Gastric Feeding as an Extension of an Established Enteral Nutrition Protocol  ,
* Department of Surgery, Louisiana State
University Health Sciences Centre, New Orleans, Louisiana; Correspondence: Frederick A. Moore, MD, Department of Surgery, University of Texas, Houston School of Medicine, 6431 Fannin, Suite 4.264, Houston, TX 77030. Electronic mail may be sent to Frederick.A.Moore{at}uth.tmc.edu. Background: Indiscriminate gastric feeding in ICU patients imposes unacceptable risks of aspiration. Believing that a subset of ICU patients can be fed safely via the stomach, we have developed a protocol to identify appropriate patients and guide the bedside clinician in how to safely and effectively feed via the stomach. Methods: A literature search was done to identify appropriate medical literature. High grade evidence along with local expert opinions were used to develop a protocol. This protocol has been refined and implemented. Results: Based on perceived risk of aspiration, patients are assigned enteral access (ie, stomach vs. distal postpyloric). Enteral formula is selected based on patient characteristics. It is then advanced by a standard protocol with specific precautions while monitoring for symptoms of intolerance. Management of intolerance is dictated by the type and severity of intolerance. Conclusion: We have implemented a gastric feeding into a subset of our ICU patients. Gastric feeding requires certain precautions but appears to be safe. With more experience and better understanding of the pathogenesis gastroparesis, we believe that most ICU patients should be able to safely feed into the stomach. This is logistically easier than postpyloric feeding and offers physiologic advantages. Early enteral nutrition (EN) has been shown to improve outcome of critically ill surgical patients, and thus, after initial stabilization, EN becomes an important process of care.1 We therefore developed an EN protocol for the Shock Trauma Intensive Care Unit (STICU) at the Memorial Hermann Hospital in Houston, Texas.2 According to the original studies that documented improved outcomes, this protocol emphasizes feeding distal to the ligament of Treitz. High-risk patients are identified, and if a surgical jejunostomy is not present, a nasojejunal tube (NJ) is placed. The appropriate formula (based on patient characteristics) is started at 15 mL/h and advanced by 15 mL/h every 12 hours. This advancement regimen is associated with variable symptoms of intolerance, but with careful monitoring and intervention, full-dose EN is achieved in over 85% of high-risk patients.3 Despite our success, we acknowledge that most ICU clinicians feed into the stomach.4–7 We believe that liberal use of gastric feeding in ICU patients, however, imposes an unacceptable risk of aspiration, which is a major contributing factor to ventilator-associated pneumonia (VAP).8,9 A number of prospective randomized trials (PRTs) have recently attempted to address this issue.10–18 These PRTs are underpowered, but when combined they consistently show that postpyloric feeding compared with gastric feeding achieves shorter times to targeted goals but fail to document a consistent reduction in VAP. One confounding variable is that the postpyloric tubes used in some of these PRTs were too short (eg, 109 cm) to reach the ligament of Treitz, and consequently the patients are being fed into the duodenum (often the proximal portions). Recent studies have shown that critically ill patients have a high incidence of duodenogastric reflux, so it is not surprising that there has been difficulty in documenting a reduction in VAP.15,16,19,20 Regardless of this debate, we believe that subset of our ICU patients (yet to be accurately defined) can be adequately and safely fed via the stomach. This would be advantageous for several reasons, including (a) provision of stress gastritis prophylaxis; (b) stimulation of neurohumoral responses that should enhance gut function; and (c) elimination of the rare but devastating complication of nonocclusive bowel necrosis that is associated with jejunal feeding.21 Our purpose, therefore, was to extend our current EN protocol to include a gastric feeding arm. Using the best available data combined with local expert opinion, we have developed a procedure to identify eligible candidates and have included additional precautions, monitoring, and interventions to the existing protocol for those patients who are being fed into the stomach or proximal duodenum.
An English medical literature search was done to ascertain (1) criteria for determining patients' risk for aspiration; (2) the relationship between gastric feeding and VAP; and (3) parameters of intolerance to EN. This search was conducted through MEDLINE and EMBASE using the terms gastric, enteral, feeding, EN, nutrition, nutrition support, aspiration, and complication. The literature was reviewed and articles were selected based on pertinence to the development of the protocol and strength of the evidence. Information from high-grade evidence was distilled into a working protocol. The protocol was then reviewed by a multidisciplinary committee of local experts, and changes were made according to specific practice patterns. The finalized protocol was then implemented.
Hemodynamically stable patients with functional gastrointestinal tracts and no history of delayed gastric emptying or aspiration are considered candidates for gastric feeding and are assessed for risk of aspiration. According to aspiration risk (Table 1), patients are assigned type of enteral access (Fig. 1).8,22 Patients with 0 or 1 risk factor have their nasogastric (NG) tube output assessed. If the output is <500 mL over 12 hours, the decompression sump tube is removed and a nonsumped 12 to 14 French feeding tube placed. After its position is confirmed by plain x-ray, feeding is initiated. All patients are positioned with the head of the bed (HOB) elevated to at least 30°, and receive oral care with chlorhexidine mouth washing.23–25 For patients with 1 risk factor, additional preventive measures include tight glucose control, correction of electrolyte abnormalities, and minimization of narcotic dosing.26–29 Patients who have (a)  2 major risk factors; (b) history of gastroparesis;
or (c) history of aspiration should be fed past the ligament of Treitz. If a
surgical jejunostomy is not present, the bedside nurse blindly passes a
10-French 140-cm NJ tube (Corpak, VIASYS Healthcare Medsystems Division,
Wheeling, IL), and an abdominal plain film x-ray is obtained to confirm
position. We require that the tube has passed over the spine twice
(Fig. 2). If this has not
occurred, the end of the tube is in the proximal duodenum. In our ventilated
patient, we would remove this tube and have our procedure team endoscopically
place an NJ tube beyond the ligament of Treitz. In a nonintubated patient,
endoscopy is not as easy and, therefore, if the patient did not have history
of gastroparesis or aspiration, we would feed into the proximal duodenum and
monitor the patient as a high-risk gastric feeding patient.
Formula selection for NG and NJ feeding is based on the same criteria (Table 2). Full-strength formula is started at 15 mL/h and advanced by 15 mL/h every 12 hours until 60 mL/h is reached. We hold this rate at 60 mL/h for 24 hours to assure tolerance and then advance to a patient-specific goal that is determined by the STICU dietitian. Gastric feedings are stopped any time the HOB needs to be <30°, except during patient care maneuvers. They are also held 4 hours before undergoing an anesthetic for an operative procedure (may be restarted immediately postoperatively at the previous rate) and 4 hours before endotracheal extubation. We do not stop gastric feedings for diagnostic tests, nursing care, or routine bedside procedures unless ordered by a physician.
Gastric residual volumes (GRV) are checked every 4 hours for the first 5 days and then every 12 hours thereafter if the patient has been tolerating gastric feedings. GRV monitoring returns to every 4 hours if the clinical course deteriorates. The fluid withdrawn when measuring GRV is returned to the patient if GRV <500 mL. During gastric feeding, the patient is also monitored every 12-hour shift for other standard signs/symptoms of intolerance and managed according to the guidelines outlined in Table 3.
Most clinicians now accept that EN is preferred to parenteral nutrition, but considerable debate exists over the best route to administer EN. The original studies that documented the superiority of EN used surgical jejunostomies.1 This route is not available in most ICU patients. In our surgery/trauma ICU, obtaining NJ tube placement is relatively easy, but we acknowledge this is not the case in most ICUs. By default, ease of obtaining access has made gastric feeding a common practice in ICUs. Unfortunately, recent studies have documented that patients fed into the stomach on average receive only about 50% of their projected needs.4–7 Multiple factors have been implicated, but high GRV is by far the most frequent cause for stopping EN. The other major concern is that gastric feeding imposes a major risk of aspiration, which is an important contributing factor in VAP.8,9 To avoid these problems, some clinicians advocate the use of postpyloric nasoenteric tubes. Most of these tubes, however, are too short to go beyond the ligament of Treitz, and as a result, the diets are being administered into the duodenum. PRTs comparing gastric to postpyloric feeding offer conflicting results.10–18 Although these studies consistently demonstrate that postpyloric enteral feeding achieves shorter time to the target dose of nutrition, they have failed to demonstrate a consistent reduction in VAP. One confounding variable that might explain the difficulty in demonstrating a difference in VAP is the fact that critically ill patients have a high incidence of duodenogastric reflux. In a study of the antral, duodenal, and proximal jejunal motility, Tournadre et al19 demonstrated that postoperative gastroparesis occurs after major abdominal surgery (not surprising). But of note, it is associated with more rapid discoordinated duodenal contractions, of which 20% migrated in a retrograde fashion. Heyland et al15 administered radio isotope-labeled enteral formulas through a standard postpyloric feeding tube in ventilated ICU patients and documented an 80% rate of radio label reflux into the stomach, 25% reflux rate into the esophagus, and 4% reflux rate into the lung. Finally, Wilmer et al20 reported monitoring bile reflux in the esophagus of ventilated ICU patients using a fiberoptic spectrophotometer (Bilitec system, Medtronic Inc, Minneapolis, MN) that detects and quantifies bilirubin concentration. All study patients received acid-suppressive stress gastritis prophylaxis and also had 24-hour esophageal pH monitoring. Endoscopy was performed and documented erosive esophagitis in half of the patients, of which 15% had pathologic acid reflux and 100% had pathologic bile reflux. These studies provide convincing evidence that duodenogastric reflux is a common event in ICU patients and have led us to conclude that feeding into the proximal duodenum is not much different than feeding into the stomach with regard to risk of aspiration. According to our review of the literature and our working group discussion, we conclude that gastric feeding increases the risk for aspiration, and our protocol is designed to reduce this risk. First, it is important to emphasize that the current understanding of the basic pathophysiology of aspiration in ICU patients is poor and that the commonly used ICU monitors are, at best, crude. Given our success with NJ feeding, we are hesitant to indiscriminately use gastric feeding in our STICU because this may tip the risk vs benefit balance. Our first strategy, therefore, was to stratify patients by aspiration risk and to then to implement our protocol in those patients in whom gastric feeding can be safely attempted (Table 3). Our next strategy is to optimize monitoring of gut dysfunctions that contribute to aspiration. This discussion primarily focused gastroesophageal reflux (GER) and gastroparesis. Commonly used clinical monitors of GER include laboratory testing for the presence of glucose in tracheal secretions or by observing blue food dye (BFD) that has been added to the enteral formula in tracheal aspirates.30 Detection of glucose lacks specificity. False-positive results can occur with high serum glucose levels or the presence of blood in tracheal secretions. The use of BFD is poorly standardized and lacks sensitivity. More importantly, several reports document absorption of BFD in critically ill patients and associated it with death. This is presumably because of a toxic effect that BFD has on mitochondrial function.31 A recent consensus conference concluded that this practice should be abandoned.8 We conclude there are no adequate ICU monitors of GER other than observing for vomiting or regurgitation, which lack sensitivity. The HOB should be elevated >30° to decrease the risk aspiration when GER does occur. GRVs should be monitored with the presumption that a distended stomach will lead to a higher-volume GER. In regard to gastroparesis, GRV is the current standard of care monitor. Like most ICU monitors of GI function, GRV is overly simplified, poorly defined, and generally inadequate. The exact threshold is not well defined. Most centers use a threshold of 100 mL to 150 mL and assume that when GRV is less than threshold, it is safe to advance the feeding rate. There is a paucity of evidence to support this practice, and it represents a major impediment to achieving EN goals.32–34 The underlying premise is based on the assumption that the GRV reflects the volume of gastric contents, which is dependent upon the adequacy of gastric emptying. Studies have shown that GRVs do not correlate with the volume of gastric content or gastric emptying. According to the consensus conference above mentioned, we have liberalized the threshold to >200 mL to start interventions and >500 mL to stop gastric feeding and pursue post–ligament of Treitz feeding. The last strategy to enhance gastric feeding was to prevent or improve ileus. When discussing how ileus relates tolerance to EN in the ICU, it is best to view ileus in the stomach, small intestine, and colon as different entities.3,35 Colonic ileus is rarely an issue. We do not equate early success of EN with return of normal bowel movements. Diarrhea (ie, excessive colon transit) is a frequent problem, but it tends to occur late and is amenable to intervention. Surprisingly, small bowel motility and transit are remarkably well preserved. Although cramping and abdominal distention are ominous signs of intolerance, only in a small percentage of patients do these symptoms get severe enough to stop EN. On the other hand, gastric atony is a very frequent reason cited for holding EN. We conceptually think of gastroparesis in the ICU as occurring in 2 forms: (1) acute gastroparesis that occurs as a result of a neurohumoral reflex to stress insults; and (2) prolonged gastroparesis that occurs secondary to an inflammatory response in the stomach and small bowel that occurs after ischemia/reperfusion, sepsis, and surgical manipulation.36–38 Although we and others are studying the chronic form in the laboratory, the acute form appears to respond to some currently available clinical interventions. Although multiple factors are operational in the acute form, endogenously released opioids secondary to stress and exogenously administered opioids to control pain stimulate receptors in the central nervous system (CNS) and enteric nervous system (ENS) to shut down gastric motor activity.27 Recent studies indicate if opioid actions at the ENS are blocked, motility may be improved without interfering with the desired opioid actions on the CNS. An investigational opioid receptor antagonist that has limited systemic absorption after oral administration has been shown to speed the recovery of bowel function after surgery.28 Additional studies are needed, but this implicates opioids in abnormal GI motility and emphasizes the need to minimize opioid administration as much as possible. Another strategy to enhance gastric motility is to stimulate motilin or serotonin receptors (eg, 5-HT4). Although a number of agents are under investigation, erythromycin is currently available and has been used in trials to enhance the success of gastric feeding.39,40 Erythromycin is a "motilide" that acts on the motilin receptors to enhance gastric emptying. It appears to be safe, but significant tachyphylaxis occurs. In summary, we have developed an EN protocol that emphasizes post–ligament of Treitz feeding and have demonstrated this to be quite successful in high-risk patients. We acknowledge that there is a role for gastric feeding in this cohort and have added a new arm to implement gastric feeding in a subset of patients in whom aspiration risk is perceived to be low. Gastric feeding requires certain precautions, and we have liberalized the thresholds for using GRVs (200 mL and 500 mL), recognizing that the commonly used restrictive threshold (eg, 100 mL) is a major impediment to gastric feeding. In the future, to enhance success of gastric feeding we need to better understand the basic pathophysiology of gastroparesis in ICU patients. With this knowledge, we should be able to minimize insults and medications that adversely affect the process, and, with ongoing research, new agents (eg, motilin and serotonin agonists) should soon become available to "kick start" motility in acute gastroparesis. We are beginning to understand the inflammatory response that causes more prolonged cases of gastroparesis and believe that in the near future interventions will be devised to prevent this.
Nutrition in Clinical Practice, Vol. 19, No. 5,
504-510 (2004) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Top
Review of the Literature
Development of the Protocol
