Bone marrow transplant (BMT) has become an important option for treatment of many conditions, including nonmalignant diseases (1). The side effects of intensive chemotherapy, such as nausea, vomiting, mucositis, anorexia, taste alteration, and malabsorption, in combination with the underlying illness, predispose patients to suboptimal oral intake and potential malnutrition (2). Total parenteral nutrition has traditionally been used to aid in optimizing nutrition for patients with BMT; however, there is growing evidence that enteral feeding is preferable (1,3). Advantages of enteral feeding include maintaining normal gut function, providing nutrients not available through total parenteral nutrition, decreasing the risk of infection, maintaining gut-associated immune function, and cost containment (3–5).
Enteral feeding can be achieved via nasogastric tubes in patients unable to feed adequately by mouth. Nasogastric tubes are an effective way of delivering enteral nutrition (3,5,6), although there are several associated risks, including vomiting, tube dislodgement, excess discomfort because of mucositis, occlusion of the tube, and potential bleeding from placement or replacement, secondary to thrombocytopenia (2,7). Nasograstric tube feedings have been used in pediatric patients with BMT with good outcomes (8,9) and can be the preferred method for enteral nutritional support. Few studies have investigated the use of percutaneous endoscopic gastrostomy (PEG) to provide enteral feedings for this group of patients (10–12).
Gastrostomy tubes have been used in many areas of pediatrics with successful outcomes (13). The first percutaneous gastrostomy tube was placed at Rainbow Babies Children's hospital in 1979 and the technique was published in 1980 (14). The number of gastrostomy tubes being placed for long-term nutritional support in pediatric patients is increasing (15). Studies have shown good outcomes and improved nutritional status with the use of gastrostomies, but complications such as localized inflammation, infection, exuberant granulation tissue, mechanical problems, insertion site bleeding, and feeding intolerance are common (16). In pediatric oncology patients, gastrostomy tubes have not been traditionally used, mainly because of concerns of neutropenia or thrombocytopenia (7).
Recent studies of pediatric oncologic patients with gastrostomies have demonstrated optimization of nutrition, despite the inherent risks of infection (10,11). Successful weight maintenance after PEG placement in adult patients after BMT has also been reported but with a definite risk of localized infection and bacteremia, especially when receiving immunosuppressive therapy for graft versus host disease (12). It is an extremely difficult clinical decision whether the complications of placing a gastrostomy in a high-risk patient outweigh the benefits. With little data available regarding the use of PEGs in pediatric bone marrow patients, we report our experience, highlighting the incidence of complications as well as the risk factors for developing complications.
We conducted a retrospective chart review for all pediatric patients who received a PEG at the Mount Sinai Medical Center between January 2007 and December 2010. Subjects were grouped by those who had a PEG placed in anticipation of a BMT, designated the BMT group, and those who had a PEG for other indications, designated the comparison group. Both inpatient and outpatient records were reviewed for diagnosis, sex, age, weight, albumin, white blood cell count, absolute neutrophil count (ANC) at the time of PEG placement, indication for PEG, minor and major PEG complications, and treatment of complications. Follow-up data for PEG complications were obtained from outpatient visit notes and documentation by a nurse who provided follow-up phone calls to all subjects post-PEG procedure. Specific documentation of words such as “erythema, induration, purulent discharge, cellulitis” around the gastrostomy tube site was recorded as a potential indication of infection, with or without concurrence of fever.
A major complication was defined as a PEG site infection severe enough to require removal of the PEG or oral or parenteral antibiotics. This is unvarying to previous studies that have defined major complications as those requiring hospitalization (admission for IV antibiotics secondary to infection) or surgical intervention (removal of gastrostomy tube for infected fistula) (15) as well as peritonitis and localized abscess in the abdominal wall (10). Minor complications such as granuloma, or localized cellulitis, responsive to topical antibiotics, were noted but not included as a major complication. For those subjects undergoing a BMT, the time from PEG placement to BMT was also noted.
Statistical analysis was performed using SPSS 19 (SPSS Inc, Chicago, IL). Data were analyzed using descriptive statistics with percentages, median, and range. Major complications were compared between the 2 groups using χ2/Fisher exact test or Wilcoxon rank sum test when appropriate. Statistical significance was determined at P < 0.05. The institutional review board of the Mount Sinai School of Medicine approved the present study.
All PEGs were placed at the Mount Sinai Medical Center endoscopy suite using the standard pull technique, by a pediatric gastroenterologist with assistance from a pediatric gastroenterology fellow. One dose of cephalexin was given intravenously before PEG placement and 2 doses were given postplacement. The gastrostomy tube sizes used were 16, 18, or 20 F depending on the size of the child.
There were 12 subjects in the BMT group. Eleven subjects received BMT at Mount Sinai Medical Center. One subject went to an outside institution for a BMT and was excluded from the analysis. Diagnoses of subjects who received a BMT are listed in Table 1. Of these 11 subjects, there were 7 boys; the median age at BMT was 2 years (7 months–12 years), and the median time elapsed between PEG placement and BMT was 56.5 days (13–160 days). One subject had a PEG placed 67 days after BMT because of poor nutritional status and inability to tolerate oral feeds. This subject was a 12-year-old boy with acute lymphoblastic leukemia who had an ANC of 1483 cells/μL at time of PEG placement, and later died from complications of relapsed lymphoblastic leukemia. The median hospital stay for the BMT group during their admission for transplant was 37 days (21–103 days).
As seen in Table 2, there were no statistically significant differences between the 2 groups before PEG placement for age, weight z score, albumin, or white blood cell count, thus improving the comparability of the 2 groups.
Of the 30 subjects in the comparison group, only 1, a child with multiple food allergies and failure to thrive, had a major complication with cellulitis requiring systemic antibiotics. This superficial PEG site infection developed 18 days after PEG placement and responded well to a 10-day course of enteral cephalexin. There were 2 subjects in the comparison group who were neutropenic, as defined by an ANC <1500 cells/μL (one subject with ANC of 1299 cells/μL, the other with ANC of 982 cells/μL) at time of PEG placement, but they did not go on to develop any complications.
The average time of follow-up after PEG placement for the subjects in the comparison group was 6 months, with a range of 5 to 15 months. Five of the subjects were not able to be seen in the pediatric GI clinic for follow-up, as they resided in chronic care facilities and there were no complications reported based on telephonic follow-up.
The age, diagnosis, ANC at PEG placement, ANC at the time of complications, and the nature and treatment of the complications in BMT subjects are listed in Table 3. Of the 11 subjects, 4 (36%) had a major complication related to PEG requiring intravenous antibiotics and 3 of those required PEG removal because of severity of infection and poor clinical status. Two of the 4 subjects who developed a major complication were moderately neutropenic (ANC <1000 cells/μL) at the time of PEG placement. At the time of the complication, all subjects in the BMT group were neutropenic with 3 of 4 showing severe neutropenia (ANC <500).
There were no mechanical complications such as tube dislodgement or obstruction in our population, and there were no PEG-related deaths. Of note, there were 2 subjects in the comparison group who were mildly neutropenic (ANC <1500 cells/μL) at PEG placement, but they did not go on to develop any complications.
There is little dispute that for patients who cannot meet their nutritional requirements by a normal diet and who have a functioning gastrointestinal tract, enteral feedings have many advantages over parenteral nutrition. The benefits of enteral feeding may even include reduction of graft versus host disease, a potentially devastating complication of BMT (3). The preferred method of delivering enteral feeds, be it by nasogastric tube, gastrostomy, or some other type of tube, can often be debated. The decision is often based on the length of time enteral feedings will be necessary, the underlying diagnosis, patient comfort, comorbid risk factors, as well as long-term prognosis. Many clinicians have been reluctant to place gastrostomies in patients undergoing chemotherapy, whether for treatment of hematological or solid tumors, or as part of induction for BMT. Despite Barron et al (10) reporting >10 years ago the use of gastrostomy tubes in children with cancer, including 8 with BMT, there has been a continued reluctance to place gastrostomies. In Barron et al (10) study of 44 patients, only 3 developed major complications consisting of either peritonitis or abscess.
In the present study, we chose to research only patients with BMT, and to use the coexisting patients who had undergone gastrostomy placement for other indications as a comparison. The 2 groups were comparable with regard to age and nutritional status as determined by weight zscores and albumin. Despite the small group, there were statistically different rates of complication, in that 36.4% of the BMT patients had a major complication compared with 3.3% of the comparison group. The complications were infectious in nature and consisted of cellulitis around the gastrostomy site that was severe enough to require parenteral antibiotics and, in 3 subjects, removal of the gastrostomy tube. The ANC at the time of placement and at the time of the complication seemed to have a large influence. PEGs were placed during neutropenic episodes for nutritional indications. These patients were not tolerating oral feeds because of their clinical condition, including mucositis from chemotherapy. It was anticipated that the nutritional status of these patients would worsen as intense chemotherapy before BMT approached. Our retrospective review demonstrates that PEG placement, as an elective procedure, should be avoided during neutropenic episodes if possible. A possible approach to address this problem would be administration of granulocyte colony-stimulating factor to improve the ANC and decrease infectious outcomes, but this would need interdisciplinary research to investigate feasibility and efficacy.
The high infection rate in the BMT group may reflect the greater severity of illness in the BMT population, which in previous studies has been attributed mainly to poor nutritional status. Roberts et al (12) reviewed the charts of 16 patients, of which 5 were neutropenic at the time of PEG placement. After PEG placement, 11 patients developed bacterial infections, 8 of which were immunocompromised when the infection occurred. The study did not make a correlation between neutropenia at the time of PEG placement and subsequent risk of infection, but noted that the development of infection in their population was likely because of the poor nutritional status (12). In our patient population, there was no statistically significant difference in weight z scores and albumin between the BMT and the comparison group. The z scores were used to assess the deviation of subjects’ weights from the average weight of comparable children in the reference population of CDC growth charts. Although serum albumin in not the most sensitive test for malnutrition, the fact there was no difference in albumin and weight z scores in our 2 populations suggests that malnutrition was not the main cause of the increased complications in our subjects with BMT.
Superficial infections, such as localized erythema, swelling, or pus, have been cited as a minor complication that can be treated with systemic or local antibiotics (11,12). Bisgaard Pederson et al (11) reported that 12 of 32 subjects encountered 23 cases of local infections. Fourteen of the 23 infections were treated with systemic or local antibiotics. Barron et al (10) reported that 18 of their 34 patients had documented episodes of localized site inflammation; 10 of the 44 had >1 episode. These infections were treated with intravenous or oral antibiotics, but it was not clarified how many patients received intravenous or oral. A local infection is not a minor complication in an immunocompromised patient, especially if it requires treatment with systemic antibiotics. When a patient's overall clinical status declines, a PEG is another possible source from which a localized infection may lead to further complications of bacteremia. Three of 4 subjects in our study required removal of PEG because of worsening swelling and erythema around the gastrostomy tube site and poor clinical status. This raises the concern of whether PEGs should be placed in high-risk patients. Furthermore, it is plausible that the BMT group had a greater risk of nosocomial and multiresistant bacterial infections given their longer stay in the hospital. On average, the hospital stay for the control group was 1 day after PEG placement. There were only a few outliers with chronic conditions that stayed for a longer period of time and they did not go on to develop any complications from PEG placement. It may be beneficial to use a prolonged course of antibiotics for patients who are neutropenic at the time of PEG placement to reduce risk of infection. Future studies should investigate usage of a longer course of antibiotics.
PEG complications in the BMT group all occurred within 2 months of the procedure, with a mean of approximately 4 weeks. This is mostly in keeping with previous studies showing that site infections are most likely to occur in the first 2 to 3 weeks after PEG placement (13). Fortunato et al (13) reviewed 94 patients with PEG and found that infections occurred in 37% of patients, with most experiencing a single infection that occurred within the first 15 days after the tube placement. Two of the 4 subjects with BMT in our study who developed an infection did so within the first week after placement of the PEG, with the other two ranging between 1.5 and 2 months. This was also the case with the subject in the comparison group who developed cellulitis around the PEG site approximately 2 weeks after PEG placement. Based on our findings, along with previous studies, the 6-month follow-up that was used was an appropriate length of time to determine whether there were further complications.
Our study sample of subjects with BMT is small, limiting the conclusions that can be drawn; however, the high rate of complication in the BMT group was significantly higher than the comparison group. We also recognize that a retrospective chart review has limitations, which include potential biases in data collection, most relevant to subjective clinical findings. We attempted to minimize these biases by using specific key words to interpret subjective findings, such as “erythema, swelling, discharge, pus, and cellulitis.” Nonetheless, the clinician documentation bias, which may depend on severity of patient's clinical course, likely still exists.
Our findings show that the incidence of complication in subjects with BMT compared with non-BMT subjects was significant. All subjects with BMT were neutropenic at the time they developed an infection. Our findings support that ANC should be considered before placement of PEG, and the inherent increased risk of infection during episodes of neutropenia in patients with BMT should raise concern for management of nutritional support in patients with BMT via PEG. Future studies should address the usage of PEGs in patients receiving a BMT and assess the complications and outcomes of PEGs compared with nasogastric tubes in pediatric subjects who are undergoing a BMT.
1. Hastings Y, White M, Young J. Enteral nutrition and bone marrow transplantation. J Pediatr Oncol Nurse 2006; 23:103–110.
2. Thompson JL, Duffy J. Nutritional support challenges in hemaopoeitc stem cell transplant patients. Nutr Clin Pract 2008; 24:533–545.
3. Seguy D, Berthon C, Micol JB, et al. Enteral feeding and early outcomes of patients undergoing allogeneic stem cell transplantation following myeloablative conditioning. Transplantation 2006; 82:835–839.
4. Langdana A, Tully N, Molloy E, et al. Intensive enteral nutrition in paediatric bone marrow transplantation. Bone Marrow Transplant 2001; 27:741–746.
5. DeSwarte-Wallace J, Firouzbakhsh S, Finklestein JZ. Using research to change practice: enteral feedings for pediatric oncology patients. J Pediatr Oncol Nurs 2001; 18:217–223.
6. Hopman GD, Pena EG, Le Cessie S, et al. Tube feeding and bone marrow transplantation. Med Pediatr Oncol 2003; 40:375–379.
7. Lipkin AC, Lenssen P, Dickson BJ. Nutrition issues in hematopoietic stem cell transplantation: state of art. Nutr Clin Pract 2005; 20:423–459.
8. Pietsch JB, Ford C, Whitlock JA. Nasogastric tube feedings in children with high risk cancer: a pilot study. J Pediatr Hematol Oncol 1999; 21:111–114.
9. Papadopoulou A, MacDonald A, Williams MD, et al. Enteral nutrition after bone marrow transplantation. Arch Dis Child 1997; 77:131–136.
10. Barron MA, Duncan DS, Green GJ, et al. Efficacy and safety of radiologically placed gastrostomy tubes in paediatric hematology/oncology patients. Med Pediatr Oncol 2000; 34:177–182.
11. Bisgaard Pedersen AM, Kok K, Petersen G, et al. Percutaneous endoscopic gastrostomy in children with cancer. Acta Paediatr 1999; 88:849–852.
12. Roberts SR, Miller JE. Success using PEG tubes in marrow transplant recipients. Nutr Clin Pract 1998; 13:74–78.
13. Fortunato JE, Troy AL, Cuffari C, et al. Outcome after percutaneous endoscopic gastrostomy in children and young adults. J Pediatr Gastroenterol Nutr 2010; 50:390–393.
14. Gauderer MW, Ponsky JL, Izant RJ Jr. Gastrostomy without laparotomy: a percutaneous endoscopic technique. J Pediatr Surg 1980; 15:872–875.
15. Goldberg E, Barton S, Xanthopoulous, et al. A descriptive study of complications of gastrostomy tubes in children. J Pediatr Nurs 2010;24:72–80.
16. Mathew P, Bowman L, Williams R, et al. Complications and effectiveness of gastrostomy feedings in pediatric cancer patients. J Pediatr Hematol Oncol 1996; 18:81–85.