June 2015 E-Journal Club

June 1, 2015 by School of Medicine Webmaster

Greetings,We do not host a traineeship in June so that we can start taking turns travelling to the beach, mountains, or just hanging close to home where the only “tube feeding” we are concerned about is a drink from a pineapple, CamelBak or garden hose, respectively.  However, we did make time for journal club in June – our article is a long-awaited multi-center study of hypocaloric feeding in the adult ICU.

June Citation:

Arabi YM, Aldawood AS, Haddad SH, et al. Permissive Underfeeding or Standard Enteral Feeding in Critically Ill Adults. N Engl J Med. 2015 May 20. [Epub ahead of print]


This was a multicenter, randomized, unblinded study of 894 adult ICU patients comparing hypocaloric enteral nutrition (EN) with intending to meet full calorie expenditure for a maximum of 2 weeks.  The hypocaloric feeding group was scheduled to receive 40-60% of estimated calorie expenditure and the standard feeding group 70-100% of calorie expenditure (as estimated by the Penn State equation or Ireton-Jones for obese or non-ventilated patients).  The hypocaloric group received a protein supplement and water flushes so that the volume and protein provision were similar between groups and daily enteral multivitamins were recommended for both groups.  Calories provided via IV fluids and medications were accounted for during the study.

This was a pragmatic study because although the study protocol offered guidelines, the choice of feeding formula, exact protein dose (range 1.2-1.5 gm/kg), multivitamin and insulin regimen were left to the discretion and protocols of the treating team.

The primary outcome was 90-day mortality, with a large number of secondary and tertiary outcomes including ICU, hospital, 28 and 180 day mortality, SOFA scores, LOS, infections and feeding tolerance and prealbumin, among other factors.

Inclusion and Exclusion Criteria were:

Inclusion criteria:

ICU patients aged ≥ 18 years, receiving enteral feeding within 48 hours of admission and expected to stay 72 hours or more in the ICU.

Exclusion criteria:

Lack of commitment to ongoing life support (terminal Illness, DNR

order in the first 48 hours and brain death within 48 hours of admission), pre-existing condition with expected 6 month mortality > 50% EN cannot be started within 48 hours of admission, receiving PN or oral feeding, previously enrolled in the study within the same admission, pregnancy, post-liver transplant, cardiac arrest, burn patients, prisoners, age >80 years, patients on “significant” ionotropic support (except dobutamine) at following doses:

Norepinephrine >0.4μg/kg/min, Epinephrine >0.4μg/kg/min, Dopamine >10μg/kg/mi, Phenylephrine >150 μg/min, Vasopressin >0.02 unit/min or half of the above doses for patients on more than one inotrope.

Major Results:

The investigators screened 6337 patients, randomized 894 patients and analyzed data from 885 patients (14% of those screened were in the study).  The groups were well matched with no obvious imbalance between the groups after randomization. The majority of patients (75%) had a medical diagnosis with 3-4% surgical and 21-22% trauma with 96-97% of all patients requiring mechanical ventilation.  The average length of the study intervention was just over 9 days and median length of stay was 13 days in the ICU and 29 in the hospital.

The hypocaloric group actually received an average of 835 ±297 kcals/24 hours (46% of expenditure) and the standard group received 1299 ±467 kcals/24 hours (71% of expenditure).  There was no significant difference in protein provision between the groups, with mean protein intake of 58 ±24 gm/day (0.7 gm protein/kg/day).  There was a statistically significant difference in fluid balance between the groups with the hypocaloric group having a +490 ±1408 ml/day balance and the standard group +688 ±1196 ml/day balance (p<0.001).

There was no significant difference between the groups in the primary outcome of 90-day mortality (27.2% hypocaloric, 28.9% standard, p=0.58).  There was also no significant difference between the groups for any kind of mortality, LOS, mechanical ventilation or SOFA score endpoints.  There was no significant difference in nitrogen balance, body weight, C-reactive protein, serum proteins (including prealbumin), or other labs between the groups.  Average blood glucose was statistically significantly different between groups (163.8 mg/dl [9.1 mmol/l] hypocaloric versus 169.2 mg/dl [9.4 mmol/L],p = 0.04), but it is unclear that this is clinically important.  The requirement for insulin therapy (45.8% hypocaloric versus 52.7% standard, p= 0.04) and total insulin dose was significantly reduced in the hypocaloric compared to the standard group (7 units/24 hours difference, p = 0.02).

There were also no significant differences in 90-day mortality between any of the pre-specified subgroups (surgical/nonsurgical, DM/no DM, APACHE > 18/<18, severe sepsis/not severe sepsis, traumatic brain injury/no TBI, vasopressors/no vasopressors, serum glucose > 165 mg/dl on randomization/< 165 mg/dl).

Author’s Conclusions:

The authors concluded that their results demonstrate that “a strategy of enteral feeding to provide a moderate amount of calories to critically

ill adults in the presence of full protein intake was not associated with lower mortality than a strategy aimed at providing a full amount of calories.”


This was a well-designed study with randomization and concealed allocation into groups.  The authors review a number of the limitations in the discussion section, including the limited generalizability of this study because only 14% of their ICU patients were represented.  Patients had to have EN started with 48 hours after admission, so it is unclear if these results would apply to patients that do not receive nutrition support for 3-4 days after admission (and who are most likely sicker).  We also noted that just over 2/3 of these patients came from 1 hospital, which also may limit the generalizability of this study.

One of the factors that we discussed was the fact that protein provision was so low in both groups for critically ill patients (0.7 gm protein/kg/day).  Especially for patients receiving hypocaloric feeding – the study patients were receiving less than 50% of what we aim to provide our critically ill adults.  Ultimately this was a study of hypocaloric, inadequate protein feeding compared with slightly hypocaloric, inadequate protein feeding.

Another concern that our group had was the lack of accounting for how many patients actually received the recommended multivitamin.  With hypocaloric amounts of EN the micronutrient provision is usually very low; we also noted that only 2 of the listed supplements used provided any minerals – so a number of patients may have received very low intakes of nutrients such as zinc.

Although this study had a number of pre-specified subgroups we were a bit surprised that there was no accounting for the nutrition status of the patients on admission, and no analysis of different BMI subgroups (obese versus lean patients) or malnourished versus well-nourished.  This study did not exclude malnourished patients, and it would be helpful to know if some groups such as those with a very low BMI or frank malnutrition, may have had a negative response to hypocaloric feedings, while patients with increased BMI may have had a more favorable response.

Despite the limitations, this study does add to the growing body of data that suggests that a modest calorie deficit (460 kcals/day) does not seem to negatively affect patient outcome, or even nitrogen balance.  The hypocaloric intervention was continued longer than in the EDEN study (1) and patients were more diverse and in the ICU for a longer period.

However, the majority of these patients were medical ICU patients, and the average BMI of these patients was 29 kg/m2, so what this study does not address are severely malnourished or very lean patients, or patients with very large wounds or those that require multiple surgical interventions during an admission.

The “hidden” data with this study is about serum proteins.  There were no differences in serum transferrin or prealbumin between groups, so this study adds to the evidence that prealbumin is not a marker of nutrition adequacy in the IU patient.

Our Take Home Message (s)

1.   There is a growing body of data that suggest that modest differences in calorie provision do not affect patient outcome, that there may be advantages in terms of glucose control from a more hypocaloric regimen.

2.   This study adds to the data that makes it difficult to make a credible argument that a particular calorie calculation or indirect calorimetry for the first 2 weeks of the ICU admission could possibly make a difference in patient outcome.

3.   It also begs the question whether “make up” or pep-up feedings are necessary or in our patients best interest.

4.   If you are thinking of checking prealbumin as an indicator of nutrition adequacy in an ICU patient, please just flip a coin instead because it will be just as accurate and will be far more time, resource, and cost-effective.


  1.  The National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network.  Initial Trophic vs. Full Enteral Feeding in Patients with Acute Lung Injury:  The EDEN Randomized Trial.  2012;307(8):795-803.


Other News on the UVAHS GI Nutrition Website: (www.ginutrition.virginia.edu):

Latest Practical Gastroenterology article:

Sapone A, Leffler D, Mukherjee R. Non-Celiac Gluten Sensitivity Where are We Now in 2015? Practical Gastroenterology 2015;XXXIX(6):40.


Joe Krenitsky MS, RD

Carol Rees Parrish MS, RD


PS – Please feel free to forward on to friends and colleagues.