Plasma Growth Hormone Concentration

in Corticosteroid-Treated Children

HELENG. MoRms, JACQUELINE R. JORGENSEN,and SHIRLEY A. JENKINSFrom the Department of Medicine, Veterans Administration Hospital,University of Colorado School of Medicine, and Children's AsthmaResearch Institute and Hospital, Denver, Colorado

A B S T R A C T Endogenous plasma growth hor-mone concentrations were measured in 23 childrenwho were receiving daily corticosteroid therapy andin 10 control asthmatic children who had not re-ceived steroids for at least 8 months. The growthhormone concentrations were similar in 'the twogroups of patients both during the fasting stateand after insulin-induced hypoglycemia. 12 chil-dren, who were studied while receiving a largedose of prednisone and again 2 wk after steroidwithdrawal, also showed no change in growthhormone concentration in relation to corticosteroidtherapy. These findings suggest that deficiency ofgrowth hormone is not the major mechanismresponsible for the dwarfism of corticosteroid-treated children.

INTRODUCTION

The mechanisms responsible for growth retarda-tion in children given prolonged corticosteroidtherapy have not been defined. Previous studies inthis laboratory have demonstrated that corti-costeroid-treated children have impaired metabolicresponse to acute administration of human growthhormone (HGH) and show no improvement intheir retarded rate of growth during prolongedHGH administration (1). These observationshave suggested that the mechanism responsiblefor the growth retardation might be due to corti-costeroid-induced antagonism of the effects of

A preliminary report of this work was presented inpart in Clin. Res. April 1966.

Address requests for reprints to Dr. Helen G. Morris,1055 Clermont Street, Denver, Colo. 80220.

Received for publication 7 November 1966 and in re-vised form 21 September 1967.

HGH at the tissue level. On the other hand,Frantz and Rabkin and Hartog, Gaafar, andFraser have reported that corticosteroid-treatedadults show inhibition of pituitary release of GHafter hypoglycemia (2, 3) and have postulatedthat dwarfism in steroid-treated children mightbe due in part to relative deficiency of endogenousGH (2). The present study was undertaken todetermine plasma growth hormone levels in rela-tion to corticosteroid treatment in a relativelyhomogeneous population of children who sufferfrom asthma. Plasma growth hormone levels insteroid-treated children during the fasting stateand after induced hypoglycemia were comparedwith those observed in nonsteroid-treated asth-matic control children. Many of the children werestudied both during corticosteroid administrationand after total withdrawal of steroids.

METHODSPatient population. Studies were performed on 33

children, aged 8-15, who were resident patients of theChildren's Asthma Research Institute and Hospital(CARIH). 10 control asthmatic children had receivedno corticosteroid for at least 8 months. 23 children hadreceived daily corticosteroid for prolonged periods withtherapy at the dose indicated for at least 1 month.11 of the children required maintenance with stable dosesof steroid (prednisone 2.5-30 mg/day). 12 children wereable to tolerate total discontinuation of steroid and werestudied while receiving a large dose of prednisone (10-15 mg/day) and again 2 wk after steroid withdrawal.Some of the patients in the last group were studied onseveral occasions as steroid dose was being reduced (by2.5 mg every 10 days) and at intervals after discontinu-ation of steroid. The sex distribution of the populationat CARIH is reflected in our study, with the steroid-treated group containing a larger percentage of males andthe nonsteroid group containing a higher percentage offemales.

The Journal of Clinical Investigation Volume 47 1968 427

Study protocol. All studies were performed on fastingsubj ects who had been at bed rest for at least 2 hr. Mostof the children receiving corticosteroid had their usualmorning dose of prednisone withheld until completion ofthe test. After initial venipuncture, an intravenous infu-sion of normal saline was begun and continued through-out the study. All subsequent blood samples were obtainedthrough the infusion needle after the first few drops ofblood were discarded to minimize dilution error. Two con-trol samples were obtained for glucose and GH determi-nations. Crystalline insulin was then administered intra-venously at a dose of 0.1 U/kg of body weight. Repeatsamples for glucose determination were obtained 20, 30, 45,60, 90, and 120 min after insulin administration. Repeatsamples for GH determination were obtained 30, 60, 90,and 120 min after insulin administration. Some studieswere terminated before 120 min because of persisting hy-poglycemia.

A physician remained in the room continously through-out each insulin tolerance test. Glucagon and 50% glu-cose were immediately available to counteract hypogly-cemia if this seemed indicated. At the termination of thestudy, children were given sweetened juice and werefed. Close observation was continued for several hours.

Methods of analysis. Samples for blood glucose werecollected in tubes containing sodium fluoride and wereanalyzed by either the glucostat method (4) or by theAutoAnalyzer (5). The values reported are the fastingsugars, the minimum blood sugars (obtained at either 20 or30 min after insulin administration), and the 60-min values.

Samples for GH were collected in heparinized tubes,kept in an ice bath until separated in a refrigerated cen-trifuge, and then stored in a freezer. GH concentrationwas determined by the radioimmunoassay method ofGlick, Roth, Yalow, and Berson (6). Most of the plasmasamples were assayed on several occasions, and the valueshown is the mean of all determinations. The GHvaluesreported are those observed during the fasting state and60 min after insulin administration. The 60 min samplewas selected as the indicator of response to hypoglycemia,since this value was determined in all patients studied andin most subjects was the highest value observed. Glucoseand GH results obtained in the different groups of chil-dren were analyzed by Student's t test (7).

Immunoassay method. "s'I-labeled HGH (specific ac-tivity 300-500 /Ac/,gg) was prepared by the method ofGreenwood, Hunter, and Glover (8) using 5 ,ug of HGH(Wilhelmi HS 612B), 2-3 mc of "~'I (Iso-Serve, Cam-bridge Nuclear Corp., Cambridge, Mass.), and 44,ug of chloramine-T for 1 min, followed by 120 ug ofsodium metabisulfite. The labeled fraction containing theleast amount of damage after purification (starch gelelectrophoresis (9) or filtration through a SephadexG-100 column (10)) was identified by paper chromato-electrophoresis (11) and was used in the assay after dilu-tion to 60,000 cpm. Antiserum prepared in rabbits usingElrick HGH(E-20) (12) was adsorbed by precipitationwith lyophilized, pooled human serum to remove possibleantibody to human serum components and was used inthe assay at a final concentration of 1:130,000.

Standard curves were prepared for each immunoassayrun using either Wilhelmi GH (HS 612B) or the BritishMedical Research Council (MRC) "GH assay standardA" which gave identical slopes.' Unknown samples werecalculated on the basis of the MRCstandard. Each plasmasample was assayed in duplicate at three different dilu-tions, with one tube in each set of three serving as a con-trol tube. 0.1 ml of plasma or diluted plasma was usedin the incubate. All solutions were prepared with Veronalbuffer 0.1 mole/liter, pH 8.6, containing 1% normal rab-bit serum and 2.5 mg/ml of human serum albumin (TheCutter Laboratories, Berkeley, Calif.). Tubes containingeither standards or plasma samples were incubated withlabeled GH "..I and antiserum in a final volume of 1 mlfor 4-8 days at 5'C. A 0.2 ml aliquot of the incubate (and0.02 ml of pooled human serum for carrier protein in in-cubates containing standards or diluted plasma specimens)were then applied to 1U-inch wide strips of Whatman3MC chromatographic paper and subjected to chromato-electrophoresis at 30 v/strip for 45 min at room tempera-ture, which separated the radioactive label into a freepeak (F) and an antibody-bound peak (B). Paper stripswere oven-dried and then scanned for radioactivity at therate of 2 cm/min in an automatic 4 pi radiochromatographscanner equipped with an integrator recorder. The inte-grator excursions were used to determine the areas underradioactivity peaks.

Standard curves were plotted either on linear graphpaper as B/F ratio vs. concentration of HGHin milli-micrograms per milliliter, according to the method ofBerson-Yalow (11), which gave the curve shown inFig. 1, or were plotted on semilogarithmic paper as percent bound vs. log GHconcentration which gave the linearrelationship shown in Fig. 2. The latter plot was foundto be more convenient for estimating the GH concen-tration of unknown plasma samples.

Precision of the immunoassay. Since the GH valuesreported here were obtained by duplicate determinationsat several dilutions and in most instances were ana-lyzed on several occasions, the results obtained withthese plasma samples were used to estimate the pre-cision of the immunoassay in our hands. The meandifference in per cent of radioactivity in the bound frac-tion in 400 duplicate determinations (at various dilutions)was ± 3%. In terms of millimicrograms per milliliterof GH, comparison of 182 duplicate determinationswhich yielded measurable GH values showed a meandifference of ± 22%. The precision at various ranges ofGH concentration shown in Table I is similar to thatreported for GH immunoassay by other investigators(13-15). Comparison of the results obtained when 148specimens were analyzed at two different dilutions indi-cated a slight but significantly greater value at the higherdilution. Comparison of the results observed when 115specimens were analyzed on different occasions yieldedno significant differences.

'We are indebted to Dr. A. E. Wilhelmi (NationalPituitary Agency) and Dr. D. Bangham (British MedicalResearch Council) for providing the purified GH prepa-rations used for assay standards.

428 H. G. Morris, J. R. Jorgensen, and S. A. Jenkins

2.2

2.0

1.4 NRC StandordB/F Rabbit Antiserum 1:130,000

Ratio 1.2

1.0'

0.8 En

0.6. :

0.4

0.2

t p p* * , *0.2 Q4 0.6 QL8 1.0 1.2 1.4 1.6 1.8

HGH mpg/mI

FIGURE 1 Growth hormone immunoassay standard curve (2/9/66). The results areplotted as antibody-bound peak/free peak(mjg/ml).

RESULTS

We found that induction of hypoglycemia forthe purpose of measuring plasma GHconcentra-tions was a safe procedure in the asthmatic chil-dren. A single, potentially serious reaction wasencountered when one child developed Cheyne-Stokes respirations 30 min after insulin adminis-tration with a blood glucose concentration of 25mg/100 ml. He was immediately given intravenousglucose and recovered promptly; subsequent ob-servation revealed no sequellae. Not one acuteepisode of asthma occurred during the conduct of

TABLE IPrecision of GHImmunoassay

Range GH No. of Mean GH Standardconcentration samples concentration deviation*

mpg/mi mAg/mi1-9 '54 6.3 ±1.5 (24%)

10-19 65 15.3 44.0 (26%)20-29 32 24.1 ±5.1 (21%)30+ 31 48.8 -7.0 (14%)

* SD = -4D2/2N (16).

ratio (B/F ratio) vs. GH concentration

these studies, and if anything, children had lesssevere asthma during the hours after hypo-glycemia.

Table II lists the pertinent clinical features ofthe patients and the results of the glucose andGH determinations. It can be seen that most ofthe children, steroid-treated as well as controlsubjects, had measurable fasting levels of GHand that in most the concentration of GH roseafter insulin administration.

In both steroid-treated and control groups veryhigh fasting levels of GH were occasionally en-countered despite the long rest period beforecollection of initial samples. In some instances, thehigh values occurred in children who were notedat the time of the test to be particularly appre-hensive or upset. Most of the high values wereencountered in females. As noted previously andindicated in Table II, the control group containsa much higher ratio of females than the steroidgroups, which may account for the higher meanfasting value.

Fig. 3 graphically illustrates the GH valuesobserved in children during steroid treatment,

Endogenous GH in Steroid-Treated Children 429

TABLE II

Glucose and GHValues during Insulin Tolerance Tests

Glucose GH

Patient Age Sex Steroid* dose 0 Minimum: 60 min§ Fasting 60 min

mg mg/100 ml mug/mlNonlsteroid-treated control children

A. H.G. L.T. L.D. P.K. A.B. B.R. C.K. H.M. R.R. S.

1313

813.5111212111013

MMMMFFFFFF

0000000000

Children during steroid treatment and 2 wk after steroid withdrM. C.

S. E.

M. G.

D. G.

Wv. I.

S. K.

B. L.

J. MI.

G. P.

L. C.

S. F.

B. S.

15

9

9.5

10.5

7.5

14

12

11

12

12

9

12

Children on stable dose steroidF. G. 15C.H. 9

B. K. 11

L. L.P. L.D. N.R. R.J. S.C. B.D. B.L. G.

13131115

13131512

M 150

M 100

M 100

M 100

M 7.50

M 100

M 100

xi 150

M 100

F 100

F 100

F 100

M 2.5M S

M 10M 30M 15M 10M 10M 20F 5

F 10F 5

95 22 53 2.588 20 42 694 20 58 1.587 19 47 1082 26 45 488 37 59 32.597 - 81 4.587 28 78 1590 30 67 2287 24 38 23

awal100 27 72 393 32 56 188 28 64 489 35 46 1086 37 72 2.557 28 25 4.187 27 60 991 32 53 579 40 62 1274 22 38 2095 20 67 7.588 27 61 4.182 52 80 5.787 16 62 587 43 85 485 32 57 1989 32 6269 32 3883 32 7785 27 5874 30 5281 32 3697 46 7284 47 66

7390899886788883

1007881

63

241.576.53.53

28 65 5

21 62 632 58 136 77 8.526 62 14

10.531 61 132 71 1129 56 942 55 1435 54 24

* Daily dose of prednisone.+ Minimum refers to minimum plasma concentration after insulin.§ 60 min refers to glucose and GHconcentration 60 min after insulin.

430 H. G. Morris, J. R. Jorgensen, and S. A. Jenkins

2320

727542118202931

817810203022159

131830505821142218.522262212212973

464

201212112522

5.81519

*% 40-Bound 0*

30-

20-

10-

I0

O ~ ~ ~ ~~ ~. . ,. * 3 3

0.1 0.2 0.3 0.4 Q5 0.7 1.0 2.0

GH mpg/mlFIGURE 2 Growth hormone immunoassay standard curve (2/9/66). The samedata shown in Fig. 1 are plotted as per cent of bound radioactive label vs. logGHconcentration in m/Ag/ml.

shortly after steroid withdrawal, and in controlsubjects. Although there was considerable varia-bility, the mean concentrations and range of valueswere quite similar in the three groups both duringthe fasting state and after induced hypoglycemia.(Mean fasting GH concentration: control = 12.1mug/ml; steroid-treated = 8.4 mug/ml; steroid-withdrawn = 6.9 mug/ml. Mean postinsulin GHconcentration: control = 25 mug/ml; steroid-treated = 23 mug/ml; steroid-withdrawn = 31.4mug/ml.). In each of the groups the rise inplasma GHconcentration after insulin administra-tion was significant (P = 0.01-0.05). The postin-sulin values in the steroid-treated children showedmore variability than that observed in either ofthe control groups. With the exception of the twolowest postinsulin values (observed in patientswho were in the stable dose steroid group and whoreceived a small dose of prednisone on the morn-ing of the study), the distribution of values wassimilar in children who subsequently toleratedwithdrawal of steroid and those who did not.

Fig. 4 compares the GHconcentrations in indi-vidual children who were studied while receivinga large dose of prednisone and again 2 wk aftercorticosteroid withdrawal. The similarity of valuesis further shown by the data in Table III whichlists the GH concentrations observed in severalchildren who were studied serially while steroiddosage was being reduced and at intervals aftertotal withdrawal of corticosteroid. It can be seenthat the GHvalues for each individual child werefairly reproducible and that there was no consistenttrend in either fasting or postinsulin values inrelation to discontinuation of steroids.

Table IV summarizes the results of statisticalcomparison of the glucose and GH values ob-served in the different groups of children. Therewere no significant differences in GH values inchildren recently withdrawn from steroid com-pared with control children who had not receivedsteroid for many months. The standard dose ofinsulin (0.1 U/kg) produced significantly lowerminimum and 60-min blood sugars in contol sub-

Endogenous GH in Steroid-Treated Children 431

Fasting

Control Steroid Rx Withdrawn

30.

25

20-

15-

10-

5-

After Insulin

a o8 08

Control Steroid Rx Withdrawn

FIGURE 3 Plasma GH concentrations in control and steroid-treated children. Each open circle rep-

resents a single patient.

jects than those observed in children who were

receiving steroid therapy (P = <0.05). Despitethe differences in degree and duration of hypo-glycemia, there was no significant difference inmean GHvalues during either the fasting state or

after insulin administration in steroid-treated andcontrol children (fasting GHP 0.2; post-insu-lin GHP = 0.65). Similarly, comparison of theresults observed when the same subjects were

studied during steroid therapy and shortly aftersteroid withdrawal again revealed no significantdifferences in GHvalues (P = 0.60-0.85), despite

Fasting At

OHmyg/ml

Predniso Ri Withdrawn

the fact that during the poststeroid withdrawalperiod the duration of hypoglycemia was signifi-cantly greater (P = <0.001 for blood glucose at60 min).

The GHvalue at 60 min after insulin adminis-tration was used as the index of response to hypo-glycemia in comparing the various groups ofsubjects, since this value was measured in allstudies and most subjects exhibited maximumGH concentration at this time. Data at laterpoints after insulin administration are less com-

plete, since a number of studies were terminated

'ter Insulin

FIGURE 4 Plasma GH levels in indi-vidual children during steroid treat.ment and 2 wk after steroid with-

drawal. The connecting bars identifythe GH values observed in each childduring treatment and after steroidwithdrawal.

Prednisone Rx Withdrawn

432 H. G. Morris, J. R. Jorgensen, and S. A. Jenkins

30

GHmpg/ml

25

20'

0

0

.8 000

00e0 00 080 88 j00 00

15-

10-

5.

8 800

0 80 8§80O00

00

0 00~~

0

TABLE IIIGHLevels during Prednisone Therapy and after Withdrawal

Daily prednisone dosage Duration of withdrawal

15 mg 10 mg 5 mg 2.5 mg 2 wk 6 wk 10 wk

PatientL. C. 24* 4 1.5 3 11

26t 15 22 15 21

M. G. 2.5 3.1 4.130 26 22

D. G. 9 5.5 5 215 15 9 14

J. M. 4 4 2.6 19 9.514 23 42 22 21

M. C. 3 1 3 4.581 78 127 113

* Upper number of each pair = fasting GHvalue.t Lower number of each pair = value after insulin.

before 2 hr because of persistent symptoms ofhypoglycemia. However, available specimens re-vealed that mean GH concentrations at 90 minwere similar to those observed at 60 min in each

group. While most subjects showed a beginningdecrease in GH concentration, there were a fewsubjects in each group who showed a further risein GHlevel at 90 min. By 120 min the GHcon-

TABLE IVStatistical Comparison of Plasma Glucose and Growth Hormone Concentrations

in Steroid-Treated and Control Patients

Mean glucose

mg/100 ml

Recent steroid withdrawal vs. control

Steroid withdrawal(n = 12)

81.9 10.430.2 7.549.7 i 12.8

Control(n = 10)

89.5 i 4.525.1 5.856.8 i 14.8

P = 0.05P = 0.11P = 0.25

Steroid withdrawal Control(n= 12) (n= 10)6.9 ±t 6.3 12.1 10.6 P = 0.15

31.4 22.7 25 i 12.2 P = 0.45

Steroid-treated vs. control

FastingMinimum60 min

Steroid-treated(n = 23)

86.6 7.633 ±t 7.765.7 i 8.8

Control(n = 10)

89.5 ±I 4.525.1 5.856.8 14.8

P = 0.3P = 0.02P = 0.04

Same children during steroid treatment and after steroid withdrawal

On steroid Off steroid(n= 12) (n= 12)

Fasting 87.2 + 7.5 81.9 10.4 P = 0.17Minimum 34.5 9.3 30.2 + 7.5 P = 0.2360 min 68.8 9.3 49.7 12.8 P = 0.001

Steroid-treated Control(n= 23) (n= 10)8.4 6 12.1 10.6 P = 0.2

23 17.4 25 12.2 P = 0.75

On steroid Off steroid(n= 12) (n= 12)

7.4 5.9 6.9 i 6.3 P = 0.85

28 i 21.3 31.4 22.7 P = 0.6

* Minimum refers to minimum plasma glucose concn after insulin.t 60 min refers to glucose and GHconcentrations 60 min after insulin.

Endogenous GH in Steroid-Treated Children 433

FastingMinimum*60 mint

Mean GH

mug/ml

centrations were approaching preinsulin levels andagain were higher in control subjects than incorticosteroid-treated or withdrawn groups. At 90and 120 min blood glucose was similar in controland steroid-treated children but continued to belower in subjects recently withdrawn from steroid.

DISCUSSION

Corticosteroid inhibition of pituitary GH releasehas been reported in adults and postulated as themechanism responsible for growth inhibition insteroid-treated children (2). In the present study,plasma GH concentrations were measured in 33asthmatic children including 23 who had receivedprolonged corticosteroid therapy. The data indi-cate that corticosteroid-treated children secretegrowth hormone and maintain plasma GHconcen-trations which are similar to those of nonsteroid-treated control subjects, both during the fastingstate and after induced hypoglycemia. GH valuesin individual children studied during therapy andafter steroid withdrawal were found to be fairlyreproducible and showed no differences in relationto corticosteroid administration. The plasmagrowth hormone concentrations reported here aresimilar to the values which have been observedin other children (17-19).

The protocol used in the present study differedfrom those used by other investigators (2, 3) forstudying corticosteroid-treated subjects. In con-trast to previous studies (2, 3) in which plasmaGH response to induced hypoglycemia was evalu-ated within 1-2 hr after corticosteroid administra-tion, most of the present studies were performedapproximately 12-16 hr after the last dose ofsteroid (prednisone). Our data therefore provideno information about the acute effects of corti-costeroid on pituitary function. However, the dataclearly indicate that prolonged suppression of pi-tuitary GH release does not occur in the steroid-treated children. Since maintenance corticosteroidtherapy is usually administered within an 8-12 hrdaytime period and allows a 12-16 hr intervalbetween the last evening dose and the first morn-ing dose, our results suggest that corticosteroid-treated children are capable of maintaining normalplasma GH levels during a considerable portionof the day.

The demonstration of normal plasma GHlevelsin children receiving therapy in doses known to

l)e associated with growth retardation (20, 21)argues against deficiency of GH as the primarymechanism responsible for dwarfism. Similar stud-ies have shown that children with other non-pituitary forms of dwarfism also have normallevels of GH during fasting and after hypo-glycemia (17-19). However it is recognized thatpituitary release of GHin response to an artificial,severe hypoglycemic stimulus may not be com-parable to GH release under more normal condi-tions. It still remains to be demonstrated whetherthe mean daily plasma concentration to whichtissues are exposed is similar in corticosteroid andother dwarfed children as compared with normalcontrols.

It was of interest that the mean GH valueswere similar in steroid-treated and control childrendespite significant differences in degree and dura-tion of hypoglycemia. It has been shown in adultsthat small doses of insulin may not produce suffi-cient hypoglycemia to cause maximal release ofpituitary GH (2, 22). On the other hand, the datapresented by Greenwood, Landon, and Stamp sug-gest that there is a maximal stimulus, above whichgreater degree or duration of hypoglycemia doesnot result in further rise in plasma GH concen-tration (22). The results of the present studysuggest that the degree of hypoglycemia achievedin the steroid-treated subjects is sufficient to causemaximal release of GH in most children, and thatmore severe hypoglycemia represents a supra-maximal stimulus. It is possible, however, that thelower postinsulin GH values observed in some

steroid-treated children are related to the lesserdegree of hypoglycemia in this group.

The current study further emphasizes the dif-ference in magnitude of GH levels in children as

compared with adults. Adults have low fastinglevels of GH (< 5 mug/ml) and show a brisk re-

sponse to hypoglycem a with postinsulin GHvaluesof > 40 mtg/ml (23). In contrast the children inthis study, as well as those reported by other in-vestigators (17-19), have higher fasting levels ofGH (6-8 mug/ml) and show a smaller magnitudeof rise in GH level (15-30 mttg/ml) after hypo-glycemia. Data reported by Cornblath, Parker,Reisner, Forbes, and Daughaday have shown thatyoung infants have even higher fasting GH levels(15-60 mn4g/ml) (24). These studies indicate thatplasma GHlevels can be expected to vary with age

434 H. G. Morris, I. R. Jorgensen, and S. A. Jenkins

and suggest that some of the mechanisms thatcontrol pituitary GH release may be different inchildren and in adults.

ACKNOWLEDGMENTSThe authors gratefully acknowledge the help and adviceof Dr. C. J. Falliers, C. J. Hlad, and Dr. David W.Talmage.

This work was supported by a V.A. Clinical Investi-gatorship and by a U. S. Public Health Service Researchgrant HD-02248. Dr. Morris is a V.A. Clinical Investi-gator.

REFERENCES

1. Morris, H. G., J. R. Jorgensen, H. Elrick, and R. E.Goldsmith. 1968. Metabolic effects of human growthhormone in corticosteroid-treated children. J. Clin.Invest. 47: 436.

2. Frantz, A. G., and M. T. Rabkin. 1964. Humangrowth hormone, clinical measurement, response tohypoglycemia, and suppression by corticosteroids. NewEngl. J. Med. 271: 1375.

3. Hartog, M., M. A. Gaafar, and R. Fraser. 1964. Effectof corticosteroids on serum growth hormone. Lancet.2: 376.

4. Cawley, L. P., F. E. Spear, and R. Kendall. 1959.Ultramicro chemical analysis of blood glucose withglucose oxidase. Am. J. Clin. Pathol. 32: 195.

5. Fingerhut, B., R. Ferzola, and W. H. Marsh. 1963.Application of a ferrocyanide phosphomolybdate reac-tion to an automated determination of serum glucose.Clin. Chim. Acta. 8: 953.

6. Glick, S. M., J. Roth, R. S. Yalow, and S. A. Berson.1963. Immunoassay of human growth hormone inplasma. Nature. 199: 784.

7. Fisher, R. A. 1954. Statistical Methods for ResearchWorkers. Hafner Publishing Co., Inc., New York.12th edition.

8. Greenwood, F. C., W. M. Hunter, and J. S. Glover.1963. The preparation of 'I-labelled human growthhormone of high specific radioactivity. Biochem. J. 89:114.

9. Smithies, 0. 1955. Zone electrophoresis in starch gel:group variation in serum proteins of normal humanadults. Biochem. J. 61: 629.

10. Reisfeld, R. A., B. G. Hallows, D. E. Williams, N. G.Brink, and S. L. Steelman. 1963. Purification of hu-man growth hormone on "Sephadex G-200." Nature.197: 1206.

11. Yalow, R. S., and S. A. Berson. 1960. Immunoassayof endogenous plasma insulin in man. J. Clin. Invest.39: 1157.

12. Elrick, H., V. Yearwood-Drayton, Y. Arai, F. Leaver,and H. G. Morris. 1963. Collection, processing andfractionation of human pituitary glands. J. Clin. Endo-crinol. Metab. 23: 694.

13. Hartog, M., M. A. Gaafar, B. Meisser, and R. Fraser.1964. Immunoassay of serum growth hormone inacromegalic patients. Brit. Med. J. 2: 1229.

14. Hunter, W. M., and F. C. Greenwood. 1964. A Radio-immunoelectrophoretic assay for human growth hor-mone. Biochem. J. 91: 43.

15. Frantz, A. G., and M. T. Rabkin. 1965. Effects ofestrogen and sex difference on secretion of humangrowth hormone. J. Clin. Endocrinol. Metab. 25: 1470.

16. Brown, J. B., B. Bulbrook, and F. C. Greenwood.1957. An evaluation of a chemical method for theestimation of oestriol, oestrone, and oestradiol-17B inhuman urine. J. Endocrinol. 16: 41.

17. Kaplan, S. L., C. A. L. Abrams, J. J. Bell, F. A.Conte, and M. M. Grumbach. 1965. Serum growthhormone response to insulin-induced hypoglycemia indisorders of growth. Sixth Pan-American Congress ofEndocrinology. (Abstr. No. 152)

18. Root, A. W., R. L. Rosenfield, A. M. Bongiovanni,and W. R. Eberlein. 1967. The plasma growth hor-mone response to insulin-induced hypoglycemia in chil-dren with retardation of growth. Pediatrics 39: 844.

19. Stimmler, L., and G. A. Brown. 1967. Growth hor-mone secretion provoked by insulin-induced hypo-glycaemia in children of short stature. Arch. DiseaseChildhood 42: 232.

20. Blodgett, F. M., L. Burgin, D. Iezzoni, D. Gribetz,and N. B. Talbot. 1956. Effects of prolonged cortisonetherapy on the statural growth, skeletal maturationand metabolic status of children. New Engl. J. Med.254: 636.

21. Falliers, C. J., L. S. Tan, J. Szentivanyi, J. R. Jorgen-sen, and S. C. Bukantz. 1963. Childhood asthma andsteroid therapy as influences on growth Am. J. Dis-eases Children. 105: 127.

22. Greenwood, F. C., J. Landon, and T. C. B. Stamp.1966. The plasma sugar, free fatty acid, cortisol andgrowth hormone response to insulin. I. In control sub-jects. J. Clin. Invest. 45: 429.

23. Roth, J., S. M. Glick, R. S. Yalow, and S. A. Ber-son. 1963. Secretion of human growth hormone:physiologic and experimental modification. Metab.Clin. Exptl. 12: 577.

24. Cornblath, M., M. L. Parker, S. H. Reisner, A. E.Forbes, and W. H. Daughaday. 1965. Secretion andmetabolism of growth hormone in premature and full-term infants. J. Clin. Endocrinol. Metab. 25: 209.

Endogenous GH in Steroid-Treated Children 435