Spectrum of gastric emptying patterns in adult patients with cyclic vomiting syndrome


  • A part of this study was presented in poster form at the American College of Gastroenterology (ACG) Meeting, Orlando, FL, USA. Oct 2008.

Address for Correspondence
Richard W. McCallum, MD, FACP, FRACP (Aust), FACG, Founding Chair, Department of Internal Medicine, Paul L. Foster School of Medicine Texas Tech University Health Sciences Center, 4800 Alberta Ave El Paso, TX 79905, USA.
Tel: +1 915 545 6626; fax: +1 915 545 6634;
e-mail: richard.mccallum@ttuhsc.edu


Background  Cyclic vomiting syndrome (CVS) in adults is a disorder characterized by recurrent and stereotypic episodes of severe nausea, vomiting and abdominal pain separated by symptom-free intervals. Both rapid and delayed gastric emptying (GE) have been observed but the reports involved small numbers of CVS patients.

Methods  We performed a retrospective study of 92 adult patients who met Rome Ш diagnostic criteria for CVS between 2003 and 2009 at the Kansas University Medical Center. Gastric emptying was measured by a standardized scintigraphic method involving a low fat (2%) isotope labeled egg white meal of 250 Kcal, with anterior and posterior gastric imaging in the standing position obtained at 0, 1, 2, 4 h after meal ingestion. Rapid GE was defined as <50% isotope retention at 1st h and/or <30% at 2nd h and delayed GE as >10% at 4 h.

Key Results  Ninety two patients were analyzed: 47 males and 45 females mean of age 37 ± 12 years (range: 20–68 years). There were 27 patients with a personal history of migraine headache, 30 with history of marijuana use, 12 had diabetes mellitus (DM) and 10 had irritable bowel syndrome (IBS) as an accompanying diagnosis. Fifty four patients (59%) met criteria for rapid GE, 25 (27%) had normal GE and 13 (14%) had slow GE. Eighty percent of patients with co-existing IBS symptoms were identified as rapid. The subset with delayed emptying was often associated with narcotics use, DM and marijuana use (P < 0.05).

Conclusions & Inferences  (i) In adult CVS patients, GE is generally either rapid or normal. (ii) Cyclic vomiting syndrome is an important new etiology to explain the finding of rapid GE on a radionuclide test. (iii) The small subset of CVS patients (14%) whose GE was slow were explained by the role of narcotics and/or marijuana.


Cyclic vomiting syndrome (CVS) is a disease characterized by recurrent episodes of incapacitating nausea and vomiting interspersed with relatively symptom-free intervals that might last anywhere from a few weeks to months.1 It was first described in the late 19th century and was essentially thought to be a disorder of childhood, but it is increasingly being recognized in adults over the last 20 years.2 The diagnostic criteria for adult CVS according to the Rome III criteria are stereotypic episodes of vomiting with the following characteristics: a sudden or acute onset; duration usually less than a week; three or more discrete episodes occurring in the prior year and absence of symptoms between attacks. Though the Rome III criteria were defined originally for the pediatric age group, actually the adult onset of CVS is defined as a separate category in Rome III classification.3 In adults abdominal pain invariably accompanies the acute episodes of nausea and vomiting. Migraine headache is reported as a frequent co-morbid disease with CVS3 as well as anxiety and/or depression.4,5 Although the diagnosis of CVS is based on Rome criteria in the setting of a clinical history of cycles of nausea, vomiting and abdominal pain some diagnostic tests are required such as, abdominal ultrasound, esophagogastroduodenoscopy and gastric emptying (GE) to rule out other differential diagnoses such as mechanical obstruction and to evaluate the co-morbid disorders.

While CVS is being increasingly recognized in the adult population, there is a paucity of data as to the GE pattern. Gastric emptying has been reported in a few studies to show that rapid emptying is common in adult CVS patients. These tended to be studies with small numbers of patients without the use of established GE criteria for identifying rapid or slow rates of emptying.6–9 Normal GE in a pediatric population using ultrasound method for measuring GE was noted.10 There are also limited reports of slow GE in pediatric11 and adult CVS populations.12 Hence, there is a need for a larger population study using a standardized isotope labeled meal to clarifying the GE profile in adult CVS.

Therefore the aims of our research in adult CVS patients were the following:

  • 1 Investigate the GE profile using the standardized 4-h isotope labeled meal.
  • 2 Identify the proportion of patients with normal, rapid and slow GE pattern based on normal values for the percent of isotope retention in each hour.
  • 3 Determine if there are demographic or clinical profile characteristics that are associated with these different GE patterns.

Patients and methods

Study participants

Patients with the primary diagnosis of CVS were eligible for participation in the study. All patients were initially referred to our academic gastrointestinal (GI) motility center for outpatient evaluation because of upper GI symptoms, and all underwent careful history taking and clinical examination, upper GI endoscopy, routine biochemistry, upper abdominal ultrasound and a gastric emptying test (GET) as part of their assessment before making the diagnosis of adult CVS. Only patients in whom we established the diagnosis of CVS based on Rome III criteria and who had no other abnormalities identified by these additional studies were included in the report. Patients with documented organic or intestinal pseudo- obstruction, primary eating or swallowing disorders, rumination syndrome, psychogenic vomiting, systemic sclerosis, thyroid and adrenal problems, chemical dependency, a diagnosis of cancer, renal failure or a positive pregnancy test were excluded. The study protocol was approved by the Human Subjects Committee at the Kansas University Medical Center.


In a retrospective study all adult patients referred to one of the investigators (RMC) in our academic GI motility center for management of nausea/vomiting in whom the diagnosis of CVS was made and whose GET results were available were included. The GET was performed as part of the initial evaluation. In addition, the patients were questioned regarding chronic marijuana smoking which was defined as daily smoking for two or more years.

Gastric emptying studies

Gastric emptying time was assessed by the standard 4-h scintigraphic method using a low fat (2%) solid meal.13–15 Scintigraphic GET was performed in the morning after an overnight fast with prokinetics and proton pump inhibitors stopped for at least 3 days and narcotic medication withheld for more than 12 h. This standardized method for GE consists of a scrambled egg substitute [120 g of Free Cholesterol & Fat Fee Egg, Sunny Fresh Foods, Inc., Monticello, MN, USA; (60 kcal)] labeled with 99mTc sulfur-colloid (1 mCi), two slices of whole wheat bread (120 kcal), 30 g jelly (75 kcal), and 120 mL of water. The meal has a total caloric value of 255 kcal (nutritional composition: 72% carbohydrate, 24% protein, 2% fat, and 2% fiber). Anterior and posterior images of the stomach in the standing position were taken immediately after eating, and then hourly for 4 h. During the 4 h of the assessment all patients remained upright, either walking or sitting. For diabetic patients undergoing GE their blood glucose level was required to be <275 mg dL−1 at the time of ingesting the test meal. There was no reclining or sleep permitted. Gastric retention of gamma counts was calculated by the Department of Nuclear Medicine using geometric and decay correction.

Delayed GE for this study was defined as isotope retention of >10% at 4 h based on normal data established for this standardized GET. Rapid emptying was defined as isotope retention <50% at 1st h and/or <30% at 2nd h. The values of normal, fast and delayed were based on a multicenter study, which provided GE values in healthy subjects of large age range and both genders utilizing data obtained in a large sample size with that same meal and methodology.9,13,14 All these GE studies were done during the remission phase of the disease.

Study protocol

Demographic data and a list of medical illnesses were recorded. The GET results were interpreted by the Department of Nuclear Medicine. Chart review and telephone interview were sometimes required for obtaining additional data.

Statistical analysis

Descriptive statistics are presented as means and standard deviations (SD) for variables measured on a continuous scale (e.g., age) and as frequencies and proportions for variables measured on a discrete scale (e.g., gender). Statistical tests were performed for comparisons of clinical characteristics among GET subgroups and measured as discrete variables, thus Fisher’s exact tests were performed for each comparison with P-values presented.


Ninety two adult patients (47 male), mean age 37 ± 12 years (range 20–68 years) with the diagnosis of CVS between 2003 and 2009 were included in this study (Table 1). Of those, 54 subjects (59%) were identified as having rapid GET based on the predefined criteria and 25 (27%) had GET within the normal range. The remaining 13 (14%) had slow GET (Table 2). The mean isotope percent retention at 1 h in the rapid group was 41 ± 6% (range 23–45%) and for delayed group was 21 ± 5 at 4 h (range 12–38%). History of migraine headache was positive in 27 (21%). Diabetes mellitus (DM), irritable bowel syndrome (IBS) meeting Rome III criteria and psychological disorder meeting DSM IV criteria as accompanying co-morbid disorder were observed, respectively in 13%, 11% and 15% patients. The mean duration of DM was 43 ± 15 months and no peripheral neuropathy or nephropathy was observed. Eight of ten (80%) of patients with IBS had accompanying rapid GET and remaining two had normal GE. Diabetes mellitus, chronic narcotics use and marijuana smoking were significantly more frequent in the group identified with delayed GET (P < 0.05). Tables 1, 2 and 3 summarize demographic, clinical profile and GET results in the 92 patients with CVS.

Table 1.   Demographic characteristics of 92 adult patients with cyclic vomiting syndrome
Characteristicn (%)
 Male47 (51)
 Female45 (49)
 Caucasian79 (85)
 African–American11 (12)
 Hispanic2 (3)
Mean age (years)37 ± 12 years (range 20–68)
Age of onset22 ± 2 years
Age of diagnosis28 ± 3 years
Psychological disorder14 (15)
 Depression5 (5)
 Anxiety disorder9 (10)
Migraine headache27 (21)
Smoking29 (31)
Chronic marijuana use30 (32)
Chronic narcotic use18 (19)
Diabetes mellitus12 (13)
Irritable bowel syndrome10 (11)
Table 2.   Summary of gastric emptying test results in 92 adult with cyclic vomiting patients
Mean GET percent retention ± SDNormal GET (n = 25)Rapid GET (n = 54)Slow GET (n = 13)
  1. GET, gastric emptying test; h, hour.

1st h84 ± 341 ± 693 ± 5
2nd h62 ± 427 ± 778 ± 6
4th h6 ± 34 ± 321 ± 5
Table 3.   Summary of clinical characteristics of 92 adult patients with cyclic vomiting syndrome and the results of gastric emptying test
CharacteristicNormal GETRapid GETSlow GETP-value*
  1. GET, gastric emptying test.

  2. *P-value calculated from Fisher’s exact method.

Age (years), Mean ± SD35 ± 437 ± 533 ± 3
Gender, %
Race/Ethnicity, %
Depression, %
Anxiety disorder, %
Migraine headache, %
Smoking, %
Chronic marijuana use, %
Chronic narcotic use, %
Diabetes mellitus, %
Irritable bowel syndrome, %


Our study, the largest reported series of adult patients with CVS in whom GE data are available shows that the most common finding was actually rapid GET observed in 59% of patients followed by normal GE which was identified in 27% of patients. Also, a subset of 14% of patients had slow GE. This data means that CVS is a new addition to the differential diagnosis of rapid GE. This new observation is very relevant for clinicians who may already suspect the diagnosis of CVS when evaluating vomiting states. Identifying rapid or normal GE would provide confirmatory evidence of CVS and set the stage for initiating appropriate medical treatment.

Gastric emptying and myoelectric activity utilizing the electrogastrogram have been investigated in adult patients with CVS, primarily during asymptomatic periods between emetic episodes. Our group previously reported that 23 of 30 (77%) adult CVS patients had an accelerated GET using 2 h criteria with less than 30% retention at 2 h.4 Similar results were noted by Fajardo et al., using similar 2 h criteria.7 They compared 64 CVS patients with 364 controls and showed a rapid early-phase GE (gastric retention<30% at 2 h) in CVS patients with an odds ratio of 13 (95% CI 5–33).7 The requirement for <50% retention at 1 h was not the criteria applied to these studies. In order to be consistent and label someone as rapid GE based on the isotope labeled egg beater meal method, there needs to be adherence to accepted criteria for rapid emptying.13–15

In the reports of slow GE in CVS the emptying studies were generally performed in the hospital where they would have been affected by narcotic medications being given to the patient. Interestingly, delayed GE has actually been found during the vomiting phase in a few patients in whom this test was able to be performed.4 This delay in GE during the symptomatic phase may represent a generalized response to central nausea and abdominal pain. Therefore, GE studies should be performed during the ‘remission’ phase when there are minimal, if any, symptoms and no narcotic medications are being received.

The presence of rapid GE was observed during the remission phase of the disease. We do not believe that this rapid GE is the cause of cyclical CVS episodes. It is an interesting observation that may explain some of the symptoms present between cyclical relapses. For example, we observed in this rapid group a low grade of nausea and dyspepsia and up to 15% had IBS symptoms. We believe this spectrum of findings could be attributed to the rapid GE inducing a prominent gastro-colic reflex in a subset of patients. These patients do not present with chronic symptoms of dumping with meals even though their GE is in a similar range to the patients with symptomatic ‘dumping syndrome’ that we recently published.14 We cannot explain why the CVS patients tolerate this rapid GE with only minor symptoms. Rapid GE does not result in CVS attacks and in fact we have actually observed impaired or slow GE during CVS attacks4. Interpreting why rapid GE is present in the vomiting-free period has led to speculation that there is an underlying autonomic dysfunction.6 This is consistent with data that rapid GE is reported more commonly than delayed GE in symptomatic patients with autonomic dysfunction.8 We have recently pursued a number of research studies to explain this finding of rapid emptying in CVS patients. One report addresses the frequency of autonomic dysfunction in adult CVS patients and indicates that although a majority [13/21 (62%)] of patients had rapid GET (<50% retention at 1 h), the frequency of abnormal autonomic nerve function was not significantly higher in the rapid GET subgroup.16

In another study, we investigated the possibility that humoral and hormonal aspects could be an explanation. The hypothesis that humoral factors may have a pivotal role in the pathogenesis of dumping was brought to light after dumping symptoms were induced experimentally in a healthy dog after transfusion of blood from a portal vein obtained from another dog with dumping.17 The initial studies indicated that serotonin and the kallikrein–kinin system may play a role in provoking dumping. Higher postprandial levels of the gut hormones such as pancreatic polypeptide, neurotensin, serotonin, enteroglucagon, peptide YY and glucagons like peptide have been documented in the patients with dumping syndrome.17,18 Therefore, post-prandial hormone release could be suggested as a cause of the rapid emptying in CVS.

Ghrelin, a hormone locally produced in gut mucosa and released before eating has been shown to accelerate GE in humans.19 It has been postulated that it could be contributing to the rapid GE status.19,20 In a recent study, we observed that serum ghrelin levels in adult CVS patients are significantly elevated compared to normal subjects. Also, the elevated serum ghrelin levels were associated with the rapid GET subset of the CVS cohort.21

Irritable bowel syndrome is more observed in patients with rapid GET, consistent with the accompanying profound gastro-colic reflex. Hence, 80% of our CVS patients with accompanying IBS symptoms had rapid GET.

Migraine headache was not significantly different in frequency among the GET groups. The current therapy for CVS is tricyclic antidepressants (TCA) often given in high doses (e.g. 1 mg kg−1).4,22 Besides having a role in reducing migraine headache, those agents address the central control mechanisms for nausea/vomiting but also decrease gut motility through their anticholinergic properties.22,23 Hence, high dose TCAs could be reducing cycles of CVS by modifying gut motility and rapid emptying as well as reducing the gastro-colic reflex and related pain.

We also identified slow GE in 13 (14%) of patients of our cohort. In this subset, chronic narcotics were being utilized in a significant number [8/13 (62%)] for pain. The combined effects of narcotics, anticholinergics, and TCAs commonly used in CVS would decrease the GET rate.24 Another possible explanation for the subset with slow GE is the effect of chronic marijuana use on slowing GE.25 In our report, the frequency of chronic marijuana smoking was significantly higher in the delayed GET group involving 11/13 patients (85%), significantly higher than in the normal (40%) and rapid (17%) GET groups. Six out of 13 of our delayed GE subset also had DM, a known cause for delayed GE which can be associated with gastroparesis symptoms. However, our diabetic subset with delayed GE had a short duration of DM (<5 years), and none had a peripheral neuropathy or nephropathy, entities usually present in the profile of diabetic gastroparesis.26 In addition a subset of diabetics with gastroparesis has been described by Abell et al. who had cyclical symptom patterns of nausea, vomiting, and degrees of abdominal pain.27 Our data would support the conclusion that CVS patients only have delayed GE as outpatients if using narcotics or are chronic marijuana smokers or if studied during hospitalizations when they are receiving narcotics for pain control.

We also demonstrated that the frequencies of co-existing psychological disorders such as depression or anxiety disorder were similar among the different GET groups. It is still unknown whether mental illness contributes to the symptoms, or coexists with CVS.

A limitation in our study is that this is a retrospective and a prospective study with a control group using the same standard GET would be ideal.

In conclusion, our data in adult CVS patients provides new information that in the largest reported series undergoing a standardized 4-h scintigraphic meal GET, the GE is usually rapid or normal. The clinical message is that when evaluating vomiting states, the finding of a rapid GE should alert the clinicians to consider CVS in the differential diagnosis. We believe that rapid GE could be considered as a new supporting criterion for the diagnosis of CVS in adult patients and recommend further prospective studies in larger populations and multiple centers with standard radionuclide 4-h GET to better evaluate this observation.


Faculty, fellows, and staff of the Division of Gastroenterology at Kansas University Medical Center, for their clinical involvement with these patients, and nuclear medicine faculty for reviewing the GE studies. Also to Dr. Patrick Tarwater, Professor of Biomedical Sciences in Texas Tech University Health Sciences Center at El Paso for his statistical advice and input.

Conflict of Interest

No conflict of interest exist.

Financial Disclosure