Dm
ORIGINAL ARTICLES
Urinary Growth Hormone Excretion During Puberty in Type 1 (Insulindependent) Diabetes Mellitus P. Hourda, J.A.
Edgeb, D.B. Dungerb, N. Dalton‘,
R. Edwards“
“North East Thames Regional lrnmunoassay Unit, St Bartholomew’s Hospita/, London, bDepartment of Paediatrics, john Radcliffe Hospital, Oxford, CNationwide Kidney Research Laboratories, Guy’s Hospital, London, UK
The excretion of urinary growth hormone was measured -y a highly sensitive direct immunoradiometric assay i n a cross-sectional study during puberty in 70 children with Type 1 (insulin-dependent) diabetes mellitus and 94 normal children. In normal children (n = 24) and diabetic children (n = 17) overnight urinary growth hormone excretion correlated significantly with the mean overnight plasma concentration ( r = 0.70, p < 0.001, and r = 0.70, p < O.OOl), indicating that urinary GH excretion reflects the circulating endogenous GH level. Overnight urinary growth hormone excretion increased during puberty. In normal and in diabetic children there was a peak in boys at genital stage 4 (both p < 0.01), and in girls at breast stage 2 (both p < 0.02). The diabetic children excreted more urinary growth hormone than the normal chidren at every pubertal stage. Excretion of albumin, retinol binding protein and N-acetyl-P-Dglucosarninidase was measured in urine from 38 diabetic children. Urinary growth hormone correlated weakly with urinary albumin ( r = 0.49, p < 0.01), retinol binding protein ( r = 0.42, p < 0.011, and Kacetyl-P-Dglucosaminidase ( r = 0.43, p < 0.01). Urinary GH excretion was not related to blood glucose control (HbA,) in boys (n= 31) or girls ( n = 39). The measurement of urinary growth hormone provides an assessment of endogenous growth hormone during puberty in normal and diabetic children. However, caution must be exercised in interpreting urinary growth hormone data from diabetic patients with increased excretion of albumin and retinol binding protein. KEY WORDS
Type 1 diabetes mellitus Tubular function
Introduction There has recently been a reappraisal of the role of growth hormone (GH) in mediating some of the day-today problems encountered in the management of Type 1 (insulin-dependent) diabetes mellitus, particularly in childhood and adolescence.’ Elevated GH concentrations - ~ result in during puberty in Type 1 d i a b e t e ~ ~may insulin insensitivity and contribute to impaired metabolic is also evidence that excessive GH c ~ n t r o l . ~There -~ secretion may be implicated in the development of the long-term microvascular complications associated with Type 1 d i a b e t e ~ . ’ , ~ ~ , ’ ~ Frequent measurements of GH levels may prove useful in the management of Type 1 diabetes and in understanding the pathological significance of GH in this disease. We have developed an immunoradiometric assay (IRMA) for the direct measurement of GH in unextracted urine and have shown that it accurately reflects physiological variations in plasma GH levels.12,’3 This has been confirmed by several other investigators
Correspondence to: P. Hourd, N.E.T.R.I.A., 51-53 Bartholornew Close, St Bartholornew’s Hospital, London EClA 7BE, UK
0742-3071 I911030237-06$05.00
0 1991 by John Wiley & Sons, Ltd.
Growth hormone Puberty Urinary albumin
who have reported the development and use of urinary GH assays, and these have highlighted the advantages of such a rnea~urement.’~-’~ In a preliminary study it was suggested that urinary G H is increased in Type 1 diabetes but the effect of renal function on urinary GH excretion was not assessed.12 Studies of urinary GH were therefore extended to assess physiological changes in G H excretion throughout puberty in normal children and children with Type 1 diabetes. The relationship of GH excretion to urinary albumin, and the tubular markers, retinol binding protein (RBP), and N-acetyl-P-D-glucosaminidase (NAG) have also been assessed.
Patients and Methods Patients In a preliminary study which examined the relationship between serum GH and urinary GH, 41 children collected 53 timed overnight urine samples during 12-h overnight plasma secretory profiles (blood was drawn continuously and collected in 15-min aliquots, 2000-0800 h). Of these subjects 24 were healthy children of normal stature (non-diabetic siblings recruited to act as controls; 11 Accepted 12 November 1990 DIABETIC MEDICINE, 1991; 8: 237-242
237
DT17
ORIGINAL ARTICLES female, 13 male; age 10.7-17.2 years), and 17 were children who had Type 1 diabetes (9 female, 8 male; age 10.7-1 5.9 years). To complete a cross-sectional study of urinary G H excretion during puberty, a further 70 timed overnight urine samples were obtained from 70 normal children (55 healthy schoolchildren and 15 non-diabetic siblings) and a further 53 urine samples from 53 children with Type 1 diabetes. Samples from a total of 94 normal children (45 male, 49 female) and 70 diabetic children (31 male, 39 female) were therefore analysed as part of this study. The mean duration of diabetes in the diabetic patients was 47 (range 1-93) months and the mean glycosylated haemoglobin (HbA,) level was 11.2 (6.7-16.6) % in the girls and 10.8 (7.8-15.9) % in the boys. Pubertal stage was assessed in all subjects by the method of Tanner” and for the purposes of analysis, breast stage was used in girls and genital stage in boys. Excretion of albumin, retinol binding protein, and Nacetyl-P-D-glucosaminidase was measured in 38 urine samples from the diabetic children (23 female, 13 male). The mean duration of diabetes was 25.2 (1-60) months and the HbA, level 11 .0 (6.9-1 6.7) % for these children. Albumin excretion was also measured in 57 urine samples from the normal children (31 female, 26 male) and retinol binding protein excretion in 67 urine samples from 22 of the normal children (8 female, 14 male). All the studies in normal and diabetic children were approved by the Central Oxford Research Ethics Committee and informed consent was obtained from the children and their parents.
Chromatography Samples of urine from two diabetic children, who had no evidence of renal dysfunction, were chromatographed as previously r e p ~ r t e d ; ’ ~1 millilitre fractions were assayed directly. Specificity for G H was confirmed by chromatographic identity of a single peak eluting in a molecular weight region corresponding to the monomeric 22 500 Dalton form of pituitary GH.
Assays Growth hormone was measured in 1 or 2 ml urine samples using a specific solid-phase immunoradiometric assay (IRMA) optimized for maximum ~ensitivity.’~ Sensitivity, calculated as 2.5 SD from zero G H concentration, ranged from 1 - 4 pU I-’. The standards were calibrated against the international Reference Preparation, 66/21 7 (National Institute for Biological Standards and Control, South Mimms, UK). The precision profile calculated on a single assay showed intra-assay coefficients of variation of 18, 6 and 3 % at concentrations of 6.25, 25, and 100 pU I-I, respectively. Urinary GH results were related to time and volume and expressed in terms of nU h-’. Serum G H concentrations were measured by an IRMA (NETRIA, London, UK); all samples from each 238
individual overnight profile were analysed in the same batch. The inter-assay coefficients of variation at G H concentrations of 3.5, 15.2, and 77.4 m U I-’ were 9.4, 7.7, and 10.4 %, respectively, and the intra-assay coefficients of variation at G H concentrations of 2.9, 14.5, and 69.4 m U I-I were 7.9, 2.0, and 3.4 %, respectively. Glycosylated haemoglobin (HbA1) was measured by electrophoresis, the normal range being 5.0-7.5 %. The concentration of albumin in urine was measured by radioimmunoassay using a commercially available kit (Diagnostic Product Corporation, Wallingford, UK); the concentration of retinol binding protein in urine by an enzyme linked immunosorbent assay using rabbit antisera (Dako, Kyoto, Japan); and the activity of N-acetyl-P-Dglucosaminidase in the urine by an automated colorimetric method using p-nitrophenyl-N-acetyl-P-D-glucosamide (Sigma, Poole, UK) as substrate.18 Urinary creatinine was measured in the Jaffe reaction. Results were expressed as a concentration ratio with respect to creatinine (urinary a1bumin:creatinine ratio in mg mmol-’ ; urinary retinol binding pr0tein:creatinine ratio in p g mmol-I, and urinary N-acetyl-P-D-glucosaminidase:creatinine ratio in pmol of p-nitro-phenylic acid h-’ mmol-’). Normal ranges were: albumin, retinol binding protein, 0.14-1.1 7 mg mmol-’; 1 .O-15.1 pg mmol-l, and N-acetyl-p-D-glucosaminidase 1.8-25.0 pmol h-’ m m ~ l - ’ . ’ ~
Specificity Validation of the IRMA for the measurement of GH in urine has been described previously.’ A negligible effect of interfering substances that may be specific to urine from diabetic patients was confirmed by recovery and dilution experiments. Recovery of known amounts (10, 20, 40, and 80 p U I F 1 ) of hGH (International Reference Preparation, 66/217) from 10 samples was: 99.2 f 9.4 (? SD) %, 93.5 f 8.7 %, 90.8 ? 6.2 %, and 88.3 9.0 %, respectively, Urine samples from 10 subjects were assayed neat and at a dilution of 1:2, 1 :4, and 1:8. Parallelism was shown for nine samples following a 1:2 dilution; results were within 1 SD of the mean (87.0 ? 7.3 %) and all were within 2 SD. Urinary GH was undetectable (< 3 pU I-’) following a 1 :4 or 1:8 dilution in these samples; this corresponded to the calculated value which was estimated below the minimum detection limit of the assay.
*
Statistical Analysis and Calculations Statistical analysis was performed using Minitab (State College, PA, USA). Serum GH, urinary GH, albumin, retinol binding protein, and N-acetyl-p-D-glucosaminidase were shown to be logarithmically normally distributed by normal order plots. Results were therefore logarithmically transformed before analysis. Associations between these variables were examined using linear regression P. HOURD ET Al.
Drn
ORIGINAL ARTICLES
and Pearson’s correlation coefficients. To eliminate the influence of sex and puberty in the correlations between urinary GH levels and HbA, the number of standard deviations of each G H level (the Z-score) from the log,, normalized mean for each pubertal stage and sex in the normal group was calculated. A t-test was used to compare the slope and intercept of the regression lines. Data from the puberty study were not transformed and were examined using non-parametric statistics. Differences between pubertal groups were evaluated with the Kruskal-Wallis one-way analysis of variance and subsequent comparisons between the prepubertal and pubertal groups were made with the Mann-Whitney U test. Renal clearance was calculated using the formula: clearance = uGH/sGH, where UGH is the urinary G H excretion rate ( n u min-’) and sGH is the mean serum CH concentration ( m u I-’) during the time of the urine collection.
ResuIts In those children who underwent overnight secretory profiles there was a highly significant correlation between mean overnight serum G H and urinary G H excretion in both normal children and in children with Type 1 diabetes mellitus; r = 0.70, p < 0.001 (n = 24) and r = 0.70, p < 0.001 (n = 171, respectively (Figure 1). There was no improvement in these correlations when urinary G H was expressed in terms of nU mmol-’ creatinine. The slope of the regression line for the normal children (urinary G H = 1.94 + 0.75 serum GH; slope SD = 0.18, intercept SD = 0.17) was greater ( p < 0.021, and the intercept lower ( p < 0.001) than that for the diabetic children (urinary G H = 2.39 0.64 serum GH; slope SD = 0.13, intercept SD = 0.17). The diabetic children had elevated levels of both mean serum G H and urinary G H when compared with puberty matched normal subjects.
+
--
5000
.-
!c
1000
-
*
me.
.
a
0
$
* * O
8 . .
C
.-C
*.
..a
3
g 0 0
0
500
0
I
0
0 . 0 0
0
3
0
0 0 0
0
100
e
00
0
L
...... . .
0
0
-
0
2
5
10
50
70
Serum GH (mu I-’
Figure 1. Relationship between urinary growth hormone excretion rate (nu h-’) and mean serum growth hormone concentration (mu I-,) during overnight secretory profiles (15min sampling, 2000-0800 h) in normal and diabetic children. Correlations in the diabetic (closed circles) and the normal (open circles) children were ( r = 0.70, p < 0.001) and ( r = 0.70, p < 0.001), respectively. When both groups of children were included correlation was r = 0.80, p < 0.001 URINARY CH IN TYPE 1 DIABETES
Renal clearance was significantly greater in the diabetic children (1.44 i 0.18 (* SE),pl min-’, n = 17) than the normal children (1.05 5 0.10 PI rnin-’, n = 24, p < 0.02) who underwent overnight secretory profiles. There was no significant correlation between the urinary concentration of G H ( p U 1 - l ) and the urinary flow rate (ml hkl) in either the diabetic or the normal population. In normal children there was a peak median urinary G H excretion noted in boys at genital stage 4 ( p < 0.001) and in girls at breast stage 2 ( p < 0.02) when compared with prepubertal levels (Figure 2). Similarly, in diabetic children there was a peak median urinary excretion at genital stage 4 in boys (p < 0.01). In girls the peak median level only reached statistical significance when breast stages 2 and 3 were combined ( p < 0.02) (Figure 2). At each pubertal stage the diabetic children as a group excreted more urinary G H than the non-diabetic children. When boys and girls were considered separately the diabetic boys excreted significantly more urinary G H than the normal boys at each pubertal stage except genital stage 1; G2 (p < 0.005), G4 (p < 0.011, and G5 (p < 0.05). A comparison of the urinary G H level at Tanner stage 3 in boys was not possible as the diabetic group contained only one subject. Similarly, the diabetic girls also excreted more urinary G H than the normal girls although the values did not reach statistical significance except at breast stage 5 (p < 0.05). In order to determine whether urinary G H excretion was related to diabetic control, as judged by HbA,, a Z-score for urinary G H excretion was calculated. Only six of the diabetic children had HbA, levels within the normal range. There was no significant correlation between the Z-score for urinary G H and HbAl in boys ( r = -0.19) or in girls ( r = 0.13). In addition, we found no significant correlation with duration of diabetes ( r = 0.24). In the diabetic children 16 % of urinary albumin, 21 % of retinol binding protein, and 76 % of N-acetyl-P-Dglucosaminidase values were greater than the upper limit of their respective ranges (log,,, 2SD above mean) for normal children. Urinary G H excretion correlated weakly with the excretion of albumin ( r = 0.49, p < 0.011, Nacetyl-P-D-glucosaminidase (r = 0.43, p < 0.01) and retinol binding protein (r = 0.42, p < 0.01) (Figure 3). However no correlation was evident in diabetic children with albumin and retinol binding protein excretion rates within the normal range (albumin < 0.92 mg mmol-’ creatinine, retinol binding protein < 12.3 pg mmol-’ creatinine), and where excretion of N-acetyl-P-D-glucosaminidase was < 106 pmol p-nitrophenyl h-’ mmol-’ creatinine. There was no significant correlation with albumin or retinol binding protein in normal subjects.
Discussion We have previously reported the development and validation of a highly sensitive IRMA for the measurement of G H in unextracted urine.I2 The specificity of the
239
DTT7
ORIGINAL ARTlCLfS
... .....
5000
t
.
looo[ T
I
:
~.~~~~~
1
1
0 01
.
0:
J
01
05
5
1
10
Albumin (mg mmol-’ creatinine) 5000
I
1
1
2
3
4
5
-!c 2 -
1000
100
. ... . ..... :. . .*.. . . . =’*. t
r
.
~
-
0
.
.
e.
0
I
.
a
1
,t L
1
.
1
~
5
1
~
~
1
10
50
~
I
~
100
500
NAG (pmol p-nitrophenylic acid h-’ mrnor’ creatinine) 5000 r
1000
-
100
-
.. *.
0
.
1
J.l
J
1
5
10
50
1W
RBP (pg rnmol-’ creatinine)
Figure 2. Urinary GH excretion ( n u hk’) in non-diabetic children of normal stature and children with Type 1 diabetes mellitus analysed by pubertal stage: genital stage 1-5 (Gl-5) in boys (top panel) and breast stage 1-5 (61-5) in girls (bottom SE for normal panel). Values are represented by mean children {open circles) and diabetic children (closed circles), and by median for normal children (open triangles) and diabetic children (closed triangles). The number of patients analysed at each pubertal stage was: normal boys G1 = 8, G2 = 12, G3 = 8,C4 = 9, and G5 = 8; normal girls 61 = 6, B2 = 6, 63 = 2, B4 = 13, and 65 = 22; diabetic boys G1 = 14, G2 = 5, G3 = 1, G4 = 4, and G5 = 7; diabetic girls B1 = 8, 62 = 4, B3 = 7, 84 = 11, and 65 = 9
assay, for urine from diabetic subjects, was confirmed by recovery and dilution experiments. Alternative or altered molecular forms of circulating G H have not been consistently found in Type 1 diabetes. In the present study, in the two subjects examined, urinary GH had a molecular size similar to the monomeric, 22 500 Dalton molecular weight form of pituitary GH. Similar chromato-
240
Figure 3. Relationship between urinary GH excretion ( n u h-’) and urinary excretion of albumin, N-acetyl-P-D-glucosaminidase (NAG) and retinol binding protein (RBP) in patients with Type 1 diabetes mellitus. Plot (log,,) of the correlation between urinary GH excretion and urinary albumin excretion ( r = 0.49, p < 0.01) (top panel), urinary N-acetyl-P-D-glucosaminidase excretion (r = 0.43, p < 0.01) (middle panel), urinary retinof binding protein excretion ( r = 0.42, p < 0.01) (bottom panel). There was no significant ( p < 0.05) correlation with urinary GH for albumin levels < 0.92 mg mmol-’ creatinine (indicated by arrow) (upper limit of normal range = 1.1 7 mg mmol-’ creatinine), N-acetyl-P-D-glucosaminidase levels < 106 yrnol p-nitrophenyl h-’ mmol-’ (upper limit of normal range = 25 ymol p-nitrophenyl h-’ mmol-’ creatinine) and retinol binding protein levels of < 12.3 pg mmol-’ creatinine (upper limit of normal range = 15.1 pg mmol-’)
graphic profiles have been found in urine from normaIl2 and acromegalic subjects12 and in patients with diabetic ketoacidosis.20 A urinary GH measurement therefore reflects circulating endogenous G H in patients with Type 1 diabetes. P. HOURD ET Al.
Dm
ORIGINAL ARTICLES
excretion. Increased urinary excretion of GH is primarily Using this assay we have shown a significant correlation between urinary GH excretion and mean plasma levels the result of either an increase in plasma concentration during 12-h secretory profiles in normal and diabetic or a decrease in the renal tubular reabsorption of G H children. The diabetic group had higher serum levels (inhibition of adsorption to tubular cell membrane or and excreted more GH in the urine than the normal tubular cell damage). Osmotic diuresis i s also reported group. Elevated urinary GH levels have also recently to influence tubular r e a b s o r p t i ~ n in ; ~our ~ study however, been reported in adults with Type 1 diabetes, when the urine flow rate for the diabetic patients was not compared with normal adults.23 Raised circulating G H significantly different from normal subjects and was not levels in poorly controlled diabetic patients have been correlated with GH concentration in either group. In the normal kidney, at filtered loads below the saturation confirmed in several s t u d i e ~ . ~ ,These ~ , ~ , elevated ~~ levels are characterized by an altered pattern of pulsatile G H level of the reabsorption process, a direct relationship release with increased pulse amplitude and with or between the plasma concentration and the urinary without increased pulse f r e q u e n ~ y . This ~,~~ is in contrast excretion of GH should occur, as we have shown in to the situation in normal puberty where G H release is Figure 1. pulse amplitude modulated and periodicity remains Since the advent of urinary GH assays it has become unchanged.24The pathophysiological significance of this increasingly clear that aberrations in renal function may be a confounding problem in the interpretation of urinary increased frequency is unclear; obviously a urinary GH measurement can only reflect the sum of the pulse G H results for the evaluation of hypothalamic-pituitary function. Indeed Matsuura et a / . 3 3have recently suggested amplitudes and would be of limited value should the frequency of pulses be diagnostically more important. that the measurement of urinary GH may be a useful We have confirmed that in children with Type 1 marker of tubular function. Investigation of patients with severe renal i n s u f f i ~ i e n c ygross , ~ ~ a l b ~ m i n u r i a chronic ,~~ diabetes mellitus there is a normal rise in urinary GH renal failure, low molecular weight proteinuria or nephexcretion during puberty. The overnight urinary G H level was elevated at each pubertal stage when compared with rotic syndrome33 have shown vastly elevated urinary GH the normal population although these raised levels were levels. It is not known whether more subtle changes in much less significant in girls. Moreover, the peak urinary renal function such as occur in Type 1 diabetes mellitus CH level in diabetic children occurred at the same affect the excretion of GH. A proportion of the diabetic children that we studied pubertal stage as in the normal children for both boys had urinary excretion of albumin, retinol binding protein, and girls. The smaller increase in urinary G H excretion in diabetic girls is of interest and requires confirmation. and N-acetyl-P-D-glucosaminidase, which was greater than the normal range (log,, 2SD above mean), in It has been suggested that GH levels are higher in diabetic patients with poor metabolic c o n t r ~ I , ~ , ~ , ~ ~keeping * ~ ~ * ~with other reported data in these subjects.” We yet we have found no correlation between the Z-score measured these proteins, and urinary GH, in single urine of urinary G H excretion and HbA, as an index of blood samples; the intra-individual variation of their excretion was therefore not accounted for, and may in part, be glucose control. Blood glucose control (fasting blood responsible for the large scatter in the data.2 % * 3 5 Despite glucose levels) has also been shown to be unrelated to this, in diabetic children urinary GH excretion correlated urinary G H excretion in Type 1 and Type 2 diabetic This may indicate that no relationship exists weakly with the excretion of all three of these proteins. between HbA, and G H secretion, but an alternative Un Ii ke reti no1 binding protein, N-acetyl-P-D-gl ucosaminexplanation might be that renal clearance of G H is idase is not filtered at the glomerulus but is liberated into the tubule lumen with renal cell damage. The variable. It has been suggested that the raised levels of CH in Type 1 diabetes mellitus may be due to reduced excretion of N-acetyl-P-D-glucosaminidase was abnormal in most of these diabetic children (76 %), which renal clearance of GH,27-29 although not all reports may be indicative of proximal tubular damage. support In our study there were significant We found no significant correlation between retinol differences between both the slope and the intercept of binding protein, albumin, and urinary G H excretion in the regression lines in the diabetic and normal groups. normal children. Similarly, when samples from diabetic Despite a steeper slope, the intercept was significantly children which had levels of retinol binding protein and lower for the normal population indicating that, within albumin greater than the upper limit of the normal range the levels of urinary G H studied, renal clearance was were omitted from the analysis no significant correlation increased in the diabetic children. This increase was ( p < 0.05) was observed. This finding would suggest that confirmed when renal clearance was calculated from the patients with renal ’leakage’ of albumin and/or retinol individual data. Increased clearance of other proteins binding protein may also ’leak’ GH; urinary excretion of has been noted in diabetic patients.” GH in these patients may therefore only partially reflect Proteins with the size and charge of monomeric GH plasma levels. undergo glomerular filtration and reabsorption in the We must conclude that although the correlation proximal tubule of the kidney.31 In patients with severe between urinary G H excretion and integrated plasma renal tubular d y s f ~ n c t i o n there ~ ~ . ~is~ a strong positive concentrations i s reasonably good, part of the variance correlation between urinary G H and p,-microglobulin URINARY CH IN TYPE 1 DIABETES
24 1
DTC7
ORIGINAL ARTICLES i n the excretion of urinary GH can be attributed t o tubular dysfunction. Caution needs t o be exercised in interpreting urinary GH data from diabetic patients w i t h elevated albumin and RBP excretion but, nevertheless, urinary GH may have a useful role in determining longterm changes in plasma GH secretion in Type 1 diabetes mellitus.
17. 18.
19.
20.
References ~~
1. 2. 3. 4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
242
Holly JMP, Amiel SA, Sandhu RR, Rees LH, Wass ]AH. The role of growth hormone in diabetes 1 Endocrinol 1988; 118: 353-364. Bloch CA, Clemons P, Sperling MA. Puberty decreases insulin sensitivity. j Pediatr 1987; 110: 481-487. Hansen AP. Serum growth hormone patterns in juvenile diabetes. Dan Med Bull 1972; 19: 1-32. Hayford JT, Danney MM, Hendrix JA, Thompson RG. Integrated concentrations of growth hormone in juvenile diabetes. Diabetes 1980; 29: 391-398. Press M, Tamborlane WV, Sherwin RS. Importance of raised growth hormone levels in mediating the metabolic derangements of diabetes. N Engl 1 Med 1984; 310: 81C815. Amiel SA, Sherwin RS, Simonsen DC, Lauritano AA, Tamborlane WV. Impaired insulin action in puberty. A contributing factor to poor glycaemic control in adolescents with diabetes. N Engll Med 1986; 315: 215-219. Gerich JE. Rationale for inhibition of growth hormone in the management of the diabetic patient. Scand j Castroenterol 1986; 21 : 154-1 57. Hindmarsh P, Di Silvio L, Pringle PI, Kurtz AB, Brook CGD. Changes in serum insulin concentration during puberty and their relationship to growth hormone. Clin Endocrinol ( 0 x 0 1988; 28: 381-388. Rizza RA, Mandarin LJ, Gerich JE. Effects of growth hormone on insulin action in man: mechanism of insulin resistance, impaired suppression of glucose production and impaired stimulation of glucose utilization. Diabetes 1982; 31: 663-669. Barnes AJ, Kohner EM, Johnston DG, Alberti KGMM. Severe retinopathy and mild carbohydrate intolerance: possible role of insulin deficiency and elevated circulating growth hormone. Lancet 1985; i: 1465-1 468. Lundbaek K, Christensen NJ, Jensen VA, et a/. Pathogenesis of diabetic angiopathy and growth hormone. Dan Med Bull 1971; 18: 1-7. Edge JA, Hourd P, Edwards R, Dunger DB. Urinary growth hormone during puberty in normal and diabetic children. Clin Endocrinol (Oxf) 1989; 30: 41 3-420. Hourd P, Edwards R. Measurement of human growth hormone in urine: development and validation of a sensitive and specific assay. 1 Endocrinol 1989; 121: 167-1 75. Albini CH, Quattrin T, Vandlen RL, MacGillivray MH. Quantitation of urinary growth hormone in children with normal and subnormal growth. Pediatr Res 1988; 23: 89-92. Okuno A, Yano K, ltoh Y, et a/. Urine growth hormone determinations compared with other methods in the assessment of growth hormone secretion. Acta Paediatr Scand Suppl 1987; 337: 74-81. Sukegawa I, Hizuka N, Takano K, et a/. Urinary growth hormone measurements are useful for evaluating
21. 22.
23.
24.
25.
26.
27.
28.
29. 30. 31.
32.
33. 34.
35.
endogenous growth hormone secretion. 1 Clin Endocrinol Metab 1988; 66: 11 19-1 123. Tanner JM. Growth at Adolescence. 3rd edn. Oxford: Blackwell, 1962. Marulin D. Rapid colourimetric assay of P-D-galactosidase and N-acetyl-P-D-glucosaminidase in human urine. Clin Chirn Acta 1976; 73: 453-461. Gibb DM, Dunger DB, Levin M, Shah V, Smith C, Barratt TM. Early markers of the renal complications of insulin dependent diabetes mellitus in children. Arch Dis Child 1989; 64: 984-991. Livesey JH, Scott RS, Donald RA. Urinary growth hormone in diabetic ketoacidosis. Horm Metab Res 1979; 1 1 : 142-7 46. Horner JM, Kemp SF, Hintz RL. Growth hormone and somatomedin in insulin dependent diabetes mellitus. 1 Clin Endocrinol Metab I98 I; 53: 1 148-1 153. Asplin CM, Faria ACS, Carlsen EC, et a/. Alterations in the pulsatile mode of growth hormone release in men and women with insulin dependent diabetes. 1 Clin Endocrinol Metab 1989; 69: 239-245. Suzuki K, Miyaka H, Suzuki T, Kajinuma H. Evaluation and clinical applications of measurement of urinary growth hormone in diabetic subjects. Diabetes 1989; 28: 1567-1 572. Hindmarsh PC, Matthews DR, Brook CGD. Growth hormone secretion in children determined by time series analysis. Clin Endocrinol ( 0 x 0 1988; 29: 35-44. Salardi S, Cacciari E, Ballardini D, et a/. Relationship between growth factors (somatomedin C and growth hormone) and body development, metabolic control and retinal changes in children and adolescents with insulin dependent diabetes mellitus. Diabetes 1986; 35: 832-836. Vigneri R, Squatrito S, Pezzino V, Filett S, Branca S, Polosa P. Growth hormone levels in diabetes: correlation with clinical control of the disease. Diabetes 1972; 25: 167-1 72. Lipman RL, Taylor AL, Conly P, Mintz DH. Metabolic clearance rate of growth hormone in juvenile diabetes mellitus. Diabetes 1972; 21 : 175-1 77. Sperling MA, Wollesen F, De Lamater PV. Daily production and metabolic clearance rate of growth hormone in juvenile diabetes mellitus. Diabetologia 1973; 9: 380-383. Taylor AL, Finster JL, Mintz DH. Metabolic clearance and production rates of human growth hormone. 1 C h lnvest 1969; 48: 2349-2358. Navalesi R, Pilo A, Vigneri R. Growth hormone kinetics in diabetic patients. Diabetes 1975; 24: 31 7-327. Maack T, Johnson V, Kau ST, Figueiredo J, Sigulem D. Renal filtration, transport, and metabolism of low molecular weight proteins: A review. Kidney Int 1979; 16: 251-270. Jung K, Pergande M, Porstmann B, Porstmann T. Diuresisdependent excretion of low molecular mass proteins in urine: B2-microglobulin, lysozyme and ribonuclease. %and 1 Clin Lab lnvest 1988; 48: 33-37. Matsuura M, Kitagawa T, Hashida S, Murakami Y. High urinary levels of human growth hormone in children with renal tubular dysfunction. Nephron 1989; 51: 561-562. Hattori N, Kato Y, Murakami Y, et a/. Urinary growth hormone levels measured by an ultrasensitive enzyme immunoassay in patients with renal insufficiency. / Clin Endocrinol 1988; 66: 727-732. Sikegawa I, Hizuka N, Takano K, et a/. Measurement of nocturnal urinary growth hormone values. Acta Endocrinol 1989; 121 : 290-296.
P. HOURD ET Al.
ORIGINAL ARTICLES
Urinary Growth Hormone Excretion During Puberty in Type 1 (Insulindependent) Diabetes Mellitus P. Hourda, J.A.
Edgeb, D.B. Dungerb, N. Dalton‘,
R. Edwards“
“North East Thames Regional lrnmunoassay Unit, St Bartholomew’s Hospita/, London, bDepartment of Paediatrics, john Radcliffe Hospital, Oxford, CNationwide Kidney Research Laboratories, Guy’s Hospital, London, UK
The excretion of urinary growth hormone was measured -y a highly sensitive direct immunoradiometric assay i n a cross-sectional study during puberty in 70 children with Type 1 (insulin-dependent) diabetes mellitus and 94 normal children. In normal children (n = 24) and diabetic children (n = 17) overnight urinary growth hormone excretion correlated significantly with the mean overnight plasma concentration ( r = 0.70, p < 0.001, and r = 0.70, p < O.OOl), indicating that urinary GH excretion reflects the circulating endogenous GH level. Overnight urinary growth hormone excretion increased during puberty. In normal and in diabetic children there was a peak in boys at genital stage 4 (both p < 0.01), and in girls at breast stage 2 (both p < 0.02). The diabetic children excreted more urinary growth hormone than the normal chidren at every pubertal stage. Excretion of albumin, retinol binding protein and N-acetyl-P-Dglucosarninidase was measured in urine from 38 diabetic children. Urinary growth hormone correlated weakly with urinary albumin ( r = 0.49, p < 0.01), retinol binding protein ( r = 0.42, p < 0.011, and Kacetyl-P-Dglucosaminidase ( r = 0.43, p < 0.01). Urinary GH excretion was not related to blood glucose control (HbA,) in boys (n= 31) or girls ( n = 39). The measurement of urinary growth hormone provides an assessment of endogenous growth hormone during puberty in normal and diabetic children. However, caution must be exercised in interpreting urinary growth hormone data from diabetic patients with increased excretion of albumin and retinol binding protein. KEY WORDS
Type 1 diabetes mellitus Tubular function
Introduction There has recently been a reappraisal of the role of growth hormone (GH) in mediating some of the day-today problems encountered in the management of Type 1 (insulin-dependent) diabetes mellitus, particularly in childhood and adolescence.’ Elevated GH concentrations - ~ result in during puberty in Type 1 d i a b e t e ~ ~may insulin insensitivity and contribute to impaired metabolic is also evidence that excessive GH c ~ n t r o l . ~There -~ secretion may be implicated in the development of the long-term microvascular complications associated with Type 1 d i a b e t e ~ . ’ , ~ ~ , ’ ~ Frequent measurements of GH levels may prove useful in the management of Type 1 diabetes and in understanding the pathological significance of GH in this disease. We have developed an immunoradiometric assay (IRMA) for the direct measurement of GH in unextracted urine and have shown that it accurately reflects physiological variations in plasma GH levels.12,’3 This has been confirmed by several other investigators
Correspondence to: P. Hourd, N.E.T.R.I.A., 51-53 Bartholornew Close, St Bartholornew’s Hospital, London EClA 7BE, UK
0742-3071 I911030237-06$05.00
0 1991 by John Wiley & Sons, Ltd.
Growth hormone Puberty Urinary albumin
who have reported the development and use of urinary GH assays, and these have highlighted the advantages of such a rnea~urement.’~-’~ In a preliminary study it was suggested that urinary G H is increased in Type 1 diabetes but the effect of renal function on urinary GH excretion was not assessed.12 Studies of urinary GH were therefore extended to assess physiological changes in G H excretion throughout puberty in normal children and children with Type 1 diabetes. The relationship of GH excretion to urinary albumin, and the tubular markers, retinol binding protein (RBP), and N-acetyl-P-D-glucosaminidase (NAG) have also been assessed.
Patients and Methods Patients In a preliminary study which examined the relationship between serum GH and urinary GH, 41 children collected 53 timed overnight urine samples during 12-h overnight plasma secretory profiles (blood was drawn continuously and collected in 15-min aliquots, 2000-0800 h). Of these subjects 24 were healthy children of normal stature (non-diabetic siblings recruited to act as controls; 11 Accepted 12 November 1990 DIABETIC MEDICINE, 1991; 8: 237-242
237
DT17
ORIGINAL ARTICLES female, 13 male; age 10.7-17.2 years), and 17 were children who had Type 1 diabetes (9 female, 8 male; age 10.7-1 5.9 years). To complete a cross-sectional study of urinary G H excretion during puberty, a further 70 timed overnight urine samples were obtained from 70 normal children (55 healthy schoolchildren and 15 non-diabetic siblings) and a further 53 urine samples from 53 children with Type 1 diabetes. Samples from a total of 94 normal children (45 male, 49 female) and 70 diabetic children (31 male, 39 female) were therefore analysed as part of this study. The mean duration of diabetes in the diabetic patients was 47 (range 1-93) months and the mean glycosylated haemoglobin (HbA,) level was 11.2 (6.7-16.6) % in the girls and 10.8 (7.8-15.9) % in the boys. Pubertal stage was assessed in all subjects by the method of Tanner” and for the purposes of analysis, breast stage was used in girls and genital stage in boys. Excretion of albumin, retinol binding protein, and Nacetyl-P-D-glucosaminidase was measured in 38 urine samples from the diabetic children (23 female, 13 male). The mean duration of diabetes was 25.2 (1-60) months and the HbA, level 11 .0 (6.9-1 6.7) % for these children. Albumin excretion was also measured in 57 urine samples from the normal children (31 female, 26 male) and retinol binding protein excretion in 67 urine samples from 22 of the normal children (8 female, 14 male). All the studies in normal and diabetic children were approved by the Central Oxford Research Ethics Committee and informed consent was obtained from the children and their parents.
Chromatography Samples of urine from two diabetic children, who had no evidence of renal dysfunction, were chromatographed as previously r e p ~ r t e d ; ’ ~1 millilitre fractions were assayed directly. Specificity for G H was confirmed by chromatographic identity of a single peak eluting in a molecular weight region corresponding to the monomeric 22 500 Dalton form of pituitary GH.
Assays Growth hormone was measured in 1 or 2 ml urine samples using a specific solid-phase immunoradiometric assay (IRMA) optimized for maximum ~ensitivity.’~ Sensitivity, calculated as 2.5 SD from zero G H concentration, ranged from 1 - 4 pU I-’. The standards were calibrated against the international Reference Preparation, 66/21 7 (National Institute for Biological Standards and Control, South Mimms, UK). The precision profile calculated on a single assay showed intra-assay coefficients of variation of 18, 6 and 3 % at concentrations of 6.25, 25, and 100 pU I-I, respectively. Urinary GH results were related to time and volume and expressed in terms of nU h-’. Serum G H concentrations were measured by an IRMA (NETRIA, London, UK); all samples from each 238
individual overnight profile were analysed in the same batch. The inter-assay coefficients of variation at G H concentrations of 3.5, 15.2, and 77.4 m U I-’ were 9.4, 7.7, and 10.4 %, respectively, and the intra-assay coefficients of variation at G H concentrations of 2.9, 14.5, and 69.4 m U I-I were 7.9, 2.0, and 3.4 %, respectively. Glycosylated haemoglobin (HbA1) was measured by electrophoresis, the normal range being 5.0-7.5 %. The concentration of albumin in urine was measured by radioimmunoassay using a commercially available kit (Diagnostic Product Corporation, Wallingford, UK); the concentration of retinol binding protein in urine by an enzyme linked immunosorbent assay using rabbit antisera (Dako, Kyoto, Japan); and the activity of N-acetyl-P-Dglucosaminidase in the urine by an automated colorimetric method using p-nitrophenyl-N-acetyl-P-D-glucosamide (Sigma, Poole, UK) as substrate.18 Urinary creatinine was measured in the Jaffe reaction. Results were expressed as a concentration ratio with respect to creatinine (urinary a1bumin:creatinine ratio in mg mmol-’ ; urinary retinol binding pr0tein:creatinine ratio in p g mmol-I, and urinary N-acetyl-P-D-glucosaminidase:creatinine ratio in pmol of p-nitro-phenylic acid h-’ mmol-’). Normal ranges were: albumin, retinol binding protein, 0.14-1.1 7 mg mmol-’; 1 .O-15.1 pg mmol-l, and N-acetyl-p-D-glucosaminidase 1.8-25.0 pmol h-’ m m ~ l - ’ . ’ ~
Specificity Validation of the IRMA for the measurement of GH in urine has been described previously.’ A negligible effect of interfering substances that may be specific to urine from diabetic patients was confirmed by recovery and dilution experiments. Recovery of known amounts (10, 20, 40, and 80 p U I F 1 ) of hGH (International Reference Preparation, 66/217) from 10 samples was: 99.2 f 9.4 (? SD) %, 93.5 f 8.7 %, 90.8 ? 6.2 %, and 88.3 9.0 %, respectively, Urine samples from 10 subjects were assayed neat and at a dilution of 1:2, 1 :4, and 1:8. Parallelism was shown for nine samples following a 1:2 dilution; results were within 1 SD of the mean (87.0 ? 7.3 %) and all were within 2 SD. Urinary GH was undetectable (< 3 pU I-’) following a 1 :4 or 1:8 dilution in these samples; this corresponded to the calculated value which was estimated below the minimum detection limit of the assay.
*
Statistical Analysis and Calculations Statistical analysis was performed using Minitab (State College, PA, USA). Serum GH, urinary GH, albumin, retinol binding protein, and N-acetyl-p-D-glucosaminidase were shown to be logarithmically normally distributed by normal order plots. Results were therefore logarithmically transformed before analysis. Associations between these variables were examined using linear regression P. HOURD ET Al.
Drn
ORIGINAL ARTICLES
and Pearson’s correlation coefficients. To eliminate the influence of sex and puberty in the correlations between urinary GH levels and HbA, the number of standard deviations of each G H level (the Z-score) from the log,, normalized mean for each pubertal stage and sex in the normal group was calculated. A t-test was used to compare the slope and intercept of the regression lines. Data from the puberty study were not transformed and were examined using non-parametric statistics. Differences between pubertal groups were evaluated with the Kruskal-Wallis one-way analysis of variance and subsequent comparisons between the prepubertal and pubertal groups were made with the Mann-Whitney U test. Renal clearance was calculated using the formula: clearance = uGH/sGH, where UGH is the urinary G H excretion rate ( n u min-’) and sGH is the mean serum CH concentration ( m u I-’) during the time of the urine collection.
ResuIts In those children who underwent overnight secretory profiles there was a highly significant correlation between mean overnight serum G H and urinary G H excretion in both normal children and in children with Type 1 diabetes mellitus; r = 0.70, p < 0.001 (n = 24) and r = 0.70, p < 0.001 (n = 171, respectively (Figure 1). There was no improvement in these correlations when urinary G H was expressed in terms of nU mmol-’ creatinine. The slope of the regression line for the normal children (urinary G H = 1.94 + 0.75 serum GH; slope SD = 0.18, intercept SD = 0.17) was greater ( p < 0.021, and the intercept lower ( p < 0.001) than that for the diabetic children (urinary G H = 2.39 0.64 serum GH; slope SD = 0.13, intercept SD = 0.17). The diabetic children had elevated levels of both mean serum G H and urinary G H when compared with puberty matched normal subjects.
+
--
5000
.-
!c
1000
-
*
me.
.
a
0
$
* * O
8 . .
C
.-C
*.
..a
3
g 0 0
0
500
0
I
0
0 . 0 0
0
3
0
0 0 0
0
100
e
00
0
L
...... . .
0
0
-
0
2
5
10
50
70
Serum GH (mu I-’
Figure 1. Relationship between urinary growth hormone excretion rate (nu h-’) and mean serum growth hormone concentration (mu I-,) during overnight secretory profiles (15min sampling, 2000-0800 h) in normal and diabetic children. Correlations in the diabetic (closed circles) and the normal (open circles) children were ( r = 0.70, p < 0.001) and ( r = 0.70, p < 0.001), respectively. When both groups of children were included correlation was r = 0.80, p < 0.001 URINARY CH IN TYPE 1 DIABETES
Renal clearance was significantly greater in the diabetic children (1.44 i 0.18 (* SE),pl min-’, n = 17) than the normal children (1.05 5 0.10 PI rnin-’, n = 24, p < 0.02) who underwent overnight secretory profiles. There was no significant correlation between the urinary concentration of G H ( p U 1 - l ) and the urinary flow rate (ml hkl) in either the diabetic or the normal population. In normal children there was a peak median urinary G H excretion noted in boys at genital stage 4 ( p < 0.001) and in girls at breast stage 2 ( p < 0.02) when compared with prepubertal levels (Figure 2). Similarly, in diabetic children there was a peak median urinary excretion at genital stage 4 in boys (p < 0.01). In girls the peak median level only reached statistical significance when breast stages 2 and 3 were combined ( p < 0.02) (Figure 2). At each pubertal stage the diabetic children as a group excreted more urinary G H than the non-diabetic children. When boys and girls were considered separately the diabetic boys excreted significantly more urinary G H than the normal boys at each pubertal stage except genital stage 1; G2 (p < 0.005), G4 (p < 0.011, and G5 (p < 0.05). A comparison of the urinary G H level at Tanner stage 3 in boys was not possible as the diabetic group contained only one subject. Similarly, the diabetic girls also excreted more urinary G H than the normal girls although the values did not reach statistical significance except at breast stage 5 (p < 0.05). In order to determine whether urinary G H excretion was related to diabetic control, as judged by HbA,, a Z-score for urinary G H excretion was calculated. Only six of the diabetic children had HbA, levels within the normal range. There was no significant correlation between the Z-score for urinary G H and HbAl in boys ( r = -0.19) or in girls ( r = 0.13). In addition, we found no significant correlation with duration of diabetes ( r = 0.24). In the diabetic children 16 % of urinary albumin, 21 % of retinol binding protein, and 76 % of N-acetyl-P-Dglucosaminidase values were greater than the upper limit of their respective ranges (log,,, 2SD above mean) for normal children. Urinary G H excretion correlated weakly with the excretion of albumin ( r = 0.49, p < 0.011, Nacetyl-P-D-glucosaminidase (r = 0.43, p < 0.01) and retinol binding protein (r = 0.42, p < 0.01) (Figure 3). However no correlation was evident in diabetic children with albumin and retinol binding protein excretion rates within the normal range (albumin < 0.92 mg mmol-’ creatinine, retinol binding protein < 12.3 pg mmol-’ creatinine), and where excretion of N-acetyl-P-D-glucosaminidase was < 106 pmol p-nitrophenyl h-’ mmol-’ creatinine. There was no significant correlation with albumin or retinol binding protein in normal subjects.
Discussion We have previously reported the development and validation of a highly sensitive IRMA for the measurement of G H in unextracted urine.I2 The specificity of the
239
DTT7
ORIGINAL ARTlCLfS
... .....
5000
t
.
looo[ T
I
:
~.~~~~~
1
1
0 01
.
0:
J
01
05
5
1
10
Albumin (mg mmol-’ creatinine) 5000
I
1
1
2
3
4
5
-!c 2 -
1000
100
. ... . ..... :. . .*.. . . . =’*. t
r
.
~
-
0
.
.
e.
0
I
.
a
1
,t L
1
.
1
~
5
1
~
~
1
10
50
~
I
~
100
500
NAG (pmol p-nitrophenylic acid h-’ mrnor’ creatinine) 5000 r
1000
-
100
-
.. *.
0
.
1
J.l
J
1
5
10
50
1W
RBP (pg rnmol-’ creatinine)
Figure 2. Urinary GH excretion ( n u hk’) in non-diabetic children of normal stature and children with Type 1 diabetes mellitus analysed by pubertal stage: genital stage 1-5 (Gl-5) in boys (top panel) and breast stage 1-5 (61-5) in girls (bottom SE for normal panel). Values are represented by mean children {open circles) and diabetic children (closed circles), and by median for normal children (open triangles) and diabetic children (closed triangles). The number of patients analysed at each pubertal stage was: normal boys G1 = 8, G2 = 12, G3 = 8,C4 = 9, and G5 = 8; normal girls 61 = 6, B2 = 6, 63 = 2, B4 = 13, and 65 = 22; diabetic boys G1 = 14, G2 = 5, G3 = 1, G4 = 4, and G5 = 7; diabetic girls B1 = 8, 62 = 4, B3 = 7, 84 = 11, and 65 = 9
assay, for urine from diabetic subjects, was confirmed by recovery and dilution experiments. Alternative or altered molecular forms of circulating G H have not been consistently found in Type 1 diabetes. In the present study, in the two subjects examined, urinary GH had a molecular size similar to the monomeric, 22 500 Dalton molecular weight form of pituitary GH. Similar chromato-
240
Figure 3. Relationship between urinary GH excretion ( n u h-’) and urinary excretion of albumin, N-acetyl-P-D-glucosaminidase (NAG) and retinol binding protein (RBP) in patients with Type 1 diabetes mellitus. Plot (log,,) of the correlation between urinary GH excretion and urinary albumin excretion ( r = 0.49, p < 0.01) (top panel), urinary N-acetyl-P-D-glucosaminidase excretion (r = 0.43, p < 0.01) (middle panel), urinary retinof binding protein excretion ( r = 0.42, p < 0.01) (bottom panel). There was no significant ( p < 0.05) correlation with urinary GH for albumin levels < 0.92 mg mmol-’ creatinine (indicated by arrow) (upper limit of normal range = 1.1 7 mg mmol-’ creatinine), N-acetyl-P-D-glucosaminidase levels < 106 yrnol p-nitrophenyl h-’ mmol-’ (upper limit of normal range = 25 ymol p-nitrophenyl h-’ mmol-’ creatinine) and retinol binding protein levels of < 12.3 pg mmol-’ creatinine (upper limit of normal range = 15.1 pg mmol-’)
graphic profiles have been found in urine from normaIl2 and acromegalic subjects12 and in patients with diabetic ketoacidosis.20 A urinary GH measurement therefore reflects circulating endogenous G H in patients with Type 1 diabetes. P. HOURD ET Al.
Dm
ORIGINAL ARTICLES
excretion. Increased urinary excretion of GH is primarily Using this assay we have shown a significant correlation between urinary GH excretion and mean plasma levels the result of either an increase in plasma concentration during 12-h secretory profiles in normal and diabetic or a decrease in the renal tubular reabsorption of G H children. The diabetic group had higher serum levels (inhibition of adsorption to tubular cell membrane or and excreted more GH in the urine than the normal tubular cell damage). Osmotic diuresis i s also reported group. Elevated urinary GH levels have also recently to influence tubular r e a b s o r p t i ~ n in ; ~our ~ study however, been reported in adults with Type 1 diabetes, when the urine flow rate for the diabetic patients was not compared with normal adults.23 Raised circulating G H significantly different from normal subjects and was not levels in poorly controlled diabetic patients have been correlated with GH concentration in either group. In the normal kidney, at filtered loads below the saturation confirmed in several s t u d i e ~ . ~ ,These ~ , ~ , elevated ~~ levels are characterized by an altered pattern of pulsatile G H level of the reabsorption process, a direct relationship release with increased pulse amplitude and with or between the plasma concentration and the urinary without increased pulse f r e q u e n ~ y . This ~,~~ is in contrast excretion of GH should occur, as we have shown in to the situation in normal puberty where G H release is Figure 1. pulse amplitude modulated and periodicity remains Since the advent of urinary GH assays it has become unchanged.24The pathophysiological significance of this increasingly clear that aberrations in renal function may be a confounding problem in the interpretation of urinary increased frequency is unclear; obviously a urinary GH measurement can only reflect the sum of the pulse G H results for the evaluation of hypothalamic-pituitary function. Indeed Matsuura et a / . 3 3have recently suggested amplitudes and would be of limited value should the frequency of pulses be diagnostically more important. that the measurement of urinary GH may be a useful We have confirmed that in children with Type 1 marker of tubular function. Investigation of patients with severe renal i n s u f f i ~ i e n c ygross , ~ ~ a l b ~ m i n u r i a chronic ,~~ diabetes mellitus there is a normal rise in urinary GH renal failure, low molecular weight proteinuria or nephexcretion during puberty. The overnight urinary G H level was elevated at each pubertal stage when compared with rotic syndrome33 have shown vastly elevated urinary GH the normal population although these raised levels were levels. It is not known whether more subtle changes in much less significant in girls. Moreover, the peak urinary renal function such as occur in Type 1 diabetes mellitus CH level in diabetic children occurred at the same affect the excretion of GH. A proportion of the diabetic children that we studied pubertal stage as in the normal children for both boys had urinary excretion of albumin, retinol binding protein, and girls. The smaller increase in urinary G H excretion in diabetic girls is of interest and requires confirmation. and N-acetyl-P-D-glucosaminidase, which was greater than the normal range (log,, 2SD above mean), in It has been suggested that GH levels are higher in diabetic patients with poor metabolic c o n t r ~ I , ~ , ~ , ~ ~keeping * ~ ~ * ~with other reported data in these subjects.” We yet we have found no correlation between the Z-score measured these proteins, and urinary GH, in single urine of urinary G H excretion and HbA, as an index of blood samples; the intra-individual variation of their excretion was therefore not accounted for, and may in part, be glucose control. Blood glucose control (fasting blood responsible for the large scatter in the data.2 % * 3 5 Despite glucose levels) has also been shown to be unrelated to this, in diabetic children urinary GH excretion correlated urinary G H excretion in Type 1 and Type 2 diabetic This may indicate that no relationship exists weakly with the excretion of all three of these proteins. between HbA, and G H secretion, but an alternative Un Ii ke reti no1 binding protein, N-acetyl-P-D-gl ucosaminexplanation might be that renal clearance of G H is idase is not filtered at the glomerulus but is liberated into the tubule lumen with renal cell damage. The variable. It has been suggested that the raised levels of CH in Type 1 diabetes mellitus may be due to reduced excretion of N-acetyl-P-D-glucosaminidase was abnormal in most of these diabetic children (76 %), which renal clearance of GH,27-29 although not all reports may be indicative of proximal tubular damage. support In our study there were significant We found no significant correlation between retinol differences between both the slope and the intercept of binding protein, albumin, and urinary G H excretion in the regression lines in the diabetic and normal groups. normal children. Similarly, when samples from diabetic Despite a steeper slope, the intercept was significantly children which had levels of retinol binding protein and lower for the normal population indicating that, within albumin greater than the upper limit of the normal range the levels of urinary G H studied, renal clearance was were omitted from the analysis no significant correlation increased in the diabetic children. This increase was ( p < 0.05) was observed. This finding would suggest that confirmed when renal clearance was calculated from the patients with renal ’leakage’ of albumin and/or retinol individual data. Increased clearance of other proteins binding protein may also ’leak’ GH; urinary excretion of has been noted in diabetic patients.” GH in these patients may therefore only partially reflect Proteins with the size and charge of monomeric GH plasma levels. undergo glomerular filtration and reabsorption in the We must conclude that although the correlation proximal tubule of the kidney.31 In patients with severe between urinary G H excretion and integrated plasma renal tubular d y s f ~ n c t i o n there ~ ~ . ~is~ a strong positive concentrations i s reasonably good, part of the variance correlation between urinary G H and p,-microglobulin URINARY CH IN TYPE 1 DIABETES
24 1
DTC7
ORIGINAL ARTICLES i n the excretion of urinary GH can be attributed t o tubular dysfunction. Caution needs t o be exercised in interpreting urinary GH data from diabetic patients w i t h elevated albumin and RBP excretion but, nevertheless, urinary GH may have a useful role in determining longterm changes in plasma GH secretion in Type 1 diabetes mellitus.
17. 18.
19.
20.
References ~~
1. 2. 3. 4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
242
Holly JMP, Amiel SA, Sandhu RR, Rees LH, Wass ]AH. The role of growth hormone in diabetes 1 Endocrinol 1988; 118: 353-364. Bloch CA, Clemons P, Sperling MA. Puberty decreases insulin sensitivity. j Pediatr 1987; 110: 481-487. Hansen AP. Serum growth hormone patterns in juvenile diabetes. Dan Med Bull 1972; 19: 1-32. Hayford JT, Danney MM, Hendrix JA, Thompson RG. Integrated concentrations of growth hormone in juvenile diabetes. Diabetes 1980; 29: 391-398. Press M, Tamborlane WV, Sherwin RS. Importance of raised growth hormone levels in mediating the metabolic derangements of diabetes. N Engl 1 Med 1984; 310: 81C815. Amiel SA, Sherwin RS, Simonsen DC, Lauritano AA, Tamborlane WV. Impaired insulin action in puberty. A contributing factor to poor glycaemic control in adolescents with diabetes. N Engll Med 1986; 315: 215-219. Gerich JE. Rationale for inhibition of growth hormone in the management of the diabetic patient. Scand j Castroenterol 1986; 21 : 154-1 57. Hindmarsh P, Di Silvio L, Pringle PI, Kurtz AB, Brook CGD. Changes in serum insulin concentration during puberty and their relationship to growth hormone. Clin Endocrinol ( 0 x 0 1988; 28: 381-388. Rizza RA, Mandarin LJ, Gerich JE. Effects of growth hormone on insulin action in man: mechanism of insulin resistance, impaired suppression of glucose production and impaired stimulation of glucose utilization. Diabetes 1982; 31: 663-669. Barnes AJ, Kohner EM, Johnston DG, Alberti KGMM. Severe retinopathy and mild carbohydrate intolerance: possible role of insulin deficiency and elevated circulating growth hormone. Lancet 1985; i: 1465-1 468. Lundbaek K, Christensen NJ, Jensen VA, et a/. Pathogenesis of diabetic angiopathy and growth hormone. Dan Med Bull 1971; 18: 1-7. Edge JA, Hourd P, Edwards R, Dunger DB. Urinary growth hormone during puberty in normal and diabetic children. Clin Endocrinol (Oxf) 1989; 30: 41 3-420. Hourd P, Edwards R. Measurement of human growth hormone in urine: development and validation of a sensitive and specific assay. 1 Endocrinol 1989; 121: 167-1 75. Albini CH, Quattrin T, Vandlen RL, MacGillivray MH. Quantitation of urinary growth hormone in children with normal and subnormal growth. Pediatr Res 1988; 23: 89-92. Okuno A, Yano K, ltoh Y, et a/. Urine growth hormone determinations compared with other methods in the assessment of growth hormone secretion. Acta Paediatr Scand Suppl 1987; 337: 74-81. Sukegawa I, Hizuka N, Takano K, et a/. Urinary growth hormone measurements are useful for evaluating
21. 22.
23.
24.
25.
26.
27.
28.
29. 30. 31.
32.
33. 34.
35.
endogenous growth hormone secretion. 1 Clin Endocrinol Metab 1988; 66: 11 19-1 123. Tanner JM. Growth at Adolescence. 3rd edn. Oxford: Blackwell, 1962. Marulin D. Rapid colourimetric assay of P-D-galactosidase and N-acetyl-P-D-glucosaminidase in human urine. Clin Chirn Acta 1976; 73: 453-461. Gibb DM, Dunger DB, Levin M, Shah V, Smith C, Barratt TM. Early markers of the renal complications of insulin dependent diabetes mellitus in children. Arch Dis Child 1989; 64: 984-991. Livesey JH, Scott RS, Donald RA. Urinary growth hormone in diabetic ketoacidosis. Horm Metab Res 1979; 1 1 : 142-7 46. Horner JM, Kemp SF, Hintz RL. Growth hormone and somatomedin in insulin dependent diabetes mellitus. 1 Clin Endocrinol Metab I98 I; 53: 1 148-1 153. Asplin CM, Faria ACS, Carlsen EC, et a/. Alterations in the pulsatile mode of growth hormone release in men and women with insulin dependent diabetes. 1 Clin Endocrinol Metab 1989; 69: 239-245. Suzuki K, Miyaka H, Suzuki T, Kajinuma H. Evaluation and clinical applications of measurement of urinary growth hormone in diabetic subjects. Diabetes 1989; 28: 1567-1 572. Hindmarsh PC, Matthews DR, Brook CGD. Growth hormone secretion in children determined by time series analysis. Clin Endocrinol ( 0 x 0 1988; 29: 35-44. Salardi S, Cacciari E, Ballardini D, et a/. Relationship between growth factors (somatomedin C and growth hormone) and body development, metabolic control and retinal changes in children and adolescents with insulin dependent diabetes mellitus. Diabetes 1986; 35: 832-836. Vigneri R, Squatrito S, Pezzino V, Filett S, Branca S, Polosa P. Growth hormone levels in diabetes: correlation with clinical control of the disease. Diabetes 1972; 25: 167-1 72. Lipman RL, Taylor AL, Conly P, Mintz DH. Metabolic clearance rate of growth hormone in juvenile diabetes mellitus. Diabetes 1972; 21 : 175-1 77. Sperling MA, Wollesen F, De Lamater PV. Daily production and metabolic clearance rate of growth hormone in juvenile diabetes mellitus. Diabetologia 1973; 9: 380-383. Taylor AL, Finster JL, Mintz DH. Metabolic clearance and production rates of human growth hormone. 1 C h lnvest 1969; 48: 2349-2358. Navalesi R, Pilo A, Vigneri R. Growth hormone kinetics in diabetic patients. Diabetes 1975; 24: 31 7-327. Maack T, Johnson V, Kau ST, Figueiredo J, Sigulem D. Renal filtration, transport, and metabolism of low molecular weight proteins: A review. Kidney Int 1979; 16: 251-270. Jung K, Pergande M, Porstmann B, Porstmann T. Diuresisdependent excretion of low molecular mass proteins in urine: B2-microglobulin, lysozyme and ribonuclease. %and 1 Clin Lab lnvest 1988; 48: 33-37. Matsuura M, Kitagawa T, Hashida S, Murakami Y. High urinary levels of human growth hormone in children with renal tubular dysfunction. Nephron 1989; 51: 561-562. Hattori N, Kato Y, Murakami Y, et a/. Urinary growth hormone levels measured by an ultrasensitive enzyme immunoassay in patients with renal insufficiency. / Clin Endocrinol 1988; 66: 727-732. Sikegawa I, Hizuka N, Takano K, et a/. Measurement of nocturnal urinary growth hormone values. Acta Endocrinol 1989; 121 : 290-296.
P. HOURD ET Al.
