Transnasal Endoscopy for EnteralFeeding Tube Placement in Critically Ill PatientsJonathan A. Dranoff, M.D., Peter J. Angood, M.D., and Mark Topazian, M.D.Department of Internal Medicine, Division of Digestive Diseases, and Department of Surgery, Yale UniversitySchool of Medicine, New Haven, Connecticut

OBJECTIVE: Early enteral feedings may improve outcomesin critically ill patients. Recently, transnasal endoscopy withan ultrathin transnasal endoscope has been shown to be ofvalue for diagnostic endoscopy without conscious sedation.We developed a technique for the placement of postpyloricfeeding tubes in critically ill patients using transnasal en-doscopy. We describe our initial experience in a consecutiveseries of patients.

METHODS: We collected data on consecutive intensive careunit patients undergoing bedside transnasal endoscopy fornasoenteric feeding tube placement using a standardizedtechnique. Tube position was confirmed in all patients witha plain abdominal radiograph. Tube placement was deemedsuccessful if the feeding tube traversed the pylorus.

RESULTS: Transnasal endoscopy was completed in all four-teen patients, as was placement of a feeding tube. Feedingtubes were successfully placed in the jejunum or duodenumin 13 of the 14 patients (93%). Tubes remained in place from3 to 45 days (mean 16 days). Two patients required con-scious sedation during tube placement, and two ultimatelyrequired percutaneous gastrostomy.

CONCLUSIONS: Transnasal endoscopy allows simple andsuccessful postpyloric feeding tube placement at the bedsideof critically ill patients. This method can facilitate earlyenteral feeding in intensive care units. (Am J Gastroenterol1999;94:2902–2904. © 1999 by Am. Coll. of Gastroenter-ology)

INTRODUCTION

Early enteral feedings may improve outcomes in criticallyill patients (1, 2). Such feedings generally require a feedingtube. Placement of the tube beyond the pylorus may bedesirable, inasmuch as this may protect against aspirationpneumonia. However, advancement of a feeding tubethrough the pylorus can be difficult, especially when gastricmotility is impaired.

When the predicted duration of enteral feeding is#1month, a nasogastric or nasoenteric feeding tube is prefer-able to a percutaneous gastrostomy or enterotomy (3). En-doscopic placement of nasoenteric feeding tubes has previ-ously been described; however, this generally requires an

oral–nasal transfer as well conscious sedation (4). Recently,transnasal endoscopy with an ultrathin transnasal endoscopehas been shown to be of value for diagnostic endoscopywithout conscious sedation (5).

We developed a technique for the placement of post-pyloric feeding tubes in critically ill patients using transna-sal endoscopy. We describe our initial experience in con-secutive patients.

MATERIALS AND METHODS

PatientsWe prospectively collected data on consecutive patientsundergoing endoscopic transnasal feeding tube placement.Informed consent was obtained from patients or their familymembers.

Table 1 summarizes primary diagnoses for the 14 patientswho underwent attempted feeding tube placement. In somepatients, there were multiple reasons for hospitalization.Mean patient age was 49 yr (range 23–79 yr). Four patientswere female and 10 were male. All patients were located inan intensive care unit at the time of feeding tube placement.All patients but one had either an endotracheal tube ortracheostomy for respiratory support.

Table 2 summarizes indications for postpyloric feedingtube placement. Some patients had several indications for anenteral feeding tube. In three patients, the attending physi-cian’s fear of aspiration was the primary reason for postpy-loric feeding tube placement.

Several patients were referred in whom transnasal endo-scopic placement of a feeding tube was believed to becontraindicated, including patients with facial and basilarskull fractures and possible cerebrospinal fluid leak. Ofnote, neither thrombocytopenia nor coagulopathy were be-lieved to be contraindications. At the time of tube place-ment, the minimum platelet count of any patient was 38,000;the maximum prothrombin time was 22 s.

MethodsAll endoscopies were performed at the patient bedside inintensive care units. Patients were supine during the entireprocedure. A 5.5-mm transnasal fiberoptic endoscope(Olympus GIF-N30) was placed in one of the patient’snostrils. Under direct visualization the nasopharynx was

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 94, No. 10, 1999© 1999 by Am. Coll. of Gastroenterology ISSN 0002-9270/99/$20.00Published by Elsevier Science Inc. PII S0002-9270(99)00490-6

traversed and the scope was advanced to the posterior oro-pharynx. The esophagus was intubated under direct visual-ization and a diagnostic endoscopy was performed. Theendoscope was then advanced as far as possible into theduodenum. External compression and endoscopic suction ofthe stomach were used in an attempt to prevent the endo-scope from coiling in the stomach.

An 8-Fr enteric feeding tube designed for radiologicplacement (Miller Feeding Tube, Cooke, Winston-Salem,NC) with its weighted tip cut off, or a two-lumen 16-Frgastric/9-Fr jejunal feeding tube (Kangaroo Naso-JejunalFeeding/Gastric Decompression Tube, Sherwood Medical,St. Louis, MO) was immersed in warm tap water togetherwith a Teflon-coated, soft tipped guidewire. The guidewirewas passed through the biopsy port of the endoscope untilseen in the duodenum. The wire was then advanced furtherdown the lumen with the tip beyond endoscopic view untilmild resistance was encountered. The endoscope was com-pletely withdrawn from the patient, keeping the wire inplace. The lubricated feeding tube was then passed over theguidewire; slight tension was kept on the wire during ad-vancement. Once resistance was met the tube was not ad-vanced further and the wire was withdrawn. When thedouble-lumen tube was used, the jejunal tube was advancedover the wire first, and then the gastric tube was advancedover the jejunal tube.

In some patients a repeat endoscopy was performedthrough the other nostril to confirm that the tube had passedthrough the pylorus. In all cases an attempt was made toaspirate bile from the jejunal tube before the tube was used.Tube position was confirmed in all patients with a plainabdominal radiograph and in the first three patients, a con-trast injection was also performed through the feeding tubeat the time of the radiograph. If the tip was in the smallbowel, tube feeds were initiated immediately. In patients inwhom single lumen feeding tubes were placed, a nasogastrictube was often placed in the other nostril to aspirate gastriccontents.

All endoscopies and tube placements were performed byfirst or second year gastroenterology fellows not experi-

enced in the use of endoscopically placed guidewires, su-pervised by one attending endoscopist with substantial pre-vious guidewire experience as well as previous experiencewith diagnostic transnasal endoscopy.

RESULTS

Transnasal endoscopy was completed in all 14 patients, aswas placement of a feeding tube (Fig. 1). Two patientsreceived conscious sedation. Feeding tubes were success-fully placed in the jejunum (five patients), first duodenum(two patients), second duodenum (four patients), or distalduodenum (two patients) in 13 of the 14 patients (93%).Early in our experience a repeat procedure was necessary intwo patients when postprocedure radiographs showed thatthe tube did not traverse the pylorus; in both patients asecond attempt resulted in successful postpyloric tube place-ment. We subsequently began to confirm postpyloric posi-tioning of the tube with immediate endoscopy through theother nostril but encountered only one other patient in whomthe first attempt at postpyloric tube placement was unsuc-cessful; in that patient a second attempt also failed. Whenbile but not air could be aspirated from the feeding tube, thetube tip was always located in the duodenum or jejunum.

Two patients had a second tube placed 10 and 30 days

Table 1. Patient Diagnosis

Diagnosis No. of Patients

Trauma 5Sepsis 5Postsurgical complications 4Respiratory failure 12

Table 2. Reason for Endoscopic Feeding tube Placement

Reason for Feeding Tube No. of Patients

Unable to place enteric feeding tube blindly 6High gastric residuals 7Pneumonia 3Coagulopathy and/or thrombocytopenia 2Fear of aspiration 3

Figure 1. Radiograph of patient with recently placed two-lumen 16Fr gastric/9 Fr jejunal feeding tube. Small white arrow indicatesjejunal port; large black arrow indicates gastric port.

2903AJG – October, 1999 Transnasal Endoscopy for Feeding Tube Placement

later when their postpyloric tubes were accidentally re-moved. Although the initial six tubes placed were singlelumen feeding tubes designed for radiological placement,we subsequently used two-lumen, 16-Fr gastric/9-Fr jejunalfeeding tubes, which allowed the gastric port to be used formedicines and gastric venting if needed. Tubes remained inplace from 3 to 45 days, with a mean of 16 days. Six tubeslasting,7 days were either inadvertently pulled out or werepresent when the patient died. No tubes became obstructed.Four patients died in the hospital; the remainder were dis-charged, although four of the patients required inpatientrehabilitation. In only two patients were percutaneous gas-trostomy tubes subsequently placed.

DISCUSSION

In this article we describe an efficient technique for endo-scopic nasoenteric feeding tube placement. Tube placementwas simple to accomplish at the bedside of critically illpatients, was free of complications, and provided a durablemeans of enteral feeding. We were able to confirm postpy-loric placement of the feeding tubes at the bedside byaspirating bile through the tube and by immediate endos-copy through the other nostril. Patients were typical of thosewith extended admissions to a surgical or medical intensivecare unit. These patients would seem to benefit the mostfrom a short, atraumatic endoscopic procedure free of se-dation. The usable lifespan of the tubes (up to 45 days) wassurprisingly good. In fact, caregiver error was the mostcommon reason for early tube removal.

Enteral nutrition improves outcomes in critically ill pa-tients (1, 2). The theoretical advantage of postpyloric feed-ing tube placement is avoidance of aspiration pneumoniacaused by regurgitated gastric contents. Although there areno conclusive data supporting nasoenteric feeding over na-sogastric feeding, there is evidence that gastric emptying isimpaired in critical illness, suggesting a predisposition toreflux of gastric feedings (6).

Other methods of nasoenteric feeding tube placementhave met with varied success. Blind placement of a naso-enteric tube is not consistently successful (7). Moreover,complications of misguided feeding tube placement can bedevastating (8). Fluoroscopic guidance of feeding tubeplacement can be time-consuming and requires transporta-tion of critically ill, mechanically ventilated patients to afluoroscopy suite (9). Endoscopic methods for placing suchtubes have been described, but require a nasal transfer afterplacement of an oroenteric feeding tube or guidewire (10).

Transnasal endoscopes have been used for diagnosticunsedated endoscopy for several years (5). We found that atransnasal approach to feeding tube placement permitted

rapid and easy enteral feeding tube placement at the patientbedside. Critically ill patients with recent aortic or intestinaltrauma, or marked coagulopathy underwent this procedurewithout complications. The procedure we describe generallyrequired 15–20 min to perform.

There were some limitations to this method of feedingtube placement. Endoscope advancement into or beyond thesecond portion of the duodenum was necessary to succeed,and this was often the most difficult part of the procedure.Deflation of the stomach and external compression oftenhelped. In addition, some patients had thick gastric secre-tions that would not suction through the 2.1-mm workingchannel of the endoscope. Despite these limitations, enterictube placement was successful in 93% of patients.

In conclusion, transnasal endoscopy permits early, rapidpostpyloric feeding tube placement at the bedside of criti-cally ill patients. This method can facilitate early enteralfeeding in intensive care units.

Reprint requests and correspondence:Mark Topazian, M.D.,Yale University School of Medicine, Division of Digestive Dis-ease, 333 Cedar Street, 1080 LMP, New Haven, CT 06520.

Received Dec. 8, 1998; accepted May 24, 1999.

REFERENCES

1. Klein S, Kinney J, Jeejeebhoy K, et al. Nutrition support inclinical practice: Review of published data and recommenda-tions for future research directions. JPEN 1997;21:133–56.

2. Sax HC. Early nutritional support in critical illness is impor-tant. Crit Care Clinics 1996;12:661–6.

3. Lipman TO, Cass OW, Ho CS, et al. Group I: Choosing theappropriate method of placement of an enteral feeding tube inthe high-risk population. Nutr Clin Pract 1997;12:S54–5.

4. Patrick PG, Marulendra S, Kirby DF, et al. Endoscopic naso-gastric-jejunal feeding tube placement in critically ill patients.Gastrointest Endosc 1997;45:72–6.

5. Dean R, Dua K, Massey B, et al. A comparative study ofunsedated transnasal esophagogastroduodenoscopy and con-ventional EGD. Gastrointest Endosc 1996;44:422–4.

6. Heyland DK, Tougas G, King D, et al. Impaired gastric emp-tying in mechanically ventilated, critically ill patients. Inten-sive Care Med 1996;22:1339–44.

7. Zaloga GP. Bedside method for placing small bowel feedingtubes in critically ill patients. A prospective study. Chest1991;100:1643–6.

8. Gharib AM, Stern EJ, Sherbin VL, et al. Nasogastric andfeeding tubes. The importance of proper placement. PostgradMed 1996;99:165–8.

9. Gutierrez ED, Balfe DM. Fluoroscopically guided nasoentericfeeding tube placement: Results of a 1-year study. Radiology1991;178:759–62.

10. Patrick PG, Marulendra S, Kirby DF, et al. Endoscopic naso-gastric-jejunal feeding tube placement in critically ill patients.Gastrointest Endosc 1997;45:72–6.

2904 Dranoff et al. AJG – Vol. 94, No. 10, 1999