Congenital diaphragmatic hernia (CDH) occurs with an incidence of 1:2500–3000 live births (1,2). With advancements in present therapy, infants with CDH have had improving survival rates (3). Unfortunately, patients who survive the newborn period often experience many gastrointestinal-related morbidities including gastroesophageal reflux (GER), failure to thrive (FTT), and oral aversion (4). GER has been reported as the most common short- and long-term complication in patients with CDH (5). Adult data have shown that CDH-related reflux is associated with esophagitis in both symptomatic and asymptomatic individuals, resulting in esophageal stricture formation and Barrett esophagus (6). Although the pathophysiology of GER in this population is controversial, it is hypothesized that the anatomic shift of the foregut during the development of the CDH anomaly is responsible (7). FTT is another common cause of morbidity and GER is likely a contributing factor, in addition to other comorbidities such as oral aversion, chronic lung disease, and pulmonary arterial hypertension (PAH). PAH owing to lung hypoplasia is often so severe as to require the use of extracorporeal membrane oxygenation (ECMO) to support a patient after birth and before surgical correction. Many CDH survivors with significant FTT and feeding difficulties require feeding tube placement (8). Long-term data regarding predictors of FTT and feeding disorders after the newborn period in this population are lacking. This study presents data from a multidisciplinary provider team for the purpose of identifying predictors of low weight and the need for tube feedings at 1 year of age in patients with CDH.
A retrospective review of the electronic medical records of all neonatal intensive care unit (NICU) patients with CDH treated in a tertiary care pediatric hospital from January 1, 2007 to June 30, 2012 was conducted. The time period was chosen because of the availability of electronic medical records for this period and to obtain patient follow-up data up to 1 year of age. Data were also collected from the Diaphragmatic Hernia Research and Exploration Advancing Molecular Science (DHREAMS) study and chart review. DHREAMS is a prospective study of the genetic causes and natural history of CDH (9). All of the patients were identified for the database by physician notification upon patient or maternal hospital admission or prenatal diagnosis. Patients were studied after NICU discharge in a multidisciplinary CDH clinic that included a pediatric surgeon, cardiologist, gastroenterologist, pulmonologist, and a developmental pediatrician. All of the patients underwent early intervention evaluations and were referred for therapy (including feeding therapy) as deemed necessary.
Medical records were examined for the use of ECMO and to check whether the patient had PAH. The degree of PAH was graded 1 through 4 (none, mild, moderate, or severe) between 2 and 6 months of age; the grade was determined by consensus echocardiogram interpretations conducted by 2 expert cardiologists using various measures, as previously described by our center (9). Patients with moderate and severe PAH (grade 3 or 4) were considered to have clinically significant PAH and were measured as a single entity in the statistical analysis.
The medical records were also examined for pH probe results (if performed before NICU discharge). The American Academy of Pediatrics (AAP) Clinical Report on CDH recommends that a test for esophageal reflux be considered in all patients with CDH before discharge from the NICU (10). pH probe testing was performed at the clinical discretion of the NICU and surgical services during the study period. All of the pH probes were performed while patients were off antacid medications for at least 48 hours before the start of the study and meal times were excluded from the analysis. Determination of a positive pH study took into account the total number of reflux episodes, the length of the longest reflux episode, the number of long reflux episodes (lasting >5 minutes), and the reflux index as described in pediatric guidelines (11).
Follow-up data for each patient were collected at 1 year of age, which included weight z scores, taking into account the patient's corrected gestational age (if born prematurely) and sex, the need for tube feedings (gastric or jejunal), and esophageal biopsy findings (if endoscopy was performed). Low weight was defined by a weight <5th percentile for sex and age (corresponding to a z score <−1.645). Endoscopy is recommended by the AAP Clinical Report in patients with symptoms of reflux with CDH and was performed at the clinical discretion of a patient's primary gastroenterologist (10). Some patients underwent endoscopy if there was concern for silent reflux without typical symptoms. An endoscopy was considered positive for reflux esophagitis if any of the following were noted grossly or pathologically: esophageal mucosal breaks or ulceration, reactive changes, and/or basal cell hyperplasia (11).
The χ2 test, the Fisher exact test, and logistic regression were used for statistical analysis. A P <0.05 was considered statistically significant. The study was approved by the institutional review board of Columbia University.
From January 1, 2007 to June 30, 2012, 92 children were cared for in this pediatric hospital with CDH. Three of these patients were excluded because the diaphragmatic hernia was diagnosed after 3 months of age, and these 3 patients were therefore not cared for in the NICU. Of the 89 remaining, 72 (81%) survived up to 1 year (as shown in Table 1). The cause of death for the 17 deceased patients included cardiorespiratory failure (16 patients) and clinical sepsis (1). Only 3 patients did not return for follow-up visits through 1 year of age. Two patients had moved out of the country, and 1 patient had a complex social environment and was subsequently lost to this center. Therefore, 69 patients were considered in the final 1-year tube feeding analysis. One patient did not have a documented weight at the 1-year visit, and, as a result, 68 patients were included in the final weight analysis. The cohort had an equal distribution of boys and girls, with a predominance of left-sided diaphragmatic herniation (Table 1).
Of the 72 survivors to 1 year of age, the average gestational age at birth was 38.1 ± 1.7 weeks, the average birth weight was 3070.2 ± 577.3 g, and the average day of life at the time of the surgical repair was 4.4 ± 3.4 days. Ten patients (13.9%) were born with a weight <5th percentile (corrected for sex and gestational age). At 1 year of age, however, 2 of these 10 patients increased percentiles to become a normal weight and 16 additional patients born with a normal weight dropped to a weight <5th percentile. A total of 24 patients (35%) were therefore <5th percentile for weight, which was corrected for age and sex at 1 year of age. The change in weight z score was also examined, and although it was decreased overall, the change in z score from birth (−0.68 ± 1.01) to 1 year of age (−1.07 ± 1.4) was not statistically significant.
As shown in Table 1, there were 13 patients (18.1%) who survived up to 1 year of age who needed ECMO. Of these 13, 8 (61.5%) had a weight <5th percentile at 1 year of age. PAH diagnosed by echocardiogram was present to some degree in 45 of 72 patients (62.5%) on the early echocardiogram, which was performed between 14 and 45 days of life. Of the 55 patients who received a repeat echocardiogram between 2 and 6 months of life, only 4 (7.3%) had persistent, significant PAH (moderate or severe with a grade of 3 or 4) and 2 of these 4 patients (50%) had a weight <5th percentile at 1 year of age. Of note, all of the patients who did not undergo a repeat echocardiogram between 2 and 6 months of life had no or mild PAH on the initial echocardiogram. Despite the high percentage of patients with ECMO and patients with PAH between 2 and 6 months of life who had weight <5th percentile at 1 year of age, no statistically significant correlation was found between either ECMO or PAH and low weight at 1 year of age. Of the patients who required ECMO, the average number of days receiving ECMO was 8.1 ± 2.9 days. No association was found between the length of time receiving ECMO and low weight at 1 year of age. Among the 52 patients (77%) who had neither a history of ECMO nor significant PAH, 20 (38%) had a weight <5th percentile for corrected age and sex at 1 year of age.
Of the 69 patients with at least 1-year follow-up, 13 patients (18.8%) required tube feedings at 1 year of age. The indication for tube feedings was inability to take in adequate oral nutrition or wean off nasogastric tube feedings owing to oral aversion or feeding intolerance attributed to presumed reflux and/or dysmotility. More than two-thirds of the patients receiving tube feedings were being fed via a gastrostomy tube (GT) with or without a fundoplication, and the remaining patients were fed by either a gastrojejunostomy tube or a jejunostomy tube (Table 1). As shown in Table 2, CDH survivors who needed ECMO were 6 times more likely to require tube feedings at 1 year of age compared with those who did not receive ECMO. There were only 4 patients with moderate or severe PAH between 2 and 6 months of age, but this group was nearly 16 times more likely to receive tube feedings at 1 year of age compared with those with no or mild PAH (P = 0.02, confidence interval 1.5–171.2; Fisher exact test).
To check for confounding of the 2 predictors (ECMO and PAH), multivariable logistic regression was performed (Table 3). The odds of having a feeding tube at 1 year of age were increased by >10 times in the patients who required ECMO compared with those who did not, when controlling for the degree of PAH between 2 and 6 months of age. The odds of having a feeding tube at 1 year of age were increased by 26 times in the patients who had moderate or severe PAH between 2 and 6 months compared with those who had no or mild PAH when controlling for a history of ECMO. Only 1 patient who required ECMO also had significant PAH between 2 and 6 months of age; all of the others who needed ECMO had no or mild PAH between 2 and 6 months of age. The odds of needing a feeding tube at 1 year of age if both ECMO was required and PAH was moderate or severe was increased by 36 times.
Of the 25 patients who underwent pH probe testing before initial hospital discharge, 10 (40%) were abnormal. There was no association found between abnormal NICU pH probe and weight <5th percentile at 1 year of age. None of the patients with normal pH probes in the NICU required tube feedings at 1 year of age. Eight of the patients who received neonatal pH probes have undergone an upper endoscopy to date. In all of the cases, a normal NICU pH probe correlated with a normal subsequent endoscopy, whereas an abnormal NICU pH probe correlated with pathologic signs of reflux on subsequent endoscopy. Three patients with a history of ECMO had a pH probe and two-thirds were abnormal. Two patients with significant PAH had a pH probe and both were abnormal.
Patients with CDH received a variety of different medications during the study. Most patients were on treatment for reflux with a proton-pump inhibitor or histamine-2 receptor antagonist. In addition to the antacid therapy at 1 year of age, 1 patient was also receiving metoclopramide for the treatment of emesis; 1 was receiving polyethylene glycol-3350 for constipation; 4 patients were taking medications for PAH, which included sildenafil and digoxin; and 3 patients were taking inhaled budesonide for asthma.
Owing to medical and surgical advancements, patients with CDH are now more likely to survive beyond infancy, and the long-term complications related to the disease remain to be fully studied (3). The present study is one of a few to evaluate patients through 1 year of age and to specifically evaluate predictors of low weight and the need for tube feedings. Poor weight gain was found in 35% of patients with weights <5th percentile at 1 year of age, despite >86% of patients being born with a weight >5th percentile when corrected for sex and gestational age. Others have described an increased incidence in FTT in patients with CDH (4), although the mechanism has not been definitively determined. Gastrointestinal conditions such as esophageal reflux disease and dysmotility may be partially responsible, although it is also possible that these patients have a genetic susceptibility for small size or that they have an increased metabolic requirement owing to cardiorespiratory disease. Oral aversion, as a result of a complicated medical course in infancy, and esophageal reflux could also play a role. The etiology behind the low weight is likely multifactorial and determining the mechanisms requires further study.
The use of ECMO and significant PAH found between 2 and 6 months of age were risk factors for needing tube feedings at 1 year of age. Of note, this study found that the later echocardiogram, performed between 2 and 6 months of age, was a better predictor of the need for tube feedings compared with the echocardiogram performed near 1 month of age. This may have been because most PAH found on the first echocardiogram subsequently resolved. The use of ECMO and significant PAH were not associated with an increased risk for having a weight <5th percentile at 1 year of age, as was originally hypothesized. Sixty-one percent of patients receiving ECMO and 50% of patients with significant PAH at 2 and 6 months of age had a weight <5th percentile at 1 year of age, suggesting that the lack of association may be related to the high overall percentage of low weight in patients with CDH. One could also postulate that the lack of association was because patients with PAH and ECMO were deemed at a higher risk for feeding problems by their caregivers and were, therefore, more closely monitored and more likely to have feeding tubes placed, which could prevent low weight. This did not prove to be true, however, because of the 13 children who were using tube feedings at 1 year of age; 8 of these tube-fed patients (61.5%) remained <5th percentile for weight at 1 year of age. Although the details of feeding advancement and caloric intake were not available for analysis in this study, the lack of adequate weight gain in the tube-fed population could be attributed to either inadequate caloric supplementation or intolerance of adequate feeding. In the experience of this center, patients with CDH in general, and particularly those requiring tube feedings, often have difficulty tolerating adequate feeds owing to reflux symptoms, distention, or discomfort. Further prospective studies, including reflux and possibly motility testing, are, however, necessary to prove this observation.
The study was limited by the number of patients overall and the 3 patients who were lost to follow-up. This, however, remains one of the largest cohorts of patients with CDH studied. Only 4 patients in this cohort had significant and persistent PAH. Despite this, a statistically significant association between PAH and need for tube feedings at 1 year of age was found, which has not been previously described. A cohort with a larger population of patients with significant PAH would be required to further validate this finding. Although few patients underwent pH probe testing in this cohort, pH probe data in the literature in the neonatal CDH population is limited. No association was found between having an abnormal pH probe and having a weight <5th percentile at 1 year of age, which implies that reflux may not be responsible for subsequent FTT in this population. The authors believe that this conclusion cannot be made, however, because few patients in this cohort were able to undergo pH probe testing before hospital discharge owing to the use of continuous nasogastric tube feedings or respiratory compromise requiring continuous positive airway pressure; however, of note, none of the patients with normal pH probe studies in the NICU required tube feedings at 1 year of age, suggesting that not having pathologic reflux in the neonatal period may be protective. Again, a larger number of patients with CDH undergoing pH probe testing would be necessary to validate these findings and to determine whether pH probe testing could be used to stratify patients at risk for chronic reflux and FTT for the long term.
Despite improved survival in the CDH population for the last 2 decades, long-term morbidities such as reflux, feeding intolerance, and poor weight gain remain a significant problem. This study demonstrates that a history of ECMO and significant PAH are associated with an increased risk of requiring tube feedings at 1 year of age and that patients with CDH in general are at increased risk for low weight. Although the present study is one of the few to collect data on patients through 1 year of age, what is clearly needed are longer-term data that are prospectively collected on patients through childhood and early adulthood, including systematic documentation of reflux symptoms, standardized testing for reflux, oral aversion, growth curves, and cardiac and pulmonary comorbidities. A prospective evaluation would require a multicenter approach to achieve adequate power and is the only proper method for evaluating growth outcomes and reflux-related morbidities such as esophageal strictures and Barrett esophagus. Present data are limited largely because of the retrospective nature of the present literature, relatively small sample sizes, and lack of a standardized clinical approach. The need for reflux medications, pH probe tests, and endoscopy therefore are largely based on clinical discretion rather than on strong evidence. A better understanding of the long-term complications and corresponding risk factors related to CDH would allow for the development of more effective preventive measures to improve upon the available recommendations provided by the AAP Clinical Report from 2008 (10).
1. Torfs CP, Curry CJ, Bateson TF, et al. A population-based study of congenital diaphragmatic hernia. Teratology
2. Di Pace MR, Caruso AM, Farina F, et al. Evaluation of esophageal motility and reflux in children treated for congenital diaphragmatic hernia with the use of combined multichannel intraluminal impedance and pH monitoring. J Pediatr Surg
3. Boloker J, Bateman DA, Wung J-T, et al. Congenital diaphragmatic hernia in 120 infants treated consecutively with permissive hypercapniea/spontaneous respiration/elective repair. J Pediatr Surg
4. Bagolan P, Morini F. Long-term follow up of infants with congenital diaphragmatic hernia. Semin Pediatr Surg
5. D’Agostino JA, Bernbaum JC, Gerdes M, et al. Outcome for infants with congenital diaphragmatic hernia requiring extracorporeal membrane oxygenation: the first year. J Pediatr Surg
6. Vanamo K, Rintala RJ, Lindahl H, et al. Long-term gastrointestinal
morbidity in patients with congenital diaphragmatic defects. J Pediatr Surg
7. Stolar CJ, Levy JP, Dillon PW, et al. Anatomic and functional abnormalities of the esophagus in infants surviving congenital diaphragmatic hernia. Am J Surg
8. Muratore CS, Utter S, Jaksic T, et al. Nutritional morbidity in survivors of congenital diaphragmatic hernia. J Pediatr Surg
9. Wynn J, Krishnan U, Aspelund G, et al. Outcomes of congenital diaphragmatic hernia in a modern era of management. J Pediatr
10. American Academy of Pediatrics Section on Surgery, American Academy of Pediatrics Committee on Fetus and Newborn, Lally KP, et al. Postdischarge follow-up of infants with congenital diaphragmatic hernia. Pediatrics
11. Vandenplas Y, Rudolph CD, Di Lorenzo C, et al. North American Society for Pediatric Gastroenterology Hepatology and Nutrition, European Society for Pediatric Gastroenterology Hepatology and Nutrition. Pediatric gastroesophageal reflux clinical practice guidelines: joint recommendations of the North American Society for Pediatric Gastroenerology, Hepatology, and Nutrition (NASPGHAN) and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN). J Pediatr Gastroenterol Nutr