When and how to start enteral nutrition

All the guidelines support the use of early enteral nutrition. Bowel sounds and flatulence are unreliable signs of gut health, but they seem to enjoy ongoing use as triggers to commence post-operative feeding. There is no good evidence that you need to slowly increment the volume, but everybody does this anyway, probably including the CICM examiners. To be called successful, enteric feeding needs to be escalated to goal rate fairly rapidly (within 72 hours). The well nourished patient will not be harmed by up to 6 days of "trophic" slow feeds, in much the same way as they would not be harmed by a week of starvation, but it remains to be seen whether this is any better than complete bowel rest (or whether it has any of the known gut-protective benefits of EN at 50-65% of the goal rate).

The most recent appearance of this topic in the CICM Fellowship Exam was in a form of a critical evaluation, as Question 9 from the second paper of 2019. The examiners specifically wanted an analysis of the pros and cons of early nutrition. 

In summary:

The best resources for this sort of thing remain the big official guidelines:

• ASPEN Guidelines
• ESPEN Guidelines

Additionally, the CICM answer to Question 9 from the second paper of 2019 mentioned these "notable trials" which the trainees should probably know:

• EDEN (Rice et al, 2012)
• PERMiT (Arabi et al, 2015)
• EPaNIC (Casaer et al, 2011)
• TARGET (ANZICS CTG, 2018)

How early is "early" nutrition?

48 hours is the cut-off for the "early" period according to both ESPEN and ASPEN. There is some evidence that feeding results in reduced mortality and reduced length of ICU stay when it is commenced within 24-48 hours of admission.  That 48 hour window seems to be the golden time for a hypercatabolic state; starting feeds within that time appears to protect patients from infectious complications and decreases hospital stay. The theory is that by improving gut integrity one diminishes the translocation of nasty anaerobes, and thus the total systemic endotoxin load diminishes. Moreover, no studies have ever demonstrated any harm in early enteral nutrition (Heighes et al, 2016). 

When to start enteral feeding, and at what rate?

Enteral nutrition should be withheld in the severely shocked patient

There is evidence to support withholding feeds from the patients who are still in need of vast quantities of fluid and vasopressors. The complication rate from enteral nutrition in these people increases, and this is probably because all those vasopressors decrease splanchnic perfusion (pretty much all of them do this, be they catecholamines or vasopressin). Even without vasopressors, the patient undergoing vigorous fluid resuscitation should have their feeds withheld. The fluids will cause whole-body oedema. Oedematous bowel means more interstitial fluid, which means a larger distance between the capillaries, and so diffusion of oxygen and nutrients is impaired by aggressive fluid resuscitation. All of this leads to an increased risk of ischaemic gut.

Though off topic, the wisdom of using vast amounts of fluid resuscitation is debated in the chapter on fluid resuscitation for septic shock - in short, it may not be the best idea, and the rationale for it does not bear close scrutiny.

Anyway; soon after the patient is stabilised feeds may commence. How "stabilised" do they need to be? Not very.  Heighes Doig and Simpson from RNSH (2016) recommend using the Shock Index to determine when this magic moment arrives; the Shock Index being a crude measure of cardiovascular derangement (it is heart rate divided by systolic blood pressure). When the Shock Index returns to 1 (eg. a heart rate of 100 and an SBP of 100) one may commence feeds according to this rule of thumb. Mindless application of such rules is to be guarded against. One might with a little imagination conceive a situation where the patient maintains these vital signs only at the cost of double strength noradrenaline and vasopressin (in which case one could hardly call the shock state "resolved").

There is no need to wait for flatus or bowel sounds

Farting and gurgling have been long upheld as the wise surgeon's triggers for feeding, but in fact there is good evidence to support the commencement of enteral nutrition without waiting for these unreliable signs of bowel activity. The sounds made by the bowel are evidence of contractility, not absorptive capacity or mucosal integrity. One can still bourbourygmize while ones mucosa is eroded, denuded of villi and swarming with translocated bacteria. Conversely, a healthy intact bowel may be silent for the brief few seconds you spend listening to it.

The starting rate can be around 30ml/hr

Locally, people use to a protocol (2013 edition) which prescribes starting at 30ml/hr and increasing by 30 after the subsequent 4-hour period provided the NG aspirate was less than 250ml. The precise volume of initiation of feeds does not seem to be selected very scientifically. There is a surprising amount of practice variation in this area (even when you take into account all the situations which genuinely call for "trophic feeding", as discussed below). The local protocol claims to be based on the defunct 2003 version of the Canadian Clinical Practice Guidelines, which seems to recommend a starting rate of 25ml/hr (and there is no reference or explanation as to how 25ml/hr is better than 30ml/hr). The Orlando Health protocol is based on the same guidelines, but thir starting rate is 20ml/hr, also without explanation of how this is better. The Dietitians Association of Australia guidelines (2011) acknowledge that there is a wide variation of recommendations, and then go on to recommend another totally different initial rate  (40-50ml/hr).

One might ask, why not start at full goal rate? Why undefeed patients in the first few hours? The concern is high gastric aspirates, and aspiration. It is based in one prospective observational study among 49 patients (Mentec et al, 2001) which frightened everybody by finding an association between high gastic aspirate volumes (150-500ml) and increased ICU mortality, increased ICU stay and a higher incidence of pneumonia. The patients with higher gastric aspirates tended to have more vomiting and had less caloric intake. Recommendations to slowly increment the feeding rate arose from this, with the expectation that small feeding volumes should prevent large gastric residual volumes, and somehow improve mortality thereby.

However, this was contradicted by subsequent studies. Specifically, the threat of high gastric aspirate volumes was thoroughly debunked. Montejo et al (2010; the REGANE study) found no significant difference in outcome for patients with volumes as high as 500ml (neither mortality nor incidence of pneumonia were affected in this prospective randomised multi-centre study). Poulard et al (2010) abandoned the practice of meauring gastric residuals altogether, and found it made no difference in VAP. Lastly, Desachy et al (2008) performed a randomised controlled trial of immediate full rate vs. incremental feeding rate increase among one hundred patients. Even though the full rate group ended up getting high residual volumes, this did not seem to matter to the overall rate of serious adverse events (i.e. torrents of vomit or other similar embarrassments which migh cause one to stop the enteral nutrition pump for some time). Notably, nothing horrible happened to the full rate group, and in addition they all ended up getting much better nutrition (in terms of calories per day).

Some caveats must be noted. The Desachy study was small in number, and excluded a large population of patients for whom slow introduction of feeds would have been beneficial (malnourished patients), or who can reasonably be expected to have poor gastric emptying (pregnant women). Unfortunately, patients with shock of any cause were also excluded. In short, that left a group of fairly stable well-nourished intubated patients. Not only that, but they were all routinely commenced on prokinetics and managed at 30 degree bed elevation. It is probably safe to do just about anything with the nutrition of this population.

So, what do you do? The savvy fellowship exam candidate will have to acknowledge the fact that in Australia there is a strong cultural resistance to introducing feeds at goal rate. The exam answer will need to reflect the fact that the marking examiner has probably spent the last 30 years teaching people to start low and go slow. Hence a starting rate of 30ml/hr should be used for answering CICM Fellowship SAQs. There may or may not be more marks in mentioning the lack of compelling evidence for this strategy.

"Trophic" trickle feeding is safe for short periods in well-nourished patients

If the patient is severely feed-intolerant, one may wish to feed them at 10ml/hr while trying various tricks to improve feed tolerance. This way, the patient's gut mucosa may still derive some of the benefits from the nutrient mixture, particularly where mucosal integrity is concerned. Trophic signals from having something in the lumen tend to protect the villi from atrophy, and maintain the ecology of the bowel microbiome. 

Rice et al (2011, the "EDEN" trial ) have demonstrated that trophic feeding is a legitimate strategy, and safe for up to six days. The population were all ARDS patients. The dribble-fed group had a trend to slightly higher mortality, but this trend did not reach statistical significance (22.4% vs 19.6%). They of course had less diarrhoea and less episodes of high gastric residual volumes, but one might argue that this was not a meaningful finding - because how clinically significant could these problems have been if they did not have an impact on hard outcomes? More weirdly, somehow the failure to meet nutritional goals in the trophic group did not translate into any meaningful disadvantage for them. This is probably because Rice et al excluded malnourished patients from the study.

So, really the study has confirmed that well-fed healthy people will tolerate just about any level of nutritional abuse. Likely, the same population could be safely fed 200% of their goal calories, or be starved completely, or have their bowel continuously irrigated with cucumber water. The question is whether this 10ml/hr rate has any benefit for the gut mucosa. The EDEN trial addressed this only indirectly. The necessary volume of enteral nutrition required to maintain intestinal integrity remains unknown.

How fast should it escalate, and to what goal rate?

Enteral nutrition be advanced to goal rate within 72 hours for malnourished patients

The ASPEN people recommend that at least 80% of the daily goals be met by the second or third days of feeding. Generally speaking, guideline statements have historically agreed with the idea that you should advance to your goal rate as soon as it is tolerated (to do otherwise brings into question the practice of setting goal rates, and the definition of the word "goal" in a broader sense). This has greatest relevance for patients at a high risk of malnutrition.

Hypocaloric feeding is appropriate for well-nourished ICU patients in the acute stage

Several recent publications have emerged which have challenged the previously unassailable truth, that the intensivist should stampede mindlessly towards a 100% satisfaction of the estimated nutritional goal. Marik and Hooper (2015) have spared me the need to hunt down and quote all of the relevant trials by performing a handy meta-analysis. With data from 2517 patients among six trials meeting eligibility criteria, the authors could not find any difference in the risk of acquired infections, hospital mortality, ICU length of stay or ventilator-free days between patients receiving 20-50% of their goal, when compared to patients receiving 72-77% of their goal.

The pursuit of normocaloric goals took another hit with the INTACT trial (Braunschweig et al, 2015), which found higher mortality in ALI with aggressive full-bore enteral nutrition. The trial is small and single-centre; however, there was sufficient concern for the steering committe to stop it prematurely, as abundant evidence of harm had manifested. The inclusion of INTACT had led the 2015 meta-analysis by Tian et al that critically ill patients who are not malnourished really do not need aggressive enteral nutrition, and that they may even be harmed by it.

Such findings have prompted a change in expert opinions regarding timing of full-scale nutrition: for instance, Jeejeebhoy (2015) suggested that moderately hypocaloric feeding (33-66% of goal rate) might be better for the early stages of an illness, with goal rate escalation to 100% becoming more appropriate in the convalescent stages of critical illness. The most recent iteration of the Canadian guidelines (May 2015) has also supported this idea.

"Trophic" trickle feeding is safe for short periods in well-nourished patients

If the patient is severely feed-intolerant, one may wish to feed them at 10ml/hr while trying various tricks to improve feed tolerance. This way, the patient's gut mucosa may still derive some of the benefits from the nutrient mixture, particularly where mucosal integrity is concerned. Trophic signals from having something in the lumen tend to protect the villi from atrophy, and maintain the ecology of the bowel microbiome. 

Rice et al (2011, the "EDEN" trial ) have demonstrated that trophic feeding is a legitimate strategy, and safe for up to six days. The population were all ARDS patients. The dribble-fed group had a trend to slightly higher mortality, but this trend did not reach statistical significance (22.4% vs 19.6%). They of course had less diarrhoea and less episodes of high gastric residual volumes, but one might argue that this was not a meaningful finding - because how clinically significant could these problems have been if they did not have an impact on hard outcomes? More weirdly, somehow the failure to meet nutritional goals in the trophic group did not translate into any meaningful disadvantage for them. This is probably because Rice et al excluded malnourished patients from the study.

So, really the study has confirmed that well-fed healthy people will tolerate just about any level of nutritional abuse. Likely, the same population could be safely fed 200% of their goal calories, or be starved completely, or have their bowel continuously irrigated with cucumber water. The question is whether this 10ml/hr rate has any benefit for the gut mucosa. The EDEN trial addressed this only indirectly. The necessary volume of enteral nutrition required to maintain intestinal integrity remains unknown.