Enterotoxigenic Escherichia coli hyperimmune bovine colostrum for preventing travellers’ diarrhoea (TD): a randomised, double-blind, placebo-controlled trial

Summary

Background

Enterotoxigenic Escherichia coli (ETEC) is a leading cause of travellers’ diarrhoea (TD). Hyperimmune bovine colostrum (HBC) containing ETEC specific antibodies has shown efficacy in controlled settings, but despite widespread use, real-world efficacy remains uncertain.

Methods

In this randomised, double-blind, placebo-controlled superiority trial, US and UK military personnel and adult civilians (aged 18–70 years) travelling to TD-risk regions were randomised (1:1) to receive ETEC HBC (600 mg twice daily) or placebo, starting 2 days before arrival and continuing for 10–20 days while in the risk area. The primary outcome was TD, defined by standardised criteria. Stool smears collected during diarrhoeal episodes and at the end-of-prophylaxis were tested by PCR for pathogen attribution. ETEC-attributable TD was assessed as a secondary outcome. This trial is registered with ClinicalTrials.gov, NCT04605783 and is complete.

Findings

Between June 2, 2022, and April 29, 2025, 975 participants were enrolled and 850 were randomised to ETEC HBC (n = 425) or placebo (n = 425). Of these, 317 ETEC HBC and 323 placebo recipients were included in the primary efficacy analysis. TD occurred in 22.7% of ETEC HBC and 26.3% of placebo recipients (relative protective efficacy [PE] 13.7 [95% CI −13.4 to 34.31]), below the protocol-specified efficacy threshold of ≥35%. ETEC accounted for 17% of attributable TD cases, with frequent multi-organism detections. The difference in ETEC attributable TD between the groups was not statistically significant (12.1 vs. 21.5%; PE: 43.8 [−28.5 to 75.4]). Adverse events were rare and mild (ETEC HBC 2/320 vs. placebo 2/325). Both events in the ETEC HBC group (frequent urination, hives) resolved spontaneously.

Interpretation

ETEC HBC did not significantly reduce overall TD risk, with interpretation of ETEC-specific efficacy constrained by limited power. Further research is warranted to determine whether broader ETEC antigenic coverage or multi-pathogen approaches can overcome the pathogen diversity and polymicrobial infections observed in field settings.

Funding

Defense Health Agency.

Research in context.

Evidence before this study

Preventive strategies for travellers’ diarrhoea (TD) remain limited. Although enterotoxigenic Escherichia coli (ETEC) has long been a target for TD prevention, no vaccines are licenced in the USA or UK for prevention of ETEC-associated TD, and available prophylactic options are constrained by variable effectiveness, dosing burden, tolerability, cost and antimicrobial-resistance concerns. Interest has therefore increased in commercially available, off-the-shelf dietary supplements such as probiotics, prebiotics, and passive immunoprophylaxis. We searched PubMed and ClinicalTrials.gov from database inception to April 1, 2026, using the terms “travellers’ diarrhoea”, “prophylaxis”, “probiotic”, “prebiotic”, “passive immunoprophylaxis” and “hyperimmune bovine colostrum (HBC)”. The search was restricted to English-language publications with no publication-type restrictions. We identified 13 clinical trials assessing field efficacy of probiotics, prebiotics, or synbiotics, and nine randomised controlled human challenge studies evaluating passive immunoprophylaxis products targeting ETEC and/or Campylobacter. Field trials of Saccharomyces boulardii probiotics and galacto-oligosaccharide prebiotics suggest modest efficacy, with protective efficacy estimates generally in the 20–45% range; however interpretation is limited by non-standardised outcomes, imprecision, adherence-related analyses, regional variability, and absence of pathogen-specific assessment. Human challenge studies of milk- or colostrum-derived bovine immunoglobulins have reported high short-term protection against homologous ETEC challenge strains, with protective efficacy estimates of approximately 60% to >90%. These studies have typically tested antibody preparations against represented or homologous challenge strains and have used single high-inoculum exposures, short dosing windows, small samples, intensive monitoring, and diarrhoea definitions that are less aligned with contemporary TD criteria, limiting generalisability to real-world travel settings.

Added value of this study

This randomised, double-blind, placebo-controlled pragmatic field trial provides a real-world evaluation of ETEC HBC for prevention of TD in military and civilian travellers. ETEC HBC did not significantly reduce overall TD incidence, with protective efficacy estimates below the prespecified threshold for clinical relevance across analysis populations. By incorporating standardised TD definitions and molecular pathogen attribution, this study provides important insight into the aetiologic diversity of TD among cases with available specimens, including frequent detection of multiple enteropathogens. Breakthrough ETEC infections were observed in both study arms, underscoring the challenges of antigenic diversity and incomplete protection despite targeted antibody coverage. Doxycycline malaria chemoprophylaxis was identified as a strong independent modifier of diarrhoea risk; an important confounder for future prevention trials. These results provide a robust benchmark for assessing nutraceutical interventions under real-world travel conditions.

Implications of all the available evidence

In this trial, ETEC HBC did not significantly reduce overall TD risk. Single-pathogen dietary supplement approaches may be insufficient in real-world travel settings, where pathogen heterogeneity, polymicrobial infections, variable exposure intensity, adherence, and concurrent antimicrobial exposures can influence efficacy. Future strategies should prioritize broader antigenic or multi-pathogen coverage and rigorous field evaluation with clinically meaningful and pathogen-attributed outcomes.

Introduction

Travellers’ diarrhoea (TD) affects 30–70% of military and civilian travellers in high-risk regions.^1,2 Bacterial pathogens cause most cases, with diarrhoeagenic E. coli – particularly enterotoxigenic E. coli (ETEC) - as the leading cause worldwide. Despite this burden, preventive strategies for TD remain suboptimal. Dukoral (oral killed whole-cell/recombinant cholera toxin B subunit vaccine) has been evaluated for prevention of ETEC-associated TD; however, randomised trial evidence has not shown consistent or statistically significant protection against ETEC-specific or all-cause TD, and no ETEC-specific vaccines are licenced in the United Kingdom or United States.^3,4 Pharmacological prophylaxis also has drawbacks. Rifaximin offers short-term protection in Latin America and Africa where non-invasive pathogens predominate,⁵ but its efficacy is inconsistent in Asia, and routine use is discouraged due to cost and the potential for antimicrobial resistance or gut microbiome disruption.^6,7 Bismuth subsalicylate offers a non-antibiotic alternative, but its utility is limited by its benign but commonly reported adverse effects (e.g., dark stool, tongue discolouration, nausea), frequent dosing and variable effectiveness across regions.^6,8

Interest has increased in fortified natural products, such as prebiotics, probiotics, and passive immunoprophylaxis, which are marketed in the United States for maintaining “digestive health” during travel.9, 10, 11 Despite widespread use, evidence for TD preventive efficacy is lacking. A recent meta-analysis reported a modest benefit for probiotics and prebiotics, constrained by product heterogeneity, bias in trial design, and the lack of confirmatory studies, underscoring the need for high-quality trials.¹²

Passive immunoprophylaxis targets TD prevention by targeting pathogen-specific virulence factors with polyclonal antibodies, unlike the non-specific microbiome effects of prebiotics and probiotics. ETEC hyperimmune bovine colostrum (ETEC HBC; Travelan®, Immuron, Australia) is sold over the counter in the U.S., Australia, and Canada and is derived from cows immunized with 13 ETEC strains.^13,14 It contains antibodies against key antigens, including O6 and O78 polysaccharides, colonization factor antigen I (CFA)/I, CFA/II, coli surface antigen (CS)3, CS4, CS6, and heat-labile enterotoxin (LT).¹⁵ The antibodies also exhibit cross-reactivity with additional ETEC O serotypes (O44, O55, and O127), and Shigella and Salmonella O polysaccharides. ETEC HBC has demonstrated significant protective efficacy (PE) in Controlled Human Infection Models (CHIM)¹⁶: in one study the PE of ETEC HBC ranged from 76.7 to 90.9% following oral challenge with an O78 strain¹⁴; likewise, a formulation targeting CFA/I colonization factors provided 90% protection after exposure to a CFA/I-bearing ETEC strain.¹⁷

CHIM studies provide proof-of-concept data on performance under controlled conditions, but they rely on a single challenge strain, which does not reflect the antigenic diversity of ETEC or other organisms encountered during travel. Moreover, strict dosing compliance and directly observed dosing in CHIM contrasts with operational environments, where adherence may be variable. Field trials are therefore essential to estimate efficacy in a more real-world environment, where pathogen heterogeneity, mixed infections, variable adherence, and concurrent medications influence outcomes.

To address this gap, we conducted a field efficacy trial with the primary aim of evaluating the PE of ETEC HBC against TD.

Methods

Study design and ethics

This randomised, double-blind, placebo-controlled trial enrolled U.S. and UK military personnel and civilian travellers at military pre-deployment facilities and two travel clinics.

The protocol was approved by institutional review boards in the U.S. (IDCRP-152, Uniformed Services University of the Health Sciences, Bethesda, MD) and the UK (Ministry of Defence Research Ethics Committee, Birmingham, UK) (2050/MODREC/21). Written informed consent was obtained from all participants prior to enrolment. The protocol was prospectively registered at ClinicalTrials.gov (NCT04605783). The study was exempt from the requirement for submitting an Investigational New Drug application to the U.S. Food and Drug Administration and from obtaining a Clinical Trial Authorization by the UK Medicines and Healthcare products Regulatory Agency as it was not intended to support a new drug indication. The primary outcome was specified in the trial registry as “gut health disruption”; this term corresponded to standard clinical definitions of TD and is described in detail in the Outcomes section.

The study protocol (version 8.0, 18 December 2024; Appendix 2) and statistical analysis plan (SAP; Appendix 3) are provided in the Supplementary Materials.

Participants

Eligible participants were aged between 18 and 70 years of age and had planned travel for ≥10 consecutive days to regions with TD risk (including South and Southeast Asia, Africa, Central and South America, the Middle East, and parts of Eastern Europe). Exclusion criteria included gastrointestinal (GI) disorders, immunosuppression, planned use of medications or nutraceuticals (e.g., prebiotic, probiotic) during travel, recent antibiotic use or diarrhoea within three days of starting prophylaxis. Malaria chemoprophylaxis and self-treatment of diarrhoea with antibiotics or antimotility agents during travel was permitted (see Appendix 2 for full protocol and eligibility criteria). Baseline demographic information, including age and biological sex, was collected at enrolment. Sex was recorded as a biological variable. Patients or the public were not involved in the design and conduct of this research.

Randomisation and masking

Participants were randomised 1:1 to receive ETEC HBC or placebo using a computer-generated permuted block design (blocks of 4) by a programmer who was independent of the trial (SAS v9.4). Participants and all study personnel involved in product administration, clinical evaluation and data analysis were blinded to the group assignments. Eligible participants at each site were sequentially issued study product kits based on the randomisation list.

Study product and placebo

ETEC HBC powder (Immuron Ltd., Victoria, Australia) was chosen over the commercially available caplet formulation (Travelan®) to enhance adherence, as the caplet excipients function primarily as binding agents and were not expected to influence gastrointestinal bioavailability.¹⁸ The manufacturer’s recommended total daily dose of 1200 mg was used and administered as 600 mg twice daily before meals to accommodate the unpredictable meal schedules during deployment. This dosing regimen was deemed more feasible than the manufacturer’s recommendation of two caplets three times daily before meals, which posed a higher risk of noncompliance due to the increased pill burden.

Sachets were prepared by an independent pharmacy and packaged as 44 masked sachets of either 600 mg of powdered ETEC HBC or visually identical placebo and dispensed within 30 days of travel. The placebo consisted of a food-grade maltodextrin (Maltrin QD® M500; Grain Processing Corporation, Muscatine, IA, USA), a bland carbohydrate powder without known effects on gut immunity or microbiome. Active and placebo sachets were visually indistinguishable and identical with respect to dosing and packaging.

Study procedures

All participants were dispensed a study product kit containing sufficient sachets for up to 20 days of prophylaxis. Participants were instructed to begin dosing two days prior to arrival in the TD-risk region and to continue for a minimum of 10 consecutive days and a maximum of 20 days while in the TD-risk region, depending on travel duration. Participants who remained in the TD-risk region beyond 20 days discontinued prophylaxis after completion of the supplied course. They recorded product adherence, GI symptoms, and use of medications for 22 days in a travel diary. Stool samples were self-collected using Whatman® FTA Elute Micro filter-paper cards (FTA Card, GE Healthcare Life Sciences, Marlborough, MA, U.S.) 2–3 days prior to starting and ending prophylaxis and during diarrhoeal episodes, and were mailed to central laboratory for testing. Post-travel surveys captured itinerary, malaria chemoprophylaxis, diarrhoea-related risk behaviours, tolerability of the study product, concomitant medications, and GI symptoms after completion of study product.

Diaries were reviewed with study coordinators at the post-travel visit or by telephone and missing or unclear entries were actively queried and resolved where possible. Missing data were not imputed. Adverse events were recorded at the post-travel visit. Adherence was assessed solely based on participant-reported travel diary entries; no additional objective verification measures were used. Participants with no evaluable travel diary data were excluded from analyses requiring diary-based outcome assessment. Participants with partial diary completion contributed outcome data through their last evaluable diary entry. For time-to-event analyses, participants who did not develop TD were right-censored at their last completed diary entry. Participants who met TD criteria before discontinuation of diary recording were included as TD cases.

Adverse events were prospectively and systematically assessed at the post-travel visit through participant interview and review of travel diaries and were documented on standardised case report for participant safety monitoring and regulatory reporting.

Testing with TaqMan array card (TAC)

Molecular detection using TAC is highly sensitive and may identify asymptomatic carriage; therefore, a nested matched case–control analysis was performed to estimate the attributable contribution of ETEC to TD. This approach compares pathogen detection in symptomatic cases vs. asymptomatic controls to derive odds ratios and estimate attributable fractions that account for background carriage. To assess pathogen distribution in TD cases, stool samples were tested using a customized TD-specific TAC PCR (Appendix 2 p 12). FTA Cards collected within 6 days of a TD episode and end-of-prophylaxis samples from asymptomatic subjects were tested. TD cases were matched 1:2 with asymptomatic participants (with replacement), by region of travel (i.e., Indo-Pacific, South America, Africa, Central Asia and Middle East, Eastern Europe). FTA cards collected during diarrhoeal episodes were matched with FTA cards from asymptomatic subjects with a comparable storage duration of ( ± 3 months). If no suitable controls were found, the storage duration was expanded to ± 5 months of the storage duration for the TD case. Nucleic acid was extracted from six, 3 mm punches per card and tested, data files were processed using QuantStudio 7 Pro software, version 1.7.2 (Applied Biosystems), and analysis was performed as previously described.¹⁹ A threshold cycle value of <35 was used as the detection threshold.

Pathogen distributions among TD cases were summarized by study arm. Because ETEC HBC was developed to target ETEC virulence factors, pathogen-specific analyses focused on ETEC-associated TD, including estimation of attributable fractions, relative protective efficacy, and characterization of ETEC virulence factors. For pathogen attribution analyses, TD cases were matched to asymptomatic controls by geographic region of travel. Comparisons of pathogen distributions by geographic region were not performed because of the limited number of pathogen-attributed TD cases available for analysis.

Outcomes

The prespecified primary outcome was “gut health disruption”, as defined in the study protocol and ClinicalTrials.gov registration. This corresponded to the standard clinical definition of TD and was defined as the passage of three or more unformed stools within a 24-h period; or two or more unformed stools with at least one associated symptom (such as abdominal cramps, nausea, vomiting, fever, blood in stool or urge to defecate) within a 24-h period; or self-reported antibiotic treatment for diarrhoea, as recorded in the participant travel diary.

The primary outcome was analysed as a binary variable (occurrence vs. no occurrence) at the participant level and summarized as the proportion of participants experiencing gut health disruption (i.e., TD) during their individual observation period. The observation period was defined as the number of days the participant completed daily travel diary entries during time spent in a TD–risk region, up to a maximum of 22 days, regardless of whether study product use was continuous throughout that period. Each participant was considered to have met the primary outcome at the first qualifying TD episode.

Prespecified secondary efficacy outcomes included differences in TD-associated enteropathogen distribution based on TAC PCR testing and evaluation of ETEC-associated TD; the proportion of participants with TD that negatively impacted daily activities; duration of TD episode; occurrence of multiple TD episodes; delayed TD defined as onset within 7 days after discontinuing study product; and time to first unformed stool associated with a TD episode (TFUS). Additional prespecified secondary outcomes included tolerability and adherence to study product during travel.

Prespecified exploratory efficacy analyses included the TD incidence rate. Person-time at risk accrued from arrival in the TD-risk region. For TFUS analyses, participants who experienced TD contributed person-time until their first qualifying episode, whereas asymptomatic participants were censored at their last diary entry, up to 22 days. TD episodes separated by at least 48 h without meeting TD criteria were considered distinct events.

Prespecified exploratory outcomes included functional bowel disorders including irritable bowel syndrome (IBS), assessment of gut microbiome composition, antimicrobial resistance gene profiles, and host immune biomarkers. These analyses are not included in the present report and will be reported separately.

Analysis populations

Participants were included in the intention-to-treat (ITT) population if they took at least one dose of study product. The modified ITT population (mITT) comprised those who took at least one dose of study product while in the risk area. The per-protocol (PP) cohort was defined as all participants who started taking study product at least 2 days before entering a TD risk region, and for at least 10 consecutive days, without missing ≥3 consecutive doses or ≥ 4 non-consecutive doses. Given the rapid action of ETEC HBC and the short incubation of most TD pathogens, a modified PP (mPP) group was defined to include those who started dosing by the day of arrival, remained TD-free for the first 48 h, and met the same adherence criteria as the PP cohort. Subjects in the mPP and PP cohort were censored at the time of discontinuation of prophylaxis for the efficacy analyses.

Statistical analysis

A target sample size of 868 participants was calculated for this randomised superiority trial, based on an expected 20% cumulative TD incidence in the placebo arm and an estimated relative PE of at least 35% for ETEC HBC, with 80% power, a one-sided type I error rate of 5%, and 25% anticipated loss to follow-up. A one-sided significance level was prespecified to reflect the a priori hypothesis of benefit and the evaluation of a clinically meaningful protective effect threshold relevant to operational use; however, results are presented using two-sided 95% confidence intervals for interpretability. The estimated TD incidence was informed by surveillance data from travellers and deployed military personnel.^1,2 The assumed relative PE for ETEC HBC was informed by the estimated ETEC prevalence, efficacy signals from CHIM studies, and findings from field trials of similar nutraceuticals, establishing 35% as a threshold that could support its consideration as a cost-effective solution for TD prevention.^11,20,21 The enrolment period was planned for three years. No formal interim analyses or stopping guidelines for efficacy or safety were prespecified or conducted, as this was a minimal-risk study with no anticipated safety concerns. Recruitment could be discontinued prior to the planned sample size based on feasibility considerations, including enrolment rates, study timeline, and observed event incidence.

Baseline demographic and travel characteristics were summarized by study arm in each cohort. Relative PE was estimated as 100 × (1−Relative Risk). TFUS was assessed using Kaplan–Meier methods in the mITT cohort, with log-rank tests for between-arm comparisons.

Attribution of TD to ETEC was determined by computing the odds of TAC detection in TD cases compared to matched asymptomatic controls using conditional logistic regression. The attributable fraction of ETEC TD, and the 95% confidence interval (CI), was calculated as prevalence × [1 − (1/OR)].²²

Missing data were not imputed. Travel diaries were reviewed with participants at the post-travel visit or by telephone and missing or unclear entries were actively queried and resolved where possible. Participants with no evaluable travel diary data were excluded from analyses requiring diary-based outcome assessment. Participants with incomplete diaries, including diaries with intermittently missing entries, were not excluded solely because of incomplete diary completion. All available diary information was used to determine TD outcomes. Participants who met TD criteria based on available diary information were included as TD cases regardless of subsequent missing diary entries. Participants who did not meet TD criteria contributed available diary data and were right-censored at their last evaluable diary entry for time-to-event analyses. Analyses of pathogen attribution were restricted to participants who submitted stool samples within the prespecified collection windows.

In addition, the PE of ETEC HBC was examined across prespecified subgroups as an exploratory analysis, including age, sex, military status, travel region, doxycycline use, country of enrolment (UK vs. U.S.), and detection of ETEC or enteroaggregative E. coli (EAEC). Race was prespecified in the protocol but subgroup analyses by race were not performed because of small and uneven subgroup sizes, which limited interpretability. For the binary TD outcome, subgroup-specific treatment effects were summarized using risk ratios and relative PE estimates. Effect modification was evaluated using generalized linear models with a binomial distribution and log link that included treatment, each subgroup factor, and treatment × subgroup interaction terms. For each subgroup factor, a separate model was fitted; for factors with more than two levels, global interaction tests were examined where appropriate. Because of multiple comparisons, subgroup findings were interpreted as exploratory.

To explore independent risk factors for TD and assess the robustness of treatment effect estimates after adjustment, we conducted a prespecified, exploratory multivariable logistic regression model to estimate adjusted relative risks, including treatment assignment (ETEC HBC vs. placebo), age, sex, travel risk region, travel population/country of enrolment, reported TD risk behaviours, and doxycycline malaria chemoprophylaxis use. Model fit was assessed using standard diagnostic measures, including McFadden’s R², Akaike information criterion (AIC), and the Hosmer–Lemeshow goodness-of-fit test. Stepwise variable selection was not performed; covariates were retained based on clinical relevance and prespecified analytic interest.

Primary efficacy was evaluated using a one-sided α of 0.05 as specified in the design; all reported CI are two-sided 95%. All analyses were performed using SAS version 9.4 and R software version 4.03 (R Foundation for Statistical Computing).

Role of the funding source

The funder (Defense Health Agency) and ETEC HBC manufacturer had no role in study design, data collection, data analysis, data interpretation, or writing of the report. ETEC HBC was purchased from the manufacturer for the study.

Results

Between June 2, 2022, and April 29, 2025, 975 participants were enrolled and 850 were randomised to ETEC HBC (n = 425) or placebo (n = 425; Fig. 1). Participants were followed for efficacy and safety outcomes from arrival in the TD–risk region through completion of the post-travel visit, which occurred within approximately 3–8 weeks after enrolment depending on travel duration. Recruitment was discontinued after 850 randomisations due to slower-than-expected enrolment and a higher-than-expected TD attack rate (24.5%), which was sufficient to meet the prespecified power assumptions without additional enrolment.

Fig. 1.

Trial profile. Participant flow through the trial from enrolment to analysis populations. Reasons for exclusion at each stage are shown for both ETEC HBC and placebo groups. Abbreviations: ETEC HBC = enterotoxigenic Escherichia coli hyperimmune bovine colostrum; mITT = modified intention-to-treat; mPP = modified per-protocol; PP = per-protocol; TD = travellers’ diarrhoea.

Of the randomised participants, 645 (75.9%) submitted travel diaries and took at least one dose, forming the ITT population for tolerability and safety analyses. Loss to follow-up, trip cancellations, and voluntary withdrawal accounted for 20.7% (n = 176) of randomised subjects excluded from the ITT population. These exclusions were more common among U.S. (20.9%) and UK (25.1%) military personnel or participants travelling for military purposes, compared to civilian travellers (11.1%). Five participants were excluded for not consuming product in the TD risk area, leaving 317 in the ETEC HBC group and 323 in placebo for the mITT cohort for efficacy analysis. The mPP and PP population included 409 and 327 participants, respectively.

Baseline demographic and travel characteristics were similar between groups in the mITT cohort (Table 1). Most participants were military personnel (94.2% of U.S. cohort, 75.0% UK cohort), predominantly young (median age 29 years; interquartile range [IQR] 24–36), male (83.1%), and frequently prescribed doxycycline for malaria prophylaxis (49.8%). Median duration on study product during travel was 20 days (IQR 18–20). Travel region was defined based on location during the prophylaxis and diary period; although some participants reported travel to multiple regions outside this window, none travelled to more than one region during the analytic observation period.

Table 1.

Baseline demographics and trip characteristics in the modified intention-to-treat (mITT) population by intervention group.

	ETEC HBC (n = 317)	Placebo (n = 323)
Median age, years (IQR)	29 (24–36)	29 (24–35)
Male sex, n (%)	265 (83.6)	267 (82.7)
U.S. participant, n (%)	228 (71.9)	240 (74.3)
U.K participant, n (%)	89 (28.1)	83 (26.7)
Race, n (%)
Caucasian	210 (66.2)	213 (65.9)
African American/Black British	41 (12.9)	44 (13.6)
Asian	22 (6.9)	27 (8.4)
Military personnel, n (%)	280 (88.3)	290 (89.8)
Officer	52 (18.6)	53 (18.3)
Enlisted	228 (81.4)	237 (81.7)
Civilian traveller, n (%)	37 (11.7)	33 (10.2)
TD risk region for travel, n (%)b
South/South-East Asia	211 (65.6)	222 (68.7)
Africa	73 (23.0)	71 (21.9)
Eastern Europe and Middle East	12 (3.8)	14 (4.3)
Central/South America	21 (6.6)	16 (4.9)
Purpose of travel, n (%)
Military purpose	276 (87.1)	284 (87.9)
Vacation or visiting friends and relatives	25 (7.9)	22 (6.8)
Other (business travel, humanitarian or religious missions)	16 (5.0)	17 (5.3)
Duration of travel in TD risk region, Median days (IQR)	39 (27–51)	38 (26–50)
Total duration of travel, Median days (IQR)	41 (28–52)	41 (28–51)
Duration on study product in TD risk region, median days (IQR)	20 (18–21)	20 (18–21)
Malaria chemoprophylaxis, n (%)
Not prescribed any malaria chemoprophylaxis	17 (5.4)	22 (6.8)
Doxycycline used for malaria chemoprophylaxis	154 (48.9)	163 (50.6)
Other medications used for malaria chemoprophylaxis	144 (45.7)	137 (42.6)
Self-reported behavioural risk factors for TDa, n (%)	268 (84.5)	273 (84.5)

mITT = modified intention-to-treat; TD = travellers’ diarrhoea; IQR = interquartile range; ETEC HBC = enterotoxigenic Escherichia coli hyperimmune bovine colostrum.

^aTD risk behaviours included any of the following on the post-travel survey: eating from street vendors or local vendors at a military dining facility, eating meals prepared at home by local population, drinking untreated water, eating poorly cooked or raw meat or seafood, eating poorly cooked or raw vegetables or salad, or eating unpasteurized cheese or drinking unpasteurized milk.

^bTD risk region represents the participants’ location during the period of prophylaxis. No participants travelled to more than one region.

Overall, 199 participants (31%) reported loose stool during the first 20 days after arrival and 157 (24.5%) met the TD outcome (incidence rate: 45.9 per 100 person-months). Dysentery and febrile acute watery diarrhoea were rare (Appendix 1 p 1) and no participants met the TD definition based solely on antibiotic use; all cases fulfilled standard stool and symptom criteria. Attack rates were higher among civilian travellers (29.1%) and U.S. military personnel (28.9%) compared to UK military personnel (7.0%). UK military personnel were deployed to controlled training environments with limited movement outside base settings, and a lower proportion reported TD risk behaviours (49.2% vs. 92.8% in the remaining cohort reporting at least one risk behaviour), which likely explains the lower TD incidence in this group. TD occurred in 72 of 317 (22.7%) participants in the ETEC HBC arm and 85 of 323 (26.3%) in the placebo arm, yielding a relative PE of 13.7% (95% CI: −13.4 to 34.3) in the mITT cohort, which was not statistically significant (Table 2). Similar non-significant results favouring ETEC HBC were observed in mPP (PE 19.3%; 95% CI: −12.6 to 42.2) and PP cohorts (PE 13.9%; 95% CI: −25.1 to 40.8). Kaplan–Meier analysis (log-rank test) indicated no statistically significant difference in TD-free survival between the two groups (Fig. 2). Rate-based analyses accounting for observation time produced comparable results (PE 14.9%; 95% CI: −16.4 to 37.9). Most TD episodes were mild, with infrequent antibiotic and loperamide use, and no differences between arms in impact on activities, duration of symptoms, or treatment (Table 2).

Table 2.

Primary and secondary efficacy outcomes for travellers’ diarrhoea (TD) by intervention group in the modified intention-to-treat (mITT), modified per-protocol (mPP), and per-protocol (PP) populations.

	mITT cohort		mPP cohort		PP cohort
	ETEC HBC (n = 317)	Placebo (n = 323)	ETEC HBC (n = 201)	Placebo (n = 208)	ETEC HBC (n = 158)	Placebo (n = 169)
Proportion of TD cases - % (95% CI), n	22.7 (22.6–22.9); 72	26.3 (26.2–26.5); 85	22.9 (22.7–23.1); 46	28.4 (28.2–28.6); 59	23.4 (23.2–23.6); 37	27.2 (27.0–27.4); 46
Relative protective efficacy (95% CI)	13.7 (−13.4 to 34.31)		19.3 (−12.6 to 42.2)		13.9 (−25.1 to 40.8)
TD incidence rate –per 100 person-months (95% CI)	42.2 (33.5–53.1)	49.6 (40.1–61.3)	41.8 (31.3–55.8)	53.8 (41.8–69.6)	44.2 (32.0–61.0)	52.3 (39.2–69.8)
Relative Protective efficacy (95% CI)	14.9 (−16.4 to 37.9)		22.5 (−13.9 to 47.3)		15.4 (−30.4 to 45.14)
TD cases with impact on activities,b n (%)	31 (9.8)	33 (10.2)	19 (9.4)	22 (10.6)	15 (9.5)	16 (9.5)
Median TD duration, days (IQR)	1.5 (1.5–3.5)	1.5 (1–2.5)	1.5 (1.5–3.5)	1.5 (1–2.5)	1.5 (1.5–4.5)	1.5 (1–2.5)
Antibiotic treatment for TD, n	3	1	1	0	0	0
Loperamide treatment for TD, n	14	18	11	15	8	10
TD within 7 d of stopping product,a n (%)	12/272 (4.4)	15/282 (5.3)	6/167 (3.6)	12/177 (6.8)	6/127 (4.7)	11/141 (7.8)
Subjects with >1 TD episode, n (%)	16 (5.0)	25 (7.7)	7 (3.5)	16 (7.7)	7 (4.4)	13 (7.7)
ETEC positive TD cases, n (%)	7 (14.9)	13 (26.5)	6 (16.7)	10 (27.8)	5 (17.9)	8 (26.7)
Attributable fraction, %	12.1	21.5	13.6	22.6	14.5	21.7
Relative protective efficacy (95% CI)	43.8 (−28.5, 75.4)		39.9 (−47.7, 75.6)		33.0 (−80.5, 75.2)

IQR = interquartile range; TD = travellers’ diarrhoea; mITT = modified intention-to-treat; mPP = modified per-protocol; PP = per-protocol; ETEC = enterotoxigenic Escherichia coli; ETEC HBC = enterotoxigenic Escherichia coli hyperimmune bovine colostrum; CI = confidence interval.

^aDenominator is the no. of participants who stayed in a TD risk area after stopping prophylaxis.

^bDaily activity impact was recorded using a 3-point scale (1 = normal activity; 2 = decreased activity; 3 = unable to participate in activities). Due to low event counts, categories 2 and 3 were combined and analysed as “TD with impact on activities.”

Fig. 2.

Kaplan–Meier curves showing the probability of remaining free from travellers’ diarrhoea (TD) over time in the ETEC HBC and placebo groups. Differences between groups were assessed using the log-rank test. Abbreviations: TD = travellers’ diarrhoea; ETEC HBC = enterotoxigenic Escherichia coli hyperimmune bovine colostrum.

Of 157 TD cases, 96 (61.1%) submitted a stool sample within 6 days of illness; 65 (67.7%) had a pathogen detected by TAC, compared to 40.0% (60/150) among matched asymptomatic controls. Diarrhoeagenic E. coli (ETEC, EAEC, EPEC) accounted for 91% of detections, and nearly half (49.2%) of positive samples had multiple pathogens (Table 3). ETEC was detected in 20.8% of diarrhoeal samples, with an attributable fraction of 17.1%. Although ETEC-associated TD among ETEC HBC recipients was numerically lower, differences were not statistically significant (12.1% vs. 21.5%; PE: 43.77 (95% CI: −28.5, 75.4)). Breakthrough ETEC strains included antigens covered by the product (CS3, CS6, LT) and others not covered (CS20, CS21) (Appendix 1 p 2).

Table 3.

Organisms detected by TAC PCR in stool smears collected during travellers’ diarrhoea (TD) episode.^a

Pathogen type	All (n = 65)		ETEC HBC (n = 33)		Placebo (n = 32)
	Solo pathogen (n = 33)	Multi-pathogen (n = 32)	Solo pathogen (n = 18)	Multi-pathogen (n = 15)	Solo pathogen (n = 15)	Multi-pathogen (n = 17)
Enterotoxigenic E coli, n (%)	5 (15.1)	15 (46.9)	3 (16.7)	4 (26.7)	2 (13.3)	11 (64.7)
LT positive only	4	11	3	3	1	8
LT/ST positive	0	0	0	0	0	0
ST positive only	1	4	0	1	1	3
Enteroaggregative E coli, n (%)	15 (45.5)	21 (65.6)	11 (61.1)	10 (66.7)	4 (26.7)	11 (64.7)
Enteropathogenic E coli, n (%)	7 (21.2)	16 (50.0)	2 (11.1)	10 (66.7)	5 (33.3)	6 (35.3)
Enteroinvasive E coli/Shigella, n (%)	1 (3.0)	1 (3.1)	0	1 (6.7)	1 (6.7)	0
Shiga toxin-producing E coli, n (%)	0	2 (6.2)	0	1 (6.7)	0	1 (5.9)
Campylobacter spp., n (%)	4 (12.1)	8 (25.0)	1 (5.6)	3 (20.0)	3 (20.0)	5 (29.4)
Plesiomonas shigelloides, n (%)	1 (3.0)	6 (18.8)	1 (5.6)	4 (26.7)	0	2 (11.8)
Salmonella spp., n (%)	0	1 (3.1)	0	0	0	1 (5.9)
Giardia lamblia, n (%)	0	1 (3.1)	0	0	0	1 (5.9)
Aeromonas, n (%)	0	1 (3.1)	0	1 (6.7)	0	0

TD = travellers’ diarrhoea; ETEC = enterotoxigenic Escherichia coli; LT = heat-labile toxin; ST = heat-stable toxin; TAC = TaqMan Array Card; PCR = polymerase chain reaction; ETEC HBC = enterotoxigenic Escherichia coli hyperimmune bovine colostrum.

^aOnly organisms detected in ≥1 stool samples listed. Norovirus PCR testing yielded indeterminate results due to failure of internal assay controls. 31 subjects that had no pathogen detected (ETEC HBC n = 14; Placebo n = 17).

In prespecified subgroup analyses, PE did not differ significantly across demographic or travel-related factors. In post-hoc exploratory analyses, non-significant differences favouring ETEC HBC were observed among participants reporting higher-risk behaviours, those travelling to Southeast Asia, those receiving doxycycline malaria prophylaxis, U.S. military travellers, and TD cases with ETEC detected (Appendix 1 p 3). In prespecified interaction screening analyses, no treatment–risk factor interaction reached statistical significance (all p > 0.05), including interactions with reported risk behaviours (p = 0.195), sex (p = 0.324), travel region (AFRICOM p = 0.964; PACOM p = 0.356), doxycycline use (p = 0.110), study population (civilian p = 0.416; UK military p = 0.059), or age (p = 0.501). Across models, McFadden’s R² values were small, Akaike information criterion values differed modestly, and model fit was adequate, with Hosmer–Lemeshow goodness-of-fit tests indicating no evidence of lack of fit (all p > 0.05). In a final multivariable model including treatment, all prespecified factors, and treatment–factor interaction terms simultaneously, no interaction term was statistically significant; the main treatment effect remained non-significant, whereas site-level differences persisted.

Multivariable analyses showed that doxycycline malaria chemoprophylaxis use was associated with lower TD risk (relative risk [RR] 0.7; 95% CI: 0.52–0.93; p = 0.01). As with cumulative diarrhoea rates and TFUS analyses above, ETEC HBC was not associated with significant TD risk reduction (RR 0.87; 95% CI: 0.67–1.14; p = 0.32) (Table 4).

Table 4.

Multivariable analysis of risk factors associated with TD in the mITT cohort.

	Relative Risk (95% CI)	p-value
ETEC HBC	0.87 (0.67–1.14)	0.32
TD risk behaviour during travel	1.03 (0.64–1.67)	0.90
Female traveller	0.95 (0.65–1.39)	0.79
TD Risk Region (ref. Cent/South America, East Europe, Middle East)
Africa	0.73 (0.38–1.41)	0.35
South/South-East Asia	1.03 (0.58–1.82)	0.93
Took doxycycline for malaria chemoprophylaxis	0.7 (0.52–0.93)	0.01
Population (ref. U.S. military)
Civilian travel	1.0 (0.6–1.66)	0.98
UK military	0.27 (0.12–0.61)	0.002
Age >34 years (ref. age 18–34 years)	1.21 (0.92–1.6)	0.18

p-values in bold are <0.05.

TD = travellers’ diarrhoea; mITT = modified intention-to-treat; CI = confidence interval; ETEC HBC = enterotoxigenic Escherichia coli hyperimmune bovine colostrum.

ETEC HBC was well tolerated (Table 5 and Appendix 1 p 4). Adverse events were rare and similar between arms (ETEC HBC: 2/320 vs. placebo: 2/325). Both events in the ETEC HBC group (frequent urination, hives) were mild, possibly related, and resolved spontaneously.

Table 5.

Adverse events reported in the ETEC HBC vs. Placebo arms - ITT population.

Adverse Event (n)	Severity	Relationship	Outcome	Product
Itchy rash (1)	Mild	Possible	Resolved	Placebo
Lightheadedness (1)	Mild	Possible	Resolved	Placebo
Frequent urination (1)	Mild	Possible	Resolved	ETEC HBC
Hives (1)	Mild	Possible	Resolved	ETEC HBC

ETEC HBC = enterotoxigenic Escherichia coli hyperimmune bovine colostrum; ITT = intention-to-treat.

Discussion

In this randomised, multi-centre field trial, ETEC HBC did not demonstrate meaningful PE against TD in a real-world travel setting. Symptom duration, functional impact, and treatment utilisation were similar between groups, with no evidence of benefit across multiple outcome measures. Although ETEC HBC was safe and has evidence of biological activity in prior studies,^14,15 this did not translate into meaningful protection against TD in this field setting.

Previous CHIM studies using a single ETEC strain homologous to the bovine vaccine components have reported efficacy estimates of 77–91%.^14,16 The absence of PE in this field trial is likely multifactorial, reflecting differences between controlled experimental conditions and real-world exposure. Although diarrhoeagenic E. coli were the most frequently detected organisms among TD cases, ETEC was only identified as the pathogen in 17% of cases, and nearly half of positive samples demonstrated polymicrobial detections. This complexity of pathogen exposure, together with imperfect adherence in real-world settings, highlights the challenges in extrapolating CHIM results to field efficacy. Detection of enteropathogens by TAC does not establish etiologic causality, particularly in the setting of frequent polymicrobial detections and background asymptomatic carriage; findings should therefore be interpreted in the context of known limitations of multiplex molecular diagnostics in distinguishing colonization from true disease attribution.

Several additional factors likely contribute to the discrepancy between CHIM and field effectiveness. Travellers are exposed to a wide range of ETEC strains with substantial antigenic diversity, whereas CHIM studies use homologous strains. Inoculum exposure in field settings is variable and often repeated over prolonged periods, in contrast to the controlled and limited exposures in challenge studies. Co-infections with multiple enteropathogens are common in field settings but absent in CHIM models, and adherence to dosing is closely monitored in experimental settings but more variable in real-world use. Furthermore, while ETEC HBC formulation targets key antigens (O6, O78, CFA/I, CS3, CS4, CS6, LT), it does not provide complete antigenic coverage, and breakthrough infections included strains expressing both targeted and untargeted colonization factors. Because passive immunoprophylaxis relies on pathogen-specific polyclonal antibodies that block virulence factors, its effectiveness may be reduced in the context of high inoculum exposure and polymicrobial infections, and may also be influenced by degradation within the gastrointestinal tract.²³ These factors together likely explain the modest and non-significant efficacy observed.

Notably, the use of doxycycline malaria chemoprophylaxis was associated with lower TD risk in the multivariable model and subgroup analyses showed non-significant differences favouring ETEC HBC among participants receiving doxycycline. Doxycycline could plausibly influence susceptibility to TD through antimicrobial activity or microbiome-mediated mechanisms; however, these effects could not be directly evaluated in this study.²⁴ Because doxycycline use was not randomized and adherence was not fully characterised, associations involving doxycycline should be interpreted as exploratory and non-causal. Doxycycline also represents a concomitant antimicrobial exposure that could act as an effect modifier, particularly in the subgroup analyses. Importantly, doxycycline is not recommended for TD prophylaxis because of high levels of tetracycline resistance among enteric pathogens, concerns regarding adverse events, and the potential for antimicrobial resistance selection and microbiome disruption.

Evidence regarding the PE of other dietary supplements is inconsistent, highlighting both the difficulty of conducting rigorous field trials and the challenge of demonstrating efficacy in real-world settings. Two randomised trials evaluating a galacto-oligosaccharide prebiotic demonstrated a PE of approximately a 40–45% that was dependent on adherence. However, these effects were imprecise, as intention-to-treat analyses were not statistically significant and confidence intervals included the null.^11,20 Notably, in the Hasle et al. trial, PE was driven by the per-protocol and secondary analyses and was limited to short, self-limiting episodes, while neither study assessed pathogen-specific aetiology, leaving mechanisms of protection unclear. Another trial reported ∼20% PE for the probiotic Saccharomyces cerevisiae but lacked a standardised diarrhoea definition and adherence data.²¹ In contrast, the P2 trial’s multi-centre design, geographically diverse cohort, prespecified outcomes, defined analysis populations, and incorporation of TAC-based pathogen attribution collectively enabled a rigorous and robust evaluation of PE. The study also incorporated practical dosing schedules and real-world conditions, enhancing external validity.

In the context of passive immunoprophylaxis, these findings have important implications for TD immunoprophylaxis strategies. The polymicrobial and heterogeneous nature of TD in real-world settings suggests that interventions targeting a single pathogen may be insufficient to achieve meaningful reductions in disease burden. Future approaches may require broader antigenic coverage or combination strategies integrating passive immunoprophylaxis with other preventive interventions, although the effectiveness of such strategies remains to be established. Unlike conventional pharmaceuticals, HBC products have minimal systemic absorption and act locally at the mucosal surface, making traditional pharmacologic frameworks less informative for evaluating real-world effectiveness. These results underscore the importance of evaluating such interventions in field settings.

This study had several limitations. The study population consisted predominantly of young military personnel, which may limit generalizability to older populations or travellers with different risk profiles; the study was also not powered to detect differences by sex. Attrition between randomisation and follow-up, together with sub-optimal compliance with stool collection, reduced the completeness of outcome data. This may have limited power for secondary and exploratory analyses and reduced the number of evaluable samples, particularly for pathogen-specific analyses, potentially affecting precision and introducing bias. In addition, some exposures were not fully characterised. Detailed timing and adherence to doxycycline chemoprophylaxis were not systematically captured, limiting assessment of its impact on TD risk and its potential interaction with the study product. Prior Dukoral vaccination was not assessed. While Dukoral is not licenced for TD prevention in the US and UK, it is licenced for cholera in some regions, and prior vaccination among a small number of participants cannot be excluded. Subgroup findings should be interpreted cautiously, as differences observed across subgroups were not supported by treatment–factor interaction tests and should be considered hypothesis-generating. Observations in specific groups (e.g., UK travellers or participants receiving doxycycline) require confirmation in larger studies. Variability in TD attack rates across regions and populations introduced heterogeneity that may have diluted treatment effect estimates and reduced the ability to detect subgroup effects. Finally, the dosing schedule for ETEC HBC was modified from the manufacturer’s recommended regimen to improve feasibility in a field setting, which may have influenced effectiveness. Optimal dosing is difficult to establish for orally administered HBC because pharmacokinetic and pharmacodynamic relationships are not well characterised and are inherently difficult to quantify in the gastrointestinal tract.²⁵ These limitations should be interpreted in the context of a large, pragmatic trial designed to assess real-world effectiveness and are unlikely to alter interpretation of the primary outcome.

Our findings, together with the broader evidence base, suggest that immunoprophylaxis products targeting single pathogens are unlikely to provide robust protection against TD in diverse travel settings. Future work should prioritize improved antigen coverage, optimized dosing strategies, and evaluation in real-world field settings using clinically meaningful outcomes such as moderate-to-severe TD and pathogen-attributed cases. Although ETEC HBC was safe and well tolerated, routine use for TD prevention cannot be recommended based on current evidence.

Contributors

Conceptualization - David R Tribble, Tahaniyat Lalani, Chad K Porter. Data Curation - Tara Griffith, Huai-Ching Kuo, Tahaniyat Lalani, Michele Cerroni. Formal Analysis - Huai-Ching Kuo, Ningbo Jian. Funding Acquisition - David R Tribble, Tahaniyat Lalani. Methodology - David R Tribble, Tahaniyat Lalani, Chad K Porter, Tom Troth, Joanna Rimmer, Daniel Burns. Project Administration - Tara Griffith, Nathanial K Copeland, Jerry Q Barton, Christina Schofield, Rhonda E Colombo, Tom Troth, Sylviane Defres, Michelle Rizzo-Pelley, Christie A Joya. Writing original draft - Tahaniyat Lalani, Ningbo Jian, Huai-Ching Kuo. Writing - review and editing - all authors. Tahaniyat Lalani, Tara Griffith and Huai-Ching Kuo directly accessed and verified the underlying data reported in the manuscript at the level of the individual participant. All authors had full access to all the data in the study and accept responsibility to submit for publication. All authors read and approved the final version of the manuscript.

Data sharing statement

Synopsis of the study protocol and statistical analysis plan are provided in Appendix 2 and 3. The raw data sets generated in this study are not publicly available due to privacy sensitivities regarding military service member cohorts. Proposals may be submitted up to 36 months following article publication for aggregated, de-identified participant data that underlie the results reported in this article. Data will be made available beginning 3 months and ending 36 months following article publication to investigators whose proposed use of the data has been approved by an independent review committee identified for this purpose. Proposals should be directed to tlalani@idcrp.org; to gain access, data requestors will need to sign a data access agreement.

Declaration of interests

All authors declare no competing interests.