Enteral Nutrition

Abstract

This chapter first describes the types of commercial forms of liquid formula diets that are commonly available in hospitals in the USA with advice on which situations they should be used, followed by a detailed description of the various forms of enteral tube feeding techniques, starting with the most simple, nasogastric, and culminating with more complex endoscopic procedures for the placement of double-lumen nasogastric decompression and distal jejunal feeding tubes and percutaneous feeding tubes. Finally, we will provide examples of where interventional tube feeding can now be used to restore gut function and obviate the need for IV feeding, thus reducing complications and improving patient outcome.

The Enteral Nutrition Formulae (LFDs)

There are a wide variety of standard and specialized enteral feeds available as illustrated in Table 8.1, but most hospitals select those they have the greatest demand for. Bottom line: most patients can be managed with standard polymeric liquid formula diets (LFD) as they are “balanced” and contain all the essential nutrients. Specialized formulae are appropriate when there is maldigestion and/or malabsorption, in severely catabolic, and in renal and hepatic failure patients. Polymeric LFDs are not only cheaper, but are also usually better tolerated. Pancreatic function is suppressed in the critically ill, but never to the degree that leads to maldigestion and malabsorption of LFDs if they are given at a constant 24 h infusion rate. The earlier enteral feeding is commenced, the lower the risk of intolerance and the greater the chance of maintaining digestion and absorption.

Table 8.1 Some of the enteral feeds commonly used in hospitals in the USA

• Polymeric: Basically, polymeric diets are liquidized food and therefore satisfy normal nutritional requirements. They are therefore used when patients cannot eat normal food for whatever reason but have normal intestinal digestion and absorption. Several commercial products are available, varying little in composition. Usually, 1,500–2,000 ml will provide total daily nutritional requirements, but the protein:calorie and fat:calorie ratios vary making some products more useful in stressed and glucose intolerant patients. The quality of protein may also vary depending on its origin, i.e., casein (animal) or soya (vegetable). The chief advantage of polymeric diets is their low cost and palatability. Because the macronutrients are in a complex form, the formulae also usually have a low osmolarity and are therefore well tolerated. Being ‘complete’ they usually contain resistant starches or fiber and therefore maintain healthy colonic function, something that has now fashionably been termed “prebiotics.” They are available in powder form for reconstitution (cheaper still) or as full liquids either in cans or in “ready-to-hang” plastic bags—alike TPN bags. Reconstituted formulae are particularly useful for longer term supplemental feeding, while factory-packaged liquid formula are advantageous for feeding the immunosuppressed and critically ill patients where a “closed system” reduces the risk of feed contamination. It must, however, be remembered that all tube feeds, like normal food, are non-sterile and they only differ in the degree of contamination.

• Semi-Elemental (ED): Here, some of the macronutrients have been predigested. The proteins are usually in small or large peptide form, the carbohydrate as maltodextrins, and the fat either low in quantity or in medium-chain triglyceride form. The advantage of these formulae is that they are easier to absorb and therefore should have advantages in patients with compromised digestive and absorptive capacity. On the other hand, the osmolarity is higher and they have to be delivered at a slow continuous rate and never in bolus form. Semi-elemental diets are not surprisingly more expensive and are probably overused in hospital practice as they are better advertised and more fashionable. It must be remembered that there is considerable digestive and absorptive reserve, such that you need to lose >90% of either function before malabsorption is encountered. Consequently, a low-cost polymeric formula is usually better tolerated even in the sickest of ICU patients provided it is given at a slow continuous rate. However, patients with pancreatic insufficiency are better managed with an elemental formula, but close supervision of blood glucose is mandatory, particularly in those with associated pancreatic endocrine dysfunction. Another indication is in the feeding of severe acute pancreatitis (SAP) for reasons of digestibility and minimization of pancreatic stimulation (see Chapter 13). Pancreatic enzyme secretion is reduced dramatically in SAP [128]. It must however be remembered that the use of an ED does not avoid pancreatic stimulation, but does reduce it by 50% [30].

• Elemental: Elemental strictly means that the diet is totally digested and can be absorbed directly. A good example of an elemental diet is a TPN solution, where protein is given as amino acids and carbohydrate as glucose (dextrose). Fat is not, however, given as free fatty acids, but as triglycerides, usually in small quantities. The problem with elemental diets is that because they are broken down to their constituents, they are hyperosmolar, and if given at a fast rate will precipitate diarrhea. Vivonex (Nestle Nutritionals) is the most commonly used elemental diet consisting of 15% of energy as protein—in the form of free amino acids, 82% as carbohydrate as maltodextrin, and 3% fat as safflower oil. The results of early studies showing that free amino acids are absorbed less well than peptides [140] together with concerns about the hyperosmolarity have led to the preferred use of the semi-elemental formulae described above.

• Immune-Enhancing Diets (IEDs): The modern nutritionist is not satisfied with using nutrition intervention simply as a “support” system, while medical practice takes care of the underlying cause of the nutritional problem: he/she seeks to use specific nutrients to treat the disease process. The recognition that glutamine is essential for immune and small intestinal function, and that arginine generates nitric oxide that is essential for macrocyte function and blood flow has led to the development of LFDs enriched with these substances, termed IEDs. Enthusiasm for their use was initially stimulated by the results of several small sized studies and their meta-analyses [141, 142], while more recent reviews have been negative [109, 143]. While glutamine enrichment is generally safe, unless the patient has liver failure in which case it could exacerbate encephalopathy, it must be remembered that it is an inessential amino acid that can be synthesized by the body from glutamate, which, in turn, is synthesized (by transamination) from a Krebs cycle intermediate a-ketoglutarate. Thus giving an excess of an inessential amino acid will result in a reciprocal reduction in essential amino acid intakes, or excessive nitrogen intake, and nitrogenous breakdown products are potentially toxic. Secondly, arginine enrichment may exacerbate the inflammatory response in septic patients [143]. A thorough review of the literature led the Canadian group to conclude that diets supplemented with arginine and other select nutrients not be used for critically ill patients [109]. Consequently, few ICUs routinely use IEDS. My personal view is that the most common nutritional problem in ICUs is inadequate feeding of food, which consists of a wide spectrum of essential nutrients, including glutamine and arginine. Consequently, outcome can be best improved by the provision of a balanced nutritional composition—which is, after all, what we evolved to survive on.

Enteral Access Techniques

Enteral feeding is used when patients cannot eat sufficient to meet requirements, when they are at risk of complications through not eating, and when they have functional GI tracts. The reader is referred back to the last section (Chapter 5) where much time was spent defining what “gut function” is and how it can be enhanced. The key point to remember is that ileus, or absent bowel sounds, or high gastric residuals do not necessarily translate into gut failure, and commencement of slow continuous enteral feeding may restore gut function. The chief problem with any form of naso-enteric feeding is that it is unsightly and uncomfortable, and given the option, many patients would opt for IV feeding. However, it must be emphasized that EN vs. PN is not a patient choice, and PN must never be employed because the patient refuses EN. It must be acknowledged that the long-term use of nasoenteric feeding can produce nasal damage, sinusitis, and sialorrhea. It is therefore our practice to convert, if practically possible, per-nasal feeding to per-cutaneous if patients require enteral feeding for periods longer than 4 weeks. The most common approach is to use endoscopic placement (or alternatively, radiologic or surgical placement), either directly into the stomach (PEG), or via the stomach into the intestine (PEG-J), or on the rare occasion to place the tube directly into the intestine (PEJ) (see Chapter 8).

Enteral Access

Although the placement of specialized enteral and parenteral feeding tubes or catheters can be performed by persons without specialized knowledge of nutritional support, just like the placement of PICC lines is generally performed by technicians as a service, it is helpful to either have someone who is attached to the team, or who develops a special relationship with the team, to place difficult tubes for enteral or IV access. For enteral access, it is advisable to have a gastroenterologist (remember, I am biased, but remember the role of the gut is to absorb food and one of the chief consequences of gut disease is disturbance in nutrition) as a team member or an associate so that he/she can develop skill through experience in placing tubes by endoscopy and/or fluoroscopy to improve nutrition. As diseases of the gut most commonly result in feeding disorders and the need for nutritional intervention, most GI divisions should have someone with an interest in nutrition support. Similarly, it is important to develop an association with a GI surgeon for surgical percutaneous or laparoscopic jejunostomy placements in patients with obstructed upper GI tracts. Previously, we developed relationships with vascular surgeons to place tunneled subclavian catheters, but nowadays it is conventional to associate with an enthusiastic interventional radiologist (IR) who has the skills and techniques to place enteral or parenteral tubes anywhere in the body you choose!

Nasogastric (NG)

The most common form of enteral access is the placement of NG tubes by nursing or house staff. Although this is usually easy, it can be difficult in sick patients, and can be associated with life-threatening complications arising from “blind” misplacement or aspiration of feeds. Consequently, a strict protocol needs to be adhered to before any tube is used for feeding, as outlined below. If the problem is simply anorexia, a nasogastric feeding tube is all that is needed. Provided gastric emptying is satisfactory, feeds can be given in bolus form in volumes up to 500 ml given 6 hourly (safer 350 ml/4h), as the stomach is designed to retain volume and slowly deliver it to the intestine via the pyloric “valve.” It must always be remembered that the usual gastric emptying rate is 1 kcal/min and the maximum 2–3 kcal/min. Thus, a standard liquid formula diet, which contains 1 kcal/ml, will empty at a maximum rate of 180 ml/h—hence it is always wise to limit NG infusion rates to <150 ml/h or boluses to <500 ml/4 h. In critically ill patients, gastric emptying is usually delayed and so slower rates should be used, not exceeding 100 ml/h, and resting volumes (gastric residual volumes or “GRVs”) should be checked regularly (4–6 h). To increase tolerance, feeds should be given as slow continuous infusions rather than bolus. Simple measures to promote gastric emptying and to prevent reflux and aspiration, such as elevate head of bed and the use of prokinetics, should be employed (see below).

Benefits: Nasogastric feeding is the most common and oldest form of interventional feeding. Its attraction is that requires little skill, and can be placed by the nursing staff on admission allowing early enteral feeding. In patients with chronic disorders with chronic malnutrition, feeding can be given in intermittent bolus form during the day, as this is easier on nursing staff and appropriate for rehabilitation units.

Complications are associated with “blind” bedside placement and include misplacement in the tracheobronchial tree, mucosal injury with bleeding, and/or perforation. It is also not uncommon for the tube to be left in the lower esophagus, again increasing the risk of aspiration. In the extreme situation, the tube can be malpositioned in the tracheobronchial tree, as illustrated in Fig. 8.1. The patient was admitted with vomiting and a history of chronic pancreatitis complicated by pseudocyst. An NG “Salem” sump tube was placed at bedside and the patient treated with opiates for abdominal and chest pain, and sent for endoscopy in the morning. Remarkably, he complained of neither breathlessness nor cough! Transnasal endoscopy was used, anticipating the need for placement of a double-lumen nasogastric decompression jejunal feeding tube. As is our protocol, the NG tube was followed beyond the retropharynx and shown to be in the right main bronchus as shown in the figure! Once removed, the stomach was examined. The outlet was compressed by the pseudocyst and the stomach contained >2 l of bile stained fluid. Clearly, the risk of aspiration was extremely high. Luckily feeding had not been initiated. This case report serves as a warning that bedside placement of NG tubes must be accompanied by (a) making sure 50–60 cm of tube goes in easily and without resistance, (b) that insufflation of 60 cc of air is well heard (with bubbling) in the epigastrium on auscultation, and (c) that an X-ray is taken to confirm correct placement before any feeding is commenced.

Fig. 8.1

Transnasal endoscopic view of an NG tube placed in the trachea, note bifurcation

Although NG feeding is understandably associated with a higher risk of reflux and micro-aspiration [144–146], it is surprisingly effective in providing enteral feeding even in some of the sickest patients in the ICU [147–149] if the following anti-reflux measures are imposed:

Measures to Minimize Aspiration Risk in NG-Fed Patients

• 1. The head of the bed should be kept elevated 30–40° to the horizontal (semi-recumbent position). This rule may have to be broken in ICU patients where frequent positional changes are required, but there is good evidence that ventilated patients are at high risk of vomiting and aspiration if they are rotated to prone position for hypoxia [150]. However another study found that if GRVs were low, patients tolerated feeding well even in the prone position [151]. Recent studies have shown that the use of nasogastric feeding together with continuous lateral rotational therapy in patients kept at 30° reduced the risk of ventilator-associated pneumonia in ICU patients compared to non-rotational management [152].

• 2. Feedings should be started slowly and given continuously (20–25 ml/h). Note: there is little point in starting with feeding rates slower than this as physiological secretion is way above this rate, e.g., 80–160 ml/h.

• 3. Bolus feeding should be avoided until good gastric emptying has been confirmed.

• 4. Gastric residuals should be checked every 4 h. If residuals are repeatedly over 400 ml it is best to place the feeding tube past the pylorus. However, if this is difficult, emptying can be improved by the use of prokinetics (e.g., metoclopramide 10 mg IV q6h) and acid secretion inhibitors (e.g., IV PPIs, pantoprazole). In patients on narcotics, Narcan 10 mg added to the feed is said to be effective, but is not commonly used for concern about other side effects—it is always better to try and wean the patient off narcotics.

The meaning and definition of high gastric residual volumes (GRV): Until recently, a GRV level of 250 ml was generally accepted as the cutoff level for “safe” NG feeding. However, a recent Spanish study has shown that there is no increased risk of aspiration in ICU patients if 500 ml is used as cutoff [153]. This has persuaded the ASPEN and American Society of Critical Care Medicine to recently revise their guidelines to 500 ml [130]. One recent study showed that aspiration was no more common in ventilated ICU patients whether a cutoff of 200 or 400 ml gastric residual volume (GRV) was used, and that gastric residuals were remarkably low (mean of 31 ml) in patients with proven aspiration [154]. This implied that aspiration pneumonia is more commonly a result of inhaled pharyngeal secretions. However, another more worrying concern is the lack of reliability of GRV volumes. For accurate measurement, the tip of the tube has to be in the “sump” of the stomach, which may vary with position, and the tip has to be free of the mucosal folds for suction to be successful. Although gastric emptying is reduced in any form of critical illness, it usually continues with a higher emptying volume, and provided the feed is given at a rate equal to or below that rate, NG feeding can be remarkably well tolerated by most ICU patients. Despite the discussion above of the results of controlled feeding trials, it would seem logical that the risk of feed aspiration would be higher in NG-fed patients. Indeed, the most recent published meta-analysis of 15 randomized controlled trials concluded that post-pyloric feeding reduced the risk of aspiration pneumonia [155]. However, earlier meta-analyses have failed to show significant differences [156], and so the benefit of post-pyloric placement must be small and suppressible by the implementation of good management practices. As previously discussed, a recent study has shown that resting and food-stimulated CCK responses are increased in ICU patients, providing a humoral explanation for delayed gastric emptying [157]. Consequently, duodenal feeding can be expected to worsen gastric emptying, indicating the potential need for simultaneous gastric decompression. Prokinetic agents, such as metoclopramide, or erythromycin, should be used in adequate IV doses (e.g., metoclopramide 10 mg IV 8 hourly) to increase gastric emptying and thus enhance feed tolerance [158]. If residuals continue to be high, the feeding tube should be placed passed the pylorus, short term PPIs should be used to suppress gastric secretions, and continuous gastric decompression should be considered.

Nasoenteric

Post-pyloric Placement

Post-pyloric placement can be performed by specialized nursing staff or the “enteral tube team” but the sicker the patient, the lower the success rate, as gastric emptying decreases progressively with the increasing severity of illness. Failure at manual placement means referral either to IR or to gastroenterology for endoscopic/fluoroscopic-guided placement.

Benefits: Increases feeding capacity in patients with poor gastric emptying, such as those patients with critical illness or diabetes [145]. If tubes are placed >40 cm down the jejunum past the ligament of Treitz, enteral feeding can be given without stimulating the pancreas—an optimal approach for the management of patients with severe acute pancreatitis [159–161]. However, it must be remembered that the avoidance of pancreatic secretion will impair digestion and so an elemental formula feed should be used when feeding at this level.

Complications: More invasive procedure, therefore more risk of mucosal injury particularly if attempted manually. The regulatory role of the stomach will be bypassed, potentially resulting in overfeeding, with diarrhea, distension, bloating, and discomfort. It is imperative to remember that bolus feeding must NEVER be used. This is appropriate for the stomach, but not the small intestine, which is designed to receive slow constant infusions of emulsified food from the stomach. Boluses of feed injected into the jejunum may produce “dumping” symptoms. Likewise, tube flushes should not be the same in the stomach and intestine: while gastric tube flushes (or “hydration fluids”) with 100–200 cc of water are commonly well tolerated, jejunal flushes should be restricted to 25 cc/4 h. Contaminated feeds are more likely to cause GI disturbances with post-pyloric infusion, as bacterial suppression by gastric acid is lost. Consequently, a sterile feed delivered through a prepackaged “closed” system should be employed.

Placement

• Manual: Many large hospitals have formed an enteral feeding team consisting of nurses with specialized training and experience in feeding tube placement. This undoubtedly increases success and reduces the risk of complications. Such personnel become good at not only placing NG feeding tubes, but also post-pyloric placement. A number of bedside techniques have been published. Transpyloric migration is facilitated if 500 cc of air is injected into the stomach when the tip of the feeding tube is in the antrum. This stimulates gastric emptying and the tube can then be gently advanced through the pylorus, after having turned the patient on his/her right side. This is usually effective in patients with normal motility and anatomy, but will fail if there is gastroparesis, distortion, or compression of upper GI tract.

• Interventional Radiology: In patients with distortion or dysmotility of the upper GI tract, the next step would be to send the patient to interventional radiology (IR) for post-pyloric guidance. The downside of this approach is that it interferes with patient management, which could be a critical issue in patients needing ICU management. Secondly, there are usual delays of 24–48 h before the procedure can be accommodated. The alternative is to use endoscopy, which can be performed immediately, if necessary, at the bedside in ICU patients, or in the endoscopy suite in more stable patients.

• Endoscopy: (a) Per Oral: Traditionally, a regular 27 French upper GI endoscope is used. The duodenum is cannulated as far as possible and then a guide wire is placed through the endoscope and advanced under vision down the duodenum and then, blindly, around and past the ligament of Treitz. The endoscope is then withdrawn leaving the guide wire in position. This leaves the proximal end of the guide wire emerging from the patient’s mouth, and another awkward procedure is required to transfer the wire through the nose before the feeding tube can be deployed. To do this, a slim (e.g., 6 French) 12 in. length of tubing is passed down the nostril into the retropharynx where it is grasped with a pair of forceps and pulled out of the mouth. The proximal end of the guide wire is then threaded through the distal (oral) end of the small tube and pushed through so that it emerges from the proximal end of the small tube out the nose. The small tube is then pulled out of the nose and the guide wire is straightened, allowing passage of the feeding tube through the nose, through the stomach, and down the duodenum. Other endoscopic techniques are to first place the feeding tube down the nose and into the stomach, and then to pass a regular EGD scope into the stomach. When the tip of the tube is identified, it is grasped with an endoscopic biopsy forceps or snare and dragged with the endoscope through the pylorus and as far down the duodenum as possible. In our experience, this rarely works as the best placement will only be within the duodenum and the tube commonly come back into the stomach with the endoscope. One solution is to use an endoscopic clip to tether the tip of the tube to the duodenal mucosa, but this produces mucosal injury.

Transnasal Endoscopic Placement of Feeding Tubes

• This is what we prefer, because

• (1) It obviates the need to perform another uncomfortable procedure at the end of a per oral endoscopic feeding tube placement to hook the feeding tube through the nose,

• (2) Because it is easier to cannulate compressed or stenosed segments of bowel under direct vision, and

• (3) Because it is safer to see where you are going.

We use a standard ultra-slim 5.8-mm EGD endoscope (made by all the major endoscope manufacturers) that will permit transnasal intubation of the GI tract [116, 162]. The smaller 5.0-mm endoscope may also be used, but it is less versatile as it only has one hand control, with up–down and no side to side tip movement. Placement of the guide wire through the endoscope is the same as described above, but this time the wire will emerge out of the nose with the endoscope, thus negating the need for an additional uncomfortable procedure. The most effective way of ensuring placement of feeding tubes far down the jejunum is to use the combination of transnasal endoscopy with fluoroscopy, as the guide wire position can be manipulated under fluoroscopy into a confirmed position far down the jejunum. However, this is only suitable for patients who can be transported to the endoscopy suite. In other patients in the ICU, the deployment of the guide wire beyond the ligament of Treitz and down the jejunum has to be done blind, which is usually not that difficult with practice. This technique is relatively easy for the average endoscopist to learn. Our GI fellows in training are usually competent after 4 weeks working with us.

The Nasogastric Decompression–Jejunal Feeding Tube (NGJ): A teaching video recording has been placed on U-Tube for further details on endoscopy and placement of NGJs (https://www.youtube.com/watch?v=YlkiY6YTnYk and https://www.youtube.com/watch?v=UjVjDkdDLW0). We find this procedure particularly useful in feeding patients with acute severe pancreatitis in the ICU without the need to transport them out of the unit, as it bypasses the compressed upper GI tract and permits placement of the tip of the feeding tube >40 cm past the ligament of Treitz where feeding does not stimulate pancreatic secretion, thereby proving enteral feeding without pancreatic stimulation [159]. However, as discussed earlier, the method is also useful in any ICU patient with persistently high GRVs, and in any patient with gastroparesis and gastric outlet obstruction, where placement of the J-tube not so far down, at or around the ligament of Treitz, is sufficient [116]. Examples of these problems and their solution with NG-J placement are discussed at the end of this chapter. We use a double-lumen tube (NG-J), which contains a shorter, thicker, outer tube, which is left proximal to the obstructed segment in order to provide proximal gastric decompression (Fig. 8.2). Two examples in the USA are the “StayPut” (Nestle Nutritionals, MN), which is a fixed system with an 18 Fr gastric decompression tube and a 9 Fr J-tube that will reach the ligament of Treitz which is most useful in general ICU patients, and the “Kangaroo-Dobhoff” system (Covidien-Kendall-Tyco Products, MA), which has a 16 Fr 97 cm outer gastric decompression tube and an inner 9 Fr 170 cm adjustable jejunal extension that permits non-stimulatory mid-jejunal (>40 cm) feeding in patients with acute pancreatitis with gastric decompression (Fig. 8.2). It is critically important to decompress the upper GI tract if there is compression and obstruction of the antrum or duodenum. If there is accumulation of fluid in the stomach proximal to the obstruction, the patient will vomit and the whole system is likely to be dislodged. In a similar way, it is an ideal method of providing enteral feeding for patients with chronic pancreatic disorders and pseudocysts who also have compressed and distorted upper GI tracts and where pancreatic stimulation is best avoided. They are commonly diabetic, which adds gastric dysmotility and exacerbates the defective gastric emptying. These patients are usually mobile, allowing placement in the GI lab with fluoroscopic assistance. Other common indications are post-partial gastrectomy and post-Whipple’s procedure, when there is functional impairment of gastric emptying resulting in vomiting after meals. Although many feeding tube systems come with a guide wire included in the kit, in our experience it is preferable to use specialized ERCP-type guide wires (e.g., JAG-wire, Wilson-Cook, NC) as they are much easier to manipulate through the loops of small intestine distal to the ligament of Treitz and do not kink. As with all procedures, the more frequently these tubes are placed, the easier it becomes. There are specific tips. For example, when withdrawing the endoscope it is very important to make sure that no loops of redundant guide wire are left within the stomach otherwise the whole system simply rolls back into the stomach when the feeding tube is advanced. Secondly, it is useful to reintroduce the endoscope through the opposing nostril while feeding the feeding tube over the guide wire and to watch the progress of the tip of the tube straight down the greater curve and through the pylorus. This ensures the feeding tube and guide wire take the shortest route to the pylorus and that there are no gastric loops.

Fig. 8.2

The NG-J feeding tube system illustrating the double tube within-a-tube conformation allowing proximal GI decompression and distal feeding into the unobstructed jejunum

We published our initial experience with these placements (chiefly StayPuts) in 57 ICU patients in 2003 [162]. All patients had been referred to us for TPN as they had high GRVs associated with respiratory failure, acute head injury, and acute pancreatitis. Placement was successful in >90% and TPN was avoided in 77%. As an additional bonus, previously unrecognized esophageal and gastric pathology was identified in 60% of patients. Recently, we updated our experience with double-lumen (NG-J) tubes anchored in place with nasal bridles, but this time we only used the Covidien adjustable tubes (Kangaroo-Dobhoffs), as the majority of patients needed pancreatic rest and so the J-tube was placed approximately 40 cm down the jejunum [116]. The study included 50 consecutive patients with obstructions chiefly due to complicated severe acute (n = 31) or chronic cystic pancreatitis (n = 11). The chief outcome parameters are shown in Table 8.2. Duration of subsequent feeding ranged from 1 to 145 days, median 25 days, and 19 patients were discharged home (HPN, see Chapter 10) with their tube feeds. Only 1 patient could not tolerate feeding and needed to be converted to parenteral feeding. Average tube life was 14 days, replacement being needed most commonly for kinking or clogging of the jejunal tube (56%) or accidental dislodgement (24%). The obstruction resolved spontaneously in 60% allowing resumption of normal eating. Of the patients with severe acute pancreatitis or pancreatic pseudocysts, pancreatic rest resulted in resolution of the disease without surgery in 87%, and need for surgery in the remainder was put off for 31–76 days. Seven patients died, most from complications of acute pancreatitis between days 1 and 31. We concluded that NGJ feeding provided a safe conservative management for critically ill patients with upper GI obstructions, reducing the need for surgery and parenteral feeding.

Table 8.2 Analysis of final outcome parameters

The Nasal Bridle

One of the most frustrating things in interventional feeding tube placement is to struggle to successfully place the tube and to be told the next day that “the tube fell out”! Many of the patients we place tubes in are confused and disorientated and it doesn’t take much effort to dislodge a tube fastened to the nose with a piece of tape. The development of nasal bridles was, therefore, for us a divine event, as it increased the life span of the average feeding tube in ICU patients from days to weeks. We currently use the AMT bridle (Applied Medical Technology, Inc., Brecksville, OH 44141 USA) that consists of two 10-cm four French flexible plastic probes with magnets at the tips of both and a thin cotton tape attached to the other end of one. After lubrication with xylocaine-containing gel, the probes are passed up each nostril until they pass the end of the septum, whence a “click” can be felt, indicating interconnection of the two magnets. The tube without the tape is then pulled out, dragging the other with the tape out. The tape then forms a bridle around the septum, and the two ends are clamped onto the feeding tube to secure it in place (Fig. 8.3). Initial fears about septal injury have not been realized, and we have some patients who go home with the bridles and feeding tubes for months on end [116].

Fig. 8.3

Illustration of a double-lumen nasogastric decompression (note bile-stained fluid being aspirated) and jejunal feeding tube (the smaller one going through) anchored in place with a nasal bridle clipped onto the tube at the nose with tape straddling the nasal septum

Percutaneous

Endoscopic Gastrostomy (PEG)

Indications: The indications for PEG placement are the same as those for enteral feeding detailed above, but, because the procedure is invasive, necessitating perforation of the bowel, and has significant potential complications, its use should only be considered in patients who will require enteral feeding for longer than 4 weeks. Continuing nasoenteric feeding beyond that time is uncomfortable and unsightly for the mobile recuperating patient, and can produce nasal injury and sinusitis. However, gastroenterologists must not simply act as technicians when asked to place PEGs, as their complications can be severe and life-threatening. Every attempt should be made to wean patients off enteral feeding and back onto food before contemplating PEGs. Although serious complications are rare in stable patients, they become increasingly more common in critically ill and debilitated patients. It must be stressed that the placement necessitates perforation of the bowel which can lead to intra-abdominal infections and bleeding. Sometimes, the family or nursing home demand PEG placement to ease the care of chronically debilitated patients, but remember that many patients will strongly resist placement—for good reasons—even if their food intake is low. Eating may be one of the few pleasures remaining for them. Any experienced gastroenterologist can place PEGs, but they must be aware of the potential risks and benefits, and only proceed if there is good evidence that the procedure will improve nutrition and quality of life. For example, they should never be placed in the terminally ill for maintaining feeding, as there is no evidence that maintaining feeding in this situation improves quality of remaining life or survival: in fact it might do the reverse by introducing complications. PEG feeding should also be avoided in the elderly with advanced dementia, who have lost the will to eat and outcome is poor. It is a bitter pill for the family to often accept that starvation is part of the dying process, but patients must be allowed to die with dignity, and not assaulted with invasive or disfiguring procedures. This will be covered in more detail in Chapter 17. The reader is referred to the vast literature on this subject [163, 164].

Placement: As mentioned above, the gastroenterologist must remember that he/she is not simply a technician, and prior to PEG placement the past medical and surgical history needs to be taken, the abdomen must be examined carefully to illicit signs of previous surgeries that might have altered their anatomies, and to feel for masses that might impair percutaneous access. It is also important to avoid placement in patients with active infections; better to get the infection under control beforehand. PEGs should also be delayed in patients with recent histories of myocardial infarction, and in patients where anticoagulation cannot be temporarily held. The procedure requires sedation and, in patients with poor respiratory reserve or cardiac failure, may result in respiratory or cardiac arrest. Consequently, it is our practice in such patients to use monitored anesthesia care so that maximum sedation can be used in safe, expert hands. Controlled studies have shown that a single IV dose of broad spectrum antibiotics (e.g., 1 g cefazolin), given immediately before the procedure, reduces the risk of more serious invasive infections [165]. A full diagnostic esophagogastroduodenoscopy (EGD) must first be performed to exclude the possibility of mucosal disease, such as gastric ulceration, and to ensure there is no gastric outlet obstruction or impairment due to the presence of distal disease. The urgency of placement must be reconsidered if ulceration is detected: it is better to postpone placement when significant ulceration (i.e., ulcers with depth) is found and to treat appropriately before trying once more. On the other hand, PEG placement might be important in the presence of superficial ulceration to commence feeding and stimulate mucosal repair, or to improve drug delivery. Infection at the abdominal incision site is universal, but usually mild, resolving with time. Placement complications such as perforation or fixation of the colon, liver, or spleen can be minimized if PEG placement is only attempted when there is good transillumination of the light on the end of the endoscope through the abdominal wall, and if a CT scan is performed beforehand (this is almost done as routine in sick patients nowadays!) to make sure there is a “window” of access to the stomach. After PEG placement, a repeat endoscopy must be performed immediately to check the position of the internal bumper and the tension between the internal and external bumpers, and to inspect for complications such as bleeding. If it is too loose, the internal bumper might migrate and create an obstruction; if too tight, a “buried bumper” may occur when the bumper is pulled beneath the mucosal surface.

Systems: A wide variety of commercial PEG systems are available. The sizes commonly used range between 18 and 28 French. In general, small diameter tubes should be avoided in patients with poor gastric emptying and the need for intragastric medication administration. If a jejunal extension is required (PEG-J), for example in patients with gastroparesis, then a wider tube will be required (e.g., 24 French), enabling cannulation with a smaller diameter (e.g., 16 French) jejunal (J) tube. In patients with severely distorted upper GI tracts, it might be necessary to use a 28 French tube as this allows cannulation with a 5-mm endoscope and placement of a guide wire well down the distorted small intestine for jejunal tube placement in a similar fashion to that described for NGJs. The placement of J extensions is described below.

The risks of PEG placement can be reduced if the following safeguards are maintained:

• 1. Use monitored anesthesia care in unstable patients.

• 2. Provide prophylactic antibiotics.

• 3. Always perform a complete EGD to rule out the presence of disease. Active deep ulceration should ideally be treated before PEG placement.

• 4. When placing the PEG make sure that good transillumination of the endoscope light through the anterior abdominal wall is seen with the overhead lights darkened. Simple palpation of the abdomen looking for indentation of the gastric wall is not good enough for PEG siting, as the colon can be transfixed between the stomach and the anterior abdominal wall and finger palpation will still be seen.

• 5. Avoid placement in patients with ascites or on peritoneal dialysis. If a PEG has to be placed in a patient with ascites, make sure the fluid is drained beforehand and the abdomen is lax.

• 6. Avoid placement in head injury patients who have had a VP shunt placed within the past week.

• 7. Ensure that a coagulopathy is not present and that the patient is not on anticoagulants. Subcutaneous heparin should be held at least 6 h prior to the procedure. If platelets are reduced below 50,000 cells/cm, provide platelet cover at the time of placement.

• 8. Do not perform the procedure single-handed. Use an endoscopic assistant (GI fellow or GI colleague) to help with placement so that sterility is maintained. Four hands on deck!

• 9. After insertion of the PEG, re-endoscope to check final placement and to detect complications.

• 10. Do not overtighten or undertighten the internal–external bumper system. The internal bumper should be flush with the mucosa and should swivel easily.

Management: Many of the late complications of PEGs can be avoided with good nursing practices. Initially, the incision site should be covered with gauze for protection and the external tube fastened to the abdominal wall with a Stat-Lock as shown in Fig. 8.4 to prevent pulling. We usually advise initial q4-h water flushes, but bolus feeding can be started by 8 h in stable patients, with good intestinal function. Leakage of the feed is unusual in this circumstance, as normal motility will propel the feed down to the duodenum, rather than out the PEG orifice. The site should be examined daily to make sure the external bumper is not too tight and that there is no infection or discharge, and cleaned with an antiseptic. With time, 4–8 weeks, the gastrocutaneous passage becomes epithelialized and becomes established by fibrosis such that if the tube is pulled or falls out, another can be safely replaced at the bedside without re-endoscopy. However, the hole closes over rapidly and so a replacement should be placed asap, within 6–8 h. In an emergency, any simple tube, such as a Foley’s catheter, can be used. If there is ever any difficulty inserting a replacement tube, it is mandatory to check the position by X-ray after contrast injection, before restarting tube feeding. Within the first month of placement, loss of the tube will necessitate replacement by endoscopy, as there is a real chance of misplacement of the manually inserted tube into the peritoneum. Healed exit sites are best cleaned with soap and water and kept dry and protected by loose gauze dressings. Some patients continue to have a serous discharge associated with “granulation” tissue as shown in Fig. 8.4. This is best treated by monthly applications of silver nitrate, contained in swabs at the end of sticks, until the red tissue retracts and the skin around the hole becomes totally healed as in Fig. 8.4.

Fig. 8.4

Left: Well-healed insertion site, managed with soap and water toilet, daily dry dressings if necessary. Right: Healed with “granulation” tissue associated with mucoid discharge. Treated with initial and then monthly applications of silver nitrate until fibrosed, barrier crèmes, daily re-dressings until dry

Percutaneous Radiological Gastrostomy

Gastrostomy placement can be performed relatively simply by interventional radiologists. However, due to manpower reasons, IR usually are only asked to place tubes when endoscopy is impossible because of stomal, pharyngeal, or esophageal obstruction, or when PEG fails, for example due to failure to observe endoscopic transillumination. Generally the technique is based on (a) the insertion of a needle into the lumen of the stomach for inflation under fluoroscopic or ultrasound guidance, (b) the placement of “fasteners,” i.e., small needles with tethers that can be deployed once inside the stomach, which hold the stomach wall against the anterior abdominal wall while (c) a trochar with graduated dilators is used to pierce the two layers, ultimately allowing (d) the placement of a gastrostomy tube with an inflatable internal bumper.

Surgical Gastrostomy

This is the classic method of gaining gastric access for feeding or decompression. Numerous comparative studies have been performed which have compared endoscopic to surgical gastrostomy placement, and all indicate that the endoscopic method is a quicker and less complicated approach to take. However, the procedure is simple and we usually refer all patients in whom safe endoscopic placement cannot be performed.

PEG-J

As mentioned above, this consists of the passage of a secondary feeding tube through the PEG and through the pylorus into the duodenum or proximal jejunum. With practice, this procedure is fairly easy to perform, but again a skilled assistant is essential, as you need four hands. The first step is to place a PEG as described above. Next, the provided guide wire is placed through the PEG tube and grasped with endoscopic forceps. The forceps are then shortened to the end of the endoscope and the tip of the endoscope is advanced with the wire through the pylorus as far down the duodenum as possible (4th part). It is always important to cannulate the duodenum before this procedure is performed to know where you are going, because the first part of the duodenum is often deformed in patients with gastric outlet obstruction needing J-tube placement! Blind advancement can injure the duodenal cap mucosa with the tip of the forceps and guide wire. Once the final position in the duodenum is achieved, the secondary jejunal feeding tube is passed over the guide wire from outside the body through the PEG, holding the wire taught to prevent looping in the stomach. It helps to advance the biopsy forceps slightly forward of the endoscope so that the tip of the J tube can be observed under direct vision to have reached the end of the guide wire where it is held by the forceps. Once the J-tube tip is seen, the biopsy forceps are slowly fed through the end of the endoscope while the endoscopy is synchronously pulled back from the duodenum into the stomach, leaving the biopsy forceps, guide wire, and J tube well down the duodenum. When the endoscope is looking at the pylorus, the guide wire is released from the forceps, and the forceps are pulled out of the endoscope. The endoscope is then brought back to the PEG internal bumper to check that the J tube goes direct from the internal PEG bumper to the pylorus, without looping in the stomach. This is good placement, and the endoscopy may now be safely withdrawn. The alternative is to perform the “exchange” and guide wire release when the endoscope is still well down the duodenum, but the J-tube often sticks to the endoscope and flips back into the stomach when the endoscope is withdrawn. In some patients with gastroparesis and large stomachs, the position of the internal PEG bumper is often facing the esophagus rather than the pylorus, and no matter what you do, the J tube will loop within the stomach before entering the pylorus. This is bad placement, as the whole tube will shortly fall back into the stomach. In this situation it can be helpful to loosen the external bumper of the PEG tube and to push the tube 5–10 cm into the stomach so that the bumper now faces the pylorus. Another alternative is to use a 28 French PEG tube and then use an ultra-slim 5.0-mm endoscope (NB: not the one we use for transnasal endoscopy which has a diameter of 5.8 mm, which is too big) which may be passed directly through the PEG and advanced through the duodenum for deployment of the guide wire and feeding tube, akin to the method described above for NG-J placement. The endoscope is then withdrawn, leaving the guide wire in the jejunum, and the J-tube is then passed over the guide wire, through the PEG and past the pylorus into the jejunum. With the rapid advancement of endoscopic technology, it can be predicted that this form of endoscopic-assisted placement will become standard with time.

Direct PEJ

In some patients with gastroparesis or surgically altered anatomy, the J-tube of a PEG-J frequently recoils back into the stomach, necessitating repeat endoscopies for repositioning. In others, a gastrostomy is impossible due to surgical resection. In these situations, a feeding jejunostomy will be needed. Most commonly this is performed surgically by open or laparoscopic technique, but it is always worth first performing an upper GI endoscopy to pass the tip of the endoscope as far as possible down the small intestine to try and detect abdominal wall transillumination, so that a direct PEJ can be placed [166]. Percutaneous endoscopic jejunostomy (PEJ) placement consists of an endoscopic procedure very similar to PEG, but involves the direct placement of the percutaneous tube into a loop of small intestine that lies immediately below the abdominal wall allowing clear transillumination of the endoscope tip through the abdominal wall after the loop is distended with air. The use of direct PEJ is more risky as the small bowel is more mobile and of smaller size, and satisfactory transillumination is not often seen. The same precautions used for standard PEG placement need to be imposed. A smaller diameter (e.g., 16–18 French) PEG tube is preferable, but systems up to 22 French often work well and do not obstruct the lumen if tethered correctly.

Surgical Jejunostomy

This is the most common form of percutaneous jejunostomy used as it is often placed at the end of a complicated abdominal surgery when there is concern that upper GI function may be impaired by the surgery and take a while to return, for example after a partial gastrectomy, a Whipple’s procedure, or any other form of complex esophageal, gastric, duodenal or pancreatic surgery. Under direct vision, a loop of small intestine is approximated to the abdominal wall and the tube is tunneled caudally to maintain position and facilitate motility.

Laparoscopic Jejunostomy

With the increasing use of and expertise in minimally invasive laparoscopic surgery, laparoscopic placement of feeding tubes is becoming more popular. The indications are those for PEJ, namely the need for long-term access and home enteral feeding, and the advantages over PEJ are that it doesn’t rely on abdominal transillumination of the bowel lumen and therefore is possible in a wider population of patients.

Recent Developments: Situations Where Enteral Feeding Should Now Be Used Instead of TPN

Dysmotility and Ileus

Classically, ileus, or “paralytic ileus” as it is commonly termed, occurs in patients following major abdominal surgery and also from critical illness. It is evidenced by the absence of bowel sounds, the typical X-ray appearance of distended intestinal loops of bowel with or without air fluid levels, nausea and vomiting, and the absence of stooling and flatus. For years, the dogma was not to give any oral or enteral feeding to patients with ileus until bowel sounds were heard, and flatus is passed. This actually perpetuated the problem, as the most effective treatment for ileus is the resumption of feeding. This practice was finally debunked by the classic study of Waldhausen et al. in 1990 [167]. During major abdominal surgery, they inserted electrodes into the stomach, jejunum and colon allowing them to pace the electrical conductivity of the intestine in the recovery period.

Remarkably, basic electrical activity showed no changes in frequency after surgery and the electrical activity was poorly associated with the presence of absence of bowel sounds. A number of studies have subsequently shown that postoperative ileus is not equated with intestinal failure, and that bowel function and nutrient absorbing capacity may be suppressed, but not absent [168]. For example, intraoperative percutaneous jejunostomy feeding tube placement allowed enteral feeding to resume within 24 h of major surgery with good tolerance and absorption, and controlled trials have shown that most patients tolerate oral intakes immediately following more moderate forms of surgery [169, 170]. Figure 8.5 illustrates a patient admitted to hospital with vomiting, abdominal pain, ARDS, and ileus due to severe acute pancreatitis. Many would consider this intestinal failure, and prescribe NG decompression, NPO, and TPN. A double-lumen NGJ tube was successfully placed by transnasal endoscopy as shown in the X-ray enabling simultaneous gastric decompression and the commencement of slow continuous (25 ml/h) jejunal feeding. Ileus resolved within 24 h, stooling returned, and TPN was never needed. The critical importance of this observation is that there are an accumulating number of studies that suggest that early enteral feeding improves outcome in ICU-ventilated patients (Table 8.3), and so attempts at enteral feeding, initially at a slow rate with NG suction, should not be delayed by X-ray or clinical evidence of ileus.

Fig. 8.5

Patient with severe acute pancreatitis, admitted with radiographic ileus: NG-J feeding tube has been positioned

Table 8.3 Investigations of early enteral feeding in medical and surgical patient populations

The etiology of ileus following abdominal surgery and in non-surgical critical illness is similar and equally complex, as recently reviewed by Caddell et al. [171]. As they suggest, a problem with multiple potential causes will have multiple potential treatments, as illustrated by Fig. 8.6. Prokinetic agents such as metoclopramide and erythromycin are effective in the short term in improving gastric emptying, but long-term use is impaired by decreased efficacy and side effects. Our experience is that virtually all these patients can be fed enterally, if a feeding tube is placed past the ligament of Treitz (i.e., in the jejunum) and provided gastric decompression is maintained. This remains one of the most distressing battlegrounds with some traditional surgeons who continue to demand PN for their postoperative patients with ileus! Education works, and the above references should be distributed. The use of an NGJ tube ensures safe gastric decompression while jejunal feeding can be started slowly (20 cc/h) and advanced with tolerance, with the resultant resolution of ileus.

Fig. 8.6

Upper GI dysfunction in critically ill patients has multiple causes, and therefore multiple treatments (Caddell et al. [171])

Inflammation, electrolyte abnormalities, edema, neurotransmitter changes and opiates all play a role. For this reason, the treatment of ileus is also complex, but improvement can be expected from any form of treatment that removes the underlying disease and promotes recovery. In practice, early enteral feeding and the minimization of narcotic use are probably the most effective forms of treatment, as luminal nutrients stimulate secretions and peristalsis, prevent intestinal inflammatory cytokine production and bacterial overgrowth, and maintain gut function. This accounts for the change in current practice to start enteral feeding early but slowly, in the course of the hospitalization. As gastric emptying is most commonly disturbed, it is best to place a feeding tube post-pyloric and commence feeding with a regular polymeric diet at a slow rate (20–25 cc/h) the day following major surgery or on admission to the ICU with any form of critical illness. As reviewed earlier, this rate may be continued for the first week providing what has been termed “trophic feeding,” to maintain gut function and mucosal integrity. Colonic function is the last to return and so early consideration, i.e., by the end of the first week, of the addition of a soluble fiber source (e.g., wheat dextrin 10–20 g/day) to the feeding schedule is wise.

Gastric Outlet Obstruction

Perhaps the most common indication for TPN up until recently was poor gastric emptying and high NG aspirates associated with critical illness. Gastric emptying is slow, as is general gut motility, in most ICU patients. Fraser and colleagues have shown that CCK levels are commonly increased in ICU patients, and jejunal feeding will increase levels further [157], leading to further reductions in gastric emptying. These physiological observations, coupled to experience in practice, have lead us to advocate the use of double-lumen feeding tubes “NG-J” tubes as described above (e.g., “StayPut,” Novartis; “Kangaroo-Dobhoff,” Kendal), with a distal jejunal feeding tube and a proximal gastric decompression tube. Our first experience with placement of double-lumen tubes by transnasal endoscopy (Fig. 7.3) was reported in 2003 on 51 consecutive ICU patients referred to us for TPN, because of failure to be able to use the gut because of “high gastric residuals and ileus” [162]. The demographics are summarized in Table 8.4 showing the variety of underlying conditions associated with disturbed upper GI function.

Table 8.4 Illustration of the wide variety of patients with upper GI dysfunction (gastroparesis) treated successfully with NGJ feeding tube placement (O’Keefe et al. [115])

Initial placement was successful in all apart from five patients with massive gastric dilatation and acute pancreatitis complicated by duodenal compression who required fluoroscopic guidance. In confirming correct tube placement, there was a near-perfect concordance between re-endoscopy and X-ray (45/46) indicating that there was no need to confirm placement with abdominal X-rays. Figure 8.7 illustrates correct placement, with the arrows marking the stomach, the pylorus, and the ligament of Treitz. Previously unrecognized upper gastrointestinal tract pathology was detected in most patients, with acute gastritis in 47, superficial gastric ulceration in 24, and erosive esophagitis in 5. These results led us to conclude that endoscopic placement of these feeding tubes in the ICU is quick, effective, and minimally disruptive of intensive therapy. In addition, it can reveal unrecognized pathology, which potentially could lead to improvements in overall medical care [162]. Our most recent published experience on 50 further ICU patients, predominantly with gastric outlet obstruction due to severe necrotizing pancreatitis, endorses this view, with only one patient failing enteral feeding and needing PN [116].

Fig. 8.7

The endoscope has been withdrawn from the patient and the double-lumen feeding tube (“StayPut,” Novartis) has been passed over the wire, through the nose, and into the correct position with the outer gastric aspiration ports within the gastric sump (arrow top right), the internal jejunal tube passing through the pylorus (left arrow) and going through the ligament of Treitz (right lower arrow) into the jejunum

Secondly, masses in the esophagus, stomach, and duodenum, or masses outside the lumen causing extrinsic compression, can compress the upper GI tract, causing obstruction. However, the bowel distal to the obstruction remains functional—if it can be accessed. The classic example is patients with acute pancreatitis or chronic pancreatitis with large pseudocysts producing extrinsic compression of the upper GI tract (Fig. 8.8). Barium studies performed on these patients commonly demonstrate “high-grade” obstruction of the outlet of the stomach or the duodenum. Great care has to be taken when attempting intubation in these patients, as the stomach is usually full of fluid and gagging may precipitate aspiration. Attempts at manual/bedside or even radiological placement of feeding tubes through the obstructed segment usually fail in this situation. The use of endoscopy allows the compressed segment of bowel to be traversed by the endoscope using direct vision. Once past the obstruction, the distal bowel is seen to be open and functional with active peristalsis, allowing placement of a guide wire through the endoscope into the functional bowel. As previously described under NGJ feeding above, the endoscope is then retracted leaving the guide wire straddling the obstructed segment. A feeding tube may then be fed over the guide wire from the nose into the functional bowel allowing resumption of full enteral feeding, with no need for TPN. At the same time, it is critically important to decompress the bowel proximal to the obstruction, as secretions will collect inducing nausea and vomiting, which might dislodge the jejunal feeding tube and lead to aspiration. In more chronic conditions, a percutaneous endoscopic jejunostomy (PEJ) can be placed directly into the patent bowel distally. Alternatively, if the patient requires surgery, a surgical jejunostomy may also be placed to gain access to functional bowel.

Fig. 8.8

Examples of gastric compression due to extrinsic compression by pancreatic masses and the restoration of gut function by transnasal endoscopic placement of a double-lumen nasogastric decompression jejunal feeding tube (NG-J). (a) endoscopic view of a large pancreatic pseudocyst compressing the stomach, (b) CT scan of the same, (c) successful placement of an NG-J feeding tube with tip 40 cm past ligament of Trietz

Severe Diarrhea

Critically ill patients given tube feeding frequently develop diarrhea, but the composition of the feed is rarely to blame. As has already been discussed under Why Might Enteral Feeding Improve Outcome? (Chapter 5, p. 42) diarrhea is usually a consequence of disturbed microbiota due to the combined effects of antibiotic therapy, PPIs, and bowel rest [34, 97]. As discussed in Chapter 5 (p. 48), recent studies have shown that the microbiota are essential for the maintenance of colonic function and health. In critical illness, they suffer a double hit from (a) the almost universal use of antibiotic therapy and (b) the use of non-residue tube feed formulae that derive them of their nutrition [34]. PPIs also disturb the composition of the microbiota and can precipitate diarrhea [172]. Diarrhea or loose stools are particularly common in patients at commencement of feeding. It is extremely important not to hold the feeding until the diarrhea resolves as this starves the microbiota further exacerbating dysbiosis and adds to the rate of nutritional deterioration. With continued feeding at a slow rate of 20–40 cc/h, the microbiota slowly recovers and the diarrhea usually improves after 3 days, allowing advancement in feeding rate to goal. If the diarrhea continues, efforts should be made to restrict antibiotic and PPI use, and to investigate other common causes of diarrhea, for example the use of sorbitol-containing liquid medications. In patients who need continued antibiotic therapy, it is important to add a probiotic to support the microbiota.

Gastroesophageal Reflux and Aspiration Risk

Although the presence of a feeding tube in the lower esophagus and stomach reduces the competency of the lower esophageal sphincter, reflux is chiefly caused by factors such as sepsis, trauma, drugs, body position, gastroparesis, esophageal dysmotility, and obesity. Aspiration in this situation is, however, preventable, even in nasogastrically fed patients, by the adoption of strict management protocols, as detailed under NG feeding above. Consequently, there is never a need for TPN in this group of patients.