National Trends and Outcomes of Pediatric Gastrostomy Tube Placement : Journal of Pediatric Gastroenterology and Nutrition

Secondary Logo

Journal Logo

Original Articles: Gastroenterology

National Trends and Outcomes of Pediatric Gastrostomy Tube Placement

Fox, David; Campagna, Elizabeth J.; Friedlander, Joel*; Partrick, David A.§; Rees, Daniel I.; Kempe, Allison

Author Information
Journal of Pediatric Gastroenterology and Nutrition 59(5):p 582-588, November 2014. | DOI: 10.1097/MPG.0000000000000468


Gastrostomy feeding tube insertion has become commonplace in the management of children with complex medical needs who require nutritional support. Procedures used for insertion are broadly divided into surgical gastrostomy tubes (G-tubes) and percutaneous endoscopic gastrostomy (PEG) tubes. PEGs were first described in 1980 when a pediatric surgeon teamed with an endoscopist to insert a feeding tube in a 4-month-old infant (1). Since then, pediatric PEGs have gained widespread acceptance, with some authors describing this procedure as the preferred method of enteral access owing to lower complication rates (2–4). In adults, PEG procedures are sometimes even done at the bedside in the intensive care unit setting for patients unable to eat secondary to an acute injury or stroke (5,6).

Surgical G-tubes have a much longer history, dating back to the late 19th century. In fact, a modified version of the Stamm gastrostomy, which was first performed in 1894, is still performed today (7). Perhaps the most important advance in surgical gastrostomy has been the use of the laparoscope during surgery, thereby increasing the efficiency and safety of the surgical gastrostomy procedure (8). G-tubes can now be inserted surgically or laparoscopically, whereas PEGs most often use gastrointestinal endoscopy. The choice between PEG and G-tube is a clinical decision made by the providers and families, but as other research in variation has shown (9–11), where a patient lives may play a role in what type of procedure is to be done.

Publicly available data on adults reveal that rates for G-tubes and PEGs have changed during the last 12 years (12), with rates of surgical G-tube insertion decreasing, whereas PEG rates are increasing, suggesting that PEGs may be replacing G-tubes in adults. Similar data in regard to pediatrics have, however, been lacking. Although estimates of national pediatric enteral access procedure rates are available (4), no previous study has analyzed trends in these rates over time, and whether the rates may be associated with improved outcomes in one procedure versus the other. The objectives of the present study were to describe trends of pediatric G-tube and PEG tube insertion over time, examine regional variation of the 2 procedures, and to compare the outcomes of inhospital infection, procedural complications, cost, and length of stay (LOS) for these 2 procedures in the pediatric population.


The analysis was based on retrospective, repeated cross-sectional data on pediatric hospital admissions obtained from the Kids’ Inpatient Database (KID) for the years 1997, 2000, 2003, 2006, and 2009. KID is an 80% sample of pediatric discharges from community, nonrehabilitation hospitals, and represents the largest collection of hospital care data in the United States. By 2009, KID included data on 7.4 million admissions from 44 states. All of the results come from survey-specific procedures that incorporate the complex sampling frame and sample discharge weights to calculate national and regional estimates. This study was approved by the Colorado Multiple institutional review board.

Study Population

We identified hospitalizations of children (<18 years) wherein the admission included G-tube placement (International Classification of Diseases-9 [ICD-9] 43.19) or PEG placement (ICD-9 43.11). Fundoplication (ICD-9 44.66 or 44.67) was an identified co-performed procedure for our analysis. Admissions with a gastrostomy-jejunostomy tube (ICD-9 44.32) or both a G-tube and PEG placement in the same admission were excluded. Age in our analysis represents age at admission, not age at time of procedure, because age at time of procedure was not available for the entire sample. Race and ethnicity are not reported separately in KID; if both are provided, ethnicity takes precedence. Diagnoses of prematurity (ICD-9 765.21–28), neurologic impairment (13), and complex chronic conditions (CCCs) (14) were used to characterize the sample and attempt to control for patient severity. Demographics and diagnostic characteristics of patients receiving a G-tube were compared with those patients receiving a PEG. Trends in diagnostic categories over time were also compared and differences were tested using regression analysis.

Population Rates

Census data were used to estimate national rates of hospitalizations for pediatric G-tube and PEG procedures for the total US population (15). These rates were also computed by region, year, and age category. Regional rates for pediatric G-tube and PEG were compared with national rates by computing the regional to national rate ratio by year.

Outcome Measures

Four outcome measures were evaluated for each admission under study: infection, surgical complication, LOS, and total cost. Outcomes for all admissions to community, nonrehabilitation hospitals were examined. To account for the influence of other procedures and diagnoses on our outcomes, we choose to examine separately outcomes for admissions wherein the G-tube or PEG was listed as the only procedure, as well as those in which it was one of a number of procedures done during that admission. For “G-tube-only” and “PEG-only” admissions, we also excluded any admission with a fundoplication.

Inhospital infections and surgical complications were identified by examination and classification of diagnostic ICD-9 codes using the surgical complication flag and the infection flag as described in the Pediatric Health Information System database. The codes from the infection and surgical complication flags were subdivided into meaningful clinical categories that we have published previously (16). We also identified abdominal procedures that may reasonably represent a complication requiring a surgical procedure following a gastrostomy procedure: exploratory laparotomy (ICD-9 54.11), reopening of laparotomy (ICD-9 54.12), incision of abdominal wall (ICD-9 54.0), or laparotomy (ICD-9 54.19).

Cost estimates, calculated using the cost-to-charge ratio, were available for 2003, 2006, and 2009. All cost data were expressed in 2009 dollars adjusted for changes in the price of hospital care and related services (17). For LOS and cost calculations we excluded admissions with an LOS of 0 (n = 137).

Adjusted Analysis

LOS and cost were modeled using log-linear multivariate regression analysis and adjusted for demographics, complexity, setting, year, and the presence of an infection or surgical complication during the admission. The log-linear specification was used because LOS and cost data were skewed. Cost was also adjusted for LOS. Multicollinearity was checked using the variance inflation factor. Results for cost and LOS were presented as the percentage difference between hospital admissions for G-tube and PEG procedures. We presented the unadjusted and adjusted results for any admission wherein a G-tube or PEG was performed and those admissions wherein G-tube or PEG was the only procedure listed. All analyses used SAS version 9.3 (SAS Institute, Cary, NC).


Study Sample Characteristics

The results of the cohort identification are shown in Figure 1. Descriptive statistics for these admissions are reported in Table 1 divided by year of discharge to show trends. The number of procedures increased in every year of the study period and the percentage of patients <1 year of age showed the same trend. Insurance status showed a trend toward more Medicaid insurance in the study population. G-tube admissions had a fundoplication during the same admission 48.5% of the time, whereas PEG admissions had a fundoplication in 8.3% of those admissions.

Cohort of discharges who underwent G-tube or PEG placement in KID during study years 1997, 2000, 2003, 2006, 2009. G-tube (ICD-9 43.19), PEG (ICD-9 43.11), and GJ-tube (ICD-9 44.32). G-tube = gastrostomy tube; GJ-tube = gastrostomy-jejunostomy tube; ICD-9 = International Classification of Diseases-9; KID = Kids’ Inpatient Database; PEG = percutaneous endoscopic gastrostomy.
Characteristics of admissions involving G-tube or PEG in children <18 years during study years

Trends in G-Tube and PEG Rates

The overall G-tube insertion rates increased from 16.6 procedures/100,000 children in 1997 to 18.5 in 2009. G-tube and PEG rates during the period under study are shown in Figure 2A, which also shows the rate at which fundoplication is combined with the G-tubes. Overall, PEG rates remained relatively stable during the study period (slope 0.05, P = 0.47), whereas G-tube rates increased from 10.0/100,000 children in 1997 to 11.9/100,000 in 2009, an average of 0.15/year (slope 0.50, P < 0.01). Of note, the rates at which fundoplication was combined with either G-tube or PEG did not show an increasing trend (G-tube with fundoplication slope −0.02 P = 0.88, PEG with fundoplication slope 0.02 P = 0.59).

A, Rate of surgical G-tube and PEG procedures in study years, KID 1997, 2000, 2003, 2006, and 2009. Rates calculated using population estimates for children 0 to 17 years from the US Census. Slope from linear regression (procedures/100,000 children/3 years)—G-tube: 0.50, P = 0.002; PEG: 0.05, P = 0.47. B, Rate of surgical G-tube procedures by age category in study years, KID 1997, 2000, 2003, 2006, and 2009. Rates calculated using population estimates by age from the US Census for each year. Slope from linear regression (procedures/100,000 children/3 years): <1 year: 7.69, P = 0.01; 1–4 years: 0.12, P = 0.37; 5–12 years: 0.02, P = 0.81; 13–17 years: −0.29, P = 0.02. G-tube = gastrostomy tube; KID = Kids’ Inpatient Database; PEG = percutaneous endoscopic gastrostomy.

Although G-tube rates among patients older than age 1 were roughly constant during the period under study, the rate increased among patients younger than 1 (Fig. 2B). Specifically, the G-tube rate for these patients increased from 96.6 G-tubes/100,000 children in 1997 to 127.3/100,000 in 2009. The rates for PEGs performed on children <1 year of age also increased, but less dramatically (40 PEGs/100,000 children in 1997 to 45.7/100,000 in 2009).

To better understand the increase in G-tube rates, the diagnoses of children undergoing G-tube procedures were examined. Using CCCs as a method of categorizing all of the listed diagnoses, a number of patterns emerged for admissions with a G-tube. During the period under study, the percentage of children with a cardiovascular chronic condition nearly doubled (from 17.4% in 1997 to 34.0% in 2009, P < 0.001); those children born prematurely also increased (11.5% in 1997 to 15% in 2009, P < 0.001) and there was an increase in the percentage of children with a genetic defect (from 18.9% in 1997 to 25.8% in 2009, P < 0.001). The percentage of children with neuromuscular conditions did not increase, although these children made up a substantial portion of the sample throughout the period under study.

Regional Variation

Admissions were divided based on the region of the country (Northeast, South, Midwest, and West), and the regional variation in the use of G-tubes and PEG was examined (Fig. 3). During the period under study, the West showed higher rate ratios for both procedures when compared with the national average, and the Midwest showed lower rate ratios for both procedures. When the sample was restricted to admissions of children <1 year of age for a G-tube, a similar pattern emerged: specifically, the West had a higher G-tube rate than the national rate, whereas the Midwest had a lower G-tube rate.

Percentage difference in the rate of surgical G-tube and PEG by region compared with the national rate, KID 1997, 2000, 2003, 2006, and 2009. Rates calculated using population estimates for children 0 to 17 years by region from the US Census for each year. G-tube = gastrostomy tube; KID = Kids’ Inpatient Database; PEG = percutaneous endoscopic gastrostomy.

To examine the effect of fundoplication upon our regional G-tube rates, we calculated the rate at which concomitant fundoplication was performed with G-tube in each region: Northeast 40%, Midwest 44%, West 47%, and South 55%. Because of these regional differences in fundoplication, we examined the trend of concomitant fundoplication in each region and found no correlation between concomitant fundoplication and regions with increasing G-tube rates. In the 2 regions where G-tube rates increased, we found that the rates of G-tube with concomitant fundoplication did not increase during the same time period (South slope 0.42, P = 0.21; Midwest slope 0.01, P = 0.97).

Outcomes: Inhospital Complications

Rates of inhospital infection and surgical complication are presented in Table 2. Overall rates infection did not differ between the 2 procedures, but when we limited the analysis to those admissions in which only the procedure in question was performed, differences emerged. The overall in-hospital infection rate for G-tube-only admissions was lower than those for PEG-only admissions and the subcategories of pneumonia, viral infection, and cellulitis were lower for G-tube-only admissions. Overall, inpatient surgical complication rates for G-tube-only and PEG-only admissions did not differ.

Infection and surgical complications of admission involving G-tube or PEG in children <18 years during study period

We attempted to identify those children who underwent operative exploration after gastrostomy using laparotomy ICD-9 codes. Of all gastrostomy procedures, <2% (n = 1120) had a laparotomy during the admission. Of the laparotomy procedures identified, the date of the laparotomy relative to the G-tube under study was not available in 50% of cases, and therefore we were unable to draw conclusions about these data.

Outcomes: LOS and Cost

Unadjusted outcomes by procedure type are reported in the left-hand portion of Table 3. G-tube was associated with more days spent in the hospital and higher costs than PEG, and those differences remained when we examined the G-tube-only and PEG-only admissions. Adjusted associations between procedure type and the outcomes under study are reported in the right-hand portion of Table 3. After adjusting for patient demographics and other factors such as neurologic impairment, G-tube-only admissions were associated with a 19% longer LOS and a 25% higher cost.

Associations between procedure type (G-tube vs PEG) and outcomes under study


Surgical G-tube insertion rates during the last decade have increased in children <1 year of age. Children with cardiovascular, respiratory, and genetic CCCs represent an increasing proportion of those receiving G-tubes, and there is pronounced regional variation in how often each enteral access procedure is performed. Admissions primarily for G-tube insertion were associated with a higher cost and longer LOS when compared with PEG admissions.

The finding of increasing G-tube rates is similar to the finding of Berry et al (18), who found an increase in the number of gastrostomy procedures during the same time period, and consistent with Simon et al (19), who found that CCCs increased the odds of having a G-tube placed. The present study confirms that of those admissions during which a G-tube is inserted, a larger percentage are associated with CCCs. It is possible that in the early part of the period under study, children with complex medical problems may have been managed with nasogastric tube feedings or longer trials of oral feeding therapy (20). Published national guidelines in 2006 for children with neurologic impairment state that nasogastric feedings be used for “short-term” support and that gastrostomy feeding initiated in the first year of life offer improved anthropometric outcomes (20). These guidelines may have played a role in the increased number of G-tubes, although they would not account for changes seen in the earlier part of the study period. It is possible that the threshold for recommending or performing gastrostomy procedures has changed, and that less invasive methods of feeding chronically ill children are being abandoned.

This study found that the majority of the increase in G-tubes is in children <1 year of age who are admitted and receive a G-tube, not the more common PEG procedure seen in adults (21,22). The relatively recent introduction of the laparoscopic-assisted G-tube may explain the increased rates of G-tube insertion in children (8,23). The laparoscope has made the G-tube procedure less invasive and perhaps more attractive to surgeons and parents (24). The relatively thin abdominal wall of the infant provides easy access to the stomach through a small incision used to perform the laparoscopic G-tube (25). The size of the laparoscopes available makes their use in infants <10 kg manageable, with some authors touting their use in children <5 kg (26). The same may not be true for endoscopes used for PEG insertion. Some pediatric gastroenterologists choose to perform PEG procedures primarily on larger patients with no previous abdominal surgeries or spinal deformities. Some centers have opted to refrain from pediatric gastroenterologist PEG placement owing to perceived increased risk of complications and the need for non–low-profile initial gastrostomy device with subsequent poor satisfaction. Unfortunately, there is no specific ICD-9 code for laparoscopic G-tube insertion, and therefore the data did not allow for an examination of the association between the laparoscope and increased G-tube rates.

Fundoplication was often combined with G-tubes in our sample, but was rarely combined with PEGs. Our finding of varying fundoplication rates between regions is consistent with another review (9), but we found no association between concomitant fundoplication and the rising G-tube rates nationally or regionally. Fundoplication is a surgical treatment for gastroesophageal reflux disease, and there is a significant center-to-center variation in how often the procedure is done (16). It is possible that the perceived need to treat or prevent gastroesophageal reflux is part of the decision to use G-tube over PEG procedures. Physicians may feel that a G-tube combined with a fundoplication is a more efficient treatment strategy (27), rather than performing a PEG, and at some point in the future performing a fundoplication for gastroesophageal reflux that may develop. Patient-specific longitudinal data would be required to confirm this hypothesis.

One explanation for the observed regional differences could be patient-level differences, but an alternative explanation relates to physician influence. Other literature on regional or small area variation indicates that decisions such as feeding tube insertion may be a preference sensitive decision, or one that varies, not according to the evidence or guidelines, but according to the preferences of providers (11,28). For a child at a particular institution the choice between procedures may have to do with the referral patterns, skill sets, and availability of resources and qualified physicians. The regional data presented may offer a macroscopic look at the diffusion of training and knowledge. Pediatric surgeons coming out of training presently are adept at the use of the laparoscope, and if the laparoscopic approach is playing a role in the increase in G-tube rates, it is possible that centers located in the West obtained those skills and infrastructure first, and the rest of the nation is catching up. We, however, do not have data to support the notion that the relative increase in G-tubes is an evidence- or outcome-based process. We hope that our data add to the discussion about the need for an examination of appropriate national outcomes measures for G-tube placement in children.

Our data on cost and LOS show better outcomes for PEG over G-tube when the sole procedure done during the admission was a gastrostomy procedure. These data are consistent with other studies in adults comparing the 2 procedures (2,29,30). It is important to note that these cost and LOS data are confined to 1 admission and we do not have data on any follow-up visits. Although recently 1-step placement of low-profile tubes has been described (31), traditionally, the placement of a PEG tube involves the sutureless insertion of a long tube that must mature before a low-profile tube (eg, MIC-KEY [Kimberly-Clark, Roswell, GA]) can be placed. Any follow-up visits for tube replacement that may include fluoroscopic confirmation and possible anesthesia/sedation for primary tube removal were not available for this analysis.

Any analysis of an administrative dataset is susceptible to coding error; however, these errors are expected to be random and, therefore, not systematically related to the outcomes under study. The data on pediatric hospital admissions were available for only 5 years (1997, 2000, 2003, 2006, and 2009), restricting the analysis of trends. The retrospective nature of the data prevented randomization, and, as a consequence, selection bias was possible. Individual patients may have been represented more than once in the dataset because patient identifiers were not available. The codes we used for classification were broad and did not provide detailed description of how the procedure was performed or the type of subspecialist that may have performed the procedure. For example, it is likely that our G-tube code includes both open and laparoscopic procedures that may have different outcomes (16). Similarly, our PEG procedures likely include those where a traditional PEG tube is placed and those where a newer low-profile PEG button is placed, and those outcomes may differ (31). The analysis performed does not include important outcome measures including patient (or parent) satisfaction, nutritional status, and outpatient complications. Furthermore, the cost data presented does not include physician charges, which may differ between G-tube and PEG insertion. The data, however, presented represent a starting point for the discussion of national outcomes for a common pediatric procedure.

Surgical G-tube rates in children <1 year of age are increasing, and the reasons for the increase may have to do with a changing population that needs more enteral access or a changing threshold for performing the procedure. Regional variation accounts for some of the increase gastrostomy procedures, but we cannot determine whether the increase in G-tube procedures represents innovation associated with improved health outcomes or the preferences of physician providers and families. Prospective studies examining decision making and health outcomes, while better controlling for case mix and posthospitalization events, are needed to better clarify the reasons for the observed trends.


The authors thank the Children's Outcomes Research Program for their assistance in supporting this research.


1. Gauderer MW, Ponsky JL, Izant RJ Jr. Gastrostomy without laparotomy: a percutaneous endoscopic technique. J Pediatr Surg 1980; 15:872–875.
2. Apelgren KN, Zambos J. Is percutaneous better than open gastrostomy? A clinical study in one surgical department. Am Surg 1989; 55:596–600.
3. Gauderer MW, Stellato TA. Gastrostomies: evolution, techniques, indications, and complications. Curr Probl Surg 1986; 23:657–719.
4. Gauderer MW. Percutaneous endoscopic gastrostomy and the evolution of contemporary long-term enteral access. Clin Nutr 2002; 21:103–110.
5. Carrillo EH, Heniford BT, Osborne DL, et al. Bedside percutaneous endoscopic gastrostomy. A safe alternative for early nutritional support in critically ill trauma patients. Surg Endosc 1997; 11:1068–1071.
6. Yarmus L, Gilbert C, Lechtzin N, et al. The safety and feasibility of interventional pulmonologists performing bedside percutaneous endoscopic gastrostomy tube placement. Chest 2013; 144:436–440.
7. Minard G. The history of surgically placed feeding tubes. Nutr Clin Pract 2006; 21:626–633.
8. Jones VS, La Hei ER, Shun A. Laparoscopic gastrostomy: the preferred method of gastrostomy in children. Pediatr Surg Int 2007; 23:1085–1089.
9. Barnhart DC, Hall M, Mahant S, et al. Effectiveness of fundoplication at the time of gastrostomy in infants with neurological impairment. JAMA Pediatr 2013; 167:911–918.
10. Finlayson SR, Birkmeyer JD, Laycock WS. Trends in surgery for gastroesophageal reflux disease: the effect of laparoscopic surgery on utilization. Surgery 2003; 133:147–153.
11. Wennberg J, Gittelsohn. Small area variations in health care delivery. Science 1973; 182:1102–1108.
12. Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project. Nationwide inpatient sample. Accessed March 20, 2013.
13. Lasser MS, Liao JG, Burd RS. National trends in the use of antireflux procedures for children. Pediatrics 2006; 118:1828–1835.
14. Feudtner C, Christakis DA, Connell FA. Pediatric deaths attributable to complex chronic conditions: a population-based study of Washington State, 1980–1997. Pediatrics 2000; 106:205–209.
15. US Census Bureau. Population estimates. Published and accessed August 15, 2013.
16. Fox D, Morrato E, Campagna EJ, et al. Outcomes of laparoscopic versus open fundoplication in children's hospitals: 2005–2008. Pediatrics 2011; 127:872–880.
17. Bureau of Labor Statistics, Department of Labor. Archived Consumer Price Index Detailed Report Information. Published 2010. Accessed September 11, 2011.
18. Berry JG, Poduri A, Bonkowsky JL, et al. Trends in resource utilization by children with neurological impairment in the United States inpatient health care system: a repeat cross-sectional study. PLoS Med 2012; 9:e1001158.
19. Simon TD, Berry J, Feudtner C, et al. Children with complex chronic conditions in inpatient hospital settings in the United States. Pediatrics 2010; 26:647–655.
20. Marchand V, Motil KJ. Nutrition support for neurologically impaired children: a clinical report of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr Gastroenterol Nutr 2006; 43:123–135.
21. Mendiratta P, Tilford JM, Prodhan P, et al. Trends in percutaneous endoscopic gastrostomy placement in the elderly from 1993 to 2003. Am J Alzheimers Dis Other Demen 2012; 27:609–613.
22. Schrag SP, Sharma R, Jaik NP, et al. Complications related to percutaneous endoscopic gastrostomy (PEG) tubes. A comprehensive clinical review. J Gastrointestin Liver Dis 2007; 16:407–418.
23. Rothenberg SS, Bealer JF, Chang JH. Primary laparoscopic placement of gastrostomy buttons for feeding tubes. A safer and simpler technique. Surg Endosc 1999; 13:995–997.
24. Wadie GM, Lobe TE. Gastroesophageal reflux disease in neurologically impaired children: the role of the gastrostomy tube. Semin Laparosc Surg 2002; 9:180–189.
25. Sandoval JA, Partrick DA. Advances in the surgical management of gastroesophageal reflux. Adv Pediatr 2010; 57:373–389.
26. Rothenberg SS, Chang JH, Bealer JF. Experience with minimally invasive surgery in infants. Am J Surg 1998; 176:654–658.
27. Partrick DA. Gastrointestinal tract feeding access and the role of fundoplication in combination with gastrostomy. Curr Opin Pediatr 2007; 19:333–337.
28. Goodman DC. Unwarranted variation in pediatric medical care. Pediatr Clin North Am 2009; 56:745–755.
29. Bankhead RR, Fisher CA, Rolandelli RH. Gastrostomy tube placement outcomes: comparison of surgical, endoscopic, and laparoscopic methods. Nutr Clin Pract 2005; 20:607–612.
30. Grant JP. Comparison of percutaneous endoscopic gastrostomy with Stamm gastrostomy. Ann Surg 1988; 207:598–603.
31. Novotny NM, Vegeler RC, Breckler FD, et al. Percutaneous endoscopic gastrostomy buttons in children: superior to tubes. J Pediatr Surg 2009; 44:1193–1196.

gastrostomy; Kids’ Inpatient Database; percutaneous endoscopic gastrostomy; surgical outcomes

© 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,