Complications of Nasogastric and Gastrostomy Tube Feedings in Children with End Stage Renal Disease

Ellen G. Wood, Timothy E. Bunchman, Renuka Khurana, Susan S. Fleming, Robert E. Lynch

Twelve children with end stage renal disease requiring dialysis received enteralfeedings via nasogastric (NG) or gastrostomy (G) tube between 1984 and 1989 for provision of adequate nutrition. Records were reviewed for frequency and types of complications seen.

Six patients, ages 1 week to 16 months received NG feedings for a total of 32 months. Complications included persistent vomiting with recurrent aspiration (2 ), persistent vomiting with peritoneal dialysis (PD) exit site leak ( 1 ), sinusitis ( 1 ), and refusal to continue NG feeds because of patient/parental anxiety ( 1 ). Three of the 6 were changed to G tubefeedings after 2 days to 3 months . The complication rate was 1 per 6.4 patient months .

Nine patients, ages 4 days to 11 years, received G tube feedings for 64 months. The complication rate was similar, 1 per 7.1 months. Complications were PD fluid leak around G tube exit site (1), G tube infection (2), G tube obstruction requiring tube replacement (3), tube migration producing intestinal obstruction ( 1 ), and gastrocutaneous fistula (2) .

Both methods were associated with similar complication rates, although somewhat different types of complications were seen. The young dialysis patient may have certain unique risks in addition to the complications generally associated with enteral feedings .

Key words

Enteral feedings, complications, children, renal failure


the Department of Pediatrics and Adolescent Medicine, St. Louis University, Cardinal Glennon Children's Hospital, St. Louis, MO.


Growth failure is an important clinical problem in management of infants and children with chronic renal failure (CRF). Multiple factors may contribute to growth retardation including anorexia, poor nutrition, acidosis, salt wasting, renal osteodystrophy, anemia, hormonal imbalance, and psychosocial deprivation (1,2).

Enteral tube feedings via nasogastric tube or gastrostomy are being used with increasing frequency as a means of delivery of adequate caloric intake to these young patients (3-5). However, little data exists regarding risks and complications with either route of administration in this unique group of patients (3-5). Both methods of enteral feeding delivery have been used at our institution during the past 5 years, thus providing the opportunity to examine the results of the delivery methods.


Twelve children with CRF received enteral feedings between 1984 and 1989 at Cardinal Glennon Children's Hospital. Initially 6 received feedings by NG tube and 6 by gastrostomy (G tube). Records of each patient were reviewed to determine types of complications and their frequency. All patients required renal replacement therapy. Type of dialysis included CAPD in 5, CCPD in 3, and both PD (CAPD or CCPD) and hemodialysis during the study period in 4.

Caloric delivery was similar in both groups, varying from 70-120% of caloric RDA for age. Three patients received intermittent feeds while 9 received continuous feedings for 8-12 hours each day.

Etiologies of renal failure included posterior or anterior urethral valves in 4, dysplasia/hypoplasia (3), urethral atresia (2), and I each with congential nephrotic syndrome and bilateral cortical necrosis. Etiology was unknown in one patient.

Data is reported as mean :t standard deviation. Statistical analysis was based on unpaired Student's t test with values of p s 0.05 considered significant.

Nasogastric tube

Age at initiation of feedings ranged from I week to 16months(4of6 < 1 yearofage). These 6 patients received NO feedings for 5.3 :t 5.1 months, for a total of 32 patient months of treatment. Three of the 6 were changed from NO to 0 tube feeds after 2 days, I week, and 3 months. Of the 3 patients receiving only NO feedings, 1 had intractable vomiting and several episodes of aspiration associated with gastroesophageal (OE) reflux. This was managed with change from NO to nasoduodenal tube. The remaining 2 patients had only intermittent vomiting that did not affect growth. One of these 2 had several episodes of sinusitis. Of the 3 patients changed to 0 tube feedings, 1 received NO feedings for 3 months. Because of frequent vomiting associated with apnea and pulmonary aspiration, and the presence of concurrent OE reflux, a Nissen fundoplication was performed and 0 tube placed. The second patient developed a PD exit site leak requiring surgical repair after 1 week of NO feeds associated with persistent vomiting. A 0 tube was subsequently placed and vomiting resolved. The last patient was changed to gastrostomy feedings because of patient and parental intolerance. The overall complication rate was 1 per 6.4 patient months.

Gastrostomy tube

Age at initiation of feedings was similar to the NO tube group, ranging from 4 days to II years (5 of 9 < 1 year of age). Ages of the 3 patients that changed from NO to 0 tube were 3 months, 8 months, and 15 months. Intractable vomiting with its subsequent complications was less often encountered while mechanical problems were more likely in this group.

One patient has leakage of PD fluid around the 0 tube exit site at 2 days and again at 3 weeks after gastrostomy placement. PD volume was temporarily decreased with subsequent resolution of the leak. However, an episode of peritonitis occurred within 48 hours of the second leak. Two patients developed local infections at the 0 tube exit. One responded to antibiotics without further problems. The other occurred in the patient previously described who underwent Nissen fundoplication with gastrostomy tube placement after recurrent vomiting and aspiration. This infant developed wound infection, 0 tube exit site infection and peritonitis from Citrobacter freundii postoperatively. He responded to antibiotics with subsequent cultures being negative. However, he died 3 months later with Citrobacter meningitis. Two patients had 3 episodes of 0 tube obstruction requiring replacement of the tube without subsequent problems. Two episodes occurred after administration of sodium polystyrene sulfonate for chronic treatment of hyperkalemia and the third after a trial of high protein formula (Traumacal, Mead Johnson). One patient had tube migration into the duodenum with vomiting and dehydration from partial intestinal obstruction. Finally, 2 patients developed small gastrocutaneous fistulas after 0 tube removal requiring I day admissions for surgical closure. The overall complication rate for this group was I per 7.1 patient months, not statistically different from the NO tube group.

Nine of the 12 children subsequently underwent renal transplantation while 3 patients died prior to transplantation of causes unrelated to feeding method (2 in 0 tube group, I in NO group). All 3 infants who died had pulmonary hypoplasia (2 required extracorporeal membrane oxygenation as a lifesaving procedure at birth). One died with meningitis as descibed above, 1 after cerebral infarction of unknown etiology, and the third from idiopathic hypertrophic subaortic stenosis.


Utilization of NO and 0 tubes for enteral feeding delivery has been described repeatedly in the pediatric literature (6-9). Each method has its own unique risks. Complications of chronic NO tube use include sinusitis, nasoseptal erosion, pulmonary placement, esophageal or gastric irritation or perforation, an increased frequency of gastroesophageal reflux, and vomiting with aspiration (6,7). In addition, psychosocial stress for parents and/or child with frequent tube placement has been reported (5). G tubes also have potential risks for gastric irritation and perforation, as well as vomiting and aspiration. Other risks include wound infection, stomal leaks, tube prolapse into the duodenum producing .intestinal obstruction, spontaneous intraperitoneal leak of gastric contents, and separation of the stomach from the anterior abdominal wall during tube reinsertion (8.9).

Complications seen in our patient population were, in most cases, those described for other populations. In addition, several complications unique to this patient population were seen. Two patients ( 1 in each group) developed PD fluid leaks either at the G tube site or at the PD exit site requiring alteration of dialysis regimen. Increased intra abdominal pressure from dialysate (10) may place these patients at higher risk for leaks. Uremia and malnutrition may also contribute. The use of high caloric, low volume feedings as well as viscous medications led to G tube obstruction on 3 occasions. Obstruction of tubing was commonly seen in the NG tube patients as well, although not listed as a complication as the risk of replacement of NG tube is less than that of G tube. This problem may be a relatively unique one for this patient population, where volume limitations and multiple medications are often necessary .GE reflux has been reported to occur in over 70% of infants with CRF (11). The combination of reflux and NG tube use clearly contributed to problems with recurrent vomiting seen in 2 of our patients. Finally, steroid therapy may have contributed to incomplete closure of G tubes sites after successful transplantation. Although each method of enteral feeding delivery was associated with significant risks, overall complication rates were similar and their frequency seems acceptable in light of the benefits gained from reasonable nutrition in this population. Certain complications associated with G tube use, obstruction and tube migration, have been subsequently avoided with changes in the medication and formula prescription (12) and better parental education. Use of G tube or the feeding gastrostomy button (13) may result in decreased anxiety on the part of patient and parents as frequent replacement of tubing is avoided.

In conclusion, both NG tube and G tube appear to be reasonable alternatives for feeding delivery in the young patient with CRF .These patients have several unique risks compared to other pediatric patients requiring enteral tube feedings.


  1. Broyer M. Growth in children with renal insuffi ciency. Ped Clin NA 1982; 29: 991-1003.
  2. Powell DR. Growth in children with chronic renal failure. The Kidney 1989; 22: 7-12.
  3. Wassner SI, Abitbol C, Alexander S, et at. Nutritional requirements for infants with renal failure. Am I Kidney Dis 1986; 7: 300-5.
  4. Strife CF, Quinlan M, Mears K, et at. Improved growth of 3 uremic children by nocturnal naso gastric feedings. AIDC 1986; 140: 438-43.
  5. Stanton EL, Telamini-Iones L. Nutrition in pediatric renal disease. Dialysis and Transplantation 1989; 4: 187-94.
  6. Andrassy RI, Page CP, Feldtman RW, et at. Continual catheter administration of an elemental diet in infants and children. Surgery 1977; 82: 205-10.
  7. Reimer SL, Michener WM, Steiger E. Nutritional support of the critically ill child. Pediatr Clin NA 1980; 27: 647-60.
  8. Haws EG, Sieber WK, Kiesewetter WB. Complications of tube gastrostomy in infants and children. 15 year review of 240 cases. Ann Surg 1966; 164: 284-90.
  9. Campbell IR, Sasaki TM. Gastrostomy in infants and children: an analysis of complications and techniques. Am Surg 1984; 40: 505-8.
  10. Bargman IM. Complications of peritoneal dialysis related to raised intra-abdominal pressure. Dialysis Transplant 1990; 19: 70-80.
  11. Ruley El, Bock GH, Kerzner B, et at. Feeding disorders and gastroesophageal reflux in infants with chronic renal failure. Pediatr Nephro11989; 3: 424-29.
  12. Schenck MH, Lynch RE, Wood EG, et at. Pretreatment of formula with sodium polystyrene sulfonate to prevent hyperkalemia in children with chronic renal failure or on dialysis. Perit Dial Intemat 1990; 10(suppll): 23.
  13. Gauderer MWL, Olsen MM, Stellato TA, et at. Feeding gastrostomy button: experience and recommendations. I Pediatr Surg 1988; 23: 24-28.