Long-term Efficacy of Low-dose Tricyclic Antidepressants for Children With Functional Gastrointestinal Disorders

Teitelbaum, Jonathan E.; Arora, Ruchi

Journal of Pediatric Gastroenterology & Nutrition: September 2011 - Volume 53 - Issue 3 - p 260–264
doi: 10.1097/MPG.0b013e318217df7c
Original Articles: Gastroenterology

Objective: The aim of the study was to determine the long-term efficacy of the tricyclic antidepressant (TCA) drugs amitriptyline or imipramine in treating functional gastrointestinal disorders (FGIDs) in pediatric patients.

Materials and Methods: A retrospective chart review of children with a diagnosis of irritable bowel syndrome, functional dyspepsia, and functional abdominal pain diagnosed on the basis of the Rome III criteria was performed. Charts were analyzed for response to the medication and the duration of the response.

Results: A total of 98 patients took a TCA for an FGID. Of the 98 patients, 77 patients (78.6%) responded to the treatment for an average of 10.73 (range 1–45) months.

Conclusions: TCAs are effective in treating FGIDs in pediatric patients for a long duration.

Department of Pediatric Gastroenterology and Nutrition, Children's Hospital at Monmouth Medical Center, Long Branch, NJ.

Address correspondence and reprint requests to Jonathan E. Teitelbaum, MD, FAAP, AGAF, 300 Second Ave, Children's Hospital at Monmouth Medical Center, Long Branch, NJ 07740-6303 (e-mail: JTeitelbaum@SBHCS.com).

Received 27 June, 2010

Accepted 28 February, 2011

The authors report no conflicts of interest.

Article Outline

Functional gastrointestinal disorders (FGIDs) encompass numerous disorders defined by the Rome III criteria (1), including irritable bowel syndrome (IBS), functional abdominal pain (FAP), and functional dyspepsia (FD). These conditions are common among school-age children (2,3) and have been shown to reduce children's quality of life and affect school attendance (4). In children with these disorders, pharmacologic treatment is mostly empirical and based on adult data. Clinical trials have proven the efficacy of low-dose tricyclic antidepressants (TCAs) for treatment of IBS in adult patients (5,6). The Functional Bowel Disorders Working Group Report from the First World Congress of Pediatric Gastroenterology, Hepatology, and Nutrition recommended the evaluation of the role of TCAs in the treatment of pain-predominant FGIDs in children (7).

To date, there have been only 2 randomized clinical trials evaluating the efficacy of the TCAs in the treatment of FGIDs in children (8,9). To our knowledge, however, there are no studies in the pediatric population that evaluate the long-term efficacy of TCAs in the treatment of FGIDs. We designed a large retrospective chart review study evaluating the long-term efficacy of TCA, amitriptyline (AMI), or imipramine (IMI) for symptomatic relief in children with FGIDs.

Back to Top | Article Outline


A retrospective chart review was performed at the pediatric gastroenterology outpatient office of a solopractitioner at the Children's Hospital at Monmouth Medical Center, an academic community-based hospital. Patients seen within the 2008 calendar year with an FGID were identified based on billing codes for any of the following: IBS, dyspepsia, other functional disorders of the intestine. These charts were then reviewed in December 2009 to determine whether a TCA had been prescribed during the course of treatment. Those prescribed a TCA were then divided into 1 of 3 functional categories: IBS, FAP, and FD on the basis of Rome III criteria (1). The charts allowed extraction and analysis of the patient data, which included age, sex, weight, diagnosis, duration of symptoms before first office visit, medication(s) recommended, dosage of medication, endoscopic procedures performed, follow-up dates, and overall response to medication during the period of office follow-up based on global assessment by the interviewing physician.

A patient was classified as having a response to the medication if he or she indicated during follow-up visits that the medication helped to decrease symptomatology to any degree. If the patient claimed to have complete resolution of symptoms for any length of time, this was termed a symptom-free period. If a patient achieved a symptom-free period for at least 2 months, then a trial off medication was considered. The treatment was deemed to have efficacy if the patient experienced a subjective decrease in his or her symptoms (ie, a response) or entered a symptom-free period. Duration of response to treatment was determined as the length of time in months from the initiation of the TCA to the date of the last office visit the patient was noted to be taking the medication.

Patients who started TCAs were followed up in the office regularly at intervals of 1 to 6 months after the start of the medication. Patients were assessed for possible adverse effects based on history. Patients were advised to take the medicine in the evening before bedtime because of the potential adverse effect of sleepiness. Electrocardiograms (ECGs) were not performed before or after initiation of the TCA.

The study was exempt from institutional review board approval at Monmouth Medical Center, and a Health Insurance Portability and Accountability Act waiver form to access patient information was granted.

Back to Top | Article Outline


During the year 2008, the gastroenterology office received 3284 outpatient visits. Of these, 516 patients in 713 visits were diagnosed as having an FGID. Upon review of these 516 patient charts, 146 children were prescribed either AMI or IMI. The patients had been started on the TCA between 2005 and the end of 2008. Patients who were not offered a TCA were not included in the present study. In general, it is the practice of the gastroenterologist not to offer TCA to patients with mild symptoms that do not affect daily activities or to patients with infrequent symptoms, generally less than once per week, in whom daily medications would not seem to be indicated.

Demographic data of these 146 children are provided in Table 1. Of the 146 prescribed a TCA, 33 (23%) did not follow up after their first visit. Fifteen patients (10%) refused to take the medication but continued to follow up in the office. The remaining 98 patients (67%), 57 of whom were girls (58%), initiated therapy with a TCA.

Before the first visit with the pediatric gastroenterologist, patients stated variable duration of symptoms. A specific duration was documented in 50 of the 98 patients who initiated TCA therapy. The number, average number of months of symptoms, and range for each FGID are as follows: IBS (n = 27, 14.7 months [range 2–120 months]), FAP (n = 16, 7.8 months [range 2–24 months]), and FD (n = 7, 16.4 months [range 2–78 months]). Of the 15 who refused TCA, 11 had a documented duration of symptoms before the first gastrointestinal (GI) appointment: IBS (n = 1, 2 months), FAP (n = 7, 6.4 months [range 2–24 months]), and FD (n = 3, 16.7 months [range 8–30 months]). Of the 33 who did not follow up, 17 had a documented duration of symptoms before the first GI appointment: IBS (n = 6, 10.2 months [range 3–24 months]) and FAP (n = 11, 18.2 months [range 2–60 months]).

Among the 98 patients who started a TCA, 49 experienced no improvement with previous trials of other medications. These included acid blockade (H2 antagonist [2], PPI [22]), anticholinergic (8), PEG 3350 (8), loperamide (1), rifaxamin (1), combination of PPI and anticholinergic (6), and anticholinergic and PEG 3350 (1). Ten of the 15 who refused TCA were provided prior medications: PPI (5), anticholinergic (1), PEG 3350 (1), PPI and anticholinergic (3). Ten of the 33 who did not follow up were provided other medications before TCA: PPI (6), anticholinergic (2), PEG 3350 (1), and loperamide (1).

Of the 98 patients who started a TCA, 19 underwent esophagogastroduodenoscopy (EGD) with biopsy. Four of these endoscopies were abnormal. One patient was noted to have celiac disease; however, IBS symptoms persisted despite a gluten-free diet and normalization of her TTG immunoglobulin (Ig) A and endoscopy; she did not respond to TCA. A second patient with IBS was noted to have erosive esophagitis; however, her symptoms persisted despite PPI and healing of her esophagus; she is doing well on a TCA. A third patient with IBS was found to have eosinophic esophagitis and his symptoms persisted despite the healing of his esophagus; he is now doing well on an elimination diet and TCA. The remaining 16 had a normal EGD with biopsy including a pancreatic rest in 1 patient. Of these 19 (7 IBS, 6 FAP, 6 FD), the response rates were similar to the larger group: 13 (68.4%) responded to the medication, 4 (21.1%) did not respond, and 2 stopped because of adverse effects from the medication. One patient who refused TCA had a normal EGD; she is doing well with a combination of a PPI and anticholinergic medication.

Of the 98 patients taking a TCA, 5 patients (5%) reported adverse effects to either AMI or IMI and stopped the medication within 1 week of initiating therapy. Adverse effects included dry mouth (2 patients), constipation (2 patients), and suicidal thoughts (1 patient). These patients were classified as nonresponders. There were no cardiac-related complaints.

The choice of AMI versus IMI was based on the physician's discretion. In general, those patients with the diarrhea form of IBS were provided AMI to take advantage of the anticholinergic effects of the drug to possibly aid in decreasing diarrhea. Those patients with constipation or normal bowel habits were often provided IMI, which has fewer anticholinergic effects. TCA dosing was prescribed initially at 10 mg regardless of weight. Patients who did not respond to the initial dosage were offered dose escalation.

Table 2 reveals the number of patients with each FGID who were assigned either AMI or IMI and their response to the medication trial. Of the 98 who took the TCA, 78.6% responded. Of the 93 patients who remained on TCA, there was an 82.8% response rate (83.3% for AMI, and 82.2% for IMI). Figure 1 reveals the length of follow-up for all of those in the study, as well as for those who responded to TCA and those who refused TCA.

Among those who tolerated the medication, the overall response rate in the IBS group was 82.4% (84.4% AMI, 78.9% IMI), or 76.5% including those who stopped because of adverse effects. The average duration of response was 12.93 months (range 2–45 months). Among the 55 patients with IBS who took TCA, 38 had diarrhea, 14 had constipation, and 3 had alternating bowel patterns. Of those who refused TCA, there were 2 patients with IBS: 1 had constipation, the other had an alternating pattern. Of the 16 patients with IBS who did not follow-up, 11 had diarrhea, 1 constipation, and 1 alternating pattern.

Overall 88.5% of those with FAP responded to a TCA (90% AMI, 87.5% IMI), or 85.2% including those who stopped because of adverse effects. The average duration of response was 8.26 months (range 2–26 months). Overall, 75% of those with FD responded to a TCA (66% AMI, 80% IMI), for an average duration of 8.67 months (range 1–21 months).

Forty-three of the 98 patients were started on the TCA before 2008. Of these, 34 did well on the medication, 10 were able to subsequently discontinue the medication (1 before 2008), 3 discontinued the medication but needed to restart the medication because of recurrence of symptoms, and 21 continued on the medication and were doing well. Of the 55 who started the TCA in 2008, 43 responded to the medication, of whom 10 were able to discontinue the medication, and 2 stopped but restarted because of recurrence of symptoms. The respective numbers of patients with each FGID by year are as follows: 2005 IBS 4; 2006 IBS 5, FAP 2; 2007 IBS 19, FAP 9, FD 3, and 2008 IBS 27, FAP 16, FD 15. The response rates and duration of response to TCA subdivided by year in which they were started are found in Table 3.

Among the patients with FGID who responded well to AMI, 27 (67.5%) responded to the starting dose of 10 mg/day. The remaining 13 patients required dose escalation (11–20 mg/kg, 2–30 mg) to achieve response. The weight range for patients taking AMI who required dose escalation was 28.7 to 150 kg (mean 55.47 kg). Of the 37 that responded well to IMI, 9 (24.3%) patients required dose escalation to 20 mg/kg to achieve response. The weight range for patients on IMI who required dose escalation was 17.6 to 100 kg (mean 47.56 kg). Seven patients did not respond to dose escalation of either AMI or IMI.

There was no patient in this cohort who changed from AMI to IMI or vice versa. Two patients were transiently changed to desipramine given what was thought to be an adverse effect to the anticholinergic properties of the TCA prescribed; however, both of these patients ultimately discontinued TCA therapy because of adverse effects.

After a symptom-free period, patients were encouraged to have a trial off medication. Overall, 26% were able to stop taking the TCA without return of symptoms, whereas 6.5% stopped and restarted the TCA. The remaining 67.5% were unable/unwilling to wean themselves off the medication but continued to have symptom relief. The breakdown of these patients based on TCA type is provided in Table 4. Those who responded to the TCA had symptomatic relief for an average of 10.73 months of follow-up (range 1–45 months). Among those that were able to successfully stop the TCA, the average length of time on medication before discontinuation was 13.36 months (range 2–27 months).

No follow-up data are available for the patients who were prescribed a TCA but did not follow-up. Among the 15 who did not take TCA (10 FAP, 2 IBS, and 3 FD), 11 are doing well, with various therapies: cognitive behavioral therapy (2), PPI (3), anticholinergic (2), PEG 3350 (1), combination of PPI and anticholinergic (1), and no therapy (2).

Back to Top | Article Outline


FGIDs are common among children. Treatment consists of either cognitive-behavioral therapy or medications including TCAs. The present study provides evidence for the long-term efficacy of TCAs in treating 3 functional disorders: IBS, FAP, and FD.

To date, only 2 prior studies have examined the efficacy of TCAs for FGID in the pediatric population (8,9). The first, by Bahar et al (8), was a randomized controlled study of 33 adolescents with IBS, in which those receiving AMI were statistically more likely to have an improved quality of life score. The second by Saps et al (9) was a randomized controlled multicenter study of 90 children with FGID receiving AMI. In this latter study, the response rate was 63%, which is similar to the overall response rate of 78.6% in our study.

Our study was the first to observe the duration of response during a period of many months. The Bahar et al study (8) evaluated the patients during a period of 11 weeks. This included 8 weeks of medication and 3 weeks of washout. The study revealed that the improvement in quality of life achieved at the end of 8 weeks of treatment persisted for another 3 weeks during the washout phase. The Saps et al study (9) assessed the effect of AMI for 4 weeks. Our study population was based on being evaluated for an FGID during 2008. Many of these patients had been prescribed a TCA in prior years, as early as 2005, thus allowing for longer follow-up data. Indeed, of those patients who responded, symptomatic relief was seen for an average of 10.73 months. Furthermore, 26% were able to stop the TCA without return of symptoms, whereas 6.5% stopped and restarted the TCA and 67.5% were unable/unwilling to wean off the medication but continued to have symptom relief.

Our study population experienced few adverse effects, with 5% stopping the TCA. Four of these were directly related to anticholergic effects that are well described (ie, constipation and dry mouth). Within the Saps et al study (9) 3 patients dropped out because of presumed medication-related adverse effects of fatigue, rash, and headache. There were no withdrawals from the Bahar et al study (8) reported. One patient in our study discontinued the medication because of suicidal thoughts and had resolution of these thoughts upon cessation of the IMI. Although suicide has been discussed as a possible adverse outcome among adolescents with major depressive disorder taking antidepressants, the literature has mainly pointed toward selective serotonin reuptake inhibitors as opposed to TCAs (10). Indeed, the causal relation in the literature is weak and does not address the low doses used for FGID, nor does it address the risk among nondepressed children.

There were no clinically relevant cardiac-related adverse effects reported in our study; however, as stated, ECGs were not performed among the patients in this study. The Bahar et al study (8) required ECG for all of the patients in the study, whereas the Saps et al study (9) left ECG testing to the discretion of the physician. Whether low-dose TCAs increase the risk of arrhythmia in this population is still unclear; however, the available data would indicate that the risk, if increased, is likely low.

The possible adverse effects and negative societal view of antidepressants have been blamed by some as an impedance to allowing physicians to provide these medications to patients with FGIDs. Indeed, Bahar et al (8) states that more than half of the eligible patients refused to participate in the study, presumably for this reason. This resulted in a lengthy enrollment period of 4 years to acquire 33 patients. Saps et al, in a multicenter study (9), had to discontinue enrollment at 90 patients after 4 years, 30 patients short of the projected number of patients predicted to have adequate power. Our study was able to identify 146 patients who had been prescribed a TCA for 3 years from a single practice. Of these, 98 (67%) agreed to a trial of TCA. All of the patients in the study were informed before prescribing that this was a low dose of an antidepressant. During the time period of this study, an unknown number of patients were offered a TCA in the office but refused, and a prescription was therefore never provided. Therefore, our study confirms that it is difficult at times to successfully prescribe this class of medication.

To date, dose-finding studies or pharmacokinetic studies using low-dose TCAs in children for FGIDs have not been reported. All of the patients, regardless of weight in our study, were started on 10 mg/day. Of the responders, 28.6% required dose escalation. The Bahar et al and Saps et al studies initiated their dosing based on weight. It is therefore unclear whether weight-adjusted dosing is required, and we would suggest, based on our experience, starting at 10 mg and increasing the dose if needed.

Patients with IBS appeared to have a somewhat longer duration of response when compared to those with FAP or FD. The reason for this was not formally studied, although this is likely because the majority of patients from 2005 to 2006 were patients with IBS (n = 9, 81.8%). These patients were taking the TCA for long periods (average 25.8 months, range 15–45 months), which likely skewed the duration data.

Our study is limited by the lack of a placebo control group. Indeed, the Saps et al (9) study, which had a response rate similar to our study, was unable to demonstrate a statistically significant difference between placebo and control. It is therefore possible that the effect seen in our patient population was related to a placebo effect. The Saps group believed that their placebo effect was enhanced by the fact that many of the centers enrolled were quaternary-care centers and the investigators were nationally recognized experts in the field. Our study was performed at a local community hospital, and the gastroenterologist typically offered the first subspecialty opinion on diagnosis and management of the FGID. The placebo effect, if it existed, was not therefore thought to be heightened by the setting.

This study was also limited by the lack of a validated survey used in assessing response to the TCA. The Bahar et al study (8) used an IBS quality of life instrument questionnaire, whereas the Saps et al study (9) used numerous questionnaires (Pain Response Inventory, Children's Depression Inventory, State-Trait Anxiety Inventory for Children, Children Somatization Inventory questionnaire, and Pediatric Functional Disability Inventory). The response in our study was based on a global assessment and represents real-life practice. Patients reported if “they were doing better.” This may have contributed to the higher response rate because the definition of “response” was not as difficult to achieve as compared with the Bahar et al and Saps et al studies.

The study is also limited by its retrospective design. This contributed to incomplete data collection for questions such as disease duration before initiation of therapy and quantification of patients offered a TCA but refusal before providing a prescription. Future prospective studies also could address issues of adherence to therapy, changes in school attendance, use of additional therapies such as hypnosis or cognitive-behavioral therapy, and presence of comorbid conditions such as depression or anxiety.

Finally, the inclusion of patients who had initiated therapy before 2008 allowed for greater length of follow-up; however, it introduced a potentially biased population of patients who did not experience adverse effects or perceived the medication as efficacious. The relative contribution of such patients is provided in Table 3.

In conclusion, our study represents the largest reported cohort of pediatric patients treated with TCAs for FGID. The results support that a high percentage of patients will benefit from these medications with few significant adverse effects. Furthermore, these medications provide long-lasting efficacy.

Back to Top | Article Outline


The authors thank Barbara Madigan, RN, and Carroll Miller, RN, for their help with chart acquisition and data collection.

Back to Top | Article Outline


1. Rasquin A, Di Lorenzo C, Forbes D, et al. Childhood functional gastrointestinal disorders: child/adolescent. Gastroenterology 2006; 130:1527–1537.
2. 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 2009; 154:322–326.
3. Hyams JS, Burke G, Davis PM, et al. Abdominal pain and irritable bowel syndrome in adolescents: a community-based study. J Pediatr 1996; 129:220–226.
4. Youssef NN, Murphy TG, Langseder AL, et al. Quality of life for children with functional abdominal pain: a comparison study of patients’ and parents’ perceptions. Pediatrics 2006; 117:54–59.
5. Rajagopalan M, Kurian G, John J. Symptom relief with amitriptyline in the irritable bowel syndrome. J Gastroentrol Hepatol 1998; 13:738–741.
6. Morgan V, Pickens D, Gautam S, et al. Amitriptyline reduces rectal pain-related activation of the anterior cingulate cortex in patients with irritable bowel syndrome. Gut 2005; 54:601–607.
7. Hyams J, Colletti R, Faure C, et al. Functional gastrointestinal disorders: working group report of the First World Congress of Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr Gastroenterol Nutr 2002; 35 (Suppl 2):S110–S117.
8. Bahar J, Collins B, Steinmetz B, et al. Double blind placebo–controlled trial of amitriptyline for the treatment of IBS in adolescents. J Pediatr 2008; 152:685–689.
9. Saps M, Youssef N, Miranda A, et al. Multicenter, radomized, placebo-controlled trial of amitriptyline in children with functional gastrointestinal disorders. J Gastroenterol 2009; 137:1261–1269.
10. Kaizar EE, Greenhouse JB, Seltman H, et al. Do antidepressants cause suicidality in children? A Bayesian meta-analysis. Clin Trials 2006; 3:73–90.

amitriptyline; functional abdominal pain; functional dyspepsia; imipramine; irritable bowel syndrome

Copyright 2011 by ESPGHAN and NASPGHAN