Enteral tolerance in critically ill patients

Enteral nutrition (EN) can maintain the structure and function of the gastrointestinal mucosa better than parenteral nutrition. In critically ill patients, EN must be discontinued or interrupted, if gastrointestinal complications, particularly vomiting and bowel movement disorders, do not resolve with appropriate management. To avoid such gastrointestinal complications, EN should be started as soon as possible with a small amount of EN first and gradually increased. EN itself may also promote intestinal peristalsis. The measures to decrease the risk of reflux and aspiration include elevation the head of the bed (30° to 45°), switch to continuous administration, administration of prokinetic drugs or narcotic antagonists to promote gastrointestinal motility, and switch to jejunal access (postpyloric route). Moreover, the control of bowel movement is also important for intensive care and management. In particular, prolonged diarrhea can cause deficiency in nutrient absorption, malnutrition, and increase in mortality. In addition, diarrhea may cause a decrease the circulating blood volume, metabolic acidosis, electrolyte abnormalities, and contamination of surgical wounds and pressure ulcers. If diarrhea occurs in critically ill patients on EN management, it is important to determine whether diarrhea is EN-related or not. After ruling out the other causes of diarrhea, the measures to prevent EN-related diarrhea include switch to continuous infusion, switch to gastric feeding, adjustment of agents that improve gastrointestinal peristalsis or laxative, administration of antidiarrheal drugs, changing the type of EN formula, and semisolidification of EN formula. One of the best ways to success for EN management is to continue as long as possible without interruption and discontinuation of EN easily by appropriate measures, even if gastrointestinal complications occur.


Introduction
The difficulties and complications related with enteral nutrition (EN) include metabolic disorders, such as increase in blood glucose level and electrolyte abnormality, but this article will mainly describe the gastrointestinal complications, particularly vomiting and bowel movement disorders (i.e., diarrhea and constipation), which are frequently observed during EN and are important to address. If such gastrointestinal symptoms do not resolve with appropriate management, EN must be discontinued or interrupted, and parenteral nutrition (PN) must be easily initiated.

Intestinal intolerance and confirmation of gastrointestinal peristalsis
Regardless of EN administration, several gastrointestinal disorders and symptoms can still occur in critically ill patients and can be precipitated by several factors, including diseases, general condition, and metabolic state before the onset, setting of respirator, and administered drugs [1]. Gastrointestinal disorders may relate to intestinal intolerance during EN. The mechanisms of gastrointestinal disorders in critically ill patients or postoperative patients can be classified as failure of mucosal barriers, attenuation of gastrointestinal peristalsis and atrophy of intestinal mucosa, decrease of gut-associated lymphatic tissue and so on [2].
Previously, ingestion can be started once peristalsis, bowel movement, or flatus is confirmed. Currently, however, early EN within 48 h of intensive care unit (ICU) admission may be initiated safely without the confirmation of these signs [3][4][5][6][7]. In fact, guidelines recommended that the decision to initiate EN should not be based on the confirmation of gastrointestinal peristalsis [8].
The presence of bowel sounds had been commonly used as a criterion for EN initiation. However, bowel sounds reflect gas movement in the intestines, and almost the same sounds can be heard when water and gas are injected via a tube inserted to the duodenum [9]. In other words, the bowel sounds cannot confirm normal functioning of the intestines, intestinal integrity, mucosal barriers, and preserved intestinal absorption. Therefore, it is important to start with a small amount of EN first, because bowel sounds cannot be used as a basis for EN initiation. Conversely, EN should be started as soon as possible, because EN itself may promote intestinal peristalsis.
There are other reasons to support early EN. The early initiation of EN including fiber may prevent atrophy of intestinal mucosa and attenuation of gastrointestinal peristalsis, because the energy substrates for intestinal mucosa are partially supplied via intraluminally. Moreover, it is believed that the early initiation of EN may prevent bacterial translocation (BT).

Monitoring of intestinal intolerance
The symptoms of patient intolerance to EN vary. It is important to comprehensively monitor for pain, abdominal distension, other clinical symptoms and findings, bowel movement or flatus, and abdominal X-ray, in order to guide the decision to continue EN or not and to avoid inappropriate discontinuation of EN [8].
Gastric residual volume (GRV) had been shown to not correlate well with the incidence of pneumonia [10][11][12], gastric emptying ability [13][14][15], and incidence of regurgitation and aspiration [16]. Decreasing the cutoff value of GRV cannot avoid such complications and may lead to inappropriate interruption, discontinuation, or reduction of the amount of EN administered [10]. Even if the GRV is less than 500 mL, EN should not be interrupted without any symptoms that indicate intestinal intolerance [17]. If GRV is restricted within 200-500 mL, EN should be carefully continued and measures to decrease the risk of aspiration should be taken, as described later.
Reignier et al. [18] reported that GRV monitoring (under 250 mL) did not affect the mortality and incidence of ventilator-associated pneumonia and infection. Poulard et al. [19] reported the same results for the incidence of complications, but the incidence of intestinal intolerance was significantly lower in the group without GRV monitoring. Regarding the criteria for GRV measure, four randomized controlled trials (RCTs) showed that the incidences of reflux, aspiration, and pneumonia did not increase when the cutoff for GRV was increased from 50-150 mL to 250-500 mL [10][11][12]17]. In addition, although the GRV measurement interval had often been set to every 4-6 h, some reports showed that there was no fixed standard and that the amount of GRV should be judged on a daily basis. Furthermore, the GRV measure may increase the risk of feeding tube occlusion, inappropriate discontinuation of EN, and the incidence of complications due to the decreased amount of EN administration [10,20].
To address the aforementioned issues, SCCM/ASPEN guidelines 2016 suggested not to include the GRV monitoring as part of daily care [21]. In clinical practice, however, several institutions still use GRV as one of the criteria to confirm intestinal intolerance and to determine EN continuation or interruption. Metheny et al. [22] reported that more than 97% of nurses assessed intolerance solely by measuring GRVs. Notably, it is important to clarify the standards for each institution and to avoid inappropriate interruption and discontinuation of EN when GRV is within 500 mL. At our institution, we set a GRV of < 300 mL/day as a guide for EN intolerance; if GRV is over 300 mL/day, we administer rikkunshito to improve gastric emptying.
A longer duration of intestinal rest can prolong the attenuation of gastrointestinal peristalsis. Since inappropriate fasting or EN discontinuation may induce the prolongation and deterioration of paralytic ileus, it is important to minimize the fasting duration for diagnostic and treatment procedures. Patient intolerance had been reported to be one third of the reason for EN interruption [23], but true intolerance represents only half of this [21,23]. Therefore, appropriate diagnosis of intestinal tolerance by the medical staff may reduce unnecessary discontinuation and withdrawal of EN.
As mentioned above, there is no useful and recommended method to monitor intestinal intolerance. During enteral nutrition management, it is important to always observe all gastrointestinal symptoms well and to confirm that these symptoms do not aggravate.

Initiation of EN and EN protocol
EN is safe and suitable to administer for patients with mild or moderate paralytic ileus, as long as the patient's hemodynamic status is stable [24]. The rate of achievement of the target amount of EN within 72 h was reported to be 30 to 85%, when EN was started after stabilization of hemodynamics, even before confirming the presence of peristaltic sounds. On the other hand, Kozar et al. reported that this rate reached 70 to 85% of the target amount of EN, when the EN protocol according to the circumstance of each facility was used [25]. The criteria to judge hemodynamic stabilization varies among facilities. At our institution, our criteria include small required administration dose of inotropic agents (e.g., ≤ 0.1 μg/kg/min of norepinephrine) or when the inotropic agents can be reduced.

Increase of EN amount and intestinal tolerance
The use of an EN protocol should be recommended to achieve the target amount of EN [8]. The factors to be set in the protocol are diverse (Table 1), but it was reported that the rate of achievement of the EN target amount increased when the ICU staff used a protocol that accounted for these factors [10,[25][26][27][28][29]. After early EN initiation, the next step is to determine whether EN can be increased systematically to reach the target amount. In general, in critically ill patients, the amount of EN administered within the first week is set within about 80% of the target amount [30]. When actively increasing the amount of EN, creation of a protocol that is suitable for the actual situation at each facility is desirable and should take into account the other protocols reported in the past [31,32]. However, the aim should be the adjustment of the EN amount along with the protocol, not to observe the protocol rigidly and inappropriately increasing the EN amount [33,34]. Development of gastrointestinal disorders due to an inappropriate increase in EN amount or flow rate to achieve the target amount may lead to necessary interruption of EN administration and would take time to achieve the target amount by careful re-initiation of EN with a small amount. Therefore, slow start and continuous administration of EN, with delayed increase and decrease of EN amount, may avoid the inappropriate interruption and discontinuation of EN, thereby enabling early achievement of the target amount.

Measures to mitigate reflux and aspiration
During EN administration, the risk of reflux and aspiration should be evaluated and prevented, especially in high-risk patients. Aspiration is one of the most notable complications of EN. The high-risk factors for aspiration are shown in Table 2 [17]. The measures to decrease the risk of reflux and aspiration are described below.
Elevation the head of the bed (30°to 45°) In critically ill patients, elevating the head of the bed is a measure without economic burden to decrease the risk of aspiration, not only during EN management, but also during artificial ventilation, and so on [35][36][37]. Compared with supine to semi-recumbent positions, elevation of the head of the bed at 30°to 45°was shown to significantly reduce the incidence of pneumonia [35]. Notably, in reality, the angle often remains less than 30°, even if the head of the bed is elevated; therefore, it is important to check the angle on a regular basis. Furthermore, the position management was reported to be thorough when the physician clearly instructed the angle [38]. However, it is important to be aware that prolonged elevation of the head of the bed during continuous EN administration may increase the risk of developing sacral pressure ulcers.

Switch to continuous administration
Intermittent infusion of EN was shown to increase the risk of aspiration pneumonia [34]. MacLeod et al. reported that although the incidence of infection and the amount of EN were not different between continuous and intermittent infusion, the ICU mortality significantly decreased with continuous infusion (7.4% vs. 13.9%) [39]. Other RCTs showed that compared with intermittent infusion, continuous infusion had similar outcomes, including the mortality, incidence of infection, and duration of hospital stay [40][41][42][43][44] but significantly earlier achievement of the target amount of EN [42]. Continuous infusion seems to alleviate the intolerance to EN. Therefore, switching to continuous infusion may become one of the measures to decrease the risk of reflux and aspiration in patients at high risk or those with intolerance to gastric EN. The incidence of vomiting may increase with continuous gastric infusion of EN, because GRV cannot be measured. Therefore, short interruption of EN infusion on a regular basis and measurement of GRV as needed are desirable to establish.
Agents, such as prokinetic drugs or narcotic antagonists, to promote motility Administration of prokinetic drugs, such as metoclopramide or erythromycin, had been shown to improve gastric emptying and intestinal intolerance [45]. Five RCTs revealed the effects of metoclopramide or erythromycin,  in comparison with those of placebo [46][47][48][49][50]. Of these, a meta-analysis of three RCTs [48][49][50] reported that the administration of prokinetic drugs decreased GRV [21]. Another study showed that both metoclopramide and erythromycin decreased GRV in a similar efficacy [51]. Moreover, compared with metoclopramide alone, combination therapy with metoclopramide and erythromycin significantly decreased GRV [52]. However, none of the studies showed differences in mortality and incidence of pneumonia between the two drugs. Prokinetic agents may be effective in patients with high risk of aspiration or those with intolerance to gastric EN. On the other hand, both metoclopramide and erythromycin had been associated with QT prolongation, which predisposes to cardiac arrhythmias [53,54]. Furthermore, it should be kept in mind that metoclopramide had been associated with adverse complications of extrapyramidal symptoms and tardive dyskinesia and that erythromycin may promote unnecessary use of antibiotics; moreover, administration of erythromycin to improve gastrointestinal motility is not covered by insurance in Japan. Therefore, these drugs should be discontinued quickly if they do not prove to be effective.
Narcotic analgesics may suppress gastrointestinal peristalsis. Administration of naloxone via a gastric tube to antagonize this adverse effect may improve the incidence of ventilator-associated pneumonia, decrease GRV, and increase the amount of EN infusion [55]. Therefore, narcotic antagonists may reduce the risk of reflux or aspiration.
These drugs have not been shown to improve outcomes, including mortality, duration of mechanical ventilation, and duration of ICU stay. In addition, various agents are used to improve gastrointestinal peristalsis in Japan. In particular, rikkunshito and mosapride citrate to promote gastric emptying, daikenchuto and prostaglandin F2α to promote intestinal peristalsis, and sodium picosulfate and magnesium oxide to promote bowel movement have been used on the basis of their pharmacological effects and experience [8]; however, evidence on the effectiveness of these drugs is not enough.

Switch to jejunal access (postpyloric route)
Switching the EN infusion route from gastric access to small intestinal (postpyloric) access had been shown to reduce the incidence of reflux [56,57], aspiration, and pneumonia [10,58,59]. A meta-analysis that included 12 RCTs [3,10,[58][59][60][61][62][63][64][65][66][67] reported that compared with gastric EN, small intestinal EN significantly reduced the incidence of pneumonia [21]. Another meta-analysis that included 7 RCTs [58-60, [65][66][67][68] showed small intestinal EN increase of the EN amount [21]. Therefore, switching the EN infusion route from gastric access to small intestinal access may become one of the measures to decrease the risk of reflux and aspiration in high-risk patients or those with intolerance to gastric EN. However, improvement of the EN amount and gastric emptying by jejunal (duodenal) EN, compared with gastric EN, were limited to patients with high severity [68] and had little effect [69][70][71]. Moreover, early jejunal EN had been reported to not reduce the incidence of pneumonia and to increase mild gastric bleeding [67]. On the other hand, gastric EN had been shown to reduce the duration of ICU stay [3,63] and the incidence of infectious complications [58,61], compared with jejunal EN.
Several methods, including X-ray fluoroscopy, auscultation, endoscopy, and ultrasound, had been reported as the methods that can be used to guide postpyloric feeding tube insertion. Although the optimal position (jejunum vs. duodenum) of the tip of the tube is controversial, insertion to the jejunum beyond the Treitz ligament had been reported to decrease the intragastric countercurrent [72]. At our institution, nasojejunal feeding tube insertion is performed at bedside using a transoral endoscope; however, regardless of the method used for postpyloric tube insertion, a gastric feeding tube is easier to insert and enables early initiation of EN. Because postpyloric tube insertion can delay EN initiation [64], early EN initiation by gastric tube should be prioritized.
Therefore, initiation of EN with jejunal feeding needs not be routine and should be switched from an initial gastric access depending on the severity of the patients' condition. Jejunal feeding should be considered in cases with jejunostomy constructed by open laparotomy, those with reflux via the gastric tube, and those with vomiting due to delayed gastric emptying despite several measures to mitigate the intolerance to gastric EN. The ESPEN guidelines strongly recommended selecting gastric access for EN initiation and to shift to postpyloric access when patients develop intolerance despite administration of prokinetic drugs or in those with high risk of aspiration [73].
For severe ARDS patients, long-duration prone position is recommended to prevent ventilator-induced lung injury [74,75]. PROSEVA study showed that prolonged (16 h) prone-positioning sessions significantly decreased mortality [76]. In such cases, EN administration in the long-duration prone position is needed. We confirmed that there was no difference in the amount of gastric reflux, regardless of gastric or jejunal access for continuous EN administration (unpublished data).

Criteria, classification, and evaluation of diarrhea
Diarrhea and constipation are common gastrointestinal symptoms. In critically ill patients, the control of bowel movement is extremely important for intensive care and management. Bacterial translocation due to the attenuation of gastrointestinal peristalsis and the stagnation of intestinal contents may cause sepsis or organ disorder [77,78]. Furthermore, diarrhea itself may induce several problems and complications, as described later. During early EN administration, observation of the status and control of bowel movement is important to achieve maximum effectiveness.
Although there are no definite criteria for diarrhea, indices, such as frequency of bowel movement > 3 to 5 times/ day or a volume of bowel movement > 200-300 g/day, had been used commonly [79,80]. In critically ill patients, various factors, including diseases; general condition; administration of drugs, such as antibiotics; presence or absence of EN administration; EN administration method; and the type of EN formula, may affect the property and volume of bowel movement. To definite the property and volume of bowel movement, evaluation tools, such as the Bristol stool form scale [81] or the King's stool chart [82], have been used. Because critically ill patients have prolonged immobility and cannot assume a bowel movement posture similar to a healthy person, bowel movement is considered to be difficult unless the stool properties are softer than usual. Therefore, in critically ill patients, soft stools should be permitted while avoiding watery stools and bowel movement should be managed as Bristol scale 4 or higher.
Based on the pathologic features, diarrhea can be classified as osmotic, exudative, secretory, or motor [79,80] (Fig. 1). In addition, diarrhea can also be classified as infectious or non-infectious, because treatment varies depending on the presence or absence of infection.

Complications and problems associated with diarrhea
Prolonged diarrhea can cause deficiency in nutrient absorption, malnutrition, and increase in mortality [80]. As the incidence of complications increases, support with PN may be necessary. In addition, diarrhea may cause a decrease in the circulating blood volume; metabolic acidosis with loss of electrolytes and bicarbonate by excretion of large quantities of digestive juices; electrolyte abnormalities with loss of potassium, magnesium, and zinc; and contamination of surgical wounds and pressure ulcers [80,83].
Strack van Schijndel et al. reported that a > 250-g/day volume of bowel movement may be used as an index of malnutrition [84]. Furthermore, Wierdsma et al. [85] reported that loss of nutrients in the feces increased as the volume of bowel movement increased and that daily measurement of bowel movement volume was important, because the risk of energy and protein deficiency may increase in patients with a bowel movement volume if > 350 g/day. Restriction of bowel movement volume to some extent is important for EN management in critically ill patients, because prolonged diarrhea is directly associated with energy deficit or negative energy balance [86], and the resulting malnutrition may impair immune function, increase the risk for infectious complications, and increase mortality. Therefore, a protocol for fecal management (Fig. 2) should be constructed according to the actual circumstances of each institution, similar to the protocols of EN initiation and decreasing the risk for aspiration. Administration of prokinetic drugs to improve the symptoms of constipation may be included in the protocol [87].

Causes of diarrhea
Since EN can maintain the structure and function of the gastrointestinal mucosa better than PN, the development of diarrhea may be suppressed. However, diarrhea often occurs after EN initiation, depending on the methods of administration, amount, flow rate, and type of EN formula. If diarrhea occurs in critically ill patients on EN management, it is important to determine the cause (Table 3). In particular, in critically ill patients, diarrhea due to the administration of antibiotics should be taken avoided. Various risk factors for Clostridioides (Clostridium) difficile-associated diarrhea, which is the most frequent cause, have reported (Table 4) [88][89][90][91][92][93]. Furthermore, the incidence of diarrhea varies according to the type of antibiotics; the risk is high for quinolones and cephalosporins and low for macrolides [88,94].

Diarrhea and EN
The incidence of EN-related diarrhea was reported to be the same, regardless of the administration route (gastric vs. postpyloric) [66,95]. However, in theory, diarrhea had been thought to be more likely to occur with the direct injection of hyperosmotic EN formula into the jejunum than into the stomach. In many cases, diarrhea can be actually improved by converting jejunal EN to gastric EN. With regard to the method of EN administration, continuous infusion with the use of EN pump was reported to suppress the development of diarrhea,

Severity of underlying diseases
Enteral nutrition (especially in postpyloric feeding) compared with intermittent infusion [41,73,96], but this effect was difficult to assess after the diarrhea has occurred [41,96]. The factors related to the development of diarrhea are the content and composition of EN nutrients, such as carbohydrates, presence or absence and type of lipid, type of nitrogen source, presence or absence of lactose and milk protein, presence or absence of dietary fiber, and osmotic pressure. However, evidence on the effects of the contents and composition of EN nutrients had been insufficient. Most of the available studies were based on the investigation of the EN formula, not if the single ingredient.

Prevention and treatment of diarrhea
The general symptomatic treatment of diarrhea comprises the administration of opioids and anticholinergic drugs and fluid replacement. If the cause of diarrhea does not seem to be EN and the presumed cause is being managed appropriately, EN should not be stopped unnecessarily and should be continued even in small amounts. The measures to prevent EN-related diarrhea are shown in Table 5.
Elemental diet, in which the nitrogen source is formulated as amino acids, is hyperosmotic and can easily cause diarrhea. On the other hand, oligomeric diet, in which the nitrogen source is formulated as peptide, may not easily cause diarrhea, but the evidence on its efficacy had been insufficient. In Japan, the available nutrition products in medicine are Twinline ® NF (EN Otsuka Pharmaceutical Co., Ltd., Iwate, Japan) and Aminoleban ® EN (Otsuka Pharmaceutical Co., Ltd., Tokyo, Japan), and the available high-density liquid diet in food are Peptino ® (Terumo, Tokyo, Japan), Peptamen ® (Nestlé Health Science, Tokyo, Japan), and Hine E-Gel ® (Otsuka Pharmaceutical Factory, Inc., Naruto, Japan). In some cases, changing to an EN formula without lactose, milk protein, and lipid may improve the diarrhea. Especially in patients who have undergone surgery, such as pancreatoduodenectomy, changes in digestive enzyme secretion may alter the function of digestion and absorption and cause diarrhea.
Water-soluble dietary fibers, compared with insoluble dietary fibers, are more effective in preventing diarrhea. In particular, pectin and guar gum increase the viscosity, delay the excretion from the stomach and absorption in the small intestine, and decrease the flow of gastrointestinal contents by resisting against gastrointestinal peristalsis. Dietary fiber-enriched EN formula is usually administered to attenuate diarrhea and to improve constipation; among these, pectin had been reported to highly and effectively prevent diarrhea [97]. Since increasing the viscosity of gastrointestinal contents improves diarrhea, a semi-solid type of EN formula may be effective in patients with gastrostomy. In critically ill patients, however, the EN formula is difficult to make into a semi-solid form, because it is infused through a thin-diameter feeding tube. Recently, Hine E-Gel ® had become commercially available; it is a high-density liquid diet that includes pectin and changes into a gel form by its reaction with gastric acid and can undergo semisolidification in the stomach, even when administered through a thindiameter tube. Moreover, viscosity-adjusted liquid food, such as Meiflow ® (Meiji Co., Ltd., Tokyo, Japan), which can be infused through a thin-diameter tube, can also prevent diarrhea. In contrast, one study reported that administration of dietary fiber-enriched EN formula had no efficacy in critically ill patients [98]. In addition, there is not enough evidence on the effectiveness of pre-/pro-/synbiotics in maintaining the bacterial flora. Further research is required.
In recent years, fecal microbiota transplantation had been carried out for several diseases, such as pseudomembranous enteritis and inflammatory bowel diseases, and has drawn attention because of its effects of decreasing the frequency of defection and improvement of stool properties. However, evidence on critically ill patients has

Conclusion
Gastrointestinal complications associated with EN, particularly vomiting and diarrhea, were described. One of the best ways to success for EN management is to continue as long as possible without interruption and discontinuation of EN easily by appropriate measures, even if gastrointestinal complications occur.