9 1992 by The Humana Press, Inc. All rights of any nature, whatsoever, reserved. 0163-4984/92/3201-:~4)399 $02.00
Serum Zinc and Somatic Growth in Children with Growth Retardation COLETTE FONS, 12 JEAN-FRP.Di~RIC BRUN,*,1 MICHELLE FUSSELLIER,2 GENEVII~VE CASSANAS,2 LUCETTE BARDET,2 AND ANDRI~ ORSETTI~
1Department of Physiology (Faculty of lVledicine), 34060 Montpellier C6dex, France; Department of Industrial Pharmaceutical Physics (Faculty of Pharmacy), 34060 Montpellier, C6dex, France Received January 31, 1991; Accepted May 2, 1991
ABSTRACT A possible role for zinc deficiency in some cases of growth retardation in southern France was investigated. Control values for zinc for 160 children (age = 12.5 _+ 2.4 yr) are 0.85 +_ 0.22 mg/L (mean ___2 SD). Twenty-five children with low serum zinc values (0.63 rag/L) were studied. Children in the two groups did not differ significantly in age, pubertal development, stature, and weight. For the 25 children whose serum values were low, we found significantly lower values for bone age delay, growth velocity in mm/month, as well as the ratio between calculated growth velocity and theoretical growth velocity for the bone age (so that zincemia was correlated to these parameters in the whole sample of 50 subjects). Nevertheless, no significant difference could be found between the two groups for serum somatomedin C, serum osteocalcin values, and GH responses to the GH stimulatory tests (exercise test, overnight sampling, insulin-induced hypoglycemia, arginine test). Therefore, low serum zinc is associated with a retardation in both somatic growth and pubertal maturation. Index Entries: Zinc; growth retardation; trace elements.
*Author to whom all correspondence and reprint requests should be addressed. Biological Trace Element Research
399
Vol. 32, 1992
400
Fons et aL
INTRODGCTION Evidence collected in several species indicates that, among the trace metals, zinc is an essential requirement for mammalian growth (1-3). Zinc deficiency in both animals and humans has been associated with short stature, abnormal hair growth, and failure of sexual maturation (4,5). In addition, zinc deficiency has been demonstrated to cause failure of normal GH secretion in both humans and rats. Conversely, zinc availability may limit the ability of GH-deficient patients to grow in response to GH therapy, though zinc supplementation has been reported to have no additional effect on the growth of GH-deficient children treated by exogenous GH (6,7). Therefore, zinc may be a limiting factor in growth-regulating mechanisms by modulating both GH release and GH action. On the other hand, evidence has been provided that many children in western countries exhibit low levels of zinc in plasma and hair, maybe because rapid growth contributes to accelerated zinc utilization (8-10). Although such children do not show the clinical signs of severe zinc deficiency as described in developing countries, it has been suggested that these borderline zinc deficiencies could result in growth delay. However, little is known about the relationships among growth, GH secretion, and zinc in European children. In a previous study (11), we reported that, in 117 children from southern France explored for short stature, serum zinc was low in 10% of the subjects and was correlated (r = 0.285 p < 0.01) to the height standard deviation score. However, the scattering of the results was important (although they were highly significant), probably because zinc is only one factor among many others involved in the complex process of growth retardation. Thus, in this study, we aimed at studying betterdefined subgroups of children. We planned to compare two matched samples of short children: 25 children with normal serum zinc, and 25 children with hypozincemia, in order to investigate whether serum zinc is related to either GH secretion and/or somatic growth.
MATERIALS AND METHODS The control group for serum zinc consisted of 160 children with normal height ( > - 2 SD), and serum frozen aliquots were assayed as described below. Blood was drawn between 9-10 AM after >2 h fasting. Age of the subjects was 12.5 +_ 2.4 yr. Fifty short children followed in our outpatient unit were selected on the following criteria. They were of short stature (below - 2 SD on the Semp6's Height charts) with no systemic disease explaining growth retardation (i.e., respiratory, digestive, cardiac, or congenital malforma-
Biological Trace Element Research
Vol. 32, 1992
Serum Zinc and Somatic Growth
401
tive disease). The sample consisted of 25 hypozincemic children (defined by the normal values observed in our control group), and 25 short children matched for age and sex, but with serum zinc values within the normal range. In these 50 children, we measured the following parameters: chronological age (years and months), bone age (Greulich and Pyle method), statural age (i.e., the theoretical age corresponding to the actual height when plotted on the height charts), and the growth velocity (mm/mo). Some derived indexes were also calculated: the difference between statural age and chronological age expressed in years and months; the delay in bone age, which corresponds to the difference between bone age and chronological age; the ratio between calculated growth velocity and theoretical growth velocity for the bone age (expressed in percentage of the latter); the height standard deviation score (given by the height charts of Setup6); the pubertal stage (according to Tanner); the B.M.I. or body mass index, which is the ratio between weight in kilograms and the square of stature. All children underwent physiological and pharmacological investigations of growth hormone (GH) secretion. The physiologic assessment consisted of a physical exercise test and nocturnal recording of spontaneous GH release during sleep. The physical exercise test on cycloergometer was submaximal allowing the heart rate to reach progressively (in 10 rain) 18{3beats per rain and maintaining it for 5 rain. Blood samples were taken, after half an hour of rest, just at the beginning and at the end of exercise and after 10 min. The nocturnal recording with EEG monitoring consisted of blood samples withdrawn every 20 min during sleep and for 8 h. The GH secretion was analyzed by the number, the amplitude, and the duration of the peaks, and also by the integrated concentration (area under the curve calculated with the trapezoidal rule and divided by the duration of exploration). As for pharmacological tests, always following the physiological ones, children underwent insulininduced hypoglycemia, with a continuous insulin infusion until monitored glycemia reached 2.2 mmol/L and a glucagon-propranolol test consisting of 0. I g/kg of propranolol per os and 1 mg of glucagon injected intramuscularly. Furthermore, 37 children underwent a GHRH iv test (2 ~g/kg body wt) according to the GHRH European Multicenter Study Group. GHRH 1-44 NH2 (SR 95228) was given by Sanofi Research (France). Pubertal stage was obtained by the measurement of dehydroepiandrosterone (DHA) sulfate (in both sexes), testosterone (in boys) and estradiol (in girls), by standardized routine radioimmunoassay techniques. Osteocalcin (GLA bone protein) was assayed with the kit OSTKPR from ORIS (International CIS, France). Somatomedin C/IGF 1 (SmC) was measured with the kit "IN-SOMC" from ORIS. Serum samples, after thawing, were extracted with ODS-silica columns before the radioimmunoassay procedure, which uses a second antibody and polyethylene
Biological Trace Element Research
Vol. 32, 1992
402
Fons et aL
glycol. Growth h o r m o n e was measured by radioimmunoassay with the "SB-HGH" kit, also obtained from the International CIS. Within assay coefficients of variation range between 9.2% (low values) and 7.7% (high values). Between assay coefficients of variation for low vs high values are 13.7 and 11.0% respectively. The lowest limit of sensitivity of the assay is 0.25 ng/mL + 0.24. Serum zinc was measured by flame atomic spectrophotometry. The lower limit of sensitivity of this method is 0.0125 rag/ L. Its coefficient of variation is 7.2% (n -- 9). Correlation analyses were performed with least squares fitting. Differences between subgroups were examined by the nonparametric (twotailed) Mann-Whitney rank test for unpaired data, since testing indicated that many parameters of the study were too far to exhibit a normal distribution. Only two parameters of the study were found to exhibit a normal distribution: growth velocity and height standard deviation score. Level of significance was defined as p < 0.05. Mean values were expressed _+ the standard error of the mean (SEM).
RESULTS The m e a n value of serum zinc in the 160 normal children was 0.85 mg/L with an SD of 0.11 mg/L. The 50 short children included in the study were divided into two subgroups with regard to their zinc serum values: one "hypozincemic" group (HZ) of 25 children with zinc serum values -< 0.63 mg/L and one "normozincemic" (NZ) with zinc serum values higher than 0.63 mg/L. The subgroups were matched for age (HZ -- 11.7 + 2.7 yr and NZ = 12 _+ 2.2 yr; n.s.), sex (HZ -- 17 boys, 8 girls; NZ = 18 boys, 7 girls); pubertal stage according to Tanner stature (HZ = 133 + 15 cms; NZ = 139 + 13 cms), and weight (HZ = 28 -+ 8 kg; NZ = 33 + 11 kg). Children in the HZ group appeared to have a stronger delay in bone age ( - 2 + 1 yr vs 0.9 + 0.7 yr; p -- 0.0001) w h e n compared to the NZ group. For the growth velocity, a significant difference (p = 0.02) was also found between HZ and NZ (4 + 1.6 m m / m o vs 5.2 + 2 mm/mo). The ratio between growth velocity and theoretical growth velocity for the bone age was lower in HZ (0.8 + 0.3%, vs 1.080 +_ 0.0004%; p = 0.004). By contrast, HZ and NZ did not differ with respect to the following parameters: IGF1-SomC, osteocalcin, and GH responses to stimulatory tests (exercise test, overnight sampling, insulin-induced hypoglycemia, arginine test). Significant correlations were found between serum zinc and the following parameters: (i) bone age delay (r = 0.477 and p -0.0005, s e e Fig. 1); (ii) the ratio between calculated growth velocity and theoretical growth velocity for the bone age (r = 0.371 and p = 0.008), as shown in Fig. 2.
Biological Trace Element Research
Vol. 32, 1992
Serum Zinc and Somatic Growth
403
il
0,8
9
i
9
O.7 o o
~6
o
o
o
l -I
Bone
[]
o
ooo I
0
o
o o
[]
115
oo
~176
o ~
I o
I
-2
Age
l
-3
Delay
I
-4
-5
(years)
Fig. 1. Correlation between bone age delay (in years) and serum zinc in the 50 subjects of the study divided into two paired subgroups according to their zincemia. Black squares: normozincemic children (n = 25); white squares: hypozincemic children (n = 25). R = 0.477, p = 0.0005.
1.8
0 -
1
o
[]
o
9 o
o
cl []
Dcl
o ooo o
0
o o
0.4
o
|
0.2 i
a
i
O.5
0.6
0.7
I
I
I
--
Q8 0.9 1 Serum Zinc mglL
Fig. 2. Correlation between growth velocity (expressed as a ratio between actual growth velocity and theoretical growth velocity for bone age as given by the height charts) and serum zinc in the 50 subjects of the study divided into two paired subgroups according to their zincemia. Black squares: normozincemic children (n -- 25); white squares: hypozincemic children (n = 25). r = 0.371, p = 0.008.
Biological Trace Element Research
VoL 32, 1992
404
Fons et aL
DISCUSSION This study shows a significant relationship between serum zinc and two aspects of somatic growth: growth velocity and bone maturation. This could suggest that moderate zinc deficiencies impair both growth and pubertal maturation, as already demonstrated in the case of more severe zinc deficiencies (12). We believe that these findings may be of clinical relevance, since zinc supplementation in such children may be a corrective treatment for growth disorders. However, one must keep in m i n d that a bone age delay is associated to some extent with a longer duration of somatic growth, which may be beneficial. Therefore, w h e t h e r a correction of this zincrelated delay is beneficial for the final stature remains to be established.
REFERENCES 1. N. F. Krebs and M. K. Hambidge, Am. J. Clin. Nutr. 43, 288 (1986). 2. J. W. Thorp, R. L. Boeckx, S. Robbins, S. Horn, and M. D. Fletcher, Amer. Jour. Clin. Nutr. 34, 1056 (1981). 3. J. P. Wouwe, C. J. Den Hamer, and J. B. Tricht, Eur. J. Pediat. 144, 598 (1986). 4. J. -K. Friel, R. S. Gibson, G. F. Kawash, and J. J. Watts, Pediatr. Gastroenterol. Nutr. 4, 746 (1985). 5. S. -Z. Ghavami Maibodi, P. J. Collip, M. Castro-Magana, C. Stewart, and S. Y. Chen, Ann. Nutr. Metab. 27, 214 (1983). 6. J. Collip, M. Castro-Magana, M. Petrovic, J. Thomas, T. Lheruvanky, and S. Y. Chen, Ann. Nutr. Metab. 26, 287 (1982). 7. A. -W. Root, G. Duckett, M. Sweetland, and E. Reiter, Rev. Nut. and Diet 36, 140 (1980). 8. A. S. Prasad, D. Oberleas, E. R. Miller, and W. Luecker, Nut. Rev. 41:7 (1983). 9. I. -E. Dreotsy, P. C. Grey, and P. J. Wilkins, S. Afr. Med. J. 46, 1585 (1972). 10. M. Yamaguchi and R. Yamaguchi, Bioch. Pharmacol. 35, 773 (1986). 11. J. F. Brun, I. Moynier, C. Fedou, M. Fusselier, and A. Orsetti, J. Endocrinol. Invest. 10 (suppl 4), 87 (1987). 12. A. S. Prasad, J. Pharmacol. 16, 344 (1985).
Biological Trace Element Research
Vol. 32, 1992
Serum Zinc and Somatic Growth in Children with Growth Retardation COLETTE FONS, 12 JEAN-FRP.Di~RIC BRUN,*,1 MICHELLE FUSSELLIER,2 GENEVII~VE CASSANAS,2 LUCETTE BARDET,2 AND ANDRI~ ORSETTI~
1Department of Physiology (Faculty of lVledicine), 34060 Montpellier C6dex, France; Department of Industrial Pharmaceutical Physics (Faculty of Pharmacy), 34060 Montpellier, C6dex, France Received January 31, 1991; Accepted May 2, 1991
ABSTRACT A possible role for zinc deficiency in some cases of growth retardation in southern France was investigated. Control values for zinc for 160 children (age = 12.5 _+ 2.4 yr) are 0.85 +_ 0.22 mg/L (mean ___2 SD). Twenty-five children with low serum zinc values (0.63 rag/L) were studied. Children in the two groups did not differ significantly in age, pubertal development, stature, and weight. For the 25 children whose serum values were low, we found significantly lower values for bone age delay, growth velocity in mm/month, as well as the ratio between calculated growth velocity and theoretical growth velocity for the bone age (so that zincemia was correlated to these parameters in the whole sample of 50 subjects). Nevertheless, no significant difference could be found between the two groups for serum somatomedin C, serum osteocalcin values, and GH responses to the GH stimulatory tests (exercise test, overnight sampling, insulin-induced hypoglycemia, arginine test). Therefore, low serum zinc is associated with a retardation in both somatic growth and pubertal maturation. Index Entries: Zinc; growth retardation; trace elements.
*Author to whom all correspondence and reprint requests should be addressed. Biological Trace Element Research
399
Vol. 32, 1992
400
Fons et aL
INTRODGCTION Evidence collected in several species indicates that, among the trace metals, zinc is an essential requirement for mammalian growth (1-3). Zinc deficiency in both animals and humans has been associated with short stature, abnormal hair growth, and failure of sexual maturation (4,5). In addition, zinc deficiency has been demonstrated to cause failure of normal GH secretion in both humans and rats. Conversely, zinc availability may limit the ability of GH-deficient patients to grow in response to GH therapy, though zinc supplementation has been reported to have no additional effect on the growth of GH-deficient children treated by exogenous GH (6,7). Therefore, zinc may be a limiting factor in growth-regulating mechanisms by modulating both GH release and GH action. On the other hand, evidence has been provided that many children in western countries exhibit low levels of zinc in plasma and hair, maybe because rapid growth contributes to accelerated zinc utilization (8-10). Although such children do not show the clinical signs of severe zinc deficiency as described in developing countries, it has been suggested that these borderline zinc deficiencies could result in growth delay. However, little is known about the relationships among growth, GH secretion, and zinc in European children. In a previous study (11), we reported that, in 117 children from southern France explored for short stature, serum zinc was low in 10% of the subjects and was correlated (r = 0.285 p < 0.01) to the height standard deviation score. However, the scattering of the results was important (although they were highly significant), probably because zinc is only one factor among many others involved in the complex process of growth retardation. Thus, in this study, we aimed at studying betterdefined subgroups of children. We planned to compare two matched samples of short children: 25 children with normal serum zinc, and 25 children with hypozincemia, in order to investigate whether serum zinc is related to either GH secretion and/or somatic growth.
MATERIALS AND METHODS The control group for serum zinc consisted of 160 children with normal height ( > - 2 SD), and serum frozen aliquots were assayed as described below. Blood was drawn between 9-10 AM after >2 h fasting. Age of the subjects was 12.5 +_ 2.4 yr. Fifty short children followed in our outpatient unit were selected on the following criteria. They were of short stature (below - 2 SD on the Semp6's Height charts) with no systemic disease explaining growth retardation (i.e., respiratory, digestive, cardiac, or congenital malforma-
Biological Trace Element Research
Vol. 32, 1992
Serum Zinc and Somatic Growth
401
tive disease). The sample consisted of 25 hypozincemic children (defined by the normal values observed in our control group), and 25 short children matched for age and sex, but with serum zinc values within the normal range. In these 50 children, we measured the following parameters: chronological age (years and months), bone age (Greulich and Pyle method), statural age (i.e., the theoretical age corresponding to the actual height when plotted on the height charts), and the growth velocity (mm/mo). Some derived indexes were also calculated: the difference between statural age and chronological age expressed in years and months; the delay in bone age, which corresponds to the difference between bone age and chronological age; the ratio between calculated growth velocity and theoretical growth velocity for the bone age (expressed in percentage of the latter); the height standard deviation score (given by the height charts of Setup6); the pubertal stage (according to Tanner); the B.M.I. or body mass index, which is the ratio between weight in kilograms and the square of stature. All children underwent physiological and pharmacological investigations of growth hormone (GH) secretion. The physiologic assessment consisted of a physical exercise test and nocturnal recording of spontaneous GH release during sleep. The physical exercise test on cycloergometer was submaximal allowing the heart rate to reach progressively (in 10 rain) 18{3beats per rain and maintaining it for 5 rain. Blood samples were taken, after half an hour of rest, just at the beginning and at the end of exercise and after 10 min. The nocturnal recording with EEG monitoring consisted of blood samples withdrawn every 20 min during sleep and for 8 h. The GH secretion was analyzed by the number, the amplitude, and the duration of the peaks, and also by the integrated concentration (area under the curve calculated with the trapezoidal rule and divided by the duration of exploration). As for pharmacological tests, always following the physiological ones, children underwent insulininduced hypoglycemia, with a continuous insulin infusion until monitored glycemia reached 2.2 mmol/L and a glucagon-propranolol test consisting of 0. I g/kg of propranolol per os and 1 mg of glucagon injected intramuscularly. Furthermore, 37 children underwent a GHRH iv test (2 ~g/kg body wt) according to the GHRH European Multicenter Study Group. GHRH 1-44 NH2 (SR 95228) was given by Sanofi Research (France). Pubertal stage was obtained by the measurement of dehydroepiandrosterone (DHA) sulfate (in both sexes), testosterone (in boys) and estradiol (in girls), by standardized routine radioimmunoassay techniques. Osteocalcin (GLA bone protein) was assayed with the kit OSTKPR from ORIS (International CIS, France). Somatomedin C/IGF 1 (SmC) was measured with the kit "IN-SOMC" from ORIS. Serum samples, after thawing, were extracted with ODS-silica columns before the radioimmunoassay procedure, which uses a second antibody and polyethylene
Biological Trace Element Research
Vol. 32, 1992
402
Fons et aL
glycol. Growth h o r m o n e was measured by radioimmunoassay with the "SB-HGH" kit, also obtained from the International CIS. Within assay coefficients of variation range between 9.2% (low values) and 7.7% (high values). Between assay coefficients of variation for low vs high values are 13.7 and 11.0% respectively. The lowest limit of sensitivity of the assay is 0.25 ng/mL + 0.24. Serum zinc was measured by flame atomic spectrophotometry. The lower limit of sensitivity of this method is 0.0125 rag/ L. Its coefficient of variation is 7.2% (n -- 9). Correlation analyses were performed with least squares fitting. Differences between subgroups were examined by the nonparametric (twotailed) Mann-Whitney rank test for unpaired data, since testing indicated that many parameters of the study were too far to exhibit a normal distribution. Only two parameters of the study were found to exhibit a normal distribution: growth velocity and height standard deviation score. Level of significance was defined as p < 0.05. Mean values were expressed _+ the standard error of the mean (SEM).
RESULTS The m e a n value of serum zinc in the 160 normal children was 0.85 mg/L with an SD of 0.11 mg/L. The 50 short children included in the study were divided into two subgroups with regard to their zinc serum values: one "hypozincemic" group (HZ) of 25 children with zinc serum values -< 0.63 mg/L and one "normozincemic" (NZ) with zinc serum values higher than 0.63 mg/L. The subgroups were matched for age (HZ -- 11.7 + 2.7 yr and NZ = 12 _+ 2.2 yr; n.s.), sex (HZ -- 17 boys, 8 girls; NZ = 18 boys, 7 girls); pubertal stage according to Tanner stature (HZ = 133 + 15 cms; NZ = 139 + 13 cms), and weight (HZ = 28 -+ 8 kg; NZ = 33 + 11 kg). Children in the HZ group appeared to have a stronger delay in bone age ( - 2 + 1 yr vs 0.9 + 0.7 yr; p -- 0.0001) w h e n compared to the NZ group. For the growth velocity, a significant difference (p = 0.02) was also found between HZ and NZ (4 + 1.6 m m / m o vs 5.2 + 2 mm/mo). The ratio between growth velocity and theoretical growth velocity for the bone age was lower in HZ (0.8 + 0.3%, vs 1.080 +_ 0.0004%; p = 0.004). By contrast, HZ and NZ did not differ with respect to the following parameters: IGF1-SomC, osteocalcin, and GH responses to stimulatory tests (exercise test, overnight sampling, insulin-induced hypoglycemia, arginine test). Significant correlations were found between serum zinc and the following parameters: (i) bone age delay (r = 0.477 and p -0.0005, s e e Fig. 1); (ii) the ratio between calculated growth velocity and theoretical growth velocity for the bone age (r = 0.371 and p = 0.008), as shown in Fig. 2.
Biological Trace Element Research
Vol. 32, 1992
Serum Zinc and Somatic Growth
403
il
0,8
9
i
9
O.7 o o
~6
o
o
o
l -I
Bone
[]
o
ooo I
0
o
o o
[]
115
oo
~176
o ~
I o
I
-2
Age
l
-3
Delay
I
-4
-5
(years)
Fig. 1. Correlation between bone age delay (in years) and serum zinc in the 50 subjects of the study divided into two paired subgroups according to their zincemia. Black squares: normozincemic children (n = 25); white squares: hypozincemic children (n = 25). R = 0.477, p = 0.0005.
1.8
0 -
1
o
[]
o
9 o
o
cl []
Dcl
o ooo o
0
o o
0.4
o
|
0.2 i
a
i
O.5
0.6
0.7
I
I
I
--
Q8 0.9 1 Serum Zinc mglL
Fig. 2. Correlation between growth velocity (expressed as a ratio between actual growth velocity and theoretical growth velocity for bone age as given by the height charts) and serum zinc in the 50 subjects of the study divided into two paired subgroups according to their zincemia. Black squares: normozincemic children (n -- 25); white squares: hypozincemic children (n = 25). r = 0.371, p = 0.008.
Biological Trace Element Research
VoL 32, 1992
404
Fons et aL
DISCUSSION This study shows a significant relationship between serum zinc and two aspects of somatic growth: growth velocity and bone maturation. This could suggest that moderate zinc deficiencies impair both growth and pubertal maturation, as already demonstrated in the case of more severe zinc deficiencies (12). We believe that these findings may be of clinical relevance, since zinc supplementation in such children may be a corrective treatment for growth disorders. However, one must keep in m i n d that a bone age delay is associated to some extent with a longer duration of somatic growth, which may be beneficial. Therefore, w h e t h e r a correction of this zincrelated delay is beneficial for the final stature remains to be established.
REFERENCES 1. N. F. Krebs and M. K. Hambidge, Am. J. Clin. Nutr. 43, 288 (1986). 2. J. W. Thorp, R. L. Boeckx, S. Robbins, S. Horn, and M. D. Fletcher, Amer. Jour. Clin. Nutr. 34, 1056 (1981). 3. J. P. Wouwe, C. J. Den Hamer, and J. B. Tricht, Eur. J. Pediat. 144, 598 (1986). 4. J. -K. Friel, R. S. Gibson, G. F. Kawash, and J. J. Watts, Pediatr. Gastroenterol. Nutr. 4, 746 (1985). 5. S. -Z. Ghavami Maibodi, P. J. Collip, M. Castro-Magana, C. Stewart, and S. Y. Chen, Ann. Nutr. Metab. 27, 214 (1983). 6. J. Collip, M. Castro-Magana, M. Petrovic, J. Thomas, T. Lheruvanky, and S. Y. Chen, Ann. Nutr. Metab. 26, 287 (1982). 7. A. -W. Root, G. Duckett, M. Sweetland, and E. Reiter, Rev. Nut. and Diet 36, 140 (1980). 8. A. S. Prasad, D. Oberleas, E. R. Miller, and W. Luecker, Nut. Rev. 41:7 (1983). 9. I. -E. Dreotsy, P. C. Grey, and P. J. Wilkins, S. Afr. Med. J. 46, 1585 (1972). 10. M. Yamaguchi and R. Yamaguchi, Bioch. Pharmacol. 35, 773 (1986). 11. J. F. Brun, I. Moynier, C. Fedou, M. Fusselier, and A. Orsetti, J. Endocrinol. Invest. 10 (suppl 4), 87 (1987). 12. A. S. Prasad, J. Pharmacol. 16, 344 (1985).
Biological Trace Element Research
Vol. 32, 1992
