key: cord-0006942-dernekty
authors: Castillo, Leticia; Yu, Y Ming; Marchini, J Sergio; Chapman, Thomas E; Sanchez, Melchor; Young, Vernon R; Burke, John F
title: Phenylalanine and Tyrosine Kinetics in Critically Ill Children with Sepsis
date: 1994
journal: Pediatr Res
DOI: 10.1203/00006450-199405000-00009
sha: 01721c2ebdf8f55b676efc82534684dcc414211e
doc_id: 6942
cord_uid: dernekty

ABSTRACT: To better understand the impact of severe illness on the amino acid economy and nutritional needs of pediatric patients, we studied plasma phenylalanine and tyrosine kinetics in eleven critically ill patients (six full-term newborns and five young infants). Within 48 h of the diagnosis of sepsis they were given primed constant i.v. infusions of L-[1-(13)C]phenylalanine and L-[3,3,(2)H(2)]tyrosine for 4 h. Routine nutritional support continued during this period by parentcral administration of dextrose, lipid emulsion, and an amino acid mixture low in tyrosine. Phenylalanine and tyrosine fluxes and rate of phenylalanine hydroxylation did not differ significantly between the two age groups, and so the data were combined for evaluation. For the entire group, values (μmol-kg(−1)-h(−1); mean ± SD) for phenylalanine and tyrosine fluxes and rate of phenylalanine hydroxylation were 132 ± 24, 66 ± 16, and 29 ± 12, respectively. Plasma phenylalanine to tyrosine concentration ratio was 1.67 ± 0.6. From a comparison of the rate of phenylalanine hydroxylation with measured phenylalanine intakes, it was concluded that their routine, clinical nutritional support was inadequate to achieve body phenylalanine balance. In comparison with published data, the relative rate of phenylalanine hydroxylation appears to be high. We speculate that tyrosine is a conditionally indispensable amino acid under these conditions; it would be desirable to establish the intake levels and ratio of phenylalanine to tyrosine that effectively support aromatic amino acid balance in these critically ill patients.

ABSTRACT. To better understand the impact of severe illness on the amino acid economy and nutritional needs of pediatric patients, we studied plasma phenylalanine and tyrosine kinetics in eleven critically ill patients (six fullterm newborns and five young infants). Within 4 8 h of the diagnosis of sepsis they were given primed constant i.v. infusions of L-11-'"]phenylalanine and 1,-13,3,~II~]tyrosine for 4 h. Routine nutritional support continued during this period by parenteral administration of destrose, lipid emulsion, and an amino acid mixture low in tyrosine. Phenylalanine and tyrosine fluses and rate of phenylalanine hydrosylation did not differ significantly between the two age groups, and so the data were combined for evaluation. For the entire group, values (pmol. kg-'. h-'; mean 2 SD) for phenylalanine and tyrosine fluses and rate of phenylalanine hydrosylation were 132 5 24, 66 2 16, and 29 2 12, respectively. Plasma phenylalanine to tyrosine concentration ratio was 1.67 +. 0.6. From a comparison of the rate of phenylalanine hydroxylation with measured phenylalanine intakes, it was concluded that their routine, clinical nutritional support was inadequate to achieve body phenylalanine balance. In comparison with published data, the relative rate of phenylalanine hydrosylation appears to be high. \Ye speculate that tyrosine is a conditionally indispensable amino acid under these conditions; it would be desirable to establish the intake levels and ratio of phenylalanine to tyrosine that effectively support aromatic amino acid balance in these critically ill patients. (Pediatr Res 35: 580-588, 1994) Adequate nutritional support ofcritically ill patients represents an essential component of medical care (1, 2). with well-nourished patients having better outcome and a shorter hospital stay (3, 4) . Although current nutritional regimens provided to critically ill children are based largely on the nutritional requirements for healthy children (5, 6) , these patients are often hypermetabolic and show profound changes in the pattern of energy substrate and amino acid use. It is likely that more effective nutritional therapies will evolve after a better understanding of the quantitative changes in nutrient metabolism, and the mechanisms responsible for them, under these disease conditions.

In this study we examined plasma kinetics of phenylalanine and tyrosine metabolism in critically ill neonates and infants with sepsis. These amino acids are substrates for protein synthesis and precursors of physiologically important metabolites, such as catecholamines and thyroid hormones. We hypothesized that the standard parenteral nutritional support given during the acute stage of critical illness to our patients might be insufficient to meet the aromatic amino acid requirement of these patients and that tyrosine, usually considered to be a nutritionally dispensable amino acid, becomes "conditionally" indispensable during hypermetabolic states. Therefore, an adequate exogenous supply of tyrosine would be necessary to support fully and effectively the nutritional needs of the patient. T o explore these hypotheses, we investigated plasma phenylalanine and tyrosine fluxes in 1 1 critically ill newborns and young infants with ["Clphenylalanine, [13C]tyrosine. and ['H2]tyrosine as labeled tracers. Our findings are compared with published data derived in healthy newborns using similar tracer protocols (7).

Puticnrs. Eleven critically ill patients (six full-term newborns and five young infants) were studied. Their clinical characteristics are described in Table 1 . All the newborns had appropriate weight for gestational age. The body weights of the five infants were below the 5th percentile for age (8) . All patients had a diagnosis of sepsis and were admitted to the Neonatal and Pediatric Intensive Care Units at the Massachusetts General Hospital. Patients were studied within 48 h of the diagnosis of sepsis once hemodynamic stability was achieved. Sepsis in newborns was defined by the presence of two or more of the following: history of maternal fever or chorioamnionitis, rupture of membranes greater than 18 h, signs of respiratory distress, temperature instability, changes in level of consciousness or lethargy, shock, white blood cell count greater than 20 000 cells/mm3 or less than 5000 cells/mm3, a left shift with more than 10% band forms or greater than 60% polymorphonuclear forms, or a positive blood culture or evidence of clinical sepsis syndrome (9) . Sepsis in infants was defined as bacteremic if there was a positive blood culture, or it was defined as clinical sepsis syndrome if three or more of the following were found: clinical evidence of infection (pneumonia, meningitis, for example), fever or hypothermia, tachypnea, tachycardia, or symptoms of decreased organ perfusion. such as oliguria, lethargy, or increased lactate level (lo).

In the newborn group ofsix patients, three of them had positive blood cultures at diagnosis. Patient 2 had a positive blood culture 0 PIIENYLALANINE METABOLISPI.1 IN SEPSIS 58 1 Viral and bacterial cultures remained ncgative, and a workup for metabolic diseases was also ncgative. Patients 1 and 6 had negative blood cultures; however, both received antibiotics at the refemng hospital before blood cultures were obtained, and the mother of patient 1 was treated with antibiotics for chorioamnionitis. These three patients (paticnts 1, 5, and 6), met the criteria for clinical sepsis syndrome. All patients had hemodynamic instability and required inotropic support with one (dopamine in patients 1, 3, and 5). two (dopamine and dobutamine in patients 2 and 6), or three (dopamine, dobutamine, and epinephrine in patient 4) inotropic agents. All paticnts required multiple fluid bolus for resuscitation. Disseminated intravascular coagulation was present in four paticnts (patients 2, 3, 4, and 6). All patients had variable degrees of metabolic acidosis and respiratory distress. Multiorgan failure (renal. hepatic. CNS, and respiratory) occurred in patient 4. Leukopenia was present in patients 4 and 5.

In the pediatric group, patients 7 and 10 had pneumonia and adult respiratory distress syndrome (patient 7 had a positivc blood culture to Strc~ptococc~u pnnunoniac; patient 10 had a negative bacterial blood culture at the time of the investigation and severe respiratory failure). Patient 8 had Pt~~~tr~~oq:sti.s caritiii pneumonia, diagnosed by lung biopsy, and a positive blood culture to E. coli. Her HIV status was negative at the time of her discharge from the intensive care unit, although her mother tested HIV positive. Patient 9 had Stevens-Johnson syndrome resulting from bactrim treatment for otitis media. Adult rcspiratory distress syndrome and multiorgan failure developed in this patient. but a blood culture was negative at the time of the study. Patient I I had II. it~/Iircnzrrc sepsis and severe bronchopulmonary dysplasia. Patients 9 and 10 met criteria for clinical sepsis syndrome.

All infant patients had variable degrees of multiorgan involvement: their lungs were mechanically ventilated, and they required inotropic support and had arterial and central venous lines in place. The newborn patients rcquired a servocontrol radiant warmer to maintain skin temperature at 36.5"C. Infants were studied in a room with ambient temperature of about 25°C and relative humidity of 40%.

Physiologic Stability Index ( I I), Neonatal Physiologic Stability Index (12) , and Therapeutic Intervention Score System (13) scores were obtained in each patient. These three systems objectively assess severity of disease and predict outcome (chances of death or survival). The Physiologic Stability Index and Neonatal Physiologic Stability Index estimate the degree of derangement in seven major physiologic systems (cardiovascular, respiratory, neurologic, hematologic, renal, gastrointestinal, and metabolic). After the degree of abnormality is assessed, each variable is assigned a scorc of 1. 3, or 5, indicating the clinical importance of the derangement. A scorc of 3 in a given variable mandates a change in therapy; a score of 5 indicates a life-threatening situation. The Therapeutic Intervcntion Score System evaluates the degree of intervention that a patient requires. Therapeutic interventions are classified from 1 to 4. Scorc I indicates routine intensive care interventions (ECG monitoring, supplemental oxygen, i.v. antibiotics, for example): scorc 2 indicates more invasive therapeutic or monitoring interventions, such as central venous pressure line and more than two i.v. lines or hemodialysis: score 3 indicates invasive therapeutic or monitoring interventions that may be lifesaving, such as endotracheal intubation. Scorc 4 indicates interventions in response to major events. such as cardiopulmonary resuscitation, pulmonary artery catheter placement. and cardiac pacing. These scores are routinely used in intensive care units for assessing severity of illness, with higher scores indicating a greater severity of the disease. Studj~protocol. The study protocol was approved by the Subcommittee on Human Studies at the Massachusetts General Hospital. Written informed consent was obtained from the patient's parents or guardians. We used the arterial line as the sampling source and the central venous line for tracer infusion. Isotope tracers were infused by means of a calibrated syringe pump (Hanard Apparatus, Natick, MA). Verification of isotopic purity was obtained before the infusion study. Both tracer solutions (see below) were infused at a rate of 4.7 mL. h-I, except for infants 9 and 10, who received an infusion rate of 10.74 mL. h-' because oftheir higher body weight. All patients received priming tracer doses of 5. Blood samples for determinations of plasma phenylalanine and tyrosine isotopic enrichment and free plasma amino acid concentrations were obtained at 0, 180, 190, 200, 2 10, and 240 min in all patients. The samples were placed in tubes containing heparin and centrifuged immediately, and the plasma was stored at -80°C until used for analysis. The total amount of blood withdrawn during the entire study period in each patient was less than 5 mL.

Nlitritiot~al slipport. Nutritional support could not be manipulated experimentally in this investigation and so was provided as indicated by the routine clinical care of these critically ill infants. Hence, the levels of energy and protein intake varied between subjects. Mean energy intakes for neonates and infants were 35 and 40 kcal. kg-l.d-' respectively, provided principally as a 10-20% dextrose solution and a soybean lipid emulsion (Intralipid 10 or 20%, KabiVitrum Inc., Franklin, OH). The protein needs ofeach patient were supplied by a 1.5-2.5% amino acid mixture of Novamine (Novamine, 11.4%; Clintec, Deerfield, IL). The amino acid composition of the mixture is shown in Table 2 . None of the patients received nutrients by the oral route. The duration of nutritional support before the infusion study was variable, ranging in the neonatal group from 0 to 32 h and in the infant group for 0 to 18 h. From the time of admission, some neonates and infants were never parenterally or enterally fed before the infusion study.

~I t~u l j~t i c a l mctllocls. Amino acid concentrations in plasma and in the infusates of tracers were analyzed by means of a Hewlett-Packard Amino Quant (Palo Alto, CA), which involved precolumn derivatization with orthophthaldehyde and quantitation by fluorescence.

For isotopic analysis of phenylalanine and tyrosine, 200 pL plasma was used. T o this 1 m L 2 N acetic acid was added, and then it was placed on an ion exchange column (BioRad AG50-X2 resin, Richmond, CA) (14) . The fraction containing the amino acids was eluted with ammonium hydroxide and then evaporated to dryness. A tertiary butyldimethylsilyl derivative of the amino acids was prepared in a one-step procedure by heating the residue for 1 h at 60°C with 50 pL of MTBSTFA [/I-methyl-N-(tcrtbutyldimethylsilyl)trifluoroacetamide] (Pierce, Rockford, IL) and 50 pL of acetonitrile (15) . Two to 3 pL of mixture were sufficient for gas chromatography/mass spectrometry analysis.

All measurements of plasma amino acid isotope dilutions were carried out on a Hewlett-Packard 5890 Series I1 gas chromatograph coupled to a Hewlett-Packard 5988A quadrapole mass spectrometer and a Hewlett-Packard RTE-A data system. Electron impact ionization was camed out at 120 eV.

Gas chromatography/mass spectrometry analysis of phenylalanine and tyrosine was camed out in one run on a 30 m x 0.25 m m DB 130 1 fused silica column (J and W Scientific, Folsom, CA) that was temperature programmed from 200°C to 300°C at lS°C/min. Phenylalanine and tyrosine eluted at 5.6 and 7.8 min, respectively. Selective ion monitoring of both amino acids was camed out with the M-57 fragment ion that corresponded to loss of the 1-butyl group (15) . Selective ion monitoring was carried out at m/z 336 and m/z 337 for natural and [I3C] phenylalanine, respectively. Ions at m/z 466, m/z 467, and m/z 468 were monitored for natural, ["C]. and [?Hz]tyrosine, respectively.

Because the ion clusters of the tyrosine tracee and the [I-I3C] and ['H?] tracers overlap, estimation of sample enrichments was based on multiple linear regression, where the various mass spectrometer signals were used as dependent variables and the sample enrichments served as unknown parameters. The coeflicients of the regression equations were derived from standard mixtures ranging from 0 to 8% mole fraction for ['3C]phenylalanine and [?Hz]tyrosine and 0 to 6% mole fraction for ["C] tyrosine. A further set of five standard mixtures was also prepared containing both ["C] and ['H2]tyrosine. Each sample was measured in duplicate and multiple samples measured during the isotopic plateau. Additional details concerning the determination of plasma isotope enrichments of phenylalanine and tyrosine are given in the appendix.

Enantiomeric purity of the tracers was verified by gas chromatographic analysis on a Chirasil-Val capillary column (Chrom Pack, Middelburg, The Netherlands) as the pentafluoropropionyl propyl esters (1 6).

Calc~rlatiot~s. The model of phenylalanine-tyrosine metabolism uscd here follows that described by Clarke and Bier (17) with modifications proposed by Thompson ef a/. (18) . Briefly, the amino acid (phenylalanine or tyrosine) flux (Qp or QT, respectively; pmol. kg-'. h-I) is calculated by isotope dilution, as previously described for leucine (19) :

where i is the rate of tracer infusion (pmol. kg-'. h-') and Ei and Ep are the isotopic enrichments ofthe infusate and plasma amino acid, respectively.

The rate of phenylalanine conversion to tyrosine (QPdT; pmol. kg-'. h-') (phenylalanine hydroxylation) was derived as follows: Primary and derived values were tested for significance between infant and newborn groups using analysis of covariance with age, weight. and severity of disease as variables. Adjusted means were tested for significance with a least-squares mean procedure. Dependent variables included phenylalanine flux (QI.), tyrosine flux (QI), phenylalanine hydroxylation (Qp,T), and protein turnover (Prots. Prot,, and ProtBAL). Because no statistically significant differences werc observed bctween newborns and infants, the results for both groups were combined for purposes ofcomparing these values with those reported in the literature (7). A p < 0.05 was chosen as being statistically significant. Relationships among variables werc tested for significance using Pearson's product moment correlation. The relationship between phenylalanine and tyrosine intake with phenylalanine and protcin balance was evaluated by least-squares linear regression analysis. An SAS program (Cary, NC) was used for the statistical analysis (20) .

Energy and protein intakes for the patients during the period of study are shown in Table 3 . The levels of intake were lower than those recommended for healthy newborns and infants (2 1). which generally reflects our patients' critical clinical status. Critically ill patients often require infusion of multiple drugs, blood products, and fluids at the expense of their nutritional support. Furthermore. fat administration in the presence of respiratory failure or sepsis is often restricted by attending physicians because of concerns about possible lipid-induced impairment of respiratory and immune function (22) (23) (24) .

The Plasma-free phenylalanine and tyrosine concentrations during the tracer study period are shown in Table 4 . The phenylalanine/tyrosine ratio in these patients with sepsis was increased (1.67) when compared with values rcported for hcalthy newborns (0.7 1) (25) and adults (0.82) (26, 27) .

A steady state level of isotopic enrichment in plasma of our patients was achieved during the 4-h continuous tracer infusion period as determined by lack of a statistically significant slope in the plasma pools of phenylalanine and tyrosine isotopologs (Fig.  1) . For the entire group of patients, the mean coefficients of variation for plasma isotope enrichment plateaus were 7, 5, and 11% for [13C]phenylalanine, ['Hz]tyrosine, and ["Cltyrosine, respectively. Their kinetics and the rates of whole body protcin turnover were calculated from these isotopic enrichments of plasma phenylalanine and tyrosine.

Phenylalanine and tyrosine fluxes and rates of phenylalanine hydroxylation were 137 + 17, 63 +. 13, and 34 2 6 pmol. kg-'.

h-' in the neonates and 126 f 29. 70 + 18, and 24 + 14 pmol .

kg-'. h-I in the infants (Table 5 ). Considerable variation existed within each group, and differences between group means were not statistically significant ( p > 0.1). Therefore, for our purposes.

we considered it appropriate to combine the results into one data set for the further comparisons given below. The combined group mean phenylalanine flux was 132 pmol. kg-'.h-', with a mean phenylalanine hydroxylation rate of 29 pmol. kg-'. h-', representing 22% of the phenylalanine flux. The components of phenylalanine flux are presented in Table  6 . Tyrosine intakes were very low (mean, 0.6 pmol.kg-'.h-'). Intakes of phenylalanine ranged widely, with the mean being 17.4 pmol . kg-'. h-'. Extrapolated to a daily basis, this amounts to a mean intake of 69 mg. kg-' .d-', which is in the requirement range for healthy infants when tyrosine intakes are generous (28) . but it presumably would be less than adequate (21) for tyrosine intakes that are low, as in the present study. Also, because the mean phenylalanine hydroxylation rate was 29 pmol. kg-'. h-' (equivalent to 1 15 mg. kg-'.d-') ( Table 5) . the group of patients as a whole were in negative body phenylalanine balance. In Table  6 , we summarized the rates of disappearance of phenylalanine from plasma by protein synthesis (Phes) and entry into plasma by protcin degradation (Phen). Again. these rates reflect the extent of the negative body phenylalanine balance, which was a mean of -14 pmol. kg-'. h-' for the combined groups.

The derived values for body protein turnover are presented in Table 7 . Mean protein synthesis and breakdown rates were 8.6 and 9.8 1 g protein. kg-' .d-', respectively, for the combined two study populations.

The following statistical correlations were observed in these children with sepsis: body protein balance was negatively correlated (R' = 0.5; p < 0.01) with the rate of phenylalanine hydroxylation to tyrosine and positively correlated ( R 2 = 0.4, p < 0.04) with phenylalanine intake (Fig. 2) .

We have investigated the kinetics of whole body phenylalanine and tyrosine metabolism in critically ill children with sepsis. Despite the importance of phenylalanine as an indispensable amino acid and as precursor for tyrosine, relatively few studies address phenylalanine and tyrosine kinetics in relationship to the metabolic needs and nutritional status of pediatric patients (29) (30) (31) . T o our knowledge, no similar studies appear to have been conducted previously in critically ill children with sepsis. Because we were not able to investigate a healthy cohort, our combined data for phenylalanine and tyrosine kinetics, as well as the estimates obtained for whole body protcin turnover in critically ill young infants with sepsis are compared here with those of Dcnne cJt (it. (7) , who havc rcported similar data for seven normal, term infants studied at approximately 2 1 d of age. In their study, phenylalanine and tyrosine fluxes and rate of phenylalanine hydroxylation were 76 + 12, 53 +-12. and 12 f 3, respectively, 584 CASTILLO 6 T :lL. Mean + S D Phcnylalaninc/tyrosine ratio Combined data Phenylalaninc/tyrosinc ratin cr (11. (7) . Some difficulty exists, however, in drawing a strict comparison between the present hydroxylation data and those of Denne 01 (11. (7) because pentadeuterated phenylalanine was used as tracer by these latter investigators; we have shown rccently in healthy adults that this tracer underestimates hydroxylation rates it1 I~V O by about a factor of 2.0 (37) . For this reason, the estimate of 10-1 2 pmol. kg-'. h-' for phenylalanine hydroxylation in healthy newborns (7) might be lower than the actual rates, although that is speculative and a matter for further study. The values for Prots and Prot,, in our patients with sepsis are somewhat higher than the 7 and 6 g.kg-'.d-', respectively, reported for healthy infants (25, 33) , although perhaps they are not as high as might be anticipated for hypermetabolic or acute catabolic states (34) .

Despite these difficulties, in fully comparing the present values with those appearing in the published literature, the present data point to changes in kinetics and in the relationship between phenylalanine and tyrosine metabolism in infants with sepsis versus normal infants. For example, phenylalanine fluxes in our critically ill infants are considerably higher than those reported for normal babies (7) . Furthermore, the flux ratio of phenylalanine to tyrosine was higher in our infants than in the normally growing newborns studied by Denne cf 01. (7) . This finding is consistent with previous reports (35) showing an increased plasma phenylalanine/tyrosine concentration ratio in patients for the fasted state and 86 + 13, 46 + 8, and 10 + 3 (mean + SD; pmol. kg-'.h-I), respectively, for the fed state. These invcstigators used [?H5]phenylalanine and ['H2]tyrosine as tracers. Thus, phenylalanine fluxes are higher in our patients with sepsis when compared with the healthy newborns, as might be anticipated for a hypermetabolic, catabolic condition. When the intravenously administered phcnylalanine intake is subtracted from phenylalanine flux, the "corrected," mean phenylalanine flux is 115 f 24 pmol. kg-'. h-', remaining higher than that in healthy newborns (7) . This correction for i.v. intake of phenylalanine is necessary, particularly because Wykes ct a/. (31) found that phenylalanine flux was greatly affected by the parentcral phcnylalanine supplied. Tyrosine fluxes in our group of critically ill infants with sepsis were similar or perhaps somewhat higher than those in healthy neonates (7), who were also receiving no tyrosine by i.v. feeding. Also, the hydroxylation of phenylalanine, expressed either as an absolute rate or as a percentage of the phenylalanine flux, was higher in our patients than was phenylalanine hydroxylation in newborn infants as reported by Denne with sepsis, suggesting a n increased rate of muscle breakdown, reduced rate of incorporation of phenylalanine into muscle proteins, and/or a n alteration in the rate of hepatic conversion of phenylalanine to tyrosine (26, 27, 35, 36) . Also, during protein deficiency, phenylalanine use decreases, leading to a relative increase in the plasma phenylalanine level, which again is possibly due to a decreased enzymatic conversion of phcnylalanine to tyrosine (37, 38) . Of course, the low content of tyrosine in the parenteral amino acid mixture would also be expected to raise the plasma-free phenylalanine/tyrosine ratio, although the ratio was also high in the two patients (patients 4 and 9) who were receiving glucose alone. This finding suggests a n altered aromatic amino acid homeostasis in our patients and possibly a change in the coupling between phenylalanine kinetics, hydroxylation, and tyrosine balance and oxidation. We speculate that because of the stressful or catabolic state combined with the low tyrosine intake, an impaired balance exists among phenylalanine retention, hydroxylation, and tyrosine homeostasis in these patients with sepsis. Thus, phenylalanine hydyroxylation rates may be adaptively increased in these patients but with the net effects of impairing phenylalanine balance and inadequately meeting the metabolic needs for tyrosine. It is not unusual for critically ill patients with sepsis to develop supranormal values of physiologic parameters such as cardiac output or oxygen delivery that are still insufficient to meet their needs (39) . We speculate that a similar situation may occur at the metabolic level, whereby multiple factors, such as increased protein breakdown and changes in the pattern of amino acid utilization and requirements, lead to increased rates of phenylalanine hydroxylation that are still insufficient to meet the tyrosine requirements. This hypothesis will need to be verified in follow-up investigations in which different intakes and ratios of phenylalanine and tyrosine intake are examined for their effects on phenylalanine and tyrosine kinetics. Alternatively, the altered rates of phenylalanine hydroxylation and their relationships to phenylalanine kinetics and intake might be a more general reflection of an impairment and changes in amino acid catabolism, which frequently occurs in sick and stressed patients (40, 4 1).

The clinically determined standard amount of total parenteral nutrition formulation routinely used in our critically ill patients appears to be inadequate to support their aromatic amino acid needs. From the data shown in Table 6 , intakes in the range of about 30-40 pmol . kg-'. h-' permit neutral body phenylalanine balances, suggesting that this intake level (equivalent to about 138 mg phenylalanine. kg-' . d-') is probably an appropriate target for critically ill patients during the acute stage of their disease and in the virtual absence of a tyrosine intake.

Critically ill newborns and infants with sepsis who receive limited parenteral protein and energy intakes, as dictated by routine clinical practice (34, 42) , are unable to maintain phenylalanine homeostasis, and because of the link between phenylalanine and tyrosine metabolism, it is to be expected that a deterioration in tyrosine homeostasis will occur. Indeed, tyrosine may become a "conditionally" indispensable amino acid under these circumstances despite an apparently generous supply of phenylalanine. The reason for this is that the formation of tyrosine from phenylalanine occurs principally in the liver, where this tyrosine may or may not be made as easily available in times of metabolic stress as an exogenous source of tyrosine would be for meeting the tyrosine needs of the body organs and tissues. Because of its poor solubility, most parenteral amino acid mixtures contain only a trace of tyrosine, while being designed to provide a generous phenylalanine intake; the assumption here is that the supply of phenylalanine will be sufficient to maintain an adequate availability of tyrosine where it is required. This assumption may be, and probably is, an overly simplistic assessment of phenylalanine-tyrosine interrelationships, particularly if the regulation of tissue and organ pools of phenylalanine and tyrosine occurs by way of separate systems and mechanisms and if these are, in turn, affected differently by stressful stimuli, including infection. Although another commonly used mixture (Trophamine, Kendall-McGaw Laboratories, Irvine, CA), has been supplemented with acetyl-tyrosine, the target intake level has been designed to support plasma amino acids concentrations similar to those observed in healthy full-term, breast-fed, l-moold infants (43, 44) . Although a possibly useful criterion, it is important to assess whether "normal" serum values represent an appropriate nutritional goal for the critically ill patients (45, 46) . Current estimates of amino acid requirements in human subjects are based largely on nitrogen balance studies (21) ; in the case of children, these values are supplemented with data from measurements of growth rates in response to specific nutritional treatments (21, 47) . In the case of critically ill neonates and infants, these procedures are not feasible, and we suggest, from the present experience, that application of stable isotope tracer methods might be a useful way to investigate the quantitative requirements for specific amino acids in these patients. Indeed, "C-enriched amino acid tracers have been used successfully in adults to estimate minimum intakes required to achieve body balance of the irreversible oxidative losses of nutritionally indispensable amino acids (48) . However, we recognize that for critically ill children with mechanically ventilated lungs it might be difficult to obtain accurate measurements of total l'C-labeled C02 and COz production by indirect calorimetry, especially under conditions of the high air flow rates used to meet their ventilatory needs. Measurements of "C02 in plasma is possible, but precise estimation of rates of total C 0 2 production in these patients remains difficult. For these reasons, we chose to apply the present phenylalanine-tyrosine tracer model as described by Thompson c7t al. (18) , except that we used [I-"C]phenylalanine instead of ['Hs-ringlphenylalanine as tracer, thereby obviating the need for breath collections and measurement of respiratory gas exchange rates.

In summary, we describe here, apparently for the first time. phenylalanine and tyrosine kinetics in critically ill pediatric patients with sepsis. From these observations we conclude that current clinical practice of total parenteral administration during critical illness is insufficient to meet the aromatic amino acid requirements of these patients. Under these circumstances, we also speculate that tyrosine may be a "conditionally" indispensable amino acid. Hence, consideration should now be given to defining adequate intakes of the aromatic acids and, in particular. the ratio between phenylalanine and tyrosine. We fully appreciate that the present study is limited by sample size, the variable clinical condition of our patients, lack of a concurrent control group of healthy infants, and the different, clinically determined intakes of macronutrients among the infants. Although these difficulties are commonly faced by investigators who are interested in canying out precise metabolic studies in critically ill infants, we considered it important to begin to establish in quantitative terms whether the nutritional component of these infants' care was appropriate to the metabolic changes that accompany their disease state. Our findings indicate the desirability of now investigating the impact of the level of aromatic amino acid intake, including the exogenous supply of tyrosine possibly given in dipeptide form, on phenylalanine and tyrosine kinetics and balance and their functional significance in critically ill children with sepsis.

A.cknoll'/edgments. The authors thank Dr. I. David Todrcs and the nursing and respiratory care staff of the Pediatric and Nconatal Intensive Care Units for assistance in carrying out these demanding studies. We are indebted to Dr. Ronald Kleinman for his helpful comments and to Alice Traudt for her secretarial assistance.

Cerra FB I n6 Nutrition in the critically ill: modern metabolic support in the intensive care unit

Elnogi Dll Kinney JM 1980 Influence of total parenteral nutrition of fuel utilization in injury and sepsis

Holbrook PR 19R2 Malnutrition in critically ill infants and children

Hoolbrook PR 1981 Early nutritional depletion in critically ill children

Winters RW 1972 Intravenous alimentation in pediatric patients

The amino acid requirements of children. In: Nyhan WL (ed) Amino Acid Metabolism and Genetic Variation

Liechty EA 1993 Intravenous substrate increases bot h protein synthesis and suppresses proteolysis in normal growing newborns

Bacterial and viral infections of the newborn

Sepsis, the sepsis syndrome, multi-organ failure: a plea for comparable definitions

Accurate prediction of the outcome of pediatric intensive care: a new quantitative method

Validation of two scoring systems which assess the degree of physiologic instability in critically ill newborn infants

Therapeutic interventio n scoring system

Determinati on of amino acids profiles in sampks by gas chromatography

Capillary gas chromatography of amino acids, including asparagine and glutamine sensitive gas ch romatographic-mass spectrometric and selected ion monitoring chromatographicmass spectrometric detection of the NO(S)-tertiary-butyldimethylsil yl derivatives

Amino acid determ ination by capillary gas chromatography on chirasi l-val enantiomer labeling and nitrogen-selective detection

The conversion of phenylalanine to ty rosi ne in man. Direct measurement of continuous intravenous tracer infusions of L-[ring-'H,(phenylalanine and L-[ 1-"C(tyrosine in the post absorptive state

Rapid measurement of whole body and forearm protein turnover using a (' II,]phenylalanine model

Bier DM 1980 Measurement of leucine metabolism in man from a primed. constant infusion of L-[ l-13 C]leucine

Guide for Personal Computers

Energy and protein requirements

Effects of intravenous fat emulsion on respiratory failure

Andersson S 1991 Generation of free radicals in lipid emulsion used in parenteral nutrition

The effect of intralipid on mononuclear and polymorphonuclear phagocytes

Leucine kinetics during feeding in normal newborns

Beisel WR 1976 The significance and mechanism of an increased serum phenylalanine-tyrosine ratio during infection

Pru it BA 1978 Abnormalities of phenylalanine and tyrosine kinetics

Protein requirements

Significant phenylalanine hydroxylation in l'il'o in patients with classical phenylketonuria

Acute illness in maple syrup urine di sease: dynamics of protein metabolism and implications for management

Pcncharl PB I 9Y2 Glycine. ieucine and phenylalanine nux in low birth weight infants during parenteral and enteral feeding

Phenylalanine conversion to tyrosine: comparative determinat ion with L-[ring 'H,(phenylalanine and L-[1-''C(phenylalanine as tracers in man

Protein and amino acid requirements in hum ans: metabolic basis and current recommendations

Coran AG 1991 I I armona I and metabolic response to operative stress in the neonate

Signilicancc of the changes in plasma amino-acid levels in meningococcal infection

Plasma amino acids as predictors of the severity and outcome

Amino acid metabolism in kwashiorkor. Metabolism of phenylalanine and tyrosi ne

Morley RM 1993 Hyperphenylalaninaemia and outcome in intravenously fed preterm neonates

Appel PL 1990 Therapy of shock based on pathophysiology, monitoring, and outcome prediction

Protein and amino acid metabolism after injury

Tracer studies of amino acid kinetics: a basis for improving nutritional therapy

The dTects of surgery on the nitrogen metabolism of parenterally fed human neonates

Protein and amino acid requirements

In8 Pediatric parenteral amino acid mixture in low birth weight infants

Lee TS Prospective trial of supra normal values of survivors as therapeutic goals in high risk surgical patients

Cellular oxygen metabolism during sepsis and shock: the relationship of oxygen consumption to oxygen delivery

Methods in infants nutrition research: balance and growth studies

Pellett PL 1989 A theoretical hasis for incr,·asing current estimates of the amino acid requirements in adult man. with experimental support

Calculations regarding { 13 Cj and f enrichmcllls. The isotopomcrs of tyrosine seen arising at m/z 466(M-57r in the electron impact mass spectrum of the tertiary butyldimethylsilyl derivative at the m + I (m/z 467) and m + 2 (m/z 468) ch;mnels have abundances of20% and 44%. respectively, relative to the main ion (m) at m/z 466. The isotopomcrs of the tracee therefore overlap with those of the less abundant [PC) and tyrosine tracers present in the labeled plasma samples.If only one tracer was present in the samples to be analyzed. then standards containing mixtures of natural and [uC] or [ 1 H1] tyrosine could be measured and a calibration graph constructed. The regression coefficient (slope) andy intercept (constant) could then be calculated by single linear regression. These could be used to solve for the unknown enrichments in plasma. However, in these studies two tracers are present, and their isotopomers overlap. The 13 C-tracer that is found at m/z 467 has isotopomers arising from it that contribute 44% of its signal to the channel at m/z 468 where the is predominantly found. Also, the ( 1 H1)tyrosine tracer has isotopomers arising from it at m/z 468 that arc I 0% of its signal and contribute to the channel at m/z 467 where the JJC-tracer is predominantly found. If the contributions of these isotopomcrs are not taken into account, then an overestimate of both enrichments will be made, resulting in an underestimate of tyrosine nux and an overestimate of phenylalanine hydroxylation. For instance, in these studies the corrected mean plasma [JJC) and ( 2 H 1 ]tyrosine enrichments were about 1.5 and 5%, respectively. However, because of the 44% spill-up of the (