Percutaneous endoscopic gastrostomy (PEG) is a generally accepted procedure and a well-established alternative to surgical feeding gastrostomies in adults. Also, the superiority of a PEG over nasogastric feeding tubes has been demonstrated in patients with dysphagia resulting from chronic neurologic disease, with improvements both in quality of life and nutritional status (1–4). Many articles dealing with procedure-related complications and nutritional benefits have been published, but relate in particular to adults and to short-term observations. In children and especially in severely mental-motor retarded patients, such data and long-term follow-up results are lacking because the experience has been limited to merely surgical gastrostomies. Mainly the fear of gastroesophageal reflux caused physicians to refrain from performing a PEG. In these cases, the combination of antireflux surgery and a surgical gastrostomy was highly preferred.
Having gained considerable experience with PEGs and percutaneous endoscopic gastrojejunostomies (PEJs; a jejunal tube introduced through the opening of the PEG), we decided to start with PEGs and a standardized protocol for follow-up in psychomotor retarded children and adults. This follow-up would enable us to find answers to questions such as:
- What are the procedure-related problems in the short term and, above all, in long-term follow-up?
- Is enteral support feasible and tolerated in the immediate postinsertion period and in prolonged out-patient care? and
- Are patients appropriately selected for either a PEG or PEJ procedure?
This article analyses prospectively collected data over a period of 5 years from tube placement until gastrostomy removal, death, or arrival at a date chosen to have a minimum follow-up of 6 months.
PATIENTS AND METHODS
All patients referred for PEG placement to the academic hospital were institutionalized and, in most cases, fed by a nasogastric tube. The others had an anticipated need for enteral nutrition of at least 4 weeks. Perinatal asphyxia, hereditary diseases, inborn errors of metabolism, and chromosomal aberrations were mainly responsible for the psychomotor retardation. Because of long-standing disease, severe physical deformities and kyphoscoliosis were present in adolescents and adults and, combined with repetitive pulmonary aspiration, were associated with compromised pulmonary function. A multidisciplinary team consisting of a pediatrician, a pediatric surgeon, and a gastroenterologist reviewed the clinical history of gastroesophageal reflux, aspiration, or both and the results of 24-hour pH-metry and gastric emptying studies, which were part of an experimental protocol in the younger children. In the presence of clinically significant abnormalities, the PEG procedure was abandoned. When in doubt, the endoscopic findings at the time of the procedure were helpful in deciding whether a PEG, a PEJ, or a surgical procedure was appropriate. The Ponsky-Gauderer's pull-on-string method was used, and Freka-Gastrostomy sets of two sizes (9 and 15 French) and Freka-PEJ sets (jejunal tube size 9 French introduced through a 15-French PEG; Fresenius, Bad Hamburg, Germany) were inserted (5). In eligible patients, age and general condition determined the allocation to group 1 of general anesthesia or to group 2 of conscious sedation. In children, there was nearly always a preference for the procedure under general anesthesia. They were treated in the operating room and received antibiotics only when indicated or in case of a complicated procedure. In case the endoscopic positioning failed, the pediatric surgeon intervened in the same session and exposed the anterior surface of the stomach via a small incision in the left upper abdomen. The stomach was punctured with a sheathed needle and a string inserted after removal of the needle. The endoscopist grasped the string and completed the procedure in the usual way. In group 2 patients, general anesthesia or surgical intervention was contraindicated or considered less desirable. They were treated in the endoscopy suite with local skin anesthesia and conscious sedation with either midazolam or computer-controlled propofol infusion. Every patient in the endoscopy suite received 1 g of a first-generation cephalosporin intravenously half an hour before the procedure (6). When the endoscopic procedure failed, patients were transferred to the radiology department. Helped by horizontal and vertical beam fluoroscopy, the procedure was repeated according to Russell's introducer method (7). Through the peel-away sheath of the introducer, a friction-lock Malecott Russell gastrostomy catheter (Cook Surgical, Bloomington, IN) was introduced into the stomach.
A PEJ, a jejunal tube through the opening of a PEG, was indicated when a documented history of reflux esophagitis, gastric emptying disturbances, or pulmonary aspiration was present.
All gastrostomy tubes were made of polyurethane materials. When replacement of the PEG was indicated, a balloon-tipped gastrostomy feeding tube (Kangaroo; Sherwood Medical, St. Louis, MO) or a skin-level gastrostomy balloon button (Mic-Key skin level gastrostomy; Medical Innovations Corporation, Draper, UT) was used. In every instance, the original PEG tube was removed by endoscopy (8).
Feeding was started after 14 to 20 hours, either by 24-hour continuous, nightly or intermittent (over 1–3 hours) drip feeding or by bolus administration. Those who were already receiving nasogastric tube feeding continued their feeding regimen. Medications were only withheld the day of the procedure. As soon as tube feeding was tolerated and feeding instructions were given, the patient was discharged and followed up in the out-patient department after 7 days, 4 weeks, and 12 weeks and every 6 to 12 weeks thereafter. In the first week, showers and baths were not allowed and the parents had to measure daily the length of the outer tube to check the proper position of inner and outer retention discs. They contacted the hospital if a discrepancy with the previous measurement was found. The visit after 7 days was necessary to confirm the formation of the fistulous tract and to release the rather tightly compressed outer disc. Also, the care of the tube, that is, flushing with soda water before and after each bolus feed or at every bottle exchange, was checked. The dietitian monitored the progress of feeding and intervened when the calculated preset goals for nutrition had not been reached.
Analysis and Statistics
Short-term (1–28 days) and long-term (>28 days) complications related to the PEG procedure as well as gastrointestinal complaints and food intolerance and their impact on nutritional goals were assessed. Also, the appropriateness of patient selection was reconfirmed, i.e., the anticipated need of enteral nutrition for at least 4 weeks, and in the case of PEJ, a documented history of reflux esophagitis, delayed gastric emptying, or aspiration.
The production of small amounts of serous, odorless fluid from the wound was considered normal. A rim of redness or induration pointed to an inflammatory response; a true skin infection and abscess were defined by erythema and induration and by the release of purulent, foul-smelling material, often accompanied by local pain. Every instance of inflammation or infection was scored, even when it recurred after treatment. Hypertrophic granulation tissue presented as soft and easily bleeding tissue without pain or signs of inflammation and was only reported on its first appearance. Any occurrence of nausea, vomiting, dumping, and abdominal distension was recorded for each patient, once in early follow-up and once in late follow-up. Clinically relevant diarrhea (at least 3 days of frequent and fluid stools) and constipation (less than 1 stool in 4 days) were reported during prolonged follow-up.
Continuous data were compared with the Student t test. In not normally distributed data, the nonparametric Mann-Whitney U and Kruskal-Wallis test were used. Frequencies were compared by the Yates' corrected chi-square test and Fisher exact test. Relative risks with 95% confidence intervals were calculated. A two-sided P value of less than 0.05 was considered significant.
Fifty-eight patients, 27 males and 31 females, were considered eligible for PEG or PEJ positioning (Table 1). The median ages in group 1 and group 2 patients were 4.1 and 22.0 years, respectively. The youngest patient was 6 weeks old, and the oldest was 45 years of age. Thirty-five patients, most being less than age 5, underwent the procedure under general anesthesia and 23 patients, mainly adults, with conscious sedation.
The initial endoscopic procedure was successful in 52 of the 58 patients (89.7%). A laryngeal spasm developed on introduction of the endoscope in two older patients of group 2. In one patient, the backward displacement of the tongue appeared to be responsible for the inspiratory wheezing, and the procedure was finished successfully. In the other, a second attempt with intubation under propofol sedation was successful.
In three of the six failed endoscopic procedures, a second attempt succeeded. In one of the group 1 patients, a surgical minilaparotomy to expose the stomach that was covered by interposed colon was helpful, resulting in a 100% (35/35) success rate in that group. In 5 of the 23 group 2 patients, a second and twice a third attempt in the radiology department finally succeeded in two patients and resulted in a success rate of 87.0% (20/23) and in an overall success rate of 55/58 (94.8%). In the other three patients, being unfit for general anesthesia and surgery, an interposed megacolon was present. Especially in the aged group, severe spine deformities distorted the anatomic relationship of internal viscera to the abdominal wall, which interfered with adequate transillumination. Physical deformities, combined with compromised pulmonary function, were present in 11 of 23 group 2 patients (47.8%), but only in 5 of 35 group 1 patients (14.3%). A PEG was placed in 48 and a PEJ in 7 patients.
Procedure-related and Tube-Related Complications
Short-term Results (≤ 28 Days)
There were no procedure-related deaths, and the 30-day mortality was zero. Major complications occurred in three patients (5.4%), one in group 1 and two in group 2 patients (Table 2). A pneumoperitoneum found on a plain abdominal radiograph resolved uneventfully without any intervention. Systemic antibiotics were given in the patient with a wound abscess. In the third patient, the PEG was removed after 9 days because of unclear complaints.
Minor complications (30.9%) consisted of two small, uncomplicated local bleedings and 15 wound infections, treated with topical antibiotic ointment in 9 patients, and with systemic intragastric antibiotics in 6 patients with a positive wound culture. Infectious complications occurred in a similar percentage of patients with and without antibiotic prophylaxis. Pain was present in 15 patients in the first days after the procedure and required treatment with paracetamol suppositories.
Long-term Results (>28 Days)
In a median follow-up of 612 days (range, 59–2,178 days), complications of clinical significance (24.1%) were mainly tube related. Once a PEG tube and four times a PEJ tube became dislodged and had to be repositioned in three patients. Tube dysfunction because of porosity, fracture, and clogging necessitated replacement in eight patients. Small-sized 9-French tubes clogged significantly easier than 15-French gastrostomy tubes (relative risk, 4.37; 95% confidence interval, 1.33/14.42). The rate of formula infusion did not have any influence on clogging. One patient with a PEJ pulled out both tubes and removed the balloon replacement catheter as well. Once, feeding via the PEG tube became difficult. The PEG tube had functioned well for 852 days. At endoscopy, only the small opening of the tube was seen, the main part of the inner retention disc being buried underneath the gastric mucosa. To enable feeding, a 9-French tube was placed through the 15-French PEG into the stomach and traction was applied at the outer part of the PEG tube to facilitate further migration of the inner retention disc toward the skin. Two weeks later, the retention disc could be exposed and removed through a small skin incision, and a new PEG was inserted in the same session through the same tract.
Minor complications (37.0%) consisted of skin problems caused by leakage and granulation tissue, which was cauterized with silver nitrate in nine patients.
Gastrointestinal Tolerance of Tube Feeding
In the first week, two patients had symptoms of aspiration. In late follow-up, this was the case in five additional patients (Table 3). Twice the PEG was converted into a PEJ, and in three patients, the prohibition of oral food intake, the infusion of smaller amounts of formula at a slower rate, and feeding only in daytime sufficed to stop aspiration. Two patients continued to aspirate intermittently because of noncompliant oral intake and ongoing rumination.
In the first week, complaints of nausea and vomiting in one quarter of patients interfered with the extension of feeding until full requirements. In follow-up, reports of nausea and vomiting persisted mainly in young and small children. To achieve the required caloric intake, the administration of the tube feed was changed accordingly: more intermittent feeding in daytime and less frequent feeding by bolus feeds (Table 3). Those who tolerated bolus feed were able to be fed at customary meal and between-meal times. The incidence of diarrhea was unexpectedly low, whereas constipation predominated in prolonged follow-up.
Appropriateness of Patient Selection
The clinical prerequisite of an anticipated need of enteral nutrition for at least 28 days was not met in a 31-year-old patient. Because of severe scoliosis and an abnormally high position of the stomach, the PEG procedure was difficult. Afterward, unclear complaints led to its removal after 9 days. A subsequent surgical jejunostomy was tolerated better. A 6-year-old boy did not benefit at all from the PEG. He periodically experienced nausea, vomiting, and rumination. Medical treatment did not relieve these complaints. After surgical gastropexy and gastrostomy, all complaints subsided.
With respect to the indications for the seven PEJ procedures, the correctness of the preference of a PEJ over a PEG could be established. In two patients, a severe grade 3 esophagitis, once with or once without a hiatal hernia, was present. One patient required a ventilator because of respiratory insufficiency and another patient had a history of aspiration. A disturbed gastric emptying was present in three patients. In five of seven patients, the need for a PEJ was only temporary. In two patients with severe reflux esophagitis, the PEJ could be exchanged for a PEG after healing of the esophagitis and the administration of adequate prokinetic and acid-suppressing drugs. Two patients showed improvement of gastric emptying, and the PEJ was removed and a gastric skin level button inserted. In these four patients, the PEJ functioned well for a median of 249 days (range, 106–531 days). One patient pulled out the PEJ and PEG after 35 days. The two patients left with the PEJ, one with documented aspiration and one with severely disturbed gastric emptying, had the tube in situ for 618 and 948 days, respectively.
Outcome of Patients
In the total follow-up of 106 patient years and a median follow-up of 618 days (range, 9–2,178 days), 10 patients died; 9 patients had their first PEG in situ and 1 patient had the PEG replaced before death (Table 4). Two patients died of respiratory insufficiency (14 and 27 years old) and eight patients died of miscellaneous causes such as heart failure, gastrointestinal hemorrhage, and sepsis. Thirty-one patients still needed enteral nutrition, and their tubes functioned well. One patient resumed oral intake and the PEG was removed. Fourteen tubes were exchanged, 3 times on request because of the preference of a skin level gastrostomy button and 11 times because of tube dysfunction. The first tubes as well as the replacement tubes had a long lifetime of a median 523 days (range, 9–1,704 days) and 345 days (range, 24–1,302 days), respectively. Tubes that had to be replaced because of dysfunction functioned well for a median of 366 days (range, 9–935 days). Skin-level gastrostomy balloon buttons were also durable (median, 302 days; range, 213–345 days).
Psychomotor retarded persons are predisposed to gastroesophageal reflux by a dysmotility of the esophagus, a lower esophageal sphincter pressure, gastric emptying disturbances, and an increased abdominal pressure because of spasticity, seizures, scoliosis, and abnormal posturing. In the past, antireflux surgery was combined with a feeding gastrostomy in these patients. Gradually, it became clear that the beneficial results of antireflux surgery obtained in the population at large were highly unfavorable in profoundly disabled persons. This was demonstrated by a perioperative complication rate of 7%, a late complication rate of 10%, and a recurrence of complaints in 71%, resulting in an operative failure rate of 25% and in reoperations in half of the patients (9–11). At the time, we decided to embark on PEG placements in mental-motor retarded patients, and the question of whether an endoscopically positioned gastrostomy would lead to gastroesophageal reflux was not resolved (12–16). In the midst of these disagreements and being undecided ourselves, we chose to group our patients into those in whom it was desirable to have general anesthesia because of the impossibility of the patient remaining relaxed (group 1) and into those being unfit for general anesthesia and surgery, who underwent the procedure with conscious sedation (group 2). The latter group, because of their protracted neurologic disease, were likely to be stunted and highly compromised by severe spinal deformities and by a restricted pulmonary function after repeated (aspiration) pneumonias in the past. Indeed, distorted anatomic relationships resulted in the necessity of radiologic assistance in two patients and in failures in three patients without the possibility of a surgical intervention.
Our absent procedure-related and 30-day mortality rates and a failure rate of 5% seem promising in these high-risk, neurologically severely disabled patients. However, our major and minor complication rates both in the short term and in follow-up may appear rather high. Much depends on the definition of a major complication. Marin et al. (17) defined major complications as those necessitating a surgical consultation or intervention and then found a major complication rate of 19% and a minor complication rate of 17%. Using this definition, we only had one major complication (2%) and only in follow-up, that is, the buried bumper. We did not encounter any case of a gastrocolonic fistula, a complication that is reported to occur more frequently in children because of changes in the local anatomy. When transillumination was questionable, we continued the procedure either by minilaparotomy or by radiologic vertical and horizontal beam exposure. Gauderer (18,19) and Beasley et al. (20) mentioned major complications of 7% and 10%, respectively, in their studies. When looking to minor complications, we had only 6 of 15 wound infections with a positive bacterial culture. This true wound infection rate of 11% equals the 10% finding of Beasley et al. (20).
Unexpectedly long was the lifetime of PEG tubes that functioned well for the predestined time. Even tubes that finally fractured survived for 1 year, which militates against the routine exchanges of tubes. Not appreciated by other investigators, except for Peters et al. (21), were the gastrointestinal side effects. In our study, at each visit these side effects were scored as present or absent, and therefore the numbers are high in view of the prolonged median follow-up of 618 days (range, 9–2,178 days). We observed that frequent consultations at least every 6 weeks after the first 12 weeks were needed to adjust the feeding regimen and to alleviate gastrointestinal symptoms. We assume that these frequent and scheduled visits with at any time the possibility of urgent appointments contributed to the fulfillment of nutritional requirements and to the quite impressive lifetimes of both PEG and PEJ tubes.
Four times we mistakenly opted for a PEG. In two patients who aspirated in the first week, the PEG had to be converted into a PEJ. Two patients did not benefit from the PEG positioning but fared well after a jejunostomy and after an antireflux procedure with surgical gastrostomy. This conversion rate to surgery of 3.6% is similar to the 3% of Gauderer et al. (5) and Gauderer (18) and the 3.6% of Flake et al. (22). Gauderer suggested that the PEG may serve as a screening tool to select patients who need a subsequent antireflux procedure. This was confirmed by many others who found a surprisingly low need for appropriate surgery in 3.3% to 7% of cases with gastroesophageal reflux after PEG insertion (10,20,22–24). Also, although limited, our experience and prolonged follow-up do not seem to support the routine performance of concomitant antireflux surgery in neurologically profoundly impaired patients. Our results support the use of a PEJ as an intermediate solution, notwithstanding the fact that jejunal tubes dislodge, kink, and clog notoriously easily and frequently because of the small caliber. Our figures of displacement (in four patients of whom two needed replacement) and reversible clogging (twice in the same patients) with PEJs are high but still below the numbers reported for another nonsurgical option, that is, the radiologic antegrade percutaneous gastrojejunostomy (9,24).
In conclusion, we advocate the use of PEGs in severely mental-motor retarded persons without objectivated gastroesophageal reflux. In patients with aspiration, gastroesophageal reflux, or gastric motility disturbances, a PEJ may be considered as an intermediate step before proceeding to surgery. A multidisciplinary set-up of a pediatric gastroenterologist, pediatric surgeon, and gastroenterologist should facilitate a fruitful discussion. A standardized follow-up should allow early recognition of feeding and tube-related problems and thereby contribute to optimal nutrition. The influence of PEG positioning on lower esophageal sphincter manometry, pH measurements, and gastric emptying as well as its impact on the quality of life are the subjects of future investigations.
1. Park RHR, Allison MC, Lang J, et al. Randomised comparison of percutaneous endoscopic gastrostomy
and nasogastric tube feeding
in patients with persisting neurologic dysphagia. BMJ. 1992; 304: 1406–9.
2. Wicks C, Gimbson A, Vlavianos P, et al. Assessment of the percutaneous endoscopic gastrostomy
feeding tube as part of an integrated approach to enteral feeding. Gut. 1992; 33: 613–6.
3. Panos MZ, Reilly H, Moran A, et al. Percutaneous endoscopic gastrostomy
in a general hospital: Prospective evaluation of indications, outcome, and randomised comparison of two tube designs. Gut. 1994; 35: 1551–6.
4. Norton B, Homer-Ward M, Donnelly MT, et al. A randomised prospective comparison of percutaneous endoscopic gastrostomy
and nasogastric tube feeding
after acute dysphagic stroke BMJ. 1996; 312: 13–6.
5. Gauderer MWL, Ponsky JL, Izant Jr. RJ Gastrostomy without laparotomy: A percutaneous endoscopic technique. J Pediatr Surg. 1980; 15: 872–5.
6. Jain NK, Larson DE, Schroeder KW, et al. Antibiotic prophylaxis for percutaneous endoscopic gastrostomy
. Ann Int Med. 1987; 107: 824–8.
7. Russell TR, Brotman M, Norris F. Percutaneous gastrostomy, a new simplified and cost-effective technique. Am J Surg. 1984; 148: 132–7.
8. Yaseen M, Steele MI, Grunow JE. Nonendoscopic removal of percutaneous endoscopic gastrostomy
tubes: morbidity and mortality in children. Gastrointest Endosc. 1996; 44: 235–8.
9. Albanese CTG, Towbin RB, Ulman I, et al. Percutaneous gastrojejunostomy versus Nissen fundoplication for enteral feeding of the neurologically impaired child with gastroesophageal reflux. J Pediatr. 1993; 123: 371–5.
10. Boyle JT. Nutritional management of the developmentally disabled child. Pediatric Surg Int. 1991; 6: 76–81.
11. Martinez DA, Ginn-Pease M, Caniano DA. Sequelae of antireflux surgery in profoundly disabled children. J Pediatr Surg. 1992; 27: 267–73.
12. Berezin S, Schwartz SM, Halata MS, et al. Gastroesophageal reflux secondary to gastrostomy tube placement. Am J Dis Child. 1986; 140: 699–701.
13. Jolley SG, Smith EI, Tunell WP. Protective antireflux operation with feeding gastrostomy: Experience with children. Ann Surg. 1985; 201: 736–40.
14. Mollitt DL, Golladay ES, Seibert JJ. Symptomatic gastroesophageal reflux following gastrostomy in neurologically impaired patients. Pediatrics. 1985; 75: 1124–6.
15. Johnson DA, Hacker JF, Benjamin SB, et al. Percutaneous endoscopic gastrostomy
effects on gastroesophageal reflux and lower esophageal sphincter. Am J Gastroenterol. 1987; 82: 622–4.
16. Grunow JE, Al-Hafidh AS, Tunell WP. Gastroesophageal reflux following percutaneous endoscopic gastrostomy
in children. J Pediatr Surg. 1989; 24: 42–5.
17. Marin OE, Glassman MS, Schoen BT, et al. Safety and efficacy of percutaneous endoscopic gastrostomy
in children. Am J Gastroenterol. 1994; 89: 357–61.
18. Gauderer MWL. Feeding the neurologically impaired child: evaluation and surgical option. Ped Surg. 1991; 6: 75–80.
19. Gauderer MWL. Percutaneous endoscopic gastrostomy
: A 10-year experience with 220 children. J Ped Surg. 1991; 26: 288–94.
20. Beasley SW, Catto-Smith AG, Davidson PM. How to avoid complications during percutaneous endoscopic gastrostomy
. J Pediatr Surg. 1995; 30: 671–3.
21. Peters JM, Simpson P, Tolia V. Experience with gastrojejunal feeding tubes in children. Am J Gastroenterol. 1997; 92: 476–80.
22. Flake AW, Shopene C, Ziegler MM. Antireflux gastrointestinal surgery in the neurologically handicapped child. Pediatric Surg Int. 1991; 6: 92–4.
23. Coben RM, Wintraub A, DiMarino JR, AJ et al. Gastroesophageal reflux during gastrostomy feeding. Gastroenterology. 1994; 106: 13–8.
24. Launay V, Gottrand F, Turck D, et al. Percutaneous endoscopic gastrostomy
in children: Influence on gastro-oesophageal reflux. Pediatrics. 1996; 97: 726–8.