Irritable bowel syndrome (IBS) is a chronic disorder of gut-brain interaction characterised by abdominal pain or discomfort along with altered bowel habits.¹ Worldwide, approximately 11% of the population suffers from IBS, leading to an estimated annual healthcare cost of US$ 1.5-10 billion.^2,3 According to the Rome IV criteria, IBS is defined as recurrent abdominal pain occurring at least once a week for a minimum of 3 months, accompanied by 2 or more of the following characteristics: (1) pain related to defecation, (2) changes in stool frequency, or (3) changes in stool consistency. To meet the criteria, these symptoms must have been present for at least 3 months, with an onset of at least 6 months prior to diagnosis.⁴ IBS is a chronic condition with an unpredictable pattern, which can significantly impact one’s quality of life.⁵ While there is no definitive treatment or cure for IBS, therapeutic options can address specific symptoms.⁶

The current treatment options recommended for IBS by the American College of Gastroenterology include dietary changes, soluble fiber supplementation, peppermint, chloride channel activators, guanylate cyclase activators, alosetron, mixed opioid agonists/antagonists, gut-directed psychotherapies, antibiotics (rifaximin), antidiarrheal medication (loperamide) and laxatives (linaclotide). Selective serotonin reuptake inhibitors (SSRIs) such as citalopram, fluoxetine, paroxetine, and tricyclic antidepressants (TCAs) like amitriptyline, nortriptyline, and imipramine, are also prescribed.⁷

TCAs have been prescribed off-label to treat IBS-related diarrhea for nearly 4 decades.⁸ Among all TCAs, amitriptyline stands out for its potent anticholinergic and sedative effects.⁹ It is believed that low-dose amitriptyline (25 mg to 50 mg) can alleviate IBS symptoms by reducing subjective pain or stress-related exacerbations.¹⁰ Despite their effectiveness, TCAs are underutilized in treating IBS. A web-based questionnaire sent to 275 primary care physicians revealed that less than 10% of respondents often prescribed TCAs for IBS.¹¹ Moreover, there are no head-to-head trials done to establish safety of individual TCAs for the treatment of IBS.

Several meta-analyses have evaluated the efficacy of TCAs for IBS.^12-15 However, only 1 meta-analysis has focused solely on amitriptyline.¹² This meta-analysis included 4 studies with a total of 130 participants, and found amitriptyline to be beneficial for IBS adult patients (OR, 4.18; 95% CI, 2.00 to 8.77, Z = 3.79, P < 0.001). However, the analysis was limited by the small sample size. Since then, 2 large-scale trials have been conducted to investigate the efficacy of amitriptyline for IBS.^16,17 Both studies had larger sample sizes than previous studies.

The purpose of this systematic review and meta-analysis is to assess the effectiveness of amitriptyline in treating IBS. This review provides an expanded and updated synthesis of all available randomized controlled trials (RCTs) that have investigated the efficacy of amitriptyline in treating IBS in comparison to a placebo. We hypothesize that amitriptyline has superior efficacy and comparable safety profile in managing symptoms of IBS when compared to placebo.

This systematic review and meta-analysis was carried out in accordance with the guidelines given by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) and the Cochrane Handbook for Systematic Reviews of Interventions.^18,19 The study protocol is available at the International Prospective Register of Systematic Reviews (PROSPERO) under the identifier CRD42023479361.

Search Strategy

A comprehensive search strategy was developed with a medical librarian to identify potentially relevant articles. Indexing terms and keywords with truncation were combined using Boolean and proximity operators. No language restriction was imposed on the search. The following databases were searched from inception until November 10, 2023: Medline/PubMed (Ovid), Embase (Ovid), Cochrane Central Register of Controlled Trial, and Web of Science. Covidence software was used to manage, screen, de-duplicate, and extract this systematic review. Complete search strategies for each database are in Supplementary Material.

Eligibility Criteria

Inclusion in this meta-analysis was limited to RCTs comparing the effectiveness of amitriptyline with placebo for IBS treatment and reporting at least 1 outcome of interest. The outcomes of interest were response to treatment, Irritable Bowel Syndrome Symptom Severity Scores (IBS-SSS), decrease in abdominal pain, and improvement in diarrhea. Secondary outcomes assessed were adverse events, including dry mouth, constipation, and tachycardia. Exclusion criteria included studies lacking a control group, had a different comparator or had overlapping patient populations in multiple studies.

The efficacy outcomes were reported and described in the included studies as follows:

IBS-SSS: a 5-item rating system for IBS that assesses the severity and frequency of abdominal pain, bloating or distention, satisfaction with bowel habits, and overall impact of IBS on quality of life. Each factor is rated on a scale from 0 to 100, where 100 indicates increased severity. The final score ranges from 0 to 500.

Response to treatment: a reduction of > 50% of symptoms reported by patients at the end of the study or at least 15% improvement from baseline in overall quality of life score.

Decrease in abdominal pain: this measure was reported either as a decreased number of pain symptoms or a decrease in the mean visual analog scale score for pain.

Improvement in diarrhea: improvement in diarrhea was reported as a decrease in diarrhea frequency or an improvement in at least 1 level on the Bristol Stool Chart with respect to the shape or hardness of the stool.

Data Extraction

Titles and abstracts identified from the search were independently screened and assessed based on eligibility by 2 reviewers. After screening, the same reviewers then assessed the full-text articles for eligibility. From the articles retrieved from this search, additional references were identified by a manual search among the eligible articles. Any disagreements were resolved by consensus. Data from the eligible studies were extracted and catalogued independently by 2 reviewers using electronic spreadsheets specifically designed for this meta-analysis. The data collected included the first author’s name, the year of publication, number of participants, study design, follow-up duration, criteria used for diagnosis of IBS, IBS subtypes, amitriptyline dose, age, and sex of participants. For crossover trials, data were collected for treatment and control periods separately. In cases where data were not available or were insufficient within the published reports, e-mail attempts to contact authors were made.

Statistical and Sensitivity Analysis

Statistical analysis was performed on Review Manager 5.4. Treatment effects for binary outcomes were compared using pooled odds ratio (OR) with 95% confidence intervals (CI). To pool continuous outcomes, weighted mean differences were employed. The Cochran Q test and I² statistics were used to assess heterogeneity; P-values < 0.10 and I² > 25% were considered significant for heterogeneity. The DerSimonian and Laird Random effects model was used for all of the outcomes.²⁰ R software version 4.3.2 was used to perform leave-one-out sensitivity analysis and meta-regression to assess heterogeneity.²¹

Quality Assessment

Study quality was independently assessed by 2 authors using the Revised Cochrane risk of bias (RoB-2) tool for randomized and crossover trials.²² Studies were scored as low risk of bias, high risk of bias or some concerns depending on their biases in selection, performance, detection, attrition, and reporting.²³ Any disagreements were resolved by consensus. Due to the limited number of studies (< 10), it was not possible to evaluate publication bias.²⁴

The systematic literature review yielded 832 studies after removing duplicates. After excluding 794 studies based on title and abstract screening, 38 full-text articles were reviewed for eligibility. Full-text screening identified 5 RCTs^16,17,25-27 and 2 crossover trials.^28,29 All 7 trials were included in the final analysis. No studies were captured from the manual search. See Figure 1 for the PRISMA Flow Diagram.

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flow diagram of the included studies.

Study Characteristics

Table provides a comprehensive overview of the 7 studies. We excluded a study³⁰ from Chao and Zhang¹² previous meta-analysis because that study included participants with functional dyspepsia as well as IBS. Although functional dyspepsia has overlapping symptoms with IBS, it is a distinct condition.³¹ Among patients in the included trials (n = 796), the mean ages ranged from 14.2 years to 49.2 years with 61% (486 of 796) identified as females. Among the 5 trials that specified patients according to IBS subtypes, 56% (414 of 742) of patients had IBS-diarrhea (IBS-D) subtype. Among the 2 trials reporting the IBS-SSS scores, the mean scores ranged from 253.38 to 273.4 in the amitriptyline arm and 264.64 to 272.1 in the placebo arm. Across the trials, the amitriptyline dose varied from 10 mg to 75 mg, and the study durations ranged from 8 weeks to 52 weeks. Three IBS-D studies (Bahar et al,²⁷ Li et al,¹⁷ and Vahedi et al²⁶) reported improvement in diarrhea symptoms, but used different metrics: Bahar et al²⁷ used percentage change, while Vahedi et al²⁶ and Li et al¹⁷ used dichotomous outcomes.

Table. Baseline Characteristics of the Included Studies

Author	Year	Country	Study design	Treatment period (wk)	Criteria for diagnosis	Amitriptyline doses (mg)	Participants	Age (yr)	Sex (M/F)		Mean DOS (mo)			IBS-SSS			IBS-subtype
								I/C			AMI	Placebo		AMI	Placebo		IBS-C	IBS-D	IBS-M	IBS-U
ATLANTIS	2023	England	RCT	52	Rome IV	10-30	463	49.2/47.8	148/315		10a,b			273.4	272.1	77		181	191	14
Li	2019	China	RCT	52	Rome III	25	170	41.3/42.5	88/82		14.13 ± 10.15	15.78 ± 9.7		253.38	264.64	0		139	0	0
Vahedi	2008	Iran	RCT	8	Rome II	10	54	39/33	29/21		43 ± 52	36 ± 37				0		50	0	0
Bahar	2008	USA	RCT	13	Rome II	10/20/30	33	15.3/14.2	9/24							0		33	0	0
Morgan	2005	USA	C-RCT	11	Rome II	25, 50	22	39	0/22							7		11	4
Rajagopalan	1998	Australia	RCT	12	Rome	25, 50, 75	40	35.1/34.5	11/11		59.4 ± 28.7	48 ± 23.2
Steinhart	1982	USA	C-RCT	11	ARS	50	14	41	3/11		5.07a

^aIn years.

^bMedian.

DOS, duration of symptoms; M/F, male/female; IBS-SSS, irritable bowel syndrome symptom severity score; IBS, irritable bowel syndrome; I/C, intervention/control; AMI, amitriptyline; IBS-C, constipation predominant IBS; IBS-D, diarrhea predominant IBS; IBS-M, mixed bowel habits IBS; IBS-U, unclassified IBS; USA, United States of America; RCT, randomized controlled trial; C-RCT, crossover randomized controlled trial; ARS, arbitrary rating scale.

Outcome Measures

Patients who were treated with amitriptyline showed a statistically significant higher response to treatment compared to placebo (OR, 5.30; 95% CI, 2.47 to 11.39; P < 0.001). Upon conducting a sensitivity analysis of treatment response in studies that included only IBS-D patients, the treatment response was still statistically significant for amitriptyline (OR, 7.46; 95% CI, 3.74 to 14.88; P < 0.001). Additionally, we found a significant decrease in IBS-SSS (MD, –50.72; 95% CI, –94.23 to –7.20; P = 0.020) and improvement in diarrhea (OR, 10.55; 95% CI, 2.90 to 38.41; P < 0.001). In patients treated with amitriptyline relative to placebo. A decrease in abdominal pain (OR, 5.23; 95% CI, 0.87 to 31.68; P = 0.070) was found to be marginally significant. See Figure 2 for the forest plots of all outcome measures.

Figure 2. Forests plots for efficacy outcomes. (A) Forest plot of studies comparing amitriptyline with placebo in terms of “Treatment response” outcome. (B) Sensitivity analysis of Treatment response including trials with IBS-diarrhea patients only. (C) Forest plot of studies comparing Irritable Bowel Syndrome Symptom Severity Scores. (D) Forest plot of studies comparing Decrease in Abdominal Pain. (E) Forest plot of studies comparing Improvement in Diarrhea. M-H, Mantel-Haenszel.

Adverse Events

Overall, there were no statistically significant differences in adverse effects between the placebo and amitriptyline groups. The adverse events included tachycardia (OR, 1.04; 95% CI, 0.64 to 1.69; P = 0.870), dry mouth (OR, 2.50; 95% CI, 0.51 to 12.22; P = 0.260), and constipation (OR, 1.52; 95% CI, 0.51 to 4.49; P = 0.450). See Supplementary Figures 1-3 for forest plots of tachycardia, dry mouth, and constipation. Across all studies, adverse events were mainly self-reported and assessed in a binary manner.

The ATLANTIS trial used the Antidepressant Side-Effect Checklist, a validated self-reporting instrument that measures 21 adverse reactions to antidepressants. Participants rated the severity of each item on a 4-point scale (0 = absent; 1 = mild; 2 = moderate; and 3 = severe) and indicated whether a symptom, if present, was likely a side effect of the antidepressant drug (yes or no).³² The trial found that less than 5 percent of participants treated with amitriptyline experience severed symptoms, other than constipation and diarrhea. Additionally, the trial reported a higher attrition rate in the amitriptyline group compared to the placebo group, with rates of 13% and 9% respectively.

Leave-one-out Sensitivity Analysis and Meta-regression

Leave-one-out sensitivity analysis³³ decreased the heterogeneity I² from 66% to 0% by eliminating the ATLANTIS trial for response to treatment (Fig. 3). However, moderate to severe heterogeneity persisted for the rest of the outcomes and adverse effects. Meta-regression showed that at the 3-month follow-up, there was no significant association between follow-up duration and odds ratio. However, at the 6-month follow-up, the odds ratio decreased as the follow-up duration increased (Supplementary Fig. 7 and 8). To validate our findings, we conducted a sensitivity analyses: one excluding studies with a high risk of bias (Supplementary Fig. 9), and another excluding the crossover trials (Supplementary Fig. 10). The results across all these analyses confirmed the reliability of our findings. For further details, refer to Supplementary Figures 4-6 for leave-one-out sensitivity analysis for outcomes and adverse effects, sensitivity analysis and meta-regression plots.

Figure 3. Leave-one-out sensitivity analysis for treatment response.

Quality Assessment

According to our assessment using the RoB-2, 3 trials^16,26,28 showed a low risk of bias. However, 2 trials^25,27 exhibited a high risk of bias due to problems with randomization and missing data. The studies by Bahar et al²⁷ and Rajagopalan et al²⁵ had a high risk of bias due to inadequate reporting of key methodological details. Specifically, Bahar et al²⁷ failed to describe random sequence generation, allocation concealment, blinding, and handling of incomplete outcome data, with 2 dropouts lacking explanation. Similarly, Rajagopalan et al²⁵ lacked clarity on random sequence generation and allocation concealment, with unreported reasons for equal numbers of dropouts in each group. Additionally, 2 other trials^17,29 raised some concerns due to missing outcome data due to loss of patient follow-up. See Supplementary Table for the risk of bias assessment.

This systematic review and meta-analysis found that amitriptyline, compared to placebo, improved treatment response rates, reduced severity of irritable bowel syndrome symptoms, and improved diarrhea while maintaining a similar safety profile. However, for abdominal pain, amitriptyline showed only a marginally significant improvement compared to placebo (Fig. 2). Our findings are consistent with previous meta-analyses by Chao and Zhang,¹² Ford et al,¹³ and Ruepert et al,¹⁵ all of which reported statistically significant improvements in IBS symptoms with TCAs compared to placebo. The exact mechanism by which amitriptyline alleviates IBS symptoms is unclear. It is hypothesized that amitriptyline’s potent anticholinergic properties may reduce visceral hypersensitivity and gastrointestinal motility.^34,35 Our analysis revealed a marginally significant reduction in abdominal pain, consistent with preclinical findings suggesting that TCAs may also decrease pain fibers responses in the pelvic nerve, leading to decreased sensitivity to painful colorectal distension.³⁶

According to our analysis, amitriptyline had a similar safety profile to placebo. This finding contrasts with a previous meta-analysis by Ford et al¹³ 2019 which suggested that TCAs had a higher risk of adverse effects to placebo, such as drowsiness and dry mouth. However, it is worth noting that the previous analysis compared both TCAs and SSRIs to placebo.¹³ Similarly, the most common side effects reported in our analysis were constipation, tachycardia, and dry mouth. In fact, the ATLANTIS trial reported a wide range of additional adverse effects captured by an Antidepressant Side Effect Checklist Questionnaire, but none of the reported adverse events were statistically significant for the amitriptyline arm.¹⁶ In the ATLANTIS trial, 2 Serious Adverse Events occurred in the amitriptyline arm compared to 3 in the placebo arm. Details regarding these events were not reported due to the small number of occurrences and to preserve patient anonymity. The reduced side effects in our analysis, compared to previous studies, may be attributed to the ATLANTIS trial's large sample size and relatively low dose (10-30 mg) of amitriptyline. This dose was further titrated based on side effects and symptom response, which may have reduced the incidence of adverse events. Additionally, the ATLANTIS trial used a comprehensive Antidepressant Side-Effect Checklist, potentially capturing a wider range of effects. We acknowledge that this large trial likely had a substantial influence on our overall results. None of the other studies reported Serious Adverse Events. These results could impact future IBS management, as few general practitioners currently prescribe amitriptyline for IBS.¹¹

The American College of Gastroenterology strongly recommends using TCAs to alleviate global symptoms of IBS.¹ For amitriptyline, the recommendation is to start with an initial 10 mg dose, with gradual dose adjustment based on response and tolerability. Slow up-titration may minimize side effects and improve adherence.³⁷ However, slow titration may delay therapeutic effects. In the ATLANTIS trial, titration started at 10 mg daily, and increased to a maximum of 30 mg over 3 weeks based on treatment response and tolerability. In the Rajagopalan et al²⁵’s trial, amitriptyline titration was fixed. Patients were prescribed 25 mg every night for the first week, followed by 50 mg every night for the second week, and then 75 mg every night until the end of the treatment period.²⁵ The Morgan et al²⁹ trial was also a forced titration study, but prescribed 25 mg of amitriptyline for the first week and 50 mg for the following weeks. The remaining trials did not perform titration regimens. Overall, no study reported the effects of amitriptyline dose titration in IBS. Future research should evaluate optimal dosage ranges and titration schedules. A fixed-dose RCTs would provide valuable insights into the efficacy, safety and tolerability of amitriptyline in the treatment of IBS. Additionally, studies should investigate the impact of dose titration on symptom improvement, time to clinical response, and patient adherence. By conducting such studies, researchers can develop evidence-based recommendations for personalized amitriptyline treatment plans that maximize therapeutic benefits while minimizing adverse effects in patients with IBS.

Amitriptyline is typically administered at low doses in IBS. A study by Halpert et al³⁸ found that increasing the dosage or blood levels of desipramine did not improve IBS symptoms. This suggests that lower doses may be sufficient in managing IBS symptoms with minimal side effects.³⁸ Conversely, Drossman et al³⁹ have suggested that if no side effects are present, doses may be increased to 100-150 mg. All trials analyzed in our study used sub-therapeutic doses compared to those used for treating depression. It is hypothesized that the observed benefits may be attributed to amitriptyline's high anti-cholinergic property, which reduces visceral hypersensitivity and gastrointestinal motility, particularly relevant for IBS-D. Future research should investigate the dose-dependent effects of amitriptyline in IBS, explore physician and patient perceptions of amitriptyline for IBS treatment, as well as design robust phase III trials, building on the methodology of the ATLANTIS trial, with a focus on IBS-D. Despite the American College of Gastroenterology and the American Gastroenterological Association’s recommendations, there remains a need for more clinical trials to determine the efficacy of TCAs in treating IBS.^1,40

The analyzed studies lacked participant homogeneity. For example, Morgan et al²⁹ included only adult females, and Bahar et al²⁷ included both sexes but only adolescents, while all other studies included both sexes but only adults. Sex differences play a significant role in IBS, but limitations in this study prevented a meta-analysis by sex. This underscores the need for future studies to prioritize sex-based analysis to ensure a comprehensive understanding of the disease. Also, only 2 trials ATLANTIS trial and Morgan et al²⁹ included patients other than the IBS-D population. Another example of clinical heterogeneity, in the Li et al¹⁷ study the mean duration of IBS symptoms was 14 months, while in ATLANTIS the mean duration was 120 months. The studies also lacked methodological homogeneity, such as the use of IBS-SSS in the ATLANTIS trial and Li et al,¹⁷ versus self-reporting in other studies, and differing thresholds for clinically significant improvements, like the 35-point minimum in the ATLANTIS trial. Recently, several new assessment tools have been introduced, including the IBS Quality of Life Questionnaire,⁴¹ the Digestive Health Status Instrument,⁴² and the Patient Health Questionnaire-12 Somatic Symptom Scale,⁴³ which may improve clinical homogeneity in future meta-analyses. All trials were similar in certain aspects. It is important to note that they were all double-blind, randomized, and placebo-controlled. Additionally, they consistently excluded patients with severe mental illness and those already taking psychiatric medications. Finally, except for the study by Morgan et al,²⁹ all studies were conducted in an outpatient setting.

The sensitivity analysis reported a decrease in heterogeneity largely attributed to the ATLANTIS study’s large sample size. The meta-regression showed that the ATLANTIS trial, specifically its 6-month follow-up data, is a significant contributor to the meta-analysis’s heterogeneity. Meta-regression results offer a possible explanation for this variability, suggesting that treatment efficacy may decrease over time due to factors like waning treatment adherence, resistance development, or natural disease progression. However, the 3-month follow-up data from the ATLANTIS trial did not show a significant association between follow-up duration and treatment response, potentially due to the shorter time frame. Future studies should include longer follow-up periods, similar to the ATLANTIS trial, and also consider the potential impact of varying outcome assessment methods on heterogeneity.

Limitations

Our meta-analysis has several limitations. Firstly, the inclusion of 2 crossover trials introduces the potential for a unit-of-analysis error.⁴⁴ Secondly, all trials with the exception of Steinhart et al²⁸ relied on the Rome criteria to recruit patients, but these criteria have changed significantly since their initial version in 1992. This evolution has led to inconsistencies in how patients are selected for the studies.⁴⁵ There was variability in the treatment period, assessment tools, outcome measures used, and amitriptyline dosages, which further impacts the heterogeneity of the analysis. For studies that included patients with different IBS subtypes, data on specific outcomes for each subtype were not reported, limiting our ability to perform a subgroup analysis, particularly for constipation predominant IBS patients. Finally, it is also important to interpret the findings cautiously because of the clinical and methodological differences between the studies.

In conclusion, IBS is a common chronic gut-brain disease which significantly impacts quality of life for patients. Our analysis suggests that amitriptyline may improve response to treatment, IBS-SSS score, and diarrhea while having a similar safety profile to placebo. Amitriptyline should be considered as a treatment option in IBS-D patients. Future research should investigate dose-dependent effects of amitriptyline, employ standardized outcome measures, extend treatment periods, and encompass diverse patient populations across various age groups and all IBS subtypes to enhance the generalizability of findings.

None.

None.

Minahil Iqbal: conceptualization, methodology, data collection, formal analysis, writing manuscript, and editing manuscript; Sara Hira: conceptualization, methodology, data collection, formal analysis, and supervision; Humza Saeed: conceptualization, methodology, data collection, and formal analysis; Sufyan Shahid: conceptualization, methodology, data collection, and formal analysis; Suha T Butt, Kamran Rashid, Mohammad Ahmad, Hammad Hussain, and Anzalna Mughal: conceptualization, methodology, data collection, formal analysis; Gabriel P A Costa and Fernanda Gushken: conceptualization, methodology, formal analysis, and editing manuscript; Neil Nero: methodology, data collection, and resources; Shreya Sengupta: conceptualization, methodology, resources, writing manuscript, editing manuscript, visualization, and supervision; and Akhil Anand: conceptualization, methodology, data collection, resources, writing manuscript, editing manuscript, visualization, and supervision.

Note: To access the supplementary material, table, and figures mentioned in this article, visit the online version of Journal of Neurogastroenterology and Motility at http://www.jnmjournal.org/, and at https://doi.org/10.5056/jnm24084.