Clinical Characterization of Pediatric Gastroparesis Using a Four-Hour Gastric Emptying Scintigraphy Standard

Abstract

Objectives:

4-hour gastric emptying scintigraphy (GES) is the recommended method to identify both adult and childhood gastroparesis (GP). However, previous pediatric studies have not used this standard. We sought to determine the characteristics and outcomes of children vs. adolescents with GP using the 4-hour GES evaluation.

Methods:

We performed a retrospective chart review of pediatric patients diagnosed with GP by 4-hour GES (≥10% retention at 4 hours). Demographics, body mass index, gastroparesis-related symptoms, comorbidities, etiologies, therapies (e.g., medications), healthcare utilization, and response to therapy were captured systematically. Symptoms were compared from the initial vs. last gastroenterology visit. Outcomes were categorized as: no improvement; improvement (resolution of at least one symptom while remaining on therapy); and complete resolution of symptoms.

Results:

239 subjects (12.1 ± 4.1 years [mean ± SD], 70% female) were included. The identified characteristics of childhood GP were broad with idiopathic GP being the most common etiology. Outcomes over a median of 22 months [25–75%: 9.0–45.5 months] were: 34.8% no improvement, 34.8% some improvement, and 30.3% with complete symptom resolution. Compared to younger children, adolescents had a higher female predominance (P<0.01) and were more likely to have nausea (P=0.006). Females were more likely to have abdominal pain (P=0.001), nausea (P=0.03), and a documented diagnosis of dysautonomia (P=0.03). Males were more likely to have regurgitation (P=0.006), gastroesophageal reflux disease (P=0.02), and rumination (P=0.02).

Conclusions:

Using the 4-hour GES standard, childhood GP has broad clinical characteristics and outcomes. There are several significant age and sex-based differences in childhood GP.

BACKGROUND

Gastroparesis (GP) is a gastrointestinal motility disorder with abnormally delayed stomach emptying in the absence of a mechanical outlet obstruction(1). GP-associated symptoms in both adults and children include nausea, vomiting, early satiety, bloating, and abdominal pain(2). GP is diagnosed with gastric emptying scintigraphy study (GES). Though epidemiologic information related to pediatric GP is unknown, there is a marked increase in healthcare expenditures related to the care of children with GP(3, 4).

Previous studies focused on the clinical characterization of pediatric GP have identified age-based and gender-based differences(5, 6). However, these studies based the diagnosis of GP on institution-dependent non-standardized scintigraphy techniques(5, 6). Current guidelines specify that diagnosis of GP be based on a 4-hour GES and a standardized low-fat egg-based meal(7). Though pediatric normal values are not available, it has been recommended that the 4-hour GES with adult value cut-offs also be used in children(8–10). Similar to adults, 4-hour GES provides increased sensitivity in identifying GP in children as compared to shorter duration studies(8–11).

To our knowledge, a study evaluating the clinical characteristics of pediatric patients with GP diagnosed by 4-hour GES has not been reported. Therefore, we sought to determine the clinical characteristics (symptoms, co-existing diagnoses, management, healthcare utilization, and clinical outcomes) of pediatric patients with GP diagnosed using 4-hour GES. Additionally, we sought to investigate age-based (children vs. adolescents) and sex-based differences.

METHODS

We conducted a retrospective chart review using the electronic medical records of pediatric patients who had an abnormal 4-hour solid food GES at a quaternary children’s hospital. The chart review captured patients aged 18 years and younger diagnosed with GP from 2011 to 2017. Infants and toddlers unable to consume the standardized solid meal were excluded from the study. GES was performed at the discretion of a pediatric gastroenterologist in the course of standard clinical evaluation of gastrointestinal symptoms. Records were reviewed from the initial visit to the latest clinical visit (reviewed through 4/30/2019) to the Gastroenterology, Hepatology, and Nutrition service. The following was systematically captured: demographics (age and sex); body mass index (BMI); presence or absence of gastrointestinal symptoms (nausea, vomiting, abdominal pain, weight loss, poor weight gain, constipation, early satiety, diarrhea, belching, dysphagia); co-existing diagnoses (e.g., H. pylori infection); physician or nurse practitioner-specified GP etiology; healthcare utilization (including emergency room visits and hospitalizations related to GP-associated symptoms); medical and/or surgical treatment, and outcome. The diagnosis of gastroesophageal reflux disease (GERD) was based solely on provider documentation within the medical record. Not all patients had endoscopies or documented esophagitis. The characteristics and co-morbidities included were dependent on review of Gastroenterology clinic notes and medical provider entered ICD-10 diagnostic codes within the electronic medical record. The study was approved by the Baylor College of Medicine Institutional Review Board.

Clinical outcome categories were determined based on symptoms and medication usage at the time of the last visit in comparison to the baseline visit. These categories were defined as: no improvement; moderate improvement (full resolution of at least one baseline visit GP associated gastrointestinal symptom while remaining on therapy); and complete symptom resolution (no current symptoms with or without ongoing therapies).

Gastric Emptying Scintigraphy

GES studies were performed using the standard solid meal consisting of two pieces of white toast, 120 mL of scrambled egg substitute (the equivalent of 2 large eggs), a 15g packet of jelly, and 120 mL of water(7). Technetium-99-sulfur colloid was used for all studies and mixed in the egg substitute before cooking(7). Meals were consumed within a 10-minute period, after which a baseline image was obtained. All included participants ate the entire meal and were supine during the baseline and hourly scintigraphy image captures but otherwise allowed to sit and walk between images. Anterior and posterior images were acquired concurrently and a geometric mean value of gastric contents was calculated(7). Images were taken at 1-hour intervals over a 4-hour period. If a child was unable to consume the entire meal, required a substitute liquid/formula meal, or vomited within the 4-hour duration of the GES study, they were excluded from the analysis. GES results were reported as percent gastric retention values at the 1-, 2-, 3-, and 4- hour time points of the study. Delayed gastric emptying was defined as a gastric retention value greater than 10% at 4 hours(8, 10).

Statistical Analyses

For descriptive analyses, continuous parametric measures are presented as mean ± standard deviation and non-parametric measures are presented as median [25–75% quartiles]. For age-based analysis, we divided patients into two age groups: children (2–11 years) and adolescents (12–18 years). We evaluated the correlation of demographic and clinical data on symptom resolution. Chi-square analysis was used for categorical variables and Mann-Whitney testing was used for analysis of continuous, non-parametric measures. Statistical significance was defined as a P-value of <0.05. Statistical Package for the Social Sciences (SPSS) 25.0 (IBM Corp., Armonk, NY, USA) was used for statistical analyses.

RESULTS

In the review period, a total of 828 patients underwent 4-hour GES studies, of which 239 patients had delayed gastric emptying. These 239 records were reviewed and included in this analysis. Of these, 112 were children (47%) and 127 were adolescents (53%) (Table 1). Mean age at the time of the GES study was 8.3 ± 2.2 years for children and 15.5 ± 1.6 years for adolescents; for the entire cohort, 12.2 ± 4.1 years, with a 70% female predominance. The female predominance was greater in adolescents (78%) compared to children (62%; P=0.006; Table 1).

Table 1:

Symptoms by Age and Sex

	Children n=112			Adolescents n=127			Children vs Adolescents P Value	All Male vs Females P Value
	Male n=46 (%)	Female n=66 (%)	P Value	Male n=25 (%)	Female n=102 (%)	P Value
Abdominal pain	25 (54.3)	50 (75.8)	.018	16 (64.0)	81 (79.4)	.10	0.11	0.001
Vomiting	30 (65.2)	39 (59.1)	.51	12 (48.0)	70 (68.6)	.053	0.64	0.40
Nausea	23 (50.0)	38 (57.6)	.43	13 (52.0)	73 (73.7)	.061	0.008	0.04
Constipation	24 (52.2)	33 (50.0)	.82	11 (44.0)	44 (43.1)	.94	0.24	0.62
Early satiety	13 (28.3)	22 (33.3)	.57	7 (25.0)	29 (29.3)	.66	0.82	0.83
Weight loss	8 (17.4)	9 (13.6)	.57	9 (36.0)	38 (37.3)	.91	<.001	0.52
Regurgitation	13 (28.3)	6 (9.1)	.008	2 (8.0)	8 (7.8)	.98	0.03	0.006
Diarrhea	3 (6.5)	3 (4.5)	.65	3 (12.0)	9 (8.8)	.63	0.23	0.73
Distension	1 (2.2)	3 (4.5)	.51	0 (0)	10 (9.8)	.10	0.16	0.057
Poor weight gain	4 (8.7)	7 (10.6)	.74	3 (12.0)	1 (1.0)	.005	0.03	0.14
Belching	1 (2.2)	1 (1.5)	.80	1 (4.0)	4 (3.9)	.99	0.33	0.95
Dysphagia	2 (4.3)	0 (0)	.09	1 (4.0)	4 (3.9)	.99	0.45	0.44

The percent of underweight males was greater than females (BMI ≤ 5%: 14.1 vs 5.4; 5–85%: 54.9 vs 66.7; ≥ 85%: 31.0 vs 28.0; respectively, P=0.049). BMI was similar across both age-based groups (data not shown).

Symptoms by Age and Sex (Table 1)

The most common symptoms were abdominal pain, vomiting, nausea, with a smaller proportion experiencing constipation, early satiety, and weight loss (Table 1). A greater proportion of adolescents (vs. children) reported nausea and weight loss (Table 1). In contrast, a greater proportion of children reported symptoms of regurgitation and poor weight gain. Within our sex-based comparison of symptoms, females reported abdominal pain and nausea more frequently, and males reported regurgitation more frequently (Table 1). When data was divided by both age and sex, abdominal pain was reported more in female children than male children and regurgitation was reported more in male children.

Coexisting Diagnoses (Table 2)

Table 2:

Coexisting Diagnoses

	Child n=112	Adolescent n=127	P Value	Male n=71	Female n=168	P Value
Gastroesophageal reflux disease	53 (47.3)#	58 (45.7)	0.80	41 (57.7)	70 (41.7)	0.023
Migraines	11 (9.8)	25 (19.7)	0.03	4 (5.6)	32 (19)	0.008
Autism or Developmental Delay	13 (11.6)	10 (7.9)	0.34	14 (19.7)	9 (5.4)	0.001
Prematurity	12 (10.7)	5 (3.9)	0.04	9 (12.7)	8 (4.8)	0.03
POTS*/dysautonomia	3 (2.7)	14 (11.0)	0.01	1 (1.4)	16 (9.5)	0.027
Eosinophilic Digestive Disease	7 (6.3)	9 (7.1)	0.78	4 (5.6)	12 (7.1)	0.78
Lactose Intolerance	7 (6.3)	8 (6.3)	0.99	4 (5.6)	11 (6.5)	1.00
Ehlers Danlos Syndrome	7 (6.3)	8 (6.3)	0.99	3 (4.2)	12 (7.1)	0.56
Chromosomal Abnormality	7 (6.3)	6 (4.7)	0.60	7 (9.9)	6 (3.6)	0.05
Rumination	6 (5.4)	5 (3.9)	0.41	7 (9.9)	4 (2.4)	0.018
H. pylori	1 (0.9)	8 (6.3)	0.03	0	9 (5.4)	0.047
History of Fundoplication	3 (2.7)	4 (3.1)	0.57	4 (5.6)	3 (1.8)	0.2
Diabetes Mellitus	1 (0.9)	2 (1.6)	0.55	0	3 (1.3)	0.56

^#Percent

^*Postural orthostatic tachycardia syndrome

GERD was the most common coexisting disorder for all patients and was more prevalent in males than females as was autism, history of prematurity and developmental delay. In contrast, migraines and postural orthostatic tachycardia syndrome (POTS)/dysautonomia were more frequent in females. Though at a low prevalence, significantly more adolescents and females had a coexisting diagnosis of H. pylori infection.

Presumed Etiology (Table 3)

Table 3:

Presumed Etiology of Gastroparesis in 239 Children and Adolescents

	Children n=112	Adolescents n=127	P Value	Male n=71	Female n=168	P Value
Idiopathic	93 (83)#	98 (77)	0.26	56 (78.9)	135 (80.4)	0.86
Post-infectious	12 (11)	16 (13)	0.65	8 (11.3)	20 (11.9)	0.89
Dysautonomia	1 (1)	1 (1)	0.93	1 (1.4)	1 (0.6)	0.51
Connective Tissue Disease/ Ehlers Danlos Syndrome	1 (1)	5 (4)	0.13	0	6 (3.6)	0.19
Prior Fundoplication	3 (3)	4 (3)	.83	4 (5.6)	3 (1.8)	0.20
Other*	2 (2)	3 (2)	.56	2 (2.8)	3 (1.8)	0.64

^#Percent

^*Of those within “Other,” specified etiologies included: medication side effect (2), Salmonella colitis (1), rumination syndrome (1), uremic gastritis (1).

For the entire cohort and by both age-based and sex-based comparisons, the most commonly reported presumed GP etiology was idiopathic (>75%). A significantly greater percentage of adolescents had connective tissue disease/Ehlers-Danlos syndrome though at a low prevalence. Only females had a co-existing connective tissue disease/Ehlers-Danlos syndrome.

Therapies (Supplemental Table 1)

There were 14 different therapies used for the management of GP in our cohort. Overall, there were no significant age-group or sex-based differences identified (data not shown).

Healthcare Utilization and Follow-Up Duration (Supplemental Table 2)

No significant differences between age groups or by sex in the number of hospitalizations, emergency department visits, or clinic appointments was identified (data are not shown). However, median duration of follow-up was greater for children (23.3 months) than adolescents (20.4 months, P=.03)

Clinical Outcome

Of the 239 records reviewed, 201 (84%) had at least one follow-up visit allowing for outcome data collection. The duration of follow-up was longer in children than in adolescents, with a median duration greater than 3 months, with a similar number of follow-up visits between the two groups (Supplemental Table 2).

When reviewing the whole cohort, 34.8% had no improvement, 34.8% had moderate symptom improvement, and 30.3% had complete symptom resolution. The Figure illustrates the response to therapy recorded at the last clinic visit. Overall, we did not identify a significant difference in therapeutic response between children and adolescents or between males and females with GP.

DISCUSSION

We evaluated the clinical characteristics and outcomes of children and adolescents with GP diagnosed using the recommended 4-hour GES(7). We identified a broad range of clinical symptoms, co-existing diagnoses, ascribed etiologies, therapies, and outcomes. Importantly, we identified several age- and sex-based differences in symptoms, co-existing diagnoses, and presumed etiology.

Rodriguez et al. and Waseem et al., in somewhat smaller studies (n=194 and n=124, respectively) using institutional GES standards (>60% retention at 1 hour and half-time emptying by 45–90 minutes, respectively) also demonstrated a greater female prevalence in adolescents with GP(5, 6). In contrast, Rodriguez et al. noted a similar female to male proportion in children (48:52) whereas females predominated the child age group (62:38)(5). Beyond different GES techniques, this may be explained by our exclusion of infants and young toddlers who were included in their study(5). Comparison to the Waseem et al. study is challenging, as they did not differentiate between infants and children when providing data on sex(6). The factors accounting for differences in sex distribution in children with GP remain to be determined.

The prevalence of abdominal pain in our cohort (67% of children, 76% of adolescents) was high with only slightly fewer children and adolescents reporting vomiting and nausea (Table 1). The prevalence of these three symptoms was greater than reported by Rodriguez et al. in which half or less of children and adolescents reported abdominal pain, vomiting, or nausea(5). The proportion of children and adolescents reporting nausea in the previous studies was similar (28.5% patient reported nausea in Waseem study)(5, 6). We identified that adolescents were more likely to report nausea compared with children (Table 1)(5). One potential explanation may be the increased ability of adolescents relative to younger children to interpret and verbalize nausea as a symptom.

Females, regardless of age, were more likely to report abdominal pain and nausea (Table 1), whereas males were more likely to describe regurgitation (Table 1). Overall, our data suggest that symptoms are more sex- rather than age-related; an observation that extends previous findings that did not evaluate age and sex effects directly(5, 6). These sex-based symptoms in pediatric upper gastrointestinal disorders are consistent with recent data identifying a predominance of females with functional nausea(12).

With the exception of H pylori infection, coexisting diagnosis prevalence did not differ between children and adolescents (Table 2). Most, but not all, studies suggest that H. pylori does not affect gastric emptying significantly and our data matches adult studies suggesting a low prevalence of H. pylori infection in GP(13–15).

Similar to findings for GP symptoms, there was a sex difference between cohorts. GERD was the most common coexisting diagnosis and was more frequent in males, consistent with males describing regurgitation more commonly than females (Table 2). Autism or developmental delay were found more frequently in males (Table 2), aligning with the known male preponderance in the general population(16). In contrast, migraines and POTS/dysautonomia occurred more frequently in females (Table 2). The increased frequency of migraines in females may be unique to GP as the prevalence of migraines in the general population is similar in childhood with only a small female preponderance in adolescence in contrast to the 3-fold difference in our study (Table 2)(17). The increased frequency of POTS/dysautonomia in females in our study reflects the similar population incidence of higher autonomic dysfunction in female populations(18). Our data underscores the importance of sex as a biological variable in pediatric GP. Within adult populations of GP, sex-based differences in presenting symptoms have been observed (i.e., female predominance, females with more severe nausea, abdominal pain, early satiety, constipation, and overall GP symptoms)(19, 20). Proposed mechanism(s) within adult populations for these differences include: effects of menstrual cycle, hormones, and sex-specific differences between how the gastric neuromuscular apparatus may be affected(20–22). Whether these physiologic factors play a role in children with GP remains to be determined, particularly in pubertal children. Future prospective evaluations accounting for aspects such pubertal status and timing of the menstrual cycle may further elucidate sex-based differences in symptoms and etiology.

Attributing an etiology to pediatric-onset GP is challenging, as the vast majority within our cohort had a presumed idiopathic etiology. To gain further insight into contributing pathophysiologic factors, we evaluated co-existing diagnoses. Despite idiopathic designation, there often were co-existing conditions identified that may potentially cause/exacerbate GP including dysautonomia, connective tissue disorders, and history of fundoplication(2, 23–28). Within our patient population, it is possible that certain co-existing diagnoses, such as previous fundoplication, may have treated underlying symptomatology of GP (i.e., severe reflux) rather than being a contributing cause of GP. Current studies from the NIH-sponsored Gastroparesis Consortium may elucidate etiologies of pediatric GP. The Consortium has reported in adults with idiopathic GP that gastric myenteric immune infiltration is associated with worse symptoms, with loss of anti-inflammatory macrophages and transcriptomic signature evidence of enrichment proinflammatory macrophages within gastric tissues(29–31). Whether these changes are found in pediatric patients with GP remains to be determined.

Our data suggest that physicians use a wide range of therapies to treat pediatric GP (Supplemental Table 1). We did not identify age- or sex-based differences in treatment. As in adults, therapies focused on addressing symptoms and gastric emptying using prokinetics(1). Unfortunately, there is a relative dearth of high quality randomized controlled trials evaluating therapies in children with GP(2). Nevertheless, a group of pediatric gastroenterology experts have recently provided a treatment algorithm for pediatric gastroparesis which includes many of the therapies identified in this study(24). Future studies evaluating the proposed algorithm versus other treatment approaches in pediatric GP and taking into consideration differences in symptoms, age, and sex are needed.

Over a median period approaching two years, we identified that the majority of children with GP continued to have some symptoms. The outcome for the patients in our cohort (34.8% with moderate and 30.3% with complete symptom resolution) appears to be better than that in adults followed at quaternary care centers, in which only 28% of adults showed any sign of symptomatic improvement(32). Though it remains to be investigated, we speculate there is a subgroup of children with GP who had an unidentified post-viral etiology or another transient process that occurs at a higher proportion than that found in adults with GP. Enteric nervous system adaptation and maturation in children also may serve as a potential reason for pediatric GP improvement. Previous studies in both children and adults with GP have documented complete resolution of symptoms following a post-viral etiology within a few years(33, 34).

We did not identify differences in clinical outcome in our cohort based on age or sex. In contrast, Rodriguez et al. reported that children had a more rapid recovery relative to adolescents(5). However, they defined GP based on a 90-minute study(5).

Limitations to our study primarily relate to its retrospective approach. Prescribed medications can be captured objectively retrospectively, while aspects such as symptom reporting are dependent on documentation within the medical record by the physician or nurse practitioner. Future studies using prospective questionnaires and prospectively determined clinical outcomes are needed. In particular, co-existing diagnoses were recorded based on provider documentation. We acknowledge that co-existing diagnoses, such as gastroesophageal reflux disease and dysautonomia, may have been empirically diagnosed. In addition, rather than representing a co-existing diagnosis, in some cases, they may have represented underlying symptomatology of GP. Finally, as a limitation to generalizability, the usual GES technique of using only anterior and posterior imaging to determine the ‘area of interest’ was not standardized; therefore, errors such as including the duodenum may have occurred. However, all interpreting radiologists are trained to read the images as part of standard clinical care.

There are several strengths of this study. A unique strength relates to the use of the 4-hour GES for the diagnosis of GP. The use of this accepted international standard helps with the generalization of our findings. The longitudinal nature by which children with GP were followed with a median follow-up time approaching two years is another strength which allows for more robust outcome data. An additional strength likely to increase the generalizability of the findings is that the evaluated cohort underwent standard medical care and was not a clinical trial population.

In conclusion, childhood and adolescent GP determined using a 4-hour GES has a broad clinical presentation. We identified several age and sex-based differences that warrant further prospective evaluation. Further investigations to evaluate other testing modalities for gastroparesis in pediatric populations are also necessary, particularly those that may spare radiation exposure. Physicians and other health care providers should consider assessing these broad characteristics when caring for children with gastroparesis.

Figure:

Response to therapy recorded at last clinic visit grouped by age and sex

What is known?

• Current guidelines specify that the diagnosis of gastroparesis (GP) is made by a 4-hour gastric emptying scintigraphy study (GES)

• Previous reviews of pediatric GP have made the diagnosis using institution-dependent diagnostic standards

What is new?

• We captured several age-based and sex-based differences in our pediatric cohort with GP diagnosed via 4-hour GES

• Our results suggest these demographic factors are important as they relate to the symptoms and co-existing diagnoses in children with GP.