This phase 3, randomized, double-blind, placebo-controlled, parallel group study was conducted at nine centers across India. The study was conducted between December 2018 and March 2020 in accordance with all applicable laws, rules and regulations and with the Declaration of Helsinki and the International Council for Harmonization guidelines for Good Clinical Practice. The parents or legal guardians of all children participating in the study provided their written, informed consent at the time of the child’s enrollment. The study protocol was approved by the Independent Ethics Committee (Supplementary Appendix 1; Additional File 1) at each study site.

Infants and children aged between 6 months and 5 years with acute moderate diarrhea (WHO 2005 definition) of less than 48 h duration were included in the study. Acute moderate diarrhea was defined as the passage of unusual loose or watery stools at least three times in the previous 24-h period, presentation with mild or moderate signs of dehydration (defined as ‘some dehydration’ by the WHO), and rehydration could be performed orally [13, 32].

Exclusion criteria included known hypersensitivity to B. clausii or excipients in the investigational medical product or to other probiotics, pre-existing chronic gastrointestinal disease, current presence (or history within the previous 3 months) of blood, pus or mucus in stools, persistent diarrhea (duration > 14 days), clinically significant signs or symptoms of parasitic or bacterial diarrhea, severe or persistent vomiting, severe dehydration (WHO classification) [13] or malnutrition (< 50% of average weight for age [Indian Academy of Pediatrics classification [33], or treatment with antibiotics, antiparasitics, probiotics or prebiotics (use in dairy foods was permitted but not in baby formulas), antidiarrheals, laxatives or corticosteroids (except intranasal, ophthalmic, or topical formulations) within 2 weeks prior to study enrollment. Patients with long-term use of oral or intravenous corticosteroids within 6 months of enrollment were excluded. Details of other inclusion and exclusion criteria are presented in Supplementary Appendix 2; Additional File 3.

Infants and children in India with acute diarrhea were randomized in a 1:1 ratio to treatment with either the probiotic B. clausii preparation in combination with ORT, or matching placebo (i.e., in identical plastic vials to those containing active treatment but containing only water) plus ORT for 5 days. Zinc therapy was given to both groups for 14 days. ORT and zinc treatment are recommended by WHO and represent the current standard of care for the treatment of acute diarrhea [8, 13]. Treatment was given twice a day for 5 days i.e., Day 1 (morning) to Day 5 (evening), or Day 1 (evening) to Day 6 (morning). Patients were hospitalized from Day 1 and discharged at least 6 h after diarrhea resolution or for a maximum of 5 days corresponding to 120 h from the time to randomization. If resolution of diarrhea occurred before 5 days (120 h from the time to randomization), patients continued in the study at home until study end on Day 6. These patients were supplied with a diary for follow-up entries, for the patient’s parent or legal guardian to complete. They were required to fill in the diary relating to study treatment; and behavior and perceived efficacy scales. Stool record and food intake were also recorded. Empty vials and sachets were brought along to follow-up visits.

During the hospitalization period, efficacy and safety assessments were carried out daily by medical study staff. Stool frequency and consistency, vital signs, hydration status, and feeding patterns were assessed by the study team.

One treatment kit was dispensed to each patient to cover the whole study period. Each treatment kit contained 15 mini bottles (5 mL) of either the B. clausii preparation or placebo (two mini bottles/day for 5 treatment days plus five mini bottles provided as a reserve [in case of vomiting or spillage]). Each mini bottle of the B. clausii preparation contained 2 billion spores of poly antibiotic-resistant B. clausii spores. To preserve study blinding, the B. clausii preparation and matching placebo were provided in identical 5 mL bottles labeled with a unique treatment number generated by the study sponsor. Access to treatment codes was not allowed except under circumstances that required unblinding, e.g. for reporting a suspected unexpected serious adverse reaction.

Patients swallowed the contents of two mini bottles of their assigned treatment per day, one in the morning and one in the evening, for 5 days. The doses were administered at 12 hourly intervals after correction of the dehydration, as per the protocol. The first dose was administered in the morning of Day 1 and the last (tenth) dose was administered in the evening of Day 5. If the first dose was administered in the evening of Day 1, the tenth dose was administered in the morning of Day 6. B. clausii or matching placebo could be taken with or without food or drink. If a patient vomited immediately after intake, or in case of spillage, a new intake could be taken but only one extra mini bottle per day was allowed, from the reserve bottles. ORT and zinc were administered after B. clausii/placebo, with ORT being administered for 5 days and zinc for 14 days (as recommended by the WHO). ORT (based on WHO formula) was provided in the form of sachets of 4.4 g Electral® powder containing sodium (75 mOsmol/L), potassium (20 mOsmol/L), chloride (65 mOsmol/L), citrate (10 mOsmol/L) and dextrose (75 mOsmol/L); with a total osmolarity of 245 mOsmol/L. The powder was mixed in 200 mL of boiled water or water with a low mineral content at ambient temperature to create a solution, which could be kept in a refrigerator for up to 24 h. Breastfeeding could be continued throughout rehydration. Oral zinc acetate was provided in the form of bottles containing 100 mL of ready-for-use Zinconia® syrup; this was administered as a 5 mL oral daily dose containing 20 mg elemental zinc.

Study treatment was permanently discontinued in cases of hypersensitivity reaction or severe vomiting with intravenous rehydration. For patients with worsening clinical features or requiring intravenous rehydration therapy it was advised that medical treatment was sought, and the patient was withdrawn from the study. Temporary treatment discontinuation was considered by the investigators due to suspected adverse events (AEs) or when ORT was ongoing for replacement of stool losses. Re-starting study treatment was to be done under close clinical and/or laboratory monitoring once the Investigator considered that the relationship of the event to study treatment was unlikely, and that the study selection criteria had been met. A patient’s parents/legal guardian could permanently withdraw them from study treatment at any time irrespective of the reason, or this could be the Investigator’s decision. In cases where study treatment was discontinued, the patient was to remain in the study and be assessed using the procedure normally planned for the last dosing day with study medication.

Patients’ medical histories were taken at the screening/baseline visit on Day 1 (the first day of their hospitalization). At the baseline visit, all patients underwent a physical examination and assessment of vital signs, assessment of pre-duration of diarrhea prior to study entry, stool consistency according to Bristol score, food intake and breastfeeding status, stool frequency and diaper use (younger children), dehydration status and concomitant medications. Monitoring of patients for any hypersensitivity reaction such as rash was to be performed for at least 30 min after the first and second administrations of study treatment. AE monitoring, treatment compliance, and behavior and perceived efficacy scales (see below) were performed each day from study Days 1–6.

The impact of acute diarrhea on behavior was evaluated using a new observer-reported outcome (ObsRO) questionnaire developed for this study. This was to be completed at each study visit (each day) from Days 1 to 6 prior to any meaningful interaction with site staff and any physical examination. The questions were to be answered by the same parent or legal guardian approximately at the same time of each day during the entire study and be completed before the perceived efficacy scale. This behavior scale included five items to measure the caregiver’s observations of the impact of diarrhea on the child’s comfort, sleep, daily activities and eating. Items were developed based on the work of Fischbach et al. [34]. Responses were to be made using a 5-point categorical rating scale: ‘not at all’, ‘slightly’, ‘somewhat’, ‘very’, or ‘extremely’. The perceived efficacy of study treatment among caregivers was also evaluated using a new self-reported questionnaire developed for this study. The perceived efficacy scale contained a single item asking the parental caregiver “Are you worried that your child’s diarrhea is getting worse?”, the available answers being ‘not at all’ (score 1), ‘slightly’ (score 2), ‘somewhat’ (score 3), ‘very’ (score 4), or ‘extremely’ (score 5).

The primary efficacy endpoint was the duration of acute diarrhea (in hours), as counted from the time of randomization up to recovery (the first stool recorded as normal according to Bristol Stool Scale classification (score < 5, [35]). Secondary efficacy endpoints included the frequency of stools per day, and dehydration status (WHO classification: A = no dehydration, B = some dehydration, C = severe dehydration) [13] on each day the patient was hospitalized. Frequency of stools was determined as the sum of the frequency of stools and the frequency of diapers with stools. In addition, the following exploratory endpoints were assessed including bedside stool screening for rotavirus and adenovirus (in order to evaluate the efficacy of study treatment in patients with confirmed viral diarrhea), the impact of acute diarrhea on the affected child’s behavior, the perceived efficacy of study treatment among caregivers and the time to ‘no dehydration’ and duration of acute diarrhea from first intake of study treatment.

Statistical analyses were generated using SAS® version 9.4.

In a previous study (EudraCT # 2014–004636-19), the Kaplan-Meier percentage of resolved diarrhea at 3 days was 76% in the B. clausii group and 66% in the placebo group [36]. This was translated (using exponential distribution) as a hazard ratio (HR) of 1.323, considering the diarrhea recovery as event. Based on these assumptions, a total of 401 patients with event were needed to achieve 80% power to demonstrate superiority of B. clausii over placebo by two-sided log-rank test at a 0.05 type I error rate. Considering a maximum diarrhea assessment at 5 days, a total of 462 patients in two arms (231 patients per arm) were anticipated to be needed in the present study to reach the targeted number of patients with event in order to determine whether there was any significant difference in the primary efficacy endpoint between the active treatment and placebo groups.

The primary efficacy analysis population was the intent-to-treat (ITT) population, defined as all randomized patients, analyzed according to the treatment group allocated by randomization. The safety population considered for safety analyses was the randomized population who received at least one dose or part of a dose of the double-blind investigational product.

Continuous data were summarized using the number of available data, mean, standard deviation (SD), median, minimum and maximum for each study conduct approach group. Categorical and ordinal data were summarized using the number and percentage of patients in each treatment group. Patients with missing data were not counted in the percentages; percentages were calculated using only available data. Statistical analyses were performed at the 5% significance level using two-sided tests or two-sided 95% confidence intervals (95% CIs). For the time-to-event outcomes, Kaplan-Meier estimates (including curves) were computed and the 95% CI for the median survival times provided. The primary endpoint, time to diarrhea recovery (in hours) was compared between the treatment groups using a stratified log-rank test through the LIFETEST procedure in SAS with the fixed factors: age (< 2 years; ≥2 years), viral status (viral; non-viral) and breastfeeding status (yes; no; mixed). HR and corresponding 95% CI were provided using a Cox proportional hazard model which included the same fixed factors as in the stratified log-rank test. Subgroup analyses were performed for patient subgroups based on the baseline characteristics of age (< 2 years; ≥2 years), viral status (viral; non-viral), and breastfeeding status (yes; no; mixed) using a similar Cox analysis approach as was used for the primary analysis, adding the corresponding subgroup factor and subgroup factor-by treatment interaction. Three sensitivity analyses – using: i) the same Cox model as in the primary analysis but without adjustment for all covariates, ii) the same Cox model as for the primary analysis but with an additional covariate, the duration of pre-inclusion (current episode) diarrhea, and iii) the same Cox model as for the primary analysis but with the intake of prohibited and/or rescue medications considered as an additional reason for censoring – were also performed. The analysis of safety endpoints was descriptive and based on the safety population. Treatment-emergent AEs (TEAEs) were defined as AEs that developed or worsened during the treatment period.