Children with developmental disorders and those with chronic disease appear to be at risk for chronic feeding problems and food refusal(1,2). In a recent study, children referred for feeding problems were categorized using a multidimensional classification system including anatomic abnormalities, neurologic conditions, behavioral issues, cardiac and respiratory problems and metabolic dysfunction (3). Developmental delay was a problem for 74% of all subjects, often in association with anatomic and behavioral problems. Children with psychomotor retardation display a high rate of oral dysfunction (4) with associated aversive experiences (aspiration and choking), food refusal after prolonged airway intubation or tube feeding, and persistent vomiting and retching despite aggressive medical and surgical treatment (5).
A frequent approach to the evaluation of infants and toddlers who refuse to eat has been to distinguish between medical and behavioral conditions (2,6). This dichotomization sometimes fails when there is a complex interaction among anatomic, physiologic, behavioral and environmental factors. For example a combination of intensive behavioral therapy and nutritional interventions eliminated tube feeding requirements in 47% of children with lifelong food aversion (7). We hypothesize that motor and sensory disorders may have contributed to persistent food refusal in some of the children who failed the intensive behavioral therapy. Infants may refuse to eat because it is painful or unpleasant to eat, or they fear eating will be a dysphoric experience. Abnormal antroduodenal manometry (8), gastric antral electrical dysrhythmias (9) and delayed gastric emptying (10), were associated with persistent symptoms of retching, vomiting and food aversion after fundoplication. It seems probable that the abnormalities in motility and visceral sensation existed before fundoplication, and contributed to the pathogenesis of chronic symptoms prompting surgery. The purposes of the current study were 1) to investigate the possible contribution of abnormalities in upper gastrointestinal motility and gastric sensation in medically fragile children referred for refusal to eat; 2) to pilot multidisciplinary treatment plans individualized to each patient's pathophysiology.
We studied 14 children (age range 1.5–6 years; 7 male): 11 with cerebral palsy and moderate to severe developmental delay, 1 with congenital chronic renal failure, and 2 with feeding problems that started after fundoplication but no other known disease. Twelve had received intensive neonatal care, 9 because of prematurity with either respiratory or neurologic complications and 3 because of perinatal hypoxic-ischemic encephalopathy. All presented with a life long history of food aversion, vomiting and/or retching persisting after multiple medical and surgical interventions, including one or two fundoplications (Table 1). The patient with renal failure was considered by his nephrologist as having satisfactory metabolic control. All patients had trials of drugs to treat gastroesophageal reflux disease, including cisapride, metoclopramide, H2 histamine receptor antagonists and/or proton pump inhibitors. All had received months of feeding therapy from occupational or speech therapists. We discontinued medications affecting motility at least 2 days before testing.
We evaluated all 14 patients with antroduodenal manometry and 8 patients with esophageal manometry. During the esophago-gastro-duodenoscopy for antroduodenal manometry catheter placement, the fundoplication was examined and found to be intact in all patients and there were no mucosal abnormalities seen. No biopsies were taken during the procedure for catheter placements. We performed esophageal and antroduodenal manometry with 8-lumen water-perfused catheters, with recording sites 3 cm apart, using methods reported in detail previously (11,12). For antroduodenal manometry, at least two recording sites were in the gastric antrum and the others were in the duodenum. We studied antroduodenal manometry fasting for 3 hours. If we saw a phase 3 in the first 3 hours, we fed a meal suited to the child's age and ability, and continued to record for one more hour. If there was no phase 3 in the first three hours of fasting we gave erythromycin 1 mg/kg IV over 30 minutes and continued recording for a full hour. Whether erythromycin induced a phase 3 or not, we proceeded to the meal. We measured postprandial motility for 1 hour. We fed each child with his or her usual formula through the gastrostomy or the gastro-jejunostomy. We analyzed the results by visual inspection during fasting for migrating motor complexes, after meals for postprandial patterns, and throughout for discrete qualitative or quantitative abnormalities (13). We defined postprandial antral hypomotility as no antral contractions for an hour following a complex liquid meal. True vomiting, comprised of strong, rapid retrograde duodenal and gastric contractions, could be differentiated from regurgitations caused by brief simultaneous increases in intraabdominal pressure from abdominal wall and diaphragmatic contractions. The later cause for emesis we interpreted as a voluntary or involuntary attempt to relieve pain or discomfort. We performed esophageal manometry immediately after catheter placement, so that the children were without prolonged naso-gastric intubation. We used a station pull-through technique. We assessed the percentage of swallows that propagated, considering normal a value ranging between 92 and 96% (14). We diagnosed patients with diffuse esophageal spasm according to criteria for adult patients (15). One patient had poor propagation of esophageal body contractions and incomplete relaxation of the LES and we made the diagnosis of non-specific esophageal motor disorder(16). We studied the gastric volume threshold for retching in 11 patients utilizing a computer-driven air pump accurately measuring intra-gastric volume (Distender II, Toronto, Canada). The patient's first retch signaled the retching threshold, as we inflated a compliant latex balloon at a rate of 100 ml/min. We considered as normal a volume of >20 ml/kg before retching, a volume used in previous studies as standard meal for infants (17). In 3 patients the gastrostomy orifice was too small to admit the latex balloon and 18 F catheter. In these patients we infused formula at a rate of 20 ml/kg/h. We considered the gastric threshold for retching to be the volume of formula that induced symptoms.
At the end of the testing, we chose treatment consistent with the motor and sensory abnormalities (Tables 2 and 3). A multidisciplinary treatment program included a variable combination of drip feeding, drugs for motility disorders, drugs to decrease visceral pain and non-specific arousal, and behavioral, cognitive and family therapy.
We evaluated the response to treatment 2 to 6 months after our intervention. A chart audit provided the frequency of vomiting at the time of the study and feeding method (tube feeding, tube and oral, or exclusively oral). Using a questionnaire developed for the study, interviews were conducted by a trained telephone interviewer with either parent. The questionnaire evaluated daily frequency of retching and vomiting, feeding modification, and emotional health of the child. Current symptoms of vomiting or retching were obtained by asking parents, “During the last seven days, has your child vomited or retched?” If the parent confirmed the presence of the symptoms, they were asked if their child had an infection during the last seven days that might account for the vomiting and retching and how often their child had vomited or retched during the last seven days. Data were analyzed using descriptive statistics (frequencies and percentages) to describe the results of the feeding modification and for the study overall. The Wilcoxon signed rank test for paired samples was used to compare retching before and after treatment. To assess alterations in emotional health, as perceived by the parents, the interviewers asked parents to compare their child's emotional health at the time of the study to current functioning using a five point Likert scale ranging from “a lot worse” to “much better”.
Motility and sensory studies
All the patients had gastrointestinal motility and/or sensory abnormalities (Table 2). Two patients had an esophageal motility disorder alone: 1 with a non-specific esophageal motor disorder and 1 with diffuse esophageal spasm. Five patients showed no motility disorders, but reduced gastric threshold for retching explained their symptoms. In 7 patients there was a combination of antroduodenal motility disorders (absence of phase 3 of the MMC in 2 patients, unremitting clustered duodenal contractions in 3 patients and postprandial hypomotility in 2 patients) concomitant with decreased gastric volume capacity.
One patient had a combination of diffuse esophageal spasm, a continuous band of repetitive contractions in the duodenum and reduced gastric volume capacity.
We treated two patients with diffuse esophageal spasm with both nifedipine (18,19) and dicyclomine (20). We performed prolonged (24 hours) esophageal manometry and monitored pulse and blood pressure to determine the safety and efficacy for nifedipine (0.2 mg/kg three times per day) at the start of treatment. Nifedipine reduced the amplitude and duration of the esophageal body contractions, as well as the number of simultaneous contractions for over two hours after administration without altering pulse or blood pressure.
For the patients with a decreased gastric volume capacity, the treatment included: 1) an explanation to the caretakers defining visceral hyperalgesia and brain-gut interaction, 2) instructions to avoid unnecessary painful experiences and to minimize discomfort with necessary medical interventions, 3) in 5 cases (two with concomitant motor disorders, antral hypomotility and absence of phase 3 of the MMC) placement of gastrojejunal feeding tubes for jejunal continuous drip feeding to eliminate gastric distention, and so avoid the gastric distention stimulus for retching, 4) we used medications implicated in reducing visceral pain perception, including tricyclic antidepressants, such as imipramine and amitriptyline (21,22), gabapentin (23), and ondansetron (24) in sequence or in combination. Cyproheptadine was prescribed to two patients with a partial response to imipramine alone to take advantage of its appetite stimulating property (25). For the 3 patients with a continuous band of duodenal contractions, we prescribed dicyclomine to relieve the obstruction caused by absence of relaxation (26).
After intervention, retching and vomiting episodes improved in 13 patients (93%) from a mean of 15 a day to 1.4 a day (p <0.01) (Fig 1). Parents reported 11 of 14 (80%) had emotional health improvement after intervention. None of the patients' conditions worsened. (Fig 2). Six of 14 patients (43%) advanced in the mode of feeding. Three of 14 patients advanced from tube to exclusively oral caloric intake (Table 5). Three of 14 had an oral intake of at least 10% (range 10 to 75%) of their recommended daily caloric intake (Tables 3 and 5).
Eight patients (57%) continued on drip feeding with oral intake <5% of daily caloric intake. Six of these 8 showed both emotional health improvement and a significant decrease in retching and vomiting.
The results in this series of chronically ill toddlers with feeding refusal suggest that these patients experienced treatable problems related to abnormal motility and/or sensation in the upper gastrointestinal tract.
In our group of patients, 9 of 14 had motility disorders involving the esophagus, gastric antrum and/or the duodenum. Abnormal motility might be related to abnormal modulation of the extrinsic innervation from a damaged CNS, or to hypoxic-ischemic damage to the enteric nerves.
Twelve of 14 patients showed a low gastric volume threshold for retching. Decreased gastric compliance resulting in elevated intra-gastric pressures and/or gastric hyperalgesia might account for this pattern. All patients had a history of fundoplication, some more then once. Fundoplication may be associated with a retching when there is an esophageal or gastric motility disorder, hyperalgesia, or central nervous system-mediated vomiting (27). Fundoplication might contribute to retching via at least 3 mechanisms: 1) as a sensitizing pain experience, 2) by decreasing gastric compliance, as portions of the cardia and fundus which participate in receptive relaxation are used to make the wrap and 3) if there is an incomplete LES relaxation in a patient with an underlying esophageal motility disorder swallowed food and secretions may not properly empty from the esophagus. Two factors implicate gastric hyperalgesia. First, reduced compliance after fundoplication appeared to be a transient postoperative phenomena. In a rat model, gastric compliance returned to the preoperative values in 30 days (28). Second, all the patients had symptoms preceding fundoplication. Symptoms improved in most during treatment for chronic visceral pain, which would not be expected to alter gastric compliance.
The histories of repeated painful experiences during infancy fit the description for the pathogenesis of visceral hyperalgesia, consistent with recruitment of previously silent nociceptors and sensitization of visceral afferent pathways. Studies in the newborn have shown that transient noxious stimuli result in permanent alterations in the afferent pain pathways (29), with evidence of an increased excitability of dorsal horn neurons and altered motor reflex arcs (30). Newborns have the neural pathways for detection, transmission and reaction to noxious stimuli, but lack a mature descending pain inhibitory tract (29–31). Alterations in visceral sensory processing occur as perinatal events or during infancy (29). Persistent stimulation of afferent fibers reshapes the peripheral nociceptor responses as well as the central neuron responses to pain, resulting in primary and secondary hyperalgesia (32). In this group of chronically ill patients, prolonged neonatal care, multiple painful medical procedures, gastritis and esophagitis, obstructive apnea and surgery (e.g. fundoplication or gastrostomy placement) were all potentially sensitizing experiences. These experiences may have contributed to the creation of pathologic pain pathways and central pain memories, especially in the infants with CNS disorders, who may have aberrant modulation of peripheral afferent input as part of their neurologic derangement.
One patient had nausea, true vomiting and retching even on jejunal drip feeding, which suggests that afferent pathways of visceral hyperalgesia can be also mediated by sensory receptors distal to the stomach. We studied gastric sensory sensitivity to distention because all the children had a gastrostomy that provided a nontraumatic access to the stomach. It is possible that visceral hyperalgesia, expressed as low sensory threshold to distention, affected also the esophagus and duodenum but these regions were not tested. A multidisciplinary approach to treat the sensory abnormalities resulted in less retching, advances in feeding behavior for some, and improved emotional health for many. Parents learned the importance of avoiding additional pain experiences, and of helping the child manage stress. Visceral hyperalgesia may be potentiated by learned helplessness. During treatment, parents learned behavioral strategies to assist their child in managing stress and discomfort. Parents' attitudes changed from a sense of powerlessness and worry to competence, reducing their stress and stress in the child. Continuous drip feedings bypassed the stomach, and reduced gastric distention. After several days, most children stopped retching, but in most cases jejunal drip feedings continued to provide nutrition support for up to two months. Parents accepted an explanation that continuous drip feedings were not a regression to an undesirable form of feeding but a logical step in treating visceral hyperalgesia.
The drugs for visceral hyperalgesia were intended to reduce afferent signals and/or central nervous system arousal. The tricyclic antidepressants are commonly used in adults with chronic abdominal pain (33,34) and have the advantage of once a day administration. They were given 2 hours before bedtime, increasing the dose from 0.2 mg/kg/d in 0.2 mg/kg increments 4–7 days apart, until the child sleeps comfortably through the night, waking refreshed in the morning. This endpoint is appreciated by parents, because their nights become more restful. In two cases, parents reported no improvement but increased irritability with imipramine. Imipramine was discontinued, and both these infants had an immediate and sustained improvement with gabapentin, 50 mg at night. Cyproheptadine combined with imipramine appeared to be effective in two patients who had showed only a partial response to imipramine alone (no retching but persistent refusal to eat).
This case series lacks a control group and consistent treatment protocol. Benoit et al. studied children with food refusal and included a control (non-intervention) group. None of the toddlers in the control group moved from tube to oral feeding in that study(7). It seems unlikely that children with food refusal will improve spontaneously, without treatment. Sharing these data might help to focus other clinicians towards improving the quality of life for these needy toddlers. In children with chronic illness, persistent vomiting or retching, and refusal to eat, unrecognized but treatable gastrointestinal motor and/or sensory disturbances may contribute to the symptoms. Prospective studies are necessary to establish the mechanism underlying what we defined as low gastric threshold for retching, and to find the most effective medical treatment alternatives to fundoplication for persistent vomiting in medically fragile toddlers.
The authors thank Janssen Pharmaceutica for supporting Pediatric Gastrointestinal Motility Research Fellowships for Carla Ciarla and Tsili Zangen.
1. Palmer S, Horn S. Feeding problems in children. In Palmer S, Ekvall S, eds. Pediatric Nutrition in Developmental Disorders.
Springfield, IL: Thomas, 1978:95–100.
2. Fank DA, Zeisel SA. Failure to thrive. Pediatr Clin N Am 1988; 35:1187–206.
3. Burklow KA, Phelps AN, Rudolph C, et al. Classifying complex pediatric feeding disorders J Pediatr Gastroenterol Nutr 1998; 27:143–47.
4. Reilly S, Skuse D, Poblete X. Prevalence of feeding problems and oral motor dysfunction in children with cerebral palsy: a community survey. J Pediatr 1996; 129:877–82.
5. Martinez DA, Ginn-Pease ME, Cainano DA. Recognition of recurrent gastroesophageal reflux following antireflux surgery in the neurologically disabled child: High index of suspicion and definitive evaluation. J Pediatr Surg 1992; 27:983–90.
6. Woolston JL. Eating and growth disorders in infants and children. Dev Clin Psychol Psychiatry 1991; 24:1–85.
7. Benoit D, Wang EEL, Zlotkin SH. Discontinuation of enterostomy tube feeding
by behavioral treatment in early childhood: A randomized controlled trial. J Pediatrics 2000; 137:498–503.
8. Richards CA, Andrews PLR, Spitz L, et al. Nissen fundoplication may induce gastric myoelectrical disturbance in children. J Pediatr Surg 1998; 33:1801–5.
9. Di Lorenzo C, Flores AF, Hyman PE. Intestinal motility in symptomatic children with fundoplication. J Pediatr Gastroenterol Nutr 1991; 12:169–73.
10. Papaiala JG, Wilmot D, Vane DW, et al. Increased incidence of delayed gastric emptying in children with gastroesophageal reflux. A prospective evaluation. Arch Surg 1989; 124:933–6.
11. Kahrilas PJ, Ray EC, Walter JH. American Gastroenterological Association technical review on the clinical use of esophageal manometry. Gastroenterology 1994; 107:1865–84.
12. Uc A, Hoon A, Di Lorenzo C, et al. Antroduodenal manometry in children with no upper gastrointestinal symptoms. Scand J Gastroenterol 1997; 32( 7):681–5.
13. Tomomasa T, Di Lorenzo C, Morikawa A, et al. Analysis of fasting antroduodenal manometry in children. Dig Dis Sci 1996; 41:2195–203.
14. Richter JE, Wu WC, Johnson DN, et al. Esophageal manometry in 95 healthy adult volunteers. Dig Dis Sci 1987; 32:583–92.
15. Richter JE, Castell DO. Diffuse esophageal spasm: a reappraisal. Ann Int Med 1984; 100:242–45.
16. McCord GS, Staiano A, Clouse RE. Achalasia, diffuse esophageal spasm and non-specific motor disorders. Baillieres Clin Gastroenetrol 1991; 5:307–35.
17. Siegel M, Lebenthal E, Topper W, et al. Gastric emptying in prematures of isocaloric feeding with differing osmolalities. Pediatric Res 1982;141–147.
18. Richter JE, Dalton CB, Gastell DG. Nifedipine: a potent inhibitor of contractions in the body of the human esophagus. Gastroenterology 1985; 89:549–54.
19. Maksimak M, Perlmutter DH, Winter HS. The use of nifedipine for the treatment of achalasia in children. J Ped Gastroenterol Nutr 1986; 5:883–86.
20. Fontan JP, Heldt GP, Heyman MB, et al. Esophageal spasm associated with apnea and bradicardia in an infant. Pediatrics 1984; 73:52–55.
21. Clouse RE, Lustman RA, Geisman RA, et al. Antidepressant therapy in 138 patients with irritable bowel syndrome: a five year clinical experience. Aliment Pharmacol Ther 1994; 8:409–16.
22. Peghini PL, Katz PO, Castell DO. Imipramine decreases esophageal pain perception in human male volunteers. Gut 1998; 48:807–13.
23. Rosner H, Rubin L, Kestenbaum A. Gabapentin adjunctive therapy in neuropathic pain states. Clin J Pain. 1996; 12:56–8.
24. Bonner SE, Carter M, Sauger GJ. 5-hydroxytryptamine3 receptor antagonism modulates a noxious visceral pseudo affective reflex. Neuropharmacology 1995; 34:263–7.
25. Bergen SS. Appetite stimulating properties of cyproheptadine. Amer J Dis Child. 1964; 108:270–73.
26. Farruggia G, Camilleri M, Whitehead W. Therapeutic strategies for motility disorders Gastroenterol Clin North Am 1996; 25:225–41.
27. Di Lorenzo C, Youssef YN, Sigurdsson L, et al. Visceral hyperalgesia in children with functional abdominal pain. J Pediatrics 2001; 139:838–43.
28. Bustoff-Silva J, Peres CA, Fonkalsrud EW, et al. Gastric emptying after fundoplication is dependent on changes in gastric volume and compliance. J Pediatr Surg 1999; 34:1232–35.
29. Fitzgerald M. Development of pain mechanism. Br Med Bull 1991; 47:667–75.
30. Fitzgerald M, Shaw A, McIntosh N. Postnatal development of cutaneous flexor reflex: comparative study of preterm infants and newborn rat pups. Dev Med Child Neurol 1988; 30;520–26.
31. Fitzgerald M, McIntosh N. Pain and analgesia in the newborn. Arch Dis Child 1989; 64:441–3.
32. La Motte RH, Shain CN, Simone DA. Neurogenic hyperalgesia: Psychophysical studies of underlying mechanism. J Neurophysiol 1991; 66:190–211.
33. Greenbaum DS, Mayle JE, Vanegeren LE, et al. Effects of desipramine on irritable bowel syndrome compared with atropine and placebo. Dig Dis Sci 1987; 32:257–66.
34. Mertz H, Foss R, Kodner A. Effect of amitriptyline on symptoms, sleep and visceral perception in patients with functional dyspepsia. Am J Gastroenterol 1998; 93:160–65.