Elsa D. Gutierrez, MD •¿ Dennis M. Balfe, MD
Fluoroscopically Tube Placement: Nasoenteric tube feeding is a widely used alternative to parenteral intra venous nutritional support or gas trostomy tube placement. Unmoni tored tube passage may result in complications and delays the begin fling of tube feedings. The authors studied the results of 882 fluoro scopically guided feeding tube placements in 448 patients in 1 year to determine rates of success and complications, as well as the long term outcome of this population of patients. Seven hundred sixty-four attempts (86.6%) were successful in positioning the tube distal to the third portion of the duodenum. Four major complications (three fa tal arrhythmias and one tracheo bronchial injury) were encountered. Only seven patients (2%) experi enced aspiration events that were due to positioning of the tube in the distal duodenum. Seventy-seven percent of patients required either one or two tubes; the average “¿tube life―was 7.8 days. Most reposition ings were required because of pa @tient noncompliance or inappropri ate administration of solid medica tions. Fluoroscopically guided nasoenteric tube passage is safe, eas ily performed, and highly success ful, and has resulted in widespread clinical acceptance in our institu tion. Index terms: Feeding tube, complications. Fluoroscopy
@Radiology
•¿ Nutrition
1991; 178:759-762
@ From the Mallincknodt Institute of Radiolo @ gy. Washington University School of Medicine, 510 S Kingshighway Blvd. St Louis, MO 63110.
Received July 20, 1990; revision requested tember
6; revision
received
October 10. Address reprint @
RSNA,
1991
October
Sep
9; accepted
requests to D.M.B.
Guided Results
Nasoenteric Feeding of a 1 -year Study@
S UCCESSFULtreatment of chronical ly ill patients requires the provi sion of adequate protein and caloric intake (1). In 1976, Dobbie and Hoff meister (2) found that nasoentemic in tubation was a viable alternative to intravenous panenteral delivery if the patient's gastrointestinal function was intact. Subsequently, a metal tipped polyurethane 8-F catheter has become commercially available and is used extensively for long-term nu tritional support. Early difficulties in passing this extremely flexible tube were solved with the introduction of a removable, steel stiffening stylus, which increases the rigidity of the system roughly to that of a standard
ments are performed by radiologists, either within the radiology depart ment or at bedside with use of a Carm portable fluomoscopy unit. We studied the results of fluomoscopically guided nasoenteric intubation pen formed in 448 patients in 1 year. The goals of this study were to determine our success and complication rates, to characterize the clinical indications for intubation, and to analyze the long-term performance of feeding tubes in this population, with specif ic reference to ultimate patient out come.
nasogastric
The 8-F feeding tubes (Entriflex; Bio search Medical Products, Somerville, NJ) were placed by using the following tech nique. After suitable lubrication (and ad ministration of topical anesthesia in pa tients who were awake), the tube and the
tube.
Although use of nasoentenic tubes is generally successful, complications have been reported, the vast majority of which are related to tube malposi tion (3—13).Blind tube passage, par ticulanly in patients with reduced neurosensory function, may result in perforation of the pharynx, esopha gus, or tracheobmonchial tree; the lat ten has been associated with fatal ten sion pneumothomax after tube with drawal (1 1). Coiling of the tube within the esophagus or in a hiatal hernia may lead to massive aspiration when high-volume tube feedings are begun. It has been suggested that me flux-induced
aspiration
may also oc
cur when the tube is positioned with in an atonic stomach (14). To prevent this complication,
tube feeding
is of
ten withheld until the side holes are distal to the pylorus. Since clinical as sessment of the position of the tube is unreliable, radiographic documen tation of successful tube placement is necessary. Fluoroscopically guided placement of nasoentenic feeding tubes directly circumvents the majority of problems associated with the use of these cath eters (15); accordingly, at our institu tion, virtually all feeding tube place
PATIENTS AND METHODS
stiffening
stylet were advanced without
use of fluoroscopy through either names. The patient was then turned to the right side and tube passage beyond the pharyn
gotracheal junction was fluoroscopically guided. In most patients, flexion of the neck facilitated passage beyond the hypo
pharynx into the proximal esophagus. When the metallic weight on the tube was observed to pass into the gastric lu men, 50-100 mL of air was insufflated to allow the tube to be maneuvered anteri orly; with the patient in the right decubi tus position, the tubes were easily manip
ulated out of the fundus and into the gastric antrum. At this point, some ma nipulation was usually necessary to allow the tubes to pass through the pyloric channel. Generally, the most successful
maneuver was to place the patient in a left lateral position while administering gentle upward pressure with a gloved hand along the greater curvature of the gastric
body.
When
the weighted
tip
passed into the duodenum, it could be easily advanced into the proximal jeju num. Although many of the patients re
ferred for placement of nasoenteric feed ing tubes had been receiving metoclo pramide hydrochloride (Reglan; Robins,
759
300
Cl)
The occurrence of aspiration was re corded for any patient in whom a wit nessed aspiration event was documented or in whom clinical and radiologic find ings supported a diagnosis of aspiration pneumonia. In each such case, the posi tion of the tube was documented by means of an abdominal radiograph, and this position was recorded. Patient outcome was classified as one of
200
z w
2
,
,@ 7
1
2
3
of patients
TUBES
5
requiring
6
one or
more tubes.
Richmond, Va), this drug was not rou tinely given during tube placement with in the radiology department. Eight hundred eighty-two feeding tube placements were performed in 448 pa tients during 1987. Of these, 135 proce dunes in 101 patients were performed at bedside in an intensive care unit with use of a portable C-arm fluoroscopy machine (Sireskop 2-0; Siemens, Iselin, NJ). All procedures were performed with the 8-F Entriflex catheter; the tubes were passed transnasally with use of the steel stiffen ing stylus (Biosearch Medical Products). After the tube position was fluoroscopi cally confirmed, patency was assured by injecting 5-10 mL of water. Approximate ly 90% of the tubes were placed by mesi dents in radiology; the remainder were placed by radiology staff. The mean age of the patients was 67.2 years (range, 3—101years). The medical records of all 448 patients were retrospec tively reviewed and the primary diagno sis, indication for feeding tube place ment, success of tube placement, total time of intubation, number of tubes used, complications related to tube placement, presence and number of aspiration events, and final patient outcome were recorded. We recorded the indication for enteric feeding as one of five categories: (a) se vere neumologic impairment, (b) known or suspected aspiration, (c) previous head and neck surgery, (d) severe underlying medical disease requiring nutritional sup port, and (e) prolonged
postoperative
phase requiring nutritional support. The tube placement was classified as “¿successful― if the tip was distal to the third portion of the duodenum. Place ment was classified as “¿partially success ful―if the tip was beyond the pylorus but proximal to the fourth portion of the duo denum. If the tip was proximal to the py lorus,
placement
was considered
“¿unsuc
cessful.―Patients who had undergone previous antrectomy and gastrojejunos tomy procedures were not classified ac cording to this scheme, but were recorded as “¿altered anatomy.― Total fluoroscopy time and total room time were not routinely recorded during the procedure. In the intensive care unit, fluoroscopy time did not exceed 15 mm utes.
760 •¿ Radiology
categories: of
tube
(a) death,
feeding
with
(b) discontinu reinstitution
of
4 I
Distribution
three ation
stomach was successfully decom pressed after 24 hours of nasogastnic tube suction. This allowed the tube to be successfully placed in all patients. In four patients, a Zenkem diverticu lum, a tracheoesophageal fistula, an esophageal stricture, and a gastric ul cem obstructing the pylomus were de tected because of failure of tube pas sage.
oral
feeding,
and
(c)
conversion
to
anoth
er method of providing nutrition (gas trostomy or intravenous pamenteral hy pemalimentation).
RESULTS Indications for Nasoentenic Alimentation The clinical indication for nasoen teric alimentation was for nutritional support for a severe underlying med ical disease in 202 patients (45.1%), neurologic impairment in 130 (29.0%), postoperative support in 69 (15.4%), prior laryngeal or phamynge al surgery in 24 (5.4%), and known or suspected aspiration in 23 (5.1%). The 101 patients who required bed side tube positioning differed sub stantially from the group as a whole; 82 (81%) needed nutritional support because of severe underlying medical disease, 11 (11%) had severe neumo logic impairment, four had aspima tion, and four had postoperative problems as the major indication for nasoentemic alimentation.
Success Rate of Initial Placement Initial tube placement was success ful in 764 placements (86.6%), partial ly successful in 43 (4.9%), and unsuc cessful in 65 (7.4%). Previous gastro jejunostomy had been performed in 10 patients, and these patients were classified as having altered anatomy. In the 135 procedures performed in the intensive came unit, 116 (85.9%) were successful, 1 1 (8.2%) were par tially successful, and seven (5%) were unsuccessful; one patient had under gone a previous gastrojejunostomy. The 65 tube placement failures were due to two major causes: the in ability to pass the tube through the pylomic channel (generally due to overdistention of the stomach com bined with gastric atony) in 40 place ments, and the inability to pass the tube beyond the hypopharynx (often due to the presence of an endotmache al tube) in 21. In the patients in whom the tube could not be passed through the pylomic channel, the
Repositioning
Rate
As a whole, 882 tubes were placed in 448 patients (mean, 2.0 tubes pen patient; mange, 1—12tubes). The dis tmibution of patients requiring one or more tubes is illustrated in the Fig ume. The average total time of na soentemic intubation for the entire group was 17 days (mange, 0-225 days); the average “¿tube life―(time between initial tube placement and tube repositioning on replacement) was 7.8 days (mange, 0—103days). The average total time of intubation for the group of patients in intensive came was 19.5 days (mange, 0—88days); mean tube life was 9.7 days (mange, 0-42 days). In 503 (77%) of the 654 cases in which tube repositioning was neces samy, the tube had become dislodged from its jejunal position. This was due to an uncooperative, unme strained patient in 466 cases, a bout of coughing in 11, vomiting in nine, and the subsequent placement of a nasogastnic tube in 17. In 151 of the 654 cases (23%), a well-positioned tube became clogged during normal operation. Four or more feeding tubes were required in 53 (12%) of the 448 pa tients. Twenty-three (43%) of these patients ultimately required long term gastrostomy tube placement.
Complications Four (0.4%) major complications were encountered during on immedi ately after nasoentemic tube place ment. Three patients had ammhyth mias that resulted in cardiac arrest, and one patient's tube was malposi tioned in the tracheobmonchial tree. All four complications occurred in the group of patients treated within the radiology department. The three patients with ammhythmias were more than 65 years old (two were over 90 years, the other had end-stage camdio myopathy). All three patients died of their arrhythmia. The patient with tube malposition had florid conges tive heart failure and could not tolem
March 1991
ate being placed in a supine position; the initial
phase
ment was, therefore, blindly
with
DISCUSSION
of the tube place
performed
the patient
seated
Nasoentenic on
tube feedings
come an important
the edge of the fluomoscopy table. When fluomoscopy was initiated, the
peutic approach that accompanies
malposition of the tube in the left pulmonary panenchyma was immedi ately recognized, and the tube was
eases (1,2). Commercially
withdrawn.
are two major difficulties
The patient had inter
have be
part of the thera
to the catabolic many chronic
state dis
available
tubes are designed to be placed with out fluoroscopic guidance, but theme
with that
monitored
during
the procedure. Only one instance of tracheobmonchial trauma due to mis direction occurred; this compares fa
the actual tube position is unreliable. For that reason, tube feedings are
vomably with the 1.3%-4% prevalence of pulmonary complications reported
quences. Diarrhea occurred during tube feeding in only six (1%) of the pa
generally
in two recent series (12,13). Anecdo tally, in our 8-year experience with fluomoscopically guided tube passage,
tients.
ly, several nepositionings are neces sary, and there is attendant delay in
was ascribed
intraduodenal
to cul
tune-proved Clostridium difficile infec tion in two patients; in four, it was believed to be due to intrajejunal tube feedings, and it stopped when tube use was discontinued. Aspiration Aspiration
pneumonia
that was not
present before institution feedings
developed
of tube
in 27 (6%) of the
patients. Of these, 18 had abdominal nadiographs documenting that the tube was in the stomach on esopha gus at the time of the aspiration.
piration despite
As
occurred in seven patients appropriate
tube placement;
tube position was not documented in the other two patients. The distribu tion of diagnoses
in the group
that
had aspiration was severe medical disease (1 1 patients), known previ ous aspiration
(eight
logic dysfunction and postoperative tient).
Nine
patients),
neuro
(seven patients), status (one pa
(9%) of the 101 intensive
cane unit patients developed tion.
withheld
aspira
until
of
was continuously
approach.
nal radiograph
assessment
It is to be expected that complica tions related to tube misplacement would be mare, since tube position
mittent bleeding from the tnacheo bronchial tree for the next 24 hours, but suffered no further conse
Diarrhea
First, clinical
current practice is to advise that gas trostomy or pamentemal nutrition be considered after placement of the fourth nasoentenic tube.
an abdomi
is obtained positioning.
to confirm Frequent
beginning parentemal nutrition. Sec ond, serious complications to blind
tube passage have been reported
(3-
13), particularly in patients with sen sony deficits. Fluoroscopically guided tube placement has been offered as a solution to both of these problems.
When we began to offer it at our in stitution
in 1982, there
was immedi
Two hundred
eleven (47%) pa
monitoring results in successful dis tal duodenal placement in the vast majority of cases. Failures of initial placement were due to two major fac tors: the inability to prevent the tube from entering the larynx, and the in ability to cannulate the pylorus be cause of a large, atonic stomach. In all
the patients in this study, we have encountered numerous episodes in which the major arrhythmia was pre
cases of initial failure due to gastric atony, a second ful. Most patients
attempt
was success
tolerate
long-term
nasoenteric
feedings well, and over
three-fourths intubations.
require only one or two Most mepositionings are
patient manipulates the tube, thereby dislodging it from its position. An
died while receiving tube feedings. One hundred thirteen patients (21%) ultimately required either a gastros tomy (108 patients) or intravenous parenteral feeding (five patients) to manage chronic malnutrition. In this latter group, the mean time of intu
high percentage of cases, the nursing staff inappropriately used the tube
prior
to discontinuing
nasoen
185 days); 59 (52%) of the 1 13 pa tients requiring gastrostomy had se
vere neurologic impairment. In two patients, the ultimate disposition could not be ascertained from chart review. Volume 178 •¿ Number 3
for administering
solid, finely
(all fatal) occurred
within
the year studied. This mate is also con sistent
with that of our larger,
8-year
experience. Although cardiac moni toring was not performed in any of
ceded by profound
bradycardia.
The
vagal stimulation produced by tube manipulation is probably the inciting
event. The three deaths in our series are most likely a reflection of the population studied than of the inher ent risks of the procedure; all of the fatalities
other common cause of tube failure is clogging. Reasons for tube clotting are difficult to substantiate, but in a
tenic feeding was 22.1 days (range, 1-
ing the mediastinum), and one duo denal perforation. The prevalence of the occurrence of minor anrhythmias during tube passage is unknown; three major an mhythmias
tients were discharged to home on skilled nursing facility came. One hundred twenty-two patients (27%)
bation
only a few addi
tional cases of misdirection, including one pneumothorax due to penetma tion of the left mainstem bronchus, two pyniform sinus perforations (one of which resulted in the tube enter
ate clinical acceptance; as a result, fewer than 30 nasoentemic tubes were placed without use of fluoroscopic guidance in the year studied. As would be expected, fluoroscopic
necessary because an unrestrained
Outcome
we have encountered
occurred
in elderly,
debili
tated patients with severe underlying cardiovascular disease. Our study confirms that document
ed aspiration
in patients undergoing
nasoentenic feeding is uncommon. In two-thirds of the patients who expe nienced aspiration, the tube was not in its original position; only seven (2%) patients had documented aspina
tion events with the tube in the ap propniate position. The mechanism for such aspiration is probably reflux
ination of the tube after removal shows that the powdered medicine forms a paste within the tube lumen. A small subset of patients require multiple repositionings, and altenna
of the feeding solution from the duo denum into an atonic stomach, with subsequent massive gastroesophageal neflux. Our retrospective study is unavoid ably flawed. The high degree of cli nician acceptance for the technique of fluoroscopically guided tube
tive methods for long-term nutrition are necessary. Nearly half of the pa
placement has reduced the number of nondinected tube placements to
tients who required placement of four or more tubes eventually me quired a permanent gastrostomy. Our
virtually zero. Before we can con dude that fluoroscopically guided
ground
medication
in addition
to the
feeding solution. In such cases, exam
tube placement
is superior,
we need
Radiology •¿ 761
to perform a comparison same population,
detailing
study in the the suc
cess and complication mates of blind tube passage. Moreover, in assessing our success, we assumed that a suc cessful placement was equivalent to reaching the distal duodenum; how ever, no study has confirmed that in tragastnic feedings are any less safe than intrajejunal feedings. In fact, the results of several series that dealt with gastrostomy tube feeding sug gest that intmagastnic feeding is safe unless
the patient
has known
gastro
esophageal meflux, gastric atony, or gastric outlet obstruction (16). Final ly, since a large percentage of our pa tients were very ill on unable to com
municate, it is likely that many com plications that actually occurred were not recorded (eg, mild diarrhea, ab dominal cramping, esophageal me flux, or nasal irritation would not be diagnosed in comatose patients). Despite these flaws, the results of our study indicate that fluomoscopi cally guided nasoentenic tube place ment is safe, inexpensive, and easy to perform.
Its widespread
clinical
ac
ceptance that it is agement vation.
at our institution suggests of major benefit in the man of severe nutritional depni U
9.
PR, Spirn P. Inadvertent transbronchial insertion of narrow-bone feeding tubes into the pleural space. JAMA 1984; 251:2396—2397. 10.
grad Med 1989;85:355—360. 2.
Dobbie
RP, Hoffmeisten
JA.
pump tube entenic hyperalimentation. 3.
Ghahremani
GG, Gould
RJ.
cheobronchial
4.
12.
5.
Odocha 0, Lowery Siram SM, Warner
enteral tube insen
tion. J Nat! Med Assoc 1989; 81:275-281.
6.
Siegle RL, Rabinowitz JG, Sanasohn C. Intestinal perforation secondary to nasoje junal feeding tubes. AJR 1976; 126:1229-
7.
Siemers PT, Reinke RT. Perforation
1232.
of
the nasopharynx by nasogastnic intuba tion: a rare cause of left pleural effusion and pneumomediastinum. AJR 1976; 127:341-343. 8.
Balogh
GJ, Adler
SJ, VanderWoude
of feeding tubes
Harris
MR. Huseby
JS.
Pulmonary
com
dence and prevention. Crit Care Med 13.
14.
RC, Mezghebe HM, 0G. Tracheopulmon
ary injuries following
placement
plications from nasoentenal feeding tube insertion in an intensive care unit: mci
Nasoentenic
Ghahremani GG, Turner MA, Pont RB. latrogenic intubation injuries of the upper gastrointestinal tract in adults. Gastroin test Radiol 1980; 5:1—10.
ME.
(letter). Radiology 1988;169:874-875.
feeding tubes: radiographic detection of complications. Dig Dis Sci 1986; 31:574585.
TB, Fon GT, Silverstein
11. Zeiss J, Woldenbeng LS. Inadvertent tra
Continuous
Sung Gynecol Obstet 1976; 143:273-276.
Hunter
Complications of intestinal tubes. Am Gastroenterol 1981; 76:256—261.
References 1. Cogen R, Weinryb J. Tube feeding. Post
Hand RW, Kempster M, Levy JH, Rogol
1989;17:917-919. McWey RE, Curry NS, Schabel SI, Reines HD. Complications of nasoentenic feed ing tubes. Am J Sung 1988; 155:253-257. Olivares L, Segovia A, Revuelta R. Tube feeding and lethal aspiration in neunolog ic patients:
15.
16.
a review
of 720 autopsy
cases.
Stroke 1974; 5:654—657. Grant JP, Curtas MS. Kelvin FM. Fluoro scopic placement of nasojejunal feeding tubes with immediate feeding using a nonelementa! diet. JPEN J Parenten Enter a! Nutn 1983; 7:299-303. Hicks ME, Surratt RS, Picus D, Marx MV, Lang EV. Fluoroscopically guided percu taneous gastrostomy and gastroenteros tomy: analysis of 158 consecutive cases. AJR 1990; 154:725-728.
J,
Glazer HS, Roper C, Weyman PJ. Pneu mothorax as a complication of feeding tube placement. AJR 1983; 141:1275—1277.
762 •¿ Radiology
March 1991
Fluoroscopically Tube Placement: Nasoenteric tube feeding is a widely used alternative to parenteral intra venous nutritional support or gas trostomy tube placement. Unmoni tored tube passage may result in complications and delays the begin fling of tube feedings. The authors studied the results of 882 fluoro scopically guided feeding tube placements in 448 patients in 1 year to determine rates of success and complications, as well as the long term outcome of this population of patients. Seven hundred sixty-four attempts (86.6%) were successful in positioning the tube distal to the third portion of the duodenum. Four major complications (three fa tal arrhythmias and one tracheo bronchial injury) were encountered. Only seven patients (2%) experi enced aspiration events that were due to positioning of the tube in the distal duodenum. Seventy-seven percent of patients required either one or two tubes; the average “¿tube life―was 7.8 days. Most reposition ings were required because of pa @tient noncompliance or inappropri ate administration of solid medica tions. Fluoroscopically guided nasoenteric tube passage is safe, eas ily performed, and highly success ful, and has resulted in widespread clinical acceptance in our institu tion. Index terms: Feeding tube, complications. Fluoroscopy
@Radiology
•¿ Nutrition
1991; 178:759-762
@ From the Mallincknodt Institute of Radiolo @ gy. Washington University School of Medicine, 510 S Kingshighway Blvd. St Louis, MO 63110.
Received July 20, 1990; revision requested tember
6; revision
received
October 10. Address reprint @
RSNA,
1991
October
Sep
9; accepted
requests to D.M.B.
Guided Results
Nasoenteric Feeding of a 1 -year Study@
S UCCESSFULtreatment of chronical ly ill patients requires the provi sion of adequate protein and caloric intake (1). In 1976, Dobbie and Hoff meister (2) found that nasoentemic in tubation was a viable alternative to intravenous panenteral delivery if the patient's gastrointestinal function was intact. Subsequently, a metal tipped polyurethane 8-F catheter has become commercially available and is used extensively for long-term nu tritional support. Early difficulties in passing this extremely flexible tube were solved with the introduction of a removable, steel stiffening stylus, which increases the rigidity of the system roughly to that of a standard
ments are performed by radiologists, either within the radiology depart ment or at bedside with use of a Carm portable fluomoscopy unit. We studied the results of fluomoscopically guided nasoenteric intubation pen formed in 448 patients in 1 year. The goals of this study were to determine our success and complication rates, to characterize the clinical indications for intubation, and to analyze the long-term performance of feeding tubes in this population, with specif ic reference to ultimate patient out come.
nasogastric
The 8-F feeding tubes (Entriflex; Bio search Medical Products, Somerville, NJ) were placed by using the following tech nique. After suitable lubrication (and ad ministration of topical anesthesia in pa tients who were awake), the tube and the
tube.
Although use of nasoentenic tubes is generally successful, complications have been reported, the vast majority of which are related to tube malposi tion (3—13).Blind tube passage, par ticulanly in patients with reduced neurosensory function, may result in perforation of the pharynx, esopha gus, or tracheobmonchial tree; the lat ten has been associated with fatal ten sion pneumothomax after tube with drawal (1 1). Coiling of the tube within the esophagus or in a hiatal hernia may lead to massive aspiration when high-volume tube feedings are begun. It has been suggested that me flux-induced
aspiration
may also oc
cur when the tube is positioned with in an atonic stomach (14). To prevent this complication,
tube feeding
is of
ten withheld until the side holes are distal to the pylorus. Since clinical as sessment of the position of the tube is unreliable, radiographic documen tation of successful tube placement is necessary. Fluoroscopically guided placement of nasoentenic feeding tubes directly circumvents the majority of problems associated with the use of these cath eters (15); accordingly, at our institu tion, virtually all feeding tube place
PATIENTS AND METHODS
stiffening
stylet were advanced without
use of fluoroscopy through either names. The patient was then turned to the right side and tube passage beyond the pharyn
gotracheal junction was fluoroscopically guided. In most patients, flexion of the neck facilitated passage beyond the hypo
pharynx into the proximal esophagus. When the metallic weight on the tube was observed to pass into the gastric lu men, 50-100 mL of air was insufflated to allow the tube to be maneuvered anteri orly; with the patient in the right decubi tus position, the tubes were easily manip
ulated out of the fundus and into the gastric antrum. At this point, some ma nipulation was usually necessary to allow the tubes to pass through the pyloric channel. Generally, the most successful
maneuver was to place the patient in a left lateral position while administering gentle upward pressure with a gloved hand along the greater curvature of the gastric
body.
When
the weighted
tip
passed into the duodenum, it could be easily advanced into the proximal jeju num. Although many of the patients re
ferred for placement of nasoenteric feed ing tubes had been receiving metoclo pramide hydrochloride (Reglan; Robins,
759
300
Cl)
The occurrence of aspiration was re corded for any patient in whom a wit nessed aspiration event was documented or in whom clinical and radiologic find ings supported a diagnosis of aspiration pneumonia. In each such case, the posi tion of the tube was documented by means of an abdominal radiograph, and this position was recorded. Patient outcome was classified as one of
200
z w
2
,
,@ 7
1
2
3
of patients
TUBES
5
requiring
6
one or
more tubes.
Richmond, Va), this drug was not rou tinely given during tube placement with in the radiology department. Eight hundred eighty-two feeding tube placements were performed in 448 pa tients during 1987. Of these, 135 proce dunes in 101 patients were performed at bedside in an intensive care unit with use of a portable C-arm fluoroscopy machine (Sireskop 2-0; Siemens, Iselin, NJ). All procedures were performed with the 8-F Entriflex catheter; the tubes were passed transnasally with use of the steel stiffen ing stylus (Biosearch Medical Products). After the tube position was fluoroscopi cally confirmed, patency was assured by injecting 5-10 mL of water. Approximate ly 90% of the tubes were placed by mesi dents in radiology; the remainder were placed by radiology staff. The mean age of the patients was 67.2 years (range, 3—101years). The medical records of all 448 patients were retrospec tively reviewed and the primary diagno sis, indication for feeding tube place ment, success of tube placement, total time of intubation, number of tubes used, complications related to tube placement, presence and number of aspiration events, and final patient outcome were recorded. We recorded the indication for enteric feeding as one of five categories: (a) se vere neumologic impairment, (b) known or suspected aspiration, (c) previous head and neck surgery, (d) severe underlying medical disease requiring nutritional sup port, and (e) prolonged
postoperative
phase requiring nutritional support. The tube placement was classified as “¿successful― if the tip was distal to the third portion of the duodenum. Place ment was classified as “¿partially success ful―if the tip was beyond the pylorus but proximal to the fourth portion of the duo denum. If the tip was proximal to the py lorus,
placement
was considered
“¿unsuc
cessful.―Patients who had undergone previous antrectomy and gastrojejunos tomy procedures were not classified ac cording to this scheme, but were recorded as “¿altered anatomy.― Total fluoroscopy time and total room time were not routinely recorded during the procedure. In the intensive care unit, fluoroscopy time did not exceed 15 mm utes.
760 •¿ Radiology
categories: of
tube
(a) death,
feeding
with
(b) discontinu reinstitution
of
4 I
Distribution
three ation
stomach was successfully decom pressed after 24 hours of nasogastnic tube suction. This allowed the tube to be successfully placed in all patients. In four patients, a Zenkem diverticu lum, a tracheoesophageal fistula, an esophageal stricture, and a gastric ul cem obstructing the pylomus were de tected because of failure of tube pas sage.
oral
feeding,
and
(c)
conversion
to
anoth
er method of providing nutrition (gas trostomy or intravenous pamenteral hy pemalimentation).
RESULTS Indications for Nasoentenic Alimentation The clinical indication for nasoen teric alimentation was for nutritional support for a severe underlying med ical disease in 202 patients (45.1%), neurologic impairment in 130 (29.0%), postoperative support in 69 (15.4%), prior laryngeal or phamynge al surgery in 24 (5.4%), and known or suspected aspiration in 23 (5.1%). The 101 patients who required bed side tube positioning differed sub stantially from the group as a whole; 82 (81%) needed nutritional support because of severe underlying medical disease, 11 (11%) had severe neumo logic impairment, four had aspima tion, and four had postoperative problems as the major indication for nasoentemic alimentation.
Success Rate of Initial Placement Initial tube placement was success ful in 764 placements (86.6%), partial ly successful in 43 (4.9%), and unsuc cessful in 65 (7.4%). Previous gastro jejunostomy had been performed in 10 patients, and these patients were classified as having altered anatomy. In the 135 procedures performed in the intensive came unit, 116 (85.9%) were successful, 1 1 (8.2%) were par tially successful, and seven (5%) were unsuccessful; one patient had under gone a previous gastrojejunostomy. The 65 tube placement failures were due to two major causes: the in ability to pass the tube through the pylomic channel (generally due to overdistention of the stomach com bined with gastric atony) in 40 place ments, and the inability to pass the tube beyond the hypopharynx (often due to the presence of an endotmache al tube) in 21. In the patients in whom the tube could not be passed through the pylomic channel, the
Repositioning
Rate
As a whole, 882 tubes were placed in 448 patients (mean, 2.0 tubes pen patient; mange, 1—12tubes). The dis tmibution of patients requiring one or more tubes is illustrated in the Fig ume. The average total time of na soentemic intubation for the entire group was 17 days (mange, 0-225 days); the average “¿tube life―(time between initial tube placement and tube repositioning on replacement) was 7.8 days (mange, 0—103days). The average total time of intubation for the group of patients in intensive came was 19.5 days (mange, 0—88days); mean tube life was 9.7 days (mange, 0-42 days). In 503 (77%) of the 654 cases in which tube repositioning was neces samy, the tube had become dislodged from its jejunal position. This was due to an uncooperative, unme strained patient in 466 cases, a bout of coughing in 11, vomiting in nine, and the subsequent placement of a nasogastnic tube in 17. In 151 of the 654 cases (23%), a well-positioned tube became clogged during normal operation. Four or more feeding tubes were required in 53 (12%) of the 448 pa tients. Twenty-three (43%) of these patients ultimately required long term gastrostomy tube placement.
Complications Four (0.4%) major complications were encountered during on immedi ately after nasoentemic tube place ment. Three patients had ammhyth mias that resulted in cardiac arrest, and one patient's tube was malposi tioned in the tracheobmonchial tree. All four complications occurred in the group of patients treated within the radiology department. The three patients with ammhythmias were more than 65 years old (two were over 90 years, the other had end-stage camdio myopathy). All three patients died of their arrhythmia. The patient with tube malposition had florid conges tive heart failure and could not tolem
March 1991
ate being placed in a supine position; the initial
phase
ment was, therefore, blindly
with
DISCUSSION
of the tube place
performed
the patient
seated
Nasoentenic on
tube feedings
come an important
the edge of the fluomoscopy table. When fluomoscopy was initiated, the
peutic approach that accompanies
malposition of the tube in the left pulmonary panenchyma was immedi ately recognized, and the tube was
eases (1,2). Commercially
withdrawn.
are two major difficulties
The patient had inter
have be
part of the thera
to the catabolic many chronic
state dis
available
tubes are designed to be placed with out fluoroscopic guidance, but theme
with that
monitored
during
the procedure. Only one instance of tracheobmonchial trauma due to mis direction occurred; this compares fa
the actual tube position is unreliable. For that reason, tube feedings are
vomably with the 1.3%-4% prevalence of pulmonary complications reported
quences. Diarrhea occurred during tube feeding in only six (1%) of the pa
generally
in two recent series (12,13). Anecdo tally, in our 8-year experience with fluomoscopically guided tube passage,
tients.
ly, several nepositionings are neces sary, and there is attendant delay in
was ascribed
intraduodenal
to cul
tune-proved Clostridium difficile infec tion in two patients; in four, it was believed to be due to intrajejunal tube feedings, and it stopped when tube use was discontinued. Aspiration Aspiration
pneumonia
that was not
present before institution feedings
developed
of tube
in 27 (6%) of the
patients. Of these, 18 had abdominal nadiographs documenting that the tube was in the stomach on esopha gus at the time of the aspiration.
piration despite
As
occurred in seven patients appropriate
tube placement;
tube position was not documented in the other two patients. The distribu tion of diagnoses
in the group
that
had aspiration was severe medical disease (1 1 patients), known previ ous aspiration
(eight
logic dysfunction and postoperative tient).
Nine
patients),
neuro
(seven patients), status (one pa
(9%) of the 101 intensive
cane unit patients developed tion.
withheld
aspira
until
of
was continuously
approach.
nal radiograph
assessment
It is to be expected that complica tions related to tube misplacement would be mare, since tube position
mittent bleeding from the tnacheo bronchial tree for the next 24 hours, but suffered no further conse
Diarrhea
First, clinical
current practice is to advise that gas trostomy or pamentemal nutrition be considered after placement of the fourth nasoentenic tube.
an abdomi
is obtained positioning.
to confirm Frequent
beginning parentemal nutrition. Sec ond, serious complications to blind
tube passage have been reported
(3-
13), particularly in patients with sen sony deficits. Fluoroscopically guided tube placement has been offered as a solution to both of these problems.
When we began to offer it at our in stitution
in 1982, there
was immedi
Two hundred
eleven (47%) pa
monitoring results in successful dis tal duodenal placement in the vast majority of cases. Failures of initial placement were due to two major fac tors: the inability to prevent the tube from entering the larynx, and the in ability to cannulate the pylorus be cause of a large, atonic stomach. In all
the patients in this study, we have encountered numerous episodes in which the major arrhythmia was pre
cases of initial failure due to gastric atony, a second ful. Most patients
attempt
was success
tolerate
long-term
nasoenteric
feedings well, and over
three-fourths intubations.
require only one or two Most mepositionings are
patient manipulates the tube, thereby dislodging it from its position. An
died while receiving tube feedings. One hundred thirteen patients (21%) ultimately required either a gastros tomy (108 patients) or intravenous parenteral feeding (five patients) to manage chronic malnutrition. In this latter group, the mean time of intu
high percentage of cases, the nursing staff inappropriately used the tube
prior
to discontinuing
nasoen
185 days); 59 (52%) of the 1 13 pa tients requiring gastrostomy had se
vere neurologic impairment. In two patients, the ultimate disposition could not be ascertained from chart review. Volume 178 •¿ Number 3
for administering
solid, finely
(all fatal) occurred
within
the year studied. This mate is also con sistent
with that of our larger,
8-year
experience. Although cardiac moni toring was not performed in any of
ceded by profound
bradycardia.
The
vagal stimulation produced by tube manipulation is probably the inciting
event. The three deaths in our series are most likely a reflection of the population studied than of the inher ent risks of the procedure; all of the fatalities
other common cause of tube failure is clogging. Reasons for tube clotting are difficult to substantiate, but in a
tenic feeding was 22.1 days (range, 1-
ing the mediastinum), and one duo denal perforation. The prevalence of the occurrence of minor anrhythmias during tube passage is unknown; three major an mhythmias
tients were discharged to home on skilled nursing facility came. One hundred twenty-two patients (27%)
bation
only a few addi
tional cases of misdirection, including one pneumothorax due to penetma tion of the left mainstem bronchus, two pyniform sinus perforations (one of which resulted in the tube enter
ate clinical acceptance; as a result, fewer than 30 nasoentemic tubes were placed without use of fluoroscopic guidance in the year studied. As would be expected, fluoroscopic
necessary because an unrestrained
Outcome
we have encountered
occurred
in elderly,
debili
tated patients with severe underlying cardiovascular disease. Our study confirms that document
ed aspiration
in patients undergoing
nasoentenic feeding is uncommon. In two-thirds of the patients who expe nienced aspiration, the tube was not in its original position; only seven (2%) patients had documented aspina
tion events with the tube in the ap propniate position. The mechanism for such aspiration is probably reflux
ination of the tube after removal shows that the powdered medicine forms a paste within the tube lumen. A small subset of patients require multiple repositionings, and altenna
of the feeding solution from the duo denum into an atonic stomach, with subsequent massive gastroesophageal neflux. Our retrospective study is unavoid ably flawed. The high degree of cli nician acceptance for the technique of fluoroscopically guided tube
tive methods for long-term nutrition are necessary. Nearly half of the pa
placement has reduced the number of nondinected tube placements to
tients who required placement of four or more tubes eventually me quired a permanent gastrostomy. Our
virtually zero. Before we can con dude that fluoroscopically guided
ground
medication
in addition
to the
feeding solution. In such cases, exam
tube placement
is superior,
we need
Radiology •¿ 761
to perform a comparison same population,
detailing
study in the the suc
cess and complication mates of blind tube passage. Moreover, in assessing our success, we assumed that a suc cessful placement was equivalent to reaching the distal duodenum; how ever, no study has confirmed that in tragastnic feedings are any less safe than intrajejunal feedings. In fact, the results of several series that dealt with gastrostomy tube feeding sug gest that intmagastnic feeding is safe unless
the patient
has known
gastro
esophageal meflux, gastric atony, or gastric outlet obstruction (16). Final ly, since a large percentage of our pa tients were very ill on unable to com
municate, it is likely that many com plications that actually occurred were not recorded (eg, mild diarrhea, ab dominal cramping, esophageal me flux, or nasal irritation would not be diagnosed in comatose patients). Despite these flaws, the results of our study indicate that fluomoscopi cally guided nasoentenic tube place ment is safe, inexpensive, and easy to perform.
Its widespread
clinical
ac
ceptance that it is agement vation.
at our institution suggests of major benefit in the man of severe nutritional depni U
9.
PR, Spirn P. Inadvertent transbronchial insertion of narrow-bone feeding tubes into the pleural space. JAMA 1984; 251:2396—2397. 10.
grad Med 1989;85:355—360. 2.
Dobbie
RP, Hoffmeisten
JA.
pump tube entenic hyperalimentation. 3.
Ghahremani
GG, Gould
RJ.
cheobronchial
4.
12.
5.
Odocha 0, Lowery Siram SM, Warner
enteral tube insen
tion. J Nat! Med Assoc 1989; 81:275-281.
6.
Siegle RL, Rabinowitz JG, Sanasohn C. Intestinal perforation secondary to nasoje junal feeding tubes. AJR 1976; 126:1229-
7.
Siemers PT, Reinke RT. Perforation
1232.
of
the nasopharynx by nasogastnic intuba tion: a rare cause of left pleural effusion and pneumomediastinum. AJR 1976; 127:341-343. 8.
Balogh
GJ, Adler
SJ, VanderWoude
of feeding tubes
Harris
MR. Huseby
JS.
Pulmonary
com
dence and prevention. Crit Care Med 13.
14.
RC, Mezghebe HM, 0G. Tracheopulmon
ary injuries following
placement
plications from nasoentenal feeding tube insertion in an intensive care unit: mci
Nasoentenic
Ghahremani GG, Turner MA, Pont RB. latrogenic intubation injuries of the upper gastrointestinal tract in adults. Gastroin test Radiol 1980; 5:1—10.
ME.
(letter). Radiology 1988;169:874-875.
feeding tubes: radiographic detection of complications. Dig Dis Sci 1986; 31:574585.
TB, Fon GT, Silverstein
11. Zeiss J, Woldenbeng LS. Inadvertent tra
Continuous
Sung Gynecol Obstet 1976; 143:273-276.
Hunter
Complications of intestinal tubes. Am Gastroenterol 1981; 76:256—261.
References 1. Cogen R, Weinryb J. Tube feeding. Post
Hand RW, Kempster M, Levy JH, Rogol
1989;17:917-919. McWey RE, Curry NS, Schabel SI, Reines HD. Complications of nasoentenic feed ing tubes. Am J Sung 1988; 155:253-257. Olivares L, Segovia A, Revuelta R. Tube feeding and lethal aspiration in neunolog ic patients:
15.
16.
a review
of 720 autopsy
cases.
Stroke 1974; 5:654—657. Grant JP, Curtas MS. Kelvin FM. Fluoro scopic placement of nasojejunal feeding tubes with immediate feeding using a nonelementa! diet. JPEN J Parenten Enter a! Nutn 1983; 7:299-303. Hicks ME, Surratt RS, Picus D, Marx MV, Lang EV. Fluoroscopically guided percu taneous gastrostomy and gastroenteros tomy: analysis of 158 consecutive cases. AJR 1990; 154:725-728.
J,
Glazer HS, Roper C, Weyman PJ. Pneu mothorax as a complication of feeding tube placement. AJR 1983; 141:1275—1277.
762 •¿ Radiology
March 1991
