A probiotic containing L. plantarum and P. acidilactici strains for treating upper respiratory infection in children aged 6 months to 5 years: a randomized, double-blind trial

Abstract

Upper respiratory infections (URIs) are highly prevalent, especially among young children. We aimed to investigate the efficacy and safety of a specific probiotic formula in shortening the duration of pediatric URIs of likely viral origin. A randomized, double-blind clinical trial was conducted in children aged 6 months to 5 years, with URI with pharyngitis as diagnosed by a specialist. Participants were allocated to probiotic (strains L. plantarum KABP022, KABP023, and KABP033 and P. acidilactici KABP021) or placebo, twice daily for 15 days, and further followed up at 30 and 60 days from baseline. Antipyretic (ibuprofen) and antihistamine (cetirizine) were allowed as comedication and controlled. Co-primary outcomes were days of fever (axillary temperature > 37.5 °C) and days of pain/discomfort (Face, Legs, Activity, Cry, and Consolability [FLACC] score > 3), as recorded in a diary. Causative viruses were identified by PCR in nasopharyngeal swabs. Seventy-five children were randomized and included in the analysis. The most frequent causative agents were rhinoviruses, coronaviruses, and respiratory syncytial virus. Compared to placebo, the durations of fever and of pain/discomfort were reduced by 1.1 days (95% CI 0.5–1.7; p = 0.001) and by 0.7 days (95% CI 0.2–1.2; p = 0.006) in the probiotic group, respectively. Sensitivity analyses by detected virus, age, sex, and recent URI episodes produced similar results. The use of comedication and other symptoms did not differ between study groups. No significant differences were found in the number of adverse events between the groups.

Conclusion: Intervention with this probiotic formula was associated with a shorter duration of fever and pain/discomfort in pediatric URIs with pharyngitis.

Trial registration: Prospectively registered on Oct 5^th, 2023, as ISRCTN-66122782.

What is Known: • Acute upper respiratory tract infections (URIs) are highly prevalent in children, but due to their frequent viral etiology, treatment is often symptomatic only. Some probiotics have shown promising results against URIs, but evidence mostly relies on untrained, subjective diagnosis by participants or their parents. Recently, a probiotic formula containing strains L. plantarum KABP022, KABP023, and KABP033 and P. acidilactici KABP021 reduced viral load and duration of symptoms in a randomized trial in 300 adults with laboratory-confirmed Covid19, against placebo.

What is New: • In children with URI with pharyngitis, as diagnosed by a specialist, intervention with said probiotic formula reduced the duration both of fever and of pain/discomfort, compared to placebo. Results were robust to sensitivity analyses by age, sex, number of recent URIs, and by most prevalent causative viruses in the study population.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00431-025-06587-7.

Introduction

Acute upper respiratory tract infections (URIs) are highly prevalent, especially in children [1]. Clinical symptoms are usually self-limited, but occasionally, the infection spreads to adjacent organs, yielding more serious clinical manifestations [2]. URIs reduce the quality of life for the affected children and their parents and result in medical and non-medical costs. Causative agents in children are mainly viruses, the most prevalent since the onset of the Covid19 pandemic being rhinoviruses, respiratory syncytial virus (RSV), and SARS-CoV2 [3, 4]. Lack of antiviral therapies against most URI viruses usually limits treatment to symptom management (e.g., antipyretics, antihistamines). Besides, evidence for dietary interventions to help treat URIs is limited [5].

Probiotics are defined as “live microorganisms that when administered in adequate amounts, confer a health benefit on the host” [6]. Mounting evidence indicates a crosstalk between the gastrointestinal tract and the respiratory system, along with their respective microbiomes, referred to as the gut-lung axis [7, 8]. In this regard, recent meta-analyses found oral probiotics may help prevent or shorten upper respiratory infections such as cold and influenza, but also noted significant limitations, such as high heterogeneity, overreliance on subjective diagnosis, and unclear treatment concealment [9, 10]. Of note, heterogeneity is not unexpected, as probiotic immune effects are considered to be strain-specific [6, 11–13].

The AB21® probiotic formula includes four strains of the Lactobacillaceae family: Lactiplantibacillus plantarum KABP022, KABP023, and KABP033, plus Pediococcus acidilactici KABP021. A previous randomized clinical trial in 300 adult patients found this formula improved PCR-confirmed Covid19, according both to objective signs (e.g., viral load, digital chest X-ray) and subjective symptoms [14]. However, evidence of its effect against other viruses and in the pediatric population was lacking. In this study, we aimed to test the efficacy and safety of the same probiotic formula in shortening the duration of specialist-diagnosed URIs of likely viral origin in a pediatric population.

Methods

Study design and setting

We conducted a randomized, single-center, double-blind, placebo-controlled trial with two parallel arms (1:1 ratio). The study was conducted at Hospital Medica Sur (Mexico City, Mexico). Approval was duly obtained from the hospital’s Committee for Ethics and Research in Studies in Humans (approval ref. CB148-CEI153). The study adhered to the tenets of the Declaration of Helsinki, was prospectively registered as ISRCTN-66122782, and was reported according to Consolidated Standards of Reporting Trials (CONSORT).

Participants

We included pediatric participants, aged 6 months to 5 years, with an upper respiratory infection with concurrent fever and pharyngitis (based on objective symptoms), as diagnosed by a pediatrician (DA-P), and symptoms’ onset within 48 h of enrolment. All participants were already known and cared for by the study investigators as their usual pediatricians before entering the study. Parents had to report the child experiencing ongoing fever (> 37.5 °C) and a FLACC score (Face, Legs, Activity, Crying and Consolability) > 3 [15] and to provide written informed consent. Exclusion criteria, based both on caregiver-reported information and medical records, were (i) symptoms suggestive of bacterial infection (tonsillitis or tonsillar exudate, cervical lymphadenitis, otitis media or scarlatiniform rash); (ii) preterm infants or those weighing < 2500 g at birth; (iii) failure to thrive (as per WHO standards); (iv) use of antibiotics, antivirals, or probiotics for > 48 h within 2 weeks of study entry (applied to infants and mothers of breastfed participants); (v) history of ≥ 2 invasive infections (meningitis, osteomyelitis, sepsis); (vi) history of two or more severe respiratory infections (upper or lower) within 12 months of study entry; (vii) chronic diarrhea or gastroesophageal reflux disease; (viii) congenital heart, respiratory, or digestive disease; (ix) primary or secondary immunodeficiency; and (x) asthma.

Procedures

The study consisted of a 15-day intervention phase (Visits 1 to 2) plus a 45-day follow-up phase (Visits 3 and 4, Supplementary Fig. 1). A computer-generated 1:1 (probiotic:placebo) randomization list, of block size 4, was prepared by personnel not directly participating in the study and kept in a sealed envelope until all study subjects completed the intervention. Study investigators recruited and allocated patients and performed all study procedures. Investigators, parents, and pediatric participants were blinded to treatment allocation. Probiotic and placebo capsules looked identical, both contained an off-white tasteless powder, and both were delivered in undistinguishable, anonymous boxes.

At Visit 1 (baseline, day 0), parents of eligible participants who provided informed consent received the allocated intervention and were briefed on product administration, proper procedures for body temperature measurement and FLACC rating, and use of the paper-based patient diary for daily recording of (i) symptoms (axillary temperature (°C), FLACC score, and presence/absence of each of cough, rhinorrhea, and nasal congestion); (ii) comedication; and (iii) adverse events (AEs). A nasopharyngeal swab (Copan, Italy) was taken and stored at − 20 °C in Universal Transport Medium for posterior analysis of causative agents. At Visit 2 (day 15 from enrolment), patient diaries were returned, and investigators verified study compliance and assessed the clinical status of participants and unsolicited AEs. The clinical status of participants and unsolicited AEs were assessed again during follow-up at Visits 3 and 4 (30 and 60 days from enrolment ± 3 days, respectively). Analysis of causative agents was performed in batch at Grupo Diagnostico Aries (Mexico), using the Respiratory Flow multiplex PCR (Vitro SA, Spain). Additionally, a phone call was held between Visits 1 and 2 and between Visits 3 and 4 to promote study adherence and protocol compliance.

Interventions

All treatments were taken orally. Probiotic or equally looking placebo was taken twice daily (morning and evening). Probiotic capsules contained lyophilized Pediococcus acidilactici strain KABP021 (CECT7483) and Lactiplantibacillus plantarum strains KABP022 (CECT7484), KABP023 (CECT7485), and KABP033 (CECT30292), totaling ≥ 2 billion CFUs (colony-forming units, as per plate counting in MRS medium), in a dextrin carrier. Placebo capsules contained the dextrin carrier only. Microbial quality was verified as described elsewhere [14]. Parents were instructed to open the capsules and dissolve the contents in water or lukewarm milk right before administration.

Study participants could also receive standard of care for pediatric URIs, consisting of ibuprofen as an antipyretic drug (10 mg/kg of body weight thrice per day) and/or cetirizine as an antihistamine drug (0.25 mg/kg of body weight per day) as rescue medication. Should study participants develop symptoms of bacterial infection after study entry, investigators could prescribe rescue antibiotics: amoxicillin clavulanate (at 45 mg/kg of body weight per day) or clarithromycin (at 7.5 mg/kg of body weight per day) in case of allergy to beta-lactams. No antivirals or other comedications were used during the study.

Predefined outcomes

The co-primary study outcomes were the number of days with fever (i.e., highest daily axillary temperature > 37.5 °C) and the number of days with a pain/discomfort score > 3 (as measured daily with the FLACC score) [15] during the intervention period. Predefined secondary outcomes included (i) area under the curve (AUC) of body temperature and FLACC score during the intervention period (trapezoidal method); (ii) number of days with cough, nasal congestion, rhinorrhea, use of antipyretic (ibuprofen), use of antihistamine (cetirizine), and use of antibiotics during the intervention period; (iii) number of unscheduled medical visits during the intervention period as reported by parents at Visit 2; (iv) occurrence of each of the different symptoms (axillary temperature > 37.5 °C, FLACC > 3, cough, rhinorrhea, and nasal congestion) and their total number, as assessed by investigators at Visits 3 and 4; and (v) number of unscheduled medical visits as reported by parents at Visits 3 and 4. Because participants attended Visit 2 (day 15) at different times of the day (morning, afternoon, or evening), only data from finished days (i.e., 1 to 14) were used in outcomes related to the intervention period to ensure comparability between participants.

Power calculation and statistical analysis

Sample size was calculated to allow detecting a difference in duration of 1 day with a standard deviation (SD) of 1.2 days, a statistical power of 90%, and an alpha of 2.5% (i.e., multiplicity-adjusted for two co-primary endpoints), leading to 36 participants per group. This number was increased to 45 per group to allow for a drop-out rate of 20%.

Statistical analysis was performed on all randomized patients (intent-to-treat) by allocated group, using IBM SPSS Statistics program (version 22) without imputation for missing diary entries. Differences in baseline characteristics between groups were not formally assessed using p-values, following CONSORT 2010 and European Medicines Agency (EMA) guidelines. Differences between study groups for the co-primary endpoints (days of highest axillary temperature > 37.5 °C and of FLACC score > 3) were assessed with generalized linear models (GLM), with baseline temperature or FLACC score as covariates (respectively); normality of residuals was verified by Q-Q plots. Differences between groups for secondary endpoints were assessed by Mann–Whitney test (exact version) for continuous or ordinal data, and chi-squared test (with N-1 correction) [16] for categorical data. In post hoc exploratory analyses, GLMs for the co-primary endpoints were adjusted for detection of rhinovirus (yes/no), detection of coronaviruses (yes/no), detection of RSV (yes/no), age (up to 12 months vs. older), sex (male/female), and number of URIs in the previous 2 months. Spearman test was used to assess bivariate correlations. All tests were two-sided. Significance for the two co-primary outcomes was set at α = 0.025, while other outcomes were assessed at nominal α = 0.05.

Results

Participants

From October 2022 to February 2023, 75 pediatric participants were consecutively recruited, underwent the 15-day intervention, and were followed up until the last study visit, 2 months from randomization, and diaries were available for all of them (Fig. 1). As no dropouts were observed, recruitment was stopped before reaching 90 participants once target statistical power was achieved for the co-primary outcomes (no other protocol deviations occurred during the study). The median age was 15 months, and 39 (52.0%) were female. Other than fever (> 37.5 °C) and a FLACC score > 3 as per entry criteria, the most common URI symptom was rhinorrhea (82.7%), and most participants were already taking ibuprofen (93.3%). Baseline characteristics were well balanced between study groups (Table 1).

Fig. 1.

Participant CONSORT flow diagram

Table 1.

Baseline characteristics of study participants

	Placebo (n = 38)	Probiotic (n = 37)
Age (months; median, range)	14.5 (6–59)	15.0 (6–53)
Weight at birth (kg; mean, SD)	3.16 (0.25)	3.23 (0.31)
Sex (female; N, %)	16 (42,1%)	20 (54,1%)
Born by C-section (N, %)	16 (43.2%)	14 (37.5%)
Breastfed until 6 months of age (N, %)	32 (84.2%)	32 (86.5%)
Completed scheduled vaccines by age (N, %)	38 (100%)	37 (100%)
Current weight (kg; median, range)	10.1 (6.8–24.2)	10.1 (6.5–22.1)
Currently attending kindergarten (N, %)	32 (84.2%)	29 (78.4%)
Number of URIs within two months (mean, SD)	1.5 (1.0)	1.8 (0.8)
Body temperature at baseline (°C; mean, SD)	37.8 (0.7)	37.7 (0.8)
FLACC score at baseline (median, range)	8.0 (4–9)	7.0 (4–9)
Presence of other URI symptoms baseline:
• Rhinorrhea (N, %)	33 (86.8%)	29 (78.4%)
• Nasal congestion (N, %)	22 (57.9%)	20 (54.1%)
• Cough (N, %)	22 (57.9%)	22 (59.5%)
Using antipyretics at baseline (N, %)	35 (92.1%)	35 (94.6%)
Using antihistamines at baseline (N, %)	12 (31.6%)	10 (27.0%)
Using antibiotics at baseline or within 2 weeks (N, %)a	0 (0%)	0 (0%)
Hours since the onset of symptoms (mean, SD)	21.8 (11.8)	22.2 (12.0)
Oxygen saturation at baseline (SpO2; mean, SD)	95.4 (1.9)	95.0 (1.8)
Number of viruses detected (mean, SD)	1.4 (1.0)	1.3 (0.8)

^aEither antibiotic given to children or taken by lactating mothers if the children were breastfed

All patients provided nasopharyngeal swabs. The most frequent etiologic agents were rhinoviruses (36.0%), coronaviruses (28.0%, mostly SARS-CoV-2), and respiratory syncytial virus (RSV, 25.3%). Several patients were positive for more than one virus. The overall distribution of viruses did not differ between study groups (Supplementary Fig. 1).

Primary outcomes

There were 1.1 less days of fever (95% CI 0.5 to 1.7; p = 0.001) and 0.7 days less of pain/discomfort (95% CI 0.2 to 1.2; p = 0.006) in the probiotic group, compared to placebo (Fig. 2). Higher temperature and FLACC score at study entry resulted in more days of fever and of pain/discomfort, respectively (i.e., baseline effect, both p < 0.001).

Fig. 2.

Primary outcomes: baseline-adjusted means (and 95% CI) for days of fever (defined as axillary temperature > 37.5 °C, left) and days of pain/discomfort (defined as FLACC score > 3, right) during the intervention period, together with summary results of their generalized linear models (GLMs). Analysis per intent-to-treat by allocated group (placebo n = 38, probiotic n = 37)

Secondary outcomes

During the intervention phase, AUC for body temperature and FLACC score were significantly lower in the probiotic group vs. placebo (both p < 0.05), but no differences were found in the duration of other symptoms (rhinorrhea, nasal congestion, and cough). Likewise, days of use of antipyretics and of antihistamines did not differ between study groups. During the 6-week follow-up phase, the incidence of nasal congestion and the total number of symptoms were significantly lower in the probiotic group at Visit 3 compared to placebo (i.e., 15 days after completing the intervention, p < 0.05), but not at Visit 4. Noteworthy, all participants displaying symptoms at Visit 3 had already achieved axillary temperature < 37.5 °C and a FLACC score of 0 at Visit 2. The complete list of secondary endpoints is reported on Table 2.

Table 2.

Secondary outcomes. Median (range) or number of participants (% within group)

	Placebo (n = 38)	Probiotic (n = 37)	p-value
Intervention phase (reported at Visit 2)a
AUC temperature (°C)b	517 (505–525)	515 (505–522)	0.021
AUC FLACC scoreb	39.5 (10.0–53.5)	31.0 (7.0–51.0)	0.002
Days with cough	4 (0–14)	2 (0–13)	0.488
Days with nasal congestion	2.5 (0–12)	1 (0–12)	0.522
Days with rhinorrhea	7 (0–13)	5 (0–13)	0.267
Days of use of antipyretic (ibuprofen)	4 (2–7)	4 (2–6)	0.108
Days of use of antihistamine (cetirizine)	5.5 (0–12)	4 (0–13)	0.630
Participants using antibiotics	1 (2.6%)	0 (0%)	0.324
Participants with emergency medical visits	0 (0%)	0 (0%)	1.000
Follow-up days 16 to 30 (reported at Visit 3)
Presence of fever (> 37.5 °C)	7 (18.4%)	4 (10.8%)	0.355
Presence of pain/discomfort (FLACC > 3)	7 (18.4%)	5 (13.5%)	0.565
Presence of cough	2 (5.3%)	2 (5.4%)	0.978
Presence of nasal congestion	7 (18.4%)	1 (2.7%)	0.029
Presence of rhinorrhea	8 (21.1%)	4 (10.8%)	0.230
Number of symptomsc	4 (2–5)	3 (1–4)	0.019
Participants with emergency medical visits	1 (2.6%)	0 (0%)	0.324
Follow-up days 31 to 60 (reported at Visit 4)
Presence of fever (> 37.5 °C)	8 (21.1%)	9 (24.3%)	0.783
Presence of pain/discomfort (FLACC > 3)	8 (21.1%)	9 (24.3%)	0.737
Presence of cough	10 (26.3%)	5 (13.5%)	0.169
Presence of nasal congestion	7 (18.4%)	9 (24.3%)	0.535
Presence of rhinorrhea	8 (21.1%)	10 (27.0%)	0.547
Number of symptomsd	5 (1–5)	4 (1–5)	0.215
Participants with emergency medical visits	0 (0%)	1 (2.7%)	0.311

^aData from patient diaries, except for the number of emergency medical visits, ^btrapezoidal sum of daily values from day 1 to day 14, ^camong symptom-positive participants at Visit 3 (probiotic n = 6, placebo n = 8), ^damong symptom-positive participants at Visit 4 (probiotic n = 13, placebo n = 11)

Exploratory outcomes

Post hoc analyses were conducted to assess the impact of the most frequently detected viruses (rhinovirus, coronaviruses, and RSV), age, sex, number of recent URIs (ranging from 0 to 3) and hours of fever before enrolment on co-primary endpoints. Probiotic intervention was still superior to placebo after adjusting for each of these factors, both for fever and for pain/discomfort (p < 0.01, Supplementary Table 1). Patients positive for coronaviruses (n = 21) had 0.7 more days of fever on average (p = 0.041). Conversely, patients positive for RSV (n = 19) had 0.8 more days of pain/discomfort on average (p = 0.004). Other variables assessed did not influence the duration of fever or pain. Days of use of antipyretic correlated with days of fever (rho = 0.56, p < 0.001) and of pain/discomfort (rho = 0.71, p < 0.001), while days of use of antihistamine correlated with days of rhinorrhea (rho = 0.85, p < 0.001). Days of fever and of pain/discomfort markedly correlated with one another (rho = 0.78, p < 0.001), while days of rhinorrhea and of congestion mildly correlated with one another (rho = 0.26, p = 0.022). Bivariate correlations can be found on Supplementary Table 2.

Safety

No severe AEs or hospitalizations were reported during the study. All reported AEs were of gastrointestinal type, mild and self-limited, and their numbers did not differ between study groups (Supplementary Table 3).

Discussion

In this randomized, placebo-controlled study, probiotic intervention was associated with a shorter duration of fever and pain/discomfort in children with an upper respiratory infection (URI) of likely viral origin with pharyngitis, as diagnosed by a pediatrician. On average, the duration of fever was shortened to 1.6 days vs. 2.7 in the placebo (i.e., 1.1 fewer days, a 40% reduction), and pain/discomfort was shortened to 3.7 days vs. 4.4 in the placebo (i.e., 0.7 fewer days, a 19% reduction). Results were robust to a variety of sensitivity analyses, including type of virus and participant’s age, sex, recent history of URIs and hours since the onset of fever. Particularly, intervention was efficacious regardless of whether patients were infected with coronaviruses or not, thus generalizing results from a previous randomized trial with the same probiotic in adult patients with Covid19 (original Wuhan strain), also with specialist and molecular diagnosis. In said study, median symptom duration was shortened to 13 days vs. 18 in the placebo (i.e., 5 fewer days, a 28% reduction) [14]. Our findings are also consistent with a recent study in pediatric URIs by Bettocchi et al., in which a probiotic mixture shortened the duration of fever from 5 days in the placebo group to 3 days in the treatment group, i.e., a 40% decrease [17]. Interestingly, the absolute reduction of the duration of symptoms seems to change depending on the natural duration of the URI, yet the percent of reduction seems more comparable across studies with probiotics. This shortening of symptoms is noteworthy as antipyretics do not seem to reduce the overall duration of the disease [18], suggesting specific probiotics could be an interesting add-on to current treatments for URIs. In this regard, they could help reduce the burden on parents kept from bringing children to kindergarten due to URI symptoms, particularly fever (normally the deciding factor), a possibility which deserves further studies.

Lactiplantibacillus plantarum (formerly Lactobacillus plantarum) and closely related Lactobacillaceae bacteria such as Pediococci are common endophytes, living in wild vegetables and frequently ingested by herbivores and omnivores [19]. As such, the immune systems of the latter evolved under repeated intestinal exposure to the former, regardless of said bacteria behaving as nomadic commensals or successfully engrafting in the intestine (i.e., becoming autochthonous). A seminal randomized trial in more than 4000 newborns, comparing L. plantarum strain ATCC202195 to placebo, found a significant reduction in lower respiratory tract infections and sepsis [20]. The formula used in the current study, containing L. plantarum (strains KABP022, KABP023, and KABP033) and P. acidilactici (strain KABP021), was previously tested in a randomized clinical trial in 300 adults with PCR-confirmed, symptomatic Covid19, resulting in reduced duration of symptoms and viral load [14]. However, as the study population ranged from 18 to 60 years old, evidence was missing in pediatric patients. Besides, in vitro studies found this same probiotic formula induced type-I interferon response in phagocytic cells [21]. Noteworthy, gut microbiota modulation of the host’s tonic type-I interferons seems to mediate systemic antiviral effects in mouse models [22, 23]. In our study, a reduction in the incidence of nasal congestion and in the total number of symptoms per patient was also observed at 2 weeks, but not 6 weeks, after completing the 15-day intervention (i.e., 30 days but not 60 days from baseline). Thus, immune-stimulating effects were seemingly maintained for a limited time once probiotic ingestion stopped, a finding deserving further investigations.

Previous randomized trials of probiotics for URIs in children have mostly relied on parent/caregiver diagnosis [9, 10]. This introduces a potential confounder, every patient in the study being diagnosed by a different, untrained person. In this regard, a recent large trial found no effect of L. rhamnosus DSM33153 (a.k.a. ATCC53103) on URIs diagnosed by a physician and confirmed by laboratory tests [24], despite the same strain showing positive effects in previous trials that relied on parental diagnosis and did not identify the etiologic agent [25, 26]. Therefore, proper diagnosis by a physician and the use of molecular methods seem relevant to increase the reliability of such clinical trials. Conversely, other recent probiotic studies relying on diagnosis by specialists alone [17] or specialist and molecular methods [27] found a significant effect against URIs, using different strains. This apparent disparity may be explained by the fact that probiotic effects are often strain-specific, particularly immune effects, and should not be extrapolated from one strain to another [6, 11–13]. In this regard, URIs in our study were diagnosed by pediatricians based on clear clinical criteria (pharyngitis accompanied by fever) and confirmed by molecular methods, thereby enhancing the robustness of our findings.

No single causative agent dominated our study population. Rhinovirus (36.0%), coronaviruses (28.0%, two-thirds being SARS-CoV2), and RSV (25.3%) were the most prevalent, in line with recent studies [3, 4]. Human metapneumovirus (14.6%) and adenoviruses (10.7%) were moderately prevalent, while influenza, parainfluenza, and human bocavirus had a prevalence < 10% each. A recent study in children found infections with rhinovirus, RSV, or metapneumovirus resulted in higher severity and/or a risk of hospitalization, underscoring the importance of effective therapies for these viruses [28]. Of note, causative agents are seldom reported in URI trials with probiotics [17, 25, 26], yet their diversity introduces a potential confounder unless properly accounted for: as the number of causative agents increases, so do the chances of a random imbalance between study groups (especially in moderately sized studies). Moreover, having several infectious agents in the same study entails the possibility of high efficacy against some viruses to mask low efficacy against others. Therefore, sensitivity analyses seem desirable to rule out such confounders. In this regard, our study is a first in probiotic research for URIs, to our knowledge.

All adverse events (AEs) in the study were mild and of gastrointestinal (GI) type. Previous studies have noted that GI symptoms are common among URI patients [29, 30]. In our study, several patients reported GI symptoms already at baseline. Unfortunately, such baseline info was not collected in a systematic, solicited manner, thus preventing a proper analysis of treatment-emergent AEs for all patients. Nevertheless, AE incidence did not differ between study groups, suggesting the probiotic intervention was well-tolerated.

Several study limitations should be pointed out. First, several URI symptoms were rated as present/absent (instead of a severity scale) and were not required as entry criteria, thus reducing our capacity to detect differences between groups. A better assessment of these secondary endpoints could have been obtained with a more precise symptoms scale, such as the WURSS. However, the pediatric version of this scale has only been validated for children 4 years and older [31]. Second, a causative virus could not be identified in 10 patients (13.3%) despite using a large panel identifying up to 19 viruses. These cases could be false negatives, viruses not in our panel, and/or bacterial infections resembling viral infections; yet, they were fairly balanced between study groups. Third, the prevalence of common viruses such as influenza or adenoviruses was low in our population. Nevertheless, the most prevalent viruses in our population coincided with a recent meta-analysis [4], and sensitivity analyses indicated efficacy in our study was not dependent on these prevalent viruses. Fourth, the recording of highest daily temperature is affected by the timing of the measurement. Also, the FLACC score measures both pain and distress or discomfort [32], is validated for the clinical setting, more than its use at home, and is dependent on subjective assessment. However, parents were trained to use the FLACC scale, and the concurrent effect in both co-primary endpoints (below the multiplicity-adjusted threshold of p = 0.025) helps strengthen our conclusions. Moreover, days of use of ibuprofen markedly correlated with both endpoints, adding internal validity. Fifth, study investigators inquired about treatment compliance, but no actual counting of remaining capsules was performed. Finally, no fecal microbiota analysis was included in our study. However, the previous randomized trial with the same probiotic found significant clinical effects despite no significant taxonomical changes in fecal microbiota [14], akin to other publications [33].

In conclusion, our study suggests the specific probiotic formula containing strains P. acidilactici KABP021 and L. plantarum KABP022, KABP023, and KABP033 helps shorten the duration of fever and pain/discomfort in children with URI with pharyngitis of diverse viral origin. This formula may also reduce the severity of relapses or new infections for a limited time after completing the intervention. Said results warrant replication in larger, multi-centric, randomized trials.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AE

Adverse event

AUC

Area under the curve

CFU

Colony-forming units

CONSORT

Consolidated Standards of Reporting Trials

FLACC

Face, Legs, Activity, Crying and Consolability

GLM

Generalized linear model

RSV

Respiratory syncytial virus

SPSS

Statistical Package for the Social Sciences

URI

Upper respiratory infection

Authors' contributions

DA-P, JE-M and MR-P conceptualized and designed the study. DA-P and MGM-F recruited and treated patients, and collected data. JE-M and DA-P analyzed the data. JE-M and JVC drafted the initial manuscript. DA-P and MR-P critically reviewed the manuscript. All authors approved the final manuscript as submitted.

Funding

This work was fully sponsored by AB-Biotics, S.A. (Kaneka Corp.).

Data availability

All data used in the current publication are available (duly anonymized) from the corresponding author upon reasonable request.

Declarations

Ethics approval

The study was conducted in accordance with the Declaration of Helsinki, and approval was obtained from Hospital Medica Sur's research ethics boards.

Consent to participate

Parents of all participants provided written informed consent.

Competing interests

JE-M, JVC and MR-P are full-time employees of AB-Biotics SA (Kaneka Group), the company that funded the study. DA-P and MGM-F were funded by AB-Biotics for study procedures and declare no additional competing interests for this work.