ABSTRACT. Background: Whether the contribution of nonurinary nitrogen excretion (N^sub 2^nu) to total nitrogen excretion (N^sub 2^tot) is clinically relevant has not been tested in children in an intensive care unit. Particularly after digestive tract surgery, fecal nitrogen losses, and losses via nasogastric tubes, enterostomies and wound drains may be large. Methods: We prospectively measured urinary nitrogen excretion (N^sub 2^u) and N^sub 2^nu 4 to 6 days after digestive tract surgery in 78 newborns and infants who were given total parenteral nutrition. Results: Two hundred seven collections of excreta, each representing one 24-hour period, were obtained. Median N^sub 2^nu was 15 mg/kg/24 hours (range, 0.4-153), median N^sub 2^u 153 mg/kg/24 hours (range, 57-558), median N^sub 2^tot 179 mg/kg/24 hours (range, 72-577), and the median ratio of N^sub 2^nu and N^sub 2^u 9.9% (range, 0.2-110). The observed variations could not be attributed to differences in the severity of the underlying disease or the surgical stress. The mean difference between N^sub 2^tot and N^sub 2^u was 21 mg/kg/24 hours (95% prediction interval -20 to +63). Use of a linear regression equation that predicts N^sub 2^tot according to N^sub 2^u and the weights of other excreta eliminated bias and improved precision (95% prediction interval -34 to +34 mg/kg/24 hours). For individual measurements, however, considerable imprecision remained. Conclusions: In newborns and infants, receiving parenteral nutrition 4 to 6 days after digestive tract surgery, N^sub 2^nu is variable and not to be neglected. The only way to accurately assess N^sub 2^tot in individual patients is by measuring the nitrogen content of all excreta. (Journal of Parenteral and Enteral Nutrition 27:327-332, 2003)
Catabolism and anabolism are fundamental aspects of illness and reconvalescence, often defined as whole-body nitrogen loss and whole-body nitrogen retention or as a negative nitrogen balance and a positive nitrogen balance.1,2 Nitrogen balance studies therefore are one of the cornerstones of clinical nutrition research. When assessing whole-body nitrogen excretion, urinary nitrogen excretion (N^sub 2^u) usually is measured or derived from urinary urea excretion. Nitrogen excretion via other routes (skin, stools, wounds, drains) is often assumed to be of minor importance or predictable.3-6 In children admitted to an intensive care unit (ICU), the value of urinary urea excretion as a predictor of N^sub 2^u has been questioned.6-9 Whether nonurinary nitrogen excretion (N^sub 2^nu) is of little importance when compared with N^sub 2^u has not been tested in such children. N^sub 2^nu may be particularly large after digestive tract surgery, when body fluids may be lost through nasogastric tubes, enterostomies, wound drains, and feces. We measured N^sub 2^nu and N^sub 2^u in newborns and infants in our pediatrie surgical ICU who were receiving parenteral nutrition after digestive tract surgery and assessed the contribution of N^sub 2^nu to total nitrogen excretion (N^sub 2^tot).
MATERIALS AND METHODS
Our pediatric surgical ICU is part of a tertiary referral university children's hospital serving approximately 4.5 million inhabitants. At the time of the study, the ICU consisted of 2 essentially identical 7-bed units. Annually, 600 to 700 patients were admitted for perioperative intensive and high care. The majority were newborns and infants admitted because of a congenital anomaly that required surgery. The second most frequent reason for admission was postoperative intensive care in all pediatric age groups.
From January 1997 through December 1999, 80 newborns, infants, and children, admitted to our ICU after digestive tract surgery, entered a prospective trial that studied the effects of isonitrogenous enrichment of parenteral nutrition with glutamine. As part of this trial, nitrogen excretion was measured on days 4, 5, and 6 after surgery (day 0).
The trial was approved by the institutional review board. Inclusion criteria were gestational age >30 weeks and age or =35 weeks and age
Total parenteral nutrition was started on day 2 after surgery. Depending on age and weight, it provided 85 to 100 kcal/kg/24 hours and 1.5 to 2.5 g/kg/24 hours of amino acids (Vaminolact; Pharmacia, Sweden). Carbohydrates (dextrose; Fresenius, Germany) provided approximately 65% of nonprotein calories and fat (Intralipid 20%; Pharmacia), approximately 35%. If the patient's condition and the surgical procedure would allow it, tapering of parenteral nutrition and reintroduction of enteral nutrition was started halfway through day 6.
Excreta Collection and Laboratory Analysis
All excreta, with the exception of bronchopulmonary secretions and 1 patient's decubitus wound exudate, were collected on days 4, 5, and 6 after surgery. For the collection of urine, a transurethral catheter was used, if present. For girls, if such a catheter was not being used already, it was inserted; for boys, appropriately sized urine bags were then used. Urine collection bags were emptied, and the collected volume was measured at least once every 4 hours. Leakage of urine that could not be quantified rendered that interval's collection invalid (cf. Appendix). Collected urine was stored in a refrigerator in containers preacidified with sulfuric acid until the morning of the next day, when it was transferred to the laboratory. Fluids produced through nasogastric or nasoduodenal tube drainage, feces, and enterostomy and wound drain output were also collected and stored in a refrigerator until transferal to the laboratory. Collection of these excreta was considered invalid if leakage was observed.
In the laboratory, all collections were weighed and homogenized. Urine collections were further acidified with sulfuric acid to a pH of 2 to 4. An aliquot of every collection was then stored at -20[degrees]C until thawing and analysis. The total nitrogen content was determined by a continuous flow elemental analyzer (Carlo Erba NC-1500; Interscience BV, The Netherlands). In brief, triplicate samples are weighed in tin sample containers, freeze-dried, and combusted at 1020[degrees]C. Copper reduces the formed nitrogen oxides to elemental nitrogen gas. The nitrogen gas flows through a thermal conductivity detector that generates an electrical signal proportional to the concentration of nitrogen. This is an automation of the Dumas combustion method.12
Because we intended to analyze the contribution of N^sub 2^nu and N^sub 2^u to N^sub 2^tot, the unit of analysis was 24-hour nitrogen excretion, with one to three 24-hour intervals being available per patient. Nitrogen excretion was expressed as milligrams nitrogen per kilogram body weight per 24 hours. Values were reported as means + or - 1 SD or as medians with ranges. Values were reported only if the collections of all excreta were considered representative of the 24-hour interval in question and if other excreta than urine were produced. For the collected urine to be considered representative of the 24 hours in question, the total collection time had to equal at least 8 hours (cf. Appendix).13 Collections of other excreta had to comprise all 24 hours to be considered representative.
We analyzed the relation between N^sub 2^nu and N^sub 2^u graphically by plotting N^sub 2^nu versus N^sub 2^u, and calculated the mean contribution of N^sub 2^nu to N^sub 2^tot and a prediction interval of the contribution (mean + or - 1.96 x SD). For our purpose, mean N^sub 2^nu may be viewed as the bias resulting when N^sub 2^u is measured instead of N^sub 2^tot and the prediction interval as a measure of precision.14 We further compared measured N^sub 2^tot values with 2 sets of predicted N^sub 2^tot values. The first set consisted of the values predicted by a linear regression equation that used N^sub 2^u as independent variable. The second set consisted of the values predicted by a linear regression equation that used N^sub 2^u and the weights of other excreta (per kilogram body weight) as independent variables. All excreta were entered into this equation, provided 5 or more measurements were available. For both linear regression equations adjusted revalues were reported, which take into account that the regression models were estimated and tested on the same data. The difference between measured and predicted N^sub 2^tot (residual) was plotted against measured N^sub 2^tot to examine the effect of either regression equation on bias and precision.
Standard descriptive and comparative statistics were calculated on a Macintosh computer using StatView version 4.5. A p value
Of 240 attempted 24-hour collections, 25 were considered nonrepresentative and 8 consisted of urine only, leaving 207 collections, obtained for 78 of the 80 patients included in the trial, for analysis. Patient characteristics are summarized in Table I. Fifty patients were operated on because of a congenital anomaly; 28, because of an acquired disease. The underlying diagnoses were necrotizing enterocolitis (n = 18), duodenal obstruction (n = 14), gastroschizis (n = 7), small bowel atresia (n = 7), Hirschsprung's disease (n = 7), and other (n = 25). Of 207 collections of excreta, 74 were obtained on day 4, 67 on day 5, and 66 on day 6 after surgery. The total collection time for urine was 8 to 16 hours in 17 collections, 16 to 24 hours in 67 collections, and 24 hours in 123 collections.
Nitrogen excretion values are summarized in Table II. N^sub 2^nu, N^sub 2^u, and N^sub 2^tot all varied widely. The ratio between N^sub 2^nu and N^sub 2^u also varied widely: in 50% of all 24-hour intervals, N^sub 2^nu was 25% of N^sub 2^u, with a maximum of 110% (Table II). No correlation was found between N^sub 2^nu and N^sub 2^u (p = .71). Linear regression with N^sub 2^u as independent variable and N^sub 2^tot as dependent variable resulted in equation 1 (adjusted r^sup 2^ 0.923):
Nasoduodenal fluid collections were not entered into equation 2 because only 3 measurements were available.
The mean difference (bias) between N^sub 2^tot and N^sub 2^u was 21 mg/kg/24 hours; the 95% prediction interval, -20 to +63 mg/kg/24 hours (Fig. 1). By definition, both linear regression equations reduced the mean difference between measured and predicted N^sub 2^tot values to zero. Use of equation 1 did not affect the width of the prediction interval (95% prediction interval, -41 to +41 mg/kg/24 hours; Fig. 2A), but use of equation 2 did (95% prediction interval, -34 to +34 mg/kg/24 hours; Fig. 2B).
Multiple linear regression with PRISM and SSS as independent variables showed a small but statistically significant influence of PRISM and SSS on the logarithm of N^sub 2^tot (p
The relation between the weight of the collected excreta and their nitrogen content, assessed separately for nasogastric tube drainage, feces, enterostomy, and wound drain output, was quite variable and seemed to be somewhat nonlinear (data not shown).
In a population of newborns and infants receiving parenteral nutrition 4 to 6 days after digestive tract surgery, we have shown N^sub 2^nu and N^sub 2^u, and their ratio, to vary widely. No correlation was found between N^sub 2^nu and N^sub 2^u. The observed variations in total nitrogen excretion and in the ratio between N^sub 2^nu and N^sub 2^u could only for a small part be attributed to the observed variations in the severity of the surgical stress or the underlying disease.
Nitrogen intake minus nitrogen excretion constitutes the whole-body nitrogen balance. Especially if a patient is being fed parenterally, nitrogen intake can easily be calculated. Measuring nitrogen excretion, however, is an arduous task. In newborns and infants, acquiring accurate nitrogen excretion data is particularly difficult. The relatively small volumes of the excreta may result in relatively large effects of sampling and weighing errors. Critical illness, not having been toilet trained, and the use of urine bags may cause significant sampling errors and leakage. Although we preferred the use of transurethral catheters over urine bags, we felt we could not justify catheterization in boys solely for the purpose of this study because this procedure may cause urethral trauma.15 To avoid the leakage associated with the use of urine bags and long collection intervals, we modified the approach proposed by Boehm et al.13 In a population of low-birth-weight infants, they showed that a continuous 6-hour collection period suffices for metabolic monitoring purposes. Because our protocol did not insist on continuity of urine collection, we arbitrarily decided 8 hours to be the lower time limit of a representative urine collection. The majority of collection intervals (92%), however, comprised 16 to 24 hours. Because other excreta are not produced continuously, these collections had to comprise the full 24-hour interval. Despite our efforts, sampling and weighing errors, particularly of urine or feces, may have occurred. Such errors, though inherent to the clinical setting, would likely affect our precision but not necessarily bias our results. In this setting, validation of data is of paramount importance. However, the wide variation seen in daily nitrogen excretion, and the paucity of similar studies in comparable patient populations, make validation of our findings difficult. Still, other studies have found similarly wide variations of the daily N^sub 2^u in surgical newborns8 and infants,3 pediatric ICU patients,5 and healthy children.16 Chaloupecky et al3 studied 37 infants on the first day after cardiopulmonary bypass surgery and found N^sub 2^u values averaging 235 + or - 83 mg/kg/24 hours. Helms et al8 found N^sub 2^u in 8 preterm newborns to be 209 + or - 99 mg/kg/24 hours on days 1 to 3 after digestive tract surgery and 96 + or - 49 mg/kg/24 hours on day 7. In our study, N^sub 2^u measured on days 4 to 6 after digestive tract surgery was 169 + or - 73 mg/kg/24 hours. Moreover, we found a small but significant influence of both SSS and PRISM on total nitrogen excretion. Our findings are in keeping with those of Chaloupecky et al3 and Helms et al8 and with the notion that nitrogen excretion in infants, as in older children and adults, is proportional to the stress of surgery and to the stress imposed by the underlying disease,1,17-19 indicating that our findings are valid. The limitations of the scoring systems should, however, be noted. SSS was originally designed by Anand and Aynsley-Green11 to grade the surgical stress of several types of surgery, including cardiopulmonary bypass surgery. When using SSS solely for digestive tract surgery, its range of potential values and its discriminative power are limited. Also, the effect of surgical stress on nitrogen excretion may have worn off by day 4 to 6 after surgery. Jones et al20 and Bouwmeester et al21 have shown that resting energy expenditure and (nor)epinephrine levels in newborn patients normalize within 24 hours after surgery. Nitrogen excretion in surgical newborns may parallel resting energy expenditure as it does in adults.1 PRISM II has been validated for use in the Netherlands, but its performance as a predictor of mortality in surgical patients was not as good as that in nonsurgical patients.22,23 PRISM II should be obtained on the day of first admission to the ICU, which occasionally preceded inclusion into this study by several days or even weeks. More importantly, PRISM predicts mortality risk and as such does not always reflect the severity of illness.24 These limitations, and the wide intrinsic variation in nitrogen excretion discussed earlier, may explain why SSS and PRISM had only small effects on total nitrogen excretion.
We are not aware of any data on N^sub 2^nu in populations comparable to the one in our study. As a rule, nonurinary nitrogen loss is neglected or accounted for by a formula that converts urinary urea or N^sub 2^u to total nitrogen excretion.3-6 In one textbook on pediatric intensive care, fecal nitrogen loss is said to equal 20% of urinary excretion, but a reference is not given, precluding identification of the population the original data apply to.2 In a study by Ziegler et al16 in 123 normal children aged 1 to 11 years, fecal nitrogen excretion averaged 15% of N^sub 2^u. In our study, fecal nitrogen excretion was 10 + or - 15% of urinary excretion (data not shown), and nitrogen excretion through all excreta other than urine equaled 14 + or - 15% of urinary excretion. It should be noted that our patients were fed parenterally, which would likely yield a lower stool output when compared with a general pediatrie ICU population.6
Measuring nitrogen excretion in newborns and infants is difficult and laborious, as was mentioned earlier. For this reason, we, and others before us, have tried to find equations that derive N^sub 2^tot from easily obtainable parameters. Because urinary urea content does not reliably predict urinary nitrogen content in critically ill children6-9 and because urine generally is the main route of nitrogen excretion, we used measured N^sub 2^u as the basis of our equations. On average, measured N^sub 2^u underestimated N^sub 2^tot by 21 mg/kg/24 hours (Fig. 1). Adding this fixed amount, or a fixed percentage (in our study, 14%; Table II) to measured (urinary) nitrogen excretion is the approach others have followed in different settings.2,4,25 Whereas this approach reduces bias, it does not improve precision. Similarly, using a linear regression equation will, by definition, eliminate bias but not necessarily improve precision, as is illustrated by equation 1 (Fig. 2A). Equation 2, on the other hand, by incorporating the weights of excreta, did improve precision (Fig. 2B). We chose to use these weights because they are easy to obtain and often are obtained as part of standard patient care. Although equation 2 may suffice for group comparisons, considerable imprecision remained for individual patient assessment, which could only be avoided by measuring the nitrogen content of all excreta. Also, when deciding whether to use equation 2, the clinician should bear in mind that the resulting loss of precision is added to that of sampling and weighing of excreta. Ideally, external validation should be performed first to assess whether the regression coefficients apply to other settings and case mixes.
In conclusion, N^sub 2^nu is variable and not to be neglected when assessing total nitrogen excretion in newborns and infants receiving parenteral nutrition shortly after digestive tract surgery. Although relatively simple predictive equations may yield reasonable estimates, the only way to accurately assess total nitrogen excretion is by measuring the nitrogen content of all excreta.
1. Long CL, Schaffel N, Geiger JW, et al: Metabolic response to injury and illness: Estimation of energy and protein needs from indirect calorimetry and nitrogen balance. JPEN 3:452-456, 1979
2. Deutschman CS: Nutrition and metabolism in the critically ill child. IN Textbook of Pediatric Intensive Care, Rogers MC (ed). Williams & Wilkins, Baltimore, 1992, pp 1109-1131
3. Ghaloupecky V, Hucin B, Tlaskal T, et al: Nitrogen balance, 3-methylhistidine excretion, and plasma amino acid profile in infants after cardiac operations for congenital heart defects: The effect of early nutritional support. J Thorac Cardiovasc Surg 114:1053-1060, 1997
4. Maldonado J, Faus MJ, Bayes R, et al: Apparent nitrogen balance and 3-methylhistidine urinary excretion in intravenously fed children with trauma and infection. Eur J Clin Nutr 42:93-100, 1988
5. Mickell JJ: Urea nitrogen excretion in critically ill children. Pediatrics 70:949-955, 1982
6. Prelack K, Dwyer J, Yu YM, et al: Urinary urea nitrogen is imprecise as a predictor of protein balance in burned children. J Am Diet Assoc 97:489-495, 1997
7. Boehm KA, Helms RA, Storm MC: Assessing the validity of adjusted urinary urea nitrogen as an estimate of total urinary nitrogen in three pediatrie populations. JPEN 18:172-176, 1994
8. Helms RA, Mowatt-Larssen CA, Boehm KA, et al: Urinary nitrogen constituents in the postsurgical preterm neonate receiving parenteral nutrition. JPEN 17:68-72, 1993
9. Patterson BW, Nguyen T, Pierre E, et al: Urea and protein metabolism in burned children: Effect of dietary protein intake. Metabolism 46:573-578, 1997
10. Pollack MM, Ruttimann UE, Getson PR: Pediatric risk of mortality (PRISM) score. Crit Care Med 16:1110-1116, 1988
11. Anand KJ, Aynsley-Green A: Measuring the severity of surgical stress in newborn infants. J Pediatr Surg 23:297-305, 1988
12. Fiedler R, Proksch G: The determination of nitrogen-15 by emission and mass spectrometry in biochemical analysis: A review. Anal Chim Acta 78:1-62, 1975
13. Boehm G, Wiener M, Schmidt C, et al: Usefulness of short-term urine collection in the nutritional monitoring of low birthweight infants. Acta Paediatr 87:339-343, 1998
14. Altman DG: Practical Statistics for Medical Research. Chapman and Hall, London, 1991
15. Elder JS: Urologic disorders in infants and children. IN Nelson Textbook of Pediatrics, Behrman RE, Kliegman RM, Jenson HB (eds). W. B. Saunders Company, Philadelphia, 2000, pp 1619-1658
16. Ziegler EE, O'Donnell AM, Stearns G, et al: Nitrogen balance studies with normal children. Am J Clin Nutr 30:939-946, 1977
17. Pollack MM: Nutritional support of children in the intensive care unit. IN Textbook of Pediatric Nutrition, Suskind RM, Lewinter-Suskind L (eds). Raven Press, New York, 1993, pp 207-223
18. Shew SB, Keshen TH, Jahoor F, et al: The determinants of protein catabolism in neonates on extracorporeal membrane oxygenation. J Pediatr Surg 34:1086-1090, 1999
19. Steinhorn DM, Green TP: Severity of illness correlates with alterations in energy metabolism in the pediatric intensive care unit. Crit Care Med 19:1503-1509, 1991
20. Jones MO, Pierro A, Hammond P, et al: The metabolic response to operative stress in infants. J Pediatr Surg 28:1258-1262, 1993
21. Bouwmeester NJ, Anand KJ, van Dijk M, et al: Hormonal and metabolic stress responses after major surgery in children aged 0-3 years: A double-blind, randomized trial comparing the effects of continuous versus intermittent morphine. Br J Anaesth 87:390-399, 2001
22. Gemke RJ, Bonsel GJ, van Vught AJ: Effectiveness and efficiency of a Dutch pediatric intensive care unit: Validity and application of the Pediatric Risk of Mortality score. Crit Care Med 22:1477-1484, 1994
23. Gemke RJ, Bonsel GJ: Comparative assessment of pediatric intensive care: A national multicenter study: Pediatric Intensive Care Assessment of Outcome (PICASSO) Study Group. Crit Care Med 23:238-245, 1995
24. Shann F: Are we doing a good job: PRISM, PIM and all that. Intensive Care Med 28:105-107, 2002
25. Mackenzie TA, Clark NG, Bistrian BR, et al: A simple method for estimating nitrogen balance in hospitalized patients: A review and supporting data for a previously proposed technique. J Am Coll Nutr 4:575-581, 1985
Marcel J.I.J. Albers*; Ewout W. Steyerberg[dagger]; Trinet Rietveld[double dagger]; and Dick Tibboel*
From the * Departments of Pediatric Surgery, * Public Health, and [double dagger] Internal Medicine, Sophia Children's Hospital/Erasmus Medical Center, Rotterdam, The Netherlands
Received for publication September 20, 2002.
Accepted for publication April 25, 2003.
Correspondence: Marcel J.I.J. Albers, Department of Pediatrics, Groningen University Hospital, X4.111 PO Box 30001, 9700 RB Groningen, The Netherlands. Electronic mail may be sent to email@example.com.
Copyright American Society for Parenteral and Enteral Nutrition Sep/Oct 2003
Provided by ProQuest Information and Learning Company. All rights Reserved