Serum osteocalcin concentrations in children with chronic renalinsufficiency who are not undergoing dialysis

Serum osteocalcin concentrations in children with chronic renal insufficiency who are not undergoing dialysis A a r o n L. Friedman, MD, J o n a t h a n D. Heiliczer, MD, C a r e n M. G u n d b e r g , PhD, Robert H. K. Mak, MD, Phb, Frank G. Boineau, MD, Paul T. McEnery, MD, James C. M. C h a n , MD, a n d m e m b e r s of the G r o w t h Failure in Children With Renal Diseases Study From the Department of Pediatrics, Universityof Wisconsin, Madison, the Department of Pediatrics, Rush-Presbyterian-St.Luke's Medical Center, Chicago, the Department of Orthopedics and Rehabilitation, Yale UniversitySchool of Medicine, New Haven, the Divisionof Nephroiogy, Children's Hospitalof Los Angeles, the Department of Pediatrics,Tulane University,New Orleans, the Universityof Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio, and the Department of Pediatrics, Medical College of Virginia, Richmond This report describes the serum o s t e o c a l c i n values in children with mild to moderate, but r e l a t i v e l y stable, renal dysfuntion followed in the Growth Failure in Children With Renal Diseases Study. This report is derived from data o b t a i n e d during the control period (6 months) before the initiation of vitamin D therapy. Up to three measurements per patient were obtained. Serum osteocalcin concentration was c o m p a r e d with creatinine c l e a r a n c e (glomerular filtration rate) c a l c u l a t e d by the Schwartz formula; with serum concentrations of a l k a l i n e phosphatase, parathyroid hormone, and bicarbonate; and with the percentages of the r e c o m m e n d e d dietary a l l o w a n c e s of calories and protein ingested. By standard correlation techniques, there a p p e a r e d to be an inverse correlation between c a l c u l a t e d creatinine c l e a r a n c e and serum osteocalcin concentration, and a direct correlation between serum osteocalcin and parathyroid hormone values. However, when we e m p l o y e d a statistical technique that takes into account repeated measures in the same patient, no correlation was found between c a l c u l a t e d glomerular filtration rate and serum osteocalcin concentration, and no direct correlation was found between serum osteocalcin and parathyroid hormone values. The lack of a correlation between c a l c u l a t e d glomerular filtration rate and serum osteocalcin values may be due to large fluctuations in the serum osteocalcin concentration, even though renal function is relatively stable. (J PEDIATR4990;446:$55-9)

Osteocalcin is a bone-derived protein containing residues of the vitamin K-dependent amino acid 7-carboxyglutamic acid (bone Gla protein) and synthesized by osteoblasts)

Supported by National Institutes of Health grants R01 DK 31370 and R01 DK 32431. Reprint requests: Aaron L. Friedman, MD, Department of Pediatrics, Clinical Science Center H4/472, 600 Highland Ave., Madison, WI 53792. 9/0/17215

Osteocalcin is found mainly in bone, but nanomolar concentrations circulate in the blood, and the circulating portion has been the focus of attention as a noninvasive measure of bone turnover. 2 We report on the serum osteocalcin levels in children with mild to moderate chronic renal failure who have not undergone dialysis, and we demonstrate the correlation (or lack thereof) between serum osteocalcin values and other biochemical or nutritional measurements that might affect or reflect bone turnover in patients with chronic renal failure.

$55

S 56

I

Friedman et al.

GFR PTH RDA

Glomerular filtration rate Parathyroid hormone Recommended dietary allowance

The Journal of Pediatrics February 1990

[

METHODS All children included in this report met the entry criteria for the Growth Failure in Children With Renal Diseases (GFRD) Study, a multicenter, collaborative clinical trial. The entry criteria included (1) age 18 months to 10 years, (2) glomerular filtration rate as estimated from the Schwartz formula 3 of 20 to 75 ml/min/1.73 m 2, and (3) no previous treatment with vitamin D analogs. The collaborative study was designed to assess the relative benefits of two vitamin D analogs, dihydrotachysterol and 1,25-dihydroxy-vitamin D3. On enrolling in the study, patients undergo a 6-month control period during which growth is measured, dietary histories are obtained, and blood studies are performed. This report includes data from the control period only. The characteristics of the patient population are more completely reviewed in a companion article4 published with this report. Serum osteocalcin concentration was measured by radioimmunoassay as previously reported by Gundberg et al. 5 The normal value for children of this age is 40 + 9 ng/ml (mean _+ SD). Samples were assayed in duplicate at three different assay dilutions. The interassay variation is less than 15% and the intraassay variation less than 10%. The assay is sensitive to 0.1 ng/ml. 5 Mid-molecule C-terminal parathyroid hormone was assayed by the Nichols Institute (San Juan Capistrano, Calif.). The normal range is 50 to 300 pg/ml. Serum alkaline phosphatase levels (measured in international units per liter) and serum bicarbonate levels (measured in milliequivalents per liter) were determined at the individual institutions by standard techniques. Diet histories were obtained from the patients and their caretakers and evaluated by registered dietitians. The diet histories were used to determine the percentages of the recommended dietary allowances of calories and protein consumed by individual patients. Statistical analyses were performed by the GFRD Study Data Coordinating Center at the Medical College of Virginia, Richmond. The analyses performed include the Pearson coefficient of correlation and repeated measures analysis of variance. 6 The latter analysis allows for the fact that observations of a single patient are not independent and is preferable to ~the former, more standard analysis. RESULTS The data are a composite of all osteocalcin measurements obtained during the G F R D Study control period. Seventy-

three patients from whom 149 osteocalcin samples were obtained have completed the control period. There was no correlation between serum osteocatcin concentration and age (Fig. 1, A). There appeared to be an inverse correlation between serum osteocalcin concentration and calculated creatinine clearance (Fig. 1, B). At a calculated creatinine clearance of approximately 50 ml/min/1.73 m 2, the serum osteocalcin value rose out of the normal range. For the entire study population, the mean (___SD) serum osteocalcin value was 48 _+ 36 ng/ml, and the mean ( + SD) calculated creatinine clearance was 52 + 22 ml/min/1.73 m 2. By the Pearson coefficient of correlation, there was an inverse correlation between serum osteocalcin concentration and calculated creatinine clearance. However, if repeated measures in a single patient are taken into account, no correlation between serum osteocalcin concentration and calculated creatinine clearance was seen. The serum osteocalcin value rose with the PTH value until the latter reached very high levels, at which point the serum osteocalcin concentration seemed to reach a plateau (Fig. 1, C). The mean ( + S D ) PTH level for the entire study population was 1044 + 1427 pg/ml. As noted above, by the Pearson coefficient of correlation, a direct correlation between serum osteocalcin concentration and calculated creatinine clearance was found, but repeated measures analysis of variance did not verify this correlation. In Fig. 1, D and E, the serum osteocalcin concentrations versus the percent RDA for calories and the percent R D A for protein, respectively, are demonstrated. By the Pearson coefficient of correlation, no correlation could be found between the serum osteocalcin values and these variables. The population mean ( + S D ) for the percent RDA for calories was 80 + 21%, and the population mean ( + S D ) for the percent R D A for protein was 157+55%. Serum osteocalcin values plotted against serum bicarbonate values are shown in Fig. 2. The mean ( _+SD) serum bicarbonate value for the population was 21 _+ 4 mEq/L. The Pearson coefficient of correlation showed a weak correlation, whereas analysis of variance of repeated measures showed no significant correlation. Repeated measures analysis of variance did show a direct correlation between serum osteocalcin values and serum alkaline phosphatase activity (Fig. 1, F). The mean ( + SD) serum alkaline phosphatase activity for the population was 272 ___ 120 U / L . DISCUSSION Studies involving patients (mainly adults) with chronic renal disease have shown that the serum osteocalcin concentration increases with declining renal function, but generally not until the G F R falls to less than 20 ml/min/1.73 m2.1,79 This inverse relationship does not reflect bone turnover exclusively. Osteocalcin is filtered by the kidney,

Volume 116 Number 2

Osteocalcin in children with renal insufficiency

S S7

180

A

150 120 -z

90 -

.r

60 -

g

I.u 0

9

9

l9

:.

.

.;,-,

309 d~ 9 I

"-

...:-

I

9 I

3

I

I

5

.

.r

....-;,

g 0

90

I

7

.-

..,. -~.~ 9" ~.4:' i _

60

I

II

_

C

9

g

0

•

i

~PI i

i

i

I 5000

I

I

I

180 -

I

I

,9 I

I

GLOMERULAR FILTRATION ml/min/173sq rn

RATE,

D

9 , . ~ .-.. ,. "*...

o% I%,

9

.%

9

I

I

I

~-

";. I

60

t

~9

I

I

I O0

I

140

% RDA CALORIES

E

F

180

"

o,, 9 9

i

:,.;?.;.;..._

9

o

9

6O "

j I 0

~

.'" ;::.

so: 0

70

/

t50

90 "

0

I

I10

0 i 20

15o-" 120

"

%

I

90

3o 1

I0000

9

I

50

SERUM PARATHYROID HORMONE, pg/ml

i

.',.

I

t80

~

a

o

I

30

16~

IZO

::1.

I

".

,5o

150

w

".

".,..'l:

"1%% %

I

I0

AGE IN Y E A R S

8o~]

V

9 ..

I

.'.

, ".a,.'il~"

9 ", :

0

I

9

9- ' Y ' . . ~

30

9 1ee 4 99 I

B

120

uJ

..

9 I

z"

le

:: -

9 :

0-"

9

9

150

I I00

I ZOO %

I 300

90 60

3o o I 400

I 500

RDA PROTEIN

"" : 2 ! ! : el

0

I

0 SERUM

l

I

9 I

9 9

I

I

300

600

ALKALINE

PHOSPHATASE,

I

I

900 IU/L

Fig. I. A•S•rum•ste•ca•einp1•ttedagainstage•B•S•rum•ste•calcinva•uespl•ttedagainstGFR(•alcu]atedcreatinin• clearance). Line at 40 ng/ml defines normal mean for children of this age ( _+SD = 9 ng/ml [data from C.M.G.]). Pearson coefficient of correlation -0.46, p _<0.0001; analysis of variance of repeated measures 0.59 (not significant). C, Serum osteocalcin values plotted against mid-molecule C-terminal PTH values. Pearson coefficient of correlation 0.54, p _< 0.0001; analysis of variance of repeated measures 0.75 (no t significant). D, Serum osteocalcin values plotted against %RDA for calories. Pearson coefficient of correlation, 0.14 (not significant); analysis of variance of repeated measures 0.26 (not significant). E, Serum osteocalcin values plotted against %RDA for protein. Pearson coefficient of correlation 0.68 (not significant); analysis of variance of repeated measures 0.91 (not significant). F, Serum osteocalcin values plotted against serum: alkaline phosphatase values. Pearson coefficient of correlation 0.25 (not significant); analysis of variance of repeated measures 0.05 (p _< 0.05).

with further proximal tubule degradation, and therefore one would expect a n increase in serum osteocalcin concentration as the G F R falls. However, the reported rise in the s e r u m osteocalcin concentration, noted when the G F R falls to less t h a n 20 m l / m i n , was not linear; rather, the serum osteocalcin concentration rose sharply as the G F R fell.8.Some authors have suggested t h a t this hyperbolic increase in the ser u m osteocalcin level reflects not only the fall in G F R b u t also an increase in the a m o u n t of osteocalcin t h a t is released

from bone as a result of the more rapid bone turnover, s S e r u m osteocalcin levels are somewhat age related.I, ]0, ] 1 For example, they are higher in the neonatal period t h a n during slower growth phases. In association with the adolescent growth spurt, there is another increase in s e r u m osteocalcin values. A f t e r adolescence the levels of osteocalcin fall and remain relatively constant. After m e n o p a u s e the serum osteocalcin concentration tends to rise slightly. In men beyond the age of 60 years, osteocalcin levels have been

S 58

Friedman et al.

The Journal of Pediatrics February 1990

E

~o -I ~ m

I,

9

.,,

"

Ill

9

J .:'+.. ":s, 2~

,,

:

.'.-

..'i

9

011

9

9

.

,

s ~ 1 7" 6 .o .. ."

"'~

o'"

I

I

I

I

I

I

I

0

20

40

60

80

I00

120

I

140

I

160

I

180

SERUM OSTEOCALCIN, ng/ml

Fig. 2, Serum osteocalcin values plotted against serum bicarbonate values. ~ r s o n coefficient of correlation -0.19 (p ~ 0.055); analysis of variance of repeated measures 0.17 (not significant).

reported to increase, decrease, or remain constant. 1 There is no evidence of seasonality in circulating osteocalcin values. However, there is a circadian difference of approximately 6 ng/ml, insufficient to explain differences seen in patients with renal insufficiency (data from C.M.G.). Our data from the age group 18 months to 10 years show no correlation between age and serum osteocalcin levels. This finding is consistent with previous studies, l~ 11 Children with chronic renal disease and other metabolic bone diseases have elevated serum osteocalcin levels. 11"13 However, in a report on a small number of children with hypoparathyroidism, osteocalcin levels were generally normal or low, 11 and in children with osteoporosis the serum osteocalcin levels were variable. The data reported here constitute the most extensive information to date on serum osteocalcin values in children with chronic renal disease. The composite data demonstrate that the serum osteocalcin concentration begins to rise out of the normal range earlier in the course of renal disease in children than in adults with renal insufficiency. In our pediatric patients as a group, the serum osteocalcin values rose out of the normal range when the calculated creatinine clearance fell to less than 50 ml/min/1.73 m 2. In the report by Delmas et al. 8 on adult patients, the serum osteocalcin level rose above normal at a GFR of 20 ml/min/1.73 m 2 (determined by clearance of iothalamate labeled with iodine 125). This finding may indicate a difference between children and adults, perhaps related to bone turnover. However, the different techniques used to measure GFR may also contribute to the difference between our findings and those repe-ted by Delmas et al. Although a standard statistical technique for analysis (Pearson coefficient of correlation) did demonstrate a statisticaily significant inverse correlation between the calculated GFR and serum osteocalein values, observations

included repeated measures in the same patient, and a significant correlation was found when the repeated measure analysis was used. This result may be explained by the finding that the calculated GFR did not vary appreciably during the control period but the serum osteocalcin values did. Within a single patient, the calculated GFR measurement fluctuated approximately 20%, with the largest range (in one patient) being 14.7 to 33.3 ml/min/1.73 m 2. However, osteocalcin values within a single Patient fluctuated more frequently and to a greater degree. In one instance a patient's calculated GFR ranged from 32 to 48 ml/ min/1.73 m 2 but the osteocalcin value ranged from 2.2 to 50.5 ng/ml. In a second example the calculated GFR ranged from 39 to 48 ml/min/1.73 m 2 but the serum osteocalcin value ranged from 20 to 135 ng/ml. The repeated measures analysis takes these fluctuations into account and therefore did not show any significant correlation between calculated GFR and serum osteocalcin concentration. A direct correlation between serum osteocalcin concentration and alkaline phosphatase activity was seen. All data from previous studies of adults and children have shown an inverse correlation between serum osteocalcin concentration and GFR and, in some instances, a correlation between serum osteocalcin concentration and serum PTH concentration or alkaline phosphatase activity.8,9 There are at least tw o possible explanations for the discrepancy between our findings and those of previous studies. Our report is the first (to our knowledge) of repeated measurements of serum osteocalcin levels in a large number of pediatric patients with chronic renal insufficiency. It is unclear whether repeated measurements in other studies would have uncovered the same variability in serum osteocalcin values. Another possibility is that the levels of serum osteocatcin in children are higher than in adults, likely reflecting a higher degree of bone turnover. Our repeated sampling may

Volume 116 Number 2

Osteocalcin in children with renal insufficiency

have measured both the usual variation in bone turnover t h a t might be expected in children a n d the variation t h a t m a y be the result of their renal disease. The children's renal disease did not fluctuate appreciably during the 6 - m o n t h control period, and therefore a clear correlation between serum osteocalcin concentration and G F R was not found. This may also explain why s e r u m values for osteocalcin a n d alkaline phosphatase activity did show a direct correlation. T h e fact t h a t the s e r u m osteocalcin level varied as m u c h as it did m a y be a true reflection of the bone turnover in children with renal disease. D a t a from the therapeutic trial with vitamin D analogs m a y help to determine whether ser u m osteocalcin concentrations in children with chronic renal disease will prove to be more useful in assessing bone activity t h a n the present biochemical markers.

10.

REFERENCES

11.

1. Lian JB, Gundberg CM. Osteocalcin: biochemical considerations and clinical applications. Clin Orthop Rel Res 1988; 226:267-91. 2. Kruse K, Kracht U. Evaluation of serum osteocalcin as an index of altered bone metabolism. Eur J Pediatr 1986;145:27-33. 3. Schwartz G J, Haycock GB, Edelmann CM Jr, Spitzer A. A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine. Pediatrics 1976; 58:259-63. 4. Chan JCM, Boineau FG, Ruley E J, et al. Descriptions of the

5.

6. 7. 8.

9.

12.

13.

S 59

participating centers and patient population in the Growth Failure in Children With Renal Diseases Study. J PEDtATR 1990;116:$24-7. Gundberg CM, Hauschka PV, Lian JB, Gallop PM. Osteocalcin isolation, characterization, and detection. Methods Enzymol 1984;107:516-44. SAS user's guide: version 5, edition 433-506. Cary, N.C.: SAS Institute, 1985. Melick RA, Farrugia W, Quelch KJ. Plasma osteocalcin in man. Aust N Z J Med 1985;15:410-6. Delmas PS, Wilson DM, Mann KG, Riggs BL. Effect of renal function on plasma levels of bone-GLA-protein. J Clin Endocrinol Metab 1983;57:1028-30. Coen G, Mazzaferco S, Bonucci E, et al. Bone GLA protein in pediatric chronic renal failure: effects of 1,25-(OH)z D3 administration in a long-term follow-up. Kidney Int 1985;28: 783-90. Gundberg CM, Lian JB, Gallop PM. Measurements of gamma-carboxyglutamate and circulating osteocalein in normal children and adults. Clin Chim Acta 1983;128:1-9. Cole DEC, Carpenter TO, Gundberg CM. Serum osteocalcin concentrations in children with metabolic bone disease. J PED1ATR 1985;106:770-6. Gundberg CM, Cole DEC, Lian JB, Reade TM, Gallop PM. Serum osteocalcin in the treatment of inherited rickets with 1,25-dihydroxyvitamin D3. J Clin Endocrinol Metab 1987; 56:1063-7. Gundberg CM, Hanning RM, Liu S, Zlotkin SH, Balfe JW, Cole DEC. Clearance of osteocalcin by peritoneal dialysis in children with end-stage renal disease. Pediatr Res 1987 ;21:296300.