Persistent Renal Hypertrophy and Faster Decline of Glomerular Filtration Rate Precede the Development of Microalbuminuria in Type 1 Diabetes
Soon after the onset of type 1 diabetes, renal hypertrophy and hyperfiltration become manifest, particularly among patients who will subsequently develop diabetic nephropathy. Whether these early renal dysfunctions are involved in the pathogenesis of diabetic nephropathy is currently unclear. We evaluated, during the same day, kidney volume and glomerular filtration rate (GFR) in 146 patients with type 1 diabetes and normal renal function. All the individuals were then monitored for a mean of 9.5 ± 4.4 years for the development of microalbuminuria. Kidney volume and GFR were reevaluated in a subset of 68 patients 4 years after baseline. During follow-up, microalbuminuria developed in 27 of 146 diabetic patients. At baseline, kidney volume (312.8 ± 52.6 vs. 281.4 ± 46.1 vs. 236.8 ± 41.6 ml/1.73 m2, P < 0.05) but not GFR was increased in patients predisposed to microalbuminuria. Risk of progression was higher in patients with increased kidney volume (P = 0.0058). Patients predisposed to microalbuminuria showed a stable increase in kidney volume (P = 0.003), along with a faster decline of GFR (P = 0.01). Persistent renal hypertrophy and faster decline of GFR precede the development of microalbuminuria in type 1 diabetes. These findings support the hypothesis that renal hypertrophy precedes hyperfiltration during the development of diabetic nephropathy.
From a clinical perspective, the first sign of diabetic nephropathy is represented by an increased urinary albumin excretion rate (known as microalbuminuria) that usually develops after a few years of duration of diabetes (1). From a morphological and functional point of view, however, it is well established that kidney volume and glomerular filtration rate (GFR) increase promptly after the onset of type 1 diabetes (2,3) and that this phenomenon is particularly accentuated in the subset of individuals who, years later, will develop diabetic nephropathy (4–6). After these initial reports, a number of studies have subsequently confirmed that increased kidney volume (7–9) and faster GFR (10–12) are associated with a poor renal prognosis in patients affected by type 1 diabetes. Altogether these findings suggest that the very beginning of this renal complication might coincide with the onset of diabetes itself and that abnormal kidney volume and GFR may be part of the process.
Two hypotheses have been formulated to explain early renal dysfunctions induced by diabetes. According to the “vascular hypothesis,” hyperfiltration is the primary event, a consequence of defects in vascular control including increases in plasma flow and glomerular capillary pressure (13,14). Hypertrophy would follow, probably after compensatory changes of tubular function aimed to prevent the urinary loss of water and electrolytes (15). The “tubular hypothesis” suggests instead that the initial event is renal hypertrophy (16–18), a consequence of hyperglycemia and hyperglycemia-induced overproduction of growth factors and cytokines (19,20). In brief, renal hypertrophy resulting from hyperplasia and hypertrophy of the nephron and, in particular, of the proximal tubule (21,22) would cause increased proximal tubular reabsorption, thus reducing the amount of salt reaching the macula densa. The consequent activation of the tubuloglomerular feedback would finally result, via an increase of GFR, in the normalization of salt delivery to the distal tubule.
During the early stages of diabetes, renal hyperfiltration and hypertrophy show up simultaneously. Thus, to distinguish at this point the cause from the effect appears to be rather difficult (14).
That renal hypertrophy and hyperfiltration may in some cases dissociate from each other is, however, suggested by previous studies showing that, in patients with type 1 diabetes, renal hyperfiltration but not hypertrophy is reduced by intensive insulin treatment (23,24). More recently, indirect support for this notion has been provided by two independent studies demonstrating reduced GFR in patients with long-standing normoalbuminuria who have type 1 diabetes predisposed to nephropathy (25) and an increased kidney volume in the early phases of microalbuminuria (26).
Whether repeated evaluations of kidney volume and GFR before the development of microalbuminuria might help in clarifying their role in the pathogenesis of diabetic nephropathy is currently unknown. To explore this issue, we evaluated on two occasions kidney volume and GFR during a prospective study aimed to evaluate the rate of development of microalbuminuria in patients with type 1 diabetes and normal renal function.
RESEARCH DESIGN AND METHODS
One hundred forty-six patients with type 1 diabetes who regularly attended the outpatient clinic of the pediatric department of the San Raffaele Scientific Institute were considered for this study. To be enrolled, an individual was required to have a duration of diabetes of at least 4 years (range at baseline 4–22 years, median 9 years) and a normal albumin excretion rate (measured by immunoturbidimetric technique on a Cobas Mira autoanalyzer; Roche, Basel, Switzerland). Normoalbuminuria was defined as a median albumin excretion rate <20 μg/min in three consecutive overnight collections of sterile urine. The presence of diabetic nephropathy and proliferative retinopathy at baseline resulted in exclusion from the study. Thirty subjects not affected by diabetes, similar for age and sex distribution, were considered for control purposes.
The study was performed according to the principles expressed in the Declaration of Helsinki and was approved by the ethics committee of the San Raffaele Scientific Institute, Milano, Italy. Written informed consent was obtained from all participants. When a subject was <18 years of age, a parent gave informed consent.
Every subject attended a morning visit, when clinical features were recorded. BMI was expressed as weight in kilograms divided by the square of height in meters. The prescribed insulin dose was also recorded. Resting blood pressure (Korotkoff phases I/V) was measured once in the right arm by ordinary mercury sphygmomanometry with approximation to the nearest 2 mmHg. No subjects were receiving antihypertensive therapy. Patients with type 1 diabetes were otherwise treated only with insulin, were studied in ordinary glycemic control, and did not have any other known disease.
At the beginning of the study, measurement of kidney volume and GFR and clinical evaluation were performed during the same day in both patients with type 1 diabetes and control subjects (baseline). These same parameters were reevaluated in a subset of 68 patients with type 1 diabetes who were still normoalbuminuric 4 years after the baseline and who agreed to repeat the entire procedure (second evaluation). Three of the 146 patients with type 1 diabetes who entered the study developed microalbuminuria in the 4 years after baseline and were therefore excluded from the second evaluation. Patients with type 1 diabetes who underwent the second evaluation were not selected in any particular way.
Since the beginning of the study, clinical evaluation and measurement of microalbuminuria (in three consecutive overnight collections) were performed on a yearly basis in all subjects involved in the study until the end of follow-up or the development of microalbuminuria. Persistent microalbuminuria was defined as a median albumin excretion rate between 20 and 200 μg/min in three overnight collections of sterile urine. Once microalbuminuria was diagnosed, patients with type 1 diabetes were removed from the study and treated with ACE inhibitors or angiotensin receptor blockers.
Kidney volume was measured by ultrasound technique (27) performed by a single observer using 3.5–5 MHz convex and linear probes (SSA-250–10; Toshiba, Tokyo, Japan), who was blinded to the clinical characteristics of the subjects. Kidney volume was calculated according to the equation for an ellipsoid (28) and expressed as the mean volume of both kidneys corrected for body surface area; the mean coefficient of variation between measurements done on the same day was 5.9% (8).
GFR was estimated for each individual via 51Cr-EDTA clearance, as previously described (29,30). Briefly, 1 μCi of 51Cr-EDTA/kg body wt was given as a single injection. Thereafter, venous blood samples were drawn at 5, 15, 30, 60, 90, 120, 180, and 300 min. Radioactivity of each blood sample was measured by gamma counter. A bicompartment model was used to fit the slope of decay of EDTA-associated radioactivity, and GFR was finally expressed as milliliters per minute after correction for body surface area.
Data are shown as arithmetical means ± SD. Comparisons between groups were addressed by ANOVA, and multiple comparisons were performed with the Tukey-Kramer test (JMP software for the Apple Macintosh; SAS Institute, Cary, NC). Parametric comparisons between groups were checked by nonparametric tests for unpaired observations where appropriate. The null hypothesis was rejected at the 5% level (two tailed). Kaplan-Meier survival analysis was used to evaluate the probability that microalbuminuria would develop in patients with type 1 diabetes who had a normal or increased kidney volume; the two groups were compared by log-rank test.
A multiple regression analysis was performed to verify which variables were associated with kidney volume at baseline. GFR and kidney volume changes over time were evaluated by ANOVA for repeated measures, considering the main effects of the between-groups factor (normoalbuminuria vs. microalbuminuria) and the time factor as well as the time-group interaction.
Progression to microalbuminuria.
Altogether, the 146 patients with type 1 diabetes involved in the study were followed for a mean of 9.5 ± 4.4 years (mean ± SD, range 2–16 years). Microalbuminuria developed in 27 of 146 patients with type 1 diabetes (18.5%) after a follow-up of 8.1 ± 3.4 years (2–15 years). During this period, albumin excretion rate increased from 7.9 ± 3.9 to 110.4 ± 21.5 μg/min (P = 0.0001). Mean age and duration of diabetes at the onset of microalbuminuria were, respectively, 24.8 ± 4.0 and 17.7 ± 3.4 years.
On the other hand, in 119 of 146 patients with type 1 diabetes (81.5%), albumin excretion rate remained within the normal range (from 6.3 ± 4.4 to 7.4 ± 3.6 μg/min, P = NS) after a follow-up of 9.8 ± 4.6 years (range 2–16). To simplify reporting, the 119 patients with type 1 diabetes who, at the end of the follow-up, remained normoalbuminuric and the 27 patients with type 1 diabetes who developed microalbuminuria will be referred to, respectively, as the normoalbuminuria and microalbuminuria groups.
Clinical characteristics of the subjects at baseline.
As shown in Table 1, at baseline, control subjects and patients with type 1 diabetes in both the normoalbuminuria and microalbuminuria groups were comparable with respect to age, sex distribution, BMI, blood pressure, triglyceride levels, and creatinine levels. Normo- and microalbuminuria groups were also similar for duration of diabetes and albumin excretion rate. HbA1c (A1C) levels were slightly increased (P = 0.048) in the microalbuminuria group.
Kidney volume and GFR at baseline.
Kidney volume was increased in patients with type 1 diabetes compared with that in control subjects (287.2 ± 48.7 vs. 236.8 ± 41.6 ml/1.73 m2, P = 0.00001). As shown in Table 1 and Fig. 1A, when patients with type 1 diabetes were divided according to renal status at the end of follow-up, the microalbuminuria group showed an increased kidney volume compared with the normoalbuminuria group and with control subjects (312.8 ± 52.6 vs. 281.4 ± 46.1 vs. 236.8 ± 41.6 ml/1.73 m2, P < 0.05 for all pairs after correcting for multiple comparisons by Tukey-Kramer test).
GFR was also increased in patients with type 1 diabetes compared with that in control subjects (118.3 ± 18.9 vs. 110.7 ± 9.6 ml/min per 1.73 m2, P = 0.03). As shown in Table 1 and Fig. 1B, when patients with type 1 diabetes were divided according to renal status at the end of follow-up, the GFR of the microalbuminuria group was similar to that of the normoalbuminuria group (121.7 ± 18.3 vs. 117.6 ± 19.0 ml/min per 1.73 m2, P > 0.05), but higher than that of control subjects (121.7 ± 18.3 vs. 110.7 ± 9.6 ml/min per 1.73 m2, P < 0.05).
Effect of kidney volume on the progression to microalbuminuria.
A kidney volume >99th percentile in control subjects (327.6 ml/1.73 m2) was defined as increased and is indicated by a dotted line in Fig. 1A. Thirty-six of 146 patients with type 1 diabetes (24.6%) were shown to have an increased kidney volume. During the follow-up, microalbuminuria developed in 15 of 110 patients with type 1 diabetes with normal kidney volume (13.6%) and in 12 of 36 patients with type 1 diabetes with increased kidney volume (33.3%).
Accordingly, Kaplan-Meier survival analysis showed that the risk of progression to microalbuminuria was significantly higher in patients with type 1 diabetes and increased kidney volume when compared with those with normal kidney volume (P = 0.0058) (Fig. 2). Patients with normal and increased kidney volumes were similar for age (16.3 ± 3.9 vs. 17.3 ± 3.6 years, P = 0.10), duration of diabetes (9.3 ± 4.2 vs. 9.9 ± 4.2 years, P = 0.49), A1C (9.1 ± 1.8 vs. 9.4 ± 1.4%, P = 0.40), BMI (20.7 ± 2.8 vs. 21.1 ± 3.3 kg/m2, P = 0.30), albumin excretion rate (7.1 ± 4.4 vs. 8.3 ± 3.8 μg/min, P = 0.12), and systolic (116.8 ± 11.2 vs. 120.6 ± 11.0 mmHg, P = 0.12) and diastolic (72.7 ± 8.1 vs. 75.0 ± 5.3 mmHg, P = 0.16) blood pressure. GFR (126.3 ± 20.4 vs. 115.8 ± 16.7 ml/min per 1.73 m2, P = 0.002) was higher in patients with type 1 diabetes and increased kidney volume. In multiple regression analysis, considering all patients with type 1 diabetes and including the variables that at baseline were found to correlate to kidney volume after simple regression (GFR, age, A1C, albumin excretion rate, and BMI), only GFR (F = 12.3, P = 0.0007), A1C (F = 8.0, P = 0.0056), and age (F = 6.7, P = 0.0108) were confirmed to be independent factors associated with kidney volume.
Clinical characteristics of the subset of patients with type 1 diabetes who underwent a second evaluation of kidney volume and GFR.
Clinical characteristics of the subset of 68 (of 146) patients with type 1 diabetes who repeated the measurement of kidney volume and GFR are shown in Table 2. Both at baseline and at the second evaluation, clinical parameters were similar for the normoalbuminuria and microalbuminuria groups. The 68 patients with type 1 diabetes of this subset were younger (14.7 ± 3.2 vs. 16.7 ± 3.9 years, P = 0.004) compared with the 146 patients with type 1 diabetes considered at baseline but with similar duration of diabetes (8.3 ± 3.4 vs. 9.5 ± 4.3 years, P = 0.06). In this subset, at the end of follow-up, microalbuminuria developed in 16 of 68 patients with type 1 diabetes (23.5%).
Kidney volume and GFR at baseline and at the second evaluation in the subset of patients with type 1 diabetes.
Kaplan-Meier survival analysis confirmed both at baseline (P = 0.03) and at the second evaluation (P = 0.004) that risk of progression to microalbuminuria was significantly higher in patients with increased kidney volume. Figure 3 and Table 2 show the change in kidney volume (Fig. 3A) and GFR (Fig. 3B) of normoalbuminuria and microalbuminuria groups between baseline and the second evaluation. During this period (4 years), kidney volume remained increased in the microalbuminuria group (from 311.0 ± 59.9 to 328.2 ± 64.7 ml/1.73 m2) compared with that in the normoalbuminuria group (from 279.7 ± 43.3 to 283 ± 45.5 ml/1.73 m2) as confirmed by a significant (F = 9.5, P = 0.003) between-groups difference paralleled by a nonsignificant difference in time (F = 1.5, P = 0.2) and in time-group interaction (F = 0.9, P = 0.4).
Kidney volume at baseline was significantly correlated to kidney volume at the second evaluation (r2 = 0.31, P = 0.0001). GFR decreased significantly between baseline and the second evaluation in both groups (normoalbuminuria, from 124.4 ± 17.2 to 115.3 ± 15.6 ml/min per 1.73 m2, and microalbuminuria, from 126.1 ± 19.4 to 102.9 ± 18.2 ml/min per 1.73 m2) as confirmed by a significant (F = 27.7, P = 0.0001) time difference paralleled by a nonsignificant group difference (F = 1.8, P = 0.2). The decline was, however, more accentuated in the microalbuminuria group as indicated by a significant time-group interaction (F = 6.4, P = 0.01).
When calculated as the change (Δ) between GFR at baseline and at the second evaluation, the rate of decline of GFR was faster in the microalbuminuria group (0.48 ± 0.47 vs. 0.19 ± 0.37 ml · min−1 · month−1, P = 0.01). Finally, kidney volume at baseline was directly correlated to rate of decline of GFR between baseline and the second evaluation (r2 = 0.13, P = 0.02), in line with the hypothesis that increased kidney volume may predispose to faster decline of GFR.
The present study shows that long-term increases of kidney volume and accelerated declines of GFR cluster in the subgroup of normoalbuminuric patients with type 1 diabetes predisposed to develop microalbuminuria. This evidence supports the hypothesis that renal hypertrophy, via increased tubular reabsorption, is the primary dysfunction in the development of diabetic nephropathy and that GFR increases as a secondary event. Through the years, hypertrophy is maintained, but although the tubule survives overworking, the glomerulus progressively loses its activity.
As described above, the fact that increased kidney volume and faster GFR during the first years of diabetes are tightly linked to a poorer renal prognosis has been known for a number of years (4–12). The novel finding of our study is the demonstration that these two dysfunctions dissociate from each other through time, thus allowing the clarification of their respective role during the early phases of diabetic nephropathy.
A few considerations immediately follow from these findings. Our results support the notion that the biological mechanisms controlling kidney volume, in particular those affected by hyperglycemia (20), should be regarded as potentially involved also in the pathogenesis of nephropathy. In addition, increased kidney volume, as a consistent, reproducible, and persistent marker of diabetic nephropathy, appears to be of potential use for the early identification of individuals at risk for this complication.
Among the factors potentially involved in the pathogenesis of diabetic nephropathy, predisposition to essential hypertension has been suggested as a major candidate (31–33). Although our results do not show a direct association between kidney volume and blood pressure, we cannot currently rule out the possibility that an underlying predisposition to hypertension might nonetheless affect kidney volume. This is not only due to the fact that, unfortunately, intermediate phenotypes and genetic markers of essential hypertension were not evaluated in this study, but also based on the recent evidence that only nocturnal blood pressure measurements appear to be informative for the early identification of patients with type 1 diabetes predisposed to microalbuminuria (34).
A recent report has demonstrated that in patients with type 1 diabetes, regression of microalbuminuria reached a 6-year cumulative incidence of 58% (35). This finding may suggest that microalbuminuria may not represent a useful marker of future development of diabetic nephropathy. In our case, however, patients with type 1 diabetes predisposed to develop microalbuminuria also show a significant reduction in GFR, a marker of diabetic glomerulopathy independent on the level of microalbuminuria (25).
Finally, whether pharmacological treatment aimed to reduce kidney volume during normoalbuminuria may protect from the subsequent development of microalbuminuria is currently unknown. In animal models of diabetes, prevention of early hyperfiltration/hypertrophy has already been shown to avoid the subsequent development of diabetic nephropathy (36,37). Prospectively, a similar approach in humans could allow identification of new therapeutic approaches of use for the primary prevention of diabetic nephropathy.
In conclusion, our findings support the hypothesis that renal hypertrophy is the primary dysfunction in the development of diabetic nephropathy. Further studies are now required to evaluate whether the prevention/remission of this early renal dysfunction may beneficially affect the natural history of diabetic nephropathy.
This work was supported by grants from the Italian Ministry of Health (to G.Z.).
We are indebted to Daniela Gabellini for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received May 2, 2006.
- Accepted June 15, 2006.