0021-972X/92/7404-0865$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 0 1992 by The Endocrine Society

Vol. 74, No. 4 in U.S.A.

Printed

Recombinant Human Growth Hormone Enhances the Metabolic Efficacy of Parenteral Nutrition: A DoubleBlind, Randomized Controlled Study* THOMAS R. ZIEGLER, JOHN L. ROMBEAU, LORRAINE S. YOUNG, YUMAN FONG, MICHAEL MARANO, STEPHEN F. LOWRY, AND DOUGLAS W. WILMORE From the Departments of Medicine (T.R.Z.) and Surgery (0. W. W., Harvard Medical School, Boston, Massachusetts; the Department of University of Pennsylvania and the VA Medical Center, University Pennsylvania; and the Department of Surgery (Y.F., M.M., S.F.L.), Center, New York, New York

ABSTRACT. This multicenter, randomized, double-blind study was performed to investigate whether recombinant GH improves the efficacy of total parenteral nutrition (TPN). Fifteen stable patients requiring parenteral feeding due to gastrointestinal/pancreatic disease were studied. Constant maintenance TPN providing approximately 30 kcal/kg day and approximately 1.6 g protein/kg+day was administered during an initial 7-day baseline period. After randomization, daily SCinjections of saline (control, n = 9) or GH (10 mg/day, n = 6) were administered a 14-day treatment period as nutrient intake remained constant. Elemental balances for nitroaen (N),.._notassium (K). nhosnhorus (P), and sodium (Na) were-determined daily and’seriai blood indices, vital signs, and other clinical parameters were monitored. Nutrient balances approached equilibrium during the baseline week in both groups. With GH administration, a significant increase in N, K, and P balance occurred; in contrast, nutrient

L.S. Y.) Brigham and Women$ Hospital, Surgery (J.L.R.) the Hospital of the of Pennsylvania, Philadelphia, New York Hospital-Cornell Medical

balances did not change significantly from baseline values in control patients. The cumulative change (A) in nutrient balances from the baseline week was also significantly greater in the GHtreated patients (A N: control + 2 + 7 g vs. GH + 36 + 6. g, P < 0.005: A K: + 57 f 45 mmol vs. + 199 f 19 mmol. P < 0.03: A P: - 27 & 30 mmol us. +91 + 69 mmol, P < 0.02). ’ Plasma insulin-like growth factor-I concentrations rose 5 fold and serum cholesterol rose slightly with GH; no other significant change in group mean blood values occurred. One patient receiving GH and chronic prednisone therapy developed moderate hyperglycemia and mild peripheral edema; no other deleterious effects attributable to GH were observed. GH was well tolerated and significantly enhanced nutrient retention compared to standard parenteral feeding alone. GH improves the efficiency of parenteral nutrient utilization in patients requiring TPN. (J Clin Endocrinol Metab 74: 865-873, 1992)

A

FTER sepsis, trauma, operation, or other catabolic illnesses well described metabolic alterations occur which result in negative nitrogen balance and loss of body protein. Erosion of lean tissue is also common in hospitalized patients receiving inadequate energy or protein intake. Whereas short-term losses are rarely clinically significant, prolonged protein breakdown results in decreased resistance to infection, poor wound healing, and delayed recovery (1). Parenteral nutrition, providing adequate protein, energy, and other essential nutrients, is commonly administered to patients with gastrointestinal diseases which preclude enteral feeding. Although this approach atten-

uates protein losses, conventional nutritional support is often unable to increase or even maintain body protein, particularly during catabolic states (1). Investigators have previously attempted to modify the catabolic response to illness with the administration of anabolic hormones, including insulin (2) and pituitary derived GH (3). However, this form of GH was in short supply and not readily available to researchers. Recently, recombinant-manufactured GH has been developed and has stimulated interest in its use as an adjunct to specialized nutritional support. The overall effect of GH on protein metabolism has been well documented. For example, GH administration causes an increase in linear growth in GH-deficient children (4), whereas in normal adult subjects (5), burn patients (6-9), or elderly males (lo), exogenous GH consistently increases the overall retention of nitrogen (N) and other cellular constituents. However, few studies

Received April 24, 1991. Address requests for reprints to: Dr. Douglas W. Wilmore, Department of Surgery, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02215. Supported by a grant from Genentech Inc. and Trauma Center Grant 5P506M 36428-03. 865

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2016. at 19:13 For personal use only. No other uses without permission. . All rights reserved.

866

ZIEGLER

have evaluated GH effects in hospitalized patients requiring total parenteral nutrition (TPN). Recent studies in healthy subjects (11) and stable hospitalized adult patients (12) have documented significant improvements in elemental balances of N, potassium (K), and phosphorus (P) when GH (10 mg/day) was administered with TPN providing adequate protein. This anabolic effect occurred even with hypocaloric feedings providing 30-60% of energy requirements (11, 12). The purpose of this trial was to extend previous studies of GH to patients receiving full maintenance parenteral nutrient intake in a multicenter, prospective, randomized, double-blind format. Thus, recombinant GH or placebo was administered to adults requiring prolonged TPN as a standard component of their clinical care.

Materials

and Methods

Patients Patients were recruited from the medical and surgical services of Brigham and Women’s Hospital (Boston, MA); the VA Medical Center and the Hospital of the University of Pennsylvania (Philadelphia, PA), and the New York Hospital-Cornell Medical Center (New York, NY). Informed consent was obtained in accordance with policies established by the Institutional Review Boards at each institution. The patients were adult males and females aged 18-85 yr (see Table 1). Eligible patients had gastrointestinal (GI) and/ or pancreatic dysfunction which precluded adequate enteral feeding and all were anticipated to require TPN for at least 3 weeks. Patients were in a stable course of their illness, without uncontrolled infection or other acute non-G1 organ dysfunction. Six patients required chronic glucocorticoid therapy. Patients were ineligible for study if they had a history of: primary or metastatic cancer diagnosed within the previous 2 yr, diabetes mellitus, AIDS, cirrhosis, hepatitis, rheumatoid arthritis, or other autoimmune disorders. Criteria for patient exclusion from the final data analysis were established before study onset to ensure comparisons between clinically stable patient groups. For data analysis, patients must have completed at least 3 study weeks (pretreatment and two consecutive treatment weeks) without the development of severe organ system dysfunction [e.g. renal failure (serum creatinine > 130 pmol/L] or uncontrolled infection, characterized by poor clinical response within 48 h of appropriate treatment. Study design The patients initially received TPN during a l-week baseline evaluation period to allow for metabolic and clinical adaptation to the study protocol and the iv diet prescription. Patients were then randomized to either a treatment or control group. The two study groups were balanced for diagnosis and degree of weight loss. After the baseline period, the treatment phase commenced and either 10 mg rhGH (Genentech Inc., S. San Francisco, CA) or a similar volume of reconstituted excipient

ET AL.

JCE & M. 1992 Vol74.No4

(control) was administered by daily SC injection. This dose of GH was selected in order to study effects of pharmacologic GH doses previously shown to safely exert anabolic effects in stable (11, 12) and critically ill (6, 9) individuals. Injections of GH or saline were administered at 0830 h and continued daily for at least two consecutive weeks. Clinical care was otherwise provided as indicated by the patient’s primary physician(s). The investigators, patients, and all other clinical staff members were blinded as to the injection composition. Individual calorie and protein prescriptions were calculated for each subject at entry. Caloric requirements were based on standard methods as previously described (13). Protein (1.5 g/ kg. day) was provided as crystalline amino acids (Aminosyn I, Abbott, N. Chicago, IL). Half of nonprotein calories was given as glucose and half as fat emulsion (Intralipid 20%, Kabivitrum, Stockholm, Sweden). Adequate electrolytes, vitamins and minerals were administered in the nutrient solutions. The electrolyte concentrations were adjusted daily as appropriate to maintain normal serum levels. TPN was infused at a constant rate over 24 h daily. Patients received nothing by mouth except distilled water or other noncaloric liquids. Patient evaluations Initial anthropometric evaluation included height, weight, ideal body weight, percent weight loss from usual, mid-arm circumference, and mid-arm and subscapular skin fold (Lange Calipers, Cambridge Scientific Industrial Inc., Cambridge, MD). Evaluation of handgrip strength was also evaluated using a dynamometer (Azimow Engineering Co., Los Angeles, CA). Initial laboratory evaluation included fasting blood for complete blood count (CBC), urinalysis, serum chemistry profile, iron, transferrin, retinol binding protein, prealbumin, FFA, Tq, and TSH. Blood obtained at each institution was analyzed in a central laboratory using standard clinical laboratory techniques. Serum concentrations of prealbumin and retinol-binding protein were determined using rate nephelometry. Body weight, vital signs, caloric intake, and total intake and output of fluid, N, K, P and sodium (Na) were obtained daily. Daily 24-h total urine collections were analysed for total volume and creatinine, urea, N, Na, K, P, and glucose concentrations. All fistula drainage, nasogastric tube losses, and stool were collected in 24-h pools and analyzed for total volume and concentrations of N, Na, and K. Determinations of N and mineral intake, output, and balance were performed as previously described (12). On day 7, fasting chemistries, CBC, urinalysis, plasma total insulin-like growth factor I (IGF-I), serum GH antibodies, FFA, iron, transferrin, prealbumin, retinol-binding protein, and insulin were measured. Insulin was measured at a commercial laboratory by RIA, total IGF-I concentration (in EDTAplasma) was determined using a commercial kit (Nichols Institute, San Juan Capistrano, CA) and serum GH antibodies were measured at Genentech Inc. Anthropometric parameters were also determined on day 7. The same daily and weekly clinical and metabolic measurements were obtained during the treatment phase as in the baseline week, beginning with the first treatment day (day 8 of study). Serum Tq and TSH were also measured on the last day of the second treatment week.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2016. at 19:13 For personal use only. No other uses without permission. . All rights reserved.

GH AND

EFFICIENCY

OF PARENTERAL

The clinical course of all patients was evaluated daily by the investigators in each participating institution and all clinically significant aspects of care recorded. Prestudy guidelines were developed to facilitate responses to potential GH side effects or other medical conditions (e.g. hyperglycemia, fluid retention, infection) to assure clinical uniformity between the control and experimental groups. Statistical analysis A total of 23 patients initially entered the trial; however, eight of these individuals were excluded from data analysis using pre-hoc criteria. Four patients (3 GH; 1 control) did not complete the requisite 3 weeks of TPN; one GH patient was removed from the study after 4 days of treatment by the attending physician due to the development of hyperglycemia (< 18 mmol/L), whereas two GH and one control patients were discontinued from TPN because enteral nutrition was well tolerated. Two additional patients (1 GH; 1 control) developed renal dysfunction and uncontrolled infection respectively. Finally, two patients (1 control; 1 GH) were classified as statistical outliers. The control patient demonstrated inordinately high values for nitrogen balance (outside 2 interquartile distances from the median), while the GH patient exhibited a markedly attenuated IGF-I response to GH injections, attributable to incomplete drug delivery (this was confirmed by additional unblinded studies of GH injection in this individual demonstrating expected IGF-I responses). Although inclusion of these data did not alter the statistical significance of the results or the conclusions of this study, use of this data prevents an accurate reflection of the metabolic state of the patients studied and their respective response to treatment. Thus, the final data set was derived from 15 eligible patients ( control; 6 GH). Mean weekly values for nutrient intake, nutrient balance, vital signs, blood values, and anthropometric measurements in each study group were calculated. To determine effects on nutrient balance from baseline, the change (A) from baseline week values was determined for each treatment week (A = treatment week 1 or 2 values - basal value). The cumulative A in nutrient balances from baseline for the entire 14-day treatment period was also calculated. Comparisons between the two study groups over time used two-way repeated measures analysis of variance (ANOVA). When significant treatment effects were noted, a post-hoc Dunnett’s t test was used to compare mean baseline week values to treatment week values within the particular group. Unpaired t tests were used for data in which time was not a variable. Probability values less than 0.05 were considered statistically significant. Data are presented as mean f SEM.

Results Clinical observations The study groups were similar in gender, age, weight, height, body surface area, percent weight loss from usual weight, current weight as a percentage of ideal body weight, and primary clinical diagnosis (Table 1). Doses

TABLE

FEEDING

867

1. Clinical characteristics of the patients Characteristic

Control (n = 9) 5 female 4 male 48.3 +- 6.6

Gender Age W (20.6-80.7) bnge) 54.9 + 3.9 Wt (kg) (42.0-81.4) hwe) Ht (cm) 167 + 4 Body surface area (m”) 1.62 + 0.06 Weight loss from usual (%) 5.6 + 1.5 (O-13) b-ange) Wt per IBW (%)’ 97 + 6 Diagnosis (no.) enterocutaneous/pancreatic fistula Chronic pancreatitis short bowel syndrome gastric bypass/enteral food intolerance Chronic corticosteroid administration (no.) Mean + SEM; range in parentheses. ’ Ideal body weight.

(nGT6) 3 female 3 male 42.8 + 6.1 (28.3-65.7) 55.4 It 4.3 (42.2-70.4) 167 +4 1.63 + 0.08 7.6 + 3.0 (O-19) 96 k 4

of corticosteroids remained stable during the study in the six individuals receiving these agents. During the treatment phase, all daily placebo and GH injections were administered without adverse local or systemic allergic reactions. The mean daily GH dose in the six subjects was 0.186 f 0.014 mg/kg.day (range: 0.142 - 0.237 mg/kg.day). Mean daily pulse rate rose slightly over time in patients receiving GH (82 f 1 beats per minute baseline vs. 84 + 2 and 88 + 2 GH weeks 1 and 2; P < 0.05 vs. baseline) while other vital signs were unaltered. Anthropometric measurements were similar between groups over time (not shown). Two control patients required removal of the central venous catheter due to catheter tip infection and bacteremia, respectively, and peripheral vein feedings were used for 4 days until reinsertion of the central venous catheter. No other significant alterations in the study nutrient prescriptions occurred in either group. One GH patient developed a pseudoaneurysm of the gastroepiploic artery at the completion of the baseline week. He was removed from study and underwent operative repair; after a clinically uneventful postoperative recovery week he entered the treatment phase. Another patient receiving prednisone therapy required insulin and a reduction in the carbohydrate dose during GH to control moderate hyperglycemia (60 yr old) fed orally, cholesterol levels were reduced with administration of lower GH doses (0.03-0.12 mg/ kg.day) (21). The reason for this discrepancy is unclear but may relate to the differing nutrient substrates administered. Iatrogenic acromegaly represents a potential complication of long-term pharmacologic rhGH administration; however, symptoms or signs consistent with this disorder have not been observed in the short-term trials performed to date (11-12, 27). As a growth factor, GH also has the potential to stimulate neoplastic growth; thus, patients with malignancies have not been entered into these clinical trials. This study documents improved metabolic efficacy of TPN with GH; other important clinical endpoints such as length of hospital stay were not evaluated. In order for GH to be considered a useful adjunct to nutrition support, decreased morbidity must be documented. However, recent studies have further defined the therapeutic efficacy of this approach. In postoperative patients GH administration maintained hand grip strength which was reduced in control patients receiving identical nutrient intake (22); improved isometric strength, and exercise capacity in nonhospitalized GH-deficient adults was also documented (20,30). GH treatment significantly reduced urea generation and serum phosphorus in adults requiring maintenance hemodialysis (31). Additionally, GH administration improved wound healing in burn patients (32, 33), and significantly reduced the length of hospital stay in burned children (33). In summary, this controlled multicenter trial documents that GH administration improves the metabolic efficacy of TPN in stable adult patients. These findings

ET

JCE & M. 1992 Voll4.No4

AL.

extend previous short-term metabolic studies and suggest that recombinant GH may improve the utilization of administered nutrients. GH administration as an adjunct to TPN represents a potentially useful strategy for enhancing nutritional recovery of patients with illnesses requiring specialized metabolic/nutritional care (1). However, additional studies are necessary to document improved clinical outcomes and to further define patient groups in whom GH administration may be of benefit. Acknowledgments The authors particularly thank Barry M. Sherman M.D. Andrew Perlman M.D. and Karen Van Slett R.N. of Genentech Inc. for their

efforts

in sponsoring

the study. The

help

of

Kathleen Benfell R.Ph., and John Fanikas R.Ph. of the research pharmacy of technical assistance Ramona Faris M.S., B.S., and SusetteM.

Brigham and Women’s Hospital and the of Elaine Brown M.S., David White M.S., Lourdes Holejko B.S., Christopher Keogh

Coyle R.N., are alsogratefully appreciated. References

5.

6.

7.

8.

9.

10. 11.

Wilmore DW. Catabolic illness-Strategies for enhancing recovery. N Engl J Med. 1991;325:695-702. Woolfson AMJ. Heatlev RV. Allison SP. Insulin to inhibit urotein catabolism after injury: N Engl J Med. 1979;300:14-17. Roe CF, Kinney J. The influence of human growth hormone on energy sources in convalescence. Surg Forum. 1962;13:369-71. Collip PJ, Curti V, Thomas J, Sharma RK, Maddaiah VT, Cohn SE. Body composition changes in children receiving growth hormone. Metabolism. 1973;22:589-95. Henneman PH, Forbes AP, Moldawer M, Dempsey EF, Carroll EL. Effects of human growth hormone in man. J Clin Invest. 1960; 39:1223-38. Liljedahl S, Gemzell C, Plantin L., Birkin G. Effect of human growth hormone in patients with severe burns. Acta Chir Stand. 1961;122:1-14. Prudden JF, Pearson E, Sorhoff HS. Studies on growth hormone. II. The effect on nitrogen metabolism in severely burned patients. Surg Gynecol Obstet. i956;102:695-701. Sorhoff HS, Rozin RR, Mooty J, Lister J, Raben MA. Role of human growth hormone in response to trauma. I. Metabolic effects following burns. Ann Surg. 1967;166:739-52. Wilmore DW. Movlan JA. Bristow BF. Mason AD. Pruitt BA. Anabolic effects of