The study was following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement and its extension for network meta-analysis (Page et al., 2021). OVID Medline, Embase, and the Cochrane Central Register of Controlled were searched from inception to 1 September 2020. Screening randomized controlled trials compared eluxadoline or antispasmodics with placebo or one of antispasmodics in the management of IBS. Clinical registries (Clinicaltrials.gov) were searched for that were completed but unpublished. Studies on IBS were identified with the terms: irritable bowel syndrome and IBS. Other terms included randomized controlled trial, controlled trial, mebeverine, trimebutine, drotaverine, eluxadoline, etc., Search strategies were showed in Supplementary Table S2.

The retrieved articles were screened by two reviewers independently, firstly screening title and level and abstract, secondly reading the full text. The diagnostic criteria were developed from Rome criteria (including Rome I, II, III, and IV). Adult participants with IBS were included. Antispasmodics and eluxadoline were used as monotherapy or as adjunctive treatment in addition to usual care. All interventions lasted at least 1 week. No restrictions on the dosage of drugs and duration of treatment. RCTs for inflammatory bowel disease were excluded unless the IBS results were reported separately. Disagreement in the studies was resolved through group discussion and finally arbitrated by a third reviewer.

The primary outcome was the relief of abdominal pain. The secondary outcome was the relief of global IBS symptoms (adequate relief of global IBS symptoms or FDA-recommended endpoints). Safety indicator was measured by evaluating the treatment-related adverse events. The included eligible RCTs often used different primary endpoints. However, some of the trials adhered to FDA-recommended endpoints and reported treatment efficacy according to a composite of improvement in both abdominal pain and stool consistency, or we were able to obtain these data from the original data.

Two reviewers (Q.F.T and S.L.H) extracted data independently on study characteristics through a standardized data extraction form. First, study characteristics were extracted including author’s name, article publication year, study design, sample size, participants’ age, gender, etc., Secondly, interventions were extracted including duration and dosage of the treatments, type of medicine. Thirdly, Outcome data were extracted—follow-up time points, mean and standard deviation (SD) for continuous data.

The risk of bias used the second version of the Cochrane risk of bias (RoB 2.0) to assessed (Sterne et al., 2019), which was updated in 2008. In RoB 2.0, the bias risk assessment was divided into five parts. Each part requires one or more questions to be answered, which leads to judgments of the risk of bias for a specific study (low, some concerns, or high risk of bias). Compared with the old version of the Cochrane risk-of-bias tool, the RoB 2.0 provides an overall rating of a study which facilitates judgment of the overall quality of the included trials.

We performed a network meta-analysis using a frequentist approach (Rücker, 2012). We calculated the treatment effect of one intervention as compared with the control and its standard error. All the outcomes were dichotomous. And then, we drew a net graph to summarized comparisons between two treatment categories and multiple individual-level treatments respectively. A p-score measures the mean probability of a treatment to be the most effective one, which is important in a network meta-analysis for the purpose of informing clinical practice. P-score can be easily calculated by one-sided p-values (Rücker and Schwarzer, 2015). The risk ratio (RR) and their corresponding 95% confidence intervals (95% CIs) between treatments in each outcome were estimated. A random-effects model was used for the network meta-analysis. The consistency of this study was examined through comparison among direct, indirect, and network estimates. We checked the significance of inconsistency by z test. The global heterogeneity of the network meta-analysis was calculated by global I² statistics and tau-squared value. I² >50% or a tau-squared value > 0.36 was considered as a sign of large heterogeneity. When a network meta-analysis has large heterogeneity, we further performed design-by-treatment analysis to detect the source of heterogeneity (König et al., 2013; Krahn et al., 2013).

The screening process of the review is shown in Figure 1. A total of 496 records were found after the first search. The second screening excluded 432 records, of which 148 were duplicates. 102 records were not RCT design, 65 records were not IBS, 117 records were not targeted interventions or outcomes. Nineteen studies were excluded at full-text level screening for unavailable full-text copies (n = 9) and duplicate outcomes (n = 11). A final total of 45 articles involving 42 RCTs and 8,457 patients were included for analysis (Figure 1). The median age of the subjects was 42 years old, and 62% of participants were female. Table 1 shows the characteristics of the patients in the included RCTs.

The RoB2 showed that 10 (23.8%) RCTs were at low risk of bias in the overall assessment, whereas 31 (73.8%) trials presented some concerns. Only 1 (2.4%) trial had a high risk of bias. The overall assessment of RoB 2.0 is shown in Supplementary Figure S2.

The analysis on efficacy against abdominal pain included 19 RCTs (n = 6,852). The category-level analysis assessed two treatment categories, whereas the individual-level analysis assessed 15 treatments. The categories analysis showed that, compared with placebo, antispasmodics were more effective [RR, 1.46 (95% CI, 1.24 to 1.72); P-score = 0.94; global I² = 73.1%] (Figure 2A) but eluxadoline [RR, 1.18 (95% CI, 0.86 to 1.60)] was not.

The results of the individual-level analysis revealed drotaverine as the most effective [RR, 2.71 (95% CI, 1.70 to 4.32); P-score = 0.99; global I² = 46.1%] (Figure 3A). Drotaverine, pinaverium, ACS, and eluxadoline 100 mg were more efficient than placebo (Figure 3A). Scopolamine, otilonium, pinaverium, ACS, and different doses of eluxadoline (75 and 100 mg) had similar effects on pairwise comparisons (Table 2). However, compared with ACS and pinaverium, eluxadoline 200 mg [RR, 0.53 (95% CI, 0.33 to 0.86)] had a significantly lower success rate in relieving abdominal pain. Furthermore, eluxadoline 100 mg [RR, 1.27 (95% CI, 0.84 to 1.93)] and 75 mg [RR, 1.19 (95% CI, 0.74 to 1.91)] showed slightly higher success rates than scopolamine (Table 2).

We analyzed 24 RCTs (n = 5,399) on their the relief of global IBS symptomss. The category-level analysis assessed two treatment categories, whereas the individual-level analysis assessed 15 treatments. The category-level results showed that antispasmodics were the most effective [RR, 1.43 (95% CI, 1.15 to 1.78); P-score = 0.89; global I² = 86.7%] (Figure 2B). Furthermore, eluxadoline was less effective than antispasmodics (Figure 2B).

The individual-level results showed drotaverine was more effective than placebo [RR, 2.45 (95% CI, 1.42 to 4.22); P-score = 0.95] (Figure 3B). Drotaverine, otilonium, pinaverium, cimetropium, and eluxadoline 100 mg all showed superior the relief of global IBS symptomss over placebo (Figure 3B). Scopolamine, pinaverium, cimetropium, and different doses of eluxadoline (75, 100, and 200 mg) had similar effects on pairwise comparisons. In these comparisons, eluxadoline showed a slightly lower the relief of global IBS symptoms than pinaverium and cimetropium but a slightly higher rate than scopolamine (Table 2).

Because RCTs on eluxadoline generally used the relief of global IBS symptoms as defined by the U.S. Food and Drug Administration, which required simultaneous improvement in the daily scores for the worst abdominal pain and stool consistency on the same day for at least 50% of the days assessed. The sensitivity analysis included 4 RCTs (n = 3,950). The category-level results showed that pinaverium [RR, 1.72 (95% CI, 1.33 to 2.21)] (Supplementary Table S1) was more effective than eluxadoline, whereas the individual-level results showed that pinaverium had the highest sensitivity rate [RR, 2.10 (95% CI, 1.67 to 2.65)]. Pinaverium and eluxadoline 100 and 75 mg were all superior over placebo (Supplementary Table S1).

15 RCTs (n = 6,397) were analyzed. The category-level analysis assessed two treatment categories, whereas the individual-level analysis assessed 13 treatments. The category-level results showed that antispasmodics had lower adverse event rate [RR, 0.99 (95% CI, 0.97 to 1.02); P-score = 0.84; global I² = 34.6%] than eluxadoline [RR, 1.12 (95% CI, 1.01 to 1.25)] (Figure 4A). But they were all higher than placebo.

The individual-level results showed that mebeverine ranked the best [RR, 0.87 (95% CI, 0.72 to 1.05); P-score = 0.86; global I² = 34.6%] (Figure 4B). The adverse event rates of pinaverium, ACS, alverine, otilonium, and hyoscine all did not differ significantly from those of the different doses of eluxadoline (Table 2). However, eluxadoline 100 mg [RR, 1.22 (95% CI, 1.01 to 1.47)] had a significantly higher adverse event rate than hyoscine (Table 2). Of the different doses of eluxadoline, 100 mg had the lowest adverse event rate [RR, 1.12 (95% CI, 0.99 to 1.27); P-score = 0.26; global I² = 34.6%] (Figure 3B), whereas 75 mg had the highest [RR, 1.26 (95% CI, 0.92 to 1.73); P-score = 0.17; global I² = 34.6%] (Figure 3B).

In the study of antispasmodics, the most common adverse events were gastrointestinal symptoms like nausea, constipation, diarrhea, or bloating. These were mild, and no medical care was needed. However, in the study of eluxadoline, in addition to the common gastrointestinal symptoms, adverse events also included Pancreatitis, Spasm of the sphincter of Oddi (Lembo et al., 2016), and headache (Brenner et al., 2019).

Our network meta-analysis showed that 1) eluxadoline 100 mg and some antispasmodics (drotaverine, pinaverium, ACS, otilonium, cimetropium) were effective in relieving abdominal pain and global IBS symptoms. 2) eluxadoline had more adverse event rates than antispasmodics. The primary outcome was the relief of abdominal pain. Because abdominal pain is a common symptom for most patients with IBS—regardless of type. And antispasmodics focused on the relief of abdominal pain. The secondary outcome was the relief of global IBS symptoms, including the severity and frequency of abdominal pain, the severity of abdominal distention, dissatisfaction with bowel habits, and interference with quality of life.

Our study used the adjusted indirect treatment comparison method to answer the clinical question: Is eluxadoline as effective as antispasmodics in relieving abdominal pain? A previous study has shown that the adjusted indirect treatment comparison method minimized the bias caused by variance in the treatment effect size and a multiarm design (Rücker, 2012). We found that the effect of eluxadoline on relieving IBS symptoms was not better than that of antispasmodics. This difference is clinically significant. To the best of our knowledge, the present study is the first network meta-analysis to compare eluxadoline with antispasmodics.

Pharmacological treatments are the primary treatment option because they are convenient and effective. Systematic reviews and guidelines recommend antispasmodics as first-line pharmacologic treatments and have been used in the treatment of IBS for decades (Annaházi et al., 2014). A systematic review published in 2011 (Ruepert et al., 2011), concluded on the efficacy of antispasmodics as a treatment for IBS symptoms, especially abdominal pain. The individual subgroups included cimetropium/dicyclomine, pinaverium, and trimebutine.

Probiotics are also a treatment option in the treatment of IBS. Probiotics are attenuated bacteria, or bacterial products, that are beneficial to the host (Ford et al., 2017). Probiotics include food ingredients, such as fructose-oligosaccharides or inulin that promote the growth or activity of gut bacteria (Ford et al., 2014). There have been many RCTs of probiotics in IBS, the most effective probiotics included Bifidobacterium species and Lactobacillus plantarum. However, because of the various probiotics studied, there are some conflicting results among different trials. That limited the recommendations on which species or strain of probiotics is effective (Ford et al., 2018a). Therefore, probiotics were not included in our study.

Eluxadoline is a novel and expensive drug that was approved for the treatment of IBS-D in adults in Western countries in 2016. In 2019, eluxadoline was recommended for use as second-line therapy for improving IBS symptoms (Ford et al., 2020) at a prescribed dose of 75 mg or 100 mg twice daily (Cangemi and Lacy, 2019). In 2020, Black et al. (Black et al., 2020b) conducted a network meta-analysis to explore the efficacy of licensed pharmacological therapies for IBS-D or IBS-M often used as second-line therapy. Based on the U.S. Food and Drug Administration-recommended endpoint, they found eluxadoline 100 mg twice daily was significantly more effective than placebo in relieving global symptoms of IBS, abdominal pain, and diarrhea, whereas eluxadoline 75 mg twice daily was significantly more effective than placebo for global symptoms, but not more effective than placebo for abdominal pain. The results of the Black et al. study on eluxadoline 100 mg are consistent with our results, but not with that for eluxadoline 75 mg. We found eluxadoline 75 mg was no more effective than placebo in the two abovementioned outcomes.

Although eluxadoline has been proven effective for IBS in some studies (Dove et al., 2013b; Lembo et al., 2016; Cash et al., 2017; Keating, 2017; Özdener and Rivkin, 2017), several questions should be addressed before it could be widely used. First, our results show that commonly used first-line antispasmodics such as pinaverium and drotaverine are significantly better than eluxadoline in relieving abdominal pain in IBS. Moreover, eluxadoline is expensive and thus incurs a high medical cost. Future treatments of IBS may consist of antispasmodics in combination with other specific drugs. For example, pinaverium bromide combined with flupentixol-melitracen can effectively treat diarrhea-type IBS (Qin et al., 2019). Second, no evidence has been obtained on the long-term effects of eluxadoline after 2–3 months of and with continued treatment. This will need to be evaluated in future longitudinal studies. Third, the mechanism of action of eluxadoline is vaguely elucidated and presumed to be related to its unique combined κ- and δ-opioid receptor antagonist characteristics (Brenner et al., 2019). Further investigations into the pharmacological mechanism of eluxadoline are thus warranted. Fourth, our study showed eluxadoline had higher rates of adverse events compared with antispasmodics. Serious adverse events such as pancreatitis and sphincter of Oddi spasms were reported in previous trials (Dove et al., 2013a). Therefore, multiple safety concerns need to be satisfactorily addressed before using eluxadoline. Furthermore, eluxadoline is contraindicated for patients with biliary duct obstruction, severe liver problems, cholecystectomy, alcoholism, pancreatitis, sphincter of Oddi problems, and chronic or severe constipation. From 2015, when eluxadoline was initially approved, until 2017, 120 cases of pancreatitis were reported, some occurring after the initial dose. Of these, 76 resulted in hospitalization, two of which resulted in death (Pimentel, 2018). Thus, more long-term follow-up studies are needed to evaluate the safety of eluxadoline.

This study has several limitations. First, the duration of treatment varied among the RCTs analyzed, and most studies lacked follow-up to assess long-term outcomes. Second, the network meta-analyses of the two primary outcomes and one safety outcome showed slightly greater heterogeneity but consistent results between direct and indirect estimates. The variety of antispasmodics, including dosages and usages, might be a source of heterogeneity and prevented a more comprehensive classification and analysis. Third, in some studies, eluxadoline would be used only after all other treatments (such as loperamide) have been undertaken. This indicates that patients in the eluxadoline groups in the trials had more severe conditions at baseline. In the final results, although the composite response rate of eluxadoline was higher than that of a placebo, no significant difference was found between the interventions. Therefore, some deviations may have occurred in the indirect comparison of antispasmodic agents. Fourth, some trials we included were old, the indicators and outcomes of its evaluation may introduce bias. Finally, eluxadoline has more serious adverse events, the most commonly reported are gastrointestinal symptoms including constipation, vomiting, and nausea. Severe patients can appear pancreatitis (Dove et al., 2013a; Cash et al., 2017).