of 2, 0.5, and 1.1 mg/L and a coefficient of variation of 8, 9, and 8%, respectively. A very close correlation between radial immunodiffusion and radioimmunoassay (r ⫽ 0.98) [13] and radioimmunoassay and enzyme im-
703
munoassay (r ⫽ 0.99) [12] was documented before changing the methods. The method used for measurement of glycosylated hemoglobin A1c from venous blood samples has also changed over the years: From 1983 to 1988, high-performance liquid chromatography (HPLC) ion exchange [14] and isoelectric focusing [15] were used. These methods have a normal range of 4.1 to 6.1% and 4.1 to 6.4%. Thereafter, glycosylated hemoglobin A1c was measured with HPLC (Bio-Rad Diamat, Richmond, CA, USA) with a normal range of 4.1 to 6.4%. The correlations between this method and the two previous methods were r ⫽ 0.983 (N ⫽ 194) and r ⫽ 0.931 (N ⫽ 119), respectively. Finally, from 1994, an HPLC-based method was used (Bio-Rad Variant). The normal range remained unchanged (4.1 to 6.4%), and the coefficient of correlation between the latter and present method was r ⫽ 0.993 (N ⫽ 161). Serum cholesterol was measured with standard laboratory techniques, and the method was unchanged during the study. Arterial blood pressure was measured at each visit with a standard mercury sphygmomanometer and appropriate cuff size. The measurements were performed twice, on the right arm, after at least 10 minutes of rest in the supine position and were averaged. Diastolic blood pressure was recorded at the disappearance of Korotkoff sounds (phase V). Arterial hypertension was diagnosed according to the World Health Organization’s criteria (ⱖ160/95 mm Hg) until 1995 and thereafter according to the American Diabetes Associations criteria ⱖ140/90 mm Hg [16]. All patients visited the outpatient clinic every three to four months during the study. Blood glucose concentration, glycosylated hemoglobin A1c, albuminuria, blood pressure, and body weight were monitored, and the insulin dose and antihypertensive treatment were adjusted. Retinopathy was assessed after pupillary dilation by ophthalmoscopy and from 1991 by fundus photography and graded: nil, simplex, or proliferative diabetic retinopathy. Statistical analyses Results are expressed as means and SE, means and SD being used for descriptive information. Albuminuria is given as median (range) and logarithmically transformed before analysis because of the positively skewed distribution. In each patient, all measurements performed during the entire follow-up period were used to calculate the mean values. A linear regression analysis, least-square method, was used to determine the rate of decline in GFR for each patient. In normally distributed variables, a comparison between groups was performed by an unpaired Student t test. In non-normally distributed continuous variables, a Mann–Whitney test was used for comparison between groups. A multiple linear regression analysis with backward selection was performed with decline in GFR as
704
Hovind et al: Progression of diabetic nephropathy
Table 1. Characteristics of type 1 diabetic patients suffering from diabetic nephropathy at baseline, and clinical and laboratory data during the follow-up period
Baseline Sex M/F Age yearsb Age at onset yearsb Duration of diabetes yearsb Height cmb Retinopathy simplex/proliferative Smoking yes/no Systolic blood pressure mm Hgb Diastolic blood pressure mm Hgb Mean arterial blood pressure mm Hgb Albuminuria lg/minc Glycosylated hemoglobin A1c %b Glomerular filtration rate mL/min/1.73 m2 During follow-up periodd Rate of decline in glomerular filtration rate mL/min/yeare Systolic blood pressure mm Hge Diastolic blood pressure mm Hge Mean arterial blood pressure mm Hge Albuminuria lg/minc Glycosylated hemoglobin A1c %e Serum cholesterol mmol/Le Observation time yearsc
Total N ⫽ 301
No antihypertensive therapy N ⫽ 30
Antihypertensive therapya N ⫽ 271
192/109 36 ⫾ 11 14 ⫾ 9 22 ⫾ 8 172 ⫾ 8 99/202 164/137 140 ⫾ 19 85 ⫾ 9 103 ⫾ 11 629 (45–6177) 9.3 ⫾ 1.4 89 ⫾ 28
16/14 35 ⫾ 10 13 ⫾ 9 22 ⫾ 7 173 ⫾ 8 15/15 20/10 129 ⫾ 14 79 ⫾ 9 96 ⫾ 9 368 (204–1694) 8.9 ⫾ 1.2 106 ⫾ 20
176/95 36 ⫾ 11 14 ⫾ 9 22 ⫾ 8 172 ⫾ 9 84/187h 144/127 141 ⫾ 19f 86 ⫾ 9f 104 ⫾ 10f 691f (45–6177) 9.4 ⫾ 1.5 87 ⫾ 28f
4.0 140 82 102 557 9.3 5.7 6.7
1.9 131 77 95 384 8.9 5.0 7.0
4.3 (0.3)g 141 (0.9)f 83 (0.4)f 102 (0.5)f 585 (13–6939)h 9.3 (0.1)h 5.8 (0.1)f 6.6 (3.0–14.0)
(0.2) (0.8) (0.5) (0.4) (13–6939) (0.1) (0.1) (3.0–14.0)
(0.5) (1.9) (1.0) (1.1) (124–2358) (0.2) (0.1) (3.4–13.3)
Some patients with previously persistent albuminuria receiving antihypertensive treatment had baseline albuminuria below 200 g/min. a These patients received antihypertensive treatment during the whole or the major part of the follow-up period Data are: b means ⫾ SD, c median (range), e means (SE) d In each patient, all measurements performed during the entire follow-up period were used to calculate mean/median values f P ⬍ 0.001 as compared with patients receiving no antihypertensive therapy g P ⱕ 0.01 as compared with patients receiving no antihypertensive therapy h P ⬍ 0.05 as compared with patients receiving no antihypertensive therapy
dependent variable, including all variables with P ⬍ 0.10 in univariate analyses. A two-factor analysis of variance was used to evaluate the two hit models. The calculations were performed with a commercially available program, SPSS 8.0 (SPSS Inc., Chicago, IL, USA). To evaluate whether there were nonlinear effects of the variables on the decline in GFR, we extended the linear regression model to a model with one change point for each variable at a time. By varying the change point, we could estimate the position of the change point (threshold) and test whether inclusion of a change point improved the model [17]. These calculations were performed by the freely available software R (http://www.ci.tuwien. ac.at/R). A P value ⬍0.05 (two-sided) was considered statistically significant. RESULTS The characteristics of the patients at baseline are shown in Table 1. The demographic and clinical data of the patients in the genuine normotensive group and in the group receiving antihypertensive treatment were alike. However, the latter group had higher arterial blood pressure and albuminuria and lower GFR (P ⬍ 0.001). During the investigation period of seven (range
3 to 14) years, the mean rate of decline in GFR was 4.0 ⫾ 0.2 mL/min/year. The 30 genuine normotensive patients had a significantly lower rate of decline in glomerular filtration as compared with the hypertensive patients (1.9 ⫾ 0.5 vs. 4.3 ⫾ 0.3 mL/min/year, respectively, P ⬍ 0.01). However, the mean arterial blood pressure was also significantly lower in the normotensive group compared with the patients receiving antihypertensive medication (95 ⫾ 1 vs. 102 ⫾ 0.5 mm Hg, P ⬍ 0.001). In addition, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol were lower in the normotensive patients (P ⬍ 0.05). When comparing the 30 genuine normotensive patients with previously hypertensive patients, who on antihypertensive treatment obtained the same level of blood pressure (N ⫽ 102), there was no difference in the decline in kidney function (1.9 ⫾ 0.5 vs. 2.5 ⫾ 0.3 mL/min/year, respectively, NS). The rate of decline in GFR in patients predominantly treated with ACE inhibitors versus patients mainly treated with other blood pressure-lowering agents did not differ significantly (4.5 ⫾ 0.3 vs. 3.8 ⫾ 0.4 mL/min/year, respectively, NS). Furthermore, there was no difference in mean arterial blood pressure values between patients predominantly treated with ACE inhibitors as compared with patients mainly treated with other blood pressure-lowering agents (103 ⫾ 0.6 vs. 102 ⫾ 0.7 mm Hg, NS). No statistical
705
Hovind et al: Progression of diabetic nephropathy Table 2. Multiple linear regression analysis of progression promoters in type 1 diabetic patients suffering from diabetic nephropathy Variable
Slope (95% CI)
P value
Dependent variable Rate of decline in GFR mL/min/year Independent variables Mean arterial blood pressure per 10 mm Hg Albuminuria log10 Hemoglobin A1c % Serum cholesterol mmol/L
1.18 2.04 0.73 0.66
⬍ 0.001 ⬍ 0.001 ⬍ 0.001 0.001
(0.59–1.77) (1.07–3.01) (0.36–1.10) (0.27–1.05)
R2 adj.: 0.29 Not included in the model are age, sex, height, grade of retinopathy, smoking, and class of antihypertensive treatment.
significant differences in any of the other progression promoters were demonstrated between these two groups (data not shown). A univariate analysis of demographic, clinical, and laboratory parameters revealed a significant positive correlation between the decline in GFR (dependent variable) and systolic (r ⫽ 0.27, P ⬍ 0.001), diastolic (r ⫽ 0.37, P ⬍ 0.001) and mean arterial blood pressure (r ⫽ 0.39, P ⬍ 0.001), albuminuria (r ⫽ 0.41, P ⬍ 0.001), glycosylated hemoglobin A1c (r ⫽ 0.30, P ⬍ 0.001), and serum cholesterol (r ⫽ 0.32, P ⬍ 0.001) during the study. Men had a significantly higher rate of decline in GFR than women (4.4 ⫾ 0.3 vs. 3.4 ⫾ 0.4 mL/min/year, P ⬍ 0.04). After adjustment for other progression promoters, the rate of decline in GFR was similar in men (4.1 ⫾ 0.3 mL/min/year) and in women (4.0 ⫾ 0.4 mL/min/year, NS). Smoking, degree of retinopathy, height, age, class of antihypertensive treatment, and baseline GFR had no predictive value. When these variables were examined in combination by multiple linear regression analysis, the significant predictors of decline in GFR were high values of mean arterial blood pressure, albuminuria, glycosylated hemoglobin A1c, and serum cholesterol (r ⫽ 0.54, r2adj ⫽ 0.29, P ⱕ 0.001; Table 2). The relationship between the rate of decline in GFR and each of the four previously mentioned progression promoters separated in quintiles is shown in Figure 1. Except for albuminuria (threshold 436 g/min, P ⬍ 0.01), none of the other potentially modifiable progression promoters showed a significant threshold, suggesting that the best fit of the data is linear or curve linear. A two-hit model with mean arterial blood pressure (previously demonstrated to be a modifiable progression promoter) combined with each of the three previouslymentioned potentially modifiable risk factors for losing filtration power is shown in Figure 2. Of the 271 patients treated with antihypertensive agents, 139 patients (51%; 95% CI, 45 to 57) achieved a systolic blood pressure below 140 mm Hg during follow-up. In 234 patients (86%; 82 to 90%), diastolic blood pressure was below 90 mm Hg, whereas 128 patients (47%; 41 to 53%) achieved a blood pressure below 140
mm Hg systolic and 90 mm Hg diastolic. Correspondingly, 202 patients (75%; 69 to 80%) had a mean arterial blood pressure below 107 mm Hg during follow-up. DISCUSSION Our long-term prospective observational study in 301 type 1 diabetic patients suffering from diabetic nephropathy revealed that the prognosis of diabetic nephropathy has improved during the past decades. Genuine normotensive patients have a slow rate of decline in GFR. Furthermore, several modifiable variables—that is, arterial blood pressure, albuminuria, glycemic control, and serum cholesterol—act as progression promoters. Finally, except for albuminuria, no statistical threshold effect on the decline in GFR of any of the identified progression promoters was demonstrated. To minimize selection bias and maximize generalizability, all type 1 diabetic patients with diabetic nephropathy according to the well-defined criteria [8] who had a minimum of three years of follow-up on their kidney function were enrolled in our prospective cohort study. To obtain a valid determination of the rate of decline in GFR the following requirements should be fulfilled: The applied GFR method should have good accuracy and precision, and repeated measurements of GFR (⬃every 6 to 12 months) should be performed. The observation period should be extended to at least two years [18]. These requirements were fulfilled in our study. A major strength in prospective cohort studies (such as ours) is a long-term follow-up period of an unselected, well-defined group of patients, which maximize the generalizability of the findings. The limitations are that the identified progression promoters can only be documented as modifiable risk factors in randomized clinical intervention trials, and interpretation of the impact of different treatment modalities is limited, since different categories of patients can receive different therapy. The natural history of diabetic nephropathy is characterized by arterial blood pressure elevation and a relentless decline in GFR of approximately 12 mL/min/year [4–6]. Data from studies evaluating the rate of decline
706
Hovind et al: Progression of diabetic nephropathy
Fig. 1. Impact of mean arterial blood pressure, glycosylated hemoglobin A1c, albuminuria, and serum cholesterol on the decline in GFR (P ⬍ 0.001 for each factor).
in GFR are presented in Table 3. While the original studies did not include patients on antihypertensive treatment, such treatment was applied in 90% of our patients, and we demonstrated a much slower rate of decline in GFR (4 mL/min/year). A retrospective analysis of 158 type 1 diabetic patients with nephropathy has revealed results nearly identical to the present findings [19]. Several prospective antihypertensive treatment trials have documented a beneficial effect on albuminuria, rate of decline in GFR, progression to end-stage renal disease, and survival in type 1 diabetic patients suffering from diabetic nephropathy [8, 20–25]. Thus, arterial blood pressure seems to have a rather complex relationship with diabetic nephropathy: nephropathy raising blood pressure and blood pressure accelerating the course of nephropathy. This acceleration may be at least partly due to impaired/abolished autoregulation of GFR [26, 27], leading to enhanced transmission of systemic blood pressure to
the glomerular capillaries and therefore to glomerular capillary hypertension. Studies in experimental diabetic animals [28, 29] and more recently in humans [30] have documented intrarenal hypertension, and have stressed the importance of this hemodynamic phenomenon in the initiation and progression of diabetic and nondiabetic glomerulopathies [31]. Impaired autoregulation of glomerular pressure and systemic blood pressure elevation induce distention– contraction of glomeruli. These alterations are associated with mesangial cell stretch, that in turn stimulates the synthesis and deposition of extracellular matrix material leading to mesangial expansion and glomerulosclerosis [32, 33]. In diabetic glomeruli, the mechanical stretch effect on extracellular matrix material synthesis is markedly aggravated by the presence of hyperglycemia [33]. Proteinuria is usually considered a marker of the extent of glomerular damage, but studies in various experi-
Hovind et al: Progression of diabetic nephropathy
Fig. 2. The decline in GFR divided according to median of mean arterial blood pressure combined with median albuminuria, glycosylated hemoglobin A1c, and serum cholesterol. Within each combination of risk factors, a comparison of the four different strata revealed a significant difference (P ⬍ 0.001).
mental animal models suggest that proteinuria per se may contribute to glomerular and tubulointerstitial lesions [34]. Our study suggests that albuminuria is an independent risk factor for progression of diabetic nephropathy, a finding that confirms and extends previous results in normotensive and hypertensive type 1 and type 2 diabetic patients with nephropathy [19, 35–39]. Similarly, proteinuria is a progression promoter in nondiabetic nephropathies [40, 41]. Furthermore, intervention that has ameliorated the progression of diabetic nephropathy has always been associated with a reduction in proteinuria as reviewed by Parving [42]. Finally, it should be recalled that initial reduction in albuminuria is the best clinical guideline predicting a beneficial effect of antihyperten-
707
sive treatment on the long-term decline in kidney function in diabetic and nondiabetic nephropathies [43–45]. Development of overt diabetic nephropathy has for many years been regarded as “a point of no return” in relationship to glycemic control as reviewed by Rossing [7]. However, this statement is based on studies using inappropriate methods for monitoring kidney function (serum creatinine), long-term glycemic control (blood glucose), and often, very few patients were studied. Applying better research techniques and in some studies a larger number of type 1 diabetic patients, a significant impact of glycemic control on progression of nephropathy has been found [19, 37, 46], as also documented in our study. Furthermore, pancreas transplantation can reverse the lesions of diabetic glomerulopathy, but reversal requires more than five years of normoglycemia [47]. Our study demonstrated that elevated serum cholesterol acts as an independent progression promoter in diabetic nephropathy. This finding supports the concept advocated by Moorhead et al [48] that hyperlipidemia promotes progression in chronic renal diseases once an initial event has damaged the glomerular capillary wall, thereby allowing passage of lipids and lipoproteins into the mesangium. Several short-term studies dealing with a limited number of diabetic patients have failed to demonstrate a beneficial effect of lipid lowering treatment on the surrogate endpoint: albuminuria [7]. Long-term trials applying a principal endpoint such as the decline in GFR are still lacking. A renoprotective effect of ACE inhibitors above and beyond the effect of blood pressure lowering has been demonstrated in some clinical trials [23, 24], while others have failed to demonstrate an additional non-hemodynamic effect of ACE inhibition [abstract; Tarnow et al, Diabetologia 42(Suppl 1):A275, 1999] [25]. In our longterm prospective observational study, we could not demonstrate a renoprotective effect of treatment with ACE inhibitors in comparison with other antihypertensive treatment modalities. However, the present study was not designed to evaluate this concept since ACE inhibitors were not available until 1989. In addition, as our study did not use a randomized design, different patient categories could in fact receive different antihypertensive treatment modalities. Previous studies have documented a relationship between the degree of glomerular lesions and GFR and progression of diabetic nephropathy in type 1 and type 2 diabetic patients [49–51]. Furthermore, a progressive loss of intrinsic ultrafiltration capacity has been shown to be the predominant cause of the declining GFR in type 2 diabetic patients with nephropathy [39]. The combined contribution from the previously identified renal structural lesions and the present demonstrated progression promoters explains the vast majority of progression in diabetic nephropathy. Unfortunately, the role of genetic
708
Hovind et al: Progression of diabetic nephropathy Table 3. Observational studies and clinical trials on progression of diabetic nephropathy in type 1 diabetic patients with and without antihypertensive treatment (AHT)
Authors Mogensen et al [4] Parving et al [5] Viberti et al [6] Bjo¨rck et al [23] Lewis et al [24] Elving et al [25] Tarnow et ale Mulec et al [19] Present study
Study design
N
Follow-up years
Observational, retrospective Observational, prospective Observational, retrospective Randomized trial Randomized trial Randomized trial Randomized trial Observational, retrospective Observational, prospective
10 14 13 40 409 29 48 158 301
2.8 2.2 2.5 2.2 2.7 2.0 4.0 8.0 6.7
⌬GFR mL/min/year
Mean arterial blood pressure mm Hg
No AHT
NonACE-I
ACE-I
No AHT
NonACE-I
ACE-I
10.9 9.0 14.4 — — — — — 1.9
— — — 5.6 10.8c 3.7 5.5
— — — 2.0 8.0c 4.9 6.5
121a 111 109b — — —
— — — 103 100 101d 103
— — — 102 96 100d 100
3.8f 3.8
4.5
— 95
102f 102
103
Glomerular filtration rate (GFR) determination was based on renal or plasma clearance of filtration markers. Blood pressure measurements were based on mean of values during follow-up. a MABP at end of the study b MABP at entry to the study c Creatinine clearance, GFR calculated from percentage decrease per year using the formula: y ⫽ y0 ⫻ exp(⫺kt), decline in GFR per year from baseline to mean follow up of 2.7 years d MABP during last 6 months of follow up e Abstract [Tarnow et al., Diabetologia 42(Suppl 1): pA275, 1999] f Antihypertensive treatment varied throughout the follow-up period
factors cannot be evaluated in our study, since systematical collection of DNA first began in 1993. The available data suggest that multifactorial intervention targeting blood pressure, albuminuria, glycemic control, and hyperlipidemia will diminish progression of nephropathy and may even induce regression of diabetic nephropathy, that is, a loss in GFR equal to the natural aging process (⬃1 mL/min/year). Recently, the success of such a combined approach in delaying progression of diabetic microangiopathy has been demonstrated in type 2 diabetic patients with microalbuminuria [52]. It should be mentioned that there appear to be no substantial differences between patients with type 2 diabetes and those with type 1 diabetes with respect to initiation, progression, and treatment of diabetic nephropathy [53]. In summary, the prognosis of diabetic nephropathy has improved, probably because of effective antihypertensive treatment. Genuine normotensive patients have a very slow progression of nephropathy. Several modifiable factors, that is, arterial blood pressure, albuminuria, glycemic control, and lipids, have been identified as progression promoters. Except for albuminuria, no thresholds for the remaining risk factors were demonstrated. ACKNOWLEDGMENTS The Paul and Erna Sehested Hansen Foundation, The Per S. Henriksen Foundation, and the Danish Diabetes Association are gratefully thanked for financial support. We acknowledge the assistance of Berit R. Jensen, Birgitte V. Hansen, and Inge-Lise Rossing. Reprint requests to Peter Hovind, M.D., Steno Diabetes Center, Niels Steensens Vej 2 DK-2820 Gentofte, Denmark. E-mail: [email protected]
REFERENCES 1. Parving H, Østerby R, Ritz E: Diabetic nephropathy, in The Kidney, edited by Brenner BM, Levine S, Philadelphia, W.B. Saunders, 2000, p 1731 2. Andersen AR, Christiansen JS, Andersen JK, et al: Diabetic nephropathy in type 1 (insulin-dependent) diabetes: An epidemiological study. Diabetologia 25:496–501, 1983 3. Rossing P, Rossing K, Jacobsen P, et al: Unchanged incidence of diabetic nephropathy in IDDM patients. Diabetes 44:739–743, 1995 4. Mogensen CE: Progression of nephropathy in long-term diabetics with proteinuria and effect of initial antihypertensive treatment. Scand J Clin Lab Invest 36:383–388, 1976 5. Parving H-H, Smidt UM, Friisberg B, et al: A prospective study of glomerular filtration rate and arterial blood pressure in insulindependent diabetics with diabetic nephropathy. Diabetologia 20: 457–461, 1981 6. Viberti GC, Bilous RW, Mackintosh D, et al: Monitoring glomerular function in diabetic nephropathy. Am J Med 74:256–264, 1983 7. Rossing P: Promotion, prediction, and prevention of progression in diabetic nephropathy. Diabet Med 15:900–919, 1998 8. Parving H-H, Andersen AR, Smidt UM, et al: Early aggressive antihypertensive treatment reduces rate of decline in kidney function in diabetic nephropathy. Lancet 1:1175–1179, 1983 9. Bro¨chner-Mortensen J: A simple method for the determination of glomerular filtration rate. Scand J Clin Lab Invest 30:271–274, 1972 10. Mancini G, Carbonara AO, Heremans JF: Immunochemical quantitation of antigens by single radial immunodiffusion. Immunochemistry 2:235–254, 1965 11. Miles DW, Mogensen CE, Gundersen HJG: Radioimmunoassay for urinary albumin using a single antibody. Scand J Clin Lab Invest 26:5–11, 1970 12. Feldt-Rasmussen B, Dinesen B, Deckert M: Enzyme immunoassay: An improved determination of urinary albumin in diabetics with incipient nephropathy. Scand J Clin Lab Invest 45:539–544, 1985 13. Rossing P, Hommel E, Smidt UM, et al: Impact of arterial blood pressure and albuminuria on the progression of diabetic nephropathy in IDDM patients. Diabetes 42:715–719, 1993 14. Svendsen PA, Christiansen JS, Søegaard U, et al: Rapid changes in chromatographically determined haemoglobin A1c induced by short-term changes in glucose concentration. Diabetologia 19:130– 136, 1980
Hovind et al: Progression of diabetic nephropathy 15. Mortensen HB: Quantitative determination of hemoglobin A1c by thin layer isoelectric focusing. J Chromatogr 182:325–333, 1980 16. American Diabetes Association: Treatment of hypertension in diabetes. Diabetes Care 16:1394–1401, 1993 17. Zoccali C, Postorino M, Martorano C, et al: The “breakpoint” test, a new statistical method for studying progression of chronic renal failure. Nephrol Dial Transplant 4:101–104, 1989 18. Levey AS, Gassman J, Hall PM, et al: Assessing the progression of renal disease in clinical studies: Effects of duration of followup and regression to the mean. J Am Soc Nephrol 1:1087–1094, 1991 19. Mulec H, Blohme´ G, Gra¨nde B, Bjo¨rck S: The effect of metabolic control on rate of decline in renal function in insulin-dependent diabetes mellitus with overt diabetic nephropathy. Nephrol Dial Transplant 13:651–655, 1998 20. Mogensen CE: Long-term antihypertensive treatment inhibiting progression of diabetic nephropathy. Br Med J 285:685–688, 1982 21. Parving H-H, Andersen AR, Smidt UM, et al: Effect of antihypertensive treatment on kidney function in diabetic nephropathy. Br Med J 294:1443–1447, 1987 22. Parving H-H, Hommel E: Prognosis in diabetic nephropathy. Br Med J 299:230–233, 1989 23. Bjo¨rck S, Mulec H, Johnsen SA, et al: Renal protective effect of enalapril in diabetic nephropathy. Br Med J 304:339–343, 1992 24. Lewis E, Hunsicker L, Bain R, et al: The effect of angiotensinconverting-enzyme inhibition on diabetic nephropathy. N Engl J Med 329:1456–1462, 1993 25. Elving LD, Wetzels JFM, Van Lier HJJ, et al: Captopril and atenolol are equally effective in retarding progression of diabetic nephropathy. Diabetologia 37:604–609, 1994 26. Parving H-H, Kastrup J, Smidt UM, et al: Impaired autoregulation of glomerular filtration rate in type 1 (insulin-dependent) diabetic patients with nephropathy. Diabetologia 27:547–552, 1984 27. Christensen PK, Hansen HP, Parving H-H: Impaired autoregulation of GFR in hypertensive non-insulin dependent diabetic patients. Kidney Int 52:1369–1374, 1997 28. Hostetter TH, Troy JL, Brenner BM: Glomerular hemodynamics in experimental diabetes mellitus. Kidney Int 19:410–415, 1981 29. Jensen PK, Steven K, Blæhr H, et al: The effects of indomethacin on glomerular hemodynamics in experimental diabetes. Kidney Int 29:490–495, 1986 30. Imanishi M, Yoshioka K, Konishi Y, et al: Glomerular hypertension as one cause of albuminuria in type II diabetic patients. Diabetologia 42:999–1005, 1999 31. Hostetter TH, Rennke HG, Brenner BM: The case for intrarenal hypertension in the initiation and progression of diabetic and other glomerulopathies. Am J Med 72:375–380, 1982 32. Riser BL, Cortes P, Zhao X, et al: Intraglomerular pressure and mesangial stretching stimulate extracellular matrix formation in the rat. J Clin Invest 90:1932–1943, 1992 33. Cortes P, Riser BL, Yee J, et al: Mechanical strain of glomerular mesangial cells in the pathogenesis of glomerulosclerosis: Clinical implications. Nephrol Dial Transplant 14:1351–1354, 1999 34. Remuzzi G, Bertani T: Is glomerulosclerosis a consequence of altered glomerular permeability to macromolecules? Kidney Int 38:384–394, 1990 35. Elving LD, Wetzels JFM, De Nobel E, et al: Erythrocyte sodiumlithium countertransport is not different in type 1 (insulin-depen-
36. 37. 38. 39. 40. 41.
42. 43.
44. 45.
46. 47. 48. 49. 50. 51. 52.
53.
709
dent) diabetic patients with and without diabetic nephropathy. Diabetologia 34:126–128, 1991 Breyer JA, Bain P, Evans JK, et al: Predictors of the progression of renal insufficiency in patients with insulin-dependent diabetes and overt diabetic nephropathy. Kidney Int 50:1651–1658, 1996 Parving H-H, Smidt UM, Hommel E, et al: Effective antihypertensive treatment postpones renal insufficiency in diabetic nephropathy. Am J Kidney Dis 22:188–195, 1993 Gall M-A, Nielsen FS, Smidt UM, et al: The course of kidney function in type 2 (non-insulin-dependent) diabetic patients with diabetic nephropathy. Diabetologia 36:1071–1078, 1993 Nelson RG, Bennett PH, Beck GJ, et al: Development and progression of renal disease in Pima Indians with non-insulin-dependent diabetes mellitus. N Engl J Med 335:1636–1642, 1996 Peterson JC, Adler S, Burkart JM, et al: Blood pressure control, proteinuria, and the progression of renal disease: The modification of diet in renal disease study. Ann Intern Med 123:754–762, 1995 The Gisen Group: Randomised placebo-controlled trial of effect of ramipril on decline in glomerular filtration rate and risk of terminal renal failure in proteinuric, non-diabetic nephropathy. Lancet 349:1857–1863, 1997 Parving H-H: Renoprotection in diabetes: Genetic and nongenetic risk factors and treatment. Diabetologia 41:745–759, 1998 Rossing P, Hommel E, Smidt UM, et al: Reduction in albuminuria predicts a beneficial effect on diminishing the progression of human diabetic nephropathy during antihypertensive treatment. Diabetologia 37:511–516, 1994 Rossing P, Hommel E, Smidt UM, et al: Reduction in albuminuria predicts diminished progression in diabetic nephropathy. Kidney Int 45(Suppl 45):S145–S149, 1994 Apperloo AJ, De Zeeuw D, De Jong PE: Short-term antiproteinuric response to antihypertensive treatment predicts long-term GFR decline in patients with non-diabetic renal disease. Kidney Int 45(Suppl 45):S174–S178, 1994 Nyberg G, Blohme´ G, Norde´n G: Impact of metabolic control in progression of clinical diabetic nephropathy. Diabetologia 30:82–86, 1987 Fioretto P, Steffes MW, Sutherland DER, et al: Reversal of lesions of diabetic nephropathy after pancreas transplantation. N Engl J Med 339:69–75, 1998 Moorhead JF, El-Nahas M, Chan MK, et al: Lipid nephrotoxicity in chronic progressive glomerular and tubulo-interstitial disease. Lancet 2:1309–1311, 1982 Mauer SM, Steffes MW, Ellis EN, et al: Structural-functional relationships in diabetic nephropathy. J Clin Invest 74:1143–1155, 1984 Østerby R, Parving H-H, Hommel E, et al: Glomerular structure and function in diabetic nephropathy -early to advanced stages. Diabetes 39:1057–1063, 1990 Østerby R, Gall M-A, Schmitz A, et al: Glomerular structure and function in proteinuric type 2 (non-insulin-dependent) diabetic patients. Diabetologia 36:1064–1070, 1993 Gæde P, Vedel P, Parving H-H, et al: Intensified multifactorial intervention in patients with type 2 diabetes mellitus and microalbuminuria: The Steno type 2 randomised study. Lancet 353:617– 622, 1999 Parving H-H: Initiation and progression of diabetic nephropathy. N Engl J Med 335:1682–1683, 1996