Rates and Risks of Gastrostomy Tubes in Infants With Cleft Palate

Objectives  To review data on a cohort of infants with cleft palate to (1) determine rates of gastrostomy (G)-tube placement, (2) identify contributing comorbidities, and (3) use relative risk analyses to improve management of cleft palate in infants with feeding difficulty.

Design  Retrospective medical record review.

Setting  Tertiary care children's hospital.

Patients  Infants with cleft palate born between January 1, 2000, and December 31, 2008, without G-tubes prior to referral were included. Comorbidities were analyzed, including syndromes and chromosomopathies (syn/chrom) and cardiac, respiratory, neurologic, and gastrointestinal diagnoses. These comorbidities were analyzed independently.

Main Outcome Measure  Gastrostomy-tube placement.

Results  Of 214 infants with cleft palate, 34 required G-tubes. Of these, 19 had syn/chrom. Independent of these diagnoses, 17 infants had 1 system comorbidity and 12 had multisystem comorbidities. Of the 180 patients without G-tubes, 20 had syn/chrom. Independent of these diagnoses, 10 infants had 1 system comorbidity and 2 had multisystem comorbidities. Rates of G-tube placement ranged from 3% in infants without any comorbidity to 94% in infants with respiratory comorbidity. Relative risks of G-tube placement with syn/chrom, 1 system comorbidity, and multisystem comorbidities were 5.68 (95% confidence interval, 3.18-10.16), 21.79 (8.76-54.17), and 29.66 (12.18-72.21), respectively.

Conclusions  Diagnosis of syn/chrom or major comorbidity significantly increases risk of G-tube placement. Regardless of syn/chrom association, problems affecting the heart, respiratory system, central nervous system, and lower esophageal sphincter are the most significant risk factors, implying that particular comorbidities are more influential than a simple diagnosis of syn/chrom. These data should help identify children at greatest risk for G-tubes and those expected to overcome feeding difficulties, leading to more persistent use of nonsurgical therapy before resorting to G-tubes.

Cleft lip and palate are congenital defects that can occur alone or be associated with a wide variety of syndromes and other abnormalities. As infants await surgical correction of cleft palate, proper nutrition and growth must be maintained. Children with cleft palate have trouble generating an adequate suck because of the palatal defect. Poor suction results from insufficient seal formation between the nasopharynx and oropharynx, preventing enough negative pressure for normal food intake and bolus formation in the oral stage of swallowing.¹ At the very least, anterior or posterior closure of the oral cavity is necessary.²

Despite multiple studies attempting to quantify measures of feeding function with standardized scales, such as the Neonatal Oral Motor Assessment Scale and Great Ormond Street Measurement of Infant Feeding, as well as videofluoroscopy and electromyograms,³ there remains to be developed a good and practical protocol to determine feeding strategies. Both objectively and subjectively, the degree of difficulty in feeding infants with cleft abnormalities depends on the degree of the deformity. It is well documented that syndromic infants require more substantial feeding interventions, leading to higher rates of enteric feeding (Table 1). The rates given in Table 1 were collected incidentally by the authors of each study while examining other aspects of feeding in infants with cleft abnormalities. The purpose of our study was to (1) determine rates of gastrostomy (G)-tube placement, (2) identify contributing comorbidities, and (3) use relative risk analyses to improve management of cleft palate in infants with feeding difficulty.

The institutional review board approved the medical record analysis used for this study. Medical records of patients monitored by a multidisciplinary cleft team at the Children's Hospital and Clinics of Minnesota–Minneapolis site were retrospectively reviewed. All patients with cleft palate born at this tertiary care children's hospital between January 1, 2000, and December 31, 2008, were identified and the following inclusion criteria were used: first examination in infancy (<1 year of age, as determined by available consult notes, summary cleft clinic notes, and operative reports), no prior G-tube feeds, unrepaired cleft palate, and eventual repair by a surgeon on our cleft team. The minimum defect required for inclusion was cleft palate. There was no distinction made between infants with isolated cleft palate and those with both cleft lip and palate. The electronic medical records of these patients were analyzed for documentation of G-tube placement and comorbidities, including cardiac, neurologic, respiratory, and gastrointestinal defects both independent of and associated with syndromes and/or chromosomopathies (syn/chrom). Chromosomopathies included unnamed deletions, translocations, and partial trisomies. Major surgical procedures, such as repair of congenital heart defects, tracheostomy, distraction osteogenesis of the mandible, Nissen fundoplication, Ladd procedure, and ventriculoperitoneal shunt placement, were tracked to help qualify the severity of each disease. For example, only atrial septal defects and ventricular septal defects requiring operations were analyzed as cardiac comorbidities and only gastroesophageal reflux disease failing swallow study or requiring Nissen fundoplication were analyzed as gastrointestinal (reflux) or respiratory (aspiration) comorbidities.

A total of 214 patients met the inclusion criteria. The main reasons for exclusion were cleft palate repaired by surgeons from other groups (n = 35), absence of operative note in the electronic medical record (n = 35), prior G-tube placement (n = 7), care primarily received at other institutions (n = 5), and adoption of patient with unknown early medical history (n = 2).

The patients were separated into 2 groups based on the placement of G-tubes. Patients were then separated further into overlapping subgroups: those with syn/chrom, those with only 1 system comorbidity, and those with multisystem comorbidities. The incidence of the comorbidities was determined in each group (G-tube vs no G-tube), and comparisons were made using relative risk (RR) ratios with 95% confidence intervals.

Statistical calculations were performed according to the Principles of Biostatistics.¹³ Infants with syn/chrom, 1 system comorbidity, and multisystem comorbidities were considered to be at risk. Those at risk who required G-tubes were designated a. Those at risk who did not require G-tubes were designated b. Infants without syn/chrom or without any major system comorbidities were considered to have no risk factors. Those not at risk who required G-tubes were designated c. Those not at risk who did not require G-tubes were designated d. The RR calculations compared infants with syn/chrom vs those with no syn/chrom, 1 system comorbidity vs no comorbidities, and multisystem comorbidities vs no comorbidity according to the following equation: RR = [a /(a + b)]/[c /(c + d)]. Confidence intervals were determined by standard error, calculated as SE = square root ({ b/[a × (a + b)]}+{ d/[c × (c + d)]}), and the following equation: RR × e exp[±1.96 × (SE × RR)].

Demographic information of the 214 patients diagnosed as having cleft palate who met the inclusion criteria is displayed in Table 2. Incidence of syn/chrom is given in Table 3. Of the 214 patients, 34 infants (15.9%) required G-tubes, while the remaining 180 (84.1%) did not. In the population requiring G-tubes, 19 infants (55.9%) had syn/chrom, 17 (50.0%) had 1 system comorbidity, and 12 (35.3%) had multisystem comorbidities. Only 5 patients (14.7%) required G-tubes because of courses that were poorly described; however, documentation did show that 1 of these 5 patients had experienced a failed trial (due to infection) of G-tube placement prior to our care, and mandibular distraction osteogenesis may have failed in another infant based on the distraction and G-tube placement dates.

In the population not requiring G-tubes, 20 infants had syn/chrom (11.1%), another 10 (5.6%) had 1 system comorbidity, and 2 (1.1%) had multisystem comorbidities. The RRs of G-tube placement for the syn/chrom vs no syn/chrom population was 5.68 (95% confidence interval, 3.18-10.16). The RR for G-tube placement for the populations with 1 system comorbidity and multisystem comorbidities vs those with no comorbidities were 21.79 (95% confidence interval, 8.76-54.17) and 29.66 (12.18-72.21), respectively.

Of all comorbidities, the highest rates for G-tube placement occurred in patients with respiratory system defects, both independently as the sole diagnosis (85.7%) and in conjunction with other diagnoses (93.8%). The lowest incidence of G-tube placement was in patients without syn/chrom or comorbidity (2.9%). For further summary of results, see Tables 4,5,6,7, and 8.

This study is significant because infant feeding problems are detrimental to both patients and parents. In the child, they can lead to serious complications, such as malnutrition and failure to thrive in addition to discomfort during feeding, with nasopharyngeal regurgitation, bloating from excessive air intake, choking, coughing, gagging, and fatigue from long feedings.^1,10 Feeding problems can also increase time to surgical repair of the cleft palate.⁴ For parents, feeding is one of the most pressing concerns after learning of the cleft diagnosis,¹⁴ and some mothers experience disappointment, guilt, or confusion when faced with difficulty breastfeeding.¹⁵

Several strategies are used to diminish feeding complications before G-tube placement is necessary. However, the majority of strategies are not supported by compelling evidence. Reid¹⁶ identified 5 main categories of feeding strategies: equipment, technique, breastfeeding, prostheses, and nutrition/lactation advice. Combinations of these strategies are often used. Feeding equipment may range from compressible bottles that overcome inadequate suck, such as the Mead Johnson or Haberman feeders; specialized orthodontic nipples; and nasogastric tubes, orogastric tubes, and G-tubes.^1,7 Positional considerations, such as semiupright and upright feeding, have been proposed to reduce nasal regurgitation and poor airway protection.^1,4 Breastfeeding can be accomplished with less severe palatal defects. However, while many cleft palate studies expound the benefits of breastfeeding, most point out that breastfeeding alone in the presence of cleft palate results in inadequate oral intake. Breast milk may be provided via special feeders or, like formula, may be thickened to enhance caloric density for adequate nutrition. Prostheses can be placed inside the infant's mouth to help better seal the oropharynx during feeding. Success and adherence using these devices are questionable,¹⁷ but they have been shown to work on a case basis, and efforts have been made to make them more comfortable for the infant.¹⁸ Finally, parent education and support have been useful in countering failure to thrive.⁴ Referrals to nutrition and feeding specialists are certainly appropriate, particularly in refractory cases.

In infants with additional anatomic defects, protective measures to avoid G-tube placement may be invasive. For example, in infants with the micrognathia and glossoptosis of Pierre Robin syndrome (PRS), which make the mechanics of swallowing and the ability to protect the airway even more challenging, corrective jaw surgery may be indicated. One retrospective study⁶ of 115 patients with PRS showed that 44% underwent operative airway management and 39% underwent tracheotomy, but 42% of patients nevertheless required G-tube placement. In a study⁹ of 67 patients with PRS, children were divided into syndromic and nonsyndromic subgroups with early (within 3 months of birth) or late airway intervention. The conditions of all nonsyndromic patients with PRS and early distraction osteogenesis or tracheotomy were able to be managed without a G-tube. Overall, approximately 10% of the nonsyndromic patients with PRS required a G-tube, while 75% of the syndromic patients eventually needed G-tubes.

Despite attempts to avoid G-tube placement, some patients still require them, most often because of malnutrition, failure to thrive, gastroesophageal reflux, or aspiration (Table 8). Complications of gastrostomy include gastroesophageal reflux, infection, formation of granulation tissue, tube dislodgement, intra-abdominal leaks, gastrocolic fistulae, and oral aversion.¹⁹ Thus, it is beneficial to determine which patients are at greatest risk for requiring this intervention. The wide range of G-tube feeding rates in previous studies (Table 1) illustrates the influence of comorbidities. In our study, rates were as low as 3% in infants with cleft palate with no comorbidity and as high as 94% in infants with respiratory comorbidity.

The complications caused by the neurologic, gastrointestinal, cardiac, and respiratory comorbidities and their influence on G-tube placement are not surprising. For example, one recent study²⁰ showed that patients with motor neuron diseases who received G-tubes had better nutrition and survival than counterparts without such invasive interventions, citing well-known complications, such as diminished oral intake. This finding is consistent with those of our study, in which insufficient tone or mental capacity to orchestrate deglutition led to similar decreased oral intake. Feeding tubes have been known to aid nutrition in neurologically disabled children with esophageal dysmotility.²¹

Gastrointestinal complications in this study consisted primarily of gastroesophageal reflux, hiatal (both sliding and paraesophageal) hernias, and aspiration (which was classified under “respiratory” complications for its effect on oxygen saturation). These problems clearly contribute to other aspects of the alimentary tract problems, such as oral feeding. Some studies even suggest resorting to continued G-tube feeds if reflux symptoms are intractable to fundoplication.²²

Feeding difficulties in patients with cardiac anomalies, including the relatively common endocardial cushion defects, have been described for decades. Poor appetite, fatigue with feeding, and failure to thrive are symptoms that warrant further cardiac workup.²³ Infants with poor cardiac output often burn too many calories in the feeding process, and their nutritional status cannot keep up with the stress and discomfort of feeds. In these cases, intervention is indicated.^24,25

Respiratory comorbidities are also understandably associated with feeding difficulties, as protection of the airway is imperative in the normal swallowing mechanism. Silent aspiration leading to chronic feeding difficulty is one independent cause of complications.²⁶ Respiratory comor bidities may also be related to neurodevelopmental abnormalities,²⁷ gastroesophageal reflux,²⁸ or airway protection and aspiration complications, such as in PRS.²⁹ In this study, respiratory comorbidity was the most common factor in increasing risk for G-tube, both alone and in combination with other system comorbidities.

Srinivasan et al³⁰ conducted a study examining the indications for G-tube placement in all children and concluded that the most common causes were neurologic disorders, followed closely by congenital heart defects. Distant third and fourth indications were respiratory related, and the diagnosis of cleft palate completed the top 5 indications. However, on closer inspection, when considering only G-tubes placed at average patient age younger than 1 year, the top 3 indications listed were cleft lip/palate, congenital heart disease, and chronic lung disease at ages 0.36, 0.39, and 0.73 year, respectively. These leading causes for G-tube placement in infancy agree with our study's findings, which show that, within cleft palate indications, comorbidities affecting other systems—specifically cardiac and respiratory—and other physiologic mechanisms (ie, not just the sucking and swallowing defects) are important determinants.

One potential limitation of our study is that the feeding interventions prior to G-tube placement were not standardized. In the past decade, the evaluation process of infants with cleft palate has become fairly consistent, with cleft nurses often meeting with parents within 1 week of their child's birth and introducing such interventions as the Haberman feeder and upright feeding. In our study, use of the Medela Special Needs Feeder (formerly called the Haberman feeder) (Medela, Inc, McHenry, Illinois) was unsuccessful in 24 of the 34 infants requiring G-tubes, and modified bottles and surgical intervention for infants with PRS worked very well for those not requiring G-tubes (116 used the Haberman feeder as the bottle of choice, 7 used the Mead Johnson Cleft Palate Nurser (Mead Johnson, Evansville, Indiana), 6 used the Pigeon Cleft Palate Nurser (Galtak Houseware Ltd, Markham, Ontario, Canada), 5 used the Playtex VentAire Bottle (Playtex Products Inc, Westport, Connecticut), and 11 required distraction osteogenesis of the mandible for definitive feeding management) (data not shown). However, no conclusions can be made about the overall algorithm for decision making regarding feeds, and we hope that future studies can provide such helpful information.

This study demonstrates the importance of neurologic, gastrointestinal, cardiac, and respiratory comorbidities in indicating the need for G-tube placement in infants with cleft palate. In our study, rates were as low as 3% in infants without any comorbidity and as high as 94% in infants with respiratory comorbidity. The diagnosis of a cleft syn/chrom significantly increased the relative risk of G-tube placement, and the particular syndrome may help in determination of whether a G-tube is indicated, as the type and incidence of many syndromes differed between the 2 populations. Nevertheless, the effect of syn/chrom pales in comparison with the relative risks imposed by specific system comorbidities, both independently and in conjunction with other system involvement. Thus, it seems that which system is involved and the degree to which the system is impaired are more influential than a simple diagnosis of syn/chrom. Patients with neurologic, gastrointestinal, cardiac, and respiratory comorbidities are at increased risk for G-tube placement, with respiratory comorbidity appearing to confer the greatest risk. It is our hope that with this information, infants with cleft palate who are at highest risk for G-tube placement can be identified effectively and aggressive measures to orally feed these infants as well as those at lower risk can be undertaken successfully.

Correspondence: James D. Sidman, MD, Pediatric ENT Associates, Children's Specialty Center, 2530 Chicago Ave S, Ste 450, Minneapolis, MN 55404 (sidma001@umn.edu;cuxxx003@umn.edu).

Submitted for Publication: July 26, 2010; final revision received October 12, 2010; accepted November 25, 2010.

Author Contributions: Ms Cu and Dr Sidman had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Cu and Sidman. Acquisition of data: Cu. Analysis and interpretation of data: Cu and Sidman. Drafting of the manuscript: Cu and Sidman. Critical revision of the manuscript for important intellectual content: Cu and Sidman. Statistical analysis: Cu. Administrative, technical, and material support: Cu and Sidman. Study supervision: Sidman.

Financial Disclosure: None reported.