Cyclic vomiting syndrome (CVS) is a common functional gastrointestinal disorder characterized by recurrent episodes of nausea and vomiting. There is no definite treatment for the condition, although some medications are recommended. We aimed to compare the efficacy of amitriptyline and cyproheptadine in prophylactic therapy of CVS.
This is a single-blinded randomized clinical trial conducted during 2015–2016 in Isfahan, Iran. Sixty-four children who were 3–15 years old, with a diagnosis of CVS (based on Rome III criteria), were included in the study and were randomly divided into two groups of amitriptyline and cyproheptadine. They were followed for 6 months, looking for frequency and duration of attacks as the primary outcome.
The mean monthly frequency of attacks in the last 2 months of the study in the amitriptyline and cyproheptadine group were 0.38±0.55 and 0.59±0.71, respectively (P-value=0.197), after intervention. The mean duration of attacks between amitriptyline and cyproheptadine group were 1.41±2.86 and 1.81±2.22 h, respectively (P-value=0.212). In the amitriptyline group 65.6% of patients reported 100% remission, whereas in the cyproheptadine group 50% reported full remission (P-value=0.206).
There was no superiority of one of the medications over the other. We did not find any age-related effect on the efficacy of these medications as well.
Cyclic vomiting syndrome (CVS) is a functional gastrointestinal disorder with an unknown origin that was first described by Gee in 1882 in English (1, 2). It is now defined in the appendix of International Classification of Headache Disorders, third edition, as one of the episodic syndromes such as abdominal migraine, benign paroxysmal vertigo, and benign paroxysmal torticollis that may be associated with migraine (3, 4, 5). CVS is characterized by recurrent episodes of nausea and vomiting, which may last hours to days with symptom-free intervals, and has a stereotypic pattern of happening in each patient (6). These episodes may be associated with abdominal pain in 80% of cases, headache, or migraine in ~40%, and pallor and fatigue in almost all cases (4). The median age of CVS onset is 4.8 years in children and 35 years in adults, and it has a sex preference by girls (7, 8).
CVS is no longer known as a rare disease and even some studies suggest it as the second cause of recurrent vomiting and nausea in children after gastroesophageal reflux disease (7, 9, 10). It causes significant morbidity in the majority of patients and not only leads to losing many days of school but also causes recurrent hospital admissions (10, 11, 12). Although the etio-pathogenesis of CVS has remained unclear, recently developed theories can better explain the condition and help in treatment and management of the disease, including considering CVS as a migraine equivalent, involvement of hypothalamic–pituitary–adrenal axis, the role of mitochondrial DNA mutations and polymorphisms, and nervous system dysfunction (7, 13, 14, 15, 16, 17). Suggested treatments for CVS are based on these different theories and four different phases of the disease clinical presentation as well (inter episodic phase, prodromal phase, emetic phase, and recovery phase) (6, 7, 8).
Notable prevalence of this disease besides the limitations it brings for patients bold the necessity of an effective treatment. Although some medications have been suggested to be useful in CVS, there is no proven specific treatment for the disease yet and there is a significant lack of clinical trials on the subject (18). In this study, we aimed to compare the efficacy of two common suggested medications for prophylaxis of CVS, amitriptyline, and cyproheptadine, through a randomized clinical trial.
This study is a randomized clinical trial, which was performed from October 2015 to February 2016, in the city of Isfahan, in the central region of Iran. The inclusion criteria were defined as follows: children from 3 to 15 years old, with diagnosis of CVS (based on Rome III criteria), with normal neurological and developmental examination. Rome III criteria consist of the followings: (i) recurrent vomiting episodes, with a duration of lesser than 1 week and a wellness phase between the episodes, (ii) a stereotypical pattern of episodes, and (iii) exclusion of other organic disease that may present clinically same as CVS (1, 6, 19). The exclusion criteria were defined as the presence of known metabolic diseases and obstructive diseases of gastrointestinal or genitourinary system, and lack of consent to enter the study.
To calculate the sample size, we used the equation for comparison of two means with type I error of 5% and power of 85%. Assumptions in two groups (including variance of the mean monthly frequency of attacks and the minimum clinically significant difference between two groups) were pre-specified on the basis of previously published data and the expert opinion. We hypothesized that an equivalent outcome will be observed in groups, regarding the mean monthly frequency of attacks. The sample size was calculated at least 30 patients in each group.
We recruited subjects from Imam Mousa Sadr clinic, which is a referral pediatrics clinic in the city of Isfahan, affiliated to Isfahan University of Medical Sciences, and a private pediatric gastroenterology clinic. It should be noted that Imam Mousa Sadr clinic is the referral pediatrics clinic for Isfahan and surrounding states. Patients who fulfilled the criteria of entering the study were invited to participate in the survey. The study design, goals, and possible complications were completely explained to the parents and informed consents were obtained before the study. The study was designed single blinded so that the patients and their parents were not aware of their medication. The study was approved by the regional bioethics committee of Isfahan University of Medical Sciences and was registered in the Iranian registry of clinical trials (registration number: IRCT 2014030416844N1).
After recruitment of patients, they were randomly divided using randomized block method and each block was allocated to one of the groups of amitriptyline or cyproheptadine using numbered envelopes thereafter. Randomized block design and allocation concealment was implemented by a nurse in each center who were not involved in the study. A complete history of disease description and presentation was got then and a thorough physical examination was performed by a pediatric neurologist or a pediatric gastroenterologist in each center. We recorded patients’ characteristics including age, sex, weight, number of hospitalizations, number of monthly attacks, and duration of each attack before intervention. Patients in amitriptyline group were started on 0.5 mg kg−1 of the drug, increased to 1 mg kg−1 of the drug after 1 week (taking at night), and patients in cyproheptadine group were started on 0.1 mg kg−1 of the drug, increased to 0.2 mg kg−1 of the drug after 1 week.
Patients were visited regularly every 2 weeks after starting the medication by the pediatric neurologist or the pediatric gastroenterologist who visited them at the first time and their status was recorded in their files. The medication for 2 weeks was delivered to the patients on each visit. In each visit the following items were recorded: number of attacks, duration of each attack, and number of hospitalizations in the last 2 weeks. The follow-up duration was 6 months for each patient.
To evaluate the outcome and compare the severity of CVS symptoms after the treatment, we compared the mean monthly frequency and duration of the attacks (the vomiting phase) in four phases: before the intervention, during the first and second months of follow-up (epoch I), during the third and fourth months of follow-up (epoch II), and during the fifth and sixth months of follow-up (epoch III). The primary outcome was defined as comparison of severity of attacks in the last 2 months of drug administration. Furthermore, we investigated the response to the medication in three patterns: complete remission, which was defined as complete resolution of disease presentations (nausea and vomiting) at least in the last 2 months of the follow-up period; 50–99% remission, which was defined as decreased number of attacks or attacks duration for >50% in the last 2 months of the follow-up period; and <50% remission, which was defined as decreased number of attacks or attacks duration for <50% in the last 2 months of the follow-up period Subjects were observed for possible adverse drug reactions during the study.
Descriptive analysis was reported as mean±s.d. for interval variables and frequency (%) for categorical variables. We used Kolmogorov–Smirnov test to evaluated normality of data in each group. To compare interval variables between two groups, independent sample t-test or Mann–Whitney U-test were performed based on the normality of data. To find associations between categorical data in two groups we performed χ2-test using cross tabulation function. To compare severity of attacks before and after intervention for each drug we performed related sample t-test or Wilcoxon-signed-rank test based on the normality of data. We should note that no statistical correction was used for multiple comparisons. Statistical analysis was performed using (IBM SPSS statistics for Windows, version 19, Armonk, NY) and a P-value <0.05 was considered as significant.
A total of 70 patients were recruited initially and 6 of them were excluded due to inclusion and exclusion criteria. Sixty-four subjects remained in the study and none of them left the study or were excluded during the study. The study diagram is presented in Figure 1. A total of 64 patients formed the study population, 32 patients in each group. The mean age of patients in amitriptyline and cyproheptadine group were 8.40±2.06 and 7.56±2.54 years, respectively. Comparison of patients’ and disease characteristics between two groups showed no statistically significant difference before intervention. Baseline data of study groups are summarized in Table 1.
Table 2 shows monthly number and duration of attacks before intervention and during three epochs of the study. As it is presented, no statistically significant difference was seen between two groups regarding mean monthly frequency and duration of attacks before the intervention. We observed lower mean number of attacks during epoch II in the amitriptyline group (0.50±0.51 compared with 0.94±0.72; P-value: 0.011); however, no difference was seen in respect to mean monthly frequency (0.38±0.55 and 0.59±0.71, respectively, P-value: 0.197) and duration (0.38±0.55 and 0.59±0.71, respectively, P-value: 0.197) of attacks between amitriptyline and cyproheptadine group in the last 2 months of the study.
As we stated before, besides comparing mean frequency and duration of attacks, we evaluated outcome in three patterns of complete remission, 50–99% remission, and <50% remission. Regarding the complete remission, 21 patients (65.6%) in amitriptyline group reported complete recovery, whereas 16 patients (50%) in cyproheptadine group reported no more attacks. The mentioned difference was statistically nonsignificant (P-value: 0.311). These results are presented in Table 3.
To better evaluate the response to each drug during the follow up period, we compared the mean monthly frequency and duration of attacks between three epochs of the study. Results from Wilcoxon signed-rank tests are presented in Table 4. As it is shown, cyproheptadine seems to start its effect later than amitriptyline, as in amitriptyline group no statistically significant difference was observed between the second and the third epochs regarding frequency and duration of attacks (P-values: 0.157 and 0.138, respectively).
Regarding the overall efficacy of drugs, in the amitriptyline group the mean monthly number of attacks (1.43±0.61 and 0.38±0.55, before intervention and during epoch III) and duration of attacks (8.85±13.83 and 1.41±2.86, before intervention and during epoch III) were both significantly lower after intervention (P-value<0.001). Same results with similar significance levels were observed in the cyproheptadine group as well (1.71±0.77 and 0.59±0.71, frequency of attacks before intervention and during epoch III; 5.84±2.45 and 1.81±2.22, duration of attacks before intervention and during epoch III; P-value<0.001). Therefore, a favorable outcome was reported in both groups as presented in Table 2. It should be noted that only one of our patients in cyproheptadine group experienced hospitalization during an attack after the intervention and serum therapy with ondansetron was prescribed for him in the hospital.
In the amitriptyline group, three patients (9%) experienced one of these side effects: sedative effects, increased appetite (leading to excessive weight gain), and constipation. On the other hand, in the cyproheptadine group, two patients (6%) experienced one of these side effects: increased appetite and temporary restlessness. None of the patients experienced any serious side effect, which might lead to drug discontinuation.
Cyproheptadine, propranolol, amitriptyline, phenobarbital, and pizotifen are the most commonly used medications for prophylaxis of CVS patients and have shown high response rates with a level II evidence (7, 8, 18, 20, 21). Based on a study by Moses et al. (19), cyproheptadine and amitriptyline were the most commonly chosen medications for prophylaxis of CVS among pediatricians. However, there is no certain treatment for CVS yet and it is not clear which medication is superior to the other due to lack of comparative clinical trials (6, 18, 22, 23).
In this study, we found that amitriptyline and cyproheptadine are both effective in prophylaxis of CVS and none has a superiority to the other one. In 1996, Andersen et al. (20) conducted a retrospective study and compared the efficacy of cyproheptadine and amitriptyline in children with CVS. They reported that 16 of 22 patients (73%) treated with amitriptyline and 4 of 6 patients (66%) treated with cyproheptadine achieved complete remission of episodes of vomiting and abdominal pain with a follow-up ranging from 5 months to 10 years, whereas based on our findings 21 of 32 patients (65.6%) in the amitriptyline group and 16 of 32 patients (50%) in the cyproheptadine group reported no more attacks. They also found that four patients (18%) in the amitriptyline group and one patient (17%) in the cyproheptadine group showed a partial response to the medication, which was defined as ≥50% reduction in frequency of the attacks, whereas our findings showed that eight patients (25%) in each group experienced partial remission. The mentioned differences were not statistically significant neither in the study by Anderson et al. (20) nor in our study, which lead us to conclude that both amitriptyline and cyproheptadine are equally effective for the CVS prophylaxis. Although the median range of follow-up in the study by Anderson et al. (20) was longer, our study supports the finding in a higher level of evidence with larger number of cases.
Other retrospective studies have also reported complete remission rate of 38.3–49% for cyproheptadine and 38.5–58% for amitriptyline (19, 24). Besides these studies, the efficacy of amitriptyline has been compared with other suggested drugs for CVS, including propranolol (22, 25) and CoQ-10 (16). Haghighat et al. (25) showed a better response rate for propranolol, whereas Treepongkaruna et al. (22) reported a superiority of amitriptyline over propranolol. The study comparing the efficacy of amitriptyline and CoQ-10 reported similar level of efficacy for both medications (16). However, none of the patients who were treated with CoQ-10 experienced any side effects, whereas 50% of those treated with amitriptyline reported side effects, which lead to drug discontinuation in 21% of them (16). Our findings are not consistent with the high rate of side effects reported in this study, as we saw side effects in only 9% of our cases who were treated by amitriptyline and none of them lead to drug discontinuation. The difference may be due to the short follow-up period in our study.
As it is clear, there is evidence in the literature that amitriptyline and cyproheptadine are potentially efficacious prophylaxis for children with CVS, but no prospective trial had ever been reported. The effect sizes reported are consistent with those reported in other retrospective studies and differences could be due to differences in trial design and medication dosing. In addition, two other possibilities can be responsible for our higher response rate compared with previous studies: the recall bias that may have affected our data regarding characteristics of attacks before the intervention and parents may have reported higher frequency or duration of attacks than the real baseline; and the placebo effect due to an intensive follow-up (every 2 weeks) after starting the medication.
Another important point that we want to discuss is that North American Society for Pediatric Gastroenterology, Hepatology and Nutrition suggests cyproheptadine as the first choice in children ≤5 years and amitriptyline in children ≥5 years and adults (4, 7, 18, 26). However, we found a favorable similar response in both groups without severe side effects, although the mean age of subjects in amitriptyline and cyproheptadine groups were 8.40±2.06 and 7.56±2.54 years, respectively. This shows that both drugs can effectively reduce CVS symptoms in ages >5 years old. By this finding we suggest that cyproheptadine can be used in children >5 years old as an effective medication in CVS prophylaxis without severe side effects, as well as amitriptyline.
In this study we tried to make our findings generalizable to the target population (children with diagnosis of CVS) through randomization methods and single-blinded design; however, the limited sample size and sampling method might disrupt with this goal. Nevertheless, compared with previous studies and their limitations, our study has fewer methodological problems and our findings are more generalizable. The short period of follow-up and the recall bias due to a retrospective data collection about baseline frequency and duration of the attacks are the main limitations of our study. Another limitation of our study is the lack of control group with placebo; however, we were not able to use placebo in our patients due to ethical issues, as they were suffering from their disease. Multi-centered studies with a longer follow-up period and larger number of cases are suggested to confirm our findings.
Amitriptyline and cyproheptadine are both effective in prophylaxis of CVS and there is no superiority of one of the medications over the other. We did not find any age-related effect on the efficacy of these medications as well.
This study was funded by the vice-chancellor for research and technology of Isfahan University of Medical Sciences (grant number 292210 allocated to Dr Omid Yaghini). Dr Omid Yaghini has full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. We also thank Dr Marjan Mansourian for her help for data analysis and data interpretation.