Enteral Nutrition

Why Enteral Nutrition?

Complete bowel rest is accompanied by progressive atrophy of the intestinal mucosa, which is evident after only a few days [Ann Surg 205: 681, 1987]. Glutamine is considered the principal metabolic fuel for intestinal epithelial cells and may be especially important in this regard [Nutrition 6: 199, 1990]. Lastly, bacterial translocation has been documented during periods of bowel rest [Curr Opin Clin Nutr Metab Care 5: 679, 2002; Best Pract Res Clin Gadstroenterol 17: 397, 2003], thus enteral nutrition may protect against bacteremia [Mainous MR: The Gut Barrier, Mosby 557, 1994]

Patient Selection – indications and contraindications

In the absence of contraindications, start enteral feeds in any patient with 5 or more days of inadequate intake for most patients. For those at high risk of translocation (ex. burn victims), start feeds as soon as nutrition input is inadequate [J Parent Ent Nutr 17S: 1SA, 1993; Gastroenterology 108: 1280, 1995]. Absolute contraindications include circulatory shock, intestinal ischemia, complete bowel obstruction, or ileus. Total enteral nutrition should not be used in include partial obstruction, severe diarrhea, pancreatitis, or high-volume (> 500 mL) enterocutaneous fistulas, although partial enteral support may be possible [J Parent Ent Nutr 17S: 1SA, 1993]. In pancreatitis, feeds can be delivered directly to the jejunum.

Feeding Tubes

The currently favored feeding tubes are 8 – 10 French and more flexible, however a rigid stylet is necessary for insertion [Heart Lung 22: 213, 1993]. “Standard” feeding tubes (14 – 16 French) are no longer favored as they cause significant patient discomfort. There was also a concern about gastroesophageal reflux with standard tubes but this appears to be unfounded [Am J Resp Crit Care Med 149: 1659, 1994]

Tubes should be inserted through the nose. The distance of insertion can be approximated by measuring from the tip of the nose to the earlobe and then to the xiphoid process [Rombeau JL: Atlas of nutritional support techniques: Little, Brown: 77 1989]. In patients not on H2-blockers or PPI’s, proper placement is suggested determined by measuring the pH of an aspirated specimen (if pH < 5, you are likely in the stomach) [Nutr Clin Pract 9: 185, 1994]. Still, because of the devastating sequelae of tracheal intubation, an X-ray is recommended.

The major complication of NG tubes is tracheal placement [Heart Lung 22: 213, 1993]. ACCIDENTAL INTUBATION OF THE TRACHEA IS OFTEN ASYMPTOMATIC. Furthermore, auscultation of the upper abdomen is not a reliable method for excluding tube malposition [Heart Lung 22: 213, 1993]. Therefore, A CHEST X-RAY IS MANDATORY.

Some prefer passing tubes into the duodenum, which can sometimes be accomplished with bedside maneuvers but often requires fluoroscopy. Passage into the duodenum can be confirmed when pH rises to > 6.0 (from < 5.0) [Am J Resp Crit Care Med 149: 1659, 1994] or radiographically. The risk of aspiration in duodenal feedings is the same as for gastric feedings [J Am Coll Nutr 10: 209, 1991; J Parent Ent Nutr 16: 59, 1992], thus the increased effort to place a duodenal tube does not seem to be worth it. For further proof: 74 patients with closed head injury (GCS avg 5.2). Transpyloric passage achieved in 43%, 57% remained intragastric. There were no differences between the groups in days to full feeding, ventilator days, ICU LOS, or incidence of pneumonia, or aspiration [J Trauma 39: 1100, 1995]. 101 ICU patients were randomized to receive NG vs transpyloric feeding. There were no differences in feeding duration, ICU length of stay, incidence of nosocomial pneumonia (40% vs. 32%), MODS at discharge, or mortality (43% vs. 38%). The only statistically significant difference was that the JEN group had fewer incidence of increased gastric residuals (2% vs 49%, p <.001) [Crit Care Med 30: 796, 2002]

Feeding Formulas

Caloric density is primarily determined by carbohydrate content, and most formulas provide 1-2 kcal/L (1.0 – 1.5 is standard, 1.5 – 2.0 is high-density). Use high-calorie formulas in volume-restricted patients and those with excessive daily energy needs.

Osmolality ranges from 280 – 1100 mOsm/kg and is also determined primarily by the carbohydrate content. Hypertonic solutions should be infused into the stomach to take advantage of the native gastric secretions.

Most formulas provide 35 – 40 g/L of protein. Those that are “HN” ie “high nitrogen” or “protein-rich” usually only contain 20% more protein. Some solutions (ex. Peptamen, Vital HN) provide small peptides instead of intact protein, which facilitates absorption. These formulas promote water reabsorption from the bowel as well, which may be beneficial in patients with diarrhea [Shronts EP: Nutrition Support Handbook, Harvey Whitney 147, 1992]

Lipids are generally long-chain triglycerides derived from vegetable oils. As excessive fat intake is associated with diarrhea, these are usually limited to 30% of total calories. Pulmocare (Ross) contains 55% of calories as lipids and is designed for patients in respiratory failure (proposed benefit = lower CO2 production). It has been shown that fatty acids derived from plant sources can serve as precursors to inflammatory mediators (eicasanoids) which can be harmful [Chest 100S: 182S, 1991]. Thus, Impact (Sandoz) and Perative (Ross) use fatty acids from marine oils, that former containing arginine as well (supposedly to improve immune competence [Surgery 112: 56, 1992]) and the latter containing beta-carotene.

There is emerging data to suggest that in ICU patients these modulatory feeds may be beneficial – In one study of 100 patients with acute lung injury, randomized to receive an enteral diet enriched with eicosapentaenoic acid (EPA), gamma-linolenic acid (GLA), and antioxidants for 14 days, a significant difference was found in length of ventilation (p < .04) in favor of the EPA + GLA group. There was no between-group difference in survival [Crit Care Med 34: 1033, 2006]. A prospective, double-blind, placebo-controlled, randomized trial in three different Brazilian ICUs. The study enrolled 165 patients with severe sepsis or septic shock, randomized to be continuously tube-fed with either a diet enriched with EPA, GLA, and elevated antioxidants or an isonitrogenous and isocaloric control diet, delivered at a constant rate to achieve a minimum of 75% of basal energy expenditure x 1.3 during a minimum of 4 days. Patients who were fed with the study diet experienced a significant reduction in mortality rate compared with patients fed with the control diet, the absolute mortality reduction amounting to 19.4% (p = .037). The group who received the study diet also experienced significant improvements in oxygenation status, more ventilator-free days (13.4 +/- 1.2 vs. 5.8 +/- 1.0, p < .001), more intensive care unit (ICU)-free days (10.8 +/- 1.1 vs. 4.6 +/- 0.9, p < .001), and a lesser development of new organ dysfunctions (p < .001) [Crit Care Med 34: 2325, 2006]


Glutamine is not an essential amino acid as it is produced by skeletal muscle, but it is the primary fuel of intestinal mucosa cells [Nutrition 6: 199, 1990]. Furthermore, tissue glutamine levels decline in acute, hypercatabolic states, thus in certain patients it may actually be an essential amino acid [Nutr Rev 48: 297, 1990]. All formulas contain glutamine but very little is in the free, unbound form. AlitraQ (Ross) seems to provide adequate free glutamine, other “enhanced” formulas provide some but not enough [J Parent Ent Nutr 14S: 137S, 1990] – a study of 185 trauma SICU patients given tube feeds showed no benefit in terms of mortality or any secondary endpoint given glutamine additives [Crit Care Med 33: 2501, 2005]

Fermentable fiber (cellulose, pectin, gums) are degraded by intestinal bacteria into short chain fatty acids which can be used by large bowel mucosa [Nutr Clin Pract 5: 99, 1990]. This fiber can slow gastric emptying and bind bile salts, and thus alleviate diarrhea. Non-fermentable fiber (lignins) are not degraded but provide an osmotic force that absorbs water from the bowel lumen, also reducing diarrhea. The actual effects of adding fiber have been inconsistent, however [J Parent Ent Nutr 18: 486, 1994; Am J Clin Nutr 50: 533, 1989]

Branched chain amino acids (leucine, isoleucine, and valine) are often used in trauma patients (sparing degradation of muscle proteins) or hepatic encephalopathy (antagonize uptake of aromatic amino acids) but are unproven.

Carnitine is necessary to transport FAs into the mitochondria, and its deficiency can cause cardiomyopathy, skeletal muscle myopathy, and hypoglycemia. The recommended daily dose of carnitine is 20-30 mg/kg [Nutrition 20: 709, 2004; Am J Nephrol 25: 106, 2005] A prospective, randomized, double-blind, controlled clinical trial of 369 ICU patients randomized to receive either the IMN Impact (includes arginine, purine nucleotides and omega-3) or an isocaloric, isonitrogenous control enteral feed. There was no significant difference in hospital mortality rate between the two groups (although Impact patients had higher APACHE-II scores on admission). Patients fed with the IMN had a significant reduction in their requirement for mechanical ventilation compared with controls (10.5 to 6.0 days, p = .007) with an associated reduction in the length of hospital stay (20 to 15.5 days, p = .03) [Crit Care Med 26: 1164, 1998]

Feeding Regimen

Tube feedings are usually run for 12-16 hrs in each day (continuous infusion overstresses the bowel mucosa, leading to malabsorption and diarrhea). To determine the infusion rate, infuse the desired hourly rate worth of water over one hour (not as a bolus, as this overestimates the residual), clamp the tube for 30 min and then aspirate the residual. If the residual volume is < 50% of the infusion and < 200 mL, feeding can proceed at this rate [Rombeau JL: Atlas of nutritional support techniques: Little, Brown: 77 1989]. If the residual is excessive, duodenal or jejunal feedings may be more appropriate.

Traditionally physicians have started patients on dilute formulas at low rates and then worked up, however studies show that beginning with full feeds does not result in vomiting or diarrhea [J Parent Ent Nutr 10: 258, 1986; J Crit Illness 10: 1116, 1993]. Starter regimens are therefore unnecessary.


Tube occlusion is often caused by protein precipitation [J Parent Ent Nutr 12: 403, 1988]. Standard preventive measures include flushing with 30 mL water every 4 hours and using 10 mL water flush every time medications are instilled [Nutr Clin Pract 5: 107, 1990]. In a partially obstructed tube (ie there is some flow), try injecting warm water. If this does not work, dissolve 1 tablet Viokase (pancreatic enzyme) and 1 table NaHCO3 in 5 mL water, inject, clamp for 5 min, then rinse with warm water. Last line is insertion of a wire, which may be dangerous?

Aspiration is reported in as many as 80% of patients receiving gastric or duodenal tube feeds – this can be reduced, but not eliminated, by keeping the head of the bed at 45 degrees [Heart Lung 22: 213, 1993]. These numbers are the same for gastric vs. duodenal feedings [J Paren Ent Nutr 26: S26, 2002]. Aspiration can be detected by testing tracheal aspirates with glucose oxidase reagent strips measured by an AccuCheck meter – > 20 mg/dL is evidence of aspiration [Chest 103: 117, 1993]. In the past, some physicians put food coloring into tube feeds but this method is too insensitive to be of much use.

Diarrhea occurs in 30% of these patients, but is usually not caused by hypertonicity. Rather, a sorbitol elixir used to improve palatability is often at fault [Am J Med 88: 91, 1990; Clin Intensive Care 3: 33, 1992]. Search for sorbitol as an ingredient and try to eliminate it before adjusting the osmolality of the tube feeds. C.dif can cause diarrhea as well, but can be differentiated from osmotic/sorbitol by the stool osmolal gap (measured osmolality – 2 x [Na] – [K]). If SOG is > 160, most likely the cause is osmolality/sorbitol. If < 160, favor C.dif.

Jejunostomy Feedings

Abdominal surgery generally produces 24 – 48 hours of gastric hypomotility, but small bowel motility is often unimpaired [Br J Surg 79: 1127, 1992], allowing immediate nutritional support. Needle jejunostomies can be placed 15 – 20 cm distal to Treitz during a laparotomy. As the jejunum does not have the reserve capacity of the stomach, starter regimens are recommended – begin with 15 – 25 mL/hr and increase over the next several days [Nutr Clin Pract 9: 101, 1994]. Flush catheters with 10 mL isotonic saline every 6 hours. The major complications are diarrhea and catheter occlusion – if feedings are required for > 1 week, a needle jejunostomy (small bore tube) can be converted to a standard jejunostomy (12 French tube) in order to reduce the incidence of occlusion [Am J Surg 164: 68, 1992]

Early vs. Late Feeding

Prior data on early vs. late feeding in critically ill patients have been difficult to interpret. A metaanalysis of 15 studies examining early enteral nutrition (containing 753 subjects) suggested that early enteral nutrition was associated with a significantly lower incidence of infections (RR 0.45; p =.00006) and a reduced hospital stay (mean 2.2 days; p =.004), however there were no significant differences in mortality or noninfectious complications. The results of this meta-analysis must be interpreted with some caution because of the significant heterogeneity between studies [Marik PE and Zaloga GP. Crit Care Med 29: 2264, 2001]

A prospective study of 52 trauma patients randomized to immediate (< 24h, n = 27) vs. delayed (> 24h, n = 25) intragastric EN. On day 4, the early EN group received a greater amount of feeding because of intolerance problems occurring in the delayed-EN group (1175 +/- 485 ml vs. 803 +/- 545 ml). 33% of the early EN patients and 64% of the delayed-EN patients met the criteria for pneumonia (p = 0.050) [Clin Nutr. 23: 527, 2004]. 30 neurosurgical patients GCS 4-10 were prospectively randomized to receive Impact with fiber < 72 hours after trauma or late (administered after gastric ileus resolved), continued for 14 days or until PO. There were no significant differences between the groups in length of stay, intensive care unit (ICU) days, significant infection, or GCS score. However, major infection correlated inversely with admission GCS score (R = -0.6, p < .003) [JPEN 24: 145, 2000]

Early vs. Late Parenteral Feeding

Seven Belgian ICUs recently performed The Early Parenteral Nutrition Completing Enteral Nutrition in Adult Critically Ill Patients (EPaNIC) Study, a prospective, randomized, controlled, parallel-group, multicenter investigator-initiated trial (partially funded by Baxter) in which 4640 nutritionally at-risk patients were randomized to early (within 48 hours) vs. late (at least 8 days) initiation of parenteral nutrition (2007 to 2010). While the early nutrition group clearly received more nutrition (both in terms of total energy and percentage of target), the late group was discharged from the ICU one day earlier (p = 0.04), had a lower incidence of hypoglycemia (p = 0.001), fewer infections (22.8% vs. 26.2%, p = 0.008), and a reduction in total healthcare costs of ~ $1600 (€1,110). Mortality was the same. Importantly, mean units of insulin and average glucose in the early and late groups were 58U/107 mg/dL and 31U/102 mg/dL, respectively [Casaer MP et al. NEJM Jun 29 2011 [Epub ahead of print]; FREE Full-text at New England Journal of Medicine]

Data on Feeding Location and Aspiration (Gastric vs. Post-Pyloric)

Early data suggested that gastric emptying was impaired in the TBI [JPEN 15: 298, 1991; J Neurotrauma 16: 233, 1999], that the use of gastric feeds delayed nutrition [Neurosurgery 25: 729, 1989], and that gastric feedings lead to a 3-fold increase in aspiration pneumonia as compared to TPN [Neurosurgery 17: 784, 1985]

More recent studies suggest that transpyloric and prepyloric feeds are identical. In one, 74 TBI patients with GCS averaging 5.2 and receiving enteral nutrition had NETs placed on average 5.6 days after admission (transpyloric in 43%, 10 of which required fluoroscopy, 57% remained intragastric). There were no differences in days to full feeding, ventilator days, ICU length of stay, or incidence of pneumonia or aspiration. The routine use of adjunctive measures to promote transpyloric passage was not particularly successful, had no obvious benefit, and therefore may not be necessary [J Trauma 39: 1100, 1995]

In another, a prospective, randomized multicenter study of 101 ICU patients started on enteral nutrition within 36 hrs, NG (51) vs. NGJ tubes (50) were no different in terms of feeding duration, ICU length of stay, or mortality. The only statistically significant difference was the incidence of increased gastric residuals (49% vs. 2%, p <.001). There was no difference in mortality, pneumonia, ICU length of stay, MODS at discharge. Both groups had a similar incidence of nosocomial pneumonia [Crit Care Med 30:796, 2002]. Similar results have been seen in the MICU (admittedly this is a different patient population) – a study of 60 patients were randomized to NG vs. NJ feedings (started at 30 mL/hr) showed that NG patients received nutrition sooner (11.2 hrs vs. 27.0 hrs), achieved goal rate sooner (28.8 hrs vs. 43.0 hrs) with no difference in aspiration events. [Crit Care Med 30: 1436, 2002]

A metaanalysis of 9 prospective randomized controlled trials (including a total of 522 patients) that compared gastric with post-pyloric feeding in MICU, neurosurgical and trauma patents. There were no significant differences in the incidence of pneumonia, percentage of caloric goal achieved, mean total caloric intake, ICU length of stay, or mortality. The time to initiation of enteral nutrition was significantly less in those patients randomized to gastric feeding. However, time to reach caloric goal did not differ between groups [Crit Care 7: R46, 2003]. Another meta-analysis including 637 patients came so similar conclusions [Ho KM et al. Intensive Care Med 32: 639, 2006]. A more recent RCT of gastric vs. post-pyloric feeds in 104 critically ill patients showed no difference in complication rates but earlier feeding and less energy deficit in the gastric group [White H et al. Crit Care 13(6): R187, 2009; FREE Full-text at PubMed Central]

Gastric feeds have traditionally been started at low volumes and increased slowly – full feeds may actually be preferable as shown in a study of 82 TBI patients on mechanical ventilation, randomized to receive tube feeds starting at 15 mL/hr or starting at the estimated energy and nitrogen requirements. Full feeding patients had a higher percentage of energy (p = .0008) and nitrogen (p<.0001) requirements met by EN in the first week after injury. Neurologic outcome at 6 months was similar between groups, but there was a tendency for more intervention patients to have a good outcome at 3 months (61% vs. 39%, p = .08). Fewer intervention patients had an infective complication (61% vs. 85%, p = .02) [Crit Care Med 27: 2525, 1999]

GRV measurements have never been validated. Use of low GRV has not been shown to decrease the incidence of aspiration. Elevated GRV is frequently associated with comorbidities such as vasopressor use, sedation, sepsis, and vomiting. GRV alone appears to have little independent clinical utility. McClave et. al. performed a prospective study of 40 critically ill patients, randomized to GRV >200 mL vs. GRV >400 mL to withhold feeding and studied using microspheres. There was no relationship between GRV and aspiration or regurgitation. In addition, there was no correlation between aspiration/regurgitation and pneumonia. Most research indicates that pneumonia results from aspiration of oral microbes and not the nutritional formula or gastric contents. Overall, there was no significant difference in pneumonia, intensive care unit (ICU) length of stay, or mortality rate between groups [Critical Care Medicine 33: 449, 2005]

Tube Feeds in Neurosurgical Patients

  • There is no difference between gastric and duodenal feeds in terms of complications or outcome [J Trauma 39: 1100, 1995; Crit Care Med 30:796, 2002] in TBI patients.
  • Starting full dose tube feeds in patients provided that residuals are < 200 cc reduces infective complications, leads to faster neurologic recovery, but at 6 months does not ultimately affect outcome. [Crit Care Med 27: 2525, 1999]

EBM: Is there a benefit to postpyloric feeding?

(Editorial by Mavkel, Pazirandeh, and Bistrian (Harvard Medical School) [Crit Care Med 30: 1654, 2002])

The VA Study [NEJM 325: 525, 1991] initially showed more infectious complication in the TPN group (14.1 vs. 6.4%, p = 0.01) with similar rates of major 30-day complications and overall 90-day mortality rates. Further analysis of this and other early studies [J Trauma 29: 916, 1989; Ann Surg 215: 503, 1992] suggest that the increased rates of infection associated with the administration of TPN could be largely attributed to overfeeding and/or resultant hyperglycemia [Crit Care Clin 17: 107, 2001]. Other studies [Arch Surg 136: 933, 2001; NEJM 345: 1359, 2001] that have matched caloric intake between groups reveal no difference in clinical outcome between total enteral and parenteral nutrition. Therefore, in the year 2002, it is indisputable that when delivered appropriately, both forms of nutritional support can be equally expected to improve organ function, immune competence, and wound healing in appropriately selected patients.

PEG tubes

118 patients with moderate to severe TBI had enteral access obtained and intragastric feeding initiated on day 3. 6. Intragastric feeding was tolerated without complication in 111 of 114 (97%) patients. Five patients aspirated, but had no evidence of intolerance prior to the event. [Am J Surg 179: 168, 2000]

In a prospective study of 40 critically ill patients (intended to study residuals), the frequency of regurgitation and aspiration was less with PEG tubes compared with nasogastric tube[Critical Care Medicine 33: 449, 2005]

See Also