This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Citing Articles via Scopus Google Scholar Articles by Sheean, P. Articles by Braunschweig, C. Search for Related Content PubMed PubMed Citation Articles by Sheean, P. Articles by Braunschweig, C. Pubmed/NCBI databases Compound via MeSH Substance via MeSH Hazardous Substances DB GLUCOSE Social Bookmarking                   What's this?

The Incidence and Impact of Dextrose Dose on Hyperglycemia From Parenteral Nutrition (PN) Exposure in Hematopoietic Stem Cell Transplant (HSCT) Recipients

Department of Preventive Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL and Department of Human Nutrition, University of Illinois at Chicago, Chicago, IL

Correspondence: Carol Braunschweig, PhD, RD, 1919 W. Taylor, Department of Human Nutrition (m/c 517), University of Illinois at Chicago, Chicago IL 60612. Electronic mail may be sent to braunsch{at}uic.edu.

Background: Short-term, transient hyperglycemia is associated with adverse outcomes in acutely ill populations. Because parenteral nutrition (PN) is dextrose based, we hypothesized that exposure to PN would be associated with hyperglycemia and that greater levels of dextrose infusion would be associated with higher glucose concentrations. Our objective was to examine the temporality, incidence, and dose response from dextrose load upon hyperglycemia using several serum glucose cut points in PN vs non-PN HSCT recipients. Methods: The medical records of adults admitted for initial autologous or allogeneic hematopoietic stem cell transplant (HSCT) at 2 university-affiliated hospitals between September 1999 and December 2003 were used in this retrospective cohort. To minimize the impact of disease acuity on serum glucose, patients with diabetes mellitus, steroid administration, patients with recently treated infections, or patients who died during therapy were eliminated from the study. Serum glucose values were recorded once per day from the first morning venous blood draw (2 AM–6AM) to achieve uniformity among patients, to avoid measurements occurring more frequently among hyperglycemic patients, and to minimize the influence of oral intake. Hyperglycemia was examined using several serum glucose cut points (110, 125, 150, 175, and 200 mg/dL). Wilcoxon rank-sum tests were used to detect differences in hyperglycemic events between PN and non-PN subjects, and mixed-effects regression models were used to detect the association between PN exposure and hyperglycemia. To address the temporality and incidence of hyperglycemia between PN vs non-PN participants, before and "after" time frames were created. Preinfusion (before) and actual infusion (after) times were used for these intervals for PN patients; however, the average hospital days before (before) or during (after) PN infusion were used for comparison in non-PN recipients (ie, autologous non-PN before = hospital days 1–10, after = hospital days 11–21). Results: Of the 208 patients who qualified for inclusion 49% (n = 101/208) received PN, which provided on average 26 kcal per kg, 1.3 g of protein per kg, and 2.7 mg/kg/min of dextrose (range 1.3–3.9 mg/kg/min). The proportion of hyperglycemic days before was not different between groups; however, it was significantly greater after in PN vs non-PN patients, regardless of serum glucose cut point. A dose response between dextrose administered (mg/kg/min) and serum glucose concentrations was not seen. When longitudinally presented, the temporal relationship between serum glucose and PN initiation was reflected approximately on hospital day 9. Using regression models that account for repeated measures, the odds of having hyperglycemia (yes/no; glucose >110 mg/dL) after PN exposure were nearly 4 times (odds ratio 3.9; 95% confidence interval, 2.7–5.5) that of non-PN exposed, after controlling for donor type, race, age, and conditioning chemotherapy. PN was the only variable to significantly interact with time (p < .0001), signifying not only the change in odds over time but also as powerful evidence that PN was the causative agent of hyperglycemic events. Conclusions: The broad use of PN at levels within current clinical guidelines in HSCT adults was associated with profound hyperglycemia; however, greater dextrose dose, within the narrow levels administered in this cohort, was not associated with higher glucose concentrations.

Optimal short-term serum glucose control in hospitalized patients with and without diabetes mellitus has gained significantly more attention from clinicians since the landmark trials by Van den Berghe et al,^1,2 which demonstrated dramatic improvements in morbidity and mortality in critically ill patients with tight serum glucose control (80–110 mg/dL) compared with conventional targets (180–200 mg/dL). Stress-induced hyperglycemia is commonly observed in acutely ill patient populations due to impairments in insulin-mediated glucose uptake in the skeletal muscle and failure of insulin to suppress hepatic gluconeogenesis.³ Parenteral nutrition (PN) is frequently a component of care for these patients and, because of its dextrose base, potentially compounds the serum glucose aberrations associated with illness. Several previous investigators have reported hyperglycemia as a complication of PN use^4–10; however, these have limited applicability due to several changes in clinical practice since their publication. Additionally, few new observational investigations have been reported in PN vs non-PN patients because of the perceived bias of inherent differences between these groups for severity of illness.

We recently reported numerous adverse outcomes associated with hyperglycemia from PN exposure (glucose 110 mg/dL) in a retrospective cohort of 357 adults undergoing initial autologous or allogeneic hematopoietic stem cell transplant (HSCT).¹¹ However, the purpose of this study was to expand these findings to examine the incidence of hyperglycemia in a more narrowly defined population of PN vs non-PN HSCT recipients using several serum glucose cut points and to determine if a dose-response existed between dextrose load and increased hyperglycemic risk.

	METHODS

Top METHODS RESULTS DISCUSSION

 
Study Design
A retrospective cohort investigation was conducted using the medical records of patients who had an initial HSCT between September 1999 and December 2003 at 2 urban university hospitals. This study design was selected because it provided an efficient, inexpensive means of examining our hypotheses with minimal risks to patients, allowed for precise documentation of both the exposure (initiation and termination of PN) and the major outcome (hyperglycemia) equally for all patients, and allowed for accurate ascertainment of incidence rates, temporality of hyperglycemia, and dose-response measurements for PN exposure between groups.

Study Population
The medical records of adults admitted for initial autologous or allogeneic HSCT at 2 university-affiliated transplant centers between September 1999 and December 2003 were used in this retrospective cohort. All patients 18 years of age, admitted for an initial autologous or allogeneic HSCT, were eligible for inclusion. To minimize the impact of disease acuity on serum glucose, patients with diabetes mellitus, steroid administration, patients with recently treated infections, or patients who died during therapy were eliminated from the study. Additionally, individuals with a history of home PN administration and those admitted with an existing infection that required antimicrobial treatment or patients <18 years old were excluded. This patient population was selected because (1) PN exposure generally occurs in 50%–60% of patients; (2) PN is delivered via a central line, representing uniform exposure; and (3) in general, patients have comparable baseline cardiac, pulmonary, renal, and hepatic function to withstand transplantation.

Data Collection
Institution-specific transplant databases were obtained and screened for initial eligibility working retrospectively. Computerized and hardcopy medical records were used. Baseline data were extracted from the physician's history and physical. Preparative chemotherapy regimens were verified from the pharmacy chemotherapy dosing records. Culture results and daily blood draws were obtained from laboratory reports. Infections were defined as those that required antimicrobial treatment, as ordered and determined by the attending physician. Admission height and weight, PN dose, volume, and duration were collected from the initial nutrition evaluation and from pharmacy records.

Outcome Measures
The primary outcome measure was hyperglycemia. Glucose was recorded once per day from the first morning venous blood draw (2 AM–6AM) to achieve uniformity among patients, to avoid measurements occurring more frequently among hyperglycemic patients, and to minimize the influence of oral intake. The percentage of hospital days with serum glucose >110 mg/dL (6.1 mmol/L),^1,2 >125 mg/dL (6.9 mmol/L), 150 mg/dL (8.3 mmol/L), >175 mg/dL (9.7 mmol/L) and >200 mg/dL (11.1 mmol/L) was used to evaluate hyperglycemic events. The amount of dextrose provided was calculated as the average amount of dextrose supplied by PN (in mg) each day divided by admission weight, adjusted for obesity accordingly, divided by 1440 minutes per day.

Comparability of PN and Non-PN Participants
It was essential to demonstrate initial comparability between PN and non-PN subjects because of the direct influences of disease acuity on glucose control; however, no scale currently exists to assess severity of illness within the stem cell transplant population. Moreover, although this population frequently receives PN, necessity is largely left to physician discretion because universally accepted guidelines for PN initiation and duration are lacking. As a result, to assess potential differences, admission nutrition status, white blood cell (WBC), oral temperature, glucose concentration, and other vital clinical characteristics were evaluated at baseline between groups. Patients who appeared fundamentally different were excluded (ie, 5 patients with persistent, unexplained hyperglycemia). Further, to address the temporality and incidence of hyperglycemia between PN and non-PN participants, "before" and "after" time frames were created. Preinfusion (before) and actual infusion (after) times were used for these intervals for PN patients; however, the average hospital days before (before) or during (after) PN infusion were used for comparison in non-PN recipients. Finally, to increase the comparability between the PN and non-PN groups, patients were excluded if they were (1) more prone to hyperglycemia (ie, history of diabetes mellitus or requiring steroid administration); (2) possessed higher illness acuity (ie, death during hospitalization or treated infection before PN exposure); or (3) had atypical donor types associated with higher rates of transplant-related complications and mortality (eg, matched unrelated donor, tandem or subsequent transplant).

Data Quality
To assure reproducible and accurate data points, a systematic data collection procedure was developed where each medical record was reviewed in the same manner, beginning with the transplant discharge summary, followed by history and physical review, followed by laboratory report review, etc. Additionally, data points were identified and then verified in more than 1 location within the medical record (eg, physician documentation of infection confirmed by laboratory reports). Quality assurance was performed using 2–4 records per month by an individual not involved in the study, and discrepancies were resolved by consensus, developing methods to prevent future errors. All data were double entered and cleaned in Epi Info 6.¹² This study was approved by the institutional review boards at both university transplant centers.

Statistical Analyses
Means, medians, standard deviations (SDs), and ranges were used to examine and describe the distribution of the data. Student's t-tests and ² tests were used to detect differences in clinical and demographic characteristics between feeding strata, and Wilcoxon rank-sum tests were used to detect differences in hyperglycemic events between PN and non-PN subjects. General estimated equation models (GEEs) that accounted for repeated measures were used to ascertain the odds ratios (OR) and 95% confidence intervals (CI) for the association between PN exposure and hyperglycemia (glucose 110). Statistical analysis was conducted using the statistical program SAS (version 8).¹³

	RESULTS

Top METHODS RESULTS DISCUSSION

 
Baseline characteristics of the 208 patients who met the eligibility criteria are provided in Table I. Overall, 49% (n = 101/208) of patients received PN, which provided on average 26 kcal per kg (range 12–33 kcal/kg), 266 g of carbohydrate (range 42–426 g), and 1.3 g of protein per kg (0.6–2.0 g/kg) for a mean of 11 days (±6 days, median = 10 days). No difference was found for any baseline characteristic, except non-PN patients had significantly shorter hospital stays. Serum glucose was 133 ± 29 (range, 83–243 mg/dL; quartiles, <112, 112–126, 127–150, >150 mg/dL) vs 102 ± 12 mg/dL (range, 75–150; quartiles, <96, 96–102, 103–107, >107 mg/dL) for PN vs non-PN subjects, respectively, when multiple measures after were averaged. The proportion of hyperglycemic days before and after categorized at various glucose cut points is provided in Table II. The proportion of hyperglycemic days was not statistically different in the before time period for any glucose cut point category between PN and non-PN patients; however, in the after time period, significantly higher proportions of hyperglycemic days were found in PN compared with non-PN patients. Within PN patients, the proportion of hyperglycemic days was significantly higher after PN exposure (after) than before exposure (before), regardless of the cut point used for hyperglycemia categorization (p < .006), indicating an increase in the incidence of hyperglycemia. The opposite was seen among non-PN patients. The proportion of hyperglycemic days was significantly lower in the after time period than in the before time frame (p < .001) for all serum glucose classifications, representing a decrease in the rate of hyperglycemia rate for all categories.

For PN patients, the average dextrose infusion rate was 2.7 mg/kg/min (range, 1.3–3.9 mg/kg/min). Dextrose dose was similar for all categories of the glucose classification (Table III).

When longitudinally presented, the temporal relationship between serum glucose and PN initiation is reflected at approximately hospital day 9 (Figure 1; n = 4360 observations). Using regression models that account for repeated measures, the odds of having hyperglycemia (yes/no; glucose >110 mg/dL) after PN exposure were nearly 4 times (OR 3.9; 95% CI, 2.7–5.5) that of non-PN exposed, after controlling for donor type, race, age, and conditioning chemotherapy. PN was the only variable to significantly interact with time (p < .0001), signifying not only the change in odds over time but also as powerful evidence that PN was the causative agent of hyperglycemic events.

View larger version (7K): [in this window] [in a new window]   FIGURE 1. Glucose over time, stratified by PN.

	DISCUSSION

Top METHODS RESULTS DISCUSSION

To date, studies that have examined the incidence of hyperglycemia a priori in patients who receive PN vs non-PN recipients are lacking. Moreover, the inferences made from post hoc analysis of PN studies are problematic in that they should only be used for future hypothesis generation and not as conclusive findings. Previously we conducted a meta-analysis on prospective randomized clinical trials that examined the efficacy of PN compared with tube feeding or standard care on major clinical outcomes.¹⁴ None of the 7 studies that compared PN to standard care reported hyperglycemic events. Twenty studies examined major outcomes for tube feeding (TF) vs PN; of these, 7 reported incidents of hyperglycemia,^4–9,15 all of which occurred more frequently in participants receiving PN. Comparisons between our results and these studies are difficult due to differences in glucose monitoring and definitions for hyperglycemia used by the various authors. Specifically, in half of these investigations hyperglycemic complications were reported without specific glucose criteria,^4,5,7 whereas the others used hyperglycemia definitions of >200 mg/dL^6,8 or >250 mg/dL.⁹ Additionally, whereas these studies were similar to ours in the central delivery of PN, they included a variety of medical and surgical patient populations with different disease processes, which can influence both the rate and timing of stress-induced hyperglycemia. Moreover, none of these investigations reported standardized serum glucose monitoring. Recently Cheung et al¹⁶ retrospectively investigated the associations of hyperglycemia and mortality in 109 medical patients receiving PN and found it was a strong predictor of increased mortality. The mean serum glucose concentrations and quartile ranges in their patients (144 ± 27 mg/dL; quartiles, 124, 125–140, 141–164, and >164 mg/dL) were similar to ours; however, the uncontrolled design and failure to address disease acuity in their population limit other comparisons that can be made.

Current guidelines for parenterally fed stressed patients recommend dextrose infusions of no more than 5–7 mg/kg/min¹⁷ because stress-induced hypercatabolism reduces the ability to oxidize glucose to carbon dioxide and potentially induces lipogenesis, fatty liver, respiratory quotient elevations, hyperglycemia, glycosuria, and hyperosmolar nonketotic coma.¹⁸ Rosmarin et al¹⁹ reported when nonobese, nondiabetic subjects were provided PN at rates exceeding 4 mg/kg/min (22%, n = 23/102 patients), a significantly greater incidence for hyperglycemia (glucose >200 mg/dL; 11 mmol/L) was observed, but none was reported in patients fed below that level. On average, their patients received nearly 32 kcal/kg (range 17–45 kcal/kg), and the mean glucose concentration reported for "euglycemic" (patients that received 4 mg/kg/min) was 146 mg/dL (range 102–194). Dextrose and energy infusions for our entire PN cohort were consistent with current nutrition support practice guidelines (mean dextrose, 2.7 mg/kg/min; range, 1.3–3.9 mg/kg/min; 26 kcal/kg, range, 12–33 kcal/kg), and lower than patients in the Rosmarin et al¹⁹ study. None of our participants received over 4 mg/kg/min; 28% had mean glucose concentrations above 150 mg/dL and 3% had levels above 200 mg/dL. These levels are quite similar to the mean levels reported for the Rosmarin et al¹⁹ "euglycemic" patients, corroborating our findings of hyperglycemia occurrence with dextrose infusions <4 mg/kg/min.

Our inability to depict a dose-response relationship between dextrose load and higher levels of glucose is intriguing, suggesting that some other attributes of PN may play a vital role in our observations. Even though PN is lifesaving in patients with gut failure, the notion that PN is not a perfect science applies. These findings may reflect the absence of the protective "first pass" effect via the portal metabolism on a normally highly regulated system. Elevated glucose concentrations with dextrose infusions within clinical guidelines also suggest that the continuous infusion of dextrose compounds the insulin resistance associated with the stress-related acute-phase response. The growing body of literature documenting the untoward effects of hyperglycemia on patient outcomes dictates an urgent need to identify patients who benefit from PN and the development of guidelines that ensure administration that does not precipitate elevated glucose concentrations. The prophylactic use of insulin in PN when administered to high-stress patients, who are often common recipients of PN therapy, may be warranted.

The limitations of this study should be noted. First, the potential for the groups to be fundamentally different cannot be ruled out due to the retrospective design and the lack of a valid measure for disease acuity for this patient population. To minimize this, tight eligibility criteria that eliminated patients prone to hyperglycemia, including those who received steroids, those with diabetes, and sicker patients, were excluded. Second, data on patients' oral intake, other sources of IV dextrose, and medications other than steroids that could have affected glucose concentrations were not assessed. Third, patients classified as obese were included in this study to enhance the applications and generalizability of these findings to modern practice; however, this may have increased the incidence of hyperglycemia in both PN and non-PN recipients. Finally, insulin administration was not assessed. Because only 1% (n = 3/208) of all patients had glucose concentrations above the traditional level that triggers insulin administration (200 mg/dL), we do not believe this was a major contributor for the differences observed in hyperglycemia.

In conclusion, adult HSCT patients exposed to PN administered within clinical guidelines experienced a greater incidence of hyperglycemia, using various serum glucose categorizations, when compared with nonrecipients. No dose response between dextrose infused and mean glucose concentration was observed, suggesting either a very low threshold for dextrose administration in acutely ill HSCT patients and/or other underlying components within PN are influencing serum glucose concentrations. These data support the need for close glucose monitoring to avoid hyperglycemia during PN administration and indicate the need for further investigations to discern optimal PN practice guidelines for this population.

Received for publication January 26, 2006. Accepted for publication April 11, 2006.

1. Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med.2001; 345:1359 –1367.[Abstract/Free Full Text]
2. Van den Berghe G, Wilmer A, Hermans G, et al. Intensive insulin therapy in the medical ICU. N Engl J Med.2006; 354:449 –461.[Abstract/Free Full Text]
3. McCowen K, Malhotra A, Bistrian B. Stress-induced hyperglycemia.Crit Care Clin. 2001;17:104 –124.
4. Bower RH, Talamini MA, Sax HC, Hamilton F, Fischer JE. Postoperative enteral vs. parenteral nutrition. Arch Surg.1986; 121:1040 –1045.[Abstract/Free Full Text]
5. Gonzalez-Huix F, Fernandez-Banares F, Esteve-Comas M, et al. Enteral vs. parenteral nutrition as adjunct therapy in acute ulcerative colitis. Am J Gastroenterol.1993; 88:227 –232.[Web of Science][Medline] [Order article via Infotrieve]
6. Kalfarentzos R, Kehadias J, Mead N, Kokkinis K, Gogos CA. Enteral nutrition is superior to parenteral nutrition in severe acute pancreatitis: results of a randomized prospective trial. Br J Surg.1997; 84:1665 –1669.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
7. McClave SA, Greene LM, Snider HL, et al. Comparison of the safety of early enteral vs. parenteral nutrition in mild acute pancreatitis.JPEN J Parenter Enteral Nutr.1997; 21:14 –20.[Abstract/Free Full Text]
8. Hamaoui E, Lefkowitz R, Olender L, et al. Enteral nutrition in the early postoperative period: a new semi-elemental formula versus total parenteral nutrition. JPEN J Parenter Enteral Nutr.1990; 14:501 –507.[Abstract/Free Full Text]
9. Iovinelli G, Marsil I, Varrassi G. Nutrition support after total laryngectomy. JPEN J Parenter Enteral Nutr.1993; 17:445 –448.[Abstract/Free Full Text]
10. The VA Total Parenteral Nutrition Cooperative Study Group. Perioperative total parenteral nutrition in surgical patients: a VA cooperative study. N Engl J Med.1991; 325:525 –532.[Abstract]
11. Sheean PM, Freels SA, Helton SW, Braunschweig CA. Adverse clinical consequences from total parenteral nutrition exposure during hematopoietic stem cell transplant. Biol Blood Marrow Trans. In press.
12. Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/epiinfo/Epi6/ei6.htm. Accessed May 20, 2005.
13. SAS Institute. SAS/STAT User's Guide (Version 8). Cary, NC: SAS Institute; 2000.
14. Braunschweig C, Levy P, Sheean P, Wang X. Enteral versus parenteral nutrition: a meta-analysis. Am J Clin Nutr.2001; 74:534 –542.[Abstract/Free Full Text]
15. Lim STK, Choa RG, Lam KH, Wong J, Ong GB. Total parenteral nutrition versus gastrostomy in the preoperative preparation of patients with carcinoma of the oseophagus. Br J Surg.1981; 68:69 –72.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
16. Cheung NW, Napier B, Zaccaria C, Fletcher JP. Hyperglycemia is associated with adverse outcomes in patients receiving total parenteral nutrition. Diabetes Care.2005; 28:2367 –2371.[Abstract/Free Full Text]
17. Matarese L. Metabolic complications of parenteral nutrition therapy. In: Gottschlich MM, ed. The Science and Practice of Nutrition Support: A Case-Based Core Curriculum. Dubuque, IA: Kendall/Hunt Publishing Co; 2001:269 –286.
18. Shils ME, Brown RO. Modern Nutrition in Health and Disease. 9^th ed. Baltimore, MD: Williams & Wilkins;1999 .
19. Rosmarin DK, Wardlaw GM, Mirtallo J. Hyperglycemia associated with high, continuous infusion rates of total parenteral nutrition dextrose.Nutr Clin Pract. 1996;4:151 –156.

Journal of Parenteral and Enteral Nutrition, Vol. 30, No. 4, 345-350 (2006)
DOI: 10.1177/0148607106030004345


CiteULike    Connotea    Del.icio.us    Digg    Reddit    Technorati    What's this?

This article has been cited by other articles:

				J. L. Thompson and J. Duffy Nutrition Support Challenges in Hematopoietic Stem Cell Transplant Patients Nutr Clin Pract, October 1, 2008; 23(5): 533 - 546. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Citing Articles via Scopus Google Scholar Articles by Sheean, P. Articles by Braunschweig, C. Search for Related Content PubMed PubMed Citation Articles by Sheean, P. Articles by Braunschweig, C. Pubmed/NCBI databases Compound via MeSH Substance via MeSH Hazardous Substances DB GLUCOSE Social Bookmarking                   What's this?