In adult patients who are critically ill, what is the effect of enteral nutrition (EN) vs. parenteral nutrition (PN) on mortality?

Conclusion Statement

It is unclear whether there is a difference in mortality rate when comparing EN vs. PN in critically ill adult patients.

Evidence Summary

In a positive research quality prospective cohort study with 2,920 medical and surgical patients from 260 intensive care units in 28 countries, Kutsogiannis J et al 2011 studied 2,569 subjects who received early EN, 188 who received early supplemental PN, and 170 received late supplemental PN. 60-day mortality rate was significantly lower in those receiving early EN (27.8%) compared to those receiving early PN (34.6%) or late PN (35.3%) (P=0.02).
In a positive quality meta-analysis, Simpson F and Doig GS, 2005 included six studies for early EN with 224 subjects and five studies for late EN with 121 subjects. When PN was compared to early EN (started within 24 hours), there was no effect on mortality (OR 1.07, 97% CI 0.39-2.05, P=0.89). However, mortality was significantly less with PN compared to delayed EN (start time greater than 24 hours) (OR 0.29, 95% CI 0.12-0.70, P=0.006).
Five studies of positive quality demonstrated no difference in mortality when EN was compared to PN (Abou-Assi et al, 2002; Kalfarentzos et al, 1997; Borzotta et al, 1994; Kudsk et al, 1992; Young et al, 1987). These studies may have been underpowered to determine the effect of EN vs. PN on mortality, as each study group included fewer than 50 patients.
One neutral quality (Moore et al, 1992) and two positive quality (Heyland et al, 2003; Braunschweig et al, 2001) meta-analyses reported no difference in mortality when EN was compared to PN
Three meta-analyses showed no difference in mortality in EN vs. PN:
- Moore et al, 1992, evaluated mortality at 10 days and 30 days in 230 patients (118 received EN and 112 received PN) and found no significant difference in mortality between the groups
- Heyland et al, 2003, reported no difference in mortality in EN vs. PN in a review of 13 PRCTs (RR of 1.08; 0.7-1.65)
- Braunschweig et al, 2001, reviewed 20 studies (508 EN and 525 PN patients) and found no difference in mortality among patients who received EN vs. PN
Five studies of neutral quality (Adams et al, 1986; Hadley et al, 1986; Hadfield et al, 1995; Cerra et al, 1988; Woodcock et al, 2001) also reported no difference in mortality among patients supported with EN vs. PN. The absolute number of deaths in each study was very small and mortality may not have been related to the mode of nutrition support (e.g., Woodcock et al, 2001, reported no deaths from PN related complication but one death occurred from peritonitis, following PEG tube placement).

View full table in new window

Author, Year, Study Design, Class, Rating	Illness Severity	Number of Subjects / Actual Sample	Review Studies: Inclusion and Exclusion Criteria	Age	Sex (Percent Male)	Description of Sample	Formulas and Rate of Delivery	Timing of Feeding Per Study Protocol	Mortality Outcome
Abou-Assi S, Craig K et al, 2002 Study Design: Randomized Controlled Trial Class: A Rating:	Acute pancreatitis, not otherwise specified.	N=156	Not applicable	21 to 91 years	EN group: 61.5% PN group: 48%	Acute pancreatitis patients: 117 received no nutrition support (Group O) 20 received EN 19 received PN	Not applicable	Not applicable	No difference between groups
Adams 1986 Study Design: Randomized Controlled Trial Class: A Rating:	Significant injuries to two or more body systems, requiring emergent laparotomy.	N=46	Not applicable	18 to 60 years	EN group: 65% PN group: 69.5%	Trauma: EN=23 PN=23	Not applicable	Not applicable	EN group: One PN group: Three
Borzotta AP. 1994 Study Design: Randomized Controlled Trial Class: A Rating:	Glasgow Coma Scale of eight or less.	N=49	Not applicable	Not applicable	EN group: 30% PN group: 95%	Head injury: EN=28 PN=21	Not applicable	Not applicable	PN group: One EN group: Five
Braunschweig CL. 2001 Study Design: Meta-analysis or Systematic Review Class: M Rating:	Not applicable	27 studies met criteria	Inclusion: Only PRCTs that evaluated the effect of PN administrated at or above estimated energy needs compared with those of tube feeding or standard care on outcomes with clinical significance were reviewed. Exclusion: Studies only evaluating nutritional outcomes.	Not applicable	Not applicable	Malabsorptive syndromes, motility disorders, mechanical obstruction not immediately remedied by surgery, pre-operative state with severe under-nutrition, critically ill patients (especially those with hypermetabolism).	Not applicable	Not applicable	No difference between EN and PN.
Cerra 1988 Study Design: Randomized Controlled Trial Class: A Rating:	Persistent hypermetabolism four to six days after sepsis and surgery.	N=66	Not applicable	EN group: 56±15 PN group: 55±11 years	60%	Sepsis: EN=31 PN=35	Not applicable	Not applicable	No differences in mortality.
Hadfield RJ. 1995 Study Design: Randomized Controlled Trial Class: A Rating:	EN group: 16.9±1.2 PN group: 13.3±1.2	N=24	Not applicable	Over 18 years	71%	ICU: EN=13 PN=11			NS different for mortality: EN group: Two PN group: Six; P=0.08
Hadley MN et al 1986 Study Design: Randomized Controlled Trial Class: A Rating:	Glasgow Coma Scale of 10 or lower at six hours post-injury.	N=43	Not applicable	Median age: 27 years old for males 34 years old for females	89%	Head injury: EN=21 PN=24	Not applicable	Not applicable	EN group: Three of 21 PN group: Two of 24
Heyland DK, Dhaliwal R et al, 2003 Study Design: Meta-analysis or Systematic Review Class: M Rating:	Not applicable	14 RCTs	Inclusion: Mechanically ventilated patients. Exclusion: Pseudo-randomized trials, elective surgery patients and data reported only in abstracts, were excluded. Mechanically ventilated patients receiving either EN or PN.	Not applicable	Not applicable	ICU patients, trauma, burn	250ml: If a feeding protocol is used, 250ml gastric residual volume may optimize formula delivery. Motility agents should be considered as a strategy to improve delivery of EN.	Not applicable	Immune enhancing: NS EN vs. PN: No difference in mortality (RR=1.08; 95% CI 0.70-1.65; P=0.7).
Kalfarentzos F 1997 Study Design: Randomized Controlled Trial Class: A Rating:	APACHE II score of eight or more.	N=38	Not applicable	EN group: 23 to 70 years PN group: 31 to 72	EN group: 80% PN group: 54%	Acute pancreatitis: EN=18 PN=20	Not applicable	Not applicable	EN group: Three of 18 PN group: Two of 20
Kudsk KA. 1992 Study Design: Randomized Controlled Trial Class: A Rating:	ATI at least 15	N=96	Not applicable	EN group: 30.4±1.7 PN group: 30.6±1.4 years	Not stated	Trauma: EN=51 PN=45	Not applicable	Not applicable	EN group: One of 51 PN group: One of 45
Kutsogiannis J, Alberda C et al, 2011 Study Design: Prospective Cohort Study Class: B Rating:	Medical and surgical patients had APACHE scores (P=0.11): 22.0±7.9 early EN group 22.6±8.4 early supplemental PN group 23.3±7.9 late supplemental PN group	2,920 patients from 260 ICUs in 28 countries: 2,569 in received early EN 188 received early supplemental PN 170 received late supplemental PN	ICUs had minimum of eight beds and access to Registered Dietitian	Mean age: 58.4 years±17.9 early EN group 62.3 years±17.9 early supplemental PN group Late supplemental PN group, 56.4±17.5 years. (P=0.02)	61% male 39% female	Medical and surgical patients from 260 ICUs in 28 countries	Adequacy of calories and protein from total nutrition was highest in the early PN group (81.2% and 80.1%) and lowest in the early EN group (63.4% and 59.3%; P<0.0001).	Early EN and supplemental PN was within 48 hours of admission; late was defined as after 48 hours	60-day mortality rate was significantly lower in those receiving early EN (27.8%) compared to those receiving early PN (34.6%) or late PN (35.3%) (P=0.02)
Moore, FA, et al. 1992 Study Design: Meta-analysis or Systematic Review Class: M Rating:	Not applicable	N=8 PRCTs	Inclusion: Nutrition support initiation within 72 hours post-operative. Studies that used vivonex TEN or one of three study TPN solutions as initial post-operative feeding. Participants were moderately to severely stressed. All studies had daily documentation of complications. Exclusion: Pre-existing diseases, including advanced diabetes, chronic renal failure, cirrhosis and inflammatory bowel disease Condition precluding use of enteral nutrition: Severe head injury Conditions precluding aggressive nutritional support: Hospitalization of over 10 days before study enrollment; prior surgical procedures during hospitalization; pre-operative nutritional support; non-study nutritional solution used immediately after operation	Not applicable	Not applicable	Various high-risk surgical patient populations that by convention have been fed parenterally.	Not applicable	Not applicable	No differences in mortality
Simpson F and Doig GS, 2005 (meta-analysis) Study Design: Meta-analysis or Systematic Review Class: M Rating:	Reported only as commensurate with patients being critically ill	N=6 for early EN with 224 subjects N=5 studies for late EN with 121 subjects	Inclusion: PN and EN did not contain supplemental glutamine, arginine or other immune enhancing ingredients; published in English language priot to April 30, 2003, patients required mechanical ventilation, had average ICU LOS more than two days, severity of illness reported and was commensurate with critical illness, patients suffered from a condition requiring care in ICU; articles were rated for methodological quality by intent to treat analysis, maintenance of allocation concealment during randomization and appropriate use of blinding. Exclusion: Papers were excluded if they did not meet criteria for patient inclusion, article quality or had more than 10% loss of patients lost to follow-up	Not described	Not described	Critically ill subjects included medical, surgical and burn patients.	EN or PN	Early defined as within 48 hours of admission to ICU	When PN compared to early EN (started within 24 hours), no effect on mortality (OR 1.07, 97% CI 0.39-2.05, P=0.89) Mortality significantly less with PN compared to delayed EN (start time greater than 24 hours) (OR 0.29, 95% CI 0.12-0.70, P=0.006)
Woodcock et al 2001 Study Design: Randomized Controlled Trial Class: A Rating:	All patients who required adjuvant nutritional support	M=498	Not applicable	At least 18 years	EN group: 50% PN group: 58%	Sepsis: EN=231 PN=267	Not applicable	Not applicable	No difference in mortality in patients randomized to receive EN or PN. EN group: 37.5% PN group: 21.9%
Young B, et al 1987 Study Design: Randomized Controlled Trial Class: A Rating:	Glasgow Coma Scale (GCS) score between four and 10 during the first 24 hours after admission.	N=51	Not applicable	Mean Age EN group: 34.0±2.92 TPN group: 30.3±2.67 years	82%	Head injury: EN=28 PN=23	Not applicable	Not applicable	No differences in mortality