Original Contribution

Percutaneous feeding tube replacement in the ED—are confirmatory x-rays necessary?

Gregory Jacobson MD, Peter A. Brokish MD, Keith Wrenn MD?

Department of Emergency Medicine, Vanderbilt University School of Medicine, Nashville, TN 37232-4700, USA

Received 25 February 2008; revised 2 April 2008; accepted 3 April 2008

Abstract

Objective: No study to date has addressed whether confirmatory x-ray after ED percutaneous feeding tube (PFT) replacement is always necessary. We hypothesized that x-ray confirmation of PFT replacement is not necessary in patients lacking both tract immaturity and trauma to the tract during dislodgement or replacement. Therefore confirmatory x-rays could safely be avoided for these patients. Methods: Medical records of 113 adult PFT encounters that met inclusion criteria between December 2000 and March 2004 at an urban university hospital ED seeing approximately 50 000 adult patients per year were reviewed.

Results: Ninety-four patients (83%) presented secondary to dislodgement, and 19 patients (17%) presented secondary to malfunction. Forty-seven patients (42%) did not have confirmatory x-rays, and 66 (58%) had confirmatory x-rays. None of the patients discharged without a confirmatory x-ray returned with evidence of improper PFT placement. Of the 66 patients who had a confirmatory x-ray, 62 (94%) x-rays showed the feeding tube was in the correct location. In 4 patients, the x-ray showed either the PFT was not in the stomach or evidence of tract compromise. All of these patients had trauma to the tract and 3 of 4 were in immature tracts. There were only 19 patients of the total group of 113, however, who had neither trauma nor an immature tract.

Conclusion: Although uncommon, significant problems with PFT placement occur with the potential for catastrophic consequences. It appears that immaturity of the tract and trauma to the tract, either potential or actual, are the major risk factors for such complications.

(C) 2009

Introduction

Percutaneous endoscopic gastroscopy, first described by Gauderer et al [1] in 1980, has become a common and relatively safe method for providing long-term Enteral feedings [2-4]. The number of types and placements of percutaneous feeding tubes (PFTs) continues to increase

* Corresponding author. Tel.: +1 615 936 1157; fax: +1 615 936 1316.

E-mail address: [email protected] (K. Wrenn).

[4-7]. If the patient’s need for a PFT continues, at least 8% will eventually malfunction unexpectedly [4]. Malfunction can be caused by a tube becoming dislodged (via traction or balloon malfunction), developing a leak, becoming clogged, causing intestinal obstruction, developing a gastrocolic fistula, or developing an infection around the tube site [4,5,8-26].

The replacement of a PFT in the ED is plagued by the worry about Tract disruption and the potential for positioning intraperitoneally with catastrophic consequences. Tradition- ally, radiographic confirmation has been suggested after

0735-6757/$ - see front matter (C) 2009 doi:10.1016/j.ajem.2008.04.005

infusion of water-soluble contrast through a newly replaced tube. The available literature is largely from retrospective studies in settings other than the ED and is lacking in consistency. There is one study showing no complications in a large group of patients [27]. There are studies suggesting all replacements should be confirmed by x-ray [5,6,10,14,15] and studies asserting that x-ray confirmation is only necessary when there is a concern as to the maturity or integrity of the tract [10,13,28].

There are 2 characteristics consistently recognized as contributing to complications when replacing PFTs. These are trauma to the tract during dislodgement or replacement [6,8,10,14,15,28,29] and tract immaturity [8,11- 14,17,18,23,28]. No one disputes the importance of trauma, but there is inconsistency concerning the definition of tract immaturity. The range of opinions is from 1 week to 6 months [11-14,17,19]. One historical comparative study showed that waiting 6 months before attempting elective replacement decreased complications compared to others who waited less time [14].

A prolonged time interval between dislodgement and replacement has also been mentioned, yet with much less data for support [13,17]. Postinsertion gastric aspiration through a PFT after replacement has not been shown to be a fail-safe mechanism to ensure intragastric location [6,15,28,29]. There is literature postulating delayed tract maturity in patients with severe malnutrition [3,14,23,26,28]. Yet others have been unable to demonstrate malnutrition as a risk factor for tract disruption [6]. Malnutrition is likely to be commonplace among patients with PFTs and that would make evaluation of this variable difficult.

We hypothesized that x-ray confirmation of PFT replace- ment is not necessary in patients without either of the 2 main risk factors for tract disruption and that confirmatory x-rays could safely be avoided, thereby decreasing cost and exposure to radiation.

Methods

Study design

This was a retrospective chart review with preexisting inclusion and exclusion factors. This study was exempted by the institutional review board because it was a retrospective study in which all patient identifiers were removed from the data collection sheet after the data were collected.

Study setting and population

The medical records of all encounters between the dates 12/15/00 and 03/23/04 at an urban university hospital ED, which sees approximately 50 000 adult patients per year, were reviewed [29].

Study protocol

Charts were identified via billing records. The following ICD-9 codes were searched by the ED billing department:

536.41 (infection of gastrostomy), 536.42 (mechanical complication of gastrostomy), 536.49 (gastrostomy compli- cation), 536.40 (unspecified gastrostomy complication), and v55.1 (attention to gastrostomy).

Preexisting inclusion criteria were as follows: (1) adults greater than 18 years of age, (2) any patient with a PFT that required replacement, (3) replacement was attempted in the ED or the prehospital setting, and (4) it was felt that the PFT was in the correct position and either the patient was discharged or PFT position was confirmed radiographically. We excluded patients in whom one would not have been able to obtain a confirmatory x-ray after the tube was replaced. Reasons for this were (1) the ED or consult services failed in replacing the PFT, (2) it was determined that the patient no longer needed a PFT, or (3) the PFT was primarily replaced surgically, endoscopically, or fluoroscopically. If a patient was seen more than once for a PFT malfunction during the study period, all visits were included.

The electronic medical record was searched first, then the paper chart. The dictation of the emergency medicine attending physician was the primary document used. Other documents were used to obtain information lacking in the primary attending’s note. In order, these were any other electronic attending note in the medical record (including clinic notes and operative notes), resident notes, nursing progress and triage notes, and emergency medical services notes. If there was a contradiction in the medical record, the emergency medicine attending note information was used except in 2 situations. The final Radiology report was the primary document used to determine whether a confirmatory radiograph was done and whether it confirmed tube placement. Secondly, other attending notes and operative notes were used as primary documents in determining age of the tract and indications for the feeding tube.

Measurements

Standardized data fields included (1) the patient age; (2) sex; (3) initial reason for PFT placement; (4) tract age; (5) length of time the tube was out; (6) nature of the tube coming out; (7) type of replacement tube; (8) whether an x-ray was done for confirmation, and if so, the result; (9) whether problems were encountered on tube placement; (10) whether a dilator was used; and (11) whether there was a complication. All data fields were predetermined except the age of the patient and the reason for feeding tube, therefore limiting the ambiguity of answers.

Trauma to the tract was defined as any action that placed force or stress on the tract itself (ie, tube being pulled out with the balloon being inflated, unknown circumstances surround- ing the dislodgement of the PFT, difficultly in replacing the

PFT as evidenced by multiple attempts or unsuccessful attempts, and/or dilator use with replacing the PFT). We made the assumptions that a tube found out was likely to have come out with some traction and that any difficulty in replacement, including use of a dilator, was likely associated with some tract trauma. Tube “leak/malfunction” was defined as a tube that was still in place on arrival, but that required replacement because of a leak, looseness, or obstruction. Specifically, if a tube was “found out” with or without a balloon rupture, it was not included in this category. To include only tracts that would be considered mature to everyone’s definition in the literature, tract maturity was defined as a tract greater than 6 months of age [11-14,17-19,26].

For grouping the reasons for initial PFT placement, “neurologic” included any swallowing disorder caused by a neurologic abnormality, including gastroparesis and achala- sia. “Obstruction” included those patients with malignancies or surgeries of the swallowing tract. “Malnutrition” included patients not in the prior groups for whom a malnourished state was the primary reason for placement. Complications were defined as unsuccessful attempts in the ED or evidence

Table 1 Demographic/Patient data

of inappropriate placement either by x-ray or clinically. “Significant” complications were defined as those resulting in the need for surgery or having the potential to cause death from peritonitis.

There were 3 chart reviewers. Of the 113 charts patient visits included, chart reviewer 1 (CR1), a physician, examined 70 charts. Chart reviewer 2 (CR2), a physician, examined 13 charts. Chart reviewer 3 (CR3), a medical student, examined 49 charts. Chart reviewer 1 used the first 15 charts to help refine the data collection sheet. This was done in coordination with CR2. Once the data collection sheet was finalized all these charts were reviewed again. Chart reviewer 3 was trained over a several-hour session to ensure there was consistency in data collection. If CR3 ran into any questions they were reviewed with CR1. ? Value for agreement between CR1 and CR3 was 0.966 and between CR2 and CR3 was 1.0.

Statistical analysis was done using JMP version 5 (SAS Institute, Inc, Cary, NC). For continuous variables with a normal distribution, the mean and SD are reported. Where appropriate, 95% confidence intervals are reported. To

X-rays		No x-rays	All patients	P ?
n	66	47	113
Sex (male)	34 (52%)	23 (49%)	57 (50%)	.99
Mean age (y)	56.7 (SD 20)	56.7 (SD 22)	56.7 (SD 20)	.99
Reason
Neurologic	45 (68%)	35 (76%)	80 (71%)	.15
Obstruction	13 (20%)	11 (22%)	24 (21%)
Malnutrition	8 (12%)	1 (2%)	9 ( 8%)
Tract age
b6 mo	26 (39%)	8 (17%)	34 (30%)	.01
N6 mo	40 (61%)	39 (83%)	79 (70%)
Time out
b6 h	40 (61%)	33 (70%)	73 (65%)	.57
N6 h	13 (19%)	7 (15%)	20 (17.5%)
Unknown	13 (19%)	7 (15%)	20 (17.5%)
Trauma
Yes	45 (68%)	21 (45%)	66 (58%)	.29
No	21 (32%)	26 (55%)	47 (42%)
Problem
Leak/malfunction	9 (14%)	10 (21%)	19 (17%)	.32
Tube found out	57 (86%)	37 (79%)	94 (83%)
Problem replacing
Yes	9 (14%)	1 (2%)	10 (9%)	.03
No	57 (86%)	46 (98%)	103 (91%)
Dilator used
Yes	7 (11%)	4 (9%)	11 (10%)	.71
No	59 (89%)	43 (91%)	102 (90%)
Consults
Yes	25 (38%)	8 (17%)	33 (29%)	.02
No	41 (62%)	39 (83%)	80 (71%)
Significant complication	4 (6%)	0 (0%)	4 (4%)	.08
No risk ??	16 (34%)	10 (15%)	26 (23%)	.02
* Comparing first and second columns. ?? Tract age >=6 months and no possible trauma.

Age		Sex	Age of track	Length of time out	Nature of tube problem	Events	Possibility of or confirmation of trauma	Mature track
Patient 1	71	M	5 wk	0-6 h	Pulled out	PFT replaced by nursing home. Sent to ED for confirmation. Gastrograffin confirmatory study showed intraperitoneal extravasation. ED attempt unsuccessful. PFT placed by surgery with a dilator. Gastrograffin confirmatory study showed intraperitoneal extravasation. PFT replaced by ED. “successful.” Gastrograffin study revealed PFT in colon through well-developed fistula. No intraperitoneal extravasation. PFT replaced by surgery. Area around PFT noted to be firm. Gastrograffin study showed PFT to be in gastrocutenous fistula. No intraperitoneal extravasation.	Yes	No
Patient 2	69	F	3 mo	Unknown	Pulled out		Yes	No
Patient 3	43	M	4 mo	0-6 h	Pulled out		Yes	No
Patient 4	33	M	N6 mo	12-24 h	Pulled out		Yes	Yes
M indicates male; F, female.

compare categorical variables, contingency tables with ?2

Table 2 Patients with significant complications

testing were used.

Results

A total of 223 adult patient encounters were seen over a 39-month period for a feeding tube malfunction. Of these, 96 encounters were excluded because the tube was not a PFT or a PFT was malfunctioning but did not require replacement during the ED stay. Fourteen were excluded because interventional radiology replaced the tube primarily under fluoroscopy without an attempt in the ED.

Therefore, 113 encounters met inclusion criteria. We divided them into 2 groups based upon whether a confirmatory x-ray was performed (66 patients, 58%) or not (47 patients, 42%). Table 1 gives the relevant data for each group and the group as a whole.

There were significant differences between the group who was x-rayed and the group who was not in terms of tract age, problems replacing the tube, or use of consultants. There were fewer encounters in the group x-rayed where none of the identified major risk factors for tract disruption were present (Table 1). In addition, there was a trend toward the “no x-ray” group having more malfunctions/leaks as opposed to tubes that came out. Therefore, it appears that the group who was not x-rayed had fewer risks for significant complications.

A large majority of patients had a neurologic problem as the reason for initial tube placement (Table 1). The

presenting problem with the tube itself was most often that it had been dislodged and, in the majority of cases, trauma was specifically mentioned to have occurred (Table 1).

There were 4 significant complications, an average of just more than 1 per year or almost 4% of patients (Table 2). Two patients had intraperitoneal extravasation. All the patients with significant complications had at least one of the major risks for tract disruption during replacement and all were x- rayed postinsertion. Of the entire group, there were only 19 patients (17%) with none of these risks. In the group in whom no confirmatory x-ray was performed, a majority of patients also had either an immature tract, definite trauma, or a tube that was found out.

Discussion

Whether confirmatory x-rays after replacement of a PFT in the ED are routinely needed has not been specifically addressed. In many studies of PFT replacement in outpatient settings other than the ED, there is no information about confirmatory x-rays or the study was done on patients having “scheduled” replacements [3,4,9,14].

In comparing the indications for initial placement of PFTs, our population’s indications are similar to others, with neurologic being the majority and obstructive and malnutri- tion accounting for the rest [2-4,6,9,14]. However, our patients presented more often for tube dislodgement compared to studies where replacement occurred in the Primary care setting [4,9].

The literature is consistent in its opinion that replace- ment of a PFT in an immature tract is potentially dangerous [8,11-14,17]. Others go further and state PFT replacement in a mature tract is safe [9,13,27]. Mellinger et al [30] looked at tract formation in dogs up to 14 days, which showed that there is an increase in collagen deposit from 7 to 14 days. Yet, no research has been done on humans. The range of opinions on the age a tract must be to be considered a mature tract is 1 week to 6 months [11-14,17,19]. Fox et al [14] waited an average of 6 months before attempting elective percutaneous endo- scopic Gastrostomy tube replacements and had a decrease in tract disruptions compared to others who on average waited less time. However, just because a tract has reached a certain age does not guarantee that it is immune to disruption [6,10,14,15].

There are several reports of significant complications, including peritonitis and death, when replacement with button-type devices is performed [5,6,8,10,14,15,28]. These devices have rigid internal parts, which are pushed through the tract with traction. This suggests that inflicting trauma to the tract can lead to disruptions in the tract and thus cause placement failures. Unfortunately, important details such as age of the tract and whether trauma to the tract on dislodgment had occurred are not reported. There is one case report of a percutaneous endoscopic gastrostomy tube being dislodged by an unknown mechanism where a balloon-type gastrostomy tube was replaced in a mature tract with resultant peritonitis [10]. This author and others conclude that in cases in which the gastrostomy tube dislodgement occurs by external traction it would be prudent to determine position by either radiographic assistance or endoscopic verification [10,14].

The significant complication rate in this series was 3.5% (95% confidence interval, 1%-9%), just more than 1 patient per year. In all 4 patients with a significant complication, trauma was documented or could not be excluded, and in 3 of 4 the tract was immature, 1 was less than 3 months of age, and 2 were less than 6 months of age (Table 2). However, of the 113 total patients, there were 94 patients (83%) in whom at least one of the major risk factors was present.

It appears that the physicians in this study were actually quite adept at picking out who needed x-ray confirmation. But most of the patients who were not x-rayed had at least one of the major risk factors for complications, usually the inability to exclude trauma on dislodgment.

A new method of x-ray confirmation using air insufflation rather than contrast has been reported but was a small retrospective study from a rehabilitation hospital and did not report the specific reasons for tube replacement [31]. This method would be cheaper because it eliminates the need for contrast and easier because it would not require the involvement of a radiologist to administer the contrast and could actually be done at bedside using portable x-ray.

Conclusions

Our data show that significant complications were uncommon and in all these cases a major risk factor was present. It appears that physicians were very good at selecting patients who were at special risk and needed an x-ray because all complications were in patients who were x- rayed. On the other hand, significant complications were not rare and most of the patients had a major risk factor, whether or not an x-ray was done or a complication occurred. The incremental cost (in charges) of x-raying everyone at risk would have been $4978 over 39 months (19 patients without any major risk x $262 charge per x-ray). These costs pale in comparison to those associated with a single missed complication. It appears that trauma to the tract and tract immaturity are the major risk factors for a significant complication after replacement in an ED setting. A prospective study with attention to the major factors thought to contribute to the risk of significant complications is indicated. Perhaps then, it would be easier to make more specific recommendations about when confirmatory x-rays are needed or can be safely avoided.

References

1. Gauderer MWL, Ponsky JL, Izant Jr RJ. Gastrostomy without laparotomy: a percutaneous endoscopic technique. J Pediatr Surg 1980;15(6):872-5.
2. Nicholson FB, Korman MG, Richardson MA. Percutaneous endo- scopic gastrostomy: a review of indications, complications and outcome. J Gastroenterol Hepatol 2000;15(1):21-5.
3. Amann W, Mischinger HJ, Berger A, Rosanelli G, Schweiger W, Werkgartner G, et al. Percutaneous endoscopic gastrostomy (PEG). 8 years of clinical experience in 232 patients. Surg Endosc 1997;11(7): 741-4.
4. Koulentaki M, Reynolds N, Steinke D, Tait J, Baxter J, Vaidya K, et al. Eight years’ experience of gastrostomy tube management. Endoscopy 2002;34(12):941-5.
5. Platt MS, Roe DC. Complications following insertion and replacement of percutaneous endoscopic gastrostomy (PEG) tubes. J Forensic Sci 2000;45(4):833-5 [Erratum in: J Forensic Sci 2000 Sep;45(5):1179].
6. Romero R, Martinez FL, Robinson SY, Sullivan KM, Hart MH. Complicated PEG-to-skin level gastrostomy conversions: analysis of risk factors for tract disruption. Gastrointest Endosc 1996;44(3):230-4.
7. Grant MD, Rudberg MA, Brody JA. Gastrostomy placement and mortality among hospitalized Medicare beneficiaries. JAMA 1998;279 (24):1973-6.
8. Meeroff JC. Removal and replacement of conventional PEG catheters: the role of endoscopy. Endosc Rev 1991:31-6.
9. Kadakia SC, Cassaday M, Shaffer RT. Comparison of Foley catheter as a replacement gastrostomy tube with commercial replacement gastrostomy tube: a prospective randomized trial. Gastrointest Endosc 1994;40(2 Pt 1):188-93.
10. Tan YM, Abdullah M, Goh KL. Hemoperitoneum after accidental dislodgement and subsequent replacement of PEG tube. Gastrointest Endosc 2001;53(6):671-3.
11. Pofahl WE, Ringold F. Management of early dislodgment of percutaneous endoscopic gastrostomy tubes. Surg Laparosc Endosc Percutan Tech 1999;9(4):253-6.
12. Galat SA, Gerig KD, Porter JA, Slezak FA. Management of premature removal of the percutaneous gastrostomy. Am Surg 1990;56(11):733-6.
13. Marshall JB, Bodnarchuk G, Barthel JS. Early accidental dislodgement of PEG tubes. J Clin Gastroenterol 1994;18(3):210-2.
14. Fox VL, Abel SD, Malas S, Duggan C, Leichtner AM. Complications following percutaneous endoscopic gastrostomy and subsequent catheter replacement in children and young adults. Gastrointest Endosc 1997;45(1):64-71.
15. McQuaid KR, Little TE. Two fatal complications related to gastrostomy

“button” placement. Gastrointest Endosc 1992;38(5):601-3.

1. Wilson WC, Zenone EA, Spector H. Small intestinal perforation following replacement of a percutaneous endoscopic gastrostomy tube. Gastrointest Endosc 1990;36(1):62-3.
2. Gray RR, St Louis EL, Grosman H. Percutaneous gastrostomy and gastro-jejunostomy. Br J Radiol 1987;60(719):1067-70.
3. Patel PH, Willis M, Thomas E. Percutaneous endoscopic gastrostomy tube deterioration: how common is the problem? Am J Gastroenterol 1987;82(8):806.
4. Gowen GF. The management of complications of Foley feeding gastrostomies. Am Surg 1988;54(9):582-5.
5. Gauderer MW. Gastrostomy techniques and devices. Surg Clin North Am 1992;72(6):1285-98 [Review].
6. Gauderer MW, Picha GJ, Izant Jr RJ. The gastrostomy “button”-a simple, skin-level, nonrefluxing device for long-term enteral feedings. J Pediatr Surg 1984;19(6):803-5.
7. Ponsky JL. Techniques of percutaneous gastrostomy. New York: Igaku-Shoin Ltd; 1988. p. 63-78.
8. Foutch PG, Talbert GA, Gaines JA, Sanowski RA. The gastrostomy button: a prospective assessment of safety, success, and spectrum of use. Gastrointest Endosc 1989;35(1):41-4.
9. al Malki T, Langer JC, Thompson V, McQueen M, Lau GY, Issenman RM, et al. A prospective evaluation of the Button gastrostomy in children. Can J Surg 1991;34(3):247-50.
10. Kadakia SC, Parker A, Angueira C, Cassaday M. Failure to deflate the balloon of replacement gastrostomy tubes. Gastrointest Endosc 1993; 39(4):576-8.
11. Schrag SP, Sharma R, Jaik NP, Seamon MJ, Lukaszczyk JL, Martin ND, et al. Complications related to percutaneous endoscopic gastrostomy (PEG) tubes. A comprehensive clinical review. J Gastrointest Liver Dis 2007;16:407-18.
12. Gauderer MW, Stellato TA. Gastrostomy button: why complicate an office procedure? Gastrointest Endosc 1989;35(5):468-9.
13. Gauderer MW, Stellato TA, Wade DC. Complications related to gastrostomy button placement. Gastrointest Endosc 1993;39(3):467.
14. Faller NA. The Gastro-port: an alternative to the button. J Enterostomal Ther 1991;18(1):39-40.
15. Mellinger JD, Simon IB, Schlechter B, Lash RH, Ponsky JL. Tract formation following percutaneous endoscopic gastrostomy in an animal model. Surg Endosc 1991;5(4):189-91.
16. Burke DT, El Shami A, Heinle E, Pina BD. Comparison of gastrostomy tube replacement verification using air insufflation versus gastrograffin. Arch Phys Med Rehabil 2006;87:1530-3.