Children with cyclic vomiting syndrome: phenotypes, disease burden and mitochondrial DNA analysis

Cyclic vomiting syndrome (CVS) is characterized by episodic attacks of vomiting and symptom-free interval periods [1, 2]. Prevalence of CVS is reported to be 0.3–3.8% among children based on the Rome III criteria [3‐5]. About 46% patients have onset of disease before 3 years [6].

According to the Rome IV criteria [7], diagnosis of CVS requires occurrence of two or more periods of intense, stereotypical vomiting within 6 months. Because of possible underlying organic causes, appropriate medical evaluation should be performed [8].

Mitochondrial DNA (mtDNA) is maternally inherited [9], and variants adversely affect energy metabolism [10]. CVS is associated with mtDNA variants [11]. A3243G mtDNA mutation was found in a family with CVS [12]. Pediatric CVS were found to be associated with C16519T and G3010A polymorphism [10].

This study described clinical characteristics, disease burden and treatment of pediatric CVS. We also performed mtDNA sequencing among Chinese pediatric CVS patients.

There is a paucity of data about pediatric CVS in China. To our knowledge, this is the largest clinical study and genetic analysis on mtDNA of pediatric CVS in China.

In this study, patients were diagnosed as pediatric CVS fulfilling the Rome IV criteria. They had two or more periods of intense, unremitting nausea and paroxysmal vomiting within a six-month periods. These episodes are stereotypical and interspersed with symptom-free interval periods [7]. According to the 2008 Consensus Statement on the Diagnosis and Management of Cyclic Vomiting Syndrome by North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition, pediatric CVS generally requires at least five episodes of vomiting in a year [8]. There were statements which criticized the minimum of two attacks in both the Rome III and Rome IV criteria for lacking specificity [2]. However, studies utilizing the Rome III and IV criterion did not show significantly higher prevalence rates [4, 13]. Based on these data and the impact of CVS attacks for the quality of life of children and their families, the Rome IV working group decided that early diagnosis is important and left the minimum number of two episodes in CVS diagnosis unchanged [2, 7]. Over the last decade, there have been advancements in the pathogenesis and diagnostic criteria in pediatric CVS [14]. Therefore, we utilized the latest Rome IV in this study.

Our study showed a mean onset of 4.0±3.4 years among patients. They had recognizable prodromes and stereotypical attacks. These results were consistent with other studies on CVS in Chinese children [15‐18]. Reports involving patients in other populations also showed a similar pattern of disease onset and vomiting characteristics [19, 20]. However, our study showed that only 11.9% of patients had family history of migraine and 2.3% had personal history of migraine. The frequency of personal and/or family history of migraine in this study seemed lower than that reported in western populations [21], which was about 39–82% [22, 23]. Dong et al. reported that personal migraine history was seen only in 7.3% of Chinese children with CVS [18]. Prevalence of migraine in the general Chinese population was 9.3% in a nationwide survey [24], compared to 14.9% in United States [25]. Study found that under-diagnosis and misdiagnosis of migraine were common in China, which might be an explanation for the low frequency of personal/family history of migraine among our cohort [26].

Moreover, parent-reported coffee-ground vomitus was present in 12 (28.6%) of subjects during attacks. All of them were treated in emergency departments in local hospitals with hydration and supportive care, and half received gastric acid suppressants. Therefore, there was no delay in medical care during attacks among them. However, abortive therapy was only instituted in one patient with ondansetron. After admission to our institution, ten underwent upper endoscopy, and only one was diagnosed with esophagitis. Repetitive nausea, vomiting, and retching in severe cases might cause hematemesis because of mucosal tear (Mallory-Wise syndrome) [27]. Pediatric CVS might be under-diagnosed and misdiagnosed in local hospitals, simply as “gastritis, food poisoning” due to physicians’ inadequate knowledge about pediatric functional gastrointestinal disorders. Therefore, delay in diagnosis was common [22, 28]. Lack of abortive care might lead to prolonged emetic phases and subsequent mucosal tear in our patients, which maybe the explanation for high frequency of parent-reported coffee ground vomitus. Upper endoscopy after admission ruled out severe mucosal diseases.

There is no randomized controlled trial on treatment of CVS. Retrospective studies showed favorable clinical responses with cyproheptadine [29], propanolol and pizotifen [30]. Valproate is also considered as an effective prophylactic therapy for pediatric CVS [31]. Patients admitted to this tertiary care center generally received prophylactic medications, including cyproheptadine, valproate acid, or both. More than half of the patients had been consulted by a pediatric psychologist. Integration of psychotherapy is beneficial, as psychologic factors are associated with functional gastrointestinal disorders including CVS [32]. However, most patients referred from local hospitals in this study did not receive prophylactic therapy and multidisciplinary treatment involving pediatric psychologists before admission.

The C16519T polymorphism locates in the non-coding mtDNA control region, and is six-fold more commonly seen in pediatric CVS than in controls [10]. The G3010A polymorphism locates in the 16S-ribosomal RNA gene [10], which increases the odds ratio for CVS by 17-fold in subjects with C16519T. These two polymorphisms are not found in adult CVS patients [33]. Han et al. performed mtDNA sequencing among four Chinese children with CVS plus, (CVS with + neurocognitive disorders), in which two had A3243G mutation and two had C16519T polymorphism [15]. In this study, none A3243G mutation has been identified, 4/13 (30.8%) patients had C16519T, and 2/13 (15.4%) had G3010A polymorphism. However, considering the relatively high prevalence of C16519T and G3010A polymorphism from an in-house database of Chinese subjects and lack of frequency of these polymorphisms among large cohort of healthy controls, it is not sufficient to detect significant associations among mtDNA polymorphism with Chinese pediatric CVS patients.

In the study, there were mtDNA variants among patients associated with other diseases. Homoplasmic A15662G mutation was identified in case 3, which was previously reported in a patient with complex mitochondriopathy [34]. However, none functional analysis had been performed, and three other variants were identified in this patient (T3398C, T4216C, and G15812A) [34]. This subject did not have manifestations of mitochondrial diseases. Heteroplasmic A4136G mutation was identified in case 9, which was previously reported in a family with Leber hereditary optic neuropathy (LHON) [35]. Other mtDNA mutations were also found in these patients [35]. In another study, A4136G was described in pediatric patients with mitochondrial encephalopathies [36]. Patients harboring variants A4136G did not have LHON, but present with clinical symptoms of mitochondrial disease [36].

Because of repeated visits to emergency department during attacks and diagnostic studies to rule out organic causes, there were substantial disease burdens caused by CVS to patients and families. A national study among adults in United States showed that there were 20,952 CVS hospitalizations during a two-year period, and caused about 400 millions US dollars of hospital charges [23]. Emergency department visits were 2.3±2.0 per year among pediatric CVS patients [37]. It also imposes a major impact on quality of life [38]. Patients generally have lower health-related quality of live than healthy controls [39]. Failure to recognize CVS occurred in 93% of emergency department visits in previous report [40], which was consistent in this study. None of our patients were diagnosed as CVS in other hospitals. Also, 40 patients had multiple outpatient visits and prior hospitalizations, which further increased the disease burden.

However, our results should be interpreted with caution. This was a retrospective study from a single tertiary care center. Only 13 patients underwent mtDNA sequencing due to the retrospective setting, and mtDNA from mothers of the patients were not available. This study cannot indicate associations between mtDNA polymorphisms with pediatric CVS. Moreover, there was possible selection bias, as only cases with severe cyclic vomiting attacks had been admitted and included in the study. A prospective study involving larger number of subjects and a cohort of healthy controls from the same populations are needed to confirm the relationship between mtDNA variants and CVS.

Pediatric CVS is a chronic and debilitating disease causing considerable disease burdens to patients and family. Early clinical suspicion and prompt diagnosis by pediatricians enable patients to receive adequate symptom-tailored management.