Preventing otitis media and respiratory tract infections with the oral probiotic Streptococcus salivarius K12 Authors: Caroline Halley, Claire Honeywill, Janice Kang, Phillipa Barnes, Gordon Purdie, Gwyneth Steenson, Rae Noble-Adams, Thorsten Stanley, Kathryn Fuge, Lynn McBain, Ben Hudson, Julian Crane Published: 2026 ABSTRACT AbstractAcute otitis media (AOM) and respiratory tract infections are common in early childhood and are major contributors to antibiotic use. The oral probiotic Streptococcus salivarius K12 has demonstrated antimicrobial and immunomodulatory activity in vitro, and preventive effects in children with recurrent AOM. We aim to determine the efficacy of S. salivarius K12 in preventing AOM in the general infant population. We conducted a multicentre, double-blind, randomised, placebo-controlled trial in Wellington and Christchurch, New Zealand. Children were enrolled at 3–6 months of age, their households were randomised 1:1, and daily study product was given from age 6 to 24 months. The primary outcome was the rate of doctor-recorded AOM. Secondary outcomes included time to first AOM episode, respiratory infections, antibiotic prescribing and healthcare utilisation. Analyses were by intention to treat using mixed-effects negative binomial regression adjusted for region and, for the primary outcome, prior AOM, with random effects for general practice and family. A total of 428 children were randomised by household and 368 (86%) contributed primary outcome data (placebo n = 180; active n = 188). Doctor-recorded AOM occurred in 70/180 (39%) placebo recipients and 70/188 (37%) probiotic recipients (rate ratio 0.84, 95% CI 0.57–1.24). Time to first AOM episode did not significantly differ (hazard ratio 0.89, 95% CI 0.63–1.26). No significant differences were observed for AOM severity, respiratory infections, antibiotic prescribing or healthcare utilisation. Higher-adherence analyses were also not statistically significant. Conclusions: Daily oral S. salivarius K12 did not reduce AOM incidence or respiratory infections in children under 2 years; therefore, routine use for AOM prevention in general paediatric populations is not supported.Trial registration: Australian New Zealand Clinical Trials Registry—ACTRN12618000130268; prospectively registered on 30 January 2018, before enrolment of the first participant on 24 August 2018. What is Known:• Streptococcus salivarius K12 is a naturally occurring oral probiotic and has demonstrated antimicrobial and immunomodulatory activity in vitro, and preventive effects in children with recurrent AOM.What is New:• Whether there are preventative effects of daily use for AOM and respiratory infections in infants from the general population.Keywords: Streptococcus salivarius K12, Probiotic, Acute otitis media, Respiratory tract infections, Infants, Randomised controlled trial Introduction Probiotics have been defined by the World Health Organisation as “live microorganisms which when administered in adequate amounts confer a health benefit to the consumer” [1]. Most probiotics derive from natural associations with milk or plants or from the environment [2]. Other probiotics derive from natural human commensals that have been shown to confer a health benefit on natural carriers [3], for example S. salivarius which has discreet antimicrobial and immune modulating properties [4]. These microbial properties include inhibitory and bactericidal peptides with activity against pathogens such as S. pneumoniae, H. influenzae and M. catarrhalis [4–6]. Recently in-vitro studies have suggested that some of the anti-bacterial peptides may also have anti-viral properties. For example, they may bind parts of the spike protein of certain viruses thus decreasing cell entry and modulate antiviral immune defences [7]. Respiratory tract infections (RTI), including Acute Otitis Media (AOM), are a major driver of childhood primary care use [8] and antibiotic prescribing in children under 5 years [9, 10]. Given the in-vitro activity of S. salivarius it should be an ideal probiotic to reduce Upper Respiratory Tract Infections (URTIs), including AOM which are often viral or bacterial and often specifically involve S. pneumoniae, H. Influenzae and M. catarrhalis. Three previous clinical trials have reported reductions in AOM incidence and antibiotic prescribing following daily administration of S. salivarius K12 in children with recurrent AOM aged 3–10 years [11–13]. However, only one of these involved a double blind randomised controlled trial [11], and none assessed outcomes over more than 3 months. A more recent double-blind, placebo-controlled randomised clinical trial in children aged 2–6 years for a 6-month period did not demonstrate a significant reduction in overall AOM incidence compared with placebo, despite good tolerability and adherence [14]. Collectively, these findings suggest that the effectiveness of S. salivarius K12 may be influenced by age, timing of intervention, and underlying disease mechanisms, and that evidence in younger infants—during the period of early immune and microbiome development—remains limited. The objectives of our study were to conduct a double blinded randomised controlled trial to examine the effect of a daily probiotic, S. salivarius K12 given to children from age 6 months to 2 years to reduce episodes of AOM and RTI and antibiotic prescribing for RTIs. Methods A two-centre double blind randomised controlled trial enrolled children at 3–6 months of age and administered probiotics or placebo daily from 6 to 24 months in Wellington and Christchurch, New Zealand. The protocol was approved by the Health and Disability Ethics Committee (HDEC 16/CEN/190), overseen by the Health Research Council of New Zealand Data Monitoring Committee (DMC), and prospectively registered with ANZCTR (ACTRN12618000130268). Participants and recruitment Fifteen general practices in the Wellington region and fourteen in the Christchurch region participated in recruitment. Children aged 3–6 months were identified from practice databases, and invitation letters with study information were sent to caregivers via participating practices. Additional participants were recruited through community advertising. Caregivers were screened by telephone, and eligible children were enrolled before the intervention commenced. Randomisation and intervention started at 6 months of age. Children were eligible if they were aged 3–6 months at enrolment, resided in the wider Wellington or Christchurch regions, and were born at ≥ 28 weeks’ gestation. Exclusion criteria included regular probiotic use (≥ 3 times per week), a history of severe milk or soy allergy, declined consent for access to medical records, significant cardiac disease, immunosuppressive illness or medication use, or concurrent participation in other health research studies. Children commenced the assigned study product at 6 months of age and were followed for the 18-month intervention period until age 24 months; no post-intervention follow-up was conducted. Recruitment concluded in May 2020, and follow-up and data collection continued until late 2021. Randomisation and blinding Randomisation was conducted at the family/household level using a computer-generated permuted block sequence with randomly varying block sizes up to 10, stratified by region. Pre-prepared study packs containing either active or placebo product were labelled with a unique study identification number and allocated sequentially in order of enrolment. Eligible children from the same household were allocated to the same treatment group. Participants, caregivers, investigators, and outcome assessors were blinded to treatment allocation throughout the study. Intervention Active and placebo powders were encapsulated and contained identical excipients (trehalose, lactitol, maltodextrin, a small quantity of isomalt, and natural strawberry flavouring), with appearance, smell, taste and packaging identical. The active formulation additionally contained freeze-dried Streptococcus salivarius K12, providing at least 1.25 × 10⁹ colony-forming units (CFU) per dose at the date of manufacture. Study products were manufactured, bottled, and labelled by BLIS Technologies Ltd. Participants began one dose daily at 6 months of age and continued for 18 months, until age 24 months. Study product was supplied in three 6-month allocations: at enrolment and at 6 and 12 months after intervention commencement. Each dose was mixed with a small amount of water to make a paste applied to the mouth and gums, direct application to food or application of dry powder to the teeth and gums following oral cleaning, as per the manufacturer’s instructions. Storage, handling, and product integrity Study products were stored at ≤ 4 °C prior to distribution. Bottles exposed to heat were replaced. During use, bottles were stored at room temperature for up to 2 months provided they were not exposed to heat. Replacement bottles were provided where product was lost, overheated, or substantially spilled. Probiotic viability was periodically tested by the manufacturer on returned bottles. Protocol and adherence monitoring Participants received text message reminders to start and stop the administration of the powder, and follow-up telephone questionnaires to assess adherence. Returned bottles were collected by courier or at study visits and weighed. Differences between pre- and post-use weights were used as an objective measure of adherence. Outcomes The primary outcome was the rate of doctor-recorded AOM during the 18-month intervention period. Secondary outcomes included the number of medical visits for respiratory tract infections (RTIs), doctor prescriptions of antibiotics for RTI or OM, number of doctor diagnoses of croup, and severity of doctor-diagnosed OM in terms of bilateral disease, visible discharge from one or both ears, bilateral discharge, and referral to ear, nose and throat (ENT) services. Other secondary outcomes were the frequency of doctor-diagnosed RTI and dental caries identified from Community Oral Health records and assessed by a dental public health specialist for Wellington participants. Parent-reported outcomes included ear infections, RTIs, hospital admissions for respiratory conditions, outpatient referrals to paediatrics or ENT services, and antibiotic courses for respiratory illness. Outcome ascertainment and GP record review Outcome data were obtained from caregiver questionnaires and structured review of general practice records. General practice records were reviewed for the prespecified 18-month period that the participants were taking the powder only, and events occurring after the individual participant’s scheduled completion date were not included. Unscheduled GP visits for illness were classified as consultations and reviewed using a dedicated electronic data capture system. Study clinicians identified AOM, RTI, croup and associated antibiotic prescribing. Analyses used all participants with data available for the relevant outcome: GP-record outcomes required retrievable GP records, whereas parent-reported outcomes used completed caregiver questionnaires. Consequently, denominators differed between outcome types. No outcome data were imputed. Adverse events Adverse events were captured through scheduled questionnaires and additional reporting as required. Serious adverse events were reviewed by the DMC. COVID-19–related protocol modifications Protocol modifications were implemented in response to COVID-19 public health restrictions. COVID-19 restrictions required remote enrolment (n = 30) or follow-up for some participants, couriered study product, delayed bottle return, and use of parental report or clinical records for anthropometric data. Oral mucosal swabs could not be collected at the end-of-study visits. Sample size Based on New Zealand estimates of 499 AOM episodes per 1000 child-years in children aged under 1 year and 377 per 1000 child-years in children aged 1–2 years [15], the expected rate from age 6 to 24 months was 586 episodes per 1000 children. A total sample of 420 participants provided 90% power at the 5% significance level to detect a 50% reduction in this rate using an unclustered negative binomial regression, assuming the reported distribution of multiple episodes [15]. A 50% reduction was selected as a clinically meaningful effect and was smaller than the 70% reduction reported in an earlier non-randomised, non-placebo-controlled and unblinded study [12]. No allowance was made for attrition because withdrawal of consent for the GP-record search was initially expected to be rare. Statistical analysis Analyses were intention-to-treat. Event rates were compared using mixed-effects negative binomial regression adjusted for region, with random effects for family and general practice; primary AOM was additionally adjusted for prior ear infection. Analyses used SAS 9.4 (GLIMMIX). Fisher’s exact test was used when one group had no events and all events occurred in different families. Time to first AOM was analysed using a Cox proportional hazards model. Available outcome data were analysed without imputation; secondary outcomes were exploratory with no adjustment for multiple comparisons. Higher-adherence analyses used the same model specifications except where indicated. A post hoc sensitivity analysis compared prior GP- or nurse-practitioner-diagnosed ear infection by GP-record availability using Fisher’s exact test and logistic regression adjusted for treatment group and region; interactions assessed differences by treatment group and region. Results Progress through the study is shown in Figs. 1 and 2. Four hundred twenty-eight children were enrolled, 210 in Wellington and 218 in Christchurch, 368 (86%) had retrievable GP outcome data after the planned 18-month intervention and follow-up period. Recruitment ended in May 2020, while intervention follow-up continued until late 2021. Baseline characteristics were broadly similar between groups (Table 1), although the active group had numerically more prior ear infection (8.8% vs 5.2%) and prior antibiotic treatment for ear infection (6.0% vs 2.4%). Participant flow diagram. EOI, expression of interest. GP-data numbers show the primary-outcome population; denominators for parent-reported outcomes vary according to questionnaire availability (Table 2) Timeline of participation in the study—highlighting COVID-19 restrictions Table 1Baseline characteristicsCharacteristicPlacebo (n = 211)Active (n = 217)Female, n (%)101 (47.9%)105 (48.4%)Gestational age, weeks, median (IQR)39.7 (38.6–40.6)39.6 (38.6–40.4)Caesarean delivery, n (%)57 (27.0%)73 (33.8%)Birth weight, kg, median (IQR)3.5 (3.2–3.8)3.5 (3.2–3.9)Prior ear infection11 (5.2%)19 (8.8%)Prior bronchiolitis15 (7.1%)22 (10.1%)Prior other chest infection7 (3.3%)9 (4.1%)Prior croup4 (1.9%)9 (4.1%)Prior antibiotics for ear infection5 (2.4%)13 (6.0%)Total courses of antibiotics for ear infection, n816Prior ENT referral7 (3.3%)4 (1.8%)Eczema67 (31.8%)62 (28.6%)Maternal smoking in pregnancy9 (4.3%)7 (3.2%)Household smoking1 (0.5%)1 (0.5%)Breastfed to 6 months of age164 (77.7%)167 (77.0%)Sibship (half and full siblings)131 (62.1%)138 (63.6%)Pneumococcal vaccination198 (93.8%)196 (90.3%)Values are n (%) unless otherwise stated; gestational age and birth weight are median (interquartile range). Gestational age, delivery mode and birth weight were available for 216 participants in the active group. Baseline characteristics Values are n (%) unless otherwise stated; gestational age and birth weight are median (interquartile range). Gestational age, delivery mode and birth weight were available for 216 participants in the active group. Table 2 shows the primary and secondary outcomes. Doctor-recorded AOM occurred in 70/180 placebo participants (129 episodes) and 70/188 active participants (109 episodes), with no significant difference in event rates (adjusted rate ratio 0.84, 95% CI 0.57–1.24; p = 0.32). In a post hoc sensitivity analysis, prior GP- or nurse-practitioner-diagnosed ear infection was more common among children without available GP notes (10.9% vs 5.1%), but the adjusted association was not statistically significant (OR 2.40, 95% CI 0.91–6.35; p = 0.078), with no evidence of interaction by treatment group or region. No secondary outcome differed significantly between groups. Time to first doctor-diagnosed AOM also did not differ significantly (hazard ratio 0.89, 95% CI 0.63–1.26; p = 0.51; Fig. 3). Table 2Primary and secondary outcomesPlaceboPercent (n/N)Total episodesS. salivarius K12Percent (n/N)Total episodesK12/placebo rate ratio (95% CI) p*Primary outcomeNumber of episodes of doctor reported AOM70/180 (39%); 12970/188 (37%); 1090.84 (0.57, 1.24) p = 0.32Secondary outcomesSeverity of doctor diagnosed OM—Bilateral OM46/180 (26%); 8049/188 (26%); 831.02 (0.62, 1.68) p = 0.93- Discharge OM10/180 (6%); 186/188 (3%); 70.60 (0.17, 2.10) p = 0.35- Bilateral discharge OM3/180 (2%); 40/188 (0%); 0p = 0.12†- ENT referral7/180 (4%); 92/188 (1%); 20.29 (0.04, 2.19) p = 0.19Number of medical visits for Upper and Lower Respiratory Tract Infections (composite)48/180 (27%); 7955/188 (29%); 891.08 (0.61, 1.91) p = 0.76‡Doctor prescriptions for antibiotics for respiratory conditions (OM, URTI, LRTI)74/180 (41%); 18783/188 (44%); 1650.91 (0.61, 1.36) p = 0.57Number of doctor diagnoses with croup31/180 (17%); 3927/188 (14%); 451.10 (0.49, 2.47) p = 0.77‡Number of GP visits for OM75/180 (42%); 17375/188 (40%); 1630.93 (0.62, 1.40) p = 0.68Frequency of doctor diagnosed RTI48/180 (27%); 7255/188 (29%); 811.08 (0.64, 1.82) p = 0.74Number with dental caries at 2 years of age0/77 (0%)0/74 (0%)not estimable as zero eventsParental reported ear infections71/192 (37%); 16772/196 (37%); 1510.93 (0.62, 1.38) p = 0.68Parental reported RTI188/192 (98%); 1552193/196 (98%); 17631.05 (0.90, 1.22) p = 0.46Parental reports of hospital admissions for respiratory conditions (OM, URTI, LRTI)7/192 (4%); 78/196 (4%); 141.19 (0.33, 4.30) p = 0.76Parental reports of outpatient referrals to paediatrics or ENT for respiratory tract infections (OM, URTI, LRTI)12/170 (7%); 317/179 (4%); 270.83 (0.11, 6.04) p = 0.83Parental report of course of antibiotics taken for respiratory conditions (OM, URTI, LRTI)80/192 (42%); 20583/196 (42%); 1750.90 (0.62, 1.30) p = 0.52Values are n/N (%) and total episodes, where applicable.*Mixed-effects negative binomial regression adjusted for region, with random effects for family and medical centre; primary AOM also adjusted for prior ear infection.†Fisher’s exact test.‡Region and medical-centre terms omitted. Primary and secondary outcomes Values are n/N (%) and total episodes, where applicable. *Mixed-effects negative binomial regression adjusted for region, with random effects for family and medical centre; primary AOM also adjusted for prior ear infection. †Fisher’s exact test. ‡Region and medical-centre terms omitted. Percentage first doctor diagnosed episode of acute otitis media. (Hazard ratio 0.89 (95% CI 0.63,1.26) p = 0.51) The powder was well tolerated. Parent-reported serious adverse events (hospitalisation for ≥ 24 h) did not differ between groups (19/211 placebo; 27/217 active; p = 0.60) and were not considered related to study product by the DMC. Self-reported adherence did not differ between groups. The proportion taking powder on at least 5 days per week fell from approximately two-thirds in the first 6 months to around half at 12–18 months (all p ≥ 0.75). Bottle returns were also similar (p = 0.67); 28.4% of placebo and 30.0% of active participants returned no bottles, while 27.0% and 26.3%, respectively, returned all nine. Powder use was nearly identical (median 204.9 g placebo vs 202.2 g active; p = 0.95). Among those participants with higher adherence (Table 3), point estimates for the primary AOM outcomes favoured S. salivarius K12, including lower proportions and rates of doctor-recorded AOM and fewer GP visits for OM than in the placebo group. However, these differences were not statistically significant and the confidence intervals were wide. Table 3Primary outcome amongst those with greater treatment compliancePlaceboPercent reported, total,S. salivarius K12Percent reported, total,K12/placebo rate ratio (95% CI) p*Primary outcomeNumber of episodes of doctor reported AOM28/74 (38%); 5825/84 (30%); 380.59 (0.28, 1.27) p = 0.13Secondary outcomes – severity of AOMSeverity of doctor diagnosed OM- Bilateral OM16/74 (22%); 3516/84 (19%); 270.86 (0.31, 2.35) p = 0.70- Discharge OM5/74 (7%); 72/84 (2%); 30.65 (0.07, 5.90) p = 0.62- Bilateral discharge OM1/74 (1%); 10/84 (0%); 0p = 0.47†- ENT referral3/74 (4%); 31/84 (1%); 10.45 (0.02, 11.85) p = 0.53‡Number of GP visits for OM29/74 (39%); 7926/84 (31%); 610.74 (0.35, 1.61) p = 0.34Values are participants with at least one event/participants in the higher-adherence subgroup (%) followed by total episodes.*Mixed-effects negative binomial regression.†Fisher’s exact test with one observation per family.‡No adjustment for centre. Primary outcome amongst those with greater treatment compliance Severity of doctor diagnosed OM - Bilateral OM Values are participants with at least one event/participants in the higher-adherence subgroup (%) followed by total episodes. *Mixed-effects negative binomial regression. †Fisher’s exact test with one observation per family. ‡No adjustment for centre. Discussion This randomised, double-blind, placebo-controlled trial found no evidence that daily S. salivarius K12 from age 6 months to 2 years reduced AOM, other RTIs or antibiotic prescribing. Findings were consistently null across intention-to-treat outcomes, including time to first AOM, severity and healthcare utilisation. Higher-adherence estimates favoured S. salivarius K12 but were imprecise and not statistically significant. Our results are consistent with the placebo-controlled trial by Sarlin et al., which also found no significant reduction in AOM [14]. Together, these trials suggest that oral S. salivarius K12 does not substantially reduce AOM risk in otherwise healthy children. Earlier clinical studies reporting reductions in AOM incidence largely involved children with recurrent disease and often used alternative delivery routes, including intranasal administration of streptococcal strains [11–13]. These interventions may act through local competitive colonisation of the nasopharynx rather than systemic or oral immune modulation. In addition, several earlier studies were small, short in duration, or lacked placebo control, limiting confidence in the magnitude or generalisability of reported effects. Systematic reviews of probiotics for ear and respiratory outcomes in children have reported modest overall reductions in AOM incidence, but interpretation is complicated by substantial heterogeneity in probiotic strains, delivery methods, treatment duration, and study quality [16, 17]. Many studies include mixed probiotic formulation. Reviews of S. salivarius K12 specifically have included small, short-duration or non-randomised studies with limited control groups [18, 19]. More recent meta-analysis of probiotic, prebiotic and symbiotic interventions suggests modest reductions in AOM incidence overall but highlights heterogeneity and variable risk of bias across studies [20]. Our findings from a large, well-controlled trial therefore provide important strain-specific evidence regarding oral S. salivarius K12 in a general infant population. The absence of measurable clinical benefit is notable given the well-described antimicrobial and immunomodulatory properties of S. salivarius K12 demonstrated in vitro [7]. Although mucosal colonisation following oral administration of K12 has been documented in several studies, colonisation appears to be variable between individuals and often transient rather than stable [14, 21]. Detectable carriage may occur shortly after administration, but persistence over longer periods is inconsistent, and sustained ecological modification of the nasopharyngeal microbiota has not been reliably demonstrated. As a result, the presence of the organism on mucosal surfaces may not be sufficient in duration or magnitude to meaningfully alter pathogen transmission dynamics or host susceptibility to infection. The infant nasopharyngeal microbiome undergoes rapid ecological change, with colonisation influenced by host immunity, viral exposure, antibiotic use and environmental factors [22, 23]. AOM pathogenesis is multifactorial, involving viral-bacterial interactions, Eustachian tube function, host immune maturation and microbial community dynamics [24]. Targeting a limited number of bacterial pathogens and transient augmentation may be an insufficient way to alter disease risk if viral triggers or host susceptibility dominate. This study has several important strengths. It was a large, multicentre, double-blind, placebo-controlled randomised trial with prolonged follow-up, and 86% of randomised children contributed GP-record data for the primary outcome. Outcome ascertainment incorporated structured review of general practice medical records, reducing reliance on parental recall. Objective measures of adherence were obtained through bottle-weight monitoring and specific questionnaire assessment, and product viability was periodically confirmed. The intervention covered a substantial period of early childhood, when AOM risk is high and preventive strategies are most clinically relevant. Together, these design features strengthen the evidence regarding the lack of efficacy of oral S. salivarius K12 in this age group. Several limitations should be considered. COVID-19 public health restrictions substantially altered respiratory pathogen circulation and healthcare utilisation worldwide and in New Zealand specifically. Population lockdowns implemented to control SARS-CoV-2 transmission resulted in marked reductions in influenza and other respiratory viruses, including near-elimination during periods of strict population quarantine measures [25]. Paediatric healthcare utilisation and diagnoses of infectious diseases also declined substantially during lockdown periods, including documented reductions in acute otitis media presentations [26–28]. New Zealand studies demonstrated dramatic reductions in infant respiratory hospitalisations during the elimination strategy period [29]. The first national lockdown occurred when trial participation was at its peak (around 380 active participants) and therefore had the greatest potential to reduce exposure to respiratory pathogens or alter healthcare-seeking behaviour. In contrast, the second lockdown occurred late in follow-up when relatively few children remained active (approximately 50) and is less likely to have materially influenced overall results. The observed AOM rate was broadly consistent with the magnitude anticipated in the sample-size calculation, despite pandemic disruptions. However, COVID-19-related changes may have nevertheless reduced opportunities to observe differences between treatment groups. Other limitations may have reduced precision or introduced bias. GP outcome data were available for 368 rather than the planned 420 children; 60 (14%) lacked primary outcome data, mainly because of record-retrieval and reconsent difficulties during COVID. Complete-case analysis could be biased if record availability related to infection risk or treatment response. Prior GP- or nurse-practitioner-diagnosed ear infection was more common among children without GP notes, suggesting missingness may not have been independent of baseline AOM risk, although the estimate was imprecise and did not differ by treatment allocation or region. The study was powered to detect a 50% reduction and may have missed a more modest effect. Baseline prior ear infection and antibiotic-treated ear infection were numerically higher in the active group, but adjustment for prior ear infection did not materially alter the treatment estimate. Adherence declined over time, with approximately two-thirds taking powder on at least 5 days per week initially and around half by 12–18 months. This reflects the difficulty of sustaining a daily preventive treatment for 18 months. Higher-adherence estimates favoured S. salivarius K12 but were imprecise and should be interpreted cautiously. Clinical implications Taken with existing placebo-controlled evidence, our findings do not support routine oral S. salivarius K12 for prevention of AOM or RTIs in the general population of infants and young children. The decline in adherence later in the study also suggests that long-term daily powder administration may be difficult for families over prolonged periods. Future research directions Further research would clarify whether probiotic interventions have a role in selected populations or under different conditions. Priorities should include children at high baseline risk of recurrent AOM; alternative delivery routes that enhance nasopharyngeal colonisation; mechanistic studies of colonisation persistence and microbial community effects; and earlier or longer interventions during microbiome development. Studies should also examine reasons for lower adherence and strategies to improve it. Understanding host–microbe interactions in early childhood and infancy and identifying populations most likely to respond may help determine whether targeted microbial prevention can meaningfully reduce respiratory disease burden. Conclusions In this large randomised trial, oral S. salivarius K12 did not significantly reduce AOM, RTIs or antibiotic prescribing in children aged 6 months to 2 years. These findings do not support routine use for AOM prevention in general paediatric populations. Authors’ contributions Conceptualization: CHa, JC. Methodology: CHa, CHo, JK, PB, GP, GS, RN-A, TS, KF, LM, BH, JC. Formal analysis: GP. Investigation: CHa, CHo, JK, PB, GS, RN-A, LM, BH, JC. Writing - original draft preparation: CHa, JK, GP, JC. Writing - review and editing: CHa, CHo, JK, GP, TS, BH, JC. Funding acquisition: CHa, JC. Resources: KF. Supervision: CHa, JC. Funding Open Access funding enabled and organized by CAUL and its Member Institutions. The study was funded by a grant from the Health Research Council of New Zealand (16/071). BLIS Technologies supplied the active and placebo powder used in this trial at no cost. BLIS Technologies had no role in study design, participant recruitment, data collection, data analysis, interpretation of the results, or manuscript preparation. Data availability The data generated by this study will not be shared with the public or other investigators as permission was not sought from the participants. Declarations Ethics approval This study was conducted in accordance with the New Zealand National Ethical Standards for Health and Disability Research and Quality Improvement. Participants were informed of the study’s purpose, risks, and benefits, and a consent-to-participate declaration was completed. Written informed consent was obtained from all caregivers/parents prior to their child’s participation. Ethical approval was granted by the Health and Disability Ethics Committee (HDEC) (Approval number: 16/CEN/190). Consultation was undertaken with the Ngāi Tahu Research Consultation Committee (NTRCC) to ensure culturally safe practice in alignment with the principles of Te Tiriti o Waitangi. Competing interestst The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data generated by this study will not be shared with the public or other investigators as permission was not sought from the participants.