Gastroesophageal reflux disease (GERD) and dysfunctional swallowing are common problems in children with neurologic impairment (NI) (1–4). Both predispose these children to aspiration pneumonia (AP), the most common cause of death in this population (3,5–8). For children with NI and dysfunctional swallowing, gastrostomy tube (GT) feedings are often initiated to provide either supplemental or complete enteral nutrition. Some children with NI and dysfunctional swallowing go on to develop medically refractory GERD and require escalation of therapy to provide nutritional supplementation while maximizing airway protection. The two most common management strategies of medically refractory GERD are Nissen fundoplication with GT or placement of a transpyloric gastrojejunal tube (GJT) (9,10).
There is some controversy in the literature regarding the comparative effectiveness of fundoplication with GT versus GJT on GERD- and/or dysfunctional swallowing–related outcomes for children with NI who fail GT feedings. Three studies have compared these 2 treatment modalities and found no difference in either subsequent hospitalizations for AP or in mortality (3,11,12). In addition, 2 large studies have examined fundoplication effectiveness (without a comparative treatment arm) and found a decrease in overall GERD- and/or dysfunctional swallowing–related hospitalizations, though with mixed findings in the subgroup analysis for effect on AP hospitalizations. The benefits to older children, especially those with NI were, however, less clear (13,14). No similar studies examining the effectiveness of GJT have been performed, but would be useful in designing comparative effectiveness trials for fundoplication with GT versus GJT. The objective of our study was to determine the effect of GJT feedings in children with NI on GERD- and/or dysfunctional swallowing–related visits and their associated costs.
The present study was conducted at Primary Children's Hospital, a 289-bed children's hospital that serves both as the primary pediatric hospital for Salt Lake County and the tertiary referral center for Utah, Wyoming, Montana, Idaho, and Nevada. The hospital is owned and operated by Intermountain Healthcare Inc, a large vertically integrated not-for-profit health care system that operates 23 hospitals.
The present study was a retrospective cohort chart review based on a previously published administrative database study of children with NI and GERD managed with either a fundoplication or GJT feedings at the study hospital between December 1999 and October 2006 (3).
In the initial administrative database study, patients were identified using the Intermountain Healthcare Enterprise Data Warehouse, an organized, integrated, and searchable administrative database that stores more than 38 million patient encounters and includes clinical, laboratory, and radiologic data from all inpatient, and outpatient settings, using a unique identifier for each patient (3). Children meeting the following criteria were included: at least 1 hospitalization at Primary Children's Medical Center between January 1999 and October 2006; presence of at least 1 International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) code suggestive of NI (eg, cerebral palsy, developmental delay, seizure disorder) (15); diagnosis of GERD defined by ICD-9-CM code; and a first GJT placement (defined by an internal charge code) performed during the study period.
In the present study, the elements of patient selection criteria from the initial database study were confirmed using chart review as a gold standard (Fig. 1). GJT placement initially identified by the internal charge code was confirmed by reviewing procedure notes. Patients who upon chart review were found not to have undergone GJT placement or to have had a prior fundoplication were excluded. Chart review was used to augment the diagnosis of NI by capturing the presence of functional impairment in at least 1 of the following areas: mobility, communication, or self-care at the time of GJT placement. GERD was defined as the presence of symptoms and corroborated with testing at the discretion of the patient's physicians (16).
GJT placements were conversions from GT performed by a pediatric interventional radiologist, gastroenterologist, or surgeon (except for a single primary jejunostomy). This procedure involved passing a catheter through the GT site and into the jejunum, confirming placement using contrast injection, and securing the GJT by inflating a balloon within the stomach. Continuous GJT feedings were instituted following the procedure. Whereas some hospitals have protocols for routine GJT exchanges, our facility does not.
The primary outcome measures of the present study were rates of the total and GERD- and/or dysfunctional swallowing–related visits before and after GJT placement. All of the emergency department and radiology visits as well as all hospitalizations from any of the 23 Intermountain Healthcare hospitals were reviewed for each study patient from the time of birth until 1 year following their initial GJT placement. One year was selected for follow-up duration because it balanced a low risk of losing patients to follow-up while providing an adequate period for observation of a treatment effect. These visits were then classified as either not GERD and/or dysfunctional swallowing related or GERD and/or dysfunctional swallowing related. Examples of visits unrelated to GERD and/or dysfunctional swallowing included those for seizures, urinary tract infections, or traumas. GERD-and/or dysfunctional swallowing–related visits were further classified into visits because of one of 4 reasons, which were defined a priori:
- Procedure or feeding tube related: for example, GT/GJT placement, feeding tube occlusion, displacement, malfunction, fundoplication (post-GJT placement)
- Gastrointestinal (GI) studies: swallow studies, upper GI radiographic series, endoscopies, pH probes, or nuclear medicine scans
- GERD symptoms: for example, vomiting, hematemesis, regurgitation, feeding refusal
- Respiratory distress, which was then further classified into likely aspiration and/or bacterial pneumonia; or respiratory distress because of other reasons (eg, viral illness)
Complex or prolonged hospitalizations were categorized by an experienced hospitalist (M.K.) using clinical judgment as to whether the GERD- and/or dysfunctional swallowing–related complaint was the primary diagnosis responsible for the hospital stay. For instance, if a patient developed AP following an elective surgical procedure and remained in the hospital because of the pneumonia, that visit would be classified as GERD and/or dysfunctional swallowing related. If, however, a patient developed a GJT occlusion requiring a tube change during a hospitalization for another reason such as ventriculoperitoneal shunt infection, and the majority of that hospitalization was the result of the treatment of the infection, then that visit would be classified as not GERD and/or dysfunctional swallowing related.
Each category had detailed clinical definitions with inclusion and exclusion criteria to assure precision. As an example, the criteria for bacterial pneumonia required the patient to meet all of the following:
- Respiratory distress not solely because of upper airway disease
- at least 1 sign or symptom: fever, cough, difficulty breathing, increased secretions, labored breathing, tachypnea for age, nasal flaring, grunting, retractions, abdominal breathing, wheezes, or crackles
- oxygen requirement: either new or increased from baseline
- A new infiltrate or opacity on chest radiograph as determined by a radiologist
- Antibiotic course of at least 7 days, prescribed specifically to treat pneumonia rather than another infection
Costs were a secondary outcome of the present study and were derived from the hospital's cost-accounting system, the Standard Cost Master, which is an activity-based microcosting system across all Intermountain Healthcare hospitals (17–19). This system identifies and aggregates the variable and fixed cost components of patient activities, hospital services, and products.
Additional data collected included the following: sex, comorbid diagnoses, age at initial GJT placement, and reasons behind the decision to place the GJT. We also collected the data on pre-GJT feeding methods, GERD medications, fundoplications during 1 year post-GJT, and deaths. The study was approved by the institutional review board of the University of Utah Health Sciences Center.
After reviewing and categorizing the visits, we calculated the total and GERD- and/or dysfunctional swallowing–related visit rates per child per year. Visits included emergency department and radiology encounters as well as hospitalizations. The “pre” period was the period from birth until GJT placement. The “post” period was the 1-year period following GJT placement, or death within the same time period. The initial GJT placement encounter was excluded from the analysis. Rates of total and GERD- and/or dysfunctional swallowing–related costs were derived for the pre and post time periods.
Incident rate ratios (IRRs) were calculated by dividing the post-GJT visit rate by the pre-GJT visit rate. An IRR >1 meant that the encounter rate after GJT placement was more frequent than that before GJT placement. Annual rates of total and GERD- and/or dysfunctional swallowing–related visits per child year were calculated and compared using Poisson regression.
First GJT placement was confirmed using radiology documentation in 33 of 43 patients. Five patients were found to have had undergone a prior Nissen fundoplication at institutions outside the Intermountain Healthcare system and were, therefore, misclassified in the original database study (3). The other 5 patients were seen in interventional radiology, but not for GJT placement (Fig. 1).
The characteristics of our study population are described in Table 1. Median age of GJT placement was 14.3 months, interquartile range Q1 8.3 to Q3 31.7. Twenty-four percent of our study population was girls. Study patients had a variety of reasons for NI, including genetic syndromes (58%), seizure disorders (58%), structural brain abnormalities (33%), and hydrocephalus (24%). Most children had more than 1 cause for NI. All of the patients had age-adjusted confirmed presence of functional impairment in at least 1 of the following areas: mobility, communication, or self-care at the time of GJT placement.
All of the patients had a clinical suspicion for GERD based on the presence of at least 1 of the following symptoms: feeding difficulties (64%), vomiting (55%), failure to thrive (39%), and dysphagia (8%). Fifty-five percent of the patients had at least 1 positive GI test: upper GI radiographic series (55%), endoscopy with biopsy (12%), pH probe (6%), and nuclear medicine scan (3%). Seventy percent of the patients were treated with anti-GERD medications at the time of GJT placement: prokinetic agent (58%), proton pump inhibitor (42%), and H2 blocker (39%) (Table 1).
Dysfunctional swallowing and primary aspiration were suspected in 79% of patients based on the presence of symptoms such as witnessed aspiration, swallowing discoordination, or gagging, choking, or coughing with feeds. Thirty-three percent of the patients had confirmed primary aspiration on a swallow study. Ninety-one percent of the patients were dependent on enteral feedings at the time of GJT placement. Thirty-six percent of the patients were taking some oral feeds at the time of GJT placement (Table 1).
Study Intervention Details
In 23 cases, the conversion was performed for a GT already in place, whereas in 9 cases a GJT was placed as the initial GT. In 1 case, a surgical primary jejunostomy was performed. Thirty-three percent of patients had relative contraindications to fundoplication, which included delayed gastric emptying, positive pressure ventilation, microgastria, esophageal atresia, or a large abdominal wound at the GT site. In 3% of the cases, parental preference was cited as the reason behind choosing GJT over fundoplication. In 64% of the cases, there was no documented reason why GJT was selected rather than fundoplication with GT.
The total pre-GJT time period for the study cohort was 62 years, whereas the total post-GJT time period was 32 years. During the pre-GJT period, there were 230 total encounters, 89 (38.7%) of which were GERD and/or dysfunctional swallowing related. During the post-GJT period, there were 215 total encounters, 128 (59.5%) of which were GERD and/or dysfunctional swallowing related. Although the total number of encounters per child-year did not change significantly following GJT placement (2.27 vs 4, P = 0.245), the number of GERD- and/or dysfunctional swallowing–related ones increased significantly (1.02 vs 3.00, P = 0.027) (Table 2). IRRs for both total (IRR 1.78, 95% confidence interval [CI] 1.12–2.81) and GERD- and/or dysfunctional swallowing–related (IRR 2.88, 95% CI 1.68–4.94) visits increased following GJT placement, largely because of procedure- or feeding tube–related visits (IRR 5.36, 95% CI 2.73–10.51). There were no statistically significant differences in the respiratory distress, GERD symptoms, or GI studies (Fig. 2).
Within 1 year following the initial GJT placement, 26 (79%) of the patients returned for 122 visits for GJT replacements. Ninety (74%) visits were because of a primary problem with the GJT. Because GJTs are not routinely replaced at our center, all replacements were secondary to GJT problems. The most common reasons for GJT replacements were as follows: displacements 44 (36%), occlusions 31 (25%), and other 47 (39%). Other reasons included nonfunction/malfunction, damage, leakage, stricture, site cellulitis, placement check, change to a button, and change to a different size tube. The mean GJT replacements per patient per year was 3.6 (range 0–12). Six (18%) patients accounted for 52% of the visits.
During the pre-GJT period, total costs were $3,769,510, with $740,790 (19.65%) of the costs resulting from GERD-and/or dysfunctional swallowing–related visits. During the post-GJT period, total costs were $1,235,320 with $308,710 (24.99%) of the costs resulting from GERD- and/or dysfunctional swallowing–related visits. Median cost per child per year in the pre-GJT period was higher than that in the post-GJT period ($59,000 vs $24,000, P = 0.028). GERD- and/or dysfunctional swallowing–related costs per child per year in the pre-GJT period were unchanged from those in the post-GJT period, but were overall extremely low (pre-GJT, $1851 [3% of total costs] vs post-GJT costs $4601 [19% of total costs], P = 0.89) (Table 2).
Within 1 year following the initial GJT placement, 7 (21%) children underwent Nissen fundoplication (median time between GJT placement and fundoplication 2 months, range 2 days–5 months). The reasons were continued emesis (4 patients), persistent need for GJT replacements (1 patient), and no documented reason (2 patients).
Four (12%) children died within 1 year following the initial GJT placement. The median age of death was 39 months (mean 44 months, range 37–59 months). The median time between GJT placement and death was 7 months (mean 9 months, range 7–12 months). Two deaths involved jejunal perforation with resulting septic shock. The first case occurred 7 months following GJT placement. It was noted that the intestinal perforation was at the site of GJT termination. The second case occurred 12 months following GJT placement and involved a jejunal mesenteric defect with strangulation of a large portion of the small bowel. The other 2 deaths were unrelated to the GT. The death of the third child was thought to have been because of bowel obstruction. The fourth child passed away at home while under hospice care because of intractable seizures and apnea secondary to a brain tumor.
In our study, we found that there was an overall increase in total and GERD- and/or dysfunctional swallowing–related visits after children with NI received a GJT replacement to treat their GERD. The majority of these visits were the result of procedure-related GJT problems such as displacements or occlusions. Although the overall costs for this cohort were high (between $1 and $3.7 million) in both the pre- and post-GJT period, the GERD- and/or dysfunctional swallowing–related costs accounted for a smaller proportion (between $300,000 and $740,000). Overall median costs per child-year decreased following the GJT, whereas the GERD- and/or dysfunctional swallowing–related costs were unchanged, but overall low. Our data suggest that the procedure- or feeding tube–related visits, although high in number, were inexpensive and contributed to the lower costs after the GJ tube was placed. During the first year after GJT placement, a significant number of patients underwent additional fundoplication. Mortality was moderately high in our cohort.
The most common problems encountered following GJT placement in both our study and other studies were mechanical problems, predominantly displacements and occlusions (10,12,20). The number of patients presenting with GJT problems in our study (79%) was the same as the rate reported by Peters et al (20) and slightly lower than the rate reported by Wales et al (12). Our average number of GJT replacements per year (3.6) was higher than that (1.7) reported by Wales et al (12). Twenty-one percent of our patients underwent fundoplication within 1 year of the GJT placement, which is similar to the 25% reported by Peters et al (20) and higher than the 8.3% reported by Wales et al (12). The differences are likely accounted for by the shorter follow-up period in our study (12,20). The study population used by Wales et al was similar to ours in that only children with neurologic impairment and first GJT were included, whereas both Godbole et al (10) and Peters et al (20) included neurologically nonimpaired children and Godbole et al included children who underwent fundoplication before GJT placement (10). In addition, neither of these studies reported on pre-GJT placement visits or costs, which makes them of more limited value to evaluate the effectiveness of the GJT on those outcomes. Our study does identify a potential group of patients in whom there are an unusually high number of procedural visits. Studies with larger sample sizes may help identify modifiable risk factors of procedural complications. Intervention studies targeting both patient risk factors and more standardized approaches to GJ care should evaluate their effect on subsequent visits and costs.
Death has been reported in children with GJT with rates reported at 5.9% (11), 12.5% (12), 17% (20), and 21% (3) over varying time frames. Our mortality rate was 12% within the 1-year follow-up time frame. One study describes 1 death during GJT insertion (21) and another study (20) notes 1 death secondary to continued aspiration and another secondary to jejunal perforation. Although we did not note death during GJT insertion or resulting from aspiration, we did find 2 deaths that may have had intestinal perforation associated with GJT feedings. One study reports data on costs for children with NI and GJT at $31,390 mean cost per patient that includes length of hospitalization, cost of the procedure, and subsequent tube manipulation visits (12). Our costs were much lower ($4601 mean cost per child-year) even without inflated adjustment, and may be because the other study measured costs in the Canadian health care system. Although not specifically focused on just GJT or fundoplication, recent data estimate the charges for children with NI to be $17.7 billion for hospitalizations in the United States (15).
The present study was a retrospective chart review study with inherent limitations. Our sample size was small, and we were underpowered to detect clinically significant differences in the IRRs for some of the GERD- and/or dysfunctional swallowing–related visit subgroups. Despite our explicit criteria, some subjectivity when classifying prolonged or complicated hospitalizations remains. Only the primary hospitalization reason was classified. We were unable to assess how different screening or evaluation methods pre-GJT placement affect the subsequent clinical reasons for visits. Costs were aggregated rather than examined as line-item costs and attribution of costs with the reasons for visits may not be exact. Other outpatient primary care or gastroenterology or surgery clinic appointments were not captured in our study. Although encounters occurring outside our health care system were not captured, this population of medically complex children typically receives most or all of their care within the Intermountain Healthcare system. We do not have data on post-GJT testing such as intraluminal impedance testing; however, the data on the relation between testing and predicting outcomes are not clear (22). Finally, because this was a retrospective study, we were unable to follow other important outcomes such as child and caregiver health–related quality of life and GERD symptoms after GJT placement.
Despite these limitations, we were able to provide novel data regarding the effectiveness of GJT in children with NI who are being treated for GERD. This is an alternative treatment option that is used in some institutions that prefer this approach over that of a fundoplication with GT. Although studies have compared these 2 treatment options and found similar numbers of deaths and hospitalizations for AP, they have all been retrospective and limited in size (the largest group have slightly <400 patients) (3,11,12). Two studies have evaluated the effectiveness of a fundoplication and found them to be largely effective for children with NI, but with mixed results on subsequent rates of AP (13,14); however, neither study compared fundoplication with GJT.
Fundoplication with GT and GJT feedings are two possible options for managing GERD in children with NI who already have a GT. Although several investigators have called for a multicentered randomized controlled trial comparing the effectiveness of these 2 treatment modalities, it is unclear what the primary outcome should be (3,11–14,23,24). Hospitalizations and associated costs are important potential primary outcomes and our study, although small in the number of subjects, provides the only preliminary data to help power calculations for a future prospective trial. An adequately powered trial comparing morbidity, mortality, complications, costs, as well as child and caregiver quality of life will help answer the question of whether one approach is truly superior and provide adequate guidance to caregivers and physicians faced with this decision.
The authors are indebted to Shawna Baker, Laurie Coburn, Mike Dean, Dan Desroucher, Brent James, Gena Lattin, Peter Rosenbaum, Matt Samore, Nick Tsu, Chelsea Welch, and Paul Young for their help on the series of studies that led to this publication.
1. Sondheimer JM, Morris BA. Gastroesophageal reflux among severely retarded children. J Pediatr
2. Spitz L, Roth K, Kiely EM, et al. Operation for gastro-oesophageal reflux associated with severe mental retardation. Arch Dis Child
3. Srivastava R, Downey EC, O’Gorman M, et al. Impact of fundoplication versus gastrojejunal feeding tubes on mortality and in preventing aspiration pneumonia
in young children with neurologic impairment
who have gastroesophageal reflux disease
4. Sullivan PB, Lambert B, Rose M, et al. Prevalence and severity of feeding and nutritional problems in children with neurological impairment: Oxford Feeding Study. Dev Med Child Neurol
5. Chaney RH, Eyman RK, Miller CR. Comparison of respiratory mortality in the profoundly mentally retarded and in the less retarded. J Ment Defic Res
6. Marik PE. Aspiration pneumonitis and aspiration pneumonia
. N Engl J Med
7. Maudsley G, Hutton JL, Pharoah PO. Cause of death in cerebral palsy: a descriptive study. Arch Dis Child
8. Plioplys AV, Kasnicka I, Lewis S, et al. Survival rates among children with severe neurologic disabilities. South Med J
9. Fortunato JE, Darbari A, Mitchell SE, et al. The limitations of gastro-jejunal (G-J) feeding tubes in children: a 9-year pediatric
hospital database analysis. Am J Gastroenterol
10. Godbole P, Margabanthu G, Crabbe DC, et al. Limitations and uses of gastrojejunal feeding tubes. Arch Dis Child
11. Albanese CT, Towbin RB, Ulman I, et al. Percutaneous gastrojejunostomy versus Nissen fundoplication for enteral feeding of the neurologically impaired child with gastroesophageal reflux. J Pediatr
12. Wales PW, Diamond IR, Dutta S, et al. Fundoplication and gastrostomy versus image-guided gastrojejunal tube for enteral feeding in neurologically impaired children with gastroesophageal reflux. J Pediatr Surg
13. Goldin AB, Sawin R, Seidel KD, et al. Do antireflux operations decrease the rate of reflux-related hospitalizations in children? Pediatrics
14. Srivastava R, Berry JG, Hall M, et al. Reflux related hospital admissions after fundoplication in children with neurological impairment: retrospective cohort study. BMJ
15. 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
16. Rudolph CD, Mazur LJ, Liptak GS. Guidelines for evaluation and treatment of gastroesophageal reflux in infants and children: recommendations of the North American Society for Pediatric
Gastroenterology and Nutrition (NASPGHAN). J Pediatr Gastroenterol Nutr
2001; 32 (suppl 2):S1–S31.
17. Ampofo K, Gesteland PH, Bender J, et al. Epidemiology, complications, and cost of hospitalization in children with laboratory-confirmed influenza infection. Pediatrics
18. Evans RS, Classen DC, Stevens LE, et al. Using a hospital information system to assess the effects of adverse drug events. Proc Annu Symp Comput Appl Med Care
19. Harbarth S, Burke JP, Lloyd JF, et al. Clinical and economic outcomes of conventional amphotericin B-associated nephrotoxicity. Clin Infect Dis
20. Peters JM, Simpson P, Tolia V. Experience with gastrojejunal feeding tubes in children. Am J Gastroenterol
21. Friedman JN, Ahmed S, Connolly B, et al. Complications associated with image-guided gastrostomy and gastrojejunostomy tubes in children. Pediatrics
22. Rosen R, Levine P, Lewis J, et al. Reflux events detected by pH-MII do not determine fundoplication outcome. J Pediatr Gastroenterol Nutr
23. Burd RS, Price MR, Whalen TV. The role of protective antireflux procedures in neurologically impaired children: a decision analysis. J Pediatr Surg
24. Srivastava R, Downey EC, Feola P, et al. Quality of life of children with neurological impairment who receive a fundoplication for gastroesophageal reflux disease
. J Hosp Med
Keywords:© 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,
aspiration pneumonia; gastroesophageal reflux disease; gastrojejunal feedings; neurologic impairment; pediatric