Effects of Bifidobacterium animalis Subsp. Lactis CP‐9 on Gut Microbiota and Immune Functions in Healthy Infants Authors: Liujing Zhang, Xueke Song, Keming Zhang, Xiaoshuang Xu, Xinmei Guo, Wei Li, Xiaoling Zeng, Yingguang Xie, Liwei Nie, Li Zhao, Xuelu Dai, Jing Sun, Jiai Yan, Yiran Liu, Feng Zhang, Zhiping Sun Published: 2026 ABSTRACT ABSTRACTInfancy represents a critical period for growth and development, during which the gut microbiota and its metabolites play essential roles in nutrient absorption and maintaining health. Therefore, it is of great significance to evaluate the safety and gut microbiota‐modulating effects of specific probiotics in healthy infants. This study aimed to investigate the effect of Bifidobacterium animalis subsp. lactis CP‐9 strain (CP‐9) on healthy infants aged 6 to 36 months and its relationship with growth patterns and multidimensional indicators. In this randomized, double‐blind, placebo‐controlled trial, healthy infants were randomly assigned to receive CP‐9 or placebo for 3 months. Safety, growth parameters, complete blood counts (CBC), immunological markers and faecal microbiota were assessed before and after the intervention. All infants exhibited age‐appropriate growth in height and head circumference during the study. The CP‐9 strain significantly increased gut Bifidobacterium abundance and regulated the microbiota‐mediated polyunsaturated fatty acids biosynthesis pathway. Haematological and systemic inflammatory markers remained within normal ranges and no treatment‐related adverse changes were detected during the 3‐month intervention period. Moreover, CP‐9 improved stool characteristics and gastrointestinal tolerability, with no increase in the incidence of respiratory allergies, skin allergies, respiratory infections, or fever, which further confirmed its safety. These findings suggested that CP‐9 supplementation resulted in significant enrichment of Bifidobacterium, a beneficial gut bacterium and had no detectable adverse effects on the measured physiological parameters. Consequently, the intervention raised no detectable safety concerns in healthy infants under the conditions of this study. Given this, CP‐9 is a probiotic that does not show any obvious short‐term safety signals for infants aged 6 to 36 months. Abbreviations absolute basophilic count absolute eosinophil count lymphocyte count absolute monocyte count absolute neutrophil count basophilic granulocyte percentage C‐reactive protein eosinophil count haematocrit haemoglobin interleukin lymphocyte percentage mean corpuscular haemoglobin mean corpuscular haemoglobin concentration mean corpuscular volume monocyte percentage mean platelets volume neutrophil percentage plateletcrit platelet distribution width red cell count red cell distribution width coefficient variation red cell distribution width standard deviation tumour necrosis factor white blood cell count Introduction Healthy infants and toddlers, when adequately nourished, exhibit robust endogenous linear growth and neurodevelopmental momentum (de Onis, Garza, et al. 2004; de Onis, Onyango, et al. 2004). Against this background, the objective of nutritional intervention is not always to ‘accelerate’ but rather to ‘safeguard’ and ‘optimize’ this natural process while countering potential risks. The period following the ‘first 1000 days of life’ (i.e., late infancy and early childhood) remains a critical window for growth and development (linear growth—height, neurodevelopment—head circumference), during which malnutrition and stunting continue to be global challenges (Organ 2021). Older infants and toddlers exactly fall within the ‘transitional phase’ and early ‘stabilization phase’ as defined by the TEDDY study (Stewart et al. 2018), representing a critical period where the microbiota experiences significant changes and begins to stabilize. Interventions during this period may produce significant effects. From infancy to adulthood, the microbiota plays a fundamental role. The microbiota established in early life, especially the gut microbiota, acts as a key regulator of health and is crucial for the growth, development and nutrient absorption of infants and toddlers (Yao et al. 2021). The presence of gut microbes is essential for the normal structure, function and transcriptional development of the intestinal barrier during early life (Blanton et al. 2016; Jones et al. 2023). This emphasizes that the gut microbiota constitutes part of the ‘internal environmen’ supporting healthy growth. During the first year after birth, the genus Bifidobacterium consistently dominates the intestinal microbiota (Yatsunenko et al. 2012; Stewart et al. 2018). This genus is considered the cornerstone of the healthy infant's gut, facilitating nutrient metabolism, producing beneficial metabolites such as short‐chain fatty acids (SCFAs) (Fukuda et al. 2011) and promoting the maturation of the host immune system. A stable gut ecosystem dominated by Bifidobacterium in early life is regarded as a biomarker for healthy growth trajectories. Therefore, establishing a stable and beneficial gut microbiota is crucial for realising children's inherent growth potential (Houghteling and Walker 2015; Robertson et al. 2019). Bifidobacterium animalis subsp. lactis CP‐9 (CP‐9) is a probiotic strain isolated from breast milk. Probiotics derived from breast milk exhibit high safety and physiological adaptability in children's nutritional intervention (Maldonado‐Lobón et al. 2015). Several studies have been conducted to confirm its beneficial effects on human health. CP‐9 has been confirmed to improve the gut microbiota of infants aged 0 to 6 months and is safe (Chen et al. 2023). Although CP‐9 has been utilized in humans for more than a decade and has been demonstrated to be safe for infants aged 0 to 6 months, further assessment is still required regarding its effects on and safety for the gut health of infants. Although the association between gut microbiota dysbiosis and childhood malnutrition has been firmly established, the potential of specific probiotic interventions to actively promote growth in otherwise healthy children remains a prominent area of research. Many current probiotic clinical trials aim to demonstrate improvements in disease conditions or growth retardation. Nevertheless, evaluating probiotics in healthy populations is valuable for assessing their ability to enrich beneficial genera, safety and beneficial interactions with host physiology. Even if a probiotic strain successfully colonizes the gut of healthy infants without changing macroscopic growth rates or the immune system, does the act of enriching the relative abundance of bacterial genera in itself possess ecological and potential health value. Despite certain research advancements, the mechanisms through which specific single‐strain probiotics simultaneously regulate the gut ecosystem and contribute to the maintenance of physical development and immune homeostasis in healthy infants (aged 6–36 months) remain unclear. Therefore, this study aims to systematically evaluate enrichment of the relative abundance of the Bifidobacterium in the gut, its regulatory effects on host nutritional metabolism and its safety via an age‐stratified intervention trial and multidimensional analysis. Based on this, we will further explore whether this probiotic intervention can effectively enrich beneficial bacteria and optimize the microbial community structure, evaluate its effects on systemic inflammatory markers and analyse its effects on core growth and developmental indicators. The study will simultaneously measure the composition of the microbial community, blood immune parameters and growth and development data and employ correlation analysis to uncover their intrinsic relationships. This approach aims to elucidate alternative pathways by which probiotics may exert effects without inducing significant systemic immune fluctuations, thereby providing new evidence‐based support for the scientific and precise application of this strain during critical developmental stages of infants. Materials and Methods Ethics Statement This study is a randomized, double‐blind, placebo‐controlled, multicentre clinical trial approved by the Ethics Committees of all participating centres: the Second Affiliated Hospital of Nanjing Medical University (approval no.: 2024‐KY‐159‐02), First Teaching Hospital of Tianjin University of Traditional Chinese Medicine (TYLL2024[K] No. 060), Jining No. 01 People's Hospital (2024‐Pharm‐Expr 015), Women and Children's Health Care Hospital of Linyi (KYL‐YXLL‐2024026), Hunan Provincial Maternal and Child Health Care Hospital (JS‐2024‐10‐017), Panjin Liaohe Gemstone Flower Hospital (2024YJZ(08)). This study has been registered at the National Medical Research Registration Information System (registration number: MR‐37‐25‐009256) and Chinese Clinical Trial Registry (ChiCTR2600125253). All human subject procedures were performed in accordance with the Guidelines of the Declaration of Helsinki. All participants provided written informed consent. Protocol Sample Size Calculation The primary endpoint was defined as the change in the relative abundance of gut Bifidobacterium from baseline to the end of the 3‐month intervention. Because no previous studies reported the exact mean difference and standard deviation for this endpoint in the target population of infants aged 6–36 months, we adopted a standardized effect size approach. According to Cohen's widely accepted criteria (Cohen 1988), effect sizes of 0.2, 0.5 and 0.8 represent small, medium and large effects, respectively. A methodological guideline supports the use of standardized effect size estimates when directly applicable prior data are lacking (Zhang and Hartmann 2023). Based on these considerations, a medium effect size (Cohen's d = 0.5) was conservatively assumed for the primary endpoint. Using a two‐sided significance level of α = 0.05 and a desired power of 90%, the required sample size was calculated as 86 infants per group (intervention and control). After accounting for an anticipated 20% dropout rate, the minimum required sample size was determined to be 108 infants per group, that is 216 infants in total. To further safeguard against a potential higher dropout rate in this infant population and to preserve adequate power for age‐stratified subgroup analyses, we increased the final recruitment target to 120 infants per group, resulting in a total of 240 infants. A stratified block randomization method was utilized to assign participants to the experimental or control group at a 1:1 ratio. Stratification was performed by age group (6–12 months and 12–36 months). In each age stratum, 120 infants were enrolled, with 60 allocated to the intervention group and 60 to the control group. Thus, across the two strata, the intervention group comprised 120 infants (60 per stratum) and the control group comprised 120 infants (60 per stratum). An independent biostatistician generated random numbers using SAS software and block randomization was applied within each age stratum. Sample Collection and Processing Inclusion and Exclusion Criteria for Recruitment of Participants (1) Inclusion criteria. Younger infants aged 6–12 months; older infants aged 12–36 months; gestational age 37–42 weeks; birth weight 2500–4000 g; mothers without diagnosed metabolic disorders (e.g., diabetes) or infectious diseases (e.g., hepatitis B, HIV); good compliance with clinical trial requirements; informed consent obtained from the legal guardian. (2) Exclusion criteria. Infants with birth defects or congenital anomalies; infants with severe malnutrition or growth retardation; infants who have used antibiotics, antidiarrheal medications, probiotics, or probiotic‐containing products within 4 weeks prior to randomization; infants with severe diarrhoea or more serious complications (e.g., ulcerative or bloody stools); infants with severe cardiac disease, hepatic, renal diseases, hematologic or metabolic/genetic disorders; infants with organic gastrointestinal diseases; infants with allergies or intolerance to any components of the experimental probiotic product; infants with food allergies; parents with a history of alcoholism or substance abuse; infants who participated in other clinical trials within 6 months prior to randomization; or any other condition deemed by the investigator as unsuitable for participation in this clinical trial. (3) Exit criteria. Parents or guardians request withdrawal of the infant from the clinical trial; parents or guardians withdraw their informed consent; serious non‐compliance with the trial protocol; the infant uses other probiotics or receives antibiotic treatment during the trial period; the physician determines that the participant should be discontinued from the trial. Grouping and Intake Method of Intervention Substances This study was designed as a multicentre, randomized, controlled, double‐blind trial. A total of 240 subjects were enrolled, including 120 younger infants aged 6 to 12 months and 120 older infants aged 12 to 36 months. Informed consent forms were signed by guardians of all subject. Subjects were randomly assigned in a 1:1 ratio to either the intervention group or the control group, receiving probiotics (CP‐9, i.e., Active Probiotic Powder, provided by Jinqiao Biotechnology Co. Ltd.) or placebo (maltodextrin) intervention, respectively. The younger infants were randomly divided into two groups: the CP‐9 group (Group YI + CP‐9) and placebo group (Group YI). The older infants were randomly divided into two groups: the CP‐9 group (Group OI + CP‐9) and placebo group (Group OI). The total daily intake of CP‐9 should be greater than or equal to 5.0 × 109 CFU. Study products and control products were uniformly stored in the central warehouse, sealed and placed in a cool, dry place. Investigators distributed samples according to subject ID and the study protocol. After receiving the study product, subjects consumed the test product (1 g/pack) or control product (1 g/pack) orally once daily as a single dose for 90 days. Intervention Process and Information, Sample Collection Visit Timing Visit 1 (Baseline Visit) was conducted upon subject enrolment and the visit date was designated as Day 0. Visit 2 (Mid‐term Visit) was conducted 45 days (±3‐day window) after Visit 1. Visit 3 (Final Visit) was conducted 90 days (±3‐day window) after Visit 1. Data Collection Researchers collect demographic data at Visit 1. Physical examinations (including height, weight and head circumference) were performed at Visits 1, 2 and 3. The Bristol Stool Form Scale (BSFS) was completed at Visits 1 and 3. Functional gastrointestinal disorders (Rome IV criteria), diarrhoea, respiratory allergies, skin allergies, gastrointestinal infections, respiratory infections and dietary intake were collected. In accordance with the standardized dietary guidelines provided prior to enrollment, the parents or guardians of the participating infants were asked to complete a dietary questionnaire to record formula intake, breast milk intake and whether complementary foods had been introduced. Adverse events were continuously monitored and recorded throughout the 90‐day intervention period. Faecal Sample Collection Faecal samples were collected at Visit 1 (V1, D0) and Visit 3 (V3, D90) for gut microbiota analysis (16S rRNA gene sequencing). During collection, a dedicated faecal scoop was used to obtain mid‐section internal samples, avoiding surface contamination and degradation. Samples were placed into pre‐labelled faecal storage containers, sealed in resealable bags and immediately stored in a dry ice cooler. After collection, samples were temporarily stored in a low‐temperature freezer at −30°C to −15°C for up to 2 h. Subsequently, the stool samples were transferred to pre‐labelled stool cryopreservation containers, sealed in self‐sealing bags and immediately stored in a dry ice cooler for temporary storage. They were then stored in a low‐temperature freezer at −30°C to −15°C within 2 h of collection and transferred to an ultra‐low temperature freezer at −90°C to −60°C within 12 h for long‐term storage. Safety Evaluation The safety evaluations covered in this study encompass: growth and development indices (including height, weight and head circumference), disease and infection surveys (including the incidence rates of respiratory allergies, skin allergies and fever), the incidence of functional gastrointestinal disorders, complete blood count (CBC), adverse reactions and concomitant medications. Clinical Indexes Detection Blood samples were collected at Visit 1 and Visit 3 for CBC and biochemical tests (CRP, IL‐2, IL‐6, TNF‐α). Serum samples were collected at Visit 1 (V1, D0) and Visit 3 (V3, D90) for inflammatory cytokine and immune marker testing. After the collection of whole blood, the samples were centrifuged. The serum was then transferred into EP tubes for a liquoting and cryopreservation. The samples were temporarily stored at −90°C to −60°C as required. Batches of samples were shipped to the analytical testing centre using dry ice (−90°C to −60°C) for transport. DNA Extraction, Database Construction and Sequencing of Gut Microbiota Sequencing was performed at Shanghai Honsun Biological Technology Co. Ltd. (Shanghai, China). Total genomic DNA was extracted from frozen faecal sample using the HiPure Stool DNA Mini Kit (Magen) following the manufacturer's protocol. DNA integrity was checked by 1% agarose gel electrophoresis. The Qubit 4 (Thermo Fisher, United States) was used to determine the final DNA concentration and purity. The V3‐V4 hypervariable region of the bacterial 16S rRNA gene was amplified using the primer pair 338F (5′‐ACTCCTACGGGAGGCAGCAG‐3′) and 806R (5′‐GGACTACHVGGGTWTCTAAT‐3′). Each 20 μL reaction mixture contains 4 μL of 5 × TransStart Rapid Pfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of TransStart Rapid Pfu DNA polymerase and 10 ng of template DNA. The thermal cycling program was: initial denaturation at 95°C for 3 min, followed by 27 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s and extension at 72°C for 45 s, extension at 72°C for 10 min and ending at 4°C. PCR was performed in triplicate. The amplified products were separated by 2% agarose gel electrophoresis, excised and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA). Purified amplicons were quantified with a Qubit 4.0 fluorometer. The purified amplification products were mixed in equimolar ratios and sequenced using paired‐end sequencing on the Illumina Novaseq 6000 PE250 platform (Illumina, San Diego, United States), in accordance with the standard protocols of Honsunbio Technology Co. Ltd. (Shanghai, China). Statistical Analysis Statistical Analysis of the Population Full Analysis Set (FAS): Comprising all subjects who participated in the trial and completed at least Visit 1. The FAS set represents the primary population for efficacy analysis. Per‐Protocol Set (PPS): As a subset of the FAS set, the PPS consists of subjects who completed all trial visits without major protocol deviations during the study period and serves as the secondary population for efficacy analysis. Safety Analysis Set (SS): The safety analysis set encompasses all subjects who participated in the trial, received the study product and had safety endpoint data recorded. General Data Processing Quantitative data were expressed as n (number of subjects without missing values), mean, median and standard deviation. Qualitative data were presented in terms of frequency and percentage. Before database lock, outliers were identified as values falling outside the mean±3 standard deviation (SD) for each continuous variable, calculated separately within each age group and visit. Outliers were excluded only if deemed biologically implausible (e.g., measurement error) and after confirming that exclusion did not change the overall conclusions. Rigorous data cleaning was conducted, including verification of data entry errors, resolution of inconsistencies and standardization of formats. Missing values were not imputed during the cleaning process and data corrections and consistency checks were completed. Statistical Analysis This study was primarily hypothesis‐driven for the predefined primary outcomes: change in Bifidobacterium relative abundance from baseline to 3 months. All other outcomes (complete blood count, physiological parameters, inflammatory markers, correlation analyses) were considered exploratory. For the primary study results, a two‐sided raw p < 0.05 was considered significant. For the exploratory outcomes, the Benjamini‐Hochberg false discovery rate (FDR) correction was applied to control for multiple testing. An FDR‐adjusted p < 0.05 was considered statistically significant. Additionally, independent samples t‐tests were employed to compare groups with respect to continuous variables (e.g., gestational age, birth weight, birth length). Chi‐square tests were utilized to compare groups regarding categorical variables (e.g., sex, mode of delivery). Generalized estimating equations (GEE) were used to analyse time effects, group effects and time × group interactions for repeated measurements of growth indicators (height, weight, head circumference), blood parameters, functional gastrointestinal disorders, stool consistency scale, diarrhoea, respiratory allergies, skin allergies, gastrointestinal infections, respiratory infections, fever incidence and feeding status at different visit time points in the experimental and control groups. Intra‐group comparisons were conducted using paired t‐tests or Wilcoxon signed‐rank tests. Combined medication use and adverse reactions were analysed using Fisher's exact test. Bioinformatics Analysis The UPARSE algorithm was used to cluster high‐quality sequences into operational taxonomic units (OTUs) at a 97% similarity threshold and taxonomic annotation was performed using the SILVA database. Alpha diversity indices, namely the Shannon index and Simpson index, were utilized to characterize the bacterial community within samples. The Shannon index reflects taxa richness, while the Simpson index reflects taxa evenness. For beta diversity analysis, Principal Coordinates of Abundance (PCoA) based on the Bray‐Curtis distance matrix was applied to visualize differences in microbial community structure between groups. Permutational multivariate analysis of variance (PERMANOVA) was employed to test for significant intergroup differences. Linear discriminant analysis (LDA) effect size (LEfSe) analysis was used to screen for differentially abundant taxa between groups, identify significantly different taxa and visually present their distribution across taxonomic levels. Metabolic pathway prediction was conducted using the PICRUSt2 tool to predict microbial functions and KEGG metabolic pathways. The Spearman correlation analysis was used to explore the relationship between the abundance changes of key bacteria and the changes in growth and development indicators. Results Overview of Involvement of the Participants This study was conducted at The Second Affiliated Hospital of Nanjing Medical University, the First Teaching Hospital of Tianjin University of Traditional Chinese Medicine, Jining No. 01 People's Hospital, Women and Children's Health Care Hospital of Linyi, Hunan Provincial Maternal and Child Health Care Hospital, Panjin Liaohe Gemstone Flower Hospital from August 2024 to May 2025. A total of 240 infants aged 6–36 months were recruited. Ultimately, 240 subjects signed informed consent forms and were formally enrolled in the study. Participants were stratified according to age and randomly allocated to either the intervention group or the control group for a 90‐day intervention study. The study design and participant flow were shown in Figure 1. A total of 227 subjects completed the clinical trial, whereas 13 subjects withdrew prematurely, resulting in an overall dropout rate of 5.42%. Overview of the Participants' Involvement in the Study. Baseline of Participants Independent samples t‐tests were utilized to compare continuous variables (e.g., gestational age, birth weight, birth length) between groups, while chi‐square tests were applied for categorical variables (e.g., gender, mode of delivery). Among all the baseline demographic and sociological characteristics, only birth weight exhibited a statistically significant difference between the experimental and control groups (p = 0.036); no other variables showed significant differences (p > 0.05). These findings indicated that the two groups were well‐balanced and comparable at baseline (Table 1). TABLE 1Baseline characteristics of the subjects (infants aged 6–36 months).6–12 months infants13–36 months infantsGroup YI + CP‐9Group YI t/x2 p Group OI + CP‐9Group OI t/x2 p Gender (n, %)Male22, 44.9%18, 37.5%0.5480.45927, 56.3%26, 50.00%0.3910.532Female27, 55.1%30, 62.5%21, 43.8%26, 50.00%Gestational age (days)Mean273.22274.23−0.6870.493275.19275.130.0400.969SD7.057.226.606.75Median273.00273.00275.50275.00Minimum260.00259.00260.00263.00Maximum291.00288.00287.00290.00Birth weight (g)Mean3201.633315.83−1.5400.1273341.353198.372.1290.036* SD365.62364.88330.83339.71Median3170.003255.003310.003192.50Minimum2525.002560.002550.002505.00Maximum3980.004000.004000.003875.00Birth height (cm)Mean49.6949.96−1.8640.06550.0249.751.5930.114SD0.680.710.760.93Median50.0050.0050.0050.00Minimum47.0047.0048.0047.00Maximum51.0052.0052.0053.00Delivery mode (n, %)Natural childbirth17, 34.7%19, 39.6%0.2480.61825, 52.1%32, 61.5%0.9100.340Assisted childbirth0, 0.0%0, 0.0%0, 0.0%0, 0.0%Caesarean section32, 65.3%29, 60.4%23, 47.9%20, 38.5%Feeding way (n, %)Pure breast milk20, 40.8%15, 31.3%1.0450.593————Mixed breast milk13, 26.5%16, 33.3%——Formula milk powder without probiotics16, 32.7%17, 35.4%——Sibling (n, %)No15, 30.6%12, 25.0%0.3800.53714, 29.2%15, 28.8%0.9330.627Yes34, 69.4%36, 75.0%34, 70.8%36, 69.2%Weight visit 1 (baseline) (kg)Mean9.989.720.7280.46912.6712.72−0.0840.933SD1.891.563.761.56Median10.009.7511.5012.00Minimum7.005.508.509.00Maximum17.5013.0028.0022.00Height visit 1 (baseline) (kg)Mean74.4174.43−0.0130.98984.9284.160.4170.678SD6.877.38.279.73Median75.0073.8585.2583.80Minimum55.0058.0060.0060.00Maximum90.0090.00101.00102.00Head visit 1 (baseline) (kg)Mean44.8644.93−0.1130.91047.3447.160.4000.690SD2.413.292.172.23Median44.5044.6047.2047.10Minimum34.6031.0042.8042.50Maximum49.5054.0052.5052.00* p < 0.05. Baseline characteristics of the subjects (infants aged 6–36 months). p < 0.05. Safety Assessment Effects on Faecal Morphology of Subjects Significant time effects were observed in both younger infants aged 6 to 12 months and older infants aged 13 to 36 months (p < 0.001). As time progressed (from Visit 1 to Visit 3), the faecal morphology in both groups changed, with an increase in Type IV (normal consistency) and a decrease in Type VI (diarrhoea‐like consistency). In younger infants aged 6 to 12 months, no significant inter‐group differences or time × group interactions were observed (p > 0.05). The results indicated a significant time effect, while the between‐group differences and time × group interactions were not significant (Table S1). This suggested that the stool consistency of infants aged 6 to 12 months did indeed change over time. However, this change pattern was consistent in both the intervention and control groups, indicating that it could not be attributed to the CP‐9 intervention. Significant time effects (p < 0.001) and time × group interactions (p < 0.05) were observed in infants aged 13–36 months. The distribution of changes in stool consistency over time differed significantly between the experimental group OI + CP‐9 and control group OI (time × group interaction: Wald χ 2 = 5.543, p = 0.019). Both the group OI + CP‐9 and the group OI exhibited changes in stool consistency over time, characterised by an increase in Type IV (normal) and a decrease in Type VI (diarrhoea). The group OI + CP‐9 showed a higher proportion of Type IV (normal) stool consistency compared to the group OI, while exhibiting a lower proportion of Type VI (diarrhoea) stool consistency (Table 2). This finding suggested that the trial product may improve stool consistency in infants aged 13–36 months. TABLE 2Comparison of faecal morphology scale scores between the experimental group and control group at different follow‐up time points (SS).Faecal MorphologyInfants aged 13–36 monthsExperimental group OI + CP‐9Control group OITime effectGroup effectTime × group interactions(n, %)(n, %) WaldX 2 p WaldX 2 p WaldX 2 p Visit 1Type I1, 1.7%1, 1.8%25.589 < 0.001 *** 2.1300.1446.7320.035* Type II4, 6.7%7, 12.7%Type III7, 11.7%7, 12.7%Type IV26, 43.3%20, 36.4%Type V9, 15.0%3, 5.5%Type VI13, 21.7%17, 30.9%Type VII0, 0.0%0, 0.0%Visit 3Type I2, 3.3%0, 0.0%Type II1, 1.7%3, 5.5%Type III3, 5.0%5, 9.1%Type IV34, 56.7%31, 56.4%Type V9, 15.0%9, 16.4%Type VI7, 11.7%4, 7.3%Type VII0, 0.0%0, 0.0%* p < 0.05.*** p < 0.001. Comparison of faecal morphology scale scores between the experimental group and control group at different follow‐up time points (SS). < 0.001 *** p < 0.05. p < 0.001. Effects of CP‐9 on the Gastrointestinal Symptoms in Subjects(SS) The generalized estimating equation (GEE) was used to analyse the time effects, group effects and time × group interactions in the repeated measurement data regarding the occurrence of functional gastrointestinal disorders between the experimental group and the control group at different follow‐up times. For infants aged 6 to 12 months, there were no significant time effects in the incidence of regurgitation, rumination syndrome, cyclic vomiting syndrome, intestinal colic, functional diarrhoea, difficult defecation and functional constipation between the experimental and control groups (p > 0.05, Table S2–S15). Except for only one case of functional diarrhoea that occurred in the control group YI during Visit 1, no significant time × group interaction differences were observed in the occurrence of other symptoms. For infants aged 6–36 months, no significant group differences were observed for regurgitation, difficult defecation or functional constipation between the experimental and control groups (p > 0.05, Tables S3, S13 and S15). No cases of rumination syndrome were observed in either the OI + CP‐9 or the OI group among infants aged 13–36 months (0 cases) (Table S5). Only one case of cyclic vomiting syndrome occurred in the control group OI during Visit 3 (Table S7). Only one case of intestinal colic occurred in the OI group during Visit 2 (Table S9). One case of functional diarrhoea occurred in the experimental group OI + CP‐9 during Visit 3 (Table S11). These findings suggested that the product exhibits good gastrointestinal tolerability and safety for subjects, without increasing the risk of functional gastrointestinal disorders. Effect of CP‐9 on the Occurrence of Diseases and Infections in Subjects (SS) Among infants aged 6–12 months, there were no significant time effects, group effects and time × group interaction in the incidence of respiratory allergy, skin allergy and fever between the experimental group and the control group (p > 0.05, Tables S16–S20). Only one case of respiratory allergy occurred in the YI + CP‐9 group during Visit 1 and Visit 3, respectively. For infants aged 13–36 months, only the OI + CP‐9 group had one reported case at Visit 1 (Tables S17, S19). The time effect was observed in the incidence of fever when comparing the YI + CP‐9 group and the YI group (p < 0.001). In both groups, the occurrence of fever changed over time. However, neither significant differences between the groups nor time × group interactions were observed (p > 0.05, Table 3). These findings indicated that the CP‐9 demonstrated good tolerability and safety, without increasing the risk of respiratory allergy, skin allergy and fever. TABLE 3Comparison of fever occurrence between the experimental group and control group at different follow‐up time points (6–12 months) (SS).FeverInfants aged 6–12 monthsExperimental group YI + CP‐9 (n, %)Control group YI (n, %)Time effectGroup effectTime × group interactions WaldX 2 p WaldX 2 p WaldX 2 p Visit 1Yes14, 23.3%19, 31.7%15.441< 0.001*** 0.0020.9671.9540.376No46, 76.7%37, 61.7%Visit 2Yes1, 1.7%1, 1.7%No57, 95.0%54, 90.0%Visit 3Yes7, 11.7%12, 20.0%No50, 83.3%40, 66.7%*** p < 0.001. Comparison of fever occurrence between the experimental group and control group at different follow‐up time points (6–12 months) (SS). p < 0.001. Effect of CP‐9 on Feeding Practices in Subjects As shown in Table 4, no significant differences were observed between the CP‐9 and placebo groups at any visit time for breastfeeding status, formula milk powder feeding status and supplemental feeding of complementary food (p > 0.05). GEE analysis confirmed no significant group differences or time × group interactions for any feeding practices (p > 0.05). A significant time effect was detected only for supplemental feeding of complementary food (p < 0.05), reflecting expected developmental progression with age, but this effect was similar in both groups (time × group interaction p = 0.692). Collectively, these results indicate that dietary intake was well balanced between the CP‐9 and placebo groups throughout the intervention. TABLE 4Comparison of feeding practices between the experimental and control groups at different follow‐up time points (13–36 months).IndexInfants aged 13–36 monthsExperimental group OI + CP‐9Control group OITime effectGroup effectTime × group interactions x¯±s x¯±s WaldX 2 p WaldX 2 p WaldX 2 p Total amount of breast milk(mL)Visit 1300.00 ± 177.04499.09 ± 240.640.2270.8931.3560.2441.6290.443Visit 2376.67 ± 114.13417.14 ± 226.33Visit 3412.50 ± 342.48490.00 ± 356.79Total amount of formula milk powder (mL)Visit 1445.21 ± 227.15408.57 ± 293.760.1900.9103.2240.0731.5900.452Visit 2460.32 ± 266.42358.46 ± 227.61Visit 3460.49 ± 225.18380.00 ± 197.67Supplemental feeding of complementary food (n, %)Visit 1Yes31, 51.7%28, 46.7%17.928< 0.001* 0.2040.6510.7380.692NoNo29, 48.3%31, 51.7%Visit 2Yes37, 61.7%35, 58.3%NoNo22, 36.7%24, 40.0%Visit 3Yes46, 76.7%39, 65.0%NoNo13, 21.7%19, 31.7%* p < 0.05. Comparison of feeding practices between the experimental and control groups at different follow‐up time points (13–36 months). p < 0.05. Effect of CP‐9 on Adverse Events No statistically significant differences were observed between the experimental and control groups for infants aged either 6 to 12 months or 13 to 36 months (Table 5). This finding indicated that the CP‐9 demonstrated good tolerability and safety, without increasing the risk of adverse events. TABLE 5Summary of adverse events in medical coding (SS).Adverse eventsTotal (n, %)Infants aged 6–12 monthsInfants aged 13–36 monthsGroup YI + CP‐9 (n, %)Group YI (n, %) p Group OI + CP‐9 (n, %)Group OI (n, %) p Adverse events40, 17.2%9, 15.0%7, 12.7%0.79213, 22.0%11, 19.0%0.820Infection‐related diseasesCold19, 8.2%5, 8.3%5, 9.1%1.0005, 8.5%4, 6.9%1.00014, 6.0%4, 6.7%5, 9.1%0.7353, 5.1%3, 5.1%1.000Mycoplasma pneumonia3, 1.3%0, 0.0%0, 0.0%—2, 3.4%2, 3.4%1.000Pneumonia1, 0.4%0, 0.0%0, 0.0%—0, 0.0%0, 0.0%0.496Upper respiratory infection1, 0.4%1, 1.7%0, 0.0%1.0000, 0.0%0, 0.0%—Respiratory system, thoracic and mediastinal diseasesCough3, 1.3%2, 3.3%0, 0.0%0.4791, 1.7%0, 0.0%1.0001, 0.4%0, 0.0%0, 0.0%—1, 1.7%0, 0.0%1.000Rhinitis1, 0.4%1, 1.7%0, 0.0%1.0001, 1.7%0, 0.0%—Acute bronchitis1, 0.4%1, 1.7%0, 0.0%1.0000, 0.0%0, 0.0%—Skin diseaseAllergic dermatitis4, 1.7%0, 0.0%0, 0.0%—2, 3.4%2, 3.4%1.0004, 1.7%0, 0.0%0, 0.0%—2, 3.4%2, 3.4%1.000Immune system diseaseEgg sensitive1, 0.4%0, 0.0%1, 1.8%0.4780, 0.0%0, 0.0%—1, 0.4%0, 0.0%1, 1.8%0.4780, 0.0%0, 0.0%—Systemic diseaseFever8, 3.4%1, 1.7%1, 1.8%1.0004, 6.8%2, 3.4%0.6798, 3.4%1, 1.7%1, 1.8%1.0004, 6.8%2, 3.4%0.679Gastrointestinal system diseasesDiarrhoea5, 2.2%1, 1.7%0, 0.0%1.0001, 1.7%3, 5.2%0.3642, 0.9%1, 1.7%0, 0.0%1.0000, 0.0%1, 1.7%0.496Emesis1, 0.4%0, 0.0%0, 0.0%—0, 0.0%1, 1.7%0.496Gastroenteritis2, 0.9%0, 0.0%0, 0.0%—1, 1.7%1, 1.7%1.000 Summary of adverse events in medical coding (SS). During the trial period, among infants aged 6–12 months, adverse events were reported in 9 cases (15.0%) in the YI + CP‐9 group and 7 cases (12.7%) in the YI group and there was no significant difference between the two groups (p = 0.792). A total of 9 infants caught cold, including 4 (6.7%) in the YI + CP‐9 group and 5 (9.1%) in the YI group (p = 0.735). For infants aged 13–36 months, 13 adverse events (22.0%) were recorded in the OI + CP‐9 group and 11 (19.0%) in the OI group, with no significant difference between the two groups (p = 0.820). Regarding infections and infestations, a total of 9 cases were recorded. Specifically, the incidence of cold was 5.1% in the YI + CP‐9 group and 3.4% in the YI group (p = 1.000). Allergic dermatitis was observed in 2 cases (3.4%) in both the experimental group OI + CP‐9 and control group OI (p = 1.000). None of these adverse events were deemed to be related to the study product. Notably, 82.2% of infants did not experience any adverse events during the trial. The occurrence of adverse events was random across the treatment and control groups and no causal relationship was established between the adverse events and the study product. Changes in Clinical Indexes Growth and Development Index In this study, for the infants (6–36 months) in the experimental group, growth indicators (height and head circumference) showed a significant increase before and after the intervention (FDR‐adjusted p < 0.05, Figure 2A–C). No significant differences were observed between the experimental and control groups and weight change was not significantly influenced by time. Changes in growth and development indicators: (A) height, (B) weight and (C) head circumference in each group pre‐ and post‐intervention. *FDR‐adjusted p < 0.05, **p < 0.01 and ***p < 0.001. The growth and development indicators of all infants in the study, such as height, weight and head circumference, were all influenced by time (FDR‐adjusted p < 0.05, Figure 2A–C). All these indicators exhibited an upward trend over time. However, there were no significant differences among the groups, which suggests that the test product has good safety in terms of growth and development under condition of 90‐day intervention. Crucially, no significant between‐group differences were found in any growth parameter across both age strata. These results indicate that although all infants followed healthy, age‐appropriate growth trajectories, the probiotic intervention did not change these natural developmental patterns, thus confirming its safety regarding physical growth. Biochemical Indexes Effect of CP‐9 on the Blood Immune Indexes and CBC Monitoring systemic immunological markers is crucial during infancy, as they can reflect both the maturing immune function and potential subclinical health disturbances. In contrast, the levels of interleukin‐2 (IL‐2) and tumour necrosis factor‐alpha (TNF‐α) decreased significantly compared to the baseline (FDR‐adjusted p < 0.05, Figure 3A–J). For older infants, there were no statistically significant differences in any blood immunological or inflammatory markers between the OI + CP‐9 and OI groups before and after the intervention (Figure 3B–D). Changes of blood immune indicators and CBC: (A) CRP, (B) IL‐2, (C) IL‐6, (D) TNF‐α and (E) ANC, (F) MO, (G) ABC, (H) MCH, (I) MCHC and (J) NEUT pre‐ and post‐intervention in each group. *FDR‐adjusted p < 0.05, **p < 0.01 and ***p < 0.001. CBC analysis offers crucial insight into haematological and immunological status during the early development stage, functioning as a vital safety monitor in paediatric intervention studies. Among younger infants, the mean corpuscular haemoglobin concentration (MCHC) and mean corpuscular haemoglobin (MCH) showed a significant decrease after the CP‐9 intervention when compared to pre‐intervention levels (FDR‐adjusted p < 0.05, Figure 3H,I). Importantly, all the detected parameters above remained within the age‐appropriate reference ranges throughout the study. Although statistically significant differences were observed between the baseline and post‐intervention levels, the extent of these changes fell within the expected physiological fluctuations for infants and no clinically significant abnormalities were detected. Effect of CP‐9 on the Gut Microbiota Effect of CP‐9 on Gut Diversity and Bacterial Community Structure Alpha diversity reflects the abundance and diversity of microbial communities. After the intervention, significant differences in alpha diversity were observed in the YI + CP‐9 and YI groups compared to their baselines, while no such change was observed in the OI + CP‐9 and OI groups. Furthermore, no statistically significant differences in alpha diversity were found between the YI + CP‐9 and its control group YI, nor between experimental group OI + CP‐9 and the control group OI (Figure S1). β‐diversity, which measures compositional dissimilarity between microbial communities, provides critical insight into overall structural shifts. Comparative analysis revealed a significant difference in microbial community structure before and after the intervention in the YI + CP‐9 and YI groups, while no significant structural changes were observed in the OI + CP‐9 and OI groups over the same period (Figure 4A–D). Moreover, between‐group comparisons after the intervention showed no significant differences in community structure between the YI + CP‐9 and YI groups, nor between the OI + CP‐9 and OI groups (Figure 4E,F). PCoA of the gut microbiota in (A) YI + CP‐9, (B) YI, (C) OI + CP‐9, (D) OI groups pre‐ and post‐intervention and (E) YI, (F) OI groups inter‐group comparison post‐intervention based on the Bray‐Curtis distance of 16S rRNA V3‐V4 sequencing data and PERMANOVA statistical analysis, YI + CP‐9 V1 versus YI CP‐9 V3, p < 0.05; YI V1 versus Y3 V3, p < 0.05. Effect of CP‐9 on Key bacteria Genera Screened by LEfSe The Linear Discriminant Analysis Effect Size (LEfSe) analysis was utilized to screen out the key differential bacterial genera in each group comparison before and after the intervention. Notably, after the intervention, the abundance of Bifidobacterium was significantly enriched in both younger (YI + CP‐9) and older (OI + CP‐9) infants compared to their respective control groups (Figure 5A,B). A simultaneous enrichment in the abundance of Blautia was also observed in the older infant group. These findings suggested that CP‐9 can enhance the abundance of this beneficial genera in infants aged 6 to 36 months. In contrast, the abundance of mucin‐degrading genus Akkermansia decreased in infants (aged 6 to 36 months) after the intervention. Effect of CP‐9 on key bacteria genera screened by LEfSe analysis. Key genera of (A) YI and (B) OI after intervention compared with those before intervention. Changes of Key Pathways for Prediction of Gut Microbiota Function by Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) PICRUSt was used to predict the functional potential of the gut microbiota based on 16S rRNA data. Differential functional analysis comparing pre‐ and post‐intervention within each group identified key pathways and KEGG orthology (KO) terms (Figure 6A–I). The predicted abundance of K00059 (encoding 3‐oxoacyl‐[acyl‐carrier protein] reductase) increased in the YI + CP‐9 group after intervention (Figure 6A). This key KO is involved in ko00061 (Fatty acid biosynthesis) and ko00780 (Biotin metabolism), both of which relate to the biosynthesis of polyunsaturated fatty acids (PUFAs) and other fatty acids. These results suggested that CP‐9 supplementation was associated with an increased predicted functional potential for PUFA biosynthesis in the infant gut microbiota, based on PICRUSt inference. Microbiota‐derived fatty acids, especially certain PUFA, can be absorbed by the host and serve as metabolic supplements, potentially contributing to the cellular membrane structure and signalling processes. Differential KO of the gut microbiota in the (A) YI + CP‐9 group, (B) YI group and (C) OI group pre‐ and post‐intervention, (D) between OI + CP‐9 and OI group post‐intervention, (E) between OI + CP‐9 and OI group pre‐intervention and differential pathway of the gut microbiota in the (F) YI + CP‐9 group, (G) OI versus CP‐9 group pre‐intervention, (H) YI group pre‐ and post‐intervention and (I) OI group pre‐ and post‐intervention, respectively. Correlation Between the Clinical Index and Gut Microbiota Spearman analysis was utilized to explore the correlation between the relative abundance of key genera and clinical immune/CBC parameters (Figure 7A–D). These exploratory analyses aim to generate hypotheses regarding potential links between gut microbiota composition and systemic haematological/immunological parameters during early immune development. To control for multiple comparisons, the Benjamin‐Hochberg false discovery rate (FDR) correction was applied separately within each age stratum (younger infants, 6–12 months; older infants, 12–36 months), with a significance threshold of FDR‐adjusted p < 0.05. Correlation analysis between the key gut microbiota genus and blood immune markers in (A) YI; (B) OI group and CBC indicators in (C) YI; (D) OI infants. *FDR‐adjusted p < 0.05, **p < 0.01 and ***p < 0.001. In younger infants, after FDR correction, the abundance of Bifidobacterium was significantly negatively correlated with CRP (Figure 7A, FDR‐adjusted p = 0.028) and positively correlated with lymphocyte percentage (LYM) (Figure 7C, FDR‐adjusted p = 0.034). In older infants, Bifidobacterium abundance was significantly negatively correlated with Neutrophil Percentage (NEUT) (FDR‐adjusted p = 0.035) and positively correlated with absolute lymphocyte count (ALC) (Figure 7D, FDR‐adjusted p = 0.040). In contrast, Akkermansia abundance was positively correlated with NEUT in older infants (FDR‐adjusted p = 0.035). Discussion Bifidobacterium is widely recognized as a cornerstone of the healthy infant gut microbiota (Turroni et al. 2012), especially in breastfed infants and plays a fundamental role in supporting gut ecosystem development (Bäckhed et al. 2015; Turroni et al. 2022). Bifidobacterium can reach the lower intestine, colonize and persist within the competitive gut microenvironment. This capacity depends on evolved molecular mechanisms that mediate host‐cell interactions and facilitate colonization (Alessandri et al. 2021). Bifidobacterium are gaining increasing attention as functional foods and the safety of CP‐9 for infants under 6 months of age has been evaluated in previous studies (Chen et al. 2023). In this 90‐day trial, CP‐9 supplementation was not associated with clinically meaningful adverse events or deviations from normal growth trajectories in healthy infants aged 6–36 months. All CBC parameters and systemic inflammatory immune markers remained within age‐appropriate ranges throughout the intervention, indicating no obvious short‐term immune safety concerns. CP‐9 intervention was associated with a significant increase in Bifidobacterium abundance without any observed disruption to the host's natural developmental processes. These observations were consistent with the known stability of inflammatory markers in healthy, well‐nourished populations. The fluctuations observed here stayed within normal limits, consistent with a healthy baseline cohort (Bäckhed et al. 2015). Importantly, the absence of adverse signals over 90 days should be interpreted as preliminary evidence of short‐term tolerability, not as definitive safety confirmation. Longer‐term studies are needed to further assess safety in this population. The health‐promoting roles of these enriched genera are well‐substantiated. Research indicated that Bifidobacterium enhanced nutrient absorption and utilization, promoted weight gain in infants suffering from severe acute malnutrition and reduced intestinal inflammatory markers (Barratt et al. 2022) and its early‐life supplementation has been linked to a reduced risk of developing allergic conditions such as eczema (Ismail et al. 2016). Notably, emerging evidence suggested that different Bifidobacterium strains can act synergistically to maintain the integrity and homeostasis of the intestinal epithelial barrier (Rizzo et al. 2024). Furthermore, Blautia, a known SCFAs producer, has also been demonstrated to confer beneficial effects on the host (Freitas et al. 2022). Blautia contributes to alleviating inflammatory and metabolic disorders while exhibiting antimicrobial activity against specific microorganism (Liu et al. 2021). Their co‐presence can synergistically activate metabolic pathways, such as the acetate‐producing Wood‐Ljungdahl pathway, further amplifying beneficial outcomes. Thus, by elevating these taxa, CP‐9 helps establish a health‐promoting microbial environment during a critical developmental window. In contrast, the mucin‐degrading genus Akkermansia abundance exhibited decreased in infants (aged 6–36 months) following intervention. This change may be attributable to competitive interactions with concurrently enriched Bifidobacterium, which also utilizes host‐derived glycoconjugates such as mucin and human milk oligosaccharides (Milani et al. 2015). Increased mucin breakdown by Bifidobacterium could limit substrate availability for Akkermansia, thereby constraining its growth. Notably, Akkermansia was difficult to detect in healthy two‐month‐old infants. They begin to appear in infants aged 12 months, generally at low abundance levels (Yassour et al. 2016). Given that the intestinal mucus layer, immune system and barrier function are still developing during infancy, caution has been exercised regarding its early colonization. The functional role of Akkermansia in early life remains complex and context‐dependent. Akkermansia has been shown to correlate with disease progression in various autoimmune and neuroinflammatory disorders. Multiple studies have confirmed that increased abundance of this bacterium in patients with Parkinson's disease and multiple sclerosis positively correlates with disease severity (Jangi et al. 2016; Berer et al. 2017; Lin et al. 2019). Mechanistically, in individuals with compromised intestinal barriers, its potent mucin‐degrading capacity may further disrupt mucus layer integrity, exacerbate local inflammation and increase the risk of pathogen translocation. However, causal evidence in pathogenesis remains limited. Thus, the reduction observed in Akkermansia following probiotic intervention may reflect a modulation toward a microbial structure less reliant on mucin metabolism. This shift suggests that ecological regulation mechanisms may operate through nutritional competition, wherein CP‐9 promotes a microbial environment dominated by microbial groups with recognized health‐promoting functions while reducing those whose high abundance during infancy may indicate ecological imbalance or niche competition. This restructuring aligns with the concept of probiotics as “microbial partners” that fine‐tune rather than radically transform the gut ecosystem. PICRUSt‐based functional prediction revealed that CP‐9 intervention was associated with an increased predicted abundance of K00059, a gene encoding 3‐oxoacyl‐[acyl‐carrier protein] reductase involved in fatty acid biosynthesis pathways. The predicted increase in this pathway's functional potential suggested that CP‐9 supplementation was associated with an enhanced predicted capacity for fatty acid biosynthesis, particularly PUFAs. For instance, Bifidobacterium encodes specialised enzymes that convert dietary linoleic acid into conjugated linoleic acid (CLA), yielding anti‐inflammatory metabolites while limiting the production of pro‐inflammatory derivatives (Miyamoto et al. 2019). They serve as beneficial supplements to host metabolism and mediate intestinal epithelial function and immune regulation (Parada Venegas et al. 2019). Therefore, a clearer understanding of how microbial metabolites, particularly those derived from Bifidobacterium, influence the overall gut ecosystem is of considerable importance (Zhang et al. 2025). Notably, CP‐9 did not induce significant changes in systemic immune or inflammatory markers nor accelerate linear growth, both consistent with the stable health status of subjects in this group. However, correlation analysis revealed that Bifidobacterium abundance negatively correlated with CRP and NEUT while positively correlated with LYM and ALC, consistent with previous research (Chen et al. 2024). These patterns reflect normal immune maturation processes (Hart et al. 2004). During infancy (1–2 years), lymphocytes account for approximately 0.60 and neutrophils for approximately 0.35. Subsequently, the neutrophil proportion gradually increased, reaching a balance ratio with lymphocytes by ages 4–6 years (Weiping Wang et al. 2022). Furthermore, the observed link between Bifidobacterium and lymphocyte‐favouring profiles echoes previous findings that specific Bifidobacterium strains can enhance T lymphocyte proportions and phagocytic activity (Gill et al. 2001). These subtle but physiologically coherent associations provide additional support for the short‐term safety of CP‐9 supplementation, without implying direct effects on systemic immune parameters or host physiology. However, these exploratory correlation‐based findings do not imply causation and require validation using causal inference frameworks. Several limitations of this study should be acknowledged. First, 16S rRNA V3‐V4 sequencing does not permit strain‐level identification. Future studies employing strain‐resolved methods (e.g., metagenomic sequencing or strain‐specific qPCR) are needed to confirm strain‐specific colonization. In addition, further studies employing more advanced approaches (Maldonado‐Gómez et al. 2016) are needed to confirm the long‐term colonization of probiotic strains, to identify the core microbial constituents of the gut microenvironment and to predict colonization outcomes through functional gene profiling. Second, dietary data were collected via parental questionnaires, which may be subject to recall bias. Nonetheless, because dietary intake was well balanced between the CP‐9 and placebo groups, any potential bias was unlikely to have systematically influenced the comparison of microbiota changes. Finally, the 90‐day follow‐up period may not capture long‐term safety or ecological effects. Despite these limitations, this study provides valuable preliminary evidence for the short‐term safety and Bifidobacterium‐enriching effect of CP‐9 in healthy infants aged 6–36 months. Future research should employ direct metagenomic and metabolomic approaches, extend observation periods to assess the long‐term safety of CP‐9 and further elucidate its mechanisms of action. Conclusion In healthy infants aged 6–36 months, 90‐day supplementation with CP‐9 resulted in a significant enrichment of the relative abundance of Bifidobacterium. During the intervention period, all measured parameters, including growth trajectories, complete blood count and systemic inflammatory markers, remained within normal physiological ranges and showed no treatment‐related adverse changes. No clinically meaningful improvements in these parameters were observed, which is consistent with the healthy, well‐nourished baseline status of the study population. These findings suggest that CP‐9 was well‐tolerated in the short term and can enrich beneficial gut bacterium without overtly disrupting normal physiological parameters. However, the mechanisms underlying these findings remain to be determined and longer‐term studies are needed to confirm safety and potential benefits. Author Contributions Liujing Zhang: writing – review and editing, writing – original draft. Yingguang Xie: data curation. Liwei Nie: data curation. Xueke Song: investigation. Keming Zhang: data curation. Xuelu Dai: visualization. Xiaoshuang Xu: methodology, formal analysis. Yiran Liu: visualization. Li Zhao: data curation. Jiai Yan: investigation. Feng Zhang: supervision, conceptualization. Xinmei Guo: resources. Wei Li: methodology. Zhiping Sun: conceptualization. Xiaoling Zeng: methodology. Jing Sun: visualization. Funding This work was supported by the Innovation and Entrepreneurship Talent Program of Jiangsu Province, Enterprise Innovation Category, JSSCRC2023500. National Natural Science Foundation of China, 82370809. The funding for Medical Key Strategic Project of Wuxi Health Commission. Consent Written informed consent was obtained from the parents or legal guardians of all participants. Conflicts of Interest Author Xinmei Guo, Wei Li and Xiaoling Zeng were employed by Jinqiao Biotech Co. Ltd. The other authors declares no conflicts of interest. Supporting information Figure S1: α‐diversity analysis of gut microbiota in (A) YI + CP‐9, (B) YI, (C) OI + CP‐9 and (D) OI groups pre‐ and post‐intervention, and (E) YI, (F) OI groups inter‐group comparison post‐intervention. Table S1: Comparison of faecal morphology scale scores between the experimental group and control group at different follow‐up time points (6–12 months) (SS). Table S2: Comparison of regurgitation occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S3: Comparison of regurgitation occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S4: Comparison of rumination syndrome occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S5: Comparison of rumination syndrome occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S6: Comparison of cyclic vomiting syndrome occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S7: Comparison of cyclic vomiting syndrome occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S8: Comparison of intestinal colic occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S9: Comparison of intestinal colic occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S10: Comparison of functional diarrhoea occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S11: Comparison of functional diarrhoea occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S12: Comparison of difficult defecation occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S13: Comparison of difficult defecation occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S14: Comparison of functional constipation occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S15: Comparison of functional constipation occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S16: Comparison of respiratory allergy occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S17: Comparison of respiratory allergy occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S18: Comparison of skin allergies occurrence between the experimental and control groups at different follow‐up time points (6–12 months) (SS). Table S19: Comparison of skin allergies occurrence between the experimental and control groups at different follow‐up time points (13–36 months) (SS). Table S20: Comparison of fever occurrence between the experimental group and control group at different follow‐up time points (13–36 months) (SS).