University of Virginia Health System
Nutrition Support E-Journal Club
Our April trainees were able to enjoy Charlottesville in it's spring glory - warm days, cool nights - flowers and greenness everywhere. This month our trainees travelled from Minnesota, Michigan, and British Columbia. Our Journal Club was held in conjunction with our semi-annual Nutrition Support Forum, where dietitians from surrounding hospitals join us to discuss interesting cases, as well as recent literature....so this month we will summarize two interesting articles.
April Citation 1:
Reid CL. Poor agreement between continuous measurements of energy expenditure and routinely used prediction equations in intensive care unit patients. Clin Nutr. 2007;26(5):649-57.
This study measured energy expenditure with continuous indirect calorimetry (24 hour per day) in 27 patients requiring mechanical ventilation. Indirect calorimetry measurement was started within 48 hours of admission to the ICU and continued for 5 days. The daily variation in total energy expenditure (TEE) was reported and the measured expenditure was compared to predicted energy expenditure from Harris-Benedict, Schofield and Ireton-Jones equations and the American College of Chest Physicians recommendation of 25 kcal/kg/day. Energy expenditure was calculated for obese (BMI > 30) and underweight patients (BMI < 19) using an adjusted body weight using a 50% correction factor (average of actual and ideal weight). A 30% "stress factor" was added to the Harris-Benedict and Schofield estimates of resting energy expenditure.
Patients received nutrition support (enteral (EN) or parenteral (PN)) by continuous infusion throughout the study. The volume of feed prescribed was modified each day, based on the indirect calorimetry measurement from the previous 24-hours. The investigators attempted to feed all patients enterally, but those who failed to meet energy needs with EN alone received supplemental PN until EN provided >50% of TEE.
Accurate prediction of energy expenditure was defined as estimates that were within 80% and 110% of actual calorie expenditure, as measured by indirect calorimetry. The difference between measured and estimated energy expenditure were calculated daily for each patient and for each of the equations. In addition, the investigators reported the cumulative energy balances that would have arisen had patients been fed according to the prediction equation.
Inclusion and Exclusion Criteria were:
The inclusion criteria were patients admitted to the general ICU requiring mechanical ventilation for 5 days.
Exclusion criteria were respiratory quotient (RQ) out of physiological range, or conditions that may have prohibited indirect calorimetry testing, such as fraction of inspired oxygen (FiO2) >0.80, or a positive end-expiratory pressure (PEEP) 20 mmHg.
Major Results reported by authors:
The average day-to-day variability in TEE was 31.7 (±22.6)%. Medical patients had a numerically higher daily variation in energy expenditure (40.7 ±33.6%) than the surgical (26.2 ±11.0%) and trauma patients (26.2 ±11.0%), but this difference was not statistically different.
The Harris-Benedict, Schofield and ACCP equations provided most estimates (66%, 66% and 65%, respectively) between 80% and 110% of TEE. However, these equations would have resulted in underfeeding (<80% of TEE) in 16%, 15% and 22% of patients, respectively, and overfeeding (>110% of TEE) in 18%, 19% and 13% of patients, respectively. Using adjusted weight in the equations tended to reduce the number of ‘adequate' estimates (80-110% of TEE) and increase the frequency of underfeeding (< 80% of TEE).
The mean cumulative energy balances that would have resulted from the use of the prediction equations for 5 days were:
- Schofield - 604 (- 6493 to + 3919)
- Harris-Benedict - 789 (- 6203 to + 4219)
- Ireton-Jones - 1003 (- 8709 to + 5951)
- 25 kcal/kg - 1304 (- 6453 to + 2115).
The actual mean cumulative energy balance, with orders adjusted daily via indirect calorimetry was - 883 (-6702 to + 4791).
"Limits of agreement between the different equations and total energy expenditure values were unacceptably wide. Prediction equations may result in significant under- or overfeeding in the clinical setting."
This study addresses many of the flaws in existing research on caloric requirements in critically ill patients, namely, that doing a single indirect calorimetry measurement at a random point of an ICU stay and assuming that the single measurement is an accurate representation of a patient's caloric expenditure. The use of continuous indirect calorimetry in this study provides a more representative picture of calorie expenditure in critically ill patients, and confirms the high day to day variation in calorie expenditure. The results of this study suggest that the use of a single indirect calorimetry measurement to develop a nutrition order may lead to a greater cumulative calorie error than the use of many prediction equations.
The small number of patients in this study is one of the limitations to this study. The other limitation is that all measurements were done during the first week of admission, and the results may not apply to patients that have extended ICU stays, when they become more stable in subsequent weeks.
Perhaps the most important limitation of this study, which was addressed wonderfully in the discussion section (note shameless prod to read entire study) is that the definition of "adequate feeding" of 80-100% of energy expenditure is a complete assumption because there are NO randomized studies that tell us the level of feeding that results in the best patient outcome.
From a practical standpoint there is also no solid data to indicate at what point underfeeding or overfeeding presents a clinically significant problem - it may be that prediction equations provide sufficient accuracy to prevent any harm to the patient - especially considering the day to day variability in how much nutrition patients actually receive in the ICU setting!
Our Take Home message:
Harris-Benedict, Schofield, and Ireton-Jones and 25 kcals/kg have similar error, as defined by 80-110% energy expenditure, during the first week of admission in critically ill patients. It is unclear if this magnitude of error has any clinical significance. There are no patient outcome data to provide support for the use of one prediction equation over another.
April Citation 2:
Besselink MG, van Santvoort HC, Buskens E, et al.; Dutch Acute Pancreatitis Study Group. Probiotic prophylaxis in predicted severe acute pancreatitis: a randomized, double-blind, placebo-controlled trial. Lancet. 2008;23;371(9613):651-9.
This was a multi-center, double blind study of 298 adult patients with severe acute pancreatitis randomized to receive either a multispecies probiotic or placebo. Patients received either Ecologic 641 (Winclove Bio Industries, Amsterdam, Netherlands, containing Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus salivarius, Lactococcus lactis, Bifi dobacterium bifi dum, and Bifi dobacterium lactis in a total daily dose of 10¹° bacteria, plus cornstarch and maltodextrins) or placebo twice per day no later than 72 h after onset of symptoms of pancreatitis continuing for 28 days. All patients had a nasojejunal tube (NJ) placed for feeding, study medication or placebo, and had a 1 Kcal/mL polymeric, fiber-enriched formula advanced to goal feeding (~ 30 kcals/kg) over 4 days. When patients started oral intake the NJ was removed and the probiotic/placebo was dissolved in water and taken orally until day 28. Administration of the study product was stopped if a patient developed infected pancreatic necrosis; antibiotic prophylaxis was not given routinely in patients with necrotizing pancreatitis.
The primary endpoint for the study was the composite of infectious complications including: infected pancreatic necrosis, bacteraemia, pneumonia, urosepsis, or infected ascites, during admission and 90-day follow-up. Secondary endpoints were mortality, sequential organ failure assessment (SOFA) scores, organ failure during admission, onset of organ failure after randomization, need for surgical intervention because of infected necrosis or intra-abdominal catastrophe, hospital stay, intensive-care stay, use of antibiotics, and abdominal complaints (nausea and abdominal fullness with visual analogue scales [VAS; cutoff 3·0 on a 10-point scale], and presence of diarrhea as assessed by the patient [at days 5, 10, 14, 21, 28, and 35]).
Inclusion and Exclusion Criteria were:
Inclusion criteria were adult patients admitted with a first episode of acute pancreatitis (defined as abdominal pain in combination with serum amylase or lipase concentrations that were raised to at least three times the institutional upper limit of normal with an Acute Physiology and Chronic Health Evaluation (APACHE II) score of 8 or more, Imrie/modified Glasgow score of 3 or more or C-reactive protein over 150 mg/L).
Exclusion criteria were pancreatitis after ERCP; suspected malignancy of the pancreas or biliary tree; non-pancreatic infection or sepsis caused by a second disease; diagnosis of pancreatitis first made at operation; or a medical history of immune deficiency.
Major Results reported by authors:
There were no significant differences in the occurrence of the primary endpoint (all infections) between the two groups, nor were there any significant differences between the groups for any single infection.
There were significantly more deaths in the probiotics group than there were in the placebo group (24 vs 9) (p=0·01; relative risk 2·53, 95% CI 1·22-5·25). Most of the deaths were caused by multiorgan failure: 20/24 (83%) of those in the probiotics group and 7/9 (78%) of those in the placebo group.
Bowel ischemia was detected during operation or autopsy in nine patients in the probiotics group; eight of these patients died as a result. No cases of bowel ischemia were seen in the placebo group (p=0·004). The nine cases of bowel ischemia were all diagnosed within the first 14 days of admission in seven different hospitals; four university and three teaching hospitals. All of these patients had early onset of organ failure (median 2 days after admission, range 1-6 days).
"In patients with predicted severe acute pancreatitis, probiotic prophylaxis with this combination of probiotic strains did not reduce the risk of infectious complications and was associated with an increased risk of mortality."
"Probiotics can no longer be considered to be harmless adjuncts to enteral nutrition, especially in critically ill patients or patients at risk for non-occlusive mesenteric ischaemia."
This was a study with excellent methodology - large number of patients, randomized, double-blind and conducted in multiple centers.
It is a classic example of why we cannot be cavalier with "nutrition" interventions in the critically ill population and why it is necessary that our interventions be tested in large randomized studies BEFORE we begin to utilize them. It may also be an example of how small studies may suggest a benefit (Olah 2002, n= 45 patients),1 but adequately powered studies expose the full effects of an intervention.
Unfortunately there is no way to know if the results of this study apply to all probiotics, or if there is some particular strain or combination of probiotic strain that produces negative effects in this population.
The unexpected results (increased mortality from probiotics) of this study have been widely reported in other journals, and will likely have an influence on approval and safety monitoring of studies for years to come.
Our Take Home message:
Probiotics should not be used in severe acute pancreatitis (or critical illness for that matter) until adequate studies demonstrate safety and effectiveness. It would be optimistic to the point of foolishness to think that this lesson did not apply to other aspects of nutrition support in critically ill patients - we need to demand adequate safety and effectiveness data before nutrition interventions are used in critically ill patients.
1Oláh A, Belágyi T, Issekutz ME, et al. Randomized clinical trial of specific lactobacillus and fibre supplement to early enteral nutrition in patients with acute pancreatitis. Brit J Surg. 2002 ;89:1103-1107.
Our next Weekend Warrior 2 day mini-traineeship program is scheduled for Saturday and Sunday, May 31 - June 1. If you know anyone who might not be able to get away for our full week traineeship, please let them know about our weekend program, and to check out our website for full information-we have a few slots left!
Check out the latest Practical Gastroenterology articles/info at:
1) Schiller LR. Nutrients and Constipation: Cause or Cure? Practical Gastroenterology 2008;XXXII(4):43.
Joe Krenitsky MS, RD
Carol Parrish RD, MS
PS - Please feel free to forward this on to friends and colleagues.