- Email: [email protected]
Comparison of direct percutaneous endoscopic jejunostomy and PEG with jejunal extension
Comparison of direct percutaneous endoscopic jejunostomy and PEG with jejunal extension Andy C. Fan, MD, Todd H. Baron, MD, FACP, Ashwin Rumalla, MD, Gavin C. Harewood, MD Rochester, Minnesota, and Scottsdale, Arizona
Background: Jejunostomy tubes can be placed endoscopically by means of percutaneous gastrostomy with jejunal extension (PEG-J) or by direct percutaneous jejunostomy. These 2 techniques were retrospectively compared in patients requiring long-term jejunal feeding. Method: An endoscopy database was used to identify all patients who underwent endoscopic jejunal feeding tube placement from January 1996 to May 2001. Patients with a history of upper GI surgery were excluded. There were 56 patients with a direct percutaneous jejunostomy and 49 with a percutaneous gastrostomy with jejunal extension. Patients in the direct percutaneous jejunostomy group received a 20F direct jejunostomy tube; a 20F PEG tube with a 9F jejunal extension was used in the percutaneous gastrostomy with jejunal extension group. Medical records for the period of 6 months after establishment of jejunal access were reviewed. Complications and need for further endoscopic intervention within this time frame were recorded. The duration of feeding tube patency (number of days from established jejunal access to first endoscopic reintervention) was compared for both groups. Results: Feeding tube patency was significantly longer in patients who had a direct percutaneous jejunostomy compared with those with a percutaneous gastrostomy with jejunal extension. Within the 6-month period, 5 patients with a direct percutaneous jejunostomy required endoscopic reintervention for tube dysfunction compared with 19 patients who had a percutaneous gastrostomy with jejunal extension (p < 0.0001). Conclusions: For patients who require long-term jejunal feeding, a direct percutaneous jejunostomy with a 20F tube provides more stable jejunal access compared with a percutaneous gastrostomy with jejunal extension with a 9F extension and has a lower associated rate of endoscopic reintervention. (Gastrointest Endosc 2002;56:890-4.)
For patients who cannot maintain adequate nutrition by eating, enteral feeding is the preferred method of nutrition and provides additional benefits in terms of GI mucosal integrity, immuno-secretory function, and avoidance of infectious complications.1-4 Jejunal feeding is preferred to gastric feeding in patients with gastroparesis, gastric outlet obstruction, and other functional or structural gastric defects but with normal small bowel function distal to the defect. In some studies, postpyloric feeding decreased the rate of aspiration of gastric contents in patients at high risk for this complication.5,6 The most commonly used endoscopic approach to the establishment of jejunal feeding is placement of a jejunal extension tube through a gastrostomy (PEG-J). However, the jejunal extension tube is prone to dysfunction and proximal migration Received January 18, 2002. For revision May 8, 2002. Accepted July 7, 2002. Current affiliations: Division of Gastroenterology and Hepatology, Mayo Medical Center, Rochester, Minnesota and Mayo Clinic, Scottsdale, Scottsdale, Arizona. Reprint Requests: Todd H. Baron, MD, FACP, 13400 East Shea Blvd., Scottsdale, AZ 85259. Copyright © 2002 by the American Society for Gastrointestinal Endoscopy 0016-5107/2002/$35.00 + 0 37/1/129607 doi:10.1067/mge.2002.129607 890
GASTROINTESTINAL ENDOSCOPY
into the stomach.7-9 A method of direct percutaneous endoscopic jejunostomy (D-PEJ) has been described.10,11 This involves direct insertion of a feeding tube into the jejunum with endoscopic assistance. This technique is technically feasible and provides a more stable access for jejunal feeding.10-12 However, comparative data for D-PEJ versus PEG-J feeding are scant. The present study retrospectively compared complications and the need for endoscopic reintervention to maintain jejunal access in patients with a D-PEJ versus patients with a PEG-J at a single institution. PATIENTS AND METHODS Our endoscopy database was used to identify all consecutive patients who underwent endoscopic jejunal feeding tube placement from January 1996 to May 2001. The study was approved by our institutional review board. The D-PEJ procedure was introduced at our institution in October 1998. Before this date, PEG-J was the standard endoscopic method for establishing jejunal feeding. With the demonstration of technical feasibility and tube patency in 26 of the initial 36 patients who underwent the DPEJ procedure before October 1998,13 D-PEJ became the preferred method of endoscopic jejunal feeding and use of the PEG-J declined. However, because D-PEJ placement is relatively new, the PEG-J procedure was still being VOLUME 56, NO. 6, 2002
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
used at our institution by endoscopists less experienced in the D-PEJ technique, thus providing a sizable PEG-J group for comparison purposes. Patients with a history of upper GI surgery with altered digestive tract anatomy were excluded from the study. For example, patients who had undergone an Ivor Lewis procedure may have been suitable candidates for D-PEJ but not PEG-J because of the intrathoracic location of the stomach. Thus, 29 patients who underwent D-PEJ and 5 who had a PEG-J were excluded on the basis of prior surgery. With these exclusions, a total of 56 patients with a D-PEJ and 49 with a PEG-J were available for analysis. Of the 49 patients in the PEG-J group, 16 had feeding tubes placed before October 1998 (when D-PEJ was not routinely performed). There were also 11 patients in the PEG-J group in whom D-PEJ was attempted unsuccessfully and who subsequently had a PEG-J placed. Demographic data and indications for tube placement are shown in Table 1. In the majority of patients the indication for jejunal tube placement was gastroparesis or a high risk for aspiration. Medical records were reviewed for complications and the need for endoscopic reintervention to maintain jejunal feeding access during the 6 months after jejunal access was established. Time of feeding tube patency was defined as the time from initial insertion of the jejunal tube to endoscopic reintervention to maintain tube patency. D-PEJ was performed as described by Shike et al.10,11 The technique is a modification of the standard technique for PEG insertion. Either a pediatric colonoscope (PCF, Olympus America Inc., Melville, N.Y.) or a push enteroscope (SIF-100, Olympus) was advanced into the jejunum to create an area of discrete illumination visible on the external abdominal wall. When transillumination was transient or scattered, external pressure was exerted to establish a landmark. The optimal site for tube insertion was further confirmed by applying external pressure with a finger to create an intrajejunal indentation. Fluoroscopy was not used routinely to determine whether the tip of the endoscope was distal to the ligament of Treitz. Glucagon was administered intravenously in incremental doses of 0.25 mg during the procedure to reduce gut motility. After standard preparation of the skin and injection of a local anesthetic, a sounding needle was used to perform the initial puncture. Later it was found that grasping the sounding needle with a snare and following the same track with the stylet sheath improved the success rate of the procedure. A stylet-sheath apparatus was inserted through the abdominal wall and into the jejunum. The stylet was then removed and a thread passed through the sheath. The thread was grasped with a snare and pulled out through the mouth. A standard 20F PEG tube (MIC PEG, Ballard Medical Products, Draper, Utah) was attached to the thread, pulled through the mouth and the upper GI tract, and positioned so that its intraluminal bumper rested against the jejunal wall. The endoscope was not routinely reinserted to confirm proper tube positioning. Most commonly the tube was withdrawn until internal traction was encountered at the appropriate external skin mark. If gastric decompression was required, a PEG tube (MIC PEG, VOLUME 56, NO. 6, 2002
A Fan, T Baron, A Rumalla, et al.
Table 1. Patient demographics and indications for jejunal feeding D-PEJ n = 56 Age Mean (y) Range (y) Men Indication Gastroparesis High-aspiration risk Gastroesophageal CA Pancreatitis Pancreatic CA Esophageal stricture Carcinoid HIV lymphoma Metastatic colon CA Cholangiocarcinoma Tongue CA
PEG-J n = 49
53 21-86 31
63 20-88 28
21 15 13 2 1 1 1 1 1 — —
16 23 1 4 1 2 — — — 1 1
CA, Carcinoma; HIV, human immunodeficiency virus.
Ballard) was placed during the same procedure by using standard pull technique. The PEG-J technique was performed after placement of a standard 20F PEG tube (MIC PEG, Ballard) by using the pull technique. Then, a 9F jejunal extension tube (Bard jejunal feeding/gastric decompression tube, C.R. Bard, Inc., Billerica, Mass.) was advanced through the PEG and distal to the ligament of Treitz under fluoroscopy with one of two methods. The first involved use of a pediatric colonoscope to grasp either the jejunal extension tube itself with rat-tooth forceps or the string attached to the end with a standard biopsy forceps, and then advancement of the endoscope. The tube or string was released after the endoscope had been withdrawn through the pylorus. Placement beyond the ligament of Treitz was then confirmed fluoroscopically. The second method was a modification of a technique previously described.14 The internal bumper of the PEG was advanced to the pylorus after the external bumper was withdrawn several centimeters; the jejunal extension tube was then advanced by using fluoroscopic guidance distal to the ligament of Treitz. The external bumper was then repositioned snugly against the anterior abdominal wall. Placement distal to the ligament of Treitz was confirmed fluoroscopically. Statistical methods Cumulative tube patency in each group was described by using the Kaplan-Meier method. Patients who died or were lost to follow-up before 6 months had elapsed were censored in the analysis. The tube patency curves were compared with the log-rank test. A p value of 0.05 was considered statistically significant. The Kaplan Meier method was used to analyze tube patency for several reasons: observations can be censored for patients who die or are lost to follow-up before the measured attribute (endoscopic intervention) is reached. Also, because patients entered the study at different times, the data could be progressively censored and comparisons of patency made at GASTROINTESTINAL ENDOSCOPY
891
A Fan, T Baron, A Rumalla, et al.
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
Figure 1. Feeding tube patency. Solid line, D-PEJ; dotted line, PEG-J. different points in time to maximize use of the available data. Using the log rank test is appropriate to assess for significance of difference between the 2 patency curves, which is essentially a comparison of 2 survival curves (survival of a functioning feeding tube), with a cut-off value of p < 0.05 to indicate statistical significance. A statistical software program (NCSS Statistical Software, Kaysville, Utah) was used for analysis of the data.
RESULTS D-PEJ insertion was successful in 56 (72.7%) of 77 patients who met inclusion criteria for the study. DPEJ was not possible in 17 patients because of failure to transilluminate a satisfactory location for insertion. No complication was noted in patients who underwent jejunal punctures and unsuccessful placement of the D-PEJ tube. In 4 patients strictures prevented advancement of the endoscope distal to the ligament of Treitz. Eleven patients in whom DPEJ was unsuccessful subsequently underwent PEG-J placement. PEG-J was performed unless a jejunostomy was specifically requested by the primary provider of care, in which case the patient underwent a surgical jejunostomy. Transillumination of a satisfactory location for D-PEJ placement appeared to be more difficult in obese patients. However, when mean body mass index for patients in whom D-PEJ was successful was compared with that for patients in whom the procedure was unsuccessful (respectively, 22.97 [5.41] vs. 24.67 [4.83]), there was no significant difference (p = 0.353). PEG-J insertion was successful in 49 (92.5%) of 53 patients. Failure to transilluminate or identify a satisfactory location for PEG insertion by exerting external pressure accounted for all 4 cases of unsuccessful PEG-J placement. Fifteen patients in the D-PEJ group and 15 in the PEG-J group were censored in the tube patency 892
GASTROINTESTINAL ENDOSCOPY
Kaplan-Meier analysis because they died or were lost to follow-up during the 6-month study period. Three patients in the D-PEJ group were censored because of elective removal of the jejunal tube before 6 months (2, resolution of gastroparesis; 1, resolution of aspiration risk). One patient was also censored in the D-PEJ group because of distal ileal volvulus and necrosis 2 days after insertion of DPEJ. At surgery, it was determined that the complication was not caused by the D-PEJ tube. During the 6-month period after tube placement, 5 (13.5%) of 37 patients in the D-PEJ group required endoscopic reintervention to maintain jejunal access versus 19 (55.9%) of 34 patients in the PEG-J group. Kaplan-Meier tube patency curves are shown in Figure 1. Feeding tube patency at 6 months was significantly longer in the D-PEJ group compared with the PEG-J group (p = 0.000085). The reasons for endoscopic reintervention in individual patients in the D-PEJ group and the number of days elapsed before reintervention were as follows: tube leakage (n = 1, 31 days), infection (n = 1, 121 days), pressure ulcer at the site of insertion (n = 1, 78 days), pain at the insertion site of unclear etiology (n = 1, 102 days), and replacement with a lower-profile tube (n = 1, 33 days). Of the 19 patients in the PEGJ group of 34 patients who required endoscopic reintervention, the main reason was proximal migration of the jejunal extension tube with subsequent coiling in the stomach (n = 8, mean 33 days). Other reasons for reintervention in the PEG-J group, number of instances, and mean time to reintervention were as follows: tube leakage (n = 3, mean 66 days), occluded tube (n = 3, mean 46 days), dislodged PEG bumper (n = 2, mean 37 days), infection (n = 1, 19 days), inadvertent patient removal (n = 1, 10 days), and unspecified tube dysfunction (n = 1, 14 days). Eleven patients in whom D-PEJ placement was unsuccessful subsequently underwent PEG-J: 3 of these patients died of underlying disease in less than 6 months and one was lost to follow-up. Of the remaining 7 patients, 4 required endoscopic reintervention. Surgical jejunostomy allows placement of the tube on the antimesenteric surface of the jejunum, which is not possible with endoscopic placement. This prevents volvulus and small bowel torsion. Volvulus and torsion were not noted in any patient with a D-PEJ during the 6-month evaluation period. No patient in either the D-PEJ or PEG-J group experienced other complications (e.g., malabsorption, diarrhea, short gut syndrome). DISCUSSION For long-term stable jejunal access, D-PEJ appears to be superior to PEG-J; the patency rate is VOLUME 56, NO. 6, 2002
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
greater and need for endoscopic reintervention is significantly less. With a high success rate for insertion (72.7%), and a lower rate of complications compared with PEG-J, D-PEJ feeding is an effective form of enteral nutrition. A low frequency (3%) of tube malfunction was also noted in a study by Shike et al.10 of patients with cancer who underwent DPEJ placement. The high rate of PEG-J tube dysfunction in the present study was similar to that in other studies: Wolfsen et al.9 encountered PEG-J tube dysfunction in 53% of 75 patients (mean followup 278 days). The D-PEJ method provides direct access to the jejunum. In contrast, the jejunal extension of the PEG-J method does not have a stable anchor within the small bowel and is thus prone to proximal migration. One possible explanation for proximal migration in the present study may be the high flexibility of the 9F jejunal tube; larger-caliber, stiffer jejunal tubes placed through 28F PEG tubes may be less prone to proximal migration. The ability to use larger-diameter tubes for jejunal access is one advantage of the D-PEJ method. The PEG-J method, conversely, requires use of smaller tubes in order for the jejunal extension to fit through the gastrostomy. In a study of PEG and PEG-J tubes, Simon and Fink15 determined that tube patency was dependent on use of tubes with large diameters; tubes of smaller sizes were significantly more prone to occlusion, kinking, and other complications. Standard technique for PEG-J placement at our institution involves use of a 20F PEG with a 9F jejunal extension. Because patency correlates with tube diameter, the use of larger-bore jejunostomy tubes might have improved the patency rate for the PEGJ group of patients. In a previous study at our institution, larger-bore jejunal tubes were placed transgastrically with the aid of an ultrathin endoscope.16 The mature gastrostomy tracts, because of previous PEG placement in these patients, allowed passage of this type of endoscope through the gastrostomy for transgastric placement of 12F jejunal PEJ tubes. If simultaneous gastric decompression is not required, placement of single lumen transgastric PEJ tubes with luminal diameters up to 24F might have significantly prolonged patency compared with that achieved with 9F or 12F tubes. Other maneuvers, such as use of a jejunal tube with a pigtail tip, or insertion 40 cm distal to the ligament of Treitz, may also decrease proximal migration rates in patients with a PEG-J. Prospective data with regard to these methods are lacking. Use of endoclips to secure the jejunal tip of the PEG-J directly on the jejunal wall might also have decreased migration rates. This method, howVOLUME 56, NO. 6, 2002
A Fan, T Baron, A Rumalla, et al.
ever, does not avert the problem of occlusion, which is common with smaller-caliber tubes.17,18 Although used successfully at our institution, the D-PEJ does have some limitations. Transillumination is sometimes unattainable. Intubation of the jejunum, even in patients with normal anatomy, can sometimes prove difficult and time-consuming. The potential for complications increases in relation to longer periods of conscious sedation. Structural digestive tract lesions make passage of the endoscope even more difficult and sometimes impossible. Fluoroscopy was not used routinely during D-PEJ placement. This was based on the fact that the ligament of Treitz demarcates the retroperitoneal portion of the duodenum from the intraperitoneal jejunum and the assumption that transillumination is only possible if the endoscope tip is within an intraperitoneal segment of the intestine. In the patients in whom D-PEJ placement was unsuccessful, fluoroscopy may have been helpful in establishing a location distal to the ligament suitable for transillumination. Other aids to endoscopic insertion of jejunal tubes have been described. Shetzline et al.19 successfully inserted D-PEJ tubes in a series of 7 patients by using an enteroscope and fluoroscopy. Sharma et al.20 describe US guidance for D-PEJ tube placement. Another limitation of the D-PEJ method is the need for a separate PEG if gastric decompression is required. In the patients in whom a D-PEJ was removed electively, no complications (e.g., enterocutaneous fistula) were noted during the 6-month retrospective follow-up. However, since this study was completed, 4 patients have been encountered at our institution who had a persistent enterocutaneous fistula develop after D-PEJ removal. A pediatric colonoscope was used for the great majority of D-PEJ procedures. A push enteroscope (with or without an overtube) was used for maneuvering through difficult strictures, or when jejunal intubation could not be achieved with the pediatric colonoscope. Since completion of this retrospective study, the variable stiffness pediatric colonoscope (PCF-160AL, Olympus) has been found to be optimal for D-PEJ placement. It facilitates intubation of the jejunum because the stiffness of the insertion tube can be adjusted to prevent looping. During the period of the present retrospective study, this instrument was not in use. D-PEJ is appropriate for patients who require a stable route for long-term enteral nutrition. A majority of patients in the present study had gastroparesis or a high risk of aspiration as the indication for jejunal feeding; the need for stable enteral nutrition is paramount in the later group. In a previous study, a high rate of satisfaction was found GASTROINTESTINAL ENDOSCOPY
893
A Fan, T Baron, A Rumalla, et al.
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
among patients who received a D-PEJ.13 In a survey of patients with a D-PEJ and their caregivers, 78% reported an overall satisfaction rating of 8 or higher on a 10-point scale (1, completely dissatisfied to 10, completely satisfied). Patients who required a separate PEG for gastric decompression in addition to the D-PEJ feeding tube were equally satisfied. Although the D-PEJ appears to provide a more stable route for enteral nutrition, the PEG-J does have advantages, the major one being ease of placement relative to D-PEJ insertion. A separate PEG for gastric decompression is not needed with a PEGJ, which combines feeding jejunal and gastric decompression ports in a single tube. Although the efficacy of jejunal feeding in preventing aspiration in high-risk patients is debatable,21 jejunal feeding was recommended for all patients in the present series by nutritional support teams. The present study has certain limitations. In a retrospective analysis such as this, the lack of patient randomization can result in selection bias. An attempt was made to minimize this potential by reviewing all consecutive patients referred for endoscopic jejunal access from January 1996 to May 2001, and then applying specific inclusion/exclusion criteria. With regard to follow-up, unless the patient volunteered information regarding endoscopic intervention elsewhere, it was assumed that this did not occur during the 6-month follow-up period. However, this does not eliminate the possibility of other interventions not recorded in our records. Although randomized prospective data comparing PEG-J with D-PEJ are lacking, the results of the present retrospective study suggest that D-PEJ is the preferred method for endoscopic jejunal tube placement when long-term jejunal feeding is required. DISCLOSURE None of the investigators have financial ties to any of the product manufacturers or distributors mentioned in this article. REFERENCES 1. Wicks C, Somasundaram S, Bjarnason I, Menzies IS, Routley D, Potter D, et al. Comparison of enteral feeding and total parenteral after liver transplantation. Lancet 1994; 344:837-40. 2. Alverdy J, Chi HS, Sheldon GF. The effect of parenteral nutrition on gastrointestinal immunity: the importance of enteral stimulation. Ann Surg 1985;202:681-4.
894
GASTROINTESTINAL ENDOSCOPY
3. Levine GM, Deren JJ, Steiger E, Zinno R. Role of oral intake in maintenance of gut mass and disaccharide activity. Gastroenterol 1974;67:975-82. 4. Lowry SF. The route of feeding influences injury responses. J Trauma 1990;30:S10-5. 5. Heyland DK, Drover JW, MacDonald S, Novak F, Lam M. Effect of postpyloric feeding on gastroesophageal regurgitation and pulmonary microaspiration: results of a randomized controlled trial. Crit Care Med 2001;29:1495-501. 6. Cech AC, Morris JB, Mullen JL, Crooks GW. Long-term enteral access in aspiration-prone patients. J Intensive Care Med 1995;10:179-86. 7. Kaplan DS, Murthy UK, Linscheer WG. Percutaneous endoscopic jejunostomy: long-term follow-up of 23 patients. Percutaneous endoscopic jejunostomy: long-term follow up of 23 patients. Gastrointest Endosc 1989;35:403-6. 8. DiSario JA, Foutch PG, Sanowski RA. Poor results with percutaneous endoscopic jejunostomy. Gastrointest Endosc 1990;36:257-60. 9. Wolfsen HC, Kozarek RA, Ball TJ, Patterson DJ, Botoman VA. Tube dysfunction following percutaneous endoscopic gastrostomy and jejunostomy. Gastrointest Endosc 1991;36:261-3. 10. Shike M, Latkany L, Gerdes H, Bloch AS. Direct percutaneous endoscopic jejunostomies for enteral feeding. Gastrointest Endosc 1996;44:536-40. 11. Shike M, Latkany L. Direct percutaneous endoscopic jejunostomy. Gastrointest Endosc Clin N Am 1998;8:569-80. 12. Mellert J, Naruhn MB, Grund KE, Becker HD. Direct endoscopic percutaneous jejunostomy (EPJ). Clinical results. Surg Endosc 1994;8:867-9. 13. Rumalla A, Baron TH. Results of direct percutaneous endoscopic jejunostomy, an alternative method for providing jejunal feeding. Mayo Clin Proc 2000;75:807-10. 14. Sibille A, Glorieux D, Fauville JP, Warzee P. An easier method for percutaneous endoscopic gastrojejunostomy tube placement. Gastrointest Endosc 1998;48:514-7. 15. Simon T, Fink AS. Recent experience with percutaneous endoscopic gastrostomy/jejunostomy (PEG/J) for enteral nutrition. Surg Endosc 2000;14:436-8. 16. Adler DG, Goustout CJ, Baron TH. Percutaneous transgastric placement of jejunal feeding tubes with an ultrathin endoscope. Gastroinest Endosc 2002;55:106-10. 17. Ginsberg GG, Lipman TO, Fleisher DE. Endoscopic clip assisted placement of enteral feeding tubes. Gastrointest Endosc 1994;40:220-2. 18. Faigel DO, Kadish SL, Ginsberg GG. The difficult to place feeding tube: successful endoscopic placement using a mucosal clip. JPEN J Parenter Enter Nutr 1996;20:306-8. 19. Shetzline MA, Suhocki PV, Workman MJ. Direct percutaneous endoscopic jejunostomy with small bowel enteroscopy and fluoroscopy. Gastrointest Endosc 2001;53:633-8. 20. Sharma VK, Close T, Bynoe R, Vasudeva R. Ultrasoundassisted direct percutaneous endoscopic jejunostomy (DPEJ) tube placement. Surg Endosc 2000;14:203-4. 21. Lazarus BA, Murphy JB, Culpepper L. Aspiration associated with long-term gastric versus jejunal feeding: a critical analysis of the literature. Arch Phys Med Rehabil 1990;71:46-53.
VOLUME 56, NO. 6, 2002
Background: Jejunostomy tubes can be placed endoscopically by means of percutaneous gastrostomy with jejunal extension (PEG-J) or by direct percutaneous jejunostomy. These 2 techniques were retrospectively compared in patients requiring long-term jejunal feeding. Method: An endoscopy database was used to identify all patients who underwent endoscopic jejunal feeding tube placement from January 1996 to May 2001. Patients with a history of upper GI surgery were excluded. There were 56 patients with a direct percutaneous jejunostomy and 49 with a percutaneous gastrostomy with jejunal extension. Patients in the direct percutaneous jejunostomy group received a 20F direct jejunostomy tube; a 20F PEG tube with a 9F jejunal extension was used in the percutaneous gastrostomy with jejunal extension group. Medical records for the period of 6 months after establishment of jejunal access were reviewed. Complications and need for further endoscopic intervention within this time frame were recorded. The duration of feeding tube patency (number of days from established jejunal access to first endoscopic reintervention) was compared for both groups. Results: Feeding tube patency was significantly longer in patients who had a direct percutaneous jejunostomy compared with those with a percutaneous gastrostomy with jejunal extension. Within the 6-month period, 5 patients with a direct percutaneous jejunostomy required endoscopic reintervention for tube dysfunction compared with 19 patients who had a percutaneous gastrostomy with jejunal extension (p < 0.0001). Conclusions: For patients who require long-term jejunal feeding, a direct percutaneous jejunostomy with a 20F tube provides more stable jejunal access compared with a percutaneous gastrostomy with jejunal extension with a 9F extension and has a lower associated rate of endoscopic reintervention. (Gastrointest Endosc 2002;56:890-4.)
For patients who cannot maintain adequate nutrition by eating, enteral feeding is the preferred method of nutrition and provides additional benefits in terms of GI mucosal integrity, immuno-secretory function, and avoidance of infectious complications.1-4 Jejunal feeding is preferred to gastric feeding in patients with gastroparesis, gastric outlet obstruction, and other functional or structural gastric defects but with normal small bowel function distal to the defect. In some studies, postpyloric feeding decreased the rate of aspiration of gastric contents in patients at high risk for this complication.5,6 The most commonly used endoscopic approach to the establishment of jejunal feeding is placement of a jejunal extension tube through a gastrostomy (PEG-J). However, the jejunal extension tube is prone to dysfunction and proximal migration Received January 18, 2002. For revision May 8, 2002. Accepted July 7, 2002. Current affiliations: Division of Gastroenterology and Hepatology, Mayo Medical Center, Rochester, Minnesota and Mayo Clinic, Scottsdale, Scottsdale, Arizona. Reprint Requests: Todd H. Baron, MD, FACP, 13400 East Shea Blvd., Scottsdale, AZ 85259. Copyright © 2002 by the American Society for Gastrointestinal Endoscopy 0016-5107/2002/$35.00 + 0 37/1/129607 doi:10.1067/mge.2002.129607 890
GASTROINTESTINAL ENDOSCOPY
into the stomach.7-9 A method of direct percutaneous endoscopic jejunostomy (D-PEJ) has been described.10,11 This involves direct insertion of a feeding tube into the jejunum with endoscopic assistance. This technique is technically feasible and provides a more stable access for jejunal feeding.10-12 However, comparative data for D-PEJ versus PEG-J feeding are scant. The present study retrospectively compared complications and the need for endoscopic reintervention to maintain jejunal access in patients with a D-PEJ versus patients with a PEG-J at a single institution. PATIENTS AND METHODS Our endoscopy database was used to identify all consecutive patients who underwent endoscopic jejunal feeding tube placement from January 1996 to May 2001. The study was approved by our institutional review board. The D-PEJ procedure was introduced at our institution in October 1998. Before this date, PEG-J was the standard endoscopic method for establishing jejunal feeding. With the demonstration of technical feasibility and tube patency in 26 of the initial 36 patients who underwent the DPEJ procedure before October 1998,13 D-PEJ became the preferred method of endoscopic jejunal feeding and use of the PEG-J declined. However, because D-PEJ placement is relatively new, the PEG-J procedure was still being VOLUME 56, NO. 6, 2002
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
used at our institution by endoscopists less experienced in the D-PEJ technique, thus providing a sizable PEG-J group for comparison purposes. Patients with a history of upper GI surgery with altered digestive tract anatomy were excluded from the study. For example, patients who had undergone an Ivor Lewis procedure may have been suitable candidates for D-PEJ but not PEG-J because of the intrathoracic location of the stomach. Thus, 29 patients who underwent D-PEJ and 5 who had a PEG-J were excluded on the basis of prior surgery. With these exclusions, a total of 56 patients with a D-PEJ and 49 with a PEG-J were available for analysis. Of the 49 patients in the PEG-J group, 16 had feeding tubes placed before October 1998 (when D-PEJ was not routinely performed). There were also 11 patients in the PEG-J group in whom D-PEJ was attempted unsuccessfully and who subsequently had a PEG-J placed. Demographic data and indications for tube placement are shown in Table 1. In the majority of patients the indication for jejunal tube placement was gastroparesis or a high risk for aspiration. Medical records were reviewed for complications and the need for endoscopic reintervention to maintain jejunal feeding access during the 6 months after jejunal access was established. Time of feeding tube patency was defined as the time from initial insertion of the jejunal tube to endoscopic reintervention to maintain tube patency. D-PEJ was performed as described by Shike et al.10,11 The technique is a modification of the standard technique for PEG insertion. Either a pediatric colonoscope (PCF, Olympus America Inc., Melville, N.Y.) or a push enteroscope (SIF-100, Olympus) was advanced into the jejunum to create an area of discrete illumination visible on the external abdominal wall. When transillumination was transient or scattered, external pressure was exerted to establish a landmark. The optimal site for tube insertion was further confirmed by applying external pressure with a finger to create an intrajejunal indentation. Fluoroscopy was not used routinely to determine whether the tip of the endoscope was distal to the ligament of Treitz. Glucagon was administered intravenously in incremental doses of 0.25 mg during the procedure to reduce gut motility. After standard preparation of the skin and injection of a local anesthetic, a sounding needle was used to perform the initial puncture. Later it was found that grasping the sounding needle with a snare and following the same track with the stylet sheath improved the success rate of the procedure. A stylet-sheath apparatus was inserted through the abdominal wall and into the jejunum. The stylet was then removed and a thread passed through the sheath. The thread was grasped with a snare and pulled out through the mouth. A standard 20F PEG tube (MIC PEG, Ballard Medical Products, Draper, Utah) was attached to the thread, pulled through the mouth and the upper GI tract, and positioned so that its intraluminal bumper rested against the jejunal wall. The endoscope was not routinely reinserted to confirm proper tube positioning. Most commonly the tube was withdrawn until internal traction was encountered at the appropriate external skin mark. If gastric decompression was required, a PEG tube (MIC PEG, VOLUME 56, NO. 6, 2002
A Fan, T Baron, A Rumalla, et al.
Table 1. Patient demographics and indications for jejunal feeding D-PEJ n = 56 Age Mean (y) Range (y) Men Indication Gastroparesis High-aspiration risk Gastroesophageal CA Pancreatitis Pancreatic CA Esophageal stricture Carcinoid HIV lymphoma Metastatic colon CA Cholangiocarcinoma Tongue CA
PEG-J n = 49
53 21-86 31
63 20-88 28
21 15 13 2 1 1 1 1 1 — —
16 23 1 4 1 2 — — — 1 1
CA, Carcinoma; HIV, human immunodeficiency virus.
Ballard) was placed during the same procedure by using standard pull technique. The PEG-J technique was performed after placement of a standard 20F PEG tube (MIC PEG, Ballard) by using the pull technique. Then, a 9F jejunal extension tube (Bard jejunal feeding/gastric decompression tube, C.R. Bard, Inc., Billerica, Mass.) was advanced through the PEG and distal to the ligament of Treitz under fluoroscopy with one of two methods. The first involved use of a pediatric colonoscope to grasp either the jejunal extension tube itself with rat-tooth forceps or the string attached to the end with a standard biopsy forceps, and then advancement of the endoscope. The tube or string was released after the endoscope had been withdrawn through the pylorus. Placement beyond the ligament of Treitz was then confirmed fluoroscopically. The second method was a modification of a technique previously described.14 The internal bumper of the PEG was advanced to the pylorus after the external bumper was withdrawn several centimeters; the jejunal extension tube was then advanced by using fluoroscopic guidance distal to the ligament of Treitz. The external bumper was then repositioned snugly against the anterior abdominal wall. Placement distal to the ligament of Treitz was confirmed fluoroscopically. Statistical methods Cumulative tube patency in each group was described by using the Kaplan-Meier method. Patients who died or were lost to follow-up before 6 months had elapsed were censored in the analysis. The tube patency curves were compared with the log-rank test. A p value of 0.05 was considered statistically significant. The Kaplan Meier method was used to analyze tube patency for several reasons: observations can be censored for patients who die or are lost to follow-up before the measured attribute (endoscopic intervention) is reached. Also, because patients entered the study at different times, the data could be progressively censored and comparisons of patency made at GASTROINTESTINAL ENDOSCOPY
891
A Fan, T Baron, A Rumalla, et al.
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
Figure 1. Feeding tube patency. Solid line, D-PEJ; dotted line, PEG-J. different points in time to maximize use of the available data. Using the log rank test is appropriate to assess for significance of difference between the 2 patency curves, which is essentially a comparison of 2 survival curves (survival of a functioning feeding tube), with a cut-off value of p < 0.05 to indicate statistical significance. A statistical software program (NCSS Statistical Software, Kaysville, Utah) was used for analysis of the data.
RESULTS D-PEJ insertion was successful in 56 (72.7%) of 77 patients who met inclusion criteria for the study. DPEJ was not possible in 17 patients because of failure to transilluminate a satisfactory location for insertion. No complication was noted in patients who underwent jejunal punctures and unsuccessful placement of the D-PEJ tube. In 4 patients strictures prevented advancement of the endoscope distal to the ligament of Treitz. Eleven patients in whom DPEJ was unsuccessful subsequently underwent PEG-J placement. PEG-J was performed unless a jejunostomy was specifically requested by the primary provider of care, in which case the patient underwent a surgical jejunostomy. Transillumination of a satisfactory location for D-PEJ placement appeared to be more difficult in obese patients. However, when mean body mass index for patients in whom D-PEJ was successful was compared with that for patients in whom the procedure was unsuccessful (respectively, 22.97 [5.41] vs. 24.67 [4.83]), there was no significant difference (p = 0.353). PEG-J insertion was successful in 49 (92.5%) of 53 patients. Failure to transilluminate or identify a satisfactory location for PEG insertion by exerting external pressure accounted for all 4 cases of unsuccessful PEG-J placement. Fifteen patients in the D-PEJ group and 15 in the PEG-J group were censored in the tube patency 892
GASTROINTESTINAL ENDOSCOPY
Kaplan-Meier analysis because they died or were lost to follow-up during the 6-month study period. Three patients in the D-PEJ group were censored because of elective removal of the jejunal tube before 6 months (2, resolution of gastroparesis; 1, resolution of aspiration risk). One patient was also censored in the D-PEJ group because of distal ileal volvulus and necrosis 2 days after insertion of DPEJ. At surgery, it was determined that the complication was not caused by the D-PEJ tube. During the 6-month period after tube placement, 5 (13.5%) of 37 patients in the D-PEJ group required endoscopic reintervention to maintain jejunal access versus 19 (55.9%) of 34 patients in the PEG-J group. Kaplan-Meier tube patency curves are shown in Figure 1. Feeding tube patency at 6 months was significantly longer in the D-PEJ group compared with the PEG-J group (p = 0.000085). The reasons for endoscopic reintervention in individual patients in the D-PEJ group and the number of days elapsed before reintervention were as follows: tube leakage (n = 1, 31 days), infection (n = 1, 121 days), pressure ulcer at the site of insertion (n = 1, 78 days), pain at the insertion site of unclear etiology (n = 1, 102 days), and replacement with a lower-profile tube (n = 1, 33 days). Of the 19 patients in the PEGJ group of 34 patients who required endoscopic reintervention, the main reason was proximal migration of the jejunal extension tube with subsequent coiling in the stomach (n = 8, mean 33 days). Other reasons for reintervention in the PEG-J group, number of instances, and mean time to reintervention were as follows: tube leakage (n = 3, mean 66 days), occluded tube (n = 3, mean 46 days), dislodged PEG bumper (n = 2, mean 37 days), infection (n = 1, 19 days), inadvertent patient removal (n = 1, 10 days), and unspecified tube dysfunction (n = 1, 14 days). Eleven patients in whom D-PEJ placement was unsuccessful subsequently underwent PEG-J: 3 of these patients died of underlying disease in less than 6 months and one was lost to follow-up. Of the remaining 7 patients, 4 required endoscopic reintervention. Surgical jejunostomy allows placement of the tube on the antimesenteric surface of the jejunum, which is not possible with endoscopic placement. This prevents volvulus and small bowel torsion. Volvulus and torsion were not noted in any patient with a D-PEJ during the 6-month evaluation period. No patient in either the D-PEJ or PEG-J group experienced other complications (e.g., malabsorption, diarrhea, short gut syndrome). DISCUSSION For long-term stable jejunal access, D-PEJ appears to be superior to PEG-J; the patency rate is VOLUME 56, NO. 6, 2002
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
greater and need for endoscopic reintervention is significantly less. With a high success rate for insertion (72.7%), and a lower rate of complications compared with PEG-J, D-PEJ feeding is an effective form of enteral nutrition. A low frequency (3%) of tube malfunction was also noted in a study by Shike et al.10 of patients with cancer who underwent DPEJ placement. The high rate of PEG-J tube dysfunction in the present study was similar to that in other studies: Wolfsen et al.9 encountered PEG-J tube dysfunction in 53% of 75 patients (mean followup 278 days). The D-PEJ method provides direct access to the jejunum. In contrast, the jejunal extension of the PEG-J method does not have a stable anchor within the small bowel and is thus prone to proximal migration. One possible explanation for proximal migration in the present study may be the high flexibility of the 9F jejunal tube; larger-caliber, stiffer jejunal tubes placed through 28F PEG tubes may be less prone to proximal migration. The ability to use larger-diameter tubes for jejunal access is one advantage of the D-PEJ method. The PEG-J method, conversely, requires use of smaller tubes in order for the jejunal extension to fit through the gastrostomy. In a study of PEG and PEG-J tubes, Simon and Fink15 determined that tube patency was dependent on use of tubes with large diameters; tubes of smaller sizes were significantly more prone to occlusion, kinking, and other complications. Standard technique for PEG-J placement at our institution involves use of a 20F PEG with a 9F jejunal extension. Because patency correlates with tube diameter, the use of larger-bore jejunostomy tubes might have improved the patency rate for the PEGJ group of patients. In a previous study at our institution, larger-bore jejunal tubes were placed transgastrically with the aid of an ultrathin endoscope.16 The mature gastrostomy tracts, because of previous PEG placement in these patients, allowed passage of this type of endoscope through the gastrostomy for transgastric placement of 12F jejunal PEJ tubes. If simultaneous gastric decompression is not required, placement of single lumen transgastric PEJ tubes with luminal diameters up to 24F might have significantly prolonged patency compared with that achieved with 9F or 12F tubes. Other maneuvers, such as use of a jejunal tube with a pigtail tip, or insertion 40 cm distal to the ligament of Treitz, may also decrease proximal migration rates in patients with a PEG-J. Prospective data with regard to these methods are lacking. Use of endoclips to secure the jejunal tip of the PEG-J directly on the jejunal wall might also have decreased migration rates. This method, howVOLUME 56, NO. 6, 2002
A Fan, T Baron, A Rumalla, et al.
ever, does not avert the problem of occlusion, which is common with smaller-caliber tubes.17,18 Although used successfully at our institution, the D-PEJ does have some limitations. Transillumination is sometimes unattainable. Intubation of the jejunum, even in patients with normal anatomy, can sometimes prove difficult and time-consuming. The potential for complications increases in relation to longer periods of conscious sedation. Structural digestive tract lesions make passage of the endoscope even more difficult and sometimes impossible. Fluoroscopy was not used routinely during D-PEJ placement. This was based on the fact that the ligament of Treitz demarcates the retroperitoneal portion of the duodenum from the intraperitoneal jejunum and the assumption that transillumination is only possible if the endoscope tip is within an intraperitoneal segment of the intestine. In the patients in whom D-PEJ placement was unsuccessful, fluoroscopy may have been helpful in establishing a location distal to the ligament suitable for transillumination. Other aids to endoscopic insertion of jejunal tubes have been described. Shetzline et al.19 successfully inserted D-PEJ tubes in a series of 7 patients by using an enteroscope and fluoroscopy. Sharma et al.20 describe US guidance for D-PEJ tube placement. Another limitation of the D-PEJ method is the need for a separate PEG if gastric decompression is required. In the patients in whom a D-PEJ was removed electively, no complications (e.g., enterocutaneous fistula) were noted during the 6-month retrospective follow-up. However, since this study was completed, 4 patients have been encountered at our institution who had a persistent enterocutaneous fistula develop after D-PEJ removal. A pediatric colonoscope was used for the great majority of D-PEJ procedures. A push enteroscope (with or without an overtube) was used for maneuvering through difficult strictures, or when jejunal intubation could not be achieved with the pediatric colonoscope. Since completion of this retrospective study, the variable stiffness pediatric colonoscope (PCF-160AL, Olympus) has been found to be optimal for D-PEJ placement. It facilitates intubation of the jejunum because the stiffness of the insertion tube can be adjusted to prevent looping. During the period of the present retrospective study, this instrument was not in use. D-PEJ is appropriate for patients who require a stable route for long-term enteral nutrition. A majority of patients in the present study had gastroparesis or a high risk of aspiration as the indication for jejunal feeding; the need for stable enteral nutrition is paramount in the later group. In a previous study, a high rate of satisfaction was found GASTROINTESTINAL ENDOSCOPY
893
A Fan, T Baron, A Rumalla, et al.
Direct percutaneous endoscopic jejunostomy vs. PEG with jejunal extension
among patients who received a D-PEJ.13 In a survey of patients with a D-PEJ and their caregivers, 78% reported an overall satisfaction rating of 8 or higher on a 10-point scale (1, completely dissatisfied to 10, completely satisfied). Patients who required a separate PEG for gastric decompression in addition to the D-PEJ feeding tube were equally satisfied. Although the D-PEJ appears to provide a more stable route for enteral nutrition, the PEG-J does have advantages, the major one being ease of placement relative to D-PEJ insertion. A separate PEG for gastric decompression is not needed with a PEGJ, which combines feeding jejunal and gastric decompression ports in a single tube. Although the efficacy of jejunal feeding in preventing aspiration in high-risk patients is debatable,21 jejunal feeding was recommended for all patients in the present series by nutritional support teams. The present study has certain limitations. In a retrospective analysis such as this, the lack of patient randomization can result in selection bias. An attempt was made to minimize this potential by reviewing all consecutive patients referred for endoscopic jejunal access from January 1996 to May 2001, and then applying specific inclusion/exclusion criteria. With regard to follow-up, unless the patient volunteered information regarding endoscopic intervention elsewhere, it was assumed that this did not occur during the 6-month follow-up period. However, this does not eliminate the possibility of other interventions not recorded in our records. Although randomized prospective data comparing PEG-J with D-PEJ are lacking, the results of the present retrospective study suggest that D-PEJ is the preferred method for endoscopic jejunal tube placement when long-term jejunal feeding is required. DISCLOSURE None of the investigators have financial ties to any of the product manufacturers or distributors mentioned in this article. REFERENCES 1. Wicks C, Somasundaram S, Bjarnason I, Menzies IS, Routley D, Potter D, et al. Comparison of enteral feeding and total parenteral after liver transplantation. Lancet 1994; 344:837-40. 2. Alverdy J, Chi HS, Sheldon GF. The effect of parenteral nutrition on gastrointestinal immunity: the importance of enteral stimulation. Ann Surg 1985;202:681-4.
894
GASTROINTESTINAL ENDOSCOPY
3. Levine GM, Deren JJ, Steiger E, Zinno R. Role of oral intake in maintenance of gut mass and disaccharide activity. Gastroenterol 1974;67:975-82. 4. Lowry SF. The route of feeding influences injury responses. J Trauma 1990;30:S10-5. 5. Heyland DK, Drover JW, MacDonald S, Novak F, Lam M. Effect of postpyloric feeding on gastroesophageal regurgitation and pulmonary microaspiration: results of a randomized controlled trial. Crit Care Med 2001;29:1495-501. 6. Cech AC, Morris JB, Mullen JL, Crooks GW. Long-term enteral access in aspiration-prone patients. J Intensive Care Med 1995;10:179-86. 7. Kaplan DS, Murthy UK, Linscheer WG. Percutaneous endoscopic jejunostomy: long-term follow-up of 23 patients. Percutaneous endoscopic jejunostomy: long-term follow up of 23 patients. Gastrointest Endosc 1989;35:403-6. 8. DiSario JA, Foutch PG, Sanowski RA. Poor results with percutaneous endoscopic jejunostomy. Gastrointest Endosc 1990;36:257-60. 9. Wolfsen HC, Kozarek RA, Ball TJ, Patterson DJ, Botoman VA. Tube dysfunction following percutaneous endoscopic gastrostomy and jejunostomy. Gastrointest Endosc 1991;36:261-3. 10. Shike M, Latkany L, Gerdes H, Bloch AS. Direct percutaneous endoscopic jejunostomies for enteral feeding. Gastrointest Endosc 1996;44:536-40. 11. Shike M, Latkany L. Direct percutaneous endoscopic jejunostomy. Gastrointest Endosc Clin N Am 1998;8:569-80. 12. Mellert J, Naruhn MB, Grund KE, Becker HD. Direct endoscopic percutaneous jejunostomy (EPJ). Clinical results. Surg Endosc 1994;8:867-9. 13. Rumalla A, Baron TH. Results of direct percutaneous endoscopic jejunostomy, an alternative method for providing jejunal feeding. Mayo Clin Proc 2000;75:807-10. 14. Sibille A, Glorieux D, Fauville JP, Warzee P. An easier method for percutaneous endoscopic gastrojejunostomy tube placement. Gastrointest Endosc 1998;48:514-7. 15. Simon T, Fink AS. Recent experience with percutaneous endoscopic gastrostomy/jejunostomy (PEG/J) for enteral nutrition. Surg Endosc 2000;14:436-8. 16. Adler DG, Goustout CJ, Baron TH. Percutaneous transgastric placement of jejunal feeding tubes with an ultrathin endoscope. Gastroinest Endosc 2002;55:106-10. 17. Ginsberg GG, Lipman TO, Fleisher DE. Endoscopic clip assisted placement of enteral feeding tubes. Gastrointest Endosc 1994;40:220-2. 18. Faigel DO, Kadish SL, Ginsberg GG. The difficult to place feeding tube: successful endoscopic placement using a mucosal clip. JPEN J Parenter Enter Nutr 1996;20:306-8. 19. Shetzline MA, Suhocki PV, Workman MJ. Direct percutaneous endoscopic jejunostomy with small bowel enteroscopy and fluoroscopy. Gastrointest Endosc 2001;53:633-8. 20. Sharma VK, Close T, Bynoe R, Vasudeva R. Ultrasoundassisted direct percutaneous endoscopic jejunostomy (DPEJ) tube placement. Surg Endosc 2000;14:203-4. 21. Lazarus BA, Murphy JB, Culpepper L. Aspiration associated with long-term gastric versus jejunal feeding: a critical analysis of the literature. Arch Phys Med Rehabil 1990;71:46-53.
VOLUME 56, NO. 6, 2002