Percutaneous endoscopic gastrostomy: Avoiding complications PERRY M. SANTOS, MD, MS, and JAMES MCDONALD, DO, Springfield, Illinois
This study evaluates our complications arising directly or indirectly from placement or management of percutaneous endoscopic gastrostomy (PEG) tubes and provides recommendations for avoidance of complications. Seventy-one patients received PEG tube placement by otolaryngologists between January 1991 and May 1997. Records were reviewed for diagnoses, combined procedures, and complications. Addressing potential causes of complications prompted modification of our technique of PEG tube placement and management. Twenty-three patients received PEG for dysphagia/aspiration unrelated to neoplasia, 11 received PEG with staging endoscopy, 11 received PEG after treatment for head and neck neoplasm, and 26 received PEG at the time of primary resection. Major complications included retained PEG hub and delayed colon abscess ultimately resulting in death. Minor complications included skin abscesses, cellulitis, and early and late vasovagal response with PEG tube removal. An airway emergency, on attempted oral airway intubation, resulted in an aborted PEG attempt and constituted another complication outside the 4 groups stated above. The major complication was not found within a literature review. We have modified our management for avoidance of this complication. We believe the causes of the minor complications have been identified, and with additional modifications in our technique, we have not had any similar complications recently. The recommended techniques are discussed in detail. (Otolaryngol Head Neck Surg 1999;120:195-9.)
endoscopic gastrostomy (PEG) has become an important mainstay in enteral alimentation From the Division of Otolaryngology, Southern Illinois University School of Medicine. Presented at the Annual Meeting of the American Academy of Otolaryngology–Head and Neck Surgery, San Francisco, CA, September 7-10, 1997. Reprint requests: Perry M. Santos, MD, MS, Oklahoma Otolaryngology Associates, 4200 W Memorial Rd, Suite 606, Oklahoma City, OK 73120. Copyright © 1999 by the American Academy of Otolaryngology– Head and Neck Surgery Foundation, Inc. 0194-5998/99/$8.00 + 0 23/1/89437
for patients unable to tolerate oral intake since its first description by Gauderer and coworkers.1,2 PEG eliminates the need for chronic nasogastric tube placement with the attendant discomfort and morbidity of sinusitis, aspiration, reflux, and pharyngolaryngeal inflammation.3-6 Head and neck surgeons are reporting and performing this procedure with greater frequency in the treatment of patients with dysphagia or aspiration from neoplastic, neurologic, or traumatic causes.4,6-10 We reported no complications in the placement or management of PEG tubes in a small series of 29 patients who received PEGs in conjunction with other procedures.9 Since that time we have had some complications in the management of PEGs and have reevaluated our management technique in an effort to decrease the risk of such complications. METHODS Patients were obtained from our initial study reported in 19959 and additional patients were accrued prospectively until May 1997. In addition to a longer follow-up period for the initial 29 patients previously reported, 42 more were accrued. Patients receiving PEGs were categorized into 4 patient groups: neurologic patients with dysphagia/aspiration unrelated to neoplasia, patients undergoing PEG at the time of staging endoscopy for neoplasia, patients requiring supplemental alimentation through PEG after treatment for head and neck neoplasm, and patients receiving PEG at the time of primary resection. Contraindications for PEG included previous gastric surgery, morbid obesity, peritonitis, massive ascites, peritoneal dialysis, sepsis, and inability to transilluminate the abdominal wall. PEG tube placement is performed with the patient under general anesthesia with techniques similar to those described by Gauderer and coworkers1,2 and within our initial series.9 The patient is placed in reverse Trendelenburg position to allow the abdominal wall and contents to migrate toward the pelvis and allow greater exposure of the povidone-iodine prepared left upper abdominal quadrant. Examination of the esophagus and stomach and insufflation of the stomach are performed with a flexible video gastroscope (Olympus, Lombard, IL). Light from the endoscope is visualized externally as it is approximated to the gastric mucosa. External digital palpation of the abdominal wall over the stomach, several centimeters under the left costal margin, is confirmed through the endoscope. The palpation site is marked, and a 16-gauge 195
SANTOS and MCDONALD
Fig 1. Axial plane abdominal CT scan demonstrates the retained inner hub of the PEG tube in the sigmoid colon.
angiocatheter trocar is placed through the abdominal wall into the stomach under endoscopic visualization. If the needle is not visualized, it is removed, and another new needle is used. Once the needle is in the stomach, a wire loop is passed though the catheter and grasped with the gastroscope snare and removed through the mouth with the gastroscope. The gastrostomy tube (22F Nuport PEG System or 15F PediPEG, Sandoz Nutrition Corp, Minneapolis, MN; 20F Bard Ponsky Pull PEG, Billerica, MA; or 20F Flexiflo Inverta-PEG, Ross/Abbott Laboratories, Columbus, OH) is attached to the wire, and the entire assembly is pulled into the stomach and through the trocar site after a 1 cm dermal incision is made to allow the tube to pass through the dermis. The outer flange is cinched at a distance that allows 1 cm of extra play between the inner hub and outer flange. There should be no tenting or compression of the gastric mucosa or skin. The esophageal mucosa is examined to rule out any mucosal tears on exit of the gastroscope, and the stomach is decompressed. For those patients requiring removal of the tube, the external tube portion is cut flush with the skin, the inner hub is pushed into the stomach with a clean instrument, and the hub is allowed to pass through the gastrointestinal tract.11,12 RESULTS
Seventy-one patients received PEG tubes and were placed into 1 of 4 patient groups: (1) 23 patients with aspiration/dysphagia primarily caused by central or peripheral neurologic impairment and unrelated to neoplasia; (2) 11 patients evaluated for upper aerodigestive tract cancer by endoscopy and receiving PEG at the time of staging endoscopy; (3) 11 patients requiring adjunctive alimentation after surgical and/or radiation treatment for a head and neck neoplasm; and (4) 26 patients
Otolaryngology– Head and Neck Surgery February 1999
with squamous cell carcinoma (SCCA) who received PEGs at the time of the primary resection of the neoplasm. The average time the PEG tube was in place for all 4 groups was 4.7 months if removed and 8.4 months if left in place (based on last contact because patient died or was lost to follow-up). PEG tube placement failed in an additional 4 patients. Two cases were reported previously.9 PEG failed in the other 2 patients because of (1) exceptionally well-developed rectus abdominal musculature preventing passage of the angiocatheter through the entire abdominal wall, and (2) an obstructing hypopharyngeal lesion that began bleeding on contact with the anesthesiologist’s laryngoscope, necessitating immediate airway control by an emergent cricothyroidotomy after oral intubation attempts failed. The major complication was in a 55-year-old man with severe chronic obstructive pulmonary disease (FEV1 of 650 mL), cirrhosis, and malnutrition who had an uncomplicated PEG tube placement at the time of his staging endoscopy for a poorly differentiated T3N1 laryngeal SCCA. He subsequently underwent uncomplicated total laryngectomy and neck dissection. He received a full course of radiation therapy of 6000 cGy in 30 fractions. Between 4 and 7 months after surgery the PEG tube was removed by cutting across it at the skin level, pushing the inner hub into the stomach, and allowing it to pass through the gastrointestinal system. The patient had been tolerating and continued to tolerate a regular diet with normal bowel habits. However, more than 1 year later, acute incapacitating abdominal pain developed. Evaluation revealed a microabscess and foreign body in the sigmoid colon (Fig 1). Sigmoid resection with end-to-end anastomosis repair was performed but later broke down. Consequently, a colostomy was performed. After surgery, liver failure with massive ascites and wound separation developed. Later hepatic encephalopathy and seizures developed, and the patient died. Minor complications included 2 episodes of skin abscess within 4 to 7 days of the PEG. One of the patients was in the neurologic group with dysphagia, and the other patient had a coincident base-of-tongue biopsy for a recurrence with progressive dysphagia. The neurologic patient was already receiving antibiotics. Both responded to local incision, drainage, dressing changes, and intravenous antibiotics without sequelae. One and possibly both patients required more than 1 trocar needle pass into the stomach, and both patients were found to have the outer flange applied too tightly to the skin. It was unclear whether the tight flange was a result of placement or abdominal wall swelling from the infection. The second group of minor complications included 3 cases of cellulitis: 2 patients had cellulitis within days of PEG tube placement, and 1 had delayed
Otolaryngology– Head and Neck Surgery Volume 120 Number 2
SANTOS and MCDONALD
recurrent episodes. All episodes responded to local care and oral antibiotics. The patient with recurrent episodes of cellulitis eventually had the tube removed. The cause was not determined. Two patients had leakage around the tube and responded to a tube change. One patient had a vasovagal episode during removal of a Bard Ponsky Pull PEG tube and hub with the “traction technique.” One of the patients who had cellulitis at the PEG site within a week of placement requested removal 8 months after PEG tube placement and treatment of a T2N3 SCCA tonsil lesion. Our standard method of cutting the tube at the skin level and pushing the hub into the stomach was not successful. Resistance to pushing the hub into the stomach was noted. Therefore the patient was taken to the gastrointestinal suite, and endoscopic removal was performed under intravenous sedation. The inner hub flanges were embedded with overlying mucosa, leaving only the central tip of the hub exposed for grasping and removal with the snare (Fig 2). He suffered no ill effects. On further questioning, the patient revealed he had not used the PEG for several months and had failed to irrigate the PEG with water for the same time frame. He denied any problems with it during this non use period. There were no cases of necrotizing fasciitis or metastatic seeding to the gastrostomy tract within our series. DISCUSSION
The major complication reported herein of an asymptomatic patient with a retained PEG hub in the sigmoid colon for more than 1 year is unreported in the literature. Nonendoscopic removal of PEG tubes by cutting the tube at the skin surface, pushing the hub into the stomach, and allowing intestinal elimination has been shown to be a safe, painless, and economical means of PEG tube removal.11,12 Korula and Harma11 documented passage with abdominal radiographs of PEG hubs a mean of 10 days after cutting the tube. They noted that of 64 patients only 1, who had C-4 quadriplegia, required endoscopic removal of the hub from the pyloric channel. Our patient with the retained PEG hub within the sigmoid was noted at the time of colonoscopy to have diverticulosis. This may have contributed to the retention of the PEG hub. On the basis of this unusual complication, our policy is to obtain an abdominal radiographic evaluation 2 to 3 weeks after cutting the PEG tube and pushing the hub into the gastric cavity, unless expulsion of the PEG hub has been documented (ie, patient or caretaker has seen the PEG hub in the patient’s stool). We discuss the importance of this management with the patient and/or caretaker both before and after the PEG tube placement. We have found the other techniques, such as endoscopic removal
Fig 2. A, Endoscopist’s view through the gastroscope of the redundant mucosa circumferentially covering the hub flanges with only the central tip exposed. B, Tip of the hub was snared and removed. C, Blood clot is seen in the central portion of the redundant mucosa. Of note, the exterior portion of the PEG tube had already been cut; see text for details.
and the traction technique of the tube and inner hub through the abdominal wall, less attractive. Patients generally prefer to avoid another procedure requiring intravenous sedation, separate scheduling, and further expense. The traction technique of removal of the tube
SANTOS and MCDONALD
Table 1. Guidelines Additional recommended guidelines for PEG tube placement 1. Prophylactic antibiotics? (see text) 2. Reverse Trendelenburg 3. Single trocar needle use 4. “Approximate don’t strangulate” abdominal wall with inner hub and outer flange 5. Note centimeter marking on tube and record on chart for reference 6. If there is tumor-seeding potential, use PEG after bulky cancer resection at same procedure, preoperative push PEG, laparoscopic gastrostomy, or radiographic or open gastrostomy Additional recommended guidelines for PEG tube management 1. Delay feedings for 24 hours, or longer if any abdominal symptoms beyond normal after surgical pain 2. Irrigate PEG 3 times daily with water or after meals 3. Check centimeter marking on PEG tube 4. Remind patients or caretakers of normal PEG tube maintenance Additional recommended guidelines for PEG tube removal 1. Endoscopic removal in patients who are poor candidates for no. 2 below (ie, prolonged postoperative ileus, pseudoobstruction, or intestinal strictures/narrowing)11 2. Cut tube at skin surface technique: After tube is cut and hub is pushed into stomach, obtain abdominal radiograph at 2-3 weeks or document hub in stool 3. Traction technique of tube and hub removal through abdominal wall
and hub in the clinic was not tolerated by our patient as described above despite use of a PEG tube designed for removal by the traction technique. Newer and softer gastrostomy tubes, such as the Ross Flexiflo InvertaPEG with a roll-tip bumper, may allow for easier traction removal in the future. We do not have enough experience with the newer PEG tubes to comment. Our rate of peristomal perioperative infections, both cellulitis and skin abscesses, was 6%. Patients received perioperative antibiotics if prescribed for other medical indications (eg, treatment of aspiration pneumonia) or the procedure was combined with a cancer resection or tracheostomy. Akkersdijk et al13 prospectively evaluated the rate of infection with the “pull” PEG technique both with and without antibiotic prophylaxis, as well as the “push” PEG technique without antibiotic prophylaxis. They found fewer peristomal infections (14% vs 30%) with antibiotics. Interestingly, the push PEG technique, which avoids any instrumentation through the pharynx or esophagus and thus oral bacterial flora, did not have a lower peristomal infection rate than the pull PEG technique without prophylactic antibiotics. This suggests that oral flora are less of a concern than skin flora. Three prospective randomized double-blind studies, evaluating the same question of benefit of prophylactic antibiotics with pull PEG tube placement, showed no benefit in 2 (infection rate in prophylaxis groups
Otolaryngology– Head and Neck Surgery February 1999
29%, 13% vs placebo groups 31%, 19%14,15) and a lower infection rate in the third study (prophylaxis group 7%, 32%16). Two of these studies demonstrated a significantly lower infection rate in those patients already receiving antibiotics for another medical condition (3% and 0%, respectively).15,16 The American Society of Gastrointestinal Endoscopy17 recommends universal prophylactic antibiotics (1 g cefazolin or equivalent) 30 minutes before PEG tube placement and references the study of Jain et al16 as justification. The reported use of prophylactic antibiotics for PEG within the literature is inconsistent. Those series reporting use5,7,10 versus those not reporting use of antibiotics4,6,9 had similar skin infection rates (0% to 10%). Often the patient scheduled for PEG is already being given antibiotics or will receive them as part of the head and neck cancer surgery or tracheostomy procedure.9 With the increasing incidence of emerging antibiotic resistances, larger and more specific prospective studies would be useful to help address the value of prophylactic antibiotics in the low-risk patient and in different patient categories (ie, neoplasia, neurologia). Some authors suggest hypochlorhydria as a risk factor for developing infections after PEG and recommend discontinuing any histamine-2 receptor antagonists 24 hours before PEG as well as cleansing the mouth with an antibacterial solution such as povidone-iodine solution.18 However, 2 of the studies mentioned evaluated gastric pH and found no association with infection rates.14,15 Technique may also play an important role in decreasing the risk of infection, as demonstrated in some of our patients in whom skin infections associated with using the trocar needle developed more than once, and perhaps cinching the outer hub too tightly against the skin. We have found the single use of the trocar needle, in cases when the needle is not initially endoscopically identified, is associated with a decreased risk of skin infections. Repeated use of the trocar needle may introduce either skin or bowel organisms into the abdominal wall. In addition, by maintaining the outer flange loose with 1 cm of extra tube between the skin surface and gastric mucosa, the risk of excessive skin pressure and thus ischemia may be significantly reduced. We also recommend noting the centimeter marking on the tube in relation to the flange and recording the distance on the chart in case it slides away from the initial position. Since instituting these technical changes we have not had further perioperative skin infections to date. However, the infections are sufficiently infrequent, thus we will need more experience to determine whether these technical factors are significant. A single instance of inner hub flange gastric mucosa overgrowth occurred in a patient who had a single
Otolaryngology– Head and Neck Surgery Volume 120 Number 2
episode of cellulitis shortly after the procedure and who discontinued use or irrigation or the PEG for months. The lack of hygiene may have contributed to the mucosal overgrowth. A similar case was reported and was thought to be caused by excessive tension on the flange.19 An important potential complication, which fortunately we did not experience, that is reported with increasing frequency in the literature, is the risk of gastrostomy or abdominal wall seeding of the tumor from head and neck cancer. Eight case reports with a total of 10 cases appear to indicate that instrumentation of the upper aerodigestive tract may carry and implant tumor cells into the newly formed gastrostomy after PEG.13,20-26 However, a minority of authors theorize the metastatic mechanism is through hematogenous spread. PEG cases in which seeding was reported occurred primarily in patients with bulky hypopharyngeal, laryngeal, and esophageal cancers. One reported exception is a report of an oral cancer associated with stomal seeding.24 Recognizing the importance of adequate alimentation in patients with head and neck cancer, but considering the risk of stomal seeding, physicians should consider alternatives to standard PEG tube placement. Alternatives to the standard sequence of PEG tube placement before surgery would include PEG after tumor extirpation and before wound closure or preoperative referral for laparoscopic gastrostomy, the push PEG technique, radiographic gastrostomy, or open gastrostomy. We have performed PEG after tumor extirpation and before wound closure/reconstruction with success in 2 patients. The ideal procedure selection is dependent on patient factors of airway, tumor bulk, previous abdominal surgery, and institutional support and experience. At the present time we have incorporated the additional management techniques outlined in Table 1 for the management of PEG in our patients. In summary, as our colleagues and we gain greater experience with PEG procedures, more specific guideline implementation can be recommended. Addendum: We have obtained further experience for the removal of PEG tubes with the roll-tip bumper using the traction technique. No further vasovagal episodes have been observed in the last 15 patients, and this has become our primary means for PEG hub removal. The newer PEG tubes with the roll-tip bumper obviate the need to determine whether the hub is spontaneously expelled. REFERENCES 1. Gauderer MW, Ponsky JL, Izant RJ. Gastrostomy without laparotomy: a percutaneous endoscopic technique. J Pediatr Surg 1980; 15:872-5. 2. Ponsky JL, Gauderer MWL. Percutaneous endoscopic gastrostomy: a nonoperative technique for feeding gastrostomy. Gastrointest Endosc 1981;27:9-11. 3. Santos PM, Afrassiabi A, Weymuller EA. Risk factors associated
SANTOS and MCDONALD
5. 6. 7. 8. 9. 10. 11. 12. 13.
14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.
with prolonged intubation and laryngeal injury. Otolaryngol Head Neck Surg 1994;111:453-9. Shike M, Berner YN, Gerdes H, et al. Percutaneous endoscopic gastrostomy and jejunostomy for long-term feeding in patients with cancer of the head and neck. Otolaryngol Head Neck Surg 1989;101:549-54. Ponsky JL, Gauderer MWL. Percutaneous endoscopic gastrostomy: indications, limitations, techniques and results. World J Surg 1989;13:165-70. Gibson S, Wenig BL. Percutaneous endoscopic gastrostomy in the management of head and neck carcinoma. Laryngoscope 1992;102:977-80. Hunter JG, Lauretano L, Shellito PC. Percutaneous endoscopic gastrostomy in head and neck cancer patients. Ann Surg 1989; 210:42-6. Gibson S, Wenig BL, Watkins JL. Complications of percutaneous endoscopic gastrostomy in head and neck cancer patients. Ann Otol Rhinol Laryngol 1992;101:46-50. Selz PA, Santos PM. Percutaneous endoscopic gastrostomy, a useful tool for the otolaryngologist-head and neck surgeon. Arch Otolaryngol Head Neck Surg 1995;121:1249-52. Urban KG, Terris DJ. Percutaneous endoscopic gastrostomy by head and neck surgeons. Otolaryngol Head Neck Surg 1997;116: 489-92. Korula J, Harma C. A simple and inexpensive method of removal or replacement of gastrostomy tubes. JAMA 1991;265:1426-8. Vemula NR. Cost effective way of removing gastrostomy tubes. Am J Gastroenterol 1991;86:1096-7. Akkersdijk WL, van Bergeijk JD, van Egmond T, et al. Percutaneous endoscopic gastrostomy (PEG): comparison of push and pull methods and evaluation of antibiotic prophylaxis. Endoscopy 1995;27:313-6. Jonas SK, Neimark S, Panwalker AP. Effect of antibiotic prophylaxis in percutaneous endoscopic gastrostomy. Am J Gastroenterol 1985;80:438-40. Sturgis TM, Yancey W, Cole JC, et al. Antibiotic prophylaxis in percutaneous endoscopic gastrostomy. Am J Gastroenterol 1996; 91:2301-4. Jain NK, Larson DE, Schroeder KW, et al. Antibiotic prophylaxis for endoscopic percutaneous gastrostomy. Ann Intern Med 1987;107:824-8. American Society of Gastrointestinal Endoscopy. Antibiotic prophylaxis for gastrointestinal endoscopy statement 1995. Publication no. 1027. Manchester (MA): The society; 1995. Greif JM, Ragland JJ, Ochsner MG, et al. Fatal necrotizing fasciitis complicating percutaneous endoscopic gastrostomy. Gastrointest Endosc 1986;32:292-4. Shallman RW, Norfleet RG, Hardache JM. Percutaneous endoscopic gastrostomy feeding tube migration and impaction in the abdominal wall. Gastrointest Endosc 1988;34:367-8. Huang DT, Thomas G. Stomal seeding by percutaneous endoscopic gastrostomy in patients with head and neck cancer. Arch Otolaryngol Head Neck Surg 1992;118:658-9. Meurer MF, Kenady DE. Metastatic head and neck carcinoma in a percutaneous gastrostomy site. Head Neck 1993;15:70-3. Laccourreye O, Chabardes E, Merite-Drancy A, et al. Implantation metastasis following percutaneous endoscopic gastrostomy. J Laryngol Otol 1993;107:946-9. Schiano TD, Pfister D, Harrison L, et al. Neoplastic seeding as a complication of percutaneous endoscopic gastrostomy. Am J Gastroenterol 1994;89:131-3. Sharma P, Berry SM, Wilson K, et al. Metastatic implantation of an oral squamous-cell carcinoma of a percutaneous endoscopic gastrostomy site. Surg Endosc 1994;8:1232-5. Lee DS, Mohit-Tabatabai MA, Rush BF Jr, et al. Stomal seeding of head and neck cancer by percutaneous endoscopic gastrostomy tube placement. Ann Surg Oncol 1995;2:170-3. Becker G, Hess CF, Grund KE, et al. Abdominal wall metastasis following percutaneous endoscopic gastrostomy. Support Care Cancer 1995;3:313-6.