Add-on trimebutine-maleate does not improve tolerability or eradication in bismuth quadruple therapy for Helicobacter pylori

Abstract

BACKGROUND

Bismuth quadruple therapy (BQT) remains an important first-line regimen for Helicobacter pylori (H. pylori) eradication in regions with high clarithromycin resistance, but tolerability is limited by a heavy pill burden and a high frequency of gastrointestinal adverse events. Trimebutine maleate, a peripherally acting opioid receptor agonist that modulates gastrointestinal motility and visceral sensitivity, has been hypothesized to attenuate these symptoms.

AIM

To evaluated whether add-on trimebutine reduces adverse events and improves medication adherence and eradication rates in patients receiving 14-day BQT.

METHODS

This open-label, single-center, randomized controlled trial enrolled adults with newly diagnosed H. pylori infection at a tertiary hospital in Korea between July 2020 and October 2022. Participants were randomly assigned to receive 14-day BQT alone or with add-on trimebutine maleate 100 mg three times daily. Co-primary outcomes were the incidence of treatment-emergent adverse events and medication adherence. The secondary outcome was microbiological eradication assessed by ¹³C-Urea breath test 4-6 weeks after treatment completion.

RESULTS

A total of 132 patients were randomized (66 per arm). Baseline characteristics were balanced. Any adverse event occurred in 34/66 control patients (51.5%) and 33/66 trimebutine patients (50.0%) [absolute difference +1.5%, 95% confidence interval (CI): -15.5% to 18.6%; P > 0.999]. Treatment discontinuation due to adverse events was identical [7/66 (10.6%) in each arm]. Mean adherence was 85.0% and 81.6%, respectively (P = 0.764). Eradication rates did not differ: Intention-to-treat 75.8% versus 74.2% (P > 0.999); Per-protocol 87.7% vs 89.1% (P > 0.999). The study had 80% power to detect an absolute reduction of ≥ 24 percentage points in any-adverse-event rates; the 95%CI excluded reductions larger than 15 percentage points.

CONCLUSION

Add-on trimebutine neither reduced adverse events nor improved adherence or eradication in Korean patients receiving 14-day BQT for H. pylori infection. This study did not demonstrate sufficient evidence of benefit to support the routine co-prescription of trimebutine with BQT, and suggests that future efforts may be better directed toward mucosal- and microbiota-directed adjuncts.

Key Words: Helicobacter pylori; Eradication therapy; Bismuth quadruple therapy; Trimebutine; Adverse events

Core Tip: This is the first prospective randomized controlled trial to evaluate add-on trimebutine maleate as an adjuvant to 14-day bismuth quadruple therapy for Helicobacter pylori eradication. Despite trimebutine’s widespread off-label use, add-on treatment did not reduce treatment-emergent adverse events, improve medication adherence, or increase eradication rates, with point estimates tightly clustered around the null. Although the trial was underpowered owing to early termination, these findings do not support routine adjunctive use of trimebutine and redirect attention toward mucosa- and microbiota-directed strategies with proven benefit.

INTRODUCTION

Helicobacter pylori (H. pylori) is a gram-negative, microaerophilic bacterium that colonizes the gastric mucosa and is causally associated with chronic gastritis, peptic ulcer disease, mucosa-associated lymphoid tissue lymphoma and gastric adenocarcinoma. In Korea, where the age-standardized incidence of gastric cancer remains among the highest worldwide, the nationwide seroprevalence of H. pylori infection, although declining over the last two decades, remains above 50%[1]. Successful eradication therefore remains a cornerstone of upper gastrointestinal disease prevention in this population[2].

Clarithromycin-based triple therapy, the default first-line regimen for more than two decades, has shown progressively lower intention-to-treat (ITT) eradication rates in Korea, with contemporary cohorts reporting figures near 60%-70% owing to the rise of clarithromycin resistance, which now exceeds 30% in several series[1-3]. Both the 2020 Korean clinical practice guideline and the Maastricht VI/Florence consensus therefore recommend bismuth quadruple therapy (BQT) or tailored susceptibility-guided regimens as acceptable first-line options in regions with macrolide resistance above 15%[1,2,4]. BQT retains activity against dual-resistant H. pylori, but its clinical usefulness is tempered by a demanding four-drug schedule and by a relatively high incidence of gastrointestinal adverse events, including nausea, dysgeusia, abdominal discomfort and diarrhea, reported in 25%-60% of Korean patients[5-8].

Data from the European Registry on H. pylori Management (Hp-EuReg) have identified adequate compliance as one of the strongest modifiable determinants of eradication success (adjusted odds ratio approximately 6), and treatment-emergent adverse events are a frequent cause of poor adherence[9,10]. Strategies that specifically mitigate the gastrointestinal toxicity of BQT therefore have a clear rationale. Probiotics (particularly Saccharomyces boulardii and multi-strain Lactobacillus preparations), rebamipide and N-acetylcysteine have all been tested as tolerability- or efficacy-enhancing adjuncts, with variable success[11-14]. By contrast, no motility- or visceral-sensitivity-modulating agent has been convincingly evaluated in this setting.

Trimebutine maleate is a peripherally acting agonist at μ-, κ- and δ-opioid receptors that accelerates gastric emptying, induces premature phase III of the migrating motor complex, and modulates visceral afferent sensitivity through direct effects on smooth-muscle L-type calcium ion- and potassium ion-channels[15-19]. It is licensed in Korea, Japan and much of Europe for the symptomatic treatment of irritable bowel syndrome (IBS) and functional dyspepsia[20-23]. A single in-vitro study reported intrinsic antimicrobial activity of trimebutine against Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa and Enterococcus faecalis at minimum inhibitory concentrations (MICs) of 1024-4000 μg/mL, although H. pylori itself was explicitly not tested[24]. Small uncontrolled observations in patients with overlapping functional dyspepsia and H. pylori infection have suggested symptomatic benefit when trimebutine is combined with eradication therapy[25-27]. Given this pharmacological profile, we hypothesized that the addition of trimebutine to a standard 14-day BQT regimen would attenuate eradication-related gastrointestinal symptoms, improve medication adherence and thereby increase ITT eradication rates. We tested this hypothesis in an open-label, randomized controlled trial.

MATERIALS AND METHODS

Study design and ethics

We conducted a single-center, open-label, parallel-group, actively-controlled, randomized clinical trial at the Department of Internal Medicine, Hallym University Chuncheon Sacred Heart Hospital (Chuncheon, Republic of Korea). The study protocol was approved by the Institutional Review Board of Chuncheon Sacred Heart Hospital (No. 2020-03-008) and performed in accordance with the Declaration of Helsinki and the Korean Good Clinical Practice guidelines. The trial was prospectively registered at ClinicalTrials.gov (No. NCT04403087). All participants provided written informed consent before any study-related procedure. Reporting follows the CONSORT 2010 statement[28].

Participants

Eligible participants were adults aged 19 years or older with H. pylori infection confirmed within the preceding 3 months by a positive result on at least one of the following three tests performed during or after upper gastrointestinal endoscopy: Histological examination, rapid urease test, or ¹³C-urea breath test (UBT). Patients with any prior H. pylori eradication therapy, previous gastrectomy (except endoscopic resection), known hypersensitivity to any trial drug, or severe hepatic or renal impairment were excluded. Use of proton pump inhibitors within 2 weeks, histamine H₂-receptor antagonists within 1 week, or systemic antibiotics within 2 weeks of randomization was also an exclusion criterion, as was concomitant use of medications that could influence gastrointestinal motility or interact with the study drugs [non-steroidal anti-inflammatory drugs, anticholinergics, prokinetics, neuromodulators (including tricyclic antidepressants, selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, and gabapentinoids), sucralfate, simvastatin, lovastatin, atorvastatin, cyclosporin, human immunodeficiency virus protease inhibitors, ergot alkaloids and pramlintide]. Patients who were taking any of these agents at screening were not eligible for enrollment unless the drug had been discontinued for at least 1 month (a 1-month washout period) before randomization. Pregnant or lactating women were excluded, and women of child-bearing potential were required to use effective contraception during the study period.

Randomization and interventions

Eligible participants were randomly assigned in a 1:1 ratio to the control or trimebutine group by a study nurse using a computer-generated block randomization list (block size of six). Because of the nature of the intervention and the absence of matching placebo tablets, the trial was conducted in an open-label fashion. All participants received the same 14-day BQT backbone consisting of rabeprazole 10 mg twice daily, bismuth sub-citrate 300 mg four times daily, metronidazole 500 mg three times daily, and tetracycline 500 mg four times daily. Participants in the trimebutine group additionally received trimebutine maleate 100 mg three times daily for 14 days. No placebo was used for the control group.

Outcomes

The pre-specified co-primary outcomes were: (1) The incidence of treatment-emergent adverse events considered possibly, probably or definitely related to the eradication medications according to the Common Terminology Criteria for Adverse Events version 4.0 framework; and (2) Medication adherence, defined as the percentage of prescribed doses actually taken, assessed by pill counting and patient diary at the end-of-treatment visit. Adequate adherence was pre-specified as ≥ 80% of the prescribed doses[29,30]. The secondary outcome was microbiological eradication, assessed by a fasting ¹³C-UBT performed 4-6 weeks after completion of the 14-day regimen. Study outcomes were assessed by the attending physicians (Gong EJ and Bang CS). Because randomization was performed independently by a study nurse, the assessing physicians were blinded to the treatment allocation sequence. The secondary outcome, microbiological eradication, was determined objectively by the ¹³C-UBT and was therefore not subject to observer interpretation.

Patients were reviewed in person on day 14 of treatment (visit 2) and at week 6-8 (visit 3). Adverse events, concomitant medication changes and pill counts were recorded at each visit. Adverse events were additionally captured through patient-initiated telephone reporting between visits. The severity of each event was graded as mild, moderate or severe.

Sample size

The sample size was calculated on the basis of the co-primary outcome of any treatment-emergent adverse event. Assuming an event rate of 30% in the BQT-only arm, consistent with earlier single-centre data on 14-day BQT from our institution[5], and a hypothesized rate of 10% with add-on trimebutine, 326 patients per arm (total 652) were required to provide 80% power at a two-sided α of 0.05, allowing for 10% dropout. Because of slower-than-anticipated recruitment in large part reflecting the reduction in elective endoscopic activity during the corona virus disease 2019 (COVID-19) pandemic, which is estimated to have caused a near-60% suspension of academic and non-COVID clinical trial activity[31] enrollment was closed on 4 October 2022 after 132 participants had been randomized, corresponding to 20% of the planned sample size. Early closure at this point was a pre-specified contingency in case of prolonged recruitment difficulty; the present report therefore describes the complete analysis of the enrolled cohort.

Statistical analysis

All primary analyses followed the ITT principle: Participants lost to follow-up or who discontinued therapy were regarded as adherence and eradication failures. A per-protocol (PP) analysis of eradication was restricted to patients who completed the 14-day regimen and returned for the follow-up UBT. Categorical variables are reported as absolute numbers and percentages and were compared between arms using Fisher’s exact test. Continuous variables are presented as mean ± SD or median (interquartile range), with comparisons by Student t-test or Mann-Whitney U-test after checking normality with the Shapiro-Wilk test. 95% confidence intervals (CIs) for proportions were calculated by the Wilson method; 95%CIs for between-group differences in proportions were based on the Wald (normal-approximation) method. A two-sided P value < 0.05 was considered statistically significant. Because the trial closed before its planned sample size was reached, we additionally computed the minimum absolute reduction in any-adverse-event rate that the actual cohort of 132 patients had 80% power to detect at two-sided α = 0.05, using the observed control-arm event rate as the reference; interpretation of the null findings relied on the width of the observed CIs rather than post-hoc power calculations of observed effect sizes, which are statistically uninformative[32,33]. Analyses were conducted in Python 3.11 using SciPy (v1.11) and statsmodels (v0.14).

RESULTS

Participants

Between 4 July 2020 and 4 October 2022, 132 patients were enrolled and randomized: 66 to the control arm and 66 to the trimebutine arm (Figure 1). In the control arm, 59 patients (89.4%) completed the 14-day regimen; 7 (10.6%) discontinued therapy prematurely because of adverse events. In the trimebutine arm, 59 patients (89.4%) completed the 14-day regimen and 7 (10.6%) discontinued. Follow-up UBT could not be performed in 9 control patients (7 lost to follow-up, 1 refused, 1 discontinued therapy) and in 11 trimebutine patients (10 lost to follow-up, 1 discontinued therapy). Accordingly, the PP analysis for eradication included 57 control and 55 trimebutine participants. Baseline characteristics of participants are described in Table 1.

Figure 1 CONSORT flow diagram of participants. Between 4 July 2020 and 4 October 2022, 132 participants were randomized 1:1 to 14-day bismuth quadruple therapy alone or with add-on trimebutine maleate 100 mg three times daily. Recruitment was terminated early at 132 of the planned 652 participants (20%) owing to corona virus disease-19-related disruption of elective endoscopic activity. ITT: Intention-to-treat; PP: Per-protocol; BQT: Bismuth quadruple therapy; UBT: Urea breath test; TID: Three time daily.

Table 1 Baseline demographic and clinical characteristics, mean ± SD/n (%).

Variable	Control (n = 66)	Trimebutine (n = 66)	P value
Age, years	47.9 ± 11.1	47.8 ± 11.2	0.932
Male sex	40 (60.6)	34 (51.5)	0.381
Body mass index, kg/m2	24.1 ± 3.0	23.8 ± 3.3	0.507
Current smoker	13 (19.7)	16 (24.2)	0.679
Habitual alcohol use	37 (56.1)	38 (57.6)	0.859
Hypertension	9 (13.6)	7 (10.6)	0.791
Diabetes mellitus	3 (4.5)	5 (7.6)	0.718
Dyslipidemia	4 (6.1)	4 (6.1)	> 0.999
Family history of gastric cancer	14 (21.2)	7 (10.6)	0.152

Comparisons used Student t-test for continuous variables and Fisher’s exact test for categorical variables.

Adverse events

At least one treatment-emergent adverse event was reported by 34 of 66 patients (51.5%; 95%CI: 39.7%-63.2%) in the control arm and 33 of 66 patients (50.0%; 95%CI: 38.3%-61.7%) in the trimebutine arm (absolute difference +1.5%, 95%CI: -15.5% to 18.6%;P > 0.999; Table 2). Nausea was the most common event in both arms, occurring in 20 control (30.3%) and 22 trimebutine (33.3%) participants (P = 0.852). The incidence of diarrhea or loose stools was numerically higher in the trimebutine arm (19.7% vs 10.6%, P = 0.224). No significant between-group difference was observed for any individual adverse event. Seven patients in each arm (10.6%) discontinued treatment because of an adverse event (P > 0.999). No serious adverse event, anaphylaxis, hospitalization or death occurred during the study.

Table 2 Treatment-emergent adverse events during 14-day bismuth quadruple therapy, n (%).

Event	Control (n = 66)	Trimebutine (n = 66)	P value
Any adverse event	34 (51.5)	33 (50.0)	> 0.999
Nausea	20 (30.3)	22 (33.3)	0.852
Diarrhea/loose stool	7 (10.6)	13 (19.7)	0.224
Dizziness	5 (7.6)	1 (1.5)	0.208
Dyspepsia/bloating	5 (7.6)	3 (4.5)	0.718
Dysgeusia (bitter/metallic)	3 (4.5)	4 (6.1)	> 0.999
Headache	3 (4.5)	3 (4.5)	> 0.999
Epigastric discomfort/heartburn	3 (4.5)	4 (6.1)	> 0.999
Asthenia/fatigue	4 (6.1)	3 (4.5)	> 0.999
Vomiting	3 (4.5)	1 (1.5)	0.619
Dark stool (bismuth-related)	3 (4.5)	6 (9.1)	0.492
Abdominal pain	1 (1.5)	2 (3.0)	> 0.999
Anorexia	0 (0.0)	1 (1.5)	> 0.999
Treatment discontinuation due to AE	7 (10.6)	7 (10.6)	> 0.999
Serious adverse event	0 (0.0)	0 (0.0)

Patients could report more than one type of adverse event; percentages therefore do not sum to 100. P values from Fisher’s exact test. AE: Adverse event.

Medication adherence and eradication

Medication adherence did not differ between arms. Mean adherence was 85.0% ± 33.0% in the control arm and 81.6% ± 36.8% in the trimebutine arm (P = 0.764), and 80.3% vs 78.8% of participants achieved ≥ 80% adherence (P > 0.999; Table 3). Eradication outcomes were likewise similar. In the ITT analysis, the eradication rate was 75.8% (50/66; 95%CI: 64.2%-84.5%) in the control arm and 74.2% (49/66; 95%CI: 62.6%-83.3%) in the trimebutine arm (absolute difference +1.5%, 95%CI: -13.3% to 16.3%; P > 0.999). The corresponding PP rates were 87.7% (50/57; 95%CI: 76.8%-93.9%) and 89.1% (49/55; 95%CI: 78.2%-94.9%; absolute difference -1.4%, 95%CI: -13.2% to 10.5%; P > 0.999).

Table 3 Medication adherence and eradication outcomes, mean ± SD/n (%).

Outcome	Control (n = 66)	Trimebutine (n = 66)	P value
Mean adherence, %	85.0 ± 33.0	81.6 ± 36.8	0.764
Adherence ≥ 80%	53 (80.3)	52 (78.8)	> 0.999
Completed 14-day regimen	59 (89.4)	59 (89.4)	> 0.999
Lost to follow-up for UBT	9 (13.6)	11 (16.7)	0.809
ITT eradication n/N (%), (95%CI)	50/66 (75.8) (64.2-84.5)	49/66 (74.2) (62.6-83.3)	> 0.999
PP eradication, n/N (%), (95%CI)	50/57 (87.7) (76.8-93.9)	49/55 (89.1) (78.2-94.9)	> 0.999

95% confidence intervals for single proportions calculated by the Wilson method. Continuous variables compared by Mann-Whitney U-test; categorical variables by Fisher’s exact test. CI: Confidence interval; ITT: Intention-to-treat; PP: Per-protocol; UBT: Urea breath test.

Pre-specified minimum detectable effect and external benchmarking

With the observed control-arm event rate of 51.5%, the as-enrolled cohort of 132 patients had 80% power (two-sided α = 0.05) to detect an absolute reduction of approximately 24 percentage points in any-adverse-event rates i.e., a drop from 51.5% to ≤ 27.9% (relative reduction of approximately 46%). The observed 95%CI for the between-group difference in any-adverse-event rate (-15.5% to 18.6%) excludes absolute reductions of more than 15 percentage points, consistent with either no effect or, at most, a small effect in either direction. The observed any-adverse-event rate in the control arm (51.5%) falls within the range reported by published 10 day-14-day BQT trials and real-world cohorts (29%-67%; Table 4), indicating that the Korean BQT adverse-event burden captured in the present study was neither unusually high nor unusually low and therefore constituted a representative population in which to test an add-on tolerability strategy[6-8,34-39].

Table 4 Adverse-event and discontinuation rates with bismuth quadruple therapy in comparable randomized trials and registry cohorts.

Ref.	Setting/population	BQT-treated (n)	Any adverse event (%)	Discontinuation (%)
Present trial (control arm)	Korea, RCT, 14-day classical BQT	66	51.5	10.6
Nyssen et al[10]	Hp-EuReg, 28 countries	Approximately 22000 (any regimen; BQT subset)	Approximately 37 (BQT)	1.3 (whole cohort)1
Olmedo et al[34]	Hp-EuReg, 2013-2021, all BQT	15582	Approximately 30-40	10
Nyssen et al[35]	Hp-EuReg, single-capsule BQT (Pylera), 10-day	Approximately 2100	Approximately 29	1.7
Malfertheiner et al[6]	Europe, RCT, Pylera 10-day	218	Approximately 50	< 2
Liou et al[7]	Taiwan, RCT, 10-day BQT	540	67	10
Kim et al[36]	Korea, RCT, 10-day BQT	175	67.5	23 (dose reduction or discontinuation)
Gisbert et al[39]	Meta-review of Pylera studies	6482 (30 studies)	Approximately 29 (Pylera)/approximately 37 (classical)	Approximately 2

¹The 13% discontinuation rate in Nyssen et al[10] is the overall European Registry on Helicobacter pylori Management cohort value across all regimens; the bismuth quadruple therapy-specific discontinuation rate was not separately reported.

Superscripted numbers after author/year correspond to the reference list. Numbers are shown as reported in the original studies; registry estimates are rounded. BQT: Bismuth quadruple therapy; Hp-EuReg: European Registry on Helicobacter pylori Management; RCT: Randomized controlled trial.

DISCUSSION

In this open-label, randomized, active-controlled trial of 132 Korean adults receiving 14-day BQT for H. pylori eradication, the addition of trimebutine maleate 100 mg three times daily did not reduce the incidence of treatment-emergent adverse events, did not improve medication adherence, and did not increase ITT or PP eradication rates. Point estimates for every pre-specified outcome were nearly identical between arms: The absolute difference in any-adverse-event rate was +1.5 percentage points (95%CI: -15.5 to 18.6), the discontinuation rate was identical at 10.6%, and ITT and PP eradication differences were +1.5 and -1.4 percentage points, respectively. To our knowledge, this is the first reported randomized trial to formally test trimebutine as an adjunct to H. pylori eradication therapy; it therefore defines the first high-quality evidence base for an intervention that has been used informally in clinical practice without supporting randomized data.

The original rationale for testing trimebutine rested on two pharmacological premises: A motility/visceral-sensitivity effect and a putative antimicrobial effect. Our null findings are most coherently explained by a mechanistic mismatch between each of these premises and the actual pathophysiology of BQT-related adverse events. We discuss each in turn.

First, the motility hypothesis. Trimebutine is a weak peripheral opioid agonist that binds μ, δ and κ receptors with preferential μ affinity (selectivity ratio approximately 100:12:14 for μ:δ:κ) but is approximately 30-fold less potent than morphine. It accelerates gastric emptying, induces premature phase III of the migrating motor complex (shortening the migrating motor complex cycle from approximately 86 to approximately 33 minutes in healthy humans after 100 mg intravenously), and exerts local-anaesthetic-like sodion-channel blockade through its active metabolite N-desmethyl-trimebutine[15-18,40]. In IBS it normalizes, rather than merely suppresses, exaggerated post-prandial colonic motor activity[21,41]. Its site of action is therefore the enteric smooth muscle and extrinsic/visceral afferent pathways. In contrast, the dominant adverse events of BQT metallic taste, nausea, dizziness and encephalopathy-spectrum symptoms, caustic mucosal effects, and diarrhea arise from: (1) Direct mucosal toxicity of tetracycline (acid potential of hydrogen 2-3 in saliva when contact-dissolved, causing pill-induced esophagitis and mucosal burn)[42,43] and metronidazole (peripheral/optic neurotoxicity and central nervous system effects via penetrable nitroimidazole radicals)[44,45]; (2) Central pharmacology (chemoreceptor-trigger-zone stimulation by metronidazole; cerebral accumulation of organobismuth)[44,46]; and (3) Antibiotic-induced dysbiosis, which follows even short courses of macrolide–nitroimidazole therapy and in landmark longitudinal studies persists with loss of short-chain-fatty-acid producing Bifidobacterium and Lachnospiraceae and with a relative expansion of Proteobacteria for up to four years after a single seven-day eradication regimen[47-50]. None of these pathways is accessed by a peripheral opioid motility modulator. A drug whose locus of action is smooth-muscle contractility and visceral sensitivity is therefore pharmacologically unsuited to counteract adverse events generated by mucosal chemistry, central neuropharmacology and perturbed microbial ecology.

Second, the antimicrobial and broader clinical-rationale hypothesis. Kountouras et al[25] have, over the last two decades, proposed that co-administration of trimebutine with H. pylori eradication regimens could improve symptomatic outcomes and, through an intrinsic antimicrobial effect on enteric organisms, possibly augment eradication itself a proposal articulated in uncontrolled observations of overlapping functional gastrointestinal disorders and subsequently formalized as a mechanistic hypothesis linking trimebutine, the gut-brain axis and H. pylori related neuro-inflammatory disease[25-27]. The antimicrobial component of that hypothesis rests on a single preliminary in-vitro study by the same group, which reported MICs of 1024-4000 μg/mL against Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa and Enterococcus faecalis but that explicitly did not test H. pylori[24]. After oral administration of trimebutine 100-200 mg in humans, parent drug plasma levels are largely undetectable owing to rapid N-demethylation, whereas circulating N-desmethyl-trimebutine reaches only approximately 40-120 ng/mL (approximately 0.1-0.3 μmol/L)[51]; achievable gastric or mucosal concentrations are thus several orders of magnitude below the reported MICs against unrelated aerobic organisms and cannot be extrapolated to the microaerophilic, gastric-niche pathogen H. pylori. The present randomized trial, by contrast, provides the first direct in-vivo test of whether the combined clinical-antimicrobial rationale translates into measurable benefit on symptoms, adherence or eradication, and clearly shows that it does not: ITT and PP eradication rates were within 1.5 percentage points of each other across arms, and no prespecified symptom endpoint favoured the trimebutine group.

Our findings are also internally coherent with what trimebutine has, and has not, been shown to do in its established therapeutic settings. In IBS, the pooled effect of antispasmodics as a class corresponds to a number needed to treat of 4-5 for global symptom improvement[52,53] and trimebutine is specifically endorsed in the 2011 Cochrane antispasmodic review and the American College of Gastroenterology IBS monograph for global assessment of symptoms and abdominal pain[53,54]. In functional dyspepsia, a recent multicentre randomized trial found that trimebutine 300 mg twice daily improved dyspepsia severity scores over four weeks[22]. None of these indications, however, reflects a mucosal-toxic or dysbiotic process. The only randomized controlled trial of trimebutine in a closely related context prevention of post-haemorrhoidectomy pain also returned a null primary outcome despite documented pharmacodynamic activity on anal resting pressure[55], illustrating that measurable motor effects of trimebutine do not always translate into improvements in symptom outcomes when the underlying pathology is not motility-driven.

The null result of the present trial is particularly informative when benchmarked against adjuncts that have succeeded in reducing BQT-related adverse events. In Hp-EuReg data, overall adverse-event rates for BQT cluster around 29%-37%, with discontinuation rates of 1%-3%[7,10,34,35]; our 51.5% rate in the control arm therefore falls within the upper half of the 29%-67% range reported in published BQT randomized controlled trials and registries (Table 4), plausibly reflecting active symptom elicitation at each study visit in a Korean academic centre rather than passive spontaneous reporting typical of registries. Against this background, Saccharomyces boulardii supplementation has consistently reduced overall adverse events (relative risk = 0.44, 95%CI: 0.31-0.64) and diarrhea (relative risk = 0.51, 95%CI: 0.42-0.62) in the 11-randomized controlled trial, 2200-patient meta-analysis of Szajewska et al[11] and Szajewska et al[12], corroborating an earlier five-trial analysis by the same group, multi-strain probiotic preparations containing Lactobacillus and Bifidobacterium species have reduced adverse events and improved eradication in strain-specific meta-analyses[13,56], and rebamipide supplementation has produced modest but reproducible gains in eradication and tolerability (odds ratio = 1.75, 95%CI: 1.31-2.33)[14]. Each of these adjuncts acts on precisely the mucosal or microbial pathways that generate BQT-related symptoms; our inability to produce even a trend toward benefit with a peripheral opioid agent reinforces, by contrast, that the correct mechanistic target for BQT-symptom mitigation is the mucosa-microbiota axis rather than the motility axis.

Adherence and eradication likewise behaved in line with this mechanistic account. Hp-EuReg analyses of 38698 patients have shown that compliance is the single strongest modifiable predictor of eradication success (adjusted odds ratio = 6.3, 95%CI: 5.2-7.7), with cure rates falling by 20-40 percentage points among non-compliant patients across all major first-line regimens[9]. Because trimebutine did not alter the symptom burden, an intervention ultimately adding three tablets per day and theoretically aggravating pill fatigue had no scope to improve adherence; indeed, the proportion of patients reaching the ≥ 80% adherence threshold was fractionally lower with trimebutine (78.8% vs 80.3%, P > 0.999). Correspondingly, ITT eradication rates (75.8% and 74.2%) and PP rates (87.7% and 89.1%) were essentially identical, and are concordant with the 92.1% first-line BQT modified ITT (mITT) eradication rate reported in the contemporary Korean Registry on H. pylori Management (overall first-line mITT 81.0%) as well as with meta-analyses of 14-day BQT in high clarithromycin-resistance regions[57-59].

Interpretation of a negative trial must address the possibility that the result is an artefact of insufficient power rather than a genuine absence of effect. Our trial closed after enrolling 132 of a planned 652 participants because of the COVID-19 pandemic, which is estimated to have caused a near-60% suspension of academic and non-COVID clinical trial activity[31]. However, two methodological considerations argue that the null result is informative rather than uninformative. First, Hoenig and Heisey[32] have repeatedly emphasized that post-hoc power calculations using observed effect sizes are logically flawed; the appropriate locus of interpretation is the width of the observed CI. In our data, the 95%CI for the between-group difference in any-adverse-event rate is -15.5 to 18.6 percentage points, excluding any absolute reduction greater than 15 points. Second, we computed that the as-enrolled cohort had 80% power to detect an absolute reduction of ≥ 24 percentage points i.e., approximately the original protocol-mandated effect of 20 percentage points would have been detectable had it existed. The observed effect was not merely statistically non-significant; the point estimate was essentially zero (1.5 percentage points), a pattern characteristic of true absence of effect rather than of underpowered measurement of a real but smaller benefit. The principle that “absence of evidence is not evidence of absence” remains salient[33], but in this case the evidence points clearly toward absence of a clinically meaningful effect.

Our study has several additional limitations. First, and most importantly, the trial enrolled only approximately 20% of the calculated target sample size (132 of a planned 652 participants). This substantial shortfall means that the assumptions underlying the original sample-size calculation could not be tested as planned and that the study was considerably underpowered relative to its pre-specified design; the null findings must therefore be interpreted with this important limitation in mind. As discussed above, we have sought to mitigate this concern by basing our inference on the width of the observed CIs rather than on post-hoc power calculations, but the reduced statistical power remains a genuine constraint on the strength of the conclusions that can be drawn. Second, the open-label design, driven by the practical unavailability of matching placebo tablets, may have introduced ascertainment bias for subjective adverse events. Because active elicitation was symmetric across arms and because the effect size was essentially zero, any such bias is unlikely to have masked a real benefit; it could, however, have inflated the overall adverse-event rate relative to blinded and registry comparators. Third, the trial was single-centre and conducted in a Korean tertiary-hospital population, and results may not generalize to primary-care settings or to populations with different antibiotic-resistance profiles. Fourth, antibiotic susceptibility testing was not systematically performed; eradication failures could therefore not be formally attributed to resistance. Fifth, we used pill counts and patient diaries rather than electronic medication monitoring for adherence; these methods systematically overestimate adherence but are identical across arms and therefore unlikely to bias the between-arm comparison. Sixth, we applied the classic ≥ 80% adherence threshold[29,30]; the contemporary Hp-EuReg convention uses ≥ 90%[9], which would reclassify a modest proportion of patients in both arms but would not change the between-group comparison. Seventh, we did not administer validated patient-reported outcome instruments that might have detected sub-threshold effects of trimebutine on symptom severity rather than on dichotomous event occurrence. Eighth, we did not examine the faecal microbiome or gastric mucosa, and can therefore not directly test the dysbiosis mechanism invoked in our interpretation. Ninth, follow-up urea breath testing could not be completed in 20 of 132 randomized participants (9 control and 11 trimebutine patients; approximately 15%), predominantly because of loss to follow-up, raising the possibility of attrition bias. Two considerations limit the likely impact of this attrition on our conclusions. First, all primary analyses followed the ITT principle, in which participants who were lost to follow-up or discontinued therapy were conservatively counted as adherence and eradication failures; this approach avoids the optimistic bias that would arise from analyzing completers only. Second, the extent of attrition was similar between arms (13.6% vs 16.7%; P = 0.809), so differential drop-out is unlikely to have systematically favored either group. Nevertheless, we acknowledge that non-random loss to follow-up could still have influenced the absolute eradication estimates, and the PP results should be interpreted with this in mind.

These findings have two practical implications. First, they argue against the routine co-prescription of trimebutine with 14-day BQT for the purpose of reducing adverse events or improving eradication. Second, they redirect attention toward adjuncts acting on the mucosa-microbiota axis probiotics, rebamipide and, in specific rescue contexts, N-acetylcysteine[60]and toward regimen-level changes that have already demonstrated superior tolerability in randomized trials, including single-capsule BQT (Pylera) and potassium-competitive acid blocker (vonoprazan)-based dual or triple therapy[61-63]. Whether differentially tailored interventions for example, visceral-sensitivity directed agents combined with microbiota-directed adjuncts could improve outcomes in selected patient subgroups remains an open empirical question but lies beyond the scope of the present trial.

CONCLUSION

In this open-label randomized controlled trial of Korean adults undergoing 14-day BQT for H. pylori infection, the addition of trimebutine maleate 100 mg three times daily did not reduce the incidence of treatment-emergent adverse events, did not improve medication adherence, and did not increase eradication rates. The point estimates for every outcome were tightly clustered around the null, and the width of the observed CIs excluded clinically meaningful reductions in adverse-event rates. These findings together with mechanistic considerations and contrasting evidence for mucosa- and microbiota-directed adjuncts did not demonstrate evidence of improved eradication or tolerability with add-on trimebutine and do not, on present evidence, support its routine adjunctive use with standard BQT. However, because the trial was substantially underpowered owing to early termination, these results should be regarded as not demonstrating benefit rather than as establishing that trimebutine is ineffective, and adequately powered studies are warranted to confirm these findings. In the meantime, our data suggest that research attention may be more productively directed toward interventions targeting the mucosa-microbiota axis.