Sample Size, Rate of Recruitment, and Attrition
Sample size, rate of recruitment, and attrition varied widely between the 7 studies, ranging from 14 (16) to 90 patients (14). Five studies had a sample with female predominance, whereas 1 study did not provide sex information (16) (Table 1). Only 3 studies used statistical power calculations to assess their required sample size (11,14,15). In order to establish the “ability to recruit” children for a clinical trial, we divided the length of time of recruitment by the number of patients enrolled. Rate of recruitment constitutes an approximation because not all of the studies provided the number of months, with some of them providing only the years of recruitment. We decided to provide the information knowing that it may not be completely accurate because recruitment has been found to be a problem in placebo-controlled clinical trials in children. Four of the 7 studies provided information on length of recruitment. See et al enrolled 25 children in a 6-month period (4 per month) (15). Sadeghian et al recruited 36 children in 14 months (2.6 per month) (13), and Saps et al had the longest recruitment period, enrolling 90 children in 44 months (2 per month) (14). Bahar et al recruited 33 children between 2002 and 2005 (8 per year) (10). One study did not sustain any withdrawals, with all of the patients completing the study (15). The rest of the studies had attrition that varied from 6.7% to 19.4%.
All of the studies provided background rationale that the intervention being investigated is associated with pain reduction (10,12–15). Five of the 7 studies presented a study aim to examine the efficacy or benefit of the intervention compared with that of the placebo. The studies by Collins and Lin and Sadeghian et al stated hypotheses that the intervention would reduce children's symptoms compared with placebo given to children (11,13). Symon and Russell provided a broader aim by evaluating whether their intervention was useful as a “prophylactic drug” (16).
All of the studies used PROs to assess primary and secondary endpoints that varied among studies and included global assessment measures and/or pain reduction. The primary outcome measure in the RCT by Bahar et al was quality of life (QOL), which was self-reported by children (10). A successful outcome was defined as 15% improvement in overall QOL measured by IBS-QOL questionnaire, an adult tool that has not been validated in children (17). The authors adapted the tool by omitting questions on sexual activity. They also assessed multiple secondary outcomes with questionnaires validated in adult patients. Outcome assessments were conducted at 2, 6, 10, and 13 weeks.
Two questions recommended by the Rome II consensus for the design of clinical trials (3) were selected to assess the primary outcome in the multicenter study by Saps et al: a question on satisfactory relief—“Overall how do you feel your problem is?” (“better,” “same,” or “worse”); and a question on satisfaction with treatment—“How did the medication relieve your pain?” (“excellent,” “good,” “fair,” “poor,” or “failed”) (14). The answers to these questions were analyzed in a binary fashion: better and same versus worse and excellent and good versus fair, poor, or failed, and self-reported by children. The study analyzed multiple secondary outcome measures and provided daily diary for the patients to assess daily symptoms including a visual analog–Likert scale ranging from 0 to 100 mm (18).
Collins and Lin assessed frequency and severity of individual gastrointestinal symptoms (bloating, excess gas, incomplete evacuation, AP, diarrhea, constipation, urgency, mucus, straining, fecal incontinence) through a visual analog scale (0–10) that was completed by parents (11). The questionnaire also included 8 multiple-choice questions on the characteristics of their AP (location, frequency, duration) and its effect on daily activities. Participants in the study were also asked to rate their overall symptom improvement as a percentage (100%—complete improvement). Patients underwent a lactulose breath test before and after the trial to assess the effect of the treatment on small intestinal bacterial overgrowth.
The trial conducted by Sadeghian et al did not use validated questionnaires to assess the effect of cyproheptadine on AP (13). Primary outcomes consisted of parents’ and children's reports of frequency and intensity of AP compared with baseline and both child's and parental reports of pain progress. Both children's and parental reports of AP frequency and intensity were measured at 1 and 2 weeks by a 6-point scale (1, completely resolved to 6, worse). For pain progress, children reported using a 4-point scale (1, no pain to 4, worse), whereas parental report was a binary outcome (whether the treatment was satisfactory or not).
The trial by See et al assessed the outcomes of their clinical trial through 2 global primary endpoints that were self-reported by children: quantitative overall AP score and a binary global assessment question (15). The quantitative score, termed AP score, is scored as the sum of 3 subscores measuring pain frequency (pain score— ranging from no pain to multiple times per day), severity (severity score—measured through a validated facial scale with 9 faces, each of them with a corresponding score) (19), and the peptic index score (comprising the presence of nausea, vomiting, chest pain, epigastric pain or tenderness, decreased appetite or weight loss, relation to meals, nocturnal awakening, and pain on waking up in the morning, each of them scored as 1 point). A subanalysis of children with dyspeptic symptoms using the peptic index score (≥4) was presented as a secondary outcome.
Symon and Russell evaluated the effect of pizotifen syrup on pain severity through daily diaries and 2 nonvalidated indices: the index of severity and index of misery (16). Questionnaires were completed by children and parents.
The trial by Kline et al used change of overall symptom improvement and mean pain severity as the primary endpoints (12). Patients completed the 15-item Gastrointestinal Symptom Rating Scale (20) on day 1 and again at the end of the trial. Symptoms included AP, change in stool pattern, heartburn, nausea, and vomiting. Daily dairies were used to report changes in the severity of symptoms. Changes in symptoms were ranked 1 to 5 (1, much better to 5, much worse), and severity of pain was ranked from 1 to 5 (1, excellent to 5, much worse).
Analysis of Outcomes (Intention-to-Treat Analysis vs Per-Protocol Analysis)
Only 2 of the RCTs (11,14) analyzed their outcomes based on intention-to-treat analysis (proportion of patients achieving improvement as a percentage of the total number of patients randomized). The method of analysis was unclear in the study by Bahar et al (possibly per-protocol [PP] analysis) (10), whereas the remaining RCTs (12,13,15,16) were clearly based on PP analysis (only patients who completed the study were included in the assessment).
Risk of Bias
The analysis of risk of bias showed that none of the studies was free of risk of bias, with only 3 studies scoring 1 point each of a possible 6 with a score of 6 of 6 indicating the lowest risk of bias (Table 3). Only 2 authors responded to our request for additional information pertaining to measures taken to avoid biases such as random sequence generation, allocation concealment, and blinding of participants, personnel, and outcome assessors. The corresponding author from the study by See et al (15) could not verify the exact details of the methodology; however, the corresponding author stated that the pharmacy at Mt. Sinai Medical Center generated the randomization sequence via a random number generator.
Two studies (10,14) assessed the efficacy of amitriptyline, a tricyclic antidepressant. The rest of the studies assessed the effect of famotidine (15), rifaximin (11), cyproheptadine (16), peppermint oil (12), and pizotifen.
The study by Bahar et al found improvement in overall QOL score (10) in patients with IBS-associated diarrhea. There was a significant improvement in right lower quadrant pain at 6, 10, and 13 weeks (follow-up), but there was no consistent or significant improvement in pain in other areas. There were no significant differences in the rate of interference with schoolwork, sports, or friends, pain relief after defecation, headache, backache, nausea, dizziness, weakness, constipation, presence of mucous in the stool, tenesmus, or pain exacerbation with defecation between the amitriptyline and placebo groups. The multicenter trial of amitriptyline conducted by Saps et al showed no significant difference in satisfaction with treatment or pain relief in intention-to-treat or PP analysis between drug and placebo (14). Children in both groups experienced a significant decline in pain, interference with daily activities, somatization, and depression scores from baseline, but no difference was found between the groups. The only significant difference between both groups was in anxiety scores that were lower in the amitriptyline group. In the RCT by Collins and Lin, children receiving rifaximin and placebo group had no significant difference in individual symptoms and overall symptom improvement (11). In the study by Sadhegian et al, children receiving cyproheptadine showed a significant benefit in terms of resolution or improvement in AP intensity and frequency at weeks 1 and 2 compared with those receiving placebo (13). There was a significant difference in both self-report and parental report of child improvement between the intervention and placebo groups. In the RCT by See et al, mixed results were reported for the primary endpoints (15). Although a significant benefit in global improvement was found in famotidine compared with that in placebo, the mean AP score was not significantly different between both groups. The RCT on pizotifen by Symon and Russell found that children in the treatment group had significantly fewer days of AP and lower indices of severity and misery compared with those in the placebo group (16). Kline et al found that children in the peppermint oil group had significant reduction in severity of pain compared with those in the placebo group (12).
Mild adverse effects were noted in 4 studies, whereas 2 studies did not report any adverse effect and 1 study did not provide information on adverse effects (13,14,16).
We have conducted a methodological review of RCTs on pharmacological agents in children with AP-FGIDs. The importance of the present review transcends the enumeration of the various features of each study. The understanding of the ability to recruit in each study setting, methodology, and comparative results when different outcome measures are used is key to establish the feasibility and specific recommendations for the design of RCTs in children. Although the argument could be made that pediatric study guidelines could be based on the same principles that guide adult studies, the results of the present review suggest otherwise. The results of pediatric studies on the efficacy of amitriptyline for the treatment of AP-FGIDs are in contrast with the results of adult studies. Although adult studies on tricyclic antidepressants found a positive effect of this group of drugs in the treatment of IBS (21), a systematic review of the efficacy of antidepressants in children and adolescents with AP-FGIDs by the Cochrane group found no evidence to recommend the use of antidepressants in children (22). This contradicts the results of multiple meta-analyses in adults with IBS (23–26) and the results of an evidence-based position statement by the American College of Gastroenterology that tricyclic antidepressants are effective for adult patients with IBS (27). Conflicting results between adult and pediatric studies can be interpreted as a different effect of the drug in children and adults. We, however, cannot exclude the fact that differences in understanding or relevance attributed to global outcomes in adults and children (children may have a different concept of satisfactory relief or satisfaction with treatment than adults) may explain the opposite results. Answering this question is beyond the scope of the present review, but the likelihood of children and adults requiring different outcome measures cannot be ignored. Establishing primary endpoints to support the efficacy of a particular drug should reflect meaningful changes for each age group. Although adult studies have established minimal clinically important differences (MCIDs) (minimal clinical change that the patient identifies as meaningful) for trials on IBS (28), no studies have been published in pediatrics (data are being analyzed by our group) and none of the RCTs in our review has considered MCIDs at the time of selection of outcomes. The absence of data on pediatric MCIDs precludes us from making recommendations on primary efficacy endpoints that reflect what is clinically meaningful to children.
Many of the RCTs in the present review did not show a consistently significant benefit of the drug over placebo in some or all of the outcomes. We cannot exclude the fact that these negative results could be at least partially explained by the use of an inclusion criterion that was exclusively based on AP-FGID phenotypes ignoring the fact that children in the sample could have different pretrial expectations of success, dissimilar psychosocial milieu, and various pathophysiological mechanisms. AP-FGIDs are complex and multifactorial disorders. A better characterization of the various factors involved in the pathogenesis and pathophysiology of different groups of children could help future studies. Establishing a more focused inclusion criterion based on specific psychosocial profiles and pathophysiological mechanisms consistent with the mechanism of the drug to be studied may increase the effect size, making it possible to demonstrate a significant difference between drug and placebo with a smaller sample size. The design of studies with a smaller sample would be of particular importance considering the difficulties in recruitment found in our review. All of the studies had a modest monthly ratio of recruitment, with none of the studies been able to recruit >4 children per month. Once patients were enrolled in the study, however, there was a low attrition rate, which suggests that families who accepted to participate in the trials were not discouraged by participating in a study with a placebo arm despite the poor outcomes achieved by many patients in the studies. We found a variable placebo effect, with some studies showing a nocebo effect (10) and others a high placebo effect (14). This variation in the placebo effect among studies warrants further investigation into the various factors that may influence the placebo effect. Understanding these factors may allow designing studies comparing patients receiving the active drug with patients with a low likelihood of achieving a good response while receiving placebo, a strategy that may also allow increasing the effect size and decreasing the sample size required to find a significant difference between drug and placebo.
Despite the high prevalence of AP-FGIDs, we found that only 7 RCTs with a total of 325 children were ever published. Moreover, studies were generally small, had heterogeneous and sometimes unclear methods, and were prone to bias. Some of the studies were conducted on drugs that are now rarely used, whereas others used old definitions that are no longer used (recurrent AP) or nonvalidated criteria for diagnosis. Shortcomings in design of the various RCT underscore the need for education and the establishment of detailed guidelines for clinical trials in children. We hope that the present review serves to stimulate researchers to conduct further studies while providing information on preventable shortcomings. Future guidelines should include information on study design, minimum length of trials, inclusion/exclusion criteria, primary efficacy endpoints, MCID, validated outcome measures, and reliability and sensitivity of the outcome measures recommended to be used in each AP-FGID. Optimization of design and reporting is paramount, considering the great need for clinical trials and the important difficulties encountered by most of the authors in recruiting patients for the studies. Despite the small number of RCTs found and the shortcomings encountered in many of the studies, we found some encouraging signs that raise our hopes in the progress of pediatric neurogastroenterology in the near future. Most of the studies and the largest RCTs were conducted in the past 6 years, with 234 of 325 participating patients (72%) having been recruited in this time period, compared with those in the previous 15 years (28%).
Our review is not devoid of limitations. We have reviewed only the literature in English and Spanish. We cannot exclude the fact that our review did not report other RCTs that were published in other languages. Our review, however, included studies conducted in distant countries such as Iran that were published in English. We have not reviewed data on RCTs conducted without a placebo arm and open-label trials. Considering the high placebo effect found in some of the studies and the nocebo effect found in others that would make the interpretation of studies without a placebo arm difficult, we made the decision to limit our review to studies comparing drug with placebo. In conclusion, our review found only 7 pharmacological RCTs on AP-FGIDs in children. Most of the studies had methodological limitations and a small sample size. The studies used varied methodology of inclusion, assessment, and length. These data do not allow establishing recommendations on the design of clinical trials in AP-FGIDs in children. Studies on MCIDs, validated outcome measures, and clinical endpoints in children are needed. Standardized validated questionnaire banks such as PROMIS (29) may help clinical practice and research.
1. Saps M, Seshadri R, Sztainberg M, et al. A prospective school-based study of abdominal pain
and other common somatic complaints in children
. J Pediatr
2. Rasquin A, Di Lorenzo C, Forbes D, et al. Childhood functional gastrointestinal disorders: child/adolescent. Gastroenterology
3. Veldhuyzen van Zanten SJ, Talley NJ, Bytzer P, et al. Design of treatment trials for functional gastrointestinal disorders. Gut
1999; 45 (suppl 2):II69–II77.
4. Krause R, Ameen V, Gordon SH, et al. A randomized, double-blind, placebo-controlled study to assess efficacy and safety of 0.5 mg and 1 mg alosetron in women with severe diarrhea-predominant IBS. Am J Gastroenterol
5. Novick J, Miner P, Krause R, et al. A randomized, double-blind, placebo-controlled trial of tegaserod in female patients suffering from irritable bowel syndrome with constipation. Aliment Pharmacol Ther
9. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ
10. Bahar RJ, Collins BS, Steinmetz B, et al. Double-blind placebo-controlled trial of amitriptyline for the treatment of irritable bowel syndrome in adolescents. J Pediatr
11. Collins BS, Lin HC. Double-blind, placebo-controlled antibiotic treatment study of small intestinal bacterial overgrowth in children
with chronic abdominal pain
. J Pediatr Gastroenterol Nutr
12. Kline RM, Kline JJ, Di Palma J, et al. Enteric-coated, pH-dependent peppermint oil capsules for the treatment of irritable bowel syndrome in children
. J Pediatr
13. Sadeghian M, Farahmand F, Fallahi GH, et al. Cyproheptadine for the treatment of functional abdominal pain
in childhood: a double-blinded randomized placebo-controlled trial. Minerva Pediatr
14. Saps M, Youssef N, Miranda A, et al. Multicenter, randomized, placebo-controlled trial of amitriptyline in children
with functional gastrointestinal disorders. Gastroenterology
15. See MC, Birnbaum AH, Schechter CB, et al. Double-blind, placebo-controlled trial of famotidine in children
with abdominal pain
and dyspepsia: global and quantitative assessment. Dig Dis Sci
16. Symon DN, Russell G. Double blind placebo controlled trial of pizotifen syrup in the treatment of abdominal migraine. Arch Dis Child
17. Patrick DL, Drossman DA, Frederick IO, et al. Quality of life in persons with irritable bowel syndrome: development and validation of a new measure. Dig Dis Sci
18. Tesler MD, Savedra MC, Holzemer WL, et al. The word-graphic rating scale as a measure of children
's and adolescents’ pain intensity. Res Nurs Health
19. McGrath PA. The multidimensional assessment and management of recurrent pain syndromes in children
. Behav Res Ther
20. Svedlund J, Sjodin I, Dotevall G. GSRS—a clinical rating scale for gastrointestinal symptoms in patients with irritable bowel syndrome and peptic ulcer disease. Dig Dis Sci
21. Drossman DA, Toner BB, Whitehead WE, et al. Cognitive–behavioral therapy versus education and desipramine versus placebo for moderate to severe functional bowel disorders. Gastroenterology
22. Kaminski A, Kamper A, Thaler K, et al. Antidepressants for the treatment of abdominal pain
-related functional gastrointestinal disorders in children
and adolescents. Cochrane Database Syst Rev
23. Chao GQ, Zhang S. A meta-analysis of the therapeutic effects of amitriptyline for treating irritable bowel syndrome. Intern Med
24. Ruepert L, Quartero AO, de Wit NJ, et al. Bulking agents, antispasmodics and antidepressants for the treatment of irritable bowel syndrome. Cochrane Database Syst Rev
25. Rahimi R, Nikfar S, Rezaie A, et al. Efficacy of tricyclic antidepressants in irritable bowel syndrome: a meta-analysis. World J Gastroenterol
26. Ford AC, Talley NJ, Schoenfeld PS, et al. Efficacy of antidepressants and psychological therapies in irritable bowel syndrome: systematic review and meta-analysis. Gut
27. Brandt LJ, Chey WD, Foxx-Orenstein AE, et al. American College of Gastroenterology Task Force on Irritable Bowel Syndrome. An evidence-based position statement on the management of irritable bowel syndrome. Am J Gastroenterol
2009; 104 (suppl 1):S1–S35.
28. Spiegel B, Camilleri M, Bolus R, et al. Psychometric evaluation of patient-reported outcomes in irritable bowel syndrome randomized controlled trials: a Rome Foundation report. Gastroenterology
29. PROMIS. Dynamic tools to measure health outcomes from the patient perspective. http://www.nihpromis.org
. Accessed October 10, 2014.
Keywords:© 2015 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,
abdominal pain; children; clinical trials; functional gastrointestinal disorder