Nausea, vomiting, and abdominal pain are among the most common complaints in children. When these symptoms are persistent, one needs to consider functional causes including gastroesophageal reflux, peptic ulcer disease, functional dyspepsia, and gastroparesis. The criterion standard for the diagnosis of gastroparesis is gastric emptying scintigraphy. Although standardized testing and normal values are well defined in adults, this is not the case in pediatrics, in which techniques for testing and definitions of delayed gastric emptying (DGE) vary across institutions. Further compounding the challenge of caring for these young patients, the etiology, clinical presentation, and optimal treatment of gastroparesis in children are poorly defined, with only small case series describing the clinical presentation and outcome of diabetic (1) and postviral (2) cases. The aims of the present study were to describe the clinical presentation of gastroparesis from all of the causes in children, evaluate the response to therapy, and determine long-term outcomes.
Institutional review board approval was obtained to perform a retrospective review of patients age 21 years or younger who underwent a gastric emptying test (GET) by scintigraphy for evaluation of DGE from January 2004 to July 2007. We conducted a search for the terms “gastric emptying study,” “esophageal reflux,” “gastroparesis,” “dyspepsia,” and “vomiting” from the Research Patient Data Registry at Massachusetts General Hospital. All of the subjects underwent imaging or endoscopic evaluation and those with anatomic obstruction were excluded. DGE was defined by institutional standards as emptying of solids and/or liquids of ≤40% at 60 minutes, a definition supported by retrospective and prospective studies (3–5). Study entry was defined as the time of first GET showing DGE, and follow-up duration was defined as the time interval between study entry until the last clinical contact. Final determination regarding resolution of symptoms was made at the time of the last clinical contact. A total of 230 subjects meeting inclusion criteria were identified.
Gastric Emptying Study
GET was performed following a 4-hour fast and off promotility drugs for at least 48 hours. For liquid phase emptying, at least 1 ounce of formula or milk (mixed with technetium-99 sulfur colloid) was ingested in 5 minutes followed by supine data acquisition for 60 minutes. For the solid phase, scrambled eggs (120 g Egg Beaters) cooked with technetium-99 sulfur colloid and 2 pieces of bread with a thin layer of strawberry jelly were ingested in 10 minutes, followed by supine or upright acquisition every 30 minutes for up to 90 minutes. Infants ingested oatmeal for the test. Indium-111 was used for liquid phase (water) and technetium-99 for solid phase when both were performed simultaneously. In 23 patients, the study was performed at an outside institution using a similar protocol with the values expressed as t½ (time in minutes to empty 50%), with all being >100 minutes. These were included in the response to therapy and outcome analyses but excluded from comparison of gastric emptying values because we could not extrapolate the exact percentage of emptying at 60 minutes.
Medical records were reviewed to extract demographic data, clinical history and presentation, medical and surgical treatment, radiographic results, response to treatment, and final outcome. Reviewers were blinded to GET results at the time of data collection to avoid bias.
For descriptive analysis, continuous measures are presented as mean ± standard deviation and categorical measures are presented in counts (proportions).
Subjects were divided into 3 age groups: infants (birth–1 year), children (>1–12 years), and adolescents (>12 years). Values for gastric emptying of liquids and solids at 60 minutes were compared using the Kruskal-Wallis test. The distribution of symptoms (vomiting, abdominal pain, nausea) was compared using the χ2 test.
The response to therapy was scored as follows: 1 = no response (including those with partial improvement with need for additional therapy, or any improvement but drug had to be discontinued due to adverse effects [AEs]); 2 = significant response: resolution of symptoms while still on therapy; 3 = resolution of symptoms with medical therapy successfully discontinued. “Responders” to diet modification and medical treatment, addition of promotility drugs to medical treatment, and surgical interventions were defined as those with a score = 2 or 3, and “nonresponders” as those with a score = 1. The median gastric emptying values at 60 minutes were compared between responders and nonresponders to addition of promotility agents using the nonparametric Mann-Whitney test for independent samples. The association between response to addition of promotility drugs and baseline categorical characteristics (age category, sex, presence of symptoms [nausea, vomiting, abdominal pain], presence of mitochondrial dysfunction [MD], presence of viral prodrome) was tested using χ2 tests for proportions, whereas association with the continuous variable “duration of symptoms” was determined using a nonparametric test. Severity of DGE was evaluated between those categorical characteristics using the χ2 test.
We evaluated the individual effect on resolution of symptoms of the following variables: age, sex, vomiting, nausea, abdominal pain, duration of symptoms (time in months between symptom onset and the time GET was obtained), MD (diagnosed by muscle biopsy), postviral gastroparesis (PVG, defined as persistence of symptoms 1 month after onset of the viral illness), and response to promotility therapy. Log-rank testing with censoring time to resolution at the end of the follow-up if not observed during study was used to evaluate the individual effect of these variables. The Kaplan-Meier method was used to estimate the cumulative incidence rate of resolution. Multiple Cox proportional hazards models were used to explore the joint effect of the above-listed risk factors on time to resolution. Only those with follow-up of at least 2 months were included in this outcomes analysis.
All Age Groups
A total of 230 patients were included, and the demographics are shown in Table 1. The most common symptoms prompting evaluation were vomiting (42%), abdominal pain (35%), and nausea (29%) (Fig. 1). The most common comorbidities were PVG in 42 patients (18%), gastroesophageal reflux in 33 (14%), MD in 18 (8%), diabetes mellitus in 5 (2%), and hypothyroidism in 1 (0.4%) patient (Table 1). Duration of symptoms was longer in patients with MD and shorter in patients with PVG compared with subjects without those conditions (P < 0.001). In contrast, there was no difference between MD and PVG with respect to age, sex, and gastric emptying of liquids or solids. Additional testing performed included upper gastrointestinal series in 82 (intestinal malrotation in 2 patients, both presenting after a Ladd procedure), pH probe study in 22 (prolonged esophageal acid exposure in 13), upper gastrointestinal endoscopy in 85 (reflux esophagitis in 22, gastritis in 17, duodenitis in 2), and antroduodenal manometry in 9 (postprandial antral hypomotility in 8) patients. Gastroparesis symptoms varied by age group. The proportion of patients with vomiting was significantly higher among children than adolescents (P = 0.017), and the proportion reporting nausea and abdominal pain was significantly higher among adolescents (P = 0.006 and P = 0.014, respectively; Fig. 1). A total of 31 of 36 infants were diagnosed as having DGE based on the emptying of liquids, and 99 of 109 children and 80 of 85 adolescents were diagnosed as having DGE based on the emptying of solids at 60 minutes. Comparing GET values among age groups was only statistically significant in that emptying of liquids was significantly delayed in infants as compared with adolescents (P = 0.009) (Table 1). Severity of DGE was stratified as 20% to 40% versus <20% emptying at 60 minutes. We found no association between severity of DGE and age, sex, nausea, vomiting, abdominal pain, PVG, MD, response to promotility drugs, or resolution.
A total of 36 infants were included. The most common symptoms were vomiting in 25 (69%), failure to thrive in 5 (14%), irritability in 5 (14%), and abdominal pain in 1 (3%). The most common associated diagnoses were cow's-milk protein allergy (diagnosed clinically) in 12 (33%) and PVG in 6 (17%). Information on the response to addition of promotility therapy was available in 32, with 13 (40%) responders and 19 (59%) nonresponders. The distribution of medications used in infants was as follows: PPI in 32, metoclopramide in 29, domperidone in 4, tegaserod in 2, erythromycin in 12, and cyproheptadine in 2 patients. Outcome data were available in 36 patients, with 22 (61%) reporting resolution of symptoms at a mean duration of 10.7 ± 6.3 months and 82% reporting resolution within 1 year. All 12 infants with milk allergy were treated with a hypoallergenic formula, with 7 reporting resolution within 12 months, 4 remaining symptomatic, and 1 lost to follow-up.
A total of 109 children were included. The most common symptoms were vomiting in 49 (45%), abdominal pain in 36 (33%), and nausea in 28 (25%) patients. The most common diagnoses were PVG in 21 (19%), MD in 10 (9%), and gastroesophageal reflux in 10 (9.6%) patients. Information on the response to the addition of promotility therapy was available in 77 children: 49 (64%) responders and 28 (36%) nonresponders. The distribution of medications used in children was as follows: PPI in 81, metoclopramide in 71, domperidone in 15, tegaserod in 3, erythromycin in 13, and cyproheptadine in 7 patients. Outcome data were available in 97 (89%) patients, with 60 (62%) reporting symptom resolution at a mean duration of 9.2 ± 7.3 months (43% by 6 months, 85% by 1 year, 97% by 2 years, and 100% by 3 years).
A total of 85 adolescents were included. The most common symptoms were abdominal pain in 43 (50%), nausea in 38 (45%), and vomiting in 24 (28%) patients. The most frequent diagnoses were PVG in 15 (18%), irritable bowel syndrome in 9 (11%), depression in 7 (8%), and MD in 6 (7%) patients. Information on the response to the addition of promotility therapy was available in 53 adolescents: 28 (53%) responders and 25 (47%) nonresponders. The distribution of medications used in adolescents was as follows: PPI in 69, metoclopramide in 42, domperidone in 14, tegaserod in 15, erythromycin in 15, and cyproheptadine in 2 patients. Outcome data were available for 65 (76%) patients, with only 25 (38%) reporting resolution at a mean duration of 7.5 ± 7.3 months (60% by 6 months, 84% by 12 months, and 100% by 28 months).
Response to Therapy
Medications used included proton pump inhibitors (PPIs), promotility agents, antiemetics, and pain modulators. Medication-related AEs are listed in Table 2.
Proton Pump Inhibitors
Proton pump inhibitors (PPIs) were used in 181 (79%) patients and were the first-line drugs in 176. Dosing ranged from 1 to 2 mg · kg−1 · day−1 for a mean duration of 9.1 ± 10.2 months, with a maximum dose of 30 mg twice per day of lanzoprazole or 40 mg twice per day of omeprazole, esomeprazole, or pantoprazole. Follow-up was available in 173 (96%) patients, with 36 (20%) responders, although only 5 (3%) reported resolution of symptoms and the remainder required additional therapy. Only 5 patients reported AEs from PPI use.
Promotility therapy included medications with a solely promotility effect (tegaserod and erythromycin) and medications with promotility and antinausea effects (metoclopramide and domperidone). Of the 162 subjects with follow-up available, 90 (55%) were responders and 72 (45%) were nonresponders (Fig. 2). The most common symptoms in responders were vomiting in 45 (50%), abdominal pain in 27 (30%), and nausea in 25 (28%). We identified a tendency toward an association between a positive response to promotility therapy in children compared with adolescents and infants (P = 0.07) and in subjects with PVG (P = 0.09). No association was found between response to promotility agents and sex, presence of nausea, vomiting, or abdominal pain, duration of symptoms, or MD. Of the 90 responders, 52 (58%) reported resolution, 29 (32%) continued to require medication, and 9 (10%) were lost to follow-up. Among the 72 nonresponders to addition of promotility therapy, 25 ultimately underwent surgical procedures. The rest were either lost to follow-up or remained symptomatic through the last clinical encounter.
Metoclopramide was used in 142 patients, in 31 as first-line therapy and in the rest after a failed trial of PPI. The dose was 0.1 to 0.2 mg/kg per dose given 4 times per day to a maximum dose of 10 mg 4 times per day. Mean duration of therapy was 8.6 ± 8.7 months. Follow-up was available in 132 (93%) patients, with 26 (20%) responders and 15 (11%) of those reporting resolution. Overall, 24% of patients reported AEs. Domperidone was used in 33 patients, always after failing PPIs or metoclopramide. Dose was 0.1 to 0.2 mg/kg per dose given 4 times per day to a maximum dose of 10 mg 4 times per day. Mean duration of therapy was 7.6 SD 8.1 months. Follow-up was available in 31 (94%) patients, with 23 (74%) responders and 8 (26%) reporting symptom resolution, the highest rate being among promotility drugs. Only 6% reported AEs, the lowest rate of any drug therapy. Tegaserod was used in 20 patients, never as a first-line agent. Starting dose was 0.25 mg · kg−1 · day−1 (range 0.1–0.6 mg · kg−1 · day−1) divided in 2 doses per day, to a maximum dose of 6 mg twice per day. Mean duration of therapy was 5.4 SD 5.8 months. Follow-up was available in 18 patients, with 11 (61%) responders and 4 (22%) reporting symptom resolution. AEs were reported by 4 (20%) patients. Erythromycin was used in 40 patients and follow-up was available in 37 (93%), with 19 (51%) responders and only 2 (5%) reporting resolution. Starting dose was 3 mg/kg per dose given 4 times per day up to 10 mg/kg per dose 4 times per day up to a maximum dose of 250 mg 4 times per day. Mean duration of therapy was 4.1 SD 3.6 months. AEs occurred in 4 (10%) patients.
Cyproheptadine was used in 11 patients and follow-up was available in 8 (73%) patients, with 4 responders and no patients reporting resolution. Drowsiness occurred in 2 (22%) patients. Amitriptyline was used in 6 patients with follow-up available in 5 (83%). All were responders, with only 1 reporting resolution and 2 requiring dose adjustment because of drowsiness.
Diet modification, including small frequent meals, decreased fat intake, and lactose-free diet, was used as primary therapy in only 4 patients (3 responders and 1 nonresponder) and as adjunctive treatment in 11 patients, all of whom were responders.
Gastrostomy tubes were placed in 11 patients (most commonly for failure to thrive). All were responders and 2 reported symptom resolution, with 1 subject lost to follow-up. Pyloroplasty was performed in 6 patients; all were responders, including 2 who reported resolution. In 1 subject, pyloroplasty was performed in combination with a fundoplication after failure of PPIs and resulted in full resolution of symptoms. No AEs were reported. Jejunostomy tubes were placed in 3 patients who failed medical therapy, all reporting significant symptomatic improvement and no adverse effects. Fundoplication was performed in 5 patients, with 2 lost to follow-up. The other 3 had a gastrostomy placed at the time of fundoplication but required subsequent gastrojejunostomy tubes for feeding intolerance. No adverse events were reported.
Outcome data were available in 204 patients. Of those, 107 (52%) reported resolution, whereas 97 (48%) reported that their symptoms persisted at the end of follow-up. Median follow-up was 18 months, with a range of 2 to 84 months. Median time to achieve symptom resolution for all 204 subjects included in the survival analysis was 14 months, with a 95% CI of 12 to 24 months. The overall resolution rate was 22.4% at 6 months, 53.3% at 18 months, and 60.6% at 36 months from the Kaplan-Meier curve. Of the 107 subjects reporting resolution of symptoms, 42% reported within 6 months, 84% within 12 months, and 100% by 36 months. Figure 3 shows the cumulative resolution rate by age group.
Univariate analysis using the log-rank test showed that younger age, PVG, shorter duration of symptoms, response to addition of promotility therapy, presence of nausea, and absence of MD are associated with resolution of symptoms. In contrast, the presence of vomiting and abdominal pain showed no association, whereas male sex showed a tendency toward an association (P = 0.06) with symptom resolution. (Table 3). Multivariate survival analysis with Cox proportional hazards models demonstrated that longer duration of symptoms, older age, and presence of MD lead to lower rates of resolution, whereas a positive response to promotility therapy predicts a higher rate of resolution (Table 4).
Despite the prevalence of gastroparesis at all ages, little information is available on the clinical presentation and response to therapy in the pediatric population. The goal of the present study was to improve our understanding of childhood gastroparesis, to determine which therapies are most effective, and to identify the predictors and time course of symptom resolution. Several limitations of the present study are worth noting, including its retrospective nature and the fact that it is a single-institution study, with possible institutional biases that may affect the observations made and their applicability at other institutions. One of the greatest limitations in studying pediatric gastroparesis, however, is the lack of a standardized definition. Technical approaches to the performance of GET in children vary widely, and normative values for gastric emptying times are not available. The only data available are based on small cohorts tested with varied protocols (3,4,6,7). These are significant limitations that require prospective controlled trials. Despite this, this review of a large group of children with gastroparesis, defined as gastric emptying ≤40% at 60 minutes, identifies important clinical features that can help clinicians treat this difficult condition.
Idiopathic gastroparesis in adults is known to be more prevalent among women (8). Interestingly, we find that the male-to-female incidence changes with age. Among infants, boys predominate; in children, the sex ratio is equal; and in adolescents, girls are the majority. Clinical symptoms in children also vary with age. Infants experience more vomiting, possibly because they are less likely to report nausea or abdominal pain, which more commonly occurs in adolescents.
One of the most important factors to patients with gastroparesis and their families is knowing whether and when to expect resolution of this chronic and often debilitating condition. In the present study, the overall rate of symptom resolution was 52%, which was achieved at a median of 14 months from the time of diagnosis. In those patients in whom symptoms ultimately resolved, 84% did so by 12 months and all resolved by 36 months. Using multivariate regression, several factors were predictive of the likelihood of symptom resolution. These include younger age, shorter duration of symptoms, absence of MD, and response to promotility drugs. Adult studies have reported much lower resolution rates than we have observed in children. In a study of 21 patients with idiopathic GP, 30% had symptom resolution and were off medications at a mean follow-up of 5 years (9). Another study reported only 22% resolution of symptoms at a mean of 20 months. These findings suggest that pediatric and adult gastroparesis may be distinct disorders with varying outcomes.
PVG is a common cause of gastroparesis, accounting for 18% of all cases in the present study, with similar rates across all age groups. Reports in adults identified PVG in 18% (8) and 21% (9) of cases of idiopathic gastroparesis. In a previous retrospective case series of pediatric PVG, all 11 children reported resolution of symptoms by 24 months, with only 27% improving with promotility drugs (2). In contrast, of our 41 patients with PVG, 73% responded positively to promotility agents and 63% ultimately reported symptom resolution, findings that are similar to those seen in adults (8,9). In our study, a viral prodrome preceding gastroparesis was predictive of symptom resolution in univariate but not multivariate analysis.
The addition of promotility medications was associated with a favorable symptomatic response in 55% of subjects, independent of age. Metoclopramide use resulted in a low resolution rate and was associated with the highest rate of AEs among all of the promotility drugs, supporting a limited role for this drug in the management of gastroparetic children. Domperidone, however, produced the highest resolution rate and caused the fewest adverse effects. Consistent with our observations, others have also reported a higher incidence of neurological AEs from metoclopramide as compared to domperidone (10). More important, domperidone was used as a second-line promotility drug in all of our cases and resulted in significant improvement in cases in which metoclopramide did not. Tegaserod, a 5HT (hydroxytryptamine)-4 partial agonist removed from the market because of increased risk of stroke, heart attack, and unstable angina, resulted in a high response rate, but was associated with a high rate of adverse effects, none of which were cardiac in origin. Erythromycin produced the lowest resolution rate (5%), which is in contrast to studies in adults, in which marked clinical improvement with erythromycin has been reported (11). Other potential treatments, including cisapride, botulinum toxin injection to the pylorus (12), and gastric electrical stimulation (13,14), were not used in our patients.
We found by both univariate and multivariate regression analysis that a positive response to promotility drugs was predictive of resolution of gastroparesis symptoms. One may hypothesize that this results from the positive effects of these agents on gastric motility and emptying and/or the important antinausea effect of domperidone and metoclopramide; however, clinical trials in adults have not demonstrated a correlation between improvement in gastric emptying or motility and symptomatic relief (15–18). No improvement was found in gastric emptying times using metoclopramide in diabetic adult patients (19) and no correlation was seen between the improvement of gastric emptying after a single dose of intravenous metoclopramide and clinical symptomatic response to its chronic use (20). A significant clinical response to domperidone has been reported despite no measurable effect on gastric emptying (21,22), although others have shown improved gastric emptying after using domperidone for 6 months (23). In a pediatric study, 8 weeks of domperidone was superior to cisapride in improving symptoms as well as gastric emptying time and gastric myoelectrical disturbances (24), supporting a relation between improvements in measurable gastric function and clinical symptomatology. Small studies have also shown the efficacy of erythromycin in long-term use (25), with superiority to metoclopramide in improving both symptoms and gastric emptying time (26). More important, the methodology and gastric emptying protocols used differ across these studies. The higher rate of symptom resolution associated with promotility therapy may be explained by a milder degree of gastric dysfunction among responders; however, we found no difference in the severity of DGE between responders and nonresponders to support this. Prospective studies are needed to define standardized protocols for gastric emptying studies in children and age-based normative values, and also to define further the role of promotility drugs and other therapies in treating pediatric gastroparesis.
1. Reid B, DiLorenzo C, Travis L, et al. Diabetic gastroparesis
due to postprandial antral hypomotility in childhood. Pediatrics
2. Sigurdsson L, Flores A, Putnam PE, et al. Postviral gastroparesis
: presentation, treatment, and outcome. J Pediatr
3. Rosen PR, Treves S. The relationship of gastroesophageal reflux and gastric emptying in infants and children: concise communication. J Nucl Med
4. Montgomery M, Escobar-Billing R, Hellstrom PM, et al. Impaired gastric emptying in children with repaired esophageal atresia: a controlled study. J Pediatr Surg
5. Sachdeva P, Malhotra N, Pathikonda M, et al. Gastric emptying of solids and liquids for evaluation for gastroparesis
. Dig Dis Sci
6. Seibert JJ, Byrne WJ, Euler AR. Gastric emptying in children: unusual patterns detected by scintigraphy. AJR Am J Roentgenol
7. Signer E. Gastric emptying in newborns and young infants. Measurement of the rate of emptying using indium-113m-microcolloid. Acta Paediatr Scand
8. Parkman HP, Yates K, Hasler WL, et al. Clinical features of idiopathic gastroparesis
vary with sex, body mass, symptom onset, delay in gastric emptying, and gastroparesis
9. Bityutskiy LP, Soykan I, McCallum RW. Viral gastroparesis
: a subgroup of idiopathic gastroparesis
—clinical characteristics and long-term outcomes. Am J Gastroenterol
10. Patterson D, Abell T, Rothstein R, et al. A double-blind multicenter comparison of domperidone and metoclopramide in the treatment of diabetic patients with symptoms of gastroparesis
. Am J Gastroenterol
11. Dhir R, Richter JE. Erythromycin in the short- and long-term control of dyspepsia symptoms in patients with gastroparesis
. J Clin Gastroenterol
12. Woodward MN, Spicer RD. Intrapyloric botulinum toxin injection improves gastric emptying. J Pediatr Gastroenterol Nutr
13. Islam S, Vick LR, Runnels MJ, et al. Gastric electrical stimulation for children with intractable nausea
. J Pediatr Surg
14. Hyman P, Schropp K, Sarosiek I, et al. Feasibility and safety of gastric electrical stimulation for a child with intractable visceral pain and gastroparesis
. J Pediatr Gastroenterol Nutr
15. Anaparthy R, Pehlivanov N, Grady J, et al. Gastroparesis
-like syndrome: response to therapy and its predictors. Dig Dis Sci
16. Arts J, Caenepeel P, Verbeke K, et al. Influence of erythromycin on gastric emptying and meal related symptoms in functional dyspepsia with delayed gastric emptying
17. Sturm A, Holtmann G, Goebell H, et al. Prokinetics in patients with gastroparesis
: a systematic analysis. Digestion
18. Talley NJ. Diabetic gastropathy and prokinetics. Am J Gastroenterol
19. Schade RR, Dugas MC, Lhotsky DM, et al. Effect of metoclopramide on gastric liquid emptying in patients with diabetic gastroparesis
. Dig Dis Sci
20. Lyday WD 2nd, DiBaise JK. Metoclopramide-stimulated gastric emptying scintigraphy: does it predict symptom response to prokinetic therapy
in chronic gastroparesis
? Am J Gastroenterol
21. Horowitz M, Harding PE, Chatterton BE, et al. Acute and chronic effects of domperidone on gastric emptying in diabetic autonomic neuropathy. Dig Dis Sci
22. Koch KL, Stern RM, Stewart WR, et al. Gastric emptying and gastric myoelectrical activity in patients with diabetic gastroparesis
: effect of long-term domperidone treatment. Am J Gastroenterol
23. Soykan I, Sarosiek I, McCallum RW. The effect of chronic oral domperidone therapy on gastrointestinal symptoms, gastric emptying, and quality of life in patients with gastroparesis
. Am J Gastroenterol
24. Franzese A, Borrelli O, Corrado G, et al. Domperidone is more effective than cisapride in children with diabetic gastroparesis
. Aliment Pharmacol Ther
25. Richards RD, Davenport K, McCallum RW. The treatment of idiopathic and diabetic gastroparesis
with acute intravenous and chronic oral erythromycin. Am J Gastroenterol
26. Erbas T, Varoglu E, Erbas B, et al. Comparison of metoclopramide and erythromycin in the treatment of diabetic gastroparesis
. Diabetes Care