Replies received from 59 members (80%) and 13 network partners (41%) of the University HealthSystems Consortium using total parenteral nutrition in the first survey to determine the usual concentrations of dextrose used in total parenteral nutrition.
Result of a questionnaire sent to the same institutions as in the first survey to determine the usual concentrations of amino acids used in total parenteral nutrition.
Concentrations of dextrose in total parenteral nutrition for 2 hypothetical patients in the second survey (n=172).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Schloerb PR, Henning JF. Patterns and Problems of Adult Total Parenteral Nutrition Use in US Academic Medical Centers. Arch Surg. 1998;133(1):7–12. doi:10.1001/archsurg.133.1.7
To determine the pattern of total parenteral nutrition (TPN) use in US academic medical centers because TPN in adults may be associated with complications related to excessive glucose (dextrose) administration and a respiratory quotient greater than 1.0.
Two surveys of the University HealthSystems Consortium (n=106, 74 members and 32 network partners using TPN) to determine TPN formulas and amounts of TPN nutrients given to 2 hypothetical abdominal trauma patients (1 man and 1 woman), each of whom weighed 70 kg.
Main Outcome Measures
Amounts of dextrose, amino acids, and fat and rates of administration of TPN.
In the first survey, 80% (59/74) of members and 41% (13/32) of network partners of the University HealthSystems Consortium reported using TPN formulas with more than 20% dextrose; half used 25% dextrose. In the second survey, the mean (±SD) TPN dextrose concentrations were 190±43 and 170±45 g/L (902±204 and 807±214 mmol/L), with amino acid concentrations at 40 to 50 g/L, in the male and female patients, respectively. The amounts of amino acids and glucose given, when referred to body cell mass in the male and female patients, were equivalent. In 26% (22/86) of the institutions surveyed, the amounts of glucose given in TPN were high enough (>4.48 mg/kg per minute) to produce a respiratory quotient greater than 1.0. A standard TPN formula was derived as 4.25% amino acids, 15% dextrose, and 20% fat emulsion, at a rate to provide required calories.
Excessive TPN glucose administration, found in at least one fourth of US academic medical centers, suggests use of a TPN formula with no more than 15% dextrose, administered at a rate to provide no more than 4 mg/kg of glucose per minute.
ALTHOUGH potentially a life-saving measure, recent evidence suggests that total parenteral nutrition (TPN) may be harmful. In a study of 395 male patients who underwent abdominal or noncardiac surgery, randomized to TPN or no TPN perioperatively, there was no evidence of benefit to those patients who received TPN, except, perhaps, for a small subset identified as severely malnourished.1 This report confirmed a meta-analysis, in which it was concluded that the routine use of perioperative TPN in unselected patients is not justified, although TPN may be warranted in patients at high risk.2 In a recent carefully executed study by Brennan et al,3 117 patients undergoing major pancreatic resections were randomized to receive TPN or no TPN postoperatively. In those who received TPN, the incidence of major complications, mostly intra-abdominal complications and infectious, was doubled (P<.05), with a 4-fold increase in mortality (P>.05).
One clue to the mechanism of adverse effects of TPN was found by Guenst and Nelson4 in a correlation between rate of glucose infusion and, by indirect calorimetry, the ratio of carbon dioxide produced to oxygen consumed, or the respiratory quotient (RQ). An RQ greater than 1.0 signifies net lipogenesis, a futile metabolic event in a malnourished, hypermetabolic patient whose priorities for energy expenditure do not include fat synthesis. In this study of 140 patients receiving TPN, 4% had an RQ greater than 1.0. An RQ greater than 1.0 was associated with a mean(±SD) glucose intake of 4.48±1.88 mg/kg per minute, while an RQ less than 1.1 was associated with a glucose intake of 2.89±1.34 mg/kg per minute (P<.001). Patients in this study usually received 25% glucose (dextrose).4 Dextrose is a glucose monohydrate with only 3.42 kcal/g (14.36×103 J/g). The small differences between glucose and dextrose are not relevant to the data to be presented, and the terms will be used interchangeably in this article. Although the volumes of TPN are not specified, calculation using their average(±SD) body weight (69.3±15.8 kg for all patients) and rates of glucose infusion from their tables suggests that TPN was provided in the usual range of 1.5 to 3.0 L of 25% glucose per day. In a patient weighing 70 kg and receiving 35 mL/kg or 2.5 L/d of TPN containing 25% glucose, the glucose infusion rate would be 6.1 mg/kg of glucose per minute. This is much larger than the maximum values reported by Guenst and Nelson4 to be associated with an RQ greater than 1.0 and can also result in hyperglycemia.5,6
How common is the problem of excessive use of glucose in TPN? As stated in a recent text, "The majority of standard parenteral nutrient solutions contain dextrose concentrations between 25 to 35%."7 To determine the amount of glucose administered and the pattern of use of TPN in academic medical centers in the United States, 2 surveys were conducted of the members and network partners of the University HealthSystems Consortium (UHC), Chicago, Ill. This organization included, at the time of this study, 74 members and 32 network partners. Members are typified by university medical centers, while network partners include affiliated hospitals, some of which indicated that they did not use TPN. Although not the only university hospital group, UHC is the largest and seems to be representative of US academic medical centers. This article presents this survey information and suggests that problems may exist in the formulation and administration of TPN.
A questionnaire was sent by the UHC to pharmacy directors of all 74 members and 32 network partners. The following questions were asked:
Based on the final concentrations of adult TPN solutions prepared at your institution, what percentages of dextrose and amino acids are used most of the time?
Does your institution use fat emulsions, all in one, with dextrose or amino acids? (Yes or no?)
Each of the initial nonresponders, and some of those whose answers were considered ambiguous, were contacted by telephone by either one of us to obtain the data.
Although increased glucose concentrations in TPN may suggest the possibility of excessive glucose administration, this information does not itself prove that excessive glucose was received. To determine an indication of the amount of glucose given from the type and rate of TPN administered, questions about the specific treatment of 2 hypothetical patients were sent by us to the same UHC addressees. These patients, the day after laparotomy for duodenal and pancreatic trauma, were assumed to require TPN with calories 20% in excess of the basal energy expenditure (BEE), calculated by the Harris-Benedict equations.8 This activity factor(±SD), 1.20±0.25 times the BEE, was that found by Guenst and Nelson4 in 140 intensive care unit patients by indirect calorimetry. Respondents were asked to provide, not for attribution, TPN prescriptions (only specify amino acids, dextrose, and fat emulsion, not additives nor fluid for gastrointestinal tract or drainage losses) for the following 2 clinical examples:
A 21-year-old, 70-kg, 183-cm (6-ft) previously healthy male sustained abdomincal injury a motor vehicle abdominal injury and underwent laparotomy yesterday with suture closure of ruptured duodenum and drainage of pancreatic area. Vital signs are near normal, and TPN is to be started. Assume about 2162 kcal (90.80×105 J) are to be provided.
A 70-year-old, 70-kg, 152-cm (5-ft) previously healthy female sustained the same injury and operation and is to begin TPN the following day. Assume about 1534 kcal (64.43×105 J) are to be provided.
Please provide the following for each of the above examples: amino acids (grams per liter), dextrose (grams per liter [millimoles per liter]), 10% fat emulsion (milliliters) or 20% fat emulsion (milliliters), and target rate (milliliters per hour).
Are amino acids, dextrose, and fat emulsion administered 3 in 1? (Yes or no?)
To provide a better reference than body weight for nutrient intake in these 2 patients of presumed different body composition, total body water (TBW) was calculated using an equation derived by Watson et al9 based on sex, age, height, and weight in 723 subjects. Body cell mass (BCM) was estimated as 70% TBW, an approximate value that was found to range from an average of 74% TBW in normal men younger than 51 years to 68% TBW for older men and all women.10 With the use of these calculations, in the male patient, TBW is 42 L and BCM is 29 kg. In the obese female patient, who also weighs 70 kg, TBW is 28 L and BCM is 20 kg.
Data are expressed as the mean(±SD). Statistical significance was accepted at P<.05, using a statistical analysis program (Statistix, 4.0, Analytical Software, St Paul, Minn).
Data received from 59 (80%) of the 74 UHC members and 13 (41%) of the 32 network partners on concentrations of dextrose in TPN are summarized in Figure 1. Total parenteral nutrition solutions with 20% or greater concentrations of dextrose were reported by 56 (78%) of the 72 respondents, with almost half using 25% dextrose. The corresponding concentrations of amino acids indicated that most institutions use 4% to 5% amino acids in TPN (Figure 2).
Data received from 86 institutions, including 74 members (95% of the 78 surveyed) and 12 network partners (38% of the 32 surveyed), responding to the questions relating to TPN treatment of the 2 hypothetical patients (n=172) are summarized in Table 1. Each of the 4 nonresponding members and 20 network partners were contacted by telephone. It was determined for the 4 members that they did not use TPN (1 member), that they outsourced TPN preparation (1 member), that they would use enteral feedings in the patient examples (1 member), or that they did not treat adult patients (1 member). Follow-up telephone calls to the 20 nonresponding network partners usually elicited the reply that they did not use TPN. The responses of the 12 network partners did not differ notably from those of the members in any of the categories of the tables.
Of replies to the second survey received from 86 institutions, 68 (79.1%) were received from pharmacists, 10 (11.6%) were received from dietitians, 2 (2.3%) were received from dietitians or pharmacists, 5 (5.8%) were received from physicians, and 1 (1.2%) was received from a nurse-nutritionist. Of the 68 replies from pharmacists, 29 identified themselves as doctors of pharmacy, 4 as board-certified nutritional support pharmacists, 9 as doctors of pharmacy and board-certified nutritional support pharmacists, and 1 as a doctor of pharmacy, a board-certified nutritional support pharmacist, and a registered dietitian. Replies from physicians were usually (4 of 5) obtained by direct solicitation after repeated failure to receive institutional responses from hospital pharmacies.
The average values of amino acids and dextrose, fat emulsion, and rates of infusion are generally in accord with common practice (Table 1). However, there was wide variation in the TPN formulations, as indicated by large coefficients of variation. Of 172 requests to 86 members for TPN prescriptions for the 2 hypothetical patients, each weighing 70 kg, 83 institutions (97%) used glucose in the male patient and 26 (30%) used glucose in the female patient at rates in excess of the 2.89 mg/kg per minute reported by Guenst and Nelson4 to be associated with a RQ greater than 1.0. The wide range of dextrose concentrations in these 172 patient TPN formulations is shown in Figure 3. Eighty percent (69/86) of the respondents used more than 15% dextrose, while 38% (33/86) used more than 20% dextrose and 14% (12/86) used 25% dextrose.
At the rates of TPN administration reported (Table 1), the male patient received 2081±349 mL (30 mL/kg per day). The female patient received 1600±354 mL (23 mL/kg per day). Several respondents commented that they would give more calories than the 20% above the Harris-Benedict–derived BEE posed in the questions.
The exclusive use of 10% fat emulsion in 14 instances was highly variable. Seven institutions (8%) used both 10% and 20% fat emulsion. Most institutions, 76% (65/86), used only 20% fat emulsion. Confirming the first survey was the response indicating that 53% (46/86) of the UHC respondents use dextrose, amino acids, and fat emulsion as an all-in-one mixture.
The hypothetical male and female patients, weighing the same, 70 kg (preinjury), received amino acids at 1.50±0.24 and 1.13±0.27 g/kg body weight, respectively (Table 2). With the use of weight as a reference, the female patient received less amino acids (P<.05). To relate nutrient intakes to a more physiologic function than body weight, the data are expressed per unit BCM in Table 2. The average values for amino acid and glucose intake in these 2 diverse clinical examples, when expressed in this way, are almost equivalent (P>.05). In calculating TBW and BCM, it is important to use "dry" body weight to avoid errors attributable to fluid retention after surgery or trauma. Total calories prescribed were within 10% of the amounts stipulated in the questions and did not differ significantly (P>.05) for the 2 patients when related to BCM.
From the results of these surveys, it would seem that 26% (22/86) of these academic medical centers surveyed and using TPN are exceeding the amount of glucose found by Guenst and Nelson4 to result in an RQ greater than 1.0. Based on available clinical and animal studies, other harmful effects, including increased net lipogenesis and hyperglycemia with glycation of proteins, could ensue. There seems to be a notable lag between research and practice.
Of replies to the second survey received from 86 institutions, 68 (79%) were sent by pharmacists. Because this questionnaire was addressed to pharmacy directors, it is to be expected that a predominance of replies would come from pharmacists. Whether this reflects the actual situation in any individual practice, department, or institution may be questioned.Total parenteral nutrition dextrose values in the first and second surveys were disparate. The first survey was sent to the pharmacy directors with the request to provide usual concentrations of dextrose and amino acids in TPN, with the result that 56 (78%) of the responding institutions used more than 20% dextrose and almost half used more than 25% dextrose. Replies from the second survey, in contrast, indicated that only 14% (12/86) used 25% dextrose, although 80% (69/86) used more than 15% dextrose compared with 93% (67/72) in the first survey (P>.05). To what extent did the nature of the questions influence the answers in the second survey?
The data of Guenst and Nelson4 suggest that the glucose concentration should be limited to less than 4 mg/kg per minute. For an arbitrary value of 3.7 mg/kg per minute in a young man who is 183 cm tall and weighs 70 kg, this amount of glucose would be 373 g/d. If the water requirement for a young man is approximately 35 mL/kg per day or 2.5 L, the glucose concentration would be approximately 150 g/L (832.6 mmol/L) or 15%. If the protein (amino acid) requirement is 1.5 g/kg per day, this would be 105 g (or 42 g/L) of TPN. Many pharmacies, including members of the UHC (Table 1), use a stock amino acid solution of 8.5% amino acids, diluted 1:1 with dextrose solution. The final amino acid concentration obtained is, therefore, 4.25%. A standard TPN solution would be 4.25% amino acids and 15% dextrose plus fat, usually 200 mL of a 20% fat emulsion, to provide the remaining required calories. The caloric content of 4.25% amino acids with 15% dextrose is 697 kilocalories per liter (29.27×105J/L). With 200 mL of 20% fat emulsion, 400 additional calories are added. In 2 L of TPN, the caloric content is, therefore, 1794 kcal (75.35×105 J). The rate of administration is dependent on caloric needs. The increase or decrease of the volume of 20% fat emulsion by 50 to 100 mL will adjust for approximate lower or higher estimates of energy expenditure.
For the hypothetical male patient in the second survey, who weighed 70 kg (BEE=1802 kcal×1.2=2162 kcal [75.68×105 J×1.2=90.80 J]), a fluid intake of 35 mL/kg per day or 2.5 L and a glucose intake of 3.7 mg/kg per minute would provide a total glucose intake of 373 g, as 2.5 L of 15% dextrose, 4.25% amino acids, and approximately 250 mL of 20% fat emulsion, all in one, at 100 mL/h, to provide 2173 kcal (91.27×105 J).
For the female patient who also weighed 70 kg (BEE=1279 kilocalories×1.2=1534 kcal [53.72×105 J×1.2=64.43×105]), the same approach results in a TPN prescription of 4.25% amino acids, 15% dextrose, and 150 mL of 20% fat emulsion, all in one, at a rate of 75 mL/h or 1.8 L/d, to provide 1555 kcal (65.31×105 J).
With an activity factor of 1.2 and a 2-L TPN volume, amino acids are 20% of required calories; carbohydrates, 60%; and fat, 20%.
Hyperglycemia occurs in proportion to the rate of glucose infusion in TPN.6 The consequences of hyperglycemia associated with TPN have been reviewed by Bistrian.11 Hyperglycemia can produce glycosylation (glycation) of proteins, with resultant dysfunction. Glycation of immunoglobulin can reduce defense against infection.12 Glycation of collagen can interfere with wound healing,13 and nonenzymatic glycation of albumin can augment platelet aggregation in proportion to plasma glucose concentration.14 Acute hyperglycemia with hyperinsulinemia has been found to enhance proteolysis in normal humans.15 In an animal study of induced spinal cord ischemia,16 glucose administration with a blood glucose concentration increased to only 170 mg/dL (9.4 mmol/L) was associated with more neurologic damage than in animals that received no glucose. Although the treatment of TPN-induced hyperglycemia with insulin can correct the blood glucose level elevation if sufficient insulin is given, problems can result from this approach. These include the increased incidence of fatty liver, the possibility that blocking amino acid mobilization may adversely affect outcome, and an inability to increase glucose oxidation despite lowering serum glucose concentrations.17
Therapeutic goals include resolution of the disease process, prevention or treatment of infection and organ failure, wound healing, decreased length of hospital stay, and eventual correction of malnutrition. A fundamental question arises: Do the presumed benefits of providing the entire caloric expenditure and attempting to achieve positive nitrogen balance during acute illness outweigh the hazards of an RQ greater than 1.0 and hyperglycemia with the harmful consequences of glycation of proteins?
At the University of Kansas Medical Center, Kansas City, prior to 1988 the usual glucose concentration in TPN was 25%. This decreased to 20% approximately 5 years ago and to 15% 2 years ago, for the reasons previously outlined. Our present standard TPN formula is 4.25% amino acids, 15% dextrose with (usually) 200 mL of 20% fat emulsion, all in one, at a rate to provide required calories. Although data support a decreased incidence of hyperglycemia6 with this reduction of glucose, we do not as yet have supporting evidence relating decreased parenteral glucose intake and energy expenditure to RQ. Data about patient outcomes with this formula compared with TPN formulas used previously would be difficult to interpret.
Guenst and Nelson4 showed in 140 intensive care unit patients receiving TPN that almost half received excessive amounts of glucose, with a resultant RQ greater than 1.0 and increased net lipogenesis. An explanation for the adverse effects of TPN is suggested by the results of our surveys, showing that excessive glucose was found to be administered by at least one fourth of US academic medical centers. Available evidence suggests that the reduction of glucose administration in TPN to less than 4 mg/kg per minute would be beneficial. The wide variety of institutional responses to relatively standard indications for TPN supports the use of a standardized TPN formula with no more than 15% dextrose at rates to assure the avoidance of complications.
We thank the staff and members of the University Health-Systems Consortium for providing survey information.
Corresponding author: Paul R. Schloerb, MD, Department of Surgery, University of Kansas Medical Center, Kansas City, KS 66160 (e-mail: firstname.lastname@example.org).