Comparative efficacy of probiotic mixture Bifidobacterium longum KABP042 plus Pediococcus pentosaceus KABP041 vs. Limosilactobacillus reuteri DSM17938 in the management of infant colic: a randomized clinical trial

Abstract

We aimed to compare the efficacy of Bifidobacterium longum KABP042 + Pediococcus pentosaceus KABP041 (BL + PP) vs. Limosilactobacillus reuteri DSM17938 (LR) in alleviating the symptoms of infant colic, as commercially available formulations. A randomized, multicenter, parallel, single-blind (investigator) trial was conducted in 112 colicky infants diagnosed as per Rome IV criteria and randomly allocated to receive BL + PP orally (10⁹ colony-forming units [CFU]/day, n = 55) or LR (10⁸ CFU/day, n = 57) for 21 days. Primary study outcomes were percentage of responders (≥ 50% reduction in crying and fussing time from baseline, as reported by parents in a structured diary) and daily crying and fussing time (minutes/day) on days 7, 14, and 21 after randomization. Study groups were comparable at baseline. Responder rate was significantly higher in BP + PP group vs. LR group at days 7 (61.1% vs. 37.5%, p = 0.013) and 14 (84.6% vs. 59.3%, p = 0.004). Crying and fussing time (median [IQR]) became significantly lower in BL + PP group vs. LR group on day 7 (119 [60–210] vs. 180 [110–270]; p = 0.028), day 14 (60.0 [30–105] vs. 120 [60–180]; p = 0.017), and day 21 (29 [0–85] vs. 67 [30–165]; p = 0.011). No significant differences were found in the number of adverse events between the groups.

Conclusion: The specific formulation of B. longum KABP042 and P. pentosaceus KABP041 achieved a higher response rate and a larger reduction in crying and fussing time in colicky infants. Both probiotic interventions were well tolerated.

Trial registration: The study was retrospectively registered as NCT05271747 on February 28th, 2022.

What is Known: • L. reuteri DSM17938 (LR) is the most researched probiotic strain for infant colic against placebo in randomized, controlled clinical trials, and is recommended in various guidelines. A novel probiotic combining strains B. longum KABP042 and P. pentosaceus KABP041 (BL + PP) has also demonstrated efficacy in infant colic against placebo.

What is New: • This randomized study provides the first direct comparison of two probiotics for infant colic. BL + PP seems to be superior to LR in reducing crying time.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00431-024-05806-x.

Introduction

According to Rome IV guidelines, functional gastrointestinal disorders (FGIDs) are disorders of gut-brain interaction, reflecting a multifactorial pathophysiology likely arising from maladaptive responses to internal and external stimuli [1, 2]. Infant colic is a paroxysmal, inconsolable crying or fussing with no apparent cause, occurring in infants up to 5 months of age [2] and affecting around 20% of infants [1]. This FGID affects the quality of life of the infant and caregivers, increases the risk of postpartum depression, and can lead to premature cessation of breastfeeding and shaken baby syndrome [3]. Furthermore, although usually self-conclusive, infant colic might be associated to the onset of other FGIDs (e.g., irritable bowel syndrome) and other conditions (e.g., migraines, behavioral problems) later in life [4].

Although the etiology of infant colic is not completely understood, increased visceral hypersensitivity, gut inflammation and altered gut microbiota seem to be involved [5–7]. Notably, differences have been found in the intestinal microbiota of colicky vs. non-colicky infants: lower levels of lactobacilli and bifidobacteria [8] and higher levels of opportunistic pathogens [5, 9].

Probiotics are “live microorganisms that, when administered in adequate amounts, confer health benefits on the host” [10, 11]. High-level evidence suggests that probiotics are the most effective treatment for infant colic, especially in breastfed babies [12]. However, probiotic benefits are strain-specific [11] and efficacy in colic relief varies widely depending on the strain [13]. Moreover, different clinical trial designs limit comparison between strains. Nevertheless, Limosilactobacillus reuteri DSM17938 (formerly Lactobacillus reuteri) is the most studied probiotic for infant colic in placebo-controlled randomized trials, particularly in breastfed babies [14–17]. As such, it is recommended in the World Gastroenterology Organization and ESPGHAN guidelines [18, 19]. The combination of strains Bifidobacterium longum KABP042 and Pediococcus pentosaceus KABP041 (AB-Kolicare®) has also shown to be effective in the management of infant colic, by reducing crying and fussing time in randomized, double-blind, placebo-controlled clinical trials [20, 21]. In this study, we aimed to compare the efficacy of these two probiotic formulations in ameliorating infant colic.

Methods

Study Design

This was a randomized, multicenter, single-blind (investigator) trial with two parallel arms. We included participants recruited in six clinical sites: (i) the PROBIAP network formed by eight public primary care centers in Murcia (Spain); (ii) Hospital HM Nens in Barcelona (Spain); (iii) Hospital HM Puerta del Sur in Madrid (Spain); (iv and v) Clinica Universidad de Navarra (Pamplona and Madrid sites); and (vi) Hospital Medica Sur (Mexico City, Mexico). The study was conducted in accordance with the Declaration of Helsinki, and the approval was obtained from the research ethics boards responsible for the different research sites (Hospital Universitario Sta Maria del Rosell, Hospitales HM, Clinica Universidad de Navarra and Hospital Medica Sur). All parents provided written informed consent before participating in this study. The study was registered at ClinicalTrials.gov (NCT05271747) and is reported according to Consolidated Standards of Reporting Trials (CONSORT) [22] (Supplementary Fig. 1).

Otherwise healthy infants aged 2 to 12 weeks and diagnosed with infant colic according to the Rome IV criteria for clinical research [2] were recruited. Accordingly, parents had to report duration of crying or fussing for ≥ 3 h per day for ≥ 3 days per week during a face-to-face screening interview conducted by the investigator. Confirmation of the duration of ≥ 3 h of crying and fussing was obtained in a 48-h prospective behavioral diary. Additional inclusion criteria included gestational age ≥ 37 weeks and birth weight > 2100 g. Both exclusive breastfeeding and exclusive or predominant formula feeding (maximum of one daily breast milk intake) were allowed.

Infants meeting the following criteria were excluded: (i) receiving solid foods; (ii) receiving extensively hydrolyzed formula; (iii) having received (or whose mothers received, in those who were breastfed) antibiotics or probiotics (including infant formula with probiotics) within 2 weeks before inclusion in the study; (iv) receiving specific medications used for the treatment of FGIDs such as antacids (proton pump inhibitors or H2 blockers), polyethylene glycol, lactulose, magnesia, lactase, or simethicone, undergoing therapies related to acupuncture, homeopathy, medicinal herbs, or taking anti-inflammatory or antispasmodic drugs; and (v) parents could not appropriately comply with the study procedures.

Interventions

Participants were randomized 1:1 to receive orally 5 drops a day, for 21 days, of one of the two probiotic products: L. reuteri DSM17938 (commercial Reuteri® drops, single-strain probiotic suspension in sunflower oil plus medium-chain triglyceride oil, 10⁸ colony-forming units (CFU) per 5 drops, produced by Biogaia® (Sweden); “LR” group) or the combination of B. longum KABP042 (CECT7894) and P. pentosaceus KABP041 (CECT8830) (AB-Kolicare®, a two-strain probiotic suspension in sunflower oil, each strain loaded in equal amount and totaling 10⁹ CFU per 5 drops, produced by AB-Biotics®, S.A. (Spain); “BL + PP” group). Stability of study products throughout the intervention period was confirmed by CFU plate counting. The randomization list was generated by independent personnel using the Epidat 3.1 software, with a block size of 4 and stratification by feeding mode. Study investigators recruited and allocated patients and performed the study visits. To secure blinding of the investigators, probiotics (in their commercial formulation) were supplied in undistinguishable, closed opaque bags. Data analysts were also blinded.

At visit 0 (screening), parents of infants who initially complied with inclusion and exclusion criteria were informed about the study. Those who agreed to participate and signed the informed consent form were given a 48-h diary to confirm infant colic under Rome IV criteria and serve as the baseline (Fig. 1). During visit 1, eligible infants were randomized into one of the study groups. Parents received the allocated probiotic product and instructions for its administration. They were also instructed to daily register in a structured diary: (i) crying and fussing time, using a chart modified from Barr et al. [23] and (ii) number of depositions and regurgitations. The diaries also contained, on days 1 and 21, the Hospital Anxiety and Depression Scales (HADS) questionnaire [24] and a 6-point Likert scale on sleep quality for each parent. During visit 2 (final), diaries were returned, and adverse events (AEs) and concomitant medication received during the intervention were assessed.

Fig. 1.

Study design. AE, adverse event; CFU, colony-forming units; HADS, Hospital Anxiety and Depression Scale

Outcomes

Two primary outcomes were assessed at 7, 14, and 21 days: (1) rate of responders, defined as those infants who reduced crying and fussing time ≥ 50% from baseline [14, 15], and (2) total crying and fussing time (in minutes per day).

Secondary study outcomes included the following: frequency of regurgitations and bowel movements (at 7, 14, and 21 days), parents’ anxiety and depression scores and quality of sleep (at day 1 and 21), and incidence of AEs during the intervention period.

Statistical analysis

Sixty infants per group were estimated to be required to detect a difference between the two active intervention groups of 12 min in the evolution of the child’s crying or irritability time for 24 h, with a power of 80% and a two-tailed significance level of 5% in a T-test. This calculation assumed a standard deviation (SD) of 23.5 min, which was approximated from the largest published randomized trials in colicky infants that included either probiotic in this study [15, 21] using the $SD=Interquartilrange÷1.35$ formula [25].

Statistical analysis was performed on all patients with data available at each analysis timepoint (full analysis set) by allocated group, using SAS (Statistical Analysis System) version 9.4. Categorical variables were compared using the chi-square test, whereas the Mann–Whitney (unpaired) and Wilcoxon (paired) tests were used for continuous variables, because of significant departures from normality. In exploratory analyses, response rates at 7, 14, and 21 days were adjusted by means of Mantel–Haenszel test for (i) feeding mode (breast or formula feeding); (ii) birth type (vaginal or C-section); (iii) the presence of anxiety among caregivers on day 1 (HADS-A ≥ 11 [24]); (iv) the presence of depression among caregivers on day 1 (HADS-D ≥ 11 [24]); and (v) infant age at screening (less than 6 weeks vs. 6 weeks and older). The presence of interactions between intervention and each of said factors was assessed by means of Breslow-Day test (Tarone’s version). All comparisons used a bilateral significance threshold of 0.05.

Results

Subjects

The study was initiated on March 1st, 2019, and completed on December 30th, 2023. We randomized 112 infants, 57 allocated to LR group, and 55 to BL + PP group. During follow-up, 2 infants from LR group and 2 infants from BL + PP group dropped out, and 108 infants completed the study (Fig. 2). Baseline demographic and clinical characteristics were similar between groups, including crying and fussing time (Table 1).

Fig. 2.

Consort flow diagram of the subjects during the study. AE, adverse event; BL + PP, Bifidobacterium longum + Pediococcus pentosaceus; LR, Limosilactobacillus reuteri; n, number of subjects

Table 1.

Baseline characteristics of the study population

Baseline characteristics	L. reuteri	B. longum + P. pentosaceus
Study population, n	57	55
Spain [n (%)]	24 (42.1)	22 (40.0)
Mexico [n (%)]	33 (57.9)	33 (60.0)
Age [weeks; mean (SD)]	5.2 (2.3)	5.3 (2.2)
Sex [women; n (%))	25 (43.9)	24 (44.4)
Weight [g; mean (SD)]	4211.8 (880.8)	4276.9 (926.6)
Height [cm; mean (SD)]	53.7 (3.0)	53.6 (3.2)
Birth weight [g; mean (SD)]	3045.5 (404.5)	3087.7 (474.5)
Gestational age [weeks; mean (SD)]	38.9 (1.2)	38.8 (1.1)
C-section [n (%)]	17 (29.8)	17 (31.5)
Exclusive breastfeeding [n (%)]	18 (31.6)	18 (33.3)
Perinatal antibiotics [yes; n (%)]	0 (0)	2 (9.5)
Mother smokes [yes; n (%)]	2 (3.6)	1 (1.9)
Family history of allergy [yes; n (%)]	5 (8.8)	4 (7.4)
Vitamin D supplementation [yes; n (%)]	16 (28.1)	15 (27.8)
Crying and fussing [min; median (IQR)]	270.0 (240.0–400.0)	270.0 (235.0–358.0)

No significant differences were found between groups for each characteristic (p < 0.05)

Primary outcomes

As shown in Fig. 3, the percentage of infants that reduced their crying and fussing time by ≥ 50% from baseline (responder rate) was significantly higher in the BL + PP group when compared with LR group on days 7 (61.1% vs. 37.5%, RD = 23.6%; p = 0.013) and 14 (84.6% vs. 59.3%, RD = 25.3%; p = 0.004). A trend was observed on day 21 (92.2% vs. 80.4%, RD = 11.8%; p = 0.084).

Fig. 3.

Response rate (%) for L. reuteri and B. longum + P. pentosaceus groups at each time point. *p < 0.05, **p < 0.01, ^#p < 0.10

Median daily crying and fussing time (minutes) decreased across the intervention period. Compared with baseline, reductions at 7, 14, and 21 days were significant within each study (all p < 0.001, both groups). Notably, infants in the BL + PP group had significantly lower total crying and fussing time than those in LR group on days 7 (p = 0.028), 14 (p = 0.017), and 21 (p = 0.011) (Table 2).

Table 2.

Primary outcome: crying and fussing time (minutes/day)

	L. reuteri	B. longum + P. pentosaceus	p-value
Day 7	180.0 (110.0–270.0)	119.0 (60.0–210.0)	0.028*
Day 14	120.0 (60.0–180.0)	60.0 (30.0–105.0)	0.017*
Day 21	67.0 (30.0–165.0)	29.0 (0.0–85.0)	0.011*

Data presented as median (IQR). *Mann–Whitney p < 0.05

Secondary outcomes

Number of regurgitations and bowel movements did not differ between study groups at day 7, 14, or 21 (Table 3). Likewise, similar scores were found for parental anxiety, depression and sleep quality between groups (Supplementary Table 1).

Table 3.

Secondary outcomes: number of regurgitations and bowel movements

Number of regurgitations	L. reuteri	B. longum + P. pentosaceus	p-value
Day 7	2.0 (0.0–2.0)	2.0 (0.5–2.0)	0.866
Day 14	1.0 (0.0–2.0)	1.0 (0.0–2.0)	0.523
Day 21	0.0 (0.0–2.0)	0.0 (0.0–1.0)	0.565
Number of bowel movements	L. reuteri	B. longum + P. pentosaceus	p-value
Day 7	2.0 (1.0–4.0)	2.0 (1.0–3.0)	0.611
Day 14	2.0 (1.0–4.0)	2.0 (2.0–3.0)	0.251
Day 21	2.0 (1.0–4.0)	2.0 (1.0–3.0)	0.205

Data presented as median (IQR). No significant differences were found between groups for each characteristic (Mann–Whitney p < 0.05)

Exploratory analyses

The significant difference between study interventions on responder rates at days 7 and 14 (p < 0.05 and p < 0.01, respectively) and the statistical trend on day 21 (p < 0.10) were maintained when adjusting for either feeding mode or delivery type. The significant differences on days 7 and 14 were also maintained when adjusting for the presence of parental anxiety, the presence of parental depression, and for age of the infant. No significant interactions of intervention by any of these factors were detected (Supplementary Table 2).

Safety and tolerability

Eight AEs were reported during the study (Table 4): four (7.4%) in LR group and four (7.8%) in BL + PP group, and no subject reported more than one AE. Seven adverse events were classified as mild, one of them resulting in the infant’s withdrawal from the study (a case of constipation in the BL + PP group). Only one AE was of moderate intensity and occurred in the LR group (gastroesophageal reflux), resulting in the infant's withdrawal from the study. All AEs not resulting in withdrawal from the study resolved during the intervention period, except for the single case of intolerance to cow’s milk.

Table 4.

Number and type of adverse events (AEs) reported during the intervention. Each of the reported AEs occurred in a different subject (i.e., no subject reported more than one AE)

Intensity	Type of AE	L. reuteri (n = 57)	B. longum + P. pentosaceus (n = 55)
Mild	Bronchitis	1
	Common cold		1
	Constipation		1*
	Intolerance to cow’s milk		1
	Oral thrush (candidiasis)	1	1
	Regurgitation/vomiting	1
Moderate	Gastroesophageal reflux	1*
Total		4	4

^*Resulted in patient withdrawal from the study

Discussion

In the present study, the efficacy of a commercial formulation containing strains B. longum KABP042 and P. pentosaceus KABP041 (BL + PP) in reducing infant colic symptoms was compared with that of L. reuteri DSM17938 (LR). Meta-analyses and reviews evidence the efficacy of L. reuteri DSM17938 in the management of infant colic, particularly in breastfed infants, and several clinical guidelines support its use [3, 18, 19, 26]. This motivated us to use L. reuteri DSM17938 as comparator probiotic strain in our study. To the best of our knowledge, this is the first comparative randomized controlled study on infant colic using such active comparator.

The results showed that the responder rate was significantly higher on days 7 (absolute difference of 23.6%) and 14 (absolute difference of 25.3%) in the BL + PP group but displayed only a statistical trend on day 21 (absolute difference of 11.8%). In line with these results, the formulation containing BL + PP seemed more effective in decreasing crying and fussing time at all study timepoints (days 7, 14, and 21) than the formulation containing LR. Largest differences were observed on study days 7 and 14, where median daily crying and fussing time was 60 min lower than in the LR group (both timepoints), while it was only 38 min lower on day 21. Importantly, this study also confirmed that oral intake of these probiotics for 21 days was well tolerated.

Previous placebo-controlled studies in breastfed colicky babies have reported variable response rates for L. reuteri DSM17938 on day 7 (range 10–80%, average 30%), which became more homogenous on day 14 (range 46–96%, average 67%), and especially on day 21 (range 71–98%, average 89%) [14–17]. In our study, response rates in the LR group are similar to the average of published values at each timepoint, thus supporting external validity of our study. However, evidence of L. reuteri DSM17938 in formula-fed colicky infants is scarce. One study evaluated its efficacy in a population of breastfed and formula-fed infants diagnosed with colic compared with placebo, showing no benefit in either subpopulation [27]. These results are surprising because the lack of effect in the breastfed subpopulation was at odds with five other randomized, placebo-controlled trials [14–17, 28], as well an individual patient meta-analysis that included patients from this diverging study [3]. Moreover, a preventive study investigating the role of L. reuteri DSM17938 in a large population of both breastfed and formula-fed newborn infants found a significantly lower crying time in comparison with placebo, both after 1 and 3 months [29]. Finally, a recent multi-country, cross-sectional, observational study in a large population of healthy infants indicated that those receiving formula supplemented with L. reuteri DSM17938 had less physician-confirmed colic than infants receiving standard formula [30].

Response rates in BL + PP group are close to those reported in the previous study, conducted in China, especially on days 14 and 21 [21]. The lower response rate observed in our study on day 7 (61% vs. 83%) could be explained by the higher median crying and fussing time at baseline in our population. Such variation is not unexpected, since an even larger variability in the response rate on day 7 is also observed among published clinical trials with L. reuteri DSM17938.

Feeding mode (breastfed vs. formula-fed) and delivery type (vaginal vs. C-section) have been repeatedly found to significantly influence infant’s gut microbiome composition [31–33]. Therefore, these factors could differentially affect the efficacy of probiotics belonging to different genera such as those used in our study (Bifidobacterium and Pediococcus vs. Limosilactobacillus). However, adjusting by feeding mode or birth type had a negligible impact on the difference in response rate between the two probiotic interventions in our study. Furthermore, testing for the homogeneity of the odds ratios suggests that no interactions (birth type by intervention or feeding mode by intervention) occurred in our study population. Notably, such tests for interactions tend to have lower statistical power than tests for main effects [34]. Nevertheless, all observed interaction p-values are above 0.10, and all but one are above 0.20, so even using more lenient p-value cutoffs would still result in rejecting the hypothesis of an interaction. Therefore, our study suggests that a similar improvement in efficacy occurred with the BL + PP formula compared with the LR formula independently of feeding mode and delivery type.

Other than feeding mode and delivery type, additional factors can influence crying time and could act as confounders in a clinical study. Noteworthy, our use of a randomized design coupled with a sample size of more than 100 infants should minimize the risk of such confounding. Nevertheless, it may be worth taking a deeper look at some of these factors. For instance, anxiety and depression can influence the caregiver’s ability to soothe the infant, which in turn could potentiate said conditions in a downward spiral [2]. Age of the infants also influences crying time, with a peak at around 5–6 weeks [35]. However, these factors were well-balanced between study groups. Moreover, as in the case of feeding mode or delivery type, adjusting for the presence of anxiety or depression among parents (based on HADS scale cutoffs [24]) or adjusting for age (less than 6 weeks vs. 6 weeks and older) had a negligible impact on the difference in response rate between the probiotic interventions. Also, no treatment by factor interactions were detected at p-values below 0.10. Besides, concomitant presence of other diseases or prematurity could also influence infant’s physiology and thus crying time, but these factors were excluded by entry criteria and thus could not confound our results. Similarly, family history of allergy has sometimes been proposed as a predisposing factor for colic, but only 8% of participants fulfilled this criterion (5 in LR group and 4 in BL + PP group). Likewise, perinatal antibiotic use could influence microbiota, and thus colic, but only 2 infants (BL + PP group) fulfilled this criterion. These low numbers preclude their statistical analysis but also rule out any relevant confounding effect. On a final note, as a limitation worth mentioning, we did not collect data on whether the mothers were primiparous or not, a factor that could also have influenced the perception of crying time and the ability to soothe the infant. Therefore, we were not able to directly assess its impact in our study.

Differences in the effect size between the two probiotic interventions could be attributed to their different dosages and/or strain-specific mechanisms of action. The studied probiotics are known to inhibit the growth of bacteria associated with gut dysbiosis in colicky babies by different means: L. reuteri DSM17938 produces the natural antibiotic reuterin [36], whereas B. longum KABP042 and P. pentosaceus KABP041 synthesize antimicrobial peptides such as bisin and an unidentified bacteriocin [37]. Also, P. pentosaceus KABP041 has been shown to ameliorate intestinal inflammation by modulating Treg cells in the colon and by increasing IL-10 in serum in mice [38], while L. reuteri DSM17938 has been shown to modulate Tregs without changes in IL-10 expression in colicky infants [39], thus suggesting different albeit similar immunomodulatory mechanisms. In addition, studies in different epithelial barrier models have shown that both probiotic formulations increase the expression of zonula occludens-1 [37, 40], with B. longum and P. pentosaceus having a synergistic effect. However, a recent study found that L. reuteri DSM17938 and P. pentosaceus KABP041 did not produce polyphosphate, a molecule involved in gut barrier protection, whereas B. longum KABP042 synthesized the highest amounts among all investigated strains, including other bifidobacteria [41]. This last study highlights strain-dependent mechanisms of action that may explain differences in efficacy. Nevertheless, more mechanistic research is needed to understand how the studied probiotic formulas exert their beneficial effects.

The strengths of our comparative study include (1) adequate sample size for the primary outcomes; (2) efficacy investigated in two different countries with a predominance of different ethnicities; (3) inclusion of both formula and breast-fed infants; and (4) exclusion of infants receiving other colic-related treatments (e.g., extensively hydrolyzed formula, simethicone, lactase, or herbal treatments), antibiotics, or other probiotics.

This study also has several limitations. First, no placebo arm was used. Notably, a marked placebo effect is observed in similar studies due to the natural course of infant colic. However, both probiotics have demonstrated their efficacy vs. placebo in previous studies, especially L. reuteri DSM17938, while both lacked a comparative study against another probiotic. Second, blinding of the participants was not possible because of the use of commercial formats of the study products. Of note, repackaging the products in a common dropper could likely result in a change of the volume of product contained in five drops, thus introducing a potential confounder to previous studies. Importantly, our single-blind design (for the investigator) introduced a potential source of bias, especially since study outcomes relied on diaries filled by unblinded parents/caregivers. L. reuteri DSM17938 (previously known as ATCC55730) was the first probiotic to be clinically studied for infant colic, already in 2007 [42]. This strain was already marketed in the countries where the study took place years before the arrival of strains B. longum KABP042 and P. pentosaceus KABP041 (also known as CECT7894 and CECT8330). Accordingly, one could hypothesize bias to favor L. reuteri DSM17938 since it was more widely known for infant colic, yet the study showed a higher effect by B. longum KABP042 + P. pentosaceus KABP041. Also, as discussed in previous paragraphs, response rates in our study approached those observed in double-blinded trials for the same products, suggesting bias was possibly mitigated. A reason could be both study arms receiving active treatments, of different yet similar format (small droppers) and identical posology (5 drops, once daily). Nevertheless, reliance on unblinded parents/caregivers remains a relevant limitation of our study. Third, compliance with study products could not be objectively assessed. Weighting study bottles before and after dispensing the treatment has been proposed as a method to follow compliance, but this method can sometimes produce high variability [43], especially when the volumes administered daily are very small compared with the bottle’s weight, as in our study.

Future clinical studies should aim to replicate our findings in infant colic. Importantly, assessment of different doses per probiotic formula would help clarify whether bacterial concentration plays a role in the seemingly higher efficacy of the B. longum KABP042 + P. pentosaceus KABP041 commercial product, compared with the L. reuteri DSM17938 one. Moreover, similar head-to-head studies could also be performed in functional abdominal pain, a condition where L. reuteri DSM17938 has previously shown efficacy against placebo [44, 45]. On the other hand, conducting head-to-head studies in animal models would be highly desirable, given the ethical limitations in obtaining multiple blood samples in healthy infants, let alone gut tissue biopsies. Such studies could help identify similarities and differences in the mechanism(s) of action between the probiotic formulas studied herein. In turn, this better knowledge of the mechanism(s) of action could explain differences in clinical response rate between probiotics. If the mechanism(s) of action were found to differ in vivo, this could help define patient populations more likely to benefit from each probiotic formula (i.e., a probiotic with a lower average response rate in the general population may have a higher response in a specific subpopulation). Of note, although animal models of infantile colic do not exist strictly speaking, maternal separation stress experiments have recently produced interesting results for L. reuteri DSM17938 and could be a useful research option [46].

Conclusions

In our study population, the commercial formula combining the probiotic strains B. longum KABP042 (CECT7894) and P. pentosaceus KABP041 (CECT8330) was superior to the formula containing L. reuteri DSM17938 in reducing colic symptoms. This effect seemed to be independent of the feeding and delivery modes. Both formulas were equally well tolerated. Additional randomized studies are required to replicate these findings.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AE

Adverse event

BL

Bifidobacterium longum

CFU

Colony-forming units

CONSORT

Consolidated Standards of Reporting Trials

FGID

Functional Gastrointestinal disease

HADS

Hospital Anxiety and Depression Scale

IL-10

Interleukin 10

IQR

Inter-quartile range

LR

Limosilactobacillus reuteri

PP

Pediococcus pentosaceus

RD

Rate difference

SAS

Statistical Analysis System

SD

Standard deviation

Author contribution

JV-M and JE-M designed the study. JV-M, JMM-V, DA-P, GC-R, MS-L, ACL and MGM-F recruited and assessed patients. JV-M, JMM-V, DA-P, GC-R, MS-L and ACL critically reviewed the manuscript.

Funding

The study was funded by AB-Biotics SA (Kaneka Group).

Data availability

The anonymized datasets used in the current publication are available from the corresponding author on reasonable request.

Declarations

Ethics approval

The study was conducted in accordance with the Declaration of Helsinki, and approval was obtained from the research ethics boards responsible for the different research sites (Hospital Universitario Sta Maria del Rosell, Hospitales HM, Clinica Universidad de Navarra and Hospital Medica Sur). Protocol was registered as NCT05271747.

Consent to participate

All parents provided written informed consent before participating in this study.

Competing interests

JE-M is a full-time employee of AB-Biotics SA (Kaneka Group), the company that funded the study. All other authors declare no competing interests for this work.