ELSEVIER Diabetes Research and Clinical Practice 25 (1994) 155-160

Glomerular hyperfiltration in young diabetes

Polynesians with type 2

Raymond Bruce, Michael Rutland, Tim Cundy*

Departments of Medicine and Nuclear Medicine, Auckland Hospital, Auckland 1, New Zealand

Received 6 April 1994; revision received 15 June 1994; accepted 16 June 1994

Abstract

An increase in glomerular filtration rate (hypertiltration) may be an important early event in the initiation of diabetic nephropathy but the prevalence of hyperfiltration appears to vary between different populations with type 2 diabetes. We have measured glomerular filtration rate using “Cr EDTA clearance in 15 young Polynesians (mean age 32 years), l-30 months after the initial diagnosis of type 2 diabetes and 15 control Polynesian subjects of comparable age and sex distribution. The mean glomerular filtration rate in the diabetic subjects (216 mUmin) was 57% greater than that of the controls (137.5 mUmin, P < 0.0001). About one-third of their excess in glomerular filtration rate could be accounted for by the marked obesity of the diabetic subjects, but even after correcting for body size the diabetic subjects still had a significantly higher mean glomerular filtration rate than controls (165.6 vs. 119.6 ml/min per 1.73 m*, P c 0.001); 73% of the diabetic subjects had hypertiltration (> 140 ml/min per 1.73 m*). The diabetic subjects were normotensive but nonetheless had increased rates of albumin excretion (median 61 versus 9 mg/day, P < 0.001). We conclude that hyperfiltration is common in young Polynesians with recently diagnosed type 2 diabetes. Prospective studies are needed to determine whether this early abnormality of renal function heralds the later development of overt nephropathy.

Keyworris: Diabetic nephropathy; Glomerular filtration rate; Hyperfdtration; Polynesian

1. Introduction

The increase in glomerular filtration rate which is often present early in the course of type 1 diabetes (hyperliltration) may be an important fac- tor initiating glomerular damage and predisposing to the later development of diabetic nephropathy

[l-3]. Although diabetic renal disease is also com- mon amongst subjects with type 2 diabetes, it is less certain whether hyperfiltration occurs early in its course. In New Zealand there is a high preva- lence of type 2 diabetes and diabetic nephropathy amongst the Maori and Pacific Islanders, who are of Polynesian origin. In this group the diabetes is often of early onset, with a prevalence of around

* Corresponding author, Department of Medicine, Auck- 5% amongst 40 year olds, and is strongly associ- land Hospital, Park Road, Auckland 1, New Zealand. ated with obesity [4]. Since glomerular hyperliltra-

016%8227/94/$07.00 0 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0168-8227(94)00976-2

156 R. Bruce et al. /Diabetes Res. Clin. Pratt. 2S (1994) 155-160

tion can be a feature of obesity as well as diabetes [S] young diabetic Polynesians could have very high glomerular filtration rates, which may con- tribute to their high prevalence of diabetic renal disease. We have therefore sought to determine whether hypertiltration is present early in the course of type 2 diabetes in young Polynesian subjects.

2. Subjects and methods

2.1. Subjects Fifteen Polynesian subjects with type 2 diabetes

were recruited from diabetes clinics in central and south Auckland. Subjects were eligible for the study if they were under the age of 40, their known duration of diabetes was less then 3 years and they did not have hypertension (blood pressure < 160195 and not taking antihypertensive drugs). Six subjects were managing their diabetes with diet alone, six were taking oral hypoglycaemic agents and three were using insulin. Fifteen non-diabetic Polynesian subjects acted as a control group. The control subjects were recruited from friends of the diabetic subjects, and had to be under the age of 40 and not to have hypertension as defined above.

2.2. Experimental protocol Prior to the study day each subject collected two

early morning urine samples from which we estimated the mean albumin excretion rate [6]. Urine infection and contamination was excluded using dipsticks to test for white blood cells, nitrate and protein. Albumin in the urine was measured by a sensitive immunoturbidimetric assay with an interassay C.V. of 5%. Fractional albumin excre- tion was calculated from the ratio of albumin clearance and measured glomerular filtration rate. Urinary creatinine was measured by the picric acid method using a Cobas Bio analyser (Hoffman La Roche, Basel, Switzerland), interassay C.V. of 2%.

All subjects attended an outpatient facility at 09:OO h in a non-fasted state and had their height, weight and blood pressure measured and a brief medical history recorded. From the measurements of height (m) and weight (kg) the following derived indices were obtained: body mass index: weight/ height2, body surface area: (d/height x weight)/36

[7] and fat-free mass: (weight x A) + (height2 x B) where A = 0.285, B = 12.1 for men and A = 0.287, B = 9.74 for women [8]. Blood was drawn for the measurement of urea, albumin, creatinine, glucose and fructosamine. Glomerular filtration rate was determined using a single bolus injection of “Cr-labelled EDTA with blood sam- ples taken at 2, 3 and 4 h later 191. A correction factor (x 0.93) was applied to all GFR results to approximate inulin clearance [lo].

2.3. Statistical analysis The results are expressed as the mean (with stan-

dard deviation). Mean values in the two groups were compared by Student’s t-test for unpaired samples, but where the data were not normally dis- tributed the Wilcoxon rank sum test was used and the results expressed as the median (and range). Correlations were calculated using the method of least squares and proportions were compared using the x2 test. Multiple linear regression analy- sis was used to examine the factors determining glomerular filtration rate. The data were examined using a variety of iterative procedures; backward, forward and stepwise selection to attempt to max- imise r*.

3. Results

3.1. Clinical characteristics The control subjects and diabetic subjects were

well matched in terms of mean age, age range and sex distribution. The diabetic subjects were, how- ever, significantly more obese as judged by body mass index (Table 1) and reported a significantly greater lifetime cigarette consumption (median 5.0 vs. 0.5 pack years, P = 0.03).

The diabetic subjects were studied at a median 14 months (range l-30) after the diagnosis of type 2 diabetes had been made and had, as expected, significantly higher random blood glucose and plasma fructosamine concentrations than the con- trols, but mean plasma creatinine concentrations were similar, as were mean values for systolic and diastolic blood pressure. The diabetic subjects also had a significantly greater albumin excretion rate than controls, whether this was expressed as albu- min excretion rate of fractional albumin excretion (Table 1).

R. Bruce et al. /Diabetes Rex C/in. Pratt. 25 (1994) 155-160

Table 1 Demographic, anthropometric, blood pressure and biochemical findings in control and diabetic subjects

157

Control Type 2 diabetes P values

Sex (M/F) Age (years) Body mass index (kg/m*) Body surface area (m’) Fat-free mass (kg) Systolic blood pressure

(mmf-fg) Diastolic blood pressure

(mmHg) Plasma creatinine (pmol/l) Albumin excretion rate

(mg/day) Fractional albumin excretion

(x 106) Plasma glucose (mmolll) Plasma fructosamine @molIl)

817

31 (3) 28.8 (2.9)

1.96 (0.19) 56.4 (9.5)

123 (16)

73 (12)

91 (21) 9 (4-34)

1.2 (0.6-5.3)

5.7 (0.9) 215 (14)

8/7

32 (5) 36.9 (6.2) 2.19 (0.27)

63.8 (I 1.9) 122 (14)

79 (14)

89 (13) 61 (8-411)

6.3 (0.8-56.5)

15.7 (6.2) 343 (54)

NS NS

<O.ool

<0.02 NS (= 0.07) NS

NS

NS

<O.OOl

<O.Ol

<O.col <O.ool

Values are given as the mean (with S.D.), except for the measure of albumin excretion rates which are given as the median (with range). P values refer to Student’s t-test, except for albumin excretion measures which were compared using the Wilcoxon rank sum test. NS, not significant (P > 0.05).

3.2. Glomerular filtration rate The mean glomerular filtration rate was very

markedly increased in the diabetic subjects (216 ml/mm) and some 57% greater than the mean in controls (137.5 ml/mm, P < 0.0001). Since obesity itself can increase glomerular filtration rate and the diabetic subjects were significantly more obese than the controls, we also compared the glomerular filtration rate in the two groups after various adjustments for body size. Irrespective of the correction used, the diabetic subjects still had a significantly higher mean glomerular filtration

rate than controls (P < 0.001) but the differences were reduced. When corrected to 1.73 m* surface area or expressed per kg of fat-free mass, approx- imately one-third of the excess glomernlar tiltra- tion rate in the diabetic subjects over controls was accounted for, but it remained a mean 38% and 39%, respectively, above control values (Table 2). Hyperfiltration (defined as a glomerular filtration rate of > 140 ml/rnin per 1.73 m* surface area) was present in 11 of the 15 diabetic subjects, but only 2 of the 15 control subjects (x2 = 8.69, P < 0.01). No correlation was present in either group

Table 2 A comparison of glomerular filtration rate in control and diabetic subjects

Glomerular filtration rate

Uncorrected (ml/mm) Corrected for surface area

(ml/mitt per 1.73 m2) Corrected for fat-free mass

(ml/mitt per kg)

Control Diabetes

137.5 (22.3) 216.0 (36.0) 119.6 (16.4) 165.6 (28.0)

2.48 (0.44) 3.45 (0.73)

P value

<O.OOOl <O.OOl

<O.OOl

The results are given both before and after the application of various correction factors for the differing body size between the two groups. Results are given as the mean (and S.D.). P values refer to Students t-test.

158 R. Bruce et al. /Diabetes Res. Clin. Pratt. 25 (1994) 155-160

between glomerular filtration rate and either body mass index, albumin excretion rate, blood sugar or plasma fructosamine.

In the multiple regression analyses the relative contributions of fat mass, fat-free mass, body mass index, fructosamine, glucose, systolic and diastolic blood pressure, sex, smoking, age and the presence or absence of diabetes, to the variation seen in glomerular filtration rate were assessed. When diabetic subjects and control subjects were con- sidered together the only variables significant at the 5% level were the presence of diabetes (partial r2 = 0.606) and body mass index (partial r2 =

0.070). When the presence or absence of diabetes was excluded as a variable then the significant determinants of glomerular filtration rate were body mass index (partial r2 = 0.480) and fructos- amine (partial r2 = 0.120). When the diabetic sub- jects alone were considered then only body mass index emerged as a significant determinant of glomerular filtration rate (partial r2 = 0.203). In the control subjects glomerular filtration rate was significantly related to both diastolic blood pres- sure and fat-free mass (partial r2 = 0.254 and 0.246, respectively).

4. Discussion

In this study of young Polynesians with recently diagnosed type 2 diabetes the mean glomerular filtration rate measured by an accurate isotopic technique was 56% higher than that found in age, sex and race-matched controls. The diabetic sub- jects were, however, significantly more obese than the control subjects and it is probable that both diabetes and obesity contributed to their very high glomerular filtration rates. Although the precise mechanism by which obesity increases the glomerular filtration rate is not clear, it appears to be related to the greater fat-free body mass of the obese. Correction of the glomerular filtration rate for a standard body surface area (1.73 m2 by con- vention) or for estimated fat-free body mass eliminates the effects of obesity on glomerular fil- tration rate [5,8]. When we corrected our glomerular filtration rate figures in this way then the excess in mean glomerular filtration rate in the diabetic subjects was reduced by approximately

one-third, but remained significantly elevated above control values. The definition of ‘hyper- filtration’ is somewhat arbitrary, but a value of greater than 140 ml/mm per 1.73 m2, which was the upper limit of normal in control subjects in the early studies in type 1 diabetes, has become accepted. By this criterion 73% of our diabetic sub- jects had hyperliltration, compared with only 13% of the controls. Hyperfiltration is a recognised fea- ture of type 1 diabetes, but whether or not it oc- curs in patients with type 2 diabetes has been the subject of much debate, with markedly differing results being found in different populations [ II- 181. These discrepancies are probably related to the differing ages at which various populations tend to develop diabetes. In Fig. 1 the results of studies of glomerular filtration rate in type 2 diabetes from various parts of the world are displayed [ II- 181. There is a clear negative corre- lation between the mean age of the population and the mean glomerular filtration rate (r = -0.838, P < 0.001). Thus the proportion of any popula- tion with glomerular tiltration rate > 140 ml/mm

_ 160.

“E - I? . 140. z _ E g120.

e a 100. S _ a = 80.

A

0 .

0 . 00 n

A + , E”

b0.838

0

A p=<O.OOl

+

“V .

30 io 50 $0 70

Mean age of Population (years)

Fig. 1. Mean glomerular filtration rate in published studies of type 2 diabetes, in relation to the mean age of the subjects studied. +, Denmark [ll]; 0, Italy [16]; 8, Wales [12]; A France [l?]; m USA - blacks [13]; 0, Denmark 1181; 0, USA - pima [ 141; A, New Zealand - Polynesians (this study); V, Brazil [IS]. In all the studies glomerular filtration rate was

measured by isotopic techniques, except for [11] which used creatinine clearance. In three studies [ 12,13,17] the population has been subdivided into various age groups.

R. Bruce et al. /Diabetes Res. Clin. Pratt. 25 (1994) 155-160 159

per 1.73 m2 will be lower the greater is the aver- age age of the subjects. As Marre has suggested the definition of hypertiltration, rather than being an arbitrary figure such as > 140 ml/mm per 1.73 m2, should perhaps take into account the normal age- related decline in renal function [ 171. The very high glomerular filtration rate we found in our Polynesian subjects is thus likely to be a reflection of their youth. A possible additional factor in our subjects may have been that they were non-fasting when studied and glomerular filtration rate may increase with high protein meals. However, this effect is unlikely to have raised the glomerular Iil- tration rate by more than 14% [19].

Glomerular hypertiltration may predispose to or initiate the structural and functional alterations in glomerular capillaries which develop into clini- cal diabetic nephropathy [l-3]. Although there is a high prevalence of diabetic renal disease amongst Polynesians with type 2 diabetes [20] the link with hyperfiltration is not necessarily causal. Prospec- tive studies are clearly required to determine whether hyperfiltration will ultimately lead to the acceleration of diabetic renal disease in this group. It is worth noting that even in these non-hyper- tensive, asymptomatic, subjects albumin excretion rates were significantly increased above control values, indicating that glomerular dysfunction may certainly accompany hyperfiltration. This increase in albumin excretion persisted even when the increased glomerular filtration rate of the diabetic subjects was taken into account and albuminuria was expressed as the fractional excre- tion. Prospective studies are also required to deter- mine if hyperfiltration in type 2 diabetes can be reduced by improved glycaemic control. In type 1 diabetes hyperfiltration is most marked in subjects with disease of recent onset and the short duration of diabetes in our subjects may also have con- tributed to their very high glomerular filtration rates. In both type 1 and type 2 diabetes hyper- filtration becomes less marked with increasing disease duration [ 13,2 11.

We conclude that young Polynesians with type 2 diabetes of recent onset commonly have very high rates of glomerular filtration and increased rates of albumin excretion. The high rates of glomerular filtration are partly related to obesity

in this group and correcting the measured glomerular filtration rate to a standard surface area tends to obscure the obesity-related compo- nent. The long-term significance of hyperfiltration in this population, which has a high prevalence of diabetic renal disease, needs to be assessed in lon- gitudinal studies.

Acknowledgements

We are grateful to the Maurice and Phyllis Paykel Trust for their support of this research and to the AMP Insurance Company who funded the Diabetes Research Fellowship held by Dr. R. Bruce. This study was presented at the Annual Scientific Meeting of the New Zealand Society for the Study of Diabetes, Hamilton, New Zealand, 1992.

References

Ill

121

131

141

[51

161

171

181

191

Hostetter, T.H., Rennke, H.G. and Brenner, B.M. (1982) The case for intrarenal hypertension in the initiation and progression of diabetic and other nephropathies. Am. J. Med. 72, 375-380. Viberti, G.C. and Wiseman, M.J. (1986) The kidney in diabetes: significance of the early abnormalities. Clin. Endocrinol. Metab. 15, 753-782. Rudberg, S., Persson, B. and Dahlquist, G. (1992) In- creased glomerular filtration rate as a predictor of diabetic nephropathy. An 8 year prospective study. Kid- ney Int. 41, 822-828. Scragg, R., Baker, J, Metcalf, P. and Dryson, E. (1991) Prevalence of diabetes mellitus and impaired glucose tol- erance in a New Zealand multiracial workforce. NZ Med. J. 104, 395-397. Stokholm, K.H., Brechner-Mortensen, J. and Hoilund- Carlsen, P.F. (1980) Increased glomerular filtration rate and adrenocortical function in obese women. Int. J. Obe- sity 4, 57-63. Cundy, T.F., Nixon, D., Berkahn, L. and Baker, J. (1992) Measuring the albumin excretion rate: agreement between methods and biological variability. Diabetes Med. 9, 138-143. Mosteller, R.D. (1987) Simplified calculation of body surface area. N. Engl. J. Med. 317, 1098. Salaxar, D.E. and Corcoran, G.B. (1988) Predicting creatinine clearance and renal drug clearance in obese patients from estimated fat-free body mass. Am. J. Med. 84, 1053-1060. Brphner-Mortensen, J. (1972) A simple method for the determination of glomerular filtration rate. Stand. J. Clin. Lab. Invest. 30, 271-274.

160 R. Bruce et al. /Diabetes Res. Clin. Pratt. 25 (1994) 155-160

[IO] Chantler, C., Gamett, E.S., Parsons, V. and Veall, N. (1969) Glomerular filtration rate measurement in man by the single injection method using “Cr EDTA. Clin. Sci. 37, 169-180.

[ll] Damsgaard, E.M. and Mogensen, C.E. (1986) Micro- albuminuria in elderly hyperglycaemic patients and con- trols. Diabetes Med. 3, 430-435.

[12] Vora, J.P., Dolben, J., Dean, J.D. et al. (1992) Renal he- modynamics in newly presenting non-insulin-dependent diabetes mellitus. Kidney Int. 41, 829-835.

[13] Lebovitz, H.E. and Pahnisano, J.A. (1990) Cross- sectional analysis of renal function in black Americans with NIDDM. Diabetes Care 13, 1186-l 190.

[14] Myers, B.D., Nelson, R.G., Williams, G.W. et al. (1991) Glomerular function in Pima Indians with non-insulin- dependent diabetes mellitus of recent onset. J. Clin. In- vest. 88, 524-530.

[I51 Silveiro, S.B., Friedman, R. and Gross, J.L. (1993) Glomerular hypertiltration in NIDDM patients without overt proteinuria. Diabetes Care 16, 115-l 19.

[16] Gambardella, S., Frontoni, S., Fehci, M.G. et al. (1990) Efficacy of antihypertensive treatment with indapamide

in patients with non-insulin dependent diabetes and per- sistent microalbuminuria. Am. J. Cardiol. 65,46H-50H.

[17] Marre, M., Hallab, M., Roy, J., Lejeune, J.J., Jallet, P. and Fressinaud, P. (1992) Glomerular hyperfiltration in Type I, Type II and secondary diabetes. J. Diabetes Comp. 6, 19-24.

1181 Nielsen, S., Rehhng., M., Schmita, A. and Mogensen, C.E. (1993) Systolic blood pressure relates to decline of glomerular filtration rate in Type II diabetes. Diabetes Care 16, 1427-1432.

[I91 Rodriguez-Iturbe, B., Herrera, J. and Garcia, R. (1988) Relationship between glomerular filtration rate and renal blood flow at different levels of protein-induced hyper- tension in man. Clin. Sci. 74, I l- 15.

[20] Scragg, R. (1989) Diabetes mellitus: secular trends in morbidity in New Zealand, 1966-1985. In: Contempor- ary Health Issues, Department of Health, Wellington, pp, 23-29.

[21] Bognetti, E., Meschi, F., Bonfanti, R., Gianolli, L. and Chiumello, G. (1993) Decrease of glomerular hypertiltra- tion in short-term diabetic adolescents without microalbuminuria. Diabetes Care 16, 120-124.