Protein loss leading to reduced lean body mass is recognized to contribute to the high levels of morbidity and mortality seen in critical illness. This prospective, randomized, controlled study compared the effects of conventional parenteral nutrition (TPN), glutamine-supplemented (0.4 g·kg-1·day-1) TPN (TPNGLN), and TPNGLN with combined growth hormone (GH, 0.2 IU·kg-1·day-1) and IGF-I (160 μg·kg-1·day-1) on protein metabolism in critical illness. Nineteen mechanically ventilated subjects [64 ± 3 yr, body mass index (BMI) 23.8 ± 1.3, kg/m2] were initially studied in the fasting state (study 1) and subsequently after 3 days of nutritional with/without hormonal support (study 2). All had recently been admitted to the ICU and the majority were postemergency abdominal surgery (APACHE II 17.5 ± 1.0). Protein metabolism was assessed using a primed constant infusion of [1-13C]leucine. Conventional TPN contained mixed amino acids, Intralipid, and 50% dextrose. TPNGLN, unlike TPN alone, resulted in an increase in plasma glutamine concentration (∼50%, P < 0.05). Both TPN and TPNGLN decreased the rate of protein breakdown (TPN 15%, P < 0.002; TPNGLN 16%, P < 0.05), but during these treatments the patients remained in a net negative protein balance. Combined treatment with TPNGLN + GH/IGF-I increased plasma IGF-I levels (10.3 ± 0.8 vs. 48.1 ± 9.1 nmol/l, study 1 vs. study 2, P < 0.05), and in contrast to therapy with nutrition alone, resulted in net protein gain (-0.75 ± 0.14 vs. 0.33 ± 0.12 g protein·kg-1·day-1, study 1 vs. study 2, P < 0.05). Therapy with GH/IGF-I + TPNGLN, unlike nutrition alone, resulted in net positive protein balance in a group of critically ill patients.
- intensive care
- total parenteral nutrition
critical illness associated with trauma, burns, sepsis, and major surgery results in a protein-catabolic state, characterized by loss of lean tissue (5). Excessive protein catabolism impairs immunocompetence, increases infection rates, and causes poor wound healing and muscle weakness (2, 3). These features may contribute to the high rates of morbidity and mortality observed in the critically ill (15, 45). The preservation of lean tissue has implications in terms of both patient outcome and health care expenditure for those caring for the severely ill individual, and therapies that improve protein anabolism, decrease catabolism, or both would be valuable in the treatment of serious illness.
Commonly, intravenous nutritional support [total parenteral nutrition (TPN)] is used to support the ill subject, particularly in situations where enteral feeding is inadequate or impossible (38). Use of TPN may improve outcome in some patients, although it is increasingly recognized that, despite intravenous feeding regimens, many patients remain catabolic and continue to lose whole body protein stores (36). Although glutamine is recognized to have a central role in protein metabolism, conventional TPN regimens do not contain glutamine for reasons related to sterility and solubility. However, critical illness results in altered glutamine metabolism typically characterized by decreased plasma glutamine levels (17). This observation has led to the speculation that, during times of severe illness, metabolic demand for glutamine exceeds the capacity for de novo synthesis, leading to the description of glutamine as a “conditionally essential” amino acid (37).
Newer strategies to improve protein metabolism and balance in critically ill subjects have focused on the use of specialized nutrition, including glutamine supplementation (1, 35). Studies to date indicate that the addition of intravenous or enteral glutamine reduces infection rates (16, 47) and length of hospital stay (34) and may improve mortality in intensive care unit (ICU) patients (14). The availability of recombinant growth hormone (GH) and insulin-like growth factor I (IGF-I) have led to considerable interest in the anabolic potential of these agents. Administration of GH has been shown to have protein-anabolic effects in critically ill subjects following both elective and emergency surgery (27, 23) and in patients with sepsis (21). In healthy subjects in whom hypercatabolism was induced by starvation, combined treatment with GH and IGF-I was shown to have synergistic effects on protein metabolism (22).
In a recent multicenter trial of GH treatment of critically ill patients, GH treatment was found to increase mortality. This was unexpected, since all previous studies of GH treatment in critically ill subjects had shown no adverse effects (39). The possible causes that have been suggested are that GH had an immunomodulatory effect, that the doses of GH may have been inappropriate, and that GH may have exacerbated nutritional (in particular glutamine) deficiency. Before the results of the multicenter trial became known, we had begun this randomized, controlled study to investigate the effects of TPN, TPN with the addition of glutamine (TPNGLN), and TPNGLN with combined GH/IGF-I on protein metabolism in critically ill patients. The results of the multicenter trial of GH treatment were announced during this study, which was then terminated. The numbers of patients in each group are therefore smaller than had originally been planned (originally n = 12 in each group).
MATERIALS AND METHODS
Experimental subjects. Nineteen critically ill subjects were included in the study. They were all newly admitted to the ICU, the majority following emergency abdominal surgery, and required mechanical ventilation and intravenous nutritional support. Criteria for inclusion in the study included patients newly admitted to the ICU and likely to require intravenous nutritional support for ≥3 days but without advanced renal or hepatic failure, sepsis, or active infection. Patients with known malignancy, diabetes mellitus, or other endocrine conditions, including a need for exogenous glucocorticoids, were excluded. The patients were recruited consecutively throughout an 18-mo period. Fourteen had undergone emergency abdominal surgery, four had gastrointestinal obstruction, and one was postcardiac arrest. The patients' clinical and metabolic characteristics are summarized in Table 1. The severity of illness was evaluated on the day of the first study by use of the APACHE II (20) and Therapeutic Intervention Scoring Systems (4); Table 1), and these indexes were appropriate for severely ill patients requiring cardiorespiratory and nutritional support. Intravenous sedation was used as necessary to facilitate ventilation with continuous infusions of midazolam and morphine, titrated to ensure a constant level of sedation throughout the duration of the protocol. No patients received epidural analgesia. The study was approved by the Ethics Committee, Guy's and St. Thomas' National Health Service Trust, and written informed consent was obtained from first-degree relatives of the patients.
Study design. Patients were randomized to receive TPN (n = 7), TPNGLN (n = 7), or TPNGLN with GH/IGF-I (n = 5). All subjects were first studied within 48 h of their admission in the fasting state (study 1), commencing at 0900, following a fast of >12 but ≤24 h. On completion of study 1, the patients were commenced on nutritional with or without anabolic hormone support for 72 h, with a repeat study (study 2) performed at the end of this period with continuation of the intravenous nutrition (±GH/IGF-I) during study 2. The TPN provided 30 kcal·kg-1·day-1 and consisted of a mixed amino acid infusion (Vamin 14; Pharmacia & Upjohn, Milton Keynes, UK), 50% dextrose, and 20% Intralipid (Kabivitrum, Stockholm, Sweden). Each patient received this TPN combination, which provided 0.2 g nitrogen·kg-1·day-1. In all treatment groups, a continuous intravenous insulin infusion (Actrapid; Novo Nordisk, Copenhagen, Denmark) was provided if necessary to maintain a blood glucose ≤8.0 mmol/l. The two groups randomized to TPNGLN received, in addition to an identical TPN regimen, intravenous l-glutamine (0.4 g·kg-1·day-1; Oxford Nutrition, Oxford, UK). The patients treated with the combined anabolic agents received subcutaneous recombinant human (rh)GH (0.2 IU·kg-1·day-1, Pharmacia & Upjohn) and subcutaneous rhIGF-I (160 μg·kg-1·day-1, Pharmacia & Upjohn). The rhGH was administered at 1200 as a single dose and the rhIGF-I in equal split doses at 1200 and 2200, respectively.
Protein metabolism was assessed using a stable isotope technique. On each study occasion, a primed intravenous infusion of l-[1-13C]leucine (1 mg/kg; 1 mg·kg-1·h-1; MassTrace, Somerville, MA) was administered for 240 min. The bicarbonate pool was primed with a bolus of NaHCO3 (0.2 mg/kg, MassTrace). Regular blood and breath samples were taken at baseline and steady state (210-240 min) for the measurement of the enrichment of α-ketoisocaproic acid (α-KIC), the concentration of leucine, and the enrichment of expired CO2. Blood samples were taken from indwelling arterial catheters and the breath samples from the expiratory port of the ventilator (Servo 900; Siemens, Berlin, Germany). All sedative and inotropic infusions remained constant throughout the studies. Regular blood samples were also taken for routine biochemistry and hematology and measurement of plasma glucose, insulin, GH, total IGF-I, IGF-binding protein-1 (IGFBP-1), cortisol, free thyroxine (FT4), free triiodothyronine (FT3), glucagon, and amino acids.
Analytical methods. α-KIC enrichment was measured as the quinoxalinol-tert-butyldimethylsilyl derivative under selected ion monitoring by gas chromatography (Hewlett-Packard 5890)-mass spectrometry (MSD 5971A; Hewlett-Packard, Berkshire, UK) at mass-to-charge ratios 259 and 260. Plasma α-KIC enrichment is a measure of intracellular leucine enrichment (24). 13C enrichment of breath CO2 was measured on a VG SIRA series II isotope ratio mass spectrometer (VG Isotech, Cheshire, UK, modified with a Roboprep G+ inlet system, Europa Scientific, Cheshire, UK). The CO2 production rate was measured using an online mass spectrometer (Airspec 2200; Airspec, Kent, UK) in the ICU.
Plasma glucose concentrations were measured on a model 23AM glucose analyzer (YSI, Hampshire, UK). Serum insulin concentrations were measured by an in-house double antibody radioimmunoassay. Plasma total IGF-I was measured by radioimmunoassay after an ethanol-hydrochloric acid extraction [within-assay coefficient of variation 7% (40)]. IGFBP-1 was measured by a specific immunoradiometric method as the 25.3-kDa protein by means of a commercially available kit (Diagnostic Systems Laboratories). The assay was found to be specific, and there was no interference by the presence of other human IGFBPs, proinsulin, recombinant insulin (Sigma Chemical, St. Louis, MO) and GH. The accuracy range of the assay was within 10.1% of the mean expected value. Cortisol was measured by ELISA, using the Enzymun-Test cortisol kit (Boehringer Mannheim, Sussex, UK). Free thyroid hormones were measured by a competitive immunoassay using chemiluminescence (Chiron Diagnostics, Essex, UK). Glucagon was measured using a commercial radioimmunoassay (Linco Research, St. Louis, MO). Plasma amino acids were measured by ion exchange chromatography using an Alpha Plus II (Pharmacia, Cambridge, UK) automated amino acid analyzer.
Measurements of leucine metabolism were calculated using standard isotope dilution equations. Leucine production rate (Ra; a measure of whole body protein degradation) was calculated as Ra = F[1/(APE × 0.01)], where F is the isotope infusion rate (μmol·kg-1·min-1) and APE is plasma α-KIC enrichment. Leucine disappearance rate (Rd) was assumed to be equal to Ra. Leucine oxidation rate (OX) was calculated as OX = (APECO2 × CO2Ra)/APEKIC, where CO2Ra is the production rate of CO2 and APECO2 the enrichment of expired CO2. In all cases, leucine oxidation was corrected by assuming that 80% of 13CO2 was expired (29). Nonoxidative leucine disposal (NOLD, a measure of whole body protein synthesis) was calculated as the difference between Rd and OX. During the second study, endogenous leucine Ra was calculated as the difference between the measured (total) Ra and exogenous Ra, where exogenous Ra is the known infusion rate of exogenous leucine from the TPN measured in μmol·kg-1·min-1. Net protein balance was estimated using the net leucine balance (NOLD - Ra), assuming 8 g leucine/100 g whole body protein (12).
Statistical analyses. All data are presented as means ± SE. Comparisons between groups were made using ANOVA (with Bonferroni's correction) and between study 1 and study 2 were made by standard two-tailed paired t-tests. Data that were not normally distributed were log-transformed prior to analysis. P values of <0.05 were considered significant.
Clinical. There were no differences in the baseline characteristics among the three groups. All subjects survived the duration of the study, although one patient died during her hospital stay, as shown in Table 1. TPN, TPNGLN with or without combined GH/IGF-I treatments were well tolerated, with no short-term adverse effects identified on carbohydrate metabolism, gas exchange, or renal, hepatic, and cardiovascular status. Continuous intravenous insulin infusions were used according to standard policy for our ICU for patients with plasma glucose concentrations >8.0 mmol/l. Accordingly, 6 of the 19 patients required insulin in the baseline fasting state (TPN group, 2 patients; TPNGLN group, 3 patients; TPNGLN + GH/IGF-I group, 1 patient). After introduction of intravenous feeding, 9 of 19 patients (TPN group, 3 patients; TPNGLN group, 4 patients; TPNGLN + GH/IGF-I group, 2 patients) required intravenous insulin to maintain glucose levels <8.0 mmol/l. No significant differences in requirement for insulin were evident among groups.
Hormones and metabolites. Plasma glucose concentrations were similar during study 1 in all three patient groups. The provision of TPN, TPNGLN, and TPNGLN + GH/IGF-I resulted in significant increases in plasma glucose levels (all groups P < 0.05; Table 2). The circulating insulin levels were similar throughout study 1 in the three treatment groups (Table 2). Treatment with both TPN and TPNGLN resulted in significant increases in plasma insulin concentrations (P < 0.01 and P < 0.05, respectively; Table 2), and although a similar effect was recorded in the patients randomized to TPNGLN + GH/IGF-I, this failed to achieve significance (P = 0.13).
The circulating total IGF-I levels were similar in each of the treatment groups during study 1. As expected, a large increase in plasma IGF-I concentration (∼5-fold) was observed following provision of TPNGLN + GH/IGF-I (P < 0.05; Table 2), with no changes recorded following either TPN or TPNGLN. Plasma IGFBP-1 levels were similar during study 1 in each of the patient groups, and although a significant decrease in IGFBP-1 concentration was recorded following treatment with TPN (P < 0.001; Table 2), no changes were observed following provision of TPNGLN or TPNGLN + GH/IGF-I (Table 2). Plasma glucagon and cortisol levels were similar in each of the groups during study 1, and no significant alterations were observed following any of the treatment regimens (Table 2). Levels of FT4 and FT3 were similar in the groups during study 1 and were unchanged by any of the treatment regimens (Table 2), with the exception of an increase in the plasma FT3 concentration following TPNGLN + GH/IGF-I (P < 0.05; Table 2).
Leucine metabolism. The leucine kinetic data, including the effects of the different treatment regimens, are shown in Fig. 1. There were some differences in the rates of leucine Ra (a measure of protein breakdown), oxidation, NOLD (a measure of protein synthesis), and net protein balance during study 1 in the three treatment groups (Figs. 1 and 2). Baseline leucine Ra was higher in the TPN group than in the TPNGLN and TPNGLN + GH/IGF-I groups (both P < 0.05). In addition, baseline NOLD was higher in TPN than in TPNGLN patients (P < 0.05), and leucine oxidation and net protein balance were higher in study 1 in the TPN group compared with those allocated to GH/IGF-I (both P < 0.05).
Intravenous feeding with both TPN and TPNGLN resulted in significant decreases in the rate of protein breakdown (endogenous leucine Ra, P < 0.002 and P < 0.05, respectively), and a similar trend was observed in those randomized to receive TPNGLN + GH/IGF-I (Fig. 1A, P = 0.18). Treatment with both TPN and TPNGLN resulted in an increase in the rate of leucine oxidation (P < 0.05 and P = 0.06, respectively), but leucine oxidation was unchanged following TPNGLN + GH/IGF-I (Fig. 1B). The rate of protein synthesis (NOLD) decreased following treatment with both TPN (P < 0.05) and TPNGLN (P < 0.05) but was maintained following provision of TPNGLN + GH/IGF-I (Fig. 1C). The net protein balance was markedly negative in all three groups during study 1 (Fig. 2). Treatment with both TPN and TPNGLN partially attenuated the net negative protein balance, but these changes failed to reach significance (Fig. 2). In contrast, those patients randomized to receive TPNGLN + GH/IGF-I achieved a net positive protein balance during study 2 (P < 0.05; Fig. 2). The magnitude of the decreases in endogenous leucine Ra were similar following each of the treatments, but treatment with both TPN (P < 0.01) and TPNGLN (P < 0.05) resulted in a greater decrease in NOLD compared with TPNGLN + GH/IGF-I. Similarly, the net protein balance was increased significantly following the provision of TPNGLN + GH/IGF-I compared with the effects of both TPN (P < 0.05) and TPNGLN (P < 0.05).
Amino acids. Table 3 displays the mean plasma amino acid levels during studies 1 and 2 for each of the groups. The total plasma amino acid concentrations were similar between groups during study 1 and increased following nutritional with or without hormonal treatment in each of the three groups, with the difference achieving significance in those randomized to TPNGLN. Plasma glutamine concentrations were similar in the groups throughout study 1, and although no change was recorded following provision of TPN alone, significant increases were seen after 3 days of treatment with both TPNGLN and TPNGLN + GH/IGF-I. The plasma leucine concentrations were similar in the three groups during study 1 and were unchanged following each of the treatment regimens (Table 3).
This study investigated the effects of conventional TPN with and without glutamine supplementation on protein metabolism in critically ill subjects and the effects of the addition of combined GH and IGF-I treatment. Provision of both TPN and TPNGLN partially attenuated endogenous protein breakdown, but despite these treatments the patients remained in a net protein-catabolic state. The addition of the combined anabolic hormones resulted in major improvements in protein metabolism, resulting in a net positive protein balance in this severely ill population.
Previous studies have addressed the effects of various nutritional support regimens on protein metabolism in critically ill subjects (28, 33). These have shown, in agreement with the results from the present study, that TPN improves nitrogen balance by decreasing proteolysis (18), but patients have remained in a net negative protein balance. The addition of intravenous glutamine supplementation to conventional feeding regimens has previously been shown to have greater protein-anabolic effects than isonitrogenous, isocaloric nutritional support when protein metabolism was assessed by measurement of nitrogen balance (35, 47). In contrast, net protein balance assessed using this sensitive isotopic technique was not substantially altered by provision of glutamine in the present study. Provision of both TPN and TPNGLN increased protein oxidation, whereas amino acid turnover was reduced, perhaps due in part to increased circulating insulin. In combination, these effects resulted in a reduction of the rate of protein synthesis. The net effects of these rate changes translated into small (nonsignificant) changes in net protein balance. It is conceivable that inclusion of a larger number of patients may have shown a subtle but potentially important benefit of glutamine on net protein balance. Additionally, glutamine supplementation may confer benefits through mechanisms independent of protein metabolism.
The major effect of the combined GH/IGF-I treatment was maintenance of amino acid incorporation into protein without an increase in use of amino acids as an oxidative fuel. Because endogenous protein breakdown was similarly altered by each of the treatment regimens, maintenance of protein synthesis was responsible for the net positive protein balance following treatment with combined GH/IGF-I. Both of these agents have been shown independently to reduce protein oxidation and promote protein synthesis in a variety of conditions (31, 32), and evidence from studies in both animals and humans indicates that the effects of both GH and IGF-I on protein oxidation and protein synthesis may be mediated by regulation of amino acid transporter mechanisms (9, 25). The major limitation of the technique used in the present study is that whole body protein metabolism was assessed. It is not known whether net protein gain occurs in all tissues, in particular in lean tissue and skeletal muscle. The production rate of CO2 and the recovery factor used in the modeling (29) are important determinants of the rates of protein oxidation and synthesis and, therefore, net protein balance. To limit error, we ensured that the feeds were isocaloric among groups and maintained similar supporting infusion rates and ventilator settings between the studies.
It is recognized that acute serious illness results in a state of GH resistance (30). This is characterized by low levels of circulating IGF-I and decreased IGF-I generation following GH administration (6). The precise mechanisms responsible for this resistant state remain unclear, but evidence suggests that it may be in part mediated by alterations in the circulating levels of the IGFBPs. The recognized changes include elevated levels of IGFBP-1 and reduced concentrations of IGFBP-3 (46). Experimental evidence suggests that the low levels of IGFBP-3 may be a result of protease activity that may exhibit nutritional dependence (7). Because IGFBP-1 has predominately inhibitory influences on IGF-I action, the reduction in the elevated levels of IGFBP-1 following TPN and TPNGLN + GH/IGF-I may have contributed to the observed improvements in protein metabolism. In addition, recent evidence suggests that increases in circulating endotoxin contribute to GH insensitivity by decreasing expression of hepatic GH receptors (8). These findings may partly explain the lack of effect of GH treatment on nitrogen balance and protein metabolism in some studies in the critically ill (13, 44). In addition to acting as a counterregulatory hormone increasing circulating insulin levels, GH has direct effects on adipose tissue increasing the rate of lipolysis (31). These actions combined with the known anabolic action on skeletal tissue provide the rationale for the potential use of pharmacological doses of GH in critical illness. More recently, investigators have studied the effects of combined treatment with both GH and IGF-I in several catabolic conditions. IGF-I has been shown to have a direct protein-anabolic action (32) but, unlike GH, improves insulin sensitivity (22). Because insulin is recognized to have a central role in protein metabolism by mediating a reduction in protein breakdown (41), maintenance of insulin sensitivity is recognized to be of importance in critically ill subjects. Thus the severely ill, catabolic patient group is one in which the combined approach of GH/IGF-I administration may be particularly appropriate (19). A study of the effect of combined GH/IGF-I in nutritionally deprived healthy controls has demonstrated a synergistic effect on protein metabolism compared with each agent being used in isolation (22). In human immunovirus-infected subjects, combined GH/IGF-I has been shown to promote lean body mass after 6 wk of treatment, although this effect was not sustained at 12 wk (10).
Results from recent multicenter European studies indicate that GH treatment following the onset of critical illness may result in increased mortality (39), but the mechanism(s) remains unclear. This finding was unexpected, because all previous studies of GH treatment in critically ill subjects had shown no adverse effects. It has also been suggested that the use of high GH doses in the multicenter European studies may have been inappropriate (42). It has recently been demonstrated that in prolonged critical illness there is a generalized hypothalamic-pituitary suppression. There is thus relative GH deficiency rather than the GH insensitivity seen in acute critical illness (43). Because patients in the previous studies had been in intensive care for 5-7 days, many may have been GH responsive, and the administration of high doses of GH may have evoked side effects that led to the deterioration of function of multiple organs (42). In the present small study, no increase in morbidity or mortality was observed with GH treatment, but this study differed from the multicenter investigation in that all patients were studied within 48 h of entering the ICU, all received TPNGLN, and both GH and IGF-I were administered. GH therapy has been shown to attenuate the skeletal muscle depletion of glutamine characteristic of serious illness (11), and it has been proposed that “prevention” of glutamine mobilization may contribute to loss of gut mucosal integrity, worsening sepsis, and multiorgan failure (26). These findings perhaps indicate that GH (with or without IGF-I) treatments in the critically ill, highly catabolic patient should be combined with adequate provision of substrates, including amino acids and, in particular, glutamine.
In summary, these results demonstrated that the addition of glutamine conferred no major benefits on protein metabolism over TPN alone but that the administration of combined GH/IGF-I treatment with glutamine-supplemented TPN resulted in net whole body protein anabolism in severely ill individuals. Although this study suggests that this treatment may be beneficial in terms of protein anabolism, the recent study showing an increase in mortality with GH treatment in critically ill patients must result in caution in the use of GH in these patients. It remains to be established whether specific groups of seriously ill individuals may benefit from GH/IGF-I treatment.
We are indebted to the nurses and staff of Mead Ward (ICU), St. Thomas' Hospital, for their patience and support. We are grateful to Pharmacia for supplying the GH and IGF-I. P. V. Carroll was supported by a fellowship from the Special Trustees (Endowments) St. Thomas' Hospital.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by American Physiological Society