Management of inadequate nutrition and feed intolerance

Considering that "Enteral and parenteral nutrition" is an L1 topic from Section 2.1.6 in the second edition of the CICM Syllabus for the Second Part Examination, it is surprising that feed intolerance  seems to appear in the exams a lot less often than it appears in practice. For example, 

Question 16 from the second paper of 2024 asks directly about this, whereas Question 9  from the second paper of 2001 presents us with the scenario of a trauma patient whose nutritional needs are not being met. Why are they not being met? There are a series of potential reasons. A tabulated response is offered, listing potential causes against their solutions. Following this, a long and pointless rant takes place, discussing the various finer points. Given the definition of "L1" for this edition of this syllabus includes "Relevant anatomy, Principles and practice, Interpretation,  Relevant guidelines and evidence, Controversies and risks", every effort was made to overcomplicate  this potentially straightforward area of discussion. Reintam Blaser et al (2021), if one can get a hold of it by honest means, is an excellent option to prime oneself for the CICM exam questions on feeding intolerance, but above all else the ASPEN and ESPEN guidelines statements should be the canonical resources.

Definition of feeding intolerance

"Feeding" intolerance is what this is usually referred to, but the author reverts to his underclass roots too soon, and continues to use "feed" as a noun as if referring to farm animals or the destination of a conveyer belt. In his defence, this colloquialism has permeated deep into the culture of Australian critical care and can even be found in professional documents of local government agencies. 

Unlike sepsis or obscenity, feed intolerance seems like something one should be able to objectively define, and so "I know it when I see it" does not feel like a very scientific way of approaching this enitiy. But that's what Reintam Blaser et al (2021) settled on as a "pragmatic" definition:

"a clinician has decided to reduce the amount of enteral nutrition specifically because features of gastrointestinal dysfunction appeared during enteral feeding"

Sure, but a person trying to run a clinical trial of prokinetics would complain that this is so broad as to be entirely meaningless. The problem with havingsz@222m3121e a scientific numerical definition is that nobody can agree on what the parameters are for measurement, or a threshold for whatever we do measure. The authors of the aforelinked article offer this excellent table to demonstrate the spread of definitions across papers:

This range of variation should encourage the pre-exam reader; as no variable with this much difference in published opinion could ever become the subject of an exam answer, and if it ever did, the range of acceptable correct answers would surely be so wide as to invite essentially any reasonable value. The most important take-home is that most of these listed papers used the gastric residual volume, GRV, as their benchmark. This gastrocentric world view needs to be explored in greater detail, which is a handy segue into the discussion of GI anatomy:

Relevant anatomy for feed intolerance

The author, having completed the primary exam gastrointestinal system section years ago, recalls experiencing some surprise that it contained zero reference to anatomy, but this is now made clear with the Second Part syllabus, where all of the anatomy seems to be represented in the "expected knowledge" of system topics.  The mindless application of this wording to these items occasionally leads one into weird culdesacs, and the discussion of the "relevant anatomy" for feed intolerance could be one such appendix. Nowhere else in the literature is this mentioned in any meaningful sense, suggesting that the leaders of thought in nutrition do not regard it as important. Still, if somebody had to come up with some kind of anatomical aspects, one could point to the different locations where passage through the GIT could be frustrated functonally, as well as the points at which it could be bypassed. The anatomical boundaries of this discussion will be limited to the upper GI structures, because to discuss the motility problems involving the colon and small intestine here would risk a semantically dangerous overlap with the chapter on ileus and constipation.

The relevance of the oesophagus for feed intolerance

Nasogastric feeding very effectively bypasses the oesophagus and whatever problems it may be having are therefore of no consequence to the delivery of nutrition, except where the oesophagus is so anatomically abnormal that the NG tube cannot be advanced into the stomach, or where the oesophageal sphincter is so incompetent that the stomach content sloshes up into the upper aerodigestive tract with the slightest increase in abdominal pressure.

• Oesophageal diverticula can frustrate the advance of the NGT tip. 
• Gastric hiatal hernias can cause the tube tip to be come supradiaphragmatic, while technically remaining endogastric. To blindly force the passage of the tip below the diaphragm in these cases is usually ineffective, and results in coiling. 
• For both of the above, the solution is to position the tube endoscopically. Feeding the supradipahgramatic stomach is said to be possible, but is known mostly from case reports, which means ICU trainees should not get into the habit of presenting it as a serious option in their exam answers.
• Interference of the NG tube itself with oesophageal sphincter function is a known thing (Manning et al, 2001): The tube stents the sphincter and prevents complete occlusion of the passage, encouraging the leak of gastric contents into the oesophagus. Positioning the patient with their head up to 30-45º is one strategy to guard against this, but it is not effective if the patient is planning to have a coughing fit (where intraabdominal pressures can exceed 160 H₂O).

The relevance of the stomach for feed intolerance

As discussed elsewhere, the stomach is a) no place to absorb anything, and b) susceptible to countless disabling influences, which makes it the main gatekeeper standing between the patient and your carefully crafted nutritional support strategy. Those sound like mostly functional aspects, of course; but the anatomy of the stomach does play a role. The most important anatomical matters to consider are the position of the stomach in the supine patient, and its capacity.

Gastric "pooling" is a phenomenon seen in supine ICU patients, whose position is entirely unlike the usual position of upright ambulant volunteers selected for gastric emptying studies. In the supine patient, the stomach drapes floppily over the central vertebral and vascular structures in the chest, separating into two pools and preventing proper transit. This excellent CT from Parrish & McClave (2008) demonstrates this phenomenon beautifully:

One can imagine how, in the left recovery position, all of the contents will slosh into the antrum, and no progress will be made with the feeding until the patient is rotated to the other position.

Gastric capacity also factors into nutrition decisions. The range of gastric capacity can vary dramatically. In a normal individual, one might expect something in the order of 250-300ml at rest (well, at eternal rest, in cadavers), and 0.4-1.5 ml/kg in the fasted state, but there is a common belief that it expands readily, and can comfortably stretch to accommodate vastly more contents than one might ever ordinarily ingest. This may not be the case.  Geliebter & Hashim (2001) defined the upper range of comfortable fullness as something in the order of 1100-1200ml for obese binge eaters and patients with bulemia, and perhaps 600-700ml for normal lean patients, on the basis of an experiment where they progressively added more and more saline to an intragastric balloon. The operational maximum volume for gastric capacity is variably defined, and where one finds a number it is often not matched with a reference. For example, Wickham et al (2012) mentions a volume of 4 litres as the maximum, but does not explain what kind of horrific experiment this came from.

What meaning does any of this have for feed tolerance and intolerance? Well:

• A defining feature of feed tolerance is the measurement of the gastric residual volume in patient having continuous feeds. 
• The normal range of gastric volumes is therefore an important range, because it defines what one might accept in ICU patients.
• The upper range of this normal volume is also important, because if we accept that patients can propel this volume into the small bowel, we accept this as normal (and return it when aspirated), whereas if we decide that it is not normal, we discard it (losing nutrients) and start prokinetics.

The most logical thing to move on to from here is:

Interpretation of the gastric residual volume measurements

It is important to recognise that the focus on the gastric residual volume is not guided by some sound scientific principles, but rather by convenience. We have access to this volume because we have placed a tube in it. Elke et al (2015) point out that it is "probably one of the most traditional and widely accepted nursing practices in the ICU", which is another way of saying that we do this because That Is How We Do Things Around Here. Parrish & McClave (2008) quote Thomas Paine in response: "A long habit of not thinking a thing wrong, gives it a superficial appearance of being right”.  This is the appropriate position from which to regard this practice. 

Consider: generally, healthy adults are thought to produce from 2-3 up to 5-6 litres (if being fed) of gastric juices per day, which averages out to 100-250ml or so every hour; of this the vast majority is reabsorbed by the gut. Of the total gastric content, some is gastric secretions (maybe 1500-2500ml per day) and some is swallowed saliva (maybe 1000ml/day). Lets say you infuse a goal rate of 60ml/hr of feeds, and aspirate 4-hourly. A patient whose pylorus is tightly closed will absorb nothing, and 240ml of feeds will collect in their stomach. Additionally, there might be from 400ml up to 1000ml of gastric secretions, and 160ml of saliva. This hypothetical patient would end up with 800-1400ml of stomach content. From one's ICU experience, one can recall that this sort of 4th hour residual volume is never seen in practice. Thus, some gastric emptying is almost always occurring, and some feeds are therefore propagating into the small intestine to be absorbed. Considering that we are largely interested in the transit of nutrient solution and not stomach juice, and that we can never know how much of which is making its way through the pylorus, it is remarkable that so much importance is attributed to this parameter.

Still, the CICM trainee needs to write something less dismissive in their exam answer. For them, the ASPEN and ESPEN guidelines are the definitive source for numbers:

Enteral nutrition tolerance	ASPEN	ESPEN
Gastric residual volume threshold	500ml	500ml
Aspirate frequency	not regularly	6 hourly
So it's higher. What to do with it?	Delay enteral nutrition

Experts disagree as to how large an aspirate is too large, as seen in the table far above. Another good review article from 20 years earlier confirms that there has always been confusion among the literature, with volumes from 150 to 500ml quoted. Consensus statements suggest that while anything in the range of 200-500 should be investigated (i.e., one should submit for some cognitive processing the question, "does my mostly-feed-tolerant patient have something seriously wrong with them?"), only a gastric aspirate in excess of 500ml should be a reason to stop feeds. The rationale for this answer is that inappropriate feed cessation is the most common reason for why one may not reach their goal rate of nutrition, and when the goal rate drops below 50-65%, one loses the benefits of enteral nutrition.

Return up to 250ml of gastric residual volume

So, lets say the aspirate was in excess of 500ml, and you have decided to act in response to this. How will you act? You have the choice of stopping the feeds or continuing them, and you have this bucket full of undigested stomach content (shall we pump it back into the patient, doctor?) The consequence of having a low threshold for stopping feeds and discarding the residual volume is unmet nutritional goals. Predictably, nobody can agree about how long the feeds should be held for, what rate to restart them at, and what to do with the aspirated residual volume. A review by Guo, 2015 could not find any grounds for firm recommendations among seventeen studies on the subject. If one needed some guidance,  Juvé-Udina et al (2009)  tested a sane compromise, and found that it was safe to return 250ml of the residual volume, discarding the rest. This seems to be a sensible approach, as it protects the patient from both malnutrition and regurgitation.

Prevalence of feed intolerance

Considering how much trouble we have defining this entity, it is no surprise that the rate is quotes as being between 2% and 80%. Heyland et al (2021) went with site-specific thresholds (i.e. however the participating hospitals defined it) and ended up with an incidence of 24% per episode of stay (which is actually very little). Among all patients, on day 4 and 5, the overall prevalence was 6%. Gungabissoon et al (2015), on the other hand, used a similar methodology and ended up with a 30% incidence. 

Assessment of the feed-intolerant patient

"Outline the potential aetiologies", asked the examiners in Question 16 from the second paper of 2024. That could be taken to mean "list the conditions that are related to, or causative of, feed intolerance" or it could mean "list the pathophysiological processes responsible for feed intolerance". To hedge against either possibility, both are hopefully answered by what follows.

Potential aetiologies of feed intolerance

Deane et al (2007) and Reintam Blaser et al (2021) are good resources for this material, though the latter features a terrifying diagram of a brightly coloured gastrointestinal tract. Looking at this thing is not going to be helpful directly, but indirectly it is valuable as a catalogue of physiological problems:

• Gastric aetiologies
  • Decreased gastro-oesophageal sphincter tone due to NG tube
  • Increased pyloric tone due to increased sensitivity to small bowel nutrient content
• Pancreatic and biliary aetiologies
  • Reduced (or increased!) bile salt transit into the small bowel
• Small bowel aetiologies
  • Abnormal pattern of peristaltic motility due to the influence of cytokines
• Large bowel aetiologies
  • Increased gas (eg. due to overgrowth of gut bacteria)
• Global influences on gastrointestinal function
  • Poor cardiac output and poor perfusion
  • Tissue oedema
  • Raised intraabdominal pressure
  • Impaired intestinal nervous system activity due to the effect of opioids and anticholinergic agents
  • Electrolyte disturbances (potassium and magnesium especially), acidosis
  • Abdominal surgery
  • Immobility
  • Gravity/pressure effects (eg, due to prone positioning or left lateral position with the antrum dependent)

These are probably not as important as a list of 

History suspicious for various causes of feed intolerance

Like with most "history" sections of the assessment, this is really a list of the risk factors and clues to specific etiologies:

• Drug history of opioids or anticholinergic agents
• Recent surgery (indirectly, suggestive of opioid exposure)
• Abdominal surgery or trauma
• Psychiatric comorbidities (a great likelihood of being on anticholinergic medications
• Neurological diagnosis (unconsciousness, and therefore immobility)
• Sepsis (cytokines are implicated in gut dysmotility)
• High illness severity (APACHE score correlates with delayed gastric emptying)
• Diabetes (poor autonomic function, possibly even a baseline gastric emptying delay)
• Ascites
• Mechanical ventilation

On the other hand, at some stage the college will ask for history of fedd intolerance itself, i.e. hiostory-derived features on assessment that might suggest the patient is introlerant of enteral nutrition. These would be something like:

• Vomiting
• Nausea
• Diarrhoea
• Aspiration 
• Abdominal pain 

Examination findings that suggest feed intolerance

• Abdominal distension
• Abnormal bowel sounds
• Bile-stained NG aspirate
• Large gastric residual volume (>500ml)

Radiological findings in enteral feed intolerance

These are hardly specific for enteric feed intolerance per se, and mostly represent a signal of some motility problem:

• Large gastric bubble
• Distended stomach and bowel loops
• Differential air-fluid levels or air filled loops
• Enlarged (but not inflamed) gallbladder, suggesting that it has not emptied for some time
• Contrast from last week's imaging still visible in the bowel lumen
• Specific imaging (eg. serial contrast studies to observe the rate of gastric emptying)
• CT or MR enterography

As one might expect, there is no specific resource for this, and all searches to find one tend to lead to papers that discuss investigating gut motility instead.

Complications of enteral feed intolerance

Consequences of feed intolerance can be divided into "consequences of not being fed enough" and "consequences of your obstinate insistance on feeding them anyway". As such:

• Inadequate nutrition
• Risk of aspiration
• Increased gut wall stress (and therefore risk of ischaemia and perforation)
• Erratic drug absorption (some tablets will surely be aspirated along with those 500ml 6-hourly residuals)
• Nausea and discomfort
• The ventilator consequences of abdominal distension (i.e. reduced FRC)

And then:

• Complications of TPN (lines, infected lines, LFT derangement, etc)
• Toxicity of prokinetics (QT prolongation, dystonic reactions)
• Toxicity of antiemetics
• Increased risk of gut translocation (due to lost enteric trophic stimulus)

Management options for enteral feed intolerance

A senior ICU trainee should be able to produce a tiered approach to escalating the level of nutritional aggression, and be able to generate specific timelines and branches of decisionmaking that can happen along the way to full nutritional supremacy. This can be summarised as follows:

• Address immediately worrying possibilities:
  • Exclude surgical causes for feed intolerance, eg. bowel obstruction or ischaemia
• Reverse the reversible physiological contributors:
  • Correct acidosis and electrolyte disturbance
  • Address tissue oedema
  • Correct shock and acidosis
• Encourage gut perfusion:
  • Extubate the patient, or at least wean them to a spontaneous mode to permit movement of the diaphragm, which should reduce abdominal venous congestion
  • Decompress the distended gut by NG drainage
• Minimise the negative/maximise the positive contribution from drugs:
  • Transition to non-opioid and opioid-sparing agents
  • Consider delirium management options which do not have anticholinergic effects
  • Start prokinetics (reasonably benign):
    • erythromycin and metoclopramide 
• Maximise the potential for gut transit:
  • Position the patient as upright as possible
  • Choose the smallest tube possible to reduce the stenting effects on the upper gastro-oesophageal sphincter
  • Use continuous, rather than bolus feeds
  • Reposition the NGT beyond the pylorus
  • Insert an NJT (nasojejunal tube) and keep the NGT toquestion decompress the stomach
• Reassess nutritional goals and aim to have them met within 72 hrs:
  • Determine the nutritional requirements in better detail
  • Use lower volume feeds (higher nutrient concentration)
  • Return half (at least 250ml) of high aspirates
  • Minimise interruptions to continuous feeds
• Add supplemental PN if nutritional goals remain unmet at the end of 1 week, or earlier if the patient is significantly malnourished

Prokinetic agents

A curiously autotranslated-looking article by Nam (2014) reviews the current evidence and the pharmacological arsenal available; a concise summary is also offered by LITFL. It is a summary of the CCR article by Frazer et al (2009), which is a definitive resource on the subject. A local homage to prokinetics also exists in the chapter on the management of constipation and poor gut motility in the ICU.

The conventional prokinetic agents are:

• Metoclopramide
• Domperidone
• Erythromycin
• Enteral naloxone
• Methylnaltrexone
• Neostigmine

The exotic drugs:

• Alvimopan (an opiate antagonist)
• Mitemcinal (an erythromycin-like motilin agonist)
• Ghrelin (an enteric hormone)
• Dexloxiglumide (a cholecystokinin receptor agonist)

In short, metoclopramide, erythromycin and enteral naloxone all seem to be good prokinetic agents. The dose of naloxone via the NG tube is 8mg 4 times a day, which means you require a special enteric formulation (the IV ampoules only contain 400mcg). Cisapride and tegaserod were once available, but their tendency to kill people with cardiovascular toxicity has seen them removed from the market. Neostigmine and methylnaltrexone are agents which influence lower gastrointestinal motility more than upper, and are probably poor first line choices.

Which drug to start as first line therapy? It seems both metoclopramide and erythromycin should be used together. Erythromycin is the more potent agent of the two (Nguyen et al, 2007), and there is evidence that metoclopramide may have no effect on gastric emptying alltogether (Marino et al, 2003). There is the threat of tachyphylaxis: both agents lose their effectiveness over the first week of therapy.

Post-pyloric feeding

If the NG aspirate volumes are too high, it is sensible to advance the feeding tube into the small bowel (beyond the ligament of Treitz). Post-pyloric feeding ( typically into the 3rd portion of the duodenum) is well studied - there are 3 big meta-analysis articles. Generally, they all agree that post-pyloric feeding is a good idea in patients whose gastric emptying rate is sub-par, or those who are at a particularly high risk of aspiration (e.g.. somebody who has to be nursed in a flat supine position). Does this really work? Probably not. The ENTERIC study (Davies et al, 2013) compared nasogastric and nasojejunal feeding and could not find any difference in complication rate (particularly, frequency of pneumonia), efficiency of nutrient delivery or mortality.

Added TPN to inadequate enteral nutrition

Once again the reviews by Braunschweig and Heyland support the idea that after 7 or so days of poor nutrition the benefits of TPN outweigh the risks. Poor nutrition may still occur with enteral nutrition if the nasogastric feeds are poorly tolerated. The TPN should continue as supplementation until at least 50-60% of nutritional needs are met by enteral nutrition.

The exact time of supplementation is unclear, but it seems that early TPN supplementation is probably a good idea.

The 2011 study of early TPN (Caesar et al, NEJM - the EPaNIC trial) found that "late initiation of parenteral nutrition was associated with faster recovery and fewer complications, as compared with early initiation". However, of the studied patients (4640 of them) the majority (60%) were cardiac surgical patients who do not routinely require TPN,  the investigators used a "strict" BSL target for their insulin therapy (which we know is harmful), and 5% dextrose was used as a maintenance fluid (which is bizarre and is not practiced in Australian ICUs).

The more recent (smaller, but better designed) trial by Heidegger et al (2013) demonstrated a decreased rate of infections in the group of patients who had received supplemental PN together with their (inadequate) EN. The supplemented group also had better nutrition (103% of the goal was met, as opposed to 77% in the control group). In view of this, both ESPEN and the Alfred authors recommend PN be added to EN after two days of struggling with feed tolerance.