This study utilized a large representative national database to investigate methods for gastrostomy placement in children. SG was the more commonly utilized method, and this proportion was generally consistent over time. PEG placement was more commonly used among children with private insurance and was less commonly used among infants, children at urban hospitals, and children cared for at children’s hospitals. The use of PEG varied dramatically across centers, with some utilizing exclusively PEG and others performing exclusively SG. This variability between centers was more pronounced among infant patients.
The importance of the variability identified in this study depends upon the presence or absence of variability in outcomes between the methods of gastrostomy placement. The best method of gastrostomy placement in children is currently unknown. Several prior studies have compared the outcomes of PEG and SG, including at least three meta-analyses. The first review of 22 studies with a total of 5438 patients found a significantly increased risk of major complications with PEG compared to laparoscopic gastrostomy mostly due to a difference in visceral injury between the two techniques [3
]. Another recent systematic review and meta-analysis including five retrospective studies found that while completion rates and minor complications were similar between the 550 PEG placements and 483 laparoscopic gastrostomy placements examined, significantly more bowel injuries, early tube dislodgements, and complications requiring reintervention under general anesthesia occurred after PEG placement [4
]. Finally, the most recent systematic review confirmed these findings, showing that in a total of eight studies examining 1550 patients undergoing gastrostomy placement, the odds of major complications were more than three-fold higher with PEG (5.4%) compared to laparoscopic gastrostomy (1.0%) [2
Concerns regarding the safety of PEG primarily stem from the endoscopic intra-gastric view that is utilized for PEG placement without direct visualization of adjacent organs and the inability to immediately identify injuries when they do occur, leading to delayed diagnosis and potentially increased morbidity. SG is at least theoretically safer, therefore, as it provides an intra-abdominal view to minimize injury to adjacent organs and allows for immediate identification and correction of injuries that could occur. In addition, SG allows for direct and precise placement of the gastrostomy at an optimal site on the stomach, whereas PEG placement is limited to the area on the stomach that has adequate transillumination.
Multiple studies have reported the safety of PEG placement in children, however. Brewster et al. examined 103 PEG placements performed by pediatric surgeons in children > 2 kg, identifying a total complication rate of 14%. No intra-operative complications or organ injuries were identified, although the authors acknowledged the potential limitation of their 90-day follow-up missing colonic injuries that might be identified later [9
]. A prior study using the KID found that PEG and SG had similar risks of postoperative complications and mortality in relatively uncomplicated infants and neonates in whom gastrostomy was the only procedure performed [10
]. PEG has the benefit of generally shorter operative time than SG. However, this difference must also be weighed against the second general anesthetic sometimes needed for the later exchange of a PEG tube for a low-profile button gastrostomy tube [4
]. The first gastrostomy tube change after SG can typically be done in the outpatient setting without an anesthetic.
Multiple individual centers have reported a higher complication rate with PEG placements [12
]. One review from Italy identified a higher risk of major complications after PEG (13.5%) compared to laparoscopic gastrostomy (0%) [11
]. Another study from Italy prospectively evaluated complications after PEG placement in 239 children at nine centers and found that 3.3% experienced major complications that required laparotomy—six gastrocolocutaneous fistulas, one intra-peritoneal hemorrhage due to colonic injury, and one peritonitis due to tube displacement [14
]. A review of 467 PEG and laparoscopic-assisted PEG placements in the Netherlands found that 12.6% of patients experienced major complications; none of these complications occurred in patients in whom laparoscopic assistance was employed [15
]. The risk of complications was significantly higher in patients with ventriculoperitoneal shunts, but no association with patient age or prior upper abdominal surgery was identified. A review of PEG placements at Boston Children’s Hospital found that 10.5% of patients experienced at least one major complication [16
]. Complications were again associated with the presence of a ventriculoperitoneal shunt, but smaller children (age <6 months) were found to be at lower
risk of complications.
PEG placement in younger children, especially infants, may be particularly challenging. However, Minar et al. described successful PEG placement in 39/40 infants with a mean gestational age of 29 weeks and mean weight of 3250 g at the time of procedure. Only one major complication—an esophageal injury—was reported, although the duration of follow-up was not stated [17
]. Most pediatric gastroenterologists can perform the endoscopy required for PEG placement in infants. However, the ability to perform PEG placement in infants without a trained gastroenterologist may be significantly limited. Of note, 56.1% of the PEG placements identified in this study were performed at a children’s hospital.
Safety concerns have been raised about PEG placement in younger children, highlighting the importance of patient selection. A prior study from our institution examined gastrostomy placement in infants less than one year of age and found that despite placement in a healthier cohort, PEG had more morbid and more costly complications, specifically a 3.8% risk of gastrocolic fistula, compared to laparoscopic gastrostomy [5
]. In addition, Petrosyan et al. reviewed gastrostomy placements in children less than five years of age, finding that the risk of major complications with PEG (3.3%) was significantly higher than that of laparoscopic gastrostomy with or without a concomitant fundoplication (0.7%) [6
]. We hypothesize that young children could be at increased risk of complications during PEG placement due to thinner tissues that are easier to transilluminate through, in particular the transverse colon or gastrocolic omentum. This difference could lead to more easily directly traversing the colon or at least the gastrocolic omentum, thereby pulling the thin colonic wall into the tract and leading to a delayed gastrocolocutaneous fistula.
This study was limited in its evaluation of outcomes. The KID captures mortality but few other outcomes, thereby limiting this study’s evaluation of potentially differential outcomes between PEG and SG. Outcomes can also vary depending on the specific technical aspects within each approach (PEG and SG), such as using fascial sutures, fluoroscopy, or transabdominal sutures [18
]. Those technical details could not be ascertained given the use of ICD-9 codes in this study. In addition, given this inability to differentiate between laparoscopic and open SG, no evaluation of temporal trends or center-level variability with regard to the specific SG method was possible. Despite these limitations, this study provides a comprehensive review of the methods recently used for gastrostomy placement in children in the United States.
While gastrostomy placement is a very common procedure in children, the best method of placement has not yet been fully established. The national distribution of PEG vs. SG has remained relatively stable over time. However, the method of placement varies significantly according to patient age, insurance type, hospital location, and hospital type. In addition, centers vary dramatically with regard to their method of gastrostomy, especially within the infant population. These findings emphasize the need to further evaluate the safest method of gastrostomy placement in children, in particular among the youngest patients in whom practice currently varies the most.