Cyclic vomiting syndrome (CVS) is a functional gastrointestinal disorder that affects both adults and children of all ages and presents with characteristic manifestations of acute, stereotyped and recurrent episodes of intense nausea and vomiting lasting from hours to a few days (1, 2). The worldwide prevalence of pediatric CVS is 1.9–2.3% (2), with the incidence of 3.1 per 100,000 per year (3), and the female:male ratio is 1.5:1 (1). Coexisting neuromuscular disease and developmental delay are found in up to 25% of patients with CVS, known as CVS plus (4).

CVS is strongly associated with migraine as ~25–35% of children with CVS develop migraine headache later in adulthood and there is a high prevalence of migraine headache in their families (5–9). Currently, according to the International Classification of Headache Disorders, CVS is classified as an “episodic syndrome that may be associated with migraine” (10). The etiopathogenesis of CVS remains obscure. Many underlying potential mechanisms have been postulated including mitochondrial dysfunction causing cellular energy deficit, hyperactivity of hypothalamic-pituitary-adrenal axis in response to stimulation, autonomic abnormalities, and calcium channel abnormalities (2, 6, 11, 12). Many studies supporting the hypothesis of mitochondrial dysfunction include frequent presence of metabolic abnormalities patterns of mitochondrial disorders during the episodes (6, 8), a maternal inheritance pattern (6, 9, 13, 14), and response to mitochondrial targeted therapies such as coenzyme Q10, L-carnitine and riboflavin, as well as antimigraine medications (1, 12). Boles et al. (15) reported that most cases of CVS plus had body fluid metabolites and muscle biopsies consistent with mitochondrial disorders and were strongly associated with maternal inheritance pattern.

Previous studies have shown an association between mitochondrial DNA (mtDNA) polymorphisms and pediatric-onset CVS with some conflicting results (14, 16–19). A3243G mutation was reported in a family with CVS (17). Zaki et al. (19) described 16519T polymorphism in 70% (odds ratio, 6.2) and concomitant polymorphisms of 16519T and 3010A in 29% (odds ratio, 17) of children with CVS. Subsequently, Ye et al. (18) performed mtDNA sequencing in 13 Chinese children with CVS and showed no significant differences in the prevalence of 16519T and 3010A polymorphism in children with CVS and in the control group.

The objectives of this study were to describe clinical characteristics and evaluate association of mtDNA polymorphisms 16519T and 3010A with pediatric-onset CVS and identify novel mtDNA candidate variants associated with this condition.

This observational study was conducted at the pediatric gastroenterology clinic of the Faculty of Medicine Ramathibodi Hospital from April 2019 to April 2020. Inclusion criteria were the patients diagnosed with CVS under the age of 15 years between January 2000 and October 2019. Diagnosis of CVS was according to the Rome III (20, 21) during 2006–2015 and the Rome IV (22, 23) during 2016–2019. Patients diagnosed with CVS before 2006 were retrospectively classified according to the Rome IV. Blood samples were collected from the eligible patients for genetic analysis.

Clinical and genetic studies on pediatric CVS among Asian affected population are limited and underrepresented in the literature. The existing genetic data are also conflicting which may reflect the scarcity of information. To our knowledge this is the largest cohort of Thai pediatric CVS and contribute as one of the milestone studies from Asian affected population.

In this cohort, we demonstrated similar clinical features to those in previous reports including vomiting characteristics and long duration of symptoms before diagnosis (5, 18, 25). However, the mean age of onset was slightly lower as compared to previous pediatric CVS studies (8, 18, 25). Maternal migraine headache was found in 16.7% in this study. Family history of migraine headache in pediatric CVS ranged from 11.9% in Chinese study (18) to 66–72% in western studies (8, 26). During the episodes, abnormal laboratory findings included elevated blood lactate, hyperammonemia, abnormal plasma amino acid profiles and urine organic acids.

Almost half of the patients achieved complete clinical remission at the mean age of 9.6 years, and among this group, the mean clinical course of CVS was 5.9 years. This outcome is consistent with our previous report (27) and other reports (25, 26). Three patients (14%) developed migraine after CVS remission while other studies reported around 25–35% (6, 7). Only one of the three patients underwent a mtDNA genetic test, which showed no 16519T or 3010A polymorphism.

It should be noted that CVS plus is not generally employed in daily clinical practice; however, in the present study of etiopathogenesis of CVS, it could raise concerns relating to differential diagnosis for the affected individuals, in particular pediatric population. Therefore, in this study, we decided to analyze this subgroup and search for potential distinctive etiology, if there was any. We found a higher prevalence of maternal migraine among the CVS plus group compared to non-CVS plus group (50 vs. 11.9%, P = 0.05), suggesting a higher rate of maternal inheritance in patients with CVS plus than in non-CVS plus. Patients with CVS plus had a lower complete remission rate (16.7 vs. 50%, P = 0.19); and a higher frequency of no remission (16.7 vs. 7.5%, P = 0.44). There were no statistical differences of the outcomes probably due to the small number of the patients in CVS plus group. These data suggest that pathogenesis of CVS plus may differ from non-CVS plus and may deserve more extensive investigations for etiologic diagnosis, including exome and mitochondrial genome sequencing for causal-disease relationship.

It is postulated that underlying mtDNA polymorphism might impact energy metabolism both in the resting state and a hyperexcitability state, including stress circumstance (6, 12). The mitochondrial genome encompasses 24 rRNA and tRNA genes and 13 genes encoding proteins involved in oxidative phosphorylation (complex I–V) (28). Some mitochondrial variants are judged as pathogenic or as having a “causality” effect, while others may contribute to the risk of developing multifactorial disorders or are considered to have an “association” effect.

Our study showed no significant association between mtDNA 16519T or 3010A SNPs and pediatric CVS. This result supports the finding in Chinese children with CVS (18) and contrasts the findings in Caucasian reports (16, 19). Explanations for these discrepancies include the small sample sizes, unintentional bias of subject recruitment, and differences in allele frequencies of the SNPs of interest among different ethnicities. More studies in Asian CVS children are required to confirm our findings.

The known pathogenic variants identified in this study (Table 2) have been reported in association with several conditions including Leber hereditary optic neuropathy; deafness; myopathy; cardiomyopathy; and encephalopathy, but not CVS. We have provided the first description of the association of these variants with pediatric-onset CVS. There were two (6.7%) occurrences of mtDNA T12811C variants in the CVS group compared with 4.73% in the control group, suggesting a strong correlation between this variant and CVS compared with other pathogenic SNPs.

Including our patient (Patient 47), seven patients with the mtT8528C allele have been identified (29, 30). All patients had infantile onset HCM, and about half (4/7 patients) had rapid progression leading to infantile death or requiring heart transplantation at a young age. To our knowledge, the present patient was the oldest living individual affected with mtT8528C-associated HCM. This study is the first to show an association between the mtT8528C allele and CVS, platelet dysfunction, abnormal hair, and behavioral disorders. Mitochondrial disorders caused by nuclear gene mutations are more severe than those caused by mtDNA mutations (31), with the exception of some rare variants including mtT8528C (29, 30). Given there has no previous association study of mtT8528C and CVS before, it suggests that mtT8528C and other mt pathogenic variant(s) can also has CVS as one of its related manifestations and that mt pathogenic variant(s) should be considered as possible SNPs associated with pediatric CVS regardless of the presence/absence of classical manifestation of mt disorders.

We found that patients with CVS tended to have at least one pathogenic variant when compared with the general population. Although the allele frequency of each pathogenic variant was not different from those in the control group, we cannot rule out a possible correlation between these variants and CVS/other manifestations. When the data of all SNPs were combined, the aggregate prevalence of pathogenic SNPs in the CVS group was 20% (6/30), which was higher than that in the reference population (9.5%) (P = 0.1). The absence of an association between the likely pathogenic variants and CVS is a straightforward finding because neither the individual nor aggregate prevalence of the variants differed between the two groups. Our data indicate that NGS with a mitochondrial analysis pipeline could be powerful for genomic research in CVS and related disorders.

Of the 11 VUS, one missense variant in complex I might affect mitochondrial function to some extent. Among the eight alleles found in HV segments of the control region, five were in the highly polymorphic HV1. However, part of the HV1, mt16040–16188 containing very low sequence variability, is important in mtDNA replication and possibly translation. A specific segment, mt16157–16172, is an important functional region: the termination-associated sequence (14). Therefore, the mtC16159A and mtCAA16179C variants (Table 4) might interfere with mitochondrial replication and translation. Among the variants in HV3, GCA513G of the D-loop (control region) appears to be a hotspot for somatic and germline mtDNA mutation and may reflect ancient mutation or ancestry (MITOMAP, https://www.mitomap.org/); therefore, it may or may not possess functional effects. We observed that C14668T variant was present in two patients with CVS plus and this variant was residing in the functional locus, MT-ND6 (Table 4; Supplementary Table S4). This may suggest that the variant deserves high priority for association study of mt polymorphisms with pediatric-onset CVS. Further studies are required to elucidate the functional consequences of this allele.

The strength of our study included whole mt genome being analyzed and having ethnic-comparable mt genome reference, which is important factor to avoid bias and increase reliability of the analysis for association study. However, we acknowledge some limitations. First, the clinical data were obtained by medical chart review and were incomplete because of the nature of this retrospective study. Second, although this is one of the few largest cohorts of pediatric CVS with whole mitochondrial genome analysis, the sample size was small. Third, statistical significance does not guarantee clinical relevance. Fourth, functional investigation was not performed to elaborate the pathological effects of the VUS identified. Fifthly, the presence of some individuals with CVS in the control group cannot be totally excluded because the individuals might not be aware of CVS during their childhood and the symptoms could be misinterpreted as infection related. Finally, the studied specimens were obtained from peripheral blood, which may not represent the mitochondrial genotypes in other tissues.

Because pediatric CVS is a rare condition, we urge multicenter national and/or regional networks to perform a larger cohort study with an ethnic-comparable reference population, which may lead to a stronger association study. Additionally, development of a simple and less sophisticated method of mitochondrial functional analysis is needed for research in the fields of CVS and mitochondrial disorders.

In conclusion, the mtDNA polymorphisms 16519T and 3010A were not associated with pediatric-onset CVS in this pediatric cohort. However, 5 known pathogenic variants and 11 novel VUS were shown to be associated with pediatric-onset CVS. Our data suggest that NGS with a mitochondrial analysis pipeline is a powerful tool for mtDNA polymorphism study in CVS research.

The data presented in the study are deposited in the ClinVar Database, accession numbers are shown in the Supplementary Table S6.

The studies involving human participants were reviewed and approved by Ramathibodi Hospital Human Research Ethics Committee (MURA2019/310). Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

KV collected and analyzed the clinical data, performed the bioinformatics analysis of the patient group, and prepared the manuscript draft. BS supervised the bioinformatics analysis. PL supervised and SK performed the bioinformatics analysis of the control population. SN performed the bioinformatics analysis and sequencing. TT-A assisted with the data analysis. ST obtained a research grant. ST and DW designed the overall concept of the study, supervised the data analysis, and critically revised the manuscript. All authors reviewed and approved the final manuscript.

This study was funded by the Faculty of Medicine Ramathibodi Hospital, Mahidol University (RF_62086).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.