Probiotic intake and mental health in healthy working adults: a systematic review and meta-analysis of randomized controlled trials

Abstract

Background

Workers face significant mental health challenges from stress, anxiety, and depression, impacting individuals, organizations, and society. Emerging research indicates a link between the gut microbiome and mental well-being, suggesting probiotics as a potential support. This study objectively evaluated probiotic supplementation’s effects on depression, anxiety, stress, sleep, and related biological markers in this population.

Methods

We conducted a systematic review and meta-analysis following PRISMA 2020 guidelines. EMBASE, Cochrane Library, and PubMed were searched for randomized controlled trials assessing probiotic supplementation on a range of psychological outcomes. The primary outcomes were perceived symptoms of depression, anxiety, stress, and sleep quality. The secondary outcomes were physiological markers of mental health, such as cortisol and C-reactive protein levels, in working populations. Eligible studies included healthy employed adults (≥ 18 years), without psychiatric, neurodegenerative, genetic, infectious, or endocrine disorders, including pregnancy.

Results

Twelve studies involving 3,350 participants were incorporated. Probiotic consumption had a modest yet statistically significant positive effect on subclinical psychological outcomes, including symptoms of depression, anxiety, and stress, in healthy working adults (standardised mean difference (SMD) = -0.21, 95% CI [-0.34, -0.09], p = 0.001). These findings were maintained despite moderate statistical heterogeneity that was likely due to variations in probiotic strains, dosages, and duration of supplementation used. Moreover, the probiotic interventions were associated with a statistically significant reduction in cortisol levels, a key biomarker of physiological stress (SMD = -0.26, 95% confidence interval [CI] [-0.45, -0.08], p = 0.005). Conversely, no statistically significant effects were observed for probiotic supplementation on the C-reactive protein levels, a marker of systemic inflammation. However, due to the lack of available evidence, it was impossible to draw firm conclusions about the effects of probiotics on sleep quality and biomarkers of oxidative stress.

Conclusion

This systematic review and meta-analysis provide preliminary evidence suggesting that probiotic supplementation may hold promise as an approach to improve mental well-being within working populations. However, further high-quality randomised controlled trials targeting this population are needed to determine the optimal probiotic strains, dosages, and treatment durations for addressing specific mental health outcomes.

Trial registration

PROSPERO number CRD42024510170.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40359-025-03885-5.

Introduction

Mental health, defined as “a state of well-being in which an individual realises their own abilities, can cope with the normal stresses of life, can work productively and fruitfully, and is able to make a contribution to their community” [1] is increasingly challenged in healthy working adults [2]. Among healthy working adults, rising exposure to psychosocial stressors has led to increasing levels of subclinical psychological distress, making this population particularly at risk and underscoring the need for preventive approaches [3]. The proportion of healthy working adults at risk for stress and poor mental health is substantial worldwide. In a large Canadian cohort, approximately 11% of workers were exposed to multiple psychosocial job stressors, placing them at elevated risk for burnout, stress, and cognitive strain [4]. Broader global estimates suggest that the at-risk population is considerably larger when accounting for single or combined stressors [5]. International data indicate that around one in five workers report significant stress-related symptoms despite the absence of a diagnosed mental disorder [6], a pattern consistent with European surveys in which roughly 20% of employees report experiencing work-related stress [7].

The continuing pressures faced by working people, such as demanding workloads and longer working hours, together with job insecurity and difficulties in balancing work and personal life, could have a negative impact on health [8]. Tunisia reflects this trend, with indicators of work-related stress and psychological strain increasing in recent years [9, 10]. These conditions have a substantial impact worldwide, accounting for 8.3% of total years lived with disability for individuals and 4.0% for the community [9]. It’s estimated that depression and anxiety alone cost us 12 million working days per year in lost productivity, absenteeism, presenteeism and quality of life; In dollar terms, it means a loss of about 1$ trillion in global productivity [11]. This underscores the importance of implementing strategies to cultivate a stronger and more adaptable workforce.

Over the past decade, scientists have focused on the intricate relationship between the gut and the brain. The gut microbiome, a complex ecosystem of trillions of microorganisms residing in the human gut, plays a central role in this bidirectional communication network, involving complex pathways [12]. Via a variety of mechanisms, including neurotransmitter production, immune modulation and the production of short-chain fatty acids, scientists increasingly suggest that the gut microbiome may profoundly shape human brain function and behavioural patterns [13]. Probiotics, defined as “live microorganisms which when administered in adequate amounts confer a health benefit on the host,” hold promise for configuring the gut microbiome to benefit mental health [14]. Although most researchs have studied the ability of probiotics to treat existing psychological illnesses, an increasing number of studies show that probiotics also have a protective effect on healthy individuals [15]. Probiotic supplementation may facilitate the establishment of a healthy and diverse gut microbiome, which in turn may help reduce the harmful effects of stress, improve emotion regulation, and promote mental health, even in clinically healthy individuals [16]. A variety of organisational initiatives have been implemented with a view to addressing health challenges in the workplace. These include workplace wellness programmes and employee assistance programmes [17].

Despite efforts to address mental health, inconsistent success, influenced by factors like participation rates, privacy concerns, and implementation inconsistencies, underscores the critical need for further research in this field [17, 18]. Probiotics are yet to be proven effective for mental health and considerable gaps remain in our understanding of their targeted actions among healthy working populations [19, 20]. Many of the previous studies have targeted clinical populations and/or used a wide variation in probiotic strains, doses and lengths of supplementation which complicates any inference about the potential for probiotics to prevent mental health issues in workplace settings [21]. Probiotics could offer a preventive strategy in this subgroup by modulating stress-related pathways and physiological markers (e.g., cortisol), potentially enhancing resilience and overall well-being in employment settings.

The present systematic review and meta-analysis assessed the currently available evidence base relating to probiotic use in non-clinical working adults and associated mental health outcomes. Unlike previous research focused largely on clinical or sick population, the novelty of this study lies in its focus on healthy working adults, a population that has been underexplored despite being at risk for work-related stress and poor mental health. This study evaluated the impact of probiotic supplementation on multiple psychological outcomes, including depression, anxiety, stress and sleep quality in this population. In addition, the research explored the effect of probiotic consumption on biological markers related to mental well-being such as cortisol and C-reactive protein levels, and genetic indications of oxidative stress.

Methods

Study protocol and registration

This study adhered to the PRISMA 2020 guidelines for systematic reviews and meta-analyses, prioritising a rigorous and transparent methodological approach [22]. To ensure reproducibility and minimise bias, the trial protocol was registered with the PROSPERO international register of systematic reviews on April 26, 2024, under the unique identification number CRD42024510170. The proactive registration of the a priori protocol underscores the commitment to a systematic and transparent research process, bolstering the reliability and validity of the findings.

Research question

The question of this research was: “Does probiotic supplementation influence psychological well-being and physiological outcomes in healthy working adult, assessed in terms of various indicators through a systematic review and meta-analysis of randomised controlled trials?”

Search methods

We conducted a comprehensive systematic search of multiple databases, including Embase, Scopus and PubMed via Medline, from December 2023 to April 2024. This comprehensive search strategy ensured the inclusion of a wide range of relevant studies for analysis.

The biomedical databases search was performed for studies published from January 2014 to April 2024, without any language restriction, using MeSH terms and Boolean operators according to the PICO format.

The search equations for the various databases are summarised in the table below (Table 1).

Table 1.

Database-specific search strategies

Database	Search terms
Pubmed	((((((((((probiotics[MeSH Major Topic]) OR (gastrointestinal microbiota[MeSH Major Topic])) OR (gut microbiota[MeSH Major Topic])) AND (mental health[MeSH Major Topic])) OR (mental hygiene[MeSH Major Topic])) OR (mental stress[MeSH Major Topic])) OR (oxidative stress[MeSH Major Topic])) OR (sleep disorder[MeSH Major Topic])) OR (depression[MeSH Major Topic])) AND (healthy workers[MeSH Major Topic])) OR (healthy adults[MeSH Major Topic])
EMBASE	((((((((((probiotics[MeSH Major Topic]) OR (gastrointestinal microbiota[MeSH Major Topic])) OR (gut microbiota[MeSH Major Topic])) AND (mental health[MeSH Major Topic])) OR (mental hygiene[MeSH Major Topic])) OR (mental stress[MeSH Major Topic])) OR (oxidative stress[MeSH Major Topic])) OR (sleep disorder[MeSH Major Topic])) OR (depression[MeSH Major Topic])) AND (healthy workers[MeSH Major Topic])) OR (healthy adults[MeSH Major Topic])
Cochrane Library	(probiotics or gut microbiota or intestinal microbiota) AND (mental health OR mental hygiene OR stress OR depression OR sleep disorder OR oxidative stress OR anxiety) AND (healthy workers or healthy adults)

The search terms were developed from previous reviews and refined in collaboration with a specialist librarian using the PRESS (Peer Review of Electronic Search Strategies) checklist to ensure accuracy and completeness [23].

The search plan was created through agreement among all the team researchers.

Database-Specific Search Strategies is presented in Supplementary material 1.

Additional studies were identified by screening the reference lists of relevant articles and systematic reviews, as well as by conducting supplementary targeted searches (e.g., checking key authors, recently published articles, and relevant citations during the review process).

Eligibility criteria

Inclusion and exclusion criteria were established in advance of the start of screening procedure.

Inclusion criteria

The six PICOst dimensions (participants, interventions, comparators, outcome, study design, and timing) were applied to define the eligibility criteria.

Characteristics of participants

We included studies that involved healthy participants in occupational settings, aged above 18 years old, regardless of sex and ethnicity. Participants were required to be described as “healthy” and “free of diagnosed psychiatric or affective disorders”, as well as major functional, structural, or neurodegenerative abnormalities. Studies also excluded individuals with conditions of genetic or acquired origin, infectious diseases, or endocrine and hormonal imbalances, including pregnancy.

Types of included interventions

Trials assessing the effectivity of viable microorganisms to be used as probiotics, without synbiotics single- and/or multi-strain preparations. There was any restriction on the dosage of probiotics used or the administration way (orally, inhaled, etc.).

Records were selected without limiting the intervention duration.

Comparators

To be included in this review, trials had to compare probiotics with a placebo. These excluded studies comparing the strains with each other or with an antibiotic.

Types of outcome assessments

The assessment of workers’ mental health and well-being in this systematic review posed significant methodological challenges, as these constructs are multifaceted and complex. The diverse range of outcomes evaluated in relation to work-related mental health necessitated a selective approach, wherein we, researchers, prioritised mental health indicators that were commonly measured across the included studies, regardless of the specific assessment scales employed. They fell into two categories as follows:

Psychological outcomes of mental health: subclinical psychiatric symptoms of depression, anxiety and stress levels (rated using a validated physician or any of the self-report instrument); sleep quality (rated using a validated physician or any of the self-report instrument). Subclinical psychiatric symptoms are psychological or behavioral signs that are significant enough to be detected on validated psychometric scales but do not meet the full diagnostic criteria for a mental disorder.

Physiological outcomes of mental health: The blood concentration of serum C-reactive protein (CRP) salivary cortisol levels, oxidative stress markers among workers were chosen as biomarkers of mental health.

Study design

We exclusively considered randomised controlled trials.

Timing

All records published between January 2014 and April 2024 available in digital format were retrieved for this systematic review.

Exclusion criteria

The following categories of studies were excluded from this review:

• ➢ Studies that employed probiotics in both treatment arms.

• ➢ Trials comparing or combining probiotics with antibiotics.

• ➢ Research articles investigating antibiotics, prebiotics, or bacterial lysate.

• ➢ Publications that were inaccessible or solely available in print format.

• ➢ Non-research articles such as reviews, technical notes, letters to the editor, and case reports.

• ➢ In vitro or cadaveric experiments.

• ➢ Studies that did not address any mental health outcomes.

• ➢ Studies encompassing non-worker populations, including football players, athletic teams from scholastic and university institutions, and university students.

• ➢ Studies involving patients with pre-existing medical conditions, specifically those diagnosed with recognised inflammatory diseases.

Data collection and analysis

Study selection

A screening plan was created through discussions among all reviewers before any studies were selected. The records obtained from both electronic and manual searches were uploaded to the Rayyan website [24]. This platform is a free online tool designed to assist the independent review and coding process within systematic reviews, where duplicates were identified and removed. The studies found through our predefined search strategy were then evaluated for eligibility based on their titles and abstracts by two trained and supervised reviewers.

At this point, two categories of findings were excluded from further evaluation: first, records with improper formatting, such as book chapters, assessments, editor correspondence, clinical protocols, and review articles; and second, trials that contained unrelated material, including those that did not focus on the mental health of workers or did not involve probiotics. The full texts of the remaining potentially relevant records were carefully examined by the same two researchers for final inclusion in the review.

A third review author was consulted for studies where consensus was not achieved. All reasons for exclusion were thoroughly documented. Excluded studies and their corresponding reasons for exclusion were recorded in line with the guidelines outlined in the Cochrane Handbook for Systematic Reviews of Interventions [25].

Data retrieval and management

A standardised data extraction template was used to outline and compile the details of each trial that had been previously selected [26]. This extraction form adhered to the guidelines set forth in the Cochrane Handbook for Systematic Reviews of Interventions [25] and included essential information such as author details, study design, participant demographics (including ethnicity, average age, BMI, etc.), sample size, methodology, the probiotic used, intervention duration, outcomes assessed, conclusions, and potential sources of bias. If any information regarding these aspects was unclear or missing, efforts were made to reach out to the original report authors for clarification. Any discrepancies were discussed and resolved in consensus meetings with a third review author.

Assessment of bias in the included studies

Two of the review authors independently evaluated the quality of the selected studies by identifying biases, using the Cochrane Collaboration’s tool for assessing the risk of bias in the included trials [25]. The five domains assessed were:

1. bias arising from the randomisation process

2. bias caused by deviations from the intended interventions

3. bias associated with incomplete outcome data

4. bias in the assessment of outcomes

5. bias in the selection and reporting of outcomes

All identified sources of bias were classified as having low, high, or uncertain risk, and a comprehensive risk evaluation was determined for each included study.

Discrepancies were resolved through consultation with a third author.

Measures of treatment effect

Dichotomous data

In the case of binary variables, we intended to report outcomes as relative risks with corresponding 95% confidence intervals.

Continuous data

We saw to analyse the mean difference when continuous outcomes were measured using the same instrument or scale throughout the pooled trials. Hence, in cases where outcomes were assessed with diverse tools or metrics, the standardised mean difference was contemplated as the method to combine and analyse the results across studies.

Unit of analysis issues

Studies with multiple treatment arms

Each intervention arm’s results was evaluated individually, while maintaining the same sample size for the common reference group across all comparisons (against a common control group: placebo). By analysing each intervention arm separately rather than combining them, we can minimize any potential bias or variation that may result from the diffrent treatment effects within the same trial. This method guarantees that our comparisons remain consistent and trustworthy, while also avoiding the complications that can arise from merging different intervention groups, thereby preserving the integrity of the analysis.

Cross-over studies

For best performance in cross-over trials, participants are randomly allocated to receive each of the interventions sequentially, in a specific order, depending on the prior intervention. It is a routine practice to put a washout period between the treatments in such studies. In this stage, it is possible to allow the results of the first intervention to fade before the next intervention is administered. The washout phase is an important process because it allows to reduce the possibility of residual effects hence to better infer the real effect of the second intervention which would otherwise be confounded from the first one. Due to the nature of the interventions that are under scrutiny, this is a relevant point. Therefore, we have chosen to report data from the first treatment phase only in crossover trials to guarantee the reliability of our results.

Assessment of heterogeneity

Statistical heterogeneity was assessed using either the Cochrane Chi2 (Q-test) or Tau2, with a significance level set at less than 0.1 to indicate heterogeneity. We evaluated the possibility of combining data and performing meta-analyses by examining the variability of treatment effects across studies, which is also referred to as the level of heterogeneity. For this purpose, we utilised the I² statistic, indicating the proportion of variability in effect estimates.

The following thresholds were established for evaluating heterogeneity: I² less than 49% indicates low heterogeneity, from 50% to 74% suggests moderate heterogeneity, and 75% or more signifies high heterogeneity.

In cases where significant statistical heterogeneity was detected, we aimed to identify potential contributing factors, such as differences in study quality, outcome measurement tools, reporting practices, participant demographics, types of interventions, or other relevant factors, through retrospective subgroup analyses when appropriate.

We planned to delay conducting a meta-analysis in situations where heterogeneity was classified as high (≥ 75%). In these cases, we would refrain from pooling data and instead perform a sensitivity analysis to assess the robustness of the findings.

Sensitivity analysis

When possible, we aimed to carry out a sensitivity analysis to assess the robustness of our findings and to investigate whether any variability was due to low-quality trials. To evaluate the reliability of our conclusions regarding the overall effects of probiotics, we performed the following sensitivity analyses:

• ➢ Conducting the sequential trial exclusion method or the leave-one-out sensitivity analysis [27] to better understand how each individual trial influences the overall standardized mean differences (SMD) and the associated p-values

• ➢ Excluding studies that presented a high risk of bias

• ➢ Excluding studies with over 20% missing outcome data

• ➢ Transitioning from a fixed-effect model to a random-effects model

If specific areas of sensitivity arise from the review, we would aim to engage in a constructive dialogue to gain a deeper insight into the potential reasons behind the identified uncertainties. Therefore, we believed it was crucial to approach the interpretation of the review’s findings with caution.

Assessment of reporting biases

Because of the expected limited number of studies corresponding to our inclusion criteria, it was thought that a detailed assessment of reporting biases might not be necessary. In accordance with the Cochrane Handbook for Systematic Reviews of Interventions, it would be of interest, in the future updates of this meta-analysis that would include more studies, to examine potential reporting biases, by employing funnel plots, for the publication bias assessment [25].

Data synthesis

All data were analysed on an intention-to‐treat basis; given the preventive nature of therapy in all studies. In instances where outcome variables were deemed homogenous and studies seemed aligned in terms of demographics and measured outcomes, the data were aggregated and combined from the individual trials in a way that was consistent with the guidance set out in the Cochrane Handbook for Systematic Reviews of Interventions [25]. Subsequently, an inverse-variance approach was employed for quantitative analytical purposes, utilising the Review Manager software (RevMan 5.4.1) in order to facilitate the process.

We considered a number of factors in determining whether to involve an individual study in a meta-analysis. Two of the most important were the likelihood of obtaining treatment effect data from the record and the extent of heterogeneity. If the data retrieved from the record in question proved to be missing, we opted to provide a narrative description of the results, and we concluded our analysis at the qualitative level.

In light of the I² statistic, we used both fixed-effects and random-effects models with regard to the issue of heterogeneity.

In cases where studies evaluated a common interventional approach and the demographics seemed to be homogeneous, and when the objective was to estimate a similar underlying treatment effect, that is, in the assumption of both clinical and statistical homogeneity, with an I²value less than 50%, a fixed-effect meta-analysis was considered an appropriate approach.

This would have involved using Chi² and I² estimates to combine data and calculate relevant parameters. If clinical heterogeneity suggested differing underlying treatment effects across trials, or if significant statistical heterogeneity (I² > 50%) was detected, the decision was taken not to pool the data.

When applicable, sensitivity analyses were conducted to identify potential causes, and a random-effects model was planned to be used. This allows for a more comprehensive interpretation of the results and provides an overall summary if the average treatment effect was clinically meaningful. In this case, the results were reported as the average treatment effect, along with its 95% confidence interval, utilising Tau² and I² estimates.

If there were high levels of inconsistency, with I² exceeding 75%, that remained unexplained, or if there was only one study identified, or if the overall treatment effect was considered to lack significance, the studies were not combined, and a meta-analysis was judged to be unfeasible. Instead, a narrative approach was taken to summarise the evidence and the study characteristics.

Dealing with missing data

In cases of missing numerical outcome data or when bias areas were not clearly reported, we intend to contact the records’ authors via E-mail to resolve these gaps. If the corresponding authors failed to respond, then we would provide the necessary information from the figures in the articles.

Results

Overview of studies

The systematic electronic research of the prementioned databases using the MeSH terms and Boolean operators provided us with 1956 publications. Furthermore, 23 additional articles from alternative sources were identified as potentially relevant to the review topic through manual examination of the reference lists of the included reviews. A summary of the systematic steps of the study collection have been illustrated in the PRISMA flow diagram as in Fig. 1. After discarding duplicates, the initial 1979 records were reduced to 1876 articles, which were then screened based on their titles and abstracts. Through this rigorous process, we were able to select 24 articles for a comprehensive review of their full texts. Twelve of the articles were not eligible for inclusion for the following reasons: the participants in ten trials were not working adults; one article was excluded because the intervention did not fulfil the inclusion criteria; and one paper was not suitable for inclusion because the outcomes did not comply. This left a total of twelve articles (with 3243 participants) that were eligible for inclusion in the review.

Fig. 1.

Study flow diagram in accordance with PRISMA guidelines

Characteristics of included studies

The results of the individual studies were provided in a consolidated format in Table 2, offering a comprehensive overview of the principal outcomes.

Table 2.

Characteristics of included studies examining probiotic effects on worker well-being

Author/Year/ Country	Study design	Occupation; Sample size (probiotic/total); Age (mean years intervention group vs. mean years control group)	Supplementation type; Daily dosage	Duration (days)	Outcomes in probiotics group (compared with control group)
Sato et al./2022/Japan [28]	double-blind, placebo-controlled, parallel group comparison study	Japanese workers; 30/60; 46.3 vs.47.4	3 strains probiotic capsule : Pediococcus acidilactici Lactiplantibacillus plantarum Lactobacillus plantarum; 3 × 106 CFUs*	28	• Significant increase in the mental component score of the SF-8* (P = 0.002) • Notable enhancement in mental health (P < 0.001) and role-emotional scores (P = 0.002) on the SF-8 scale
Pacifici et al./2020/Italy and Albania [29]	Randomised double-blinded placebo controlled trial	Oral and maxillofacial surgeons; 20/40; 26 vs.45	15 strains probiotic tablet: Lactobacillus plantarum Lactobacillus acidophilus Bifidobacterium infantis Lactobacillus fermentum Lactobacillus reuteri Lactobacillus casei Bifidobacterium Longum Lactobacillus rhamnosus Bifidobacterium Lactis Lactobacillus salivarius Lactobacillus paracasei Bifidobacterium bificum Lactobacillus gasseari Bifidobacterium Breve Streptococcus thermophilus; 225 × 1012 CFUs/tablet	70	• Reduced levels of salivary cortisol
Smith-Ryan et al./2019/United States of America [30]	randomised double-blinded placebo-controlled trial	Healthcare workers; 15/33; 30.5 vs. 30.2	9 strains probiotic tablet : Bifidobacterium bifidum W23, Bifidobacterium lactis W51, Bifidobacterium lactis W52, Lactobacillus acidophilus W37, Lactobacillus brevis W63, Lactobacillus casei W56, Lactobacillus salivarius W24, and Lactococcus lactis (W19 and W58); 2.5 × 106 CFUs	42	• No difference in CRP* level • No difference in depression scales • Reduction in anxiety (difference − 2.3 ± 2.6) and fatigue (− 4.8 ± 5.5)
Kokubo et al./2020/Japan [31]	Randomised, Open-Label Crossover Study	Corporate office workers; 80/153; 36.2 vs.36.4	One probiotic stain yoghurt: Lactococcus lactis Plasma; 1 × 1011 CFUs	28	• Higher levels of vigor (P = 0.02) in the POMS* Questionnaire • No differences in anxiety and depression in the POMS Questionnaire
Mohammadi et al./2015/Iran [32]	randomised, blinded three-arm parallel design	Petrochemical factory workers; 25/35; 32.vs. 32.8	2 probiotic strain yoghurt: Lactobacillus acidophilus LA5 Bifidobacterium lactis BB12; 1 × 107 CFUs 7 probiotic strains capsule M7: Lactobacillus casei;3 × 103CFUs, Lactobacillus acidophilus; 3 × 107 CFUs, Lactobacillus rhamnosus; 7 × 109CFUs, Lactobacillus bulgaricus; 5 × 108CFUs, Bifidobacterium breve; 2 × 1010CFUs Bifidobacterium longum; 1 × 109CFUs, S. thermophilus; 3 × 108 CFUs.	42	• No difference in CRP •Lower levels of plasma PC*, serum 8-dihydroguanine and iso-prostaglandin resulted by the consumption of probiotic capsule. •Lower levels of plasma PC among petrochemical workers after the consumption of probiotic yogurt for 6 weeks, but no difference in the levels of serum 8-dihydroguanine. •Higher levels of serum 8-oxo-7 in the two probiotic groups in comparison to the control group.
West et al./2020/Australia [33]	randomised blinded three-arm parallel design	Healthcare workers; 58/87; 41.8 vs. 40.7 vs. 41.6	One probiotic strain capsule: Lactobacillus acidophilus (DDS-1) or Bifidobacterium animalis subsp. Lactis (UABla-12);1 × 1010 CFUs	14	• Lower level of serum cortisol (P = 0.01) during v1 to v2 • Higher levels of sleep quality (P = 0.06), 22% decrease in the PSQI* (improved sleep) but not significant • No difference in sleep duration/time recorded on Fitbit activity • Inverse correlation at baseline between PSQI and the Connor Davidson Resilience Scale (r = − 0.21, P = 0.07)
Kinoshita et al./2019/Japan [34]	randomised, controlled, and open-label trial	women healthcare workers; 479/961; 39.3 vs. 39.4	2 probiotic strain yoghurt: Lactobacillus bulgaricus OLL1073R-1 Streptococcus thermophilus; ≥1.12 × 109 CFUs	112	• No difference in levels of CRP
Kinoshita et al./2021/Japan [35]	randomised, controlled, and open-label trial	women healthcare workers; 479/961; 39.3 vs. 39.4	2 probiotic strain yoghurt: Lactobacillus bulgaricus OLL1073R-1 Streptococcus thermophilus; ≥1.12 × 109 CFUs	112	• Higher levels of general health (P = 0.02) and vitality (P = 0.01) (according to Quality of life QOL related health SF-8 questionnaire) • No differences in other emotional and mental health • Higher levels of sleep quality (P = 0.007) (according to the PSQI questionnaire)
Shida et al./2017/Japan [36]	randomised double-blinded placebo controlled trial	Corporate office workers; 49/99; 40.5 vs. 40.6	1 probiotic stain fermented milk: Lactobacillus casei; ≥1011 CFUs	84	• Lower level of salivary cortisol (P = 0.045)
Slykerman et al./2022/New Zealand [37]	randomised, double-blind, placebo-controlled trial	Healthy nurses; 300/600; 40,74 vs. 40,08	one probiotic strain capsule : Lactobacillus rhamnosus HN001; 6 × 109 CFUs	84	•No significant difference in stress, anxiety, and psychological wellbeing
Chong et al./2019/Malaysia [38]	randomised, double-blind, placebo-controlled trial	Healthy employees at Penang and Kubang Kerian; 56/111; 31.1 vs. 32.1	one probiotic strain powder: Lactobacillus plantarum DR7, 1 × 109 CFUs	84	•Reduced symptoms of stress (P = 0.024), anxiety (P = 0.001), and total psychological scores (P = 0.022) •No difference against depression •Lower plasma cortisol levels (P = 0.008)
Lew et al./2019/Malaysia [39]	randomised, double-blind, placebo-controlled trial	Healthy employees at Penang and Kubang Kerian; 52/103; 31.1 vs. 32.1	one probiotic strain powder: Lactobacillus plantarum P8; 2 × 1010 CFUs	84	•Reduced scores of stress for the probiotic group (P = 0.048), anxiety (p = 0.031) and total psychological score (p = 0.041) as assessed by the DASS-42* questionnaire. •The effects of P8 against reduction of depression was insignificant: The P8 only marginally reduced depression as compared to placebo (P < 0.10). • Marginal difference in plasma cortisol levels (P = 0.090)

SF −8 Short Form-8 Health Survey, CRP C-Reactive Protein, POMS Plasma, PC Plasma Phosphatidylcholine, PSQI Pittsburgh Sleep Quality Index, DASS-42 Depression Anxiety and Stress Scale, CFU Colony-forming units

Seven studies [28–30, 36–39] were designed as randomised, double-blinded, placebo-controlled trials, one paper employed an open-label randomised crossover design [31], two papers were randomised trials comprising three parallel groups [32, 33], and two records were both randomised controlled and open-label trials [34, 35]. Included studies were published between 2015 and 2022 across diverse geographic regions. While five of the interventions took place in Japan [28, 31, 34–36], the rest were held respectively in Iran [32], the United States of America [30], Australia [33], New Zealand [37], Malaysia [38, 39] and/Albania [29].

The final stage of the study selection process involved a total of 3,350 subjects, with 1,710 randomly assigned to the intervention groups and 1,640 to the control groups. The participants were between the ages of 18 and 71 years. All studies did mention the participants’ gender except two [32, 38]. And among the nine records that reported the gender, 2788 (90%) of participants were women. eleven of the twelve included studies defined the working field of their participants. Although the occupational groups varied across the included studies, the majority of randomised employees (n = 2682; 84%) were healthcare workers, including otolaryngologists and maxillofacial surgeons, nurses, certified nursing assistants, emergency medical services personnel, physical therapists, occupational therapists, and other non-specified healthcare workers. The remaining occupational settings were distributed across a range of work environments, with the highest prevalence observed in corporate offices (n = 252; 8%) and the lowest in petrochemical plants (n = 35; 1%) and a Malaysian company based in Penang and Kubang Kerian (n = 214; 7%).

It emerges that approximately half of the records [28–30, 32, 34, 35] incorporated multi-strain probiotics with a variety of bacterial strains, including Lactobacillus, Bifidobacterium, and others. In the reviewed trials, participants were provided with probiotics via oral administration. However, the delivery methods varied, with the most prevalent format being tablets or capsules [28–30, 33, 37], followed by yoghurt [31, 32, 34, 35] then by powder [38, 39] and fermented milk [36] (Table 2).

The majority of studies (n = 10/12) employed commercially available products [28–32, 34, 35, 37–39]. One study made use of a yoghurt developed in their own laboratory [36], while another trial was lacking in detail regarding the accessibility of the product [33]. The intervention periods extended from 14 to 112 days, with an average duration of 59.1 days. The supplemented probiotic doses differed considerably across studies, ranging from 3 million to 225 billion colony-forming units (CFUs).

Seven trials lacked data on adherence to the probiotic intervention [29, 31, 32, 34, 35, 37–39]. When the data were available, the estimates differed. These discrepancies were observed in the percentage of individuals who were fully compliant with the prescribed regimen [28, 36], the rate of non-adherence to the prescribed dosage regimen during the treatment phase [33], and the number of days of supplementation missed during the intervention period [30]. No severe adverse effects associated with the probiotic strains administered were documented in any of the considered studies. Nevertheless, two studies reported that digestive symptoms were experienced as mild side effects [34, 35].

Assessment of bias in the included studies

Given our aim of providing the most comprehensive and unbiased overview of the available evidence, we prioritised data derived from intention-to-treat (ITT) analyses, as ITT preserves the benefits of randomisation and is recommended for minimising bias arising from deviations from intended interventions. The Cochrane Collaboration website provided the ROB2 tool for RCTs, which yielded two summary figures at the conclusion of the risk of bias assessment process. Firstly, a traffic light plot was developed to provide a visual representation of the evaluations of every risk of bias domain for every included randomised controlled trial (referred to as Fig. 2).

Fig. 2.

Risk of bias assessment for domains D1-D5: a traffic light plot analysis

The Fig. 3, is a bar graph that illustrates the proportion of studies classified under every risk of bias judgment across different domains within the final selection of records.

Fig. 3.

Weighted representation of bias categories identified in selected randomized controlled trials

As demonstrated in Table 3, the global risk of bias was rated as low in seven studies (58.3%), moderate in one study (8.3%), and high in four studies (33.3%).

Table 3.

Assessment of potential bias in the included studies

	Randomization process	Deviations from intended interventions	Missing outcome data	Measurement of the outcome	Selection of the reported result	Overall bias
Assignment to investigate (the 'intention-to-treat' effect)
Total number of study = 12
Low risk (%)	83,3	75	100	66,7	100	58,3
Some concerns (%)	0	16,7	0	0	0	8,3
High risk (%)	16,7	8,3	0	33,3	0	33,3

Outcome results of individual included studies

Psychological mental health outcomes

Subclinical-psychological symptoms: (Anxiety, depression and stress)

The seven trials evaluating the perceived stress and the symptoms of anxiety and depression employed diverse scales of outcome measures; yet, they were all subjectives self-reporting scales. These instruments included assessments of mood fluctuations, stress levels, anxiety and depression, and overall quality of life.

This report outlines five distinct scoring methodologies, beginning with the State-Trait Anxiety Inventory 6-item scale (STAI6) [37], followed by the Hospital Anxiety and Depression Scale (HADS)) [30], and then the Short Form. (SF-8) [28, 34],; additionnaly, the Profile of Mood States (POMS) [31],; and finally, the Depression, Anxiety and Stress Scale (DASS-42) [38, 39].

In the intervention groups, participants evidenced varying yet promising results in their self-reported experiences of subclinical psychological symptoms when compared to those in the placebo group. In four of the seven previously noted trials, there was a considerable improvement in psychological well-being among those who received probiotics [28, 30, 38, 39]. One study demonstrated an improvement in both anxiety and depression symptoms [28], while three others indicated a reduction in anxiety symptoms alone [30, 38, 39]. Furthermore, both of the studies that evaluated stress levels reported a statistical significance in the observed improvement [38, 39]. However, the remaining studies did not reveal any substantial differences in psychological assessments between the probiotic and placebo groups [31, 34, 37].

Sleep quality

The two studies that investigated the impact of probiotic supplementation on sleep quality deployed the Pittsburgh Sleep Quality Index (PSQI) [40]. Both of the previously referenced studies [33, 35] reported an improvement in sleep quality in the probiotic supplementation groups. Although a 22% reduction in the PSQI score was recorded in the initial trial, indicating an improvement in the participants’ sleep patterns, this result did not achieve statistical significance (p = 0.06) [33]. On the other hand, the second study documented an enhancement in sleep quality using the same PSQI measure (P = 0.007) [35]. Thus, although a reduction in PSQI scores was observed, the limited number of studies and the lack of consistent statistical significance mean that these findings should be interpreted with caution and cannot be considered definitive evidence of improvements in sleep duration or subjective sleep satisfaction.

Physiological mental health outcomes: (CRP, Cortisol, oxidative stress indicators)

This systematic review of the literature revealed that eight of the trials included in the review evaluated at least one biomarker related to physiological mental health outcomes among workers [29, 30, 32–34, 36, 38, 39].

Cortisol levels were the most frequently assessed biomarker, analysed in five trials [29, 33, 36, 38, 39], followed closely by C-reactive protein, analysed in four trials [30, 32–34]. In contrast, only one trial examined oxidative stress markers [32].

In order to comprehensively assess these oxidative stress biological indicators [41] in the population of healthy workers, Mohammadi et al. [32] selected three established biomarkers: serum 8-oxo-7,8-dihydroguanine (8-oxo-dG), protein carbonyl content, and 8-iso-prostaglandin F2α (8-iso-PGF2α). These prementioned specific biomarkers reflect damage caused to specific biomolecules as follows: 8-oxo-dG is indicative of damage to DNA, protein carbonyls point to protein oxidation, and 8-iso-PGF2α functions as a marker of lipid peroxidation.

In comparison to the control groups, the individuals who received probiotics exhibited a reduction in both plasma and salivary cortisol levels [29, 33, 36, 38, 39], as well as a decrease in plasma protein carbonyl, serum 8-iso-prostaglandin F2α (8-iso-PGF2α), and serum 8-dihydroguanine levels [32]. These findings collectively suggest a multifaceted effect of probiotics on the mitigation of stress responses and oxidative stress. However, it is pertinent to acknowledge that probiotic intervention did not significantly impact circulating C-reactive protein levels among the included trials. This indicates that while probiotics may effectively modulate the specific stress indicator, cortisol, and the oxidative stress markers, their influence on systemic inflammation, as reflected by CRP, appears limited.

Meta-analysis

Results of the impact of probiotic intake on psychological mental health outcomes

We conducted a meta-analysis of seven studies [28, 30, 34, 36–39] to evaluate the impact of probiotic supplementation on psychological well-being. The studies included a total of 1,905 healthy workers. The analysis was primarily directed towards the investigation of three principal outcomes: perceived symptoms of depression, anxiety, and stress.

The adoption of a random-effects model within the meta-analysis revealed a statistically significant improvement in the psychological outcomes in question among healthy workers who received probiotic supplementation, in comparison to those in control groups. The pooled standardised mean difference (SMD) of −0.21 (95% CI [−0.34, −0.09], p = 0.001) indicated a small to moderate effect size, suggesting that probiotic intake may lead to clinically reductions in the levels of anxiety, depression and stress experienced by the participants.

The forest plot provided a visual representation of the individual study results and the pooled effect, revealing a consistent improvement across the included studies (Fig. 4). The heterogeneity across studies was moderate (I² = 48%), and the Chi² test was statistically significant (P = 0.03), indicating that the variability in effect sizes was greater than would be expected by chance alone. This suggests that differences in intervention characteristics (e.g., strain composition, duration) or participant factors (e.g., baseline psychological status) may have contributed to the observed heterogeneity.

Fig. 4.

Forest Plot: Subclinical Anxiety, Depression and Psychological Stress Symptoms - Random-Effects Meta-Analysis of Probiotic Supplementation vs. Placebo. (A: anxiety, D: depression, str: stress)

To provide a degree of robustness to the findings, a sensitivity analysis was conducted using a fixed-effect model (Fig. 5). The results of this analysis were comparable (SMD = −0.14, 95% CI −0.22 to −0.05), providing further support for the conclusion that probiotic supplementation would have a beneficial effect on psychological outcomes in healthy workers. The I² statistic remained steady at 48% for both the random-effects and fixed-effect analyses, indicative of a negligible impact of the selected model on the assessment of heterogeneity.

Fig. 5.

Forest Plot: Mental Health Outcomes – Sensitivity analysis - Fixed-Effects Meta-Analysis of Probiotic Supplementation vs. Placebo. (A: anxiety, D: depression, str: stress)

A second sensitivity analysis was executed in accordance with the sequential trial exclusion method, which served to reaffirm the stability of the initial findings. In this regard, the standardised mean difference was found to remain statistically significant within a range of −0.25 to −0.15 (95% CI: −0.37 to −0.09, −0.25 to −0.05) even after the withdrawal of any individual study from the analysis.

This consistency across sensitivity analyses strengthens the reliability of the findings and suggests that the favourable impact of probiotics on psychological well-being is not significantly influenced by the results of any individual study.

Results of the impact of probiotic intake on the physiological mental health outcomes

Impact of probiotic intake on cortisol concentrations

A random-effects meta-analysis was carried out to explore the potential impact of probiotic supplementation on the physiological stress response in healthy working adults, with cortisol levels identified as a primary outcome measure (Fig. 6). A total of five trials [29, 33, 36, 38, 39], were identified as meeting the inclusion criteria, involving a total of 437 participants and thus providing data suitable for analysis. As illustrated in Chart 6, the standardised mean difference of −0.26 (95% CI [−0.45, −0.08]) provided by the meta-analysis suggested a statistically significant decrease in cortisol levels among participants who received probiotic intake compared to those in the control groups.

Fig. 6.

Forest Plot: hs-CRP and Cortisol Levels - Random-Effects Meta-Analysis of Probiotic Supplementation vs. Placebo (hs-CRP: High-Sensitivity C-Reactive Protein, c: probiotic capsule, y: probiotic yogurt)

Furthermore, the heterogeneity analysis yielded an I² value of 0%, indicating the paucity of significant variability in the results across the included studies. This homogeneity enhances the reliability of the combined effect size, indicating that the observed reduction in cortisol levels is likely a reproducible outcome of probiotic supplementation, rather than a consequence of variations in study populations, interventions, or methodological approaches.

Impact of probiotic intake on blood C-reactive protein concentrations

The results of our meta-analysis suggested that probiotic supplementation did not exert a statistically significant influence on CRP levels in working adults (Fig. 6). The aggregated effect size was − 0.00, with a 95% confidence interval from − 0.12 to 0.12, indicating that the intervention had a negligible impact on CRP levels. This finding was further confirmed by the overall effect test (Z = 0.01, P = 0.99), which demonstrated no statistically significant difference in CRP levels between the probiotic and control groups. The examination of heterogeneity disclosed a minimal degree of variability between the included studies (Chi² = 0.33, df = 3, P = 0.95; I² = 0%). This suggests that the findings were consistent across studies and that the evidenced null effect was not attributed to significant heterogeneity in the study populations or interventions.

Discussion

By investigating the effects of probiotic supplementation on psychological and physiological mental health outcomes in a sample of otherwise healthy workers, our study aimed to contribute to this growing body of knowledge. The primary research question that guided the research was “Does probiotic supplementation influence psychological well-being and physiological outcomes in healthy working adult?“. To address this question, we conducted a meta-analysis when quantitative data were available. When pooling was not possible, we provided a descriptive synthesis of study findings. Overall, the twelve included studies (3,350 participants) suggest that probiotics provide modest benefits in reducing subclinical symptoms of depression, anxiety, and stress, and may contribute to lowering cortisol levels. No clear effects were observed for inflammatory markers such as C-reactive protein, and evidence regarding sleep and oxidative stress remained limited. These findings indicate that probiotics could play a role as an early preventive strategy [42].

Probiotics and subclinical psychological symptoms

Our findings are consistent with preclinical reports of comparable effects in rodent models using Lactobacillus strains [43, 44]. In human studies, Wallace and Milev [45] systematically reviewed 10 randomised controlled trials and provided robust evidence that probiotic supplementation significantly reduced symptoms of anxiety, depression and stress in both healthy adults and patients with major depressive disorder. Likewise, a systematic review and meta-analysis of seven randomised controlled trials by McKean et al. [46] confirmed the positive findings regarding the beneficial effects of probiotics on psychological symptoms in individuals without pre-existing mental health conditions. Huang et al. [47] also found, through their meta-analysis of five RCTs, that probiotic supplementation significantly improved mood and reduced symptoms of depression, showing benefits in both healthy individuals and those with major depressive disorder or anxiety disorder. Three main hypotheses have been proposed in the literature to explain the observed mental health benefits of probiotics: the immune pathway, the neuroendocrine pathway and epigenetic mechanisms. First, probiotics may improve mental health by modulating the immune system, particularly by reducing systemic inflammation, which is widely considered to be a key trigger of mental disorders [48]. Evidence suggests that probiotics may modulate inflammation by reducing pro-inflammatory cytokine production, suppressing nuclear factor kappa-B activation (a key regulator of inflammatory and immune processes), and regulating the kynurenine pathway [47]. Certain probiotic strains are able to decrease pro-inflammatory kynurenine metabolites while increasing anti-inflammatory intermediates, thereby balancing serotonin and kynurenine levels through the kynurenine pathway and promoting a favourable neurochemical environment, contributing to potential mental health benefits [49]. Furthermore, it has been demonstrated that probiotics improve gut barrier integrity, thereby reducing endotoxin leak and attenuating systemic inflammation [50]. Second, probiotics might influence mood regulation by acting through the neuroendocrine pathway to modulate neurotransmitter systems, potentially by directly regulating levels of serotonin (via tryptophan), GABA, and ACTH [46, 51]. Furthermore, probiotics could indirectly impact neurotransmission by modifying the expression of neurotransmitter receptors, with some strains potentially increasing GABA uptake by modulating GABA receptor expression and thereby promoting inhibitory neurotransmission [44]. Third, emerging evidence suggests that probiotics may improve mental health via epigenetic mechanisms—modulating gene expression related to neurological processes, inflammation, and stress response—as demonstrated by Lactobacilli and Bifidobacteria altering gene methylation to potentially preserve mood and cognitive function [52]. Probiotics can also induce histone modifications, altering chromatin structure and gene transcription in ways that support mental health, potentially reshaping the genetic profile to maintain optimal mental function and emotional balance [44].

The results of our study differed from those provided by Romjin and Ricklidge [53] who, after a descriptive review of ten RCTs examining the effects of probiotics on psychological symptoms in people with psychiatric disorders, stated that the evidence for their efficacy was limited. Similarly, the systematic review by Liu et al. [54] suggested that the evidence for the efficacy of probiotics in reducing anxiety was limited. In contrast, our study used a rigorous quantitative and analytical methodology, which differs from the narrative approaches used in the previous research by Romjin and Rucklidge [53] and Liu et al. [54]. It’s worth noting that previous researches has suggested that while probiotics may have beneficial effects on mood and psychological well-being in healthy individuals, their efficacy appears to be diminished in populations with pre-existing mental [55–58] or physical health problems, such as irritable bowel syndrome and rheumatoid arthritis [46].

Moderate heterogeneity in our meta-analysis limited the robustness of conclusions, likely reflecting subtle differences in participant stress or emotional reactivity not captured by standard psychometric measures [59]. Beyond baseline mental health, genetic and epigenetic factors likely contributed to heterogeneity [60], as individual variations can shape gut microbiome composition and responses to probiotics [61].

Environmental and lifestyle variables, including diet, physical activity, sleep, and smoking, may also have influenced gut–brain interactions, yet these were insufficiently controlled or reported, limiting interpretability [62]. Differences in probiotic interventions further added to variability [63], particularly in strain specificity, dosage, duration, and delivery. Strain-dependent mechanisms, such as neurotransmitter or short-chain fatty acid production, immune modulation, and gut barrier effects, can yield divergent outcomes [64]. For example, certain strains of Bifidobacterium and Lactobacillus are known to produce gamma-aminobutyric acid, a neurotransmitter with sedating effects related to reduced anxiety, while others, such as Lactobacillus rhamnosus GG, have been shown to improve gut barrier integrity, potentially reducing inflammation and its impact on mental health [65]. Furthermore, the fact that all included studies used either Lactobacillus or Bifidobacterium species supports their hypothetical potential to promote mental health, which is consistent with existing research highlighting their relevance in this regard [66].

The choice of single-strain or multi-strain probiotics could also influence the heterogeneity of effect sizes due to fundamental differences in their composition and potential mechanisms of action [67]. Furthermore, the possibility of synergistic interactions between strains in multi-strain formulations, which may not be present with single strains, can complicate the interpretation of the observed effects [68]. Therefore, the observed effects in the multi-strain probiotic studies may be enhanced, attenuated, or otherwise altered compared to the single-strain studies, directly contributing to the variability in effect sizes [69]. In addition, the composition of the resident gut microbiome may influence the survival, colonisation and metabolic activity of the introduced probiotic strains, further contributing to the observed variability [70].

Even though the definition of probiotics required an ‘adequate dose’ to achieve a health benefit, no health authority had yet specified this ‘adequate probiotic dose’ [9, 71]. In fact, in clinical research, dose-ranging studies were standard for pharmaceuticals but rare in dietary supplements due to assumed safety [72]. Clinical trials were then conducted according to the ‘maximum tolerable dose (MTD)’ assessments, yet no human study has established an oral MTD for probiotics and as a result, most studies tended to adopt daily doses ranging from 10⁸ to 10¹¹ CFUs [73]. These dosage levels were primarily based on prior research demonstrating their efficacy [73]. Although identifying an optimal dose for a targeted health benefit may have been of scientific interest, it was not considered a fundamental requirement in probiotic research [73, 74]. The International Scientific Association for Probiotics and Prebiotics (ISAPP) recommended a daily probiotic intake of 10⁸ to 10¹¹ CFUs for efficacy, while the International Dairy Federation (IDF) has established application guidelines recommending a daily probiotic intake of 10⁹ CFUs [73, 75]. Similarly, in our study on psychological mental health outcomes, probiotic doses ranged from 10⁶ to 10¹¹ CFUs per day. As for the World Gastroenterology Organisation Global Guidelines (WGOGG) it was suggested that the supplementation of patients with 10¹⁰ CFUs per day of Bifidobacterium longum could result in a reduction of depression scores and improvement of quality of life scores in IBS patients [76]. In line with these findings, a network meta-analysis of 45 articles by Yang et al. [77] found that the combination of Lactobacillus and Bifidobacterium strains at 10⁹ to 10¹⁰ CFUs per day had a beneficial effect on depression while doses above 10¹¹ CFUs per day of Lactobacillus or Bifidobacterium strains alone could successfully alleviate anxiety. A systematic review of twenty-three RCTs involving 1401 patients by Asad et al. [78] highlights the effectiveness of high-dose single-strain probiotics, particularly Bacillus coagulans, Clostridium butyricum, Lactobacillus, and Bifidobacterium, in improving outcomes in patients with depressive disorders. This finding was further supported by a pioneer randomised controlled trial of 60 patients diagnosed with major depressive disorder (MDD), which showed that high-dose, multi-strain probiotic supplementation, added to standard treatment, significantly reduced depressive symptoms [79]. Further corroborating these results, a review conducted by Jach et al. [80], where the authors concluded that the administration of probiotics labeled as psychobiotics, should be at a dose higher than 1 billion CFUs/day, as an adjuvant to antidepressants to improved the effectiveness of antidepressant therapy.

Heterogeneity in intervention duration, ranging from 4 to 16 weeks, may have also contributed to the inconsistency in effect size across the included studies [81]. On one hand, evidence from prior systematic reviews indicates that treatment duration does not significantly modify the effect of probiotic supplementation on cortisol levels, despite substantial heterogeneity across categories [82]. These observations suggest that factors other than intervention length, such as strain specificity, population characteristics, or geographical context, are more likely to account for variability in effect size [44]. On the other hand, shorter interventions may not allow sufficient time for complex processes such as modulation of immune responses, neurotransmitter production and gut-brain signalling to translate into measurable and lasting changes in psychological well-being [83]. In addition, the literature suggests that the method of delivery of probiotics can significantly affect their bioavailability and efficacy [84]. Encapsulated probiotics are designed to protect the bacteria from stomach acid, potentially increasing the number of viable probiotics that reach the gut [85]. However, this protection may come at the cost of potential interactions with the food matrix, which could affect probiotic activity [85, 86]. In contrast, probiotics delivered via fermented foods (such as yoghurt) may have lower survival rates due to stomach acid, but may benefit from synergistic interactions with the food matrix, potentially enhancing their effects [87]. This is particularly relevant in modern work environments, where demanding schedules and disrupted sleep heighten vulnerability to stress, anxiety, and burnout [88].

Probiotics and sleep quality

A limited number of the included studies investigated the effect of probiotics on sleep quality. The results suggest a potential beneficial effect of probiotics on sleep quality, although this difference did not reach statistical significance. Although the precise mechanisms by which probiotics may improve sleep quality are not fully elucidated, emerging knowledge suggests plausible pathways. A growing body of evidence suggests that activation of the hypothalamic-pituitary-adrenal axis may mediate the relationship between work-related psychological stress and impaired sleep quality [89]. Specifically, work-related psychological stress may stimulate the release of corticotropin-releasing factor, which in activation of the hypothalamic-pituitary-adrenal axis leads to impaired sleep quality [90]. Interestingly, new guidelines suggest that the administration of probiotic microorganisms can attenuate the increase in glucocorticoid levels associated with stress, suggesting a potential modulatory role for probiotics within this stress-sleep pathway [91]. In addition, probiotics may affect the production of neurotransmitters such as GABA and serotonin, which are key regulators of the sleep-wake cycle [92]. Furthermore, beneficial bacteria consumption might improve perceived sleep quality by positively impacting broader mood-related traits, potentially independent of stress levels, highlighting the gut-brain axis’s role in regulating both emotional and physiological processes [44]. Despite these promising observations, more research is needed to further characterise the effects of probiotics on sleep parameters and to fully understand the potential underlying mechanisms of action.

Probiotics and serum C-reactive protein levels

Chronic and acute inflammation, as evidenced by increased levels of C-reactive protein, are increasingly recognised as important factors influencing mental health outcomes [93]. Exploring the potential of probiotics to modulate CRP levels is therefore an important opportunity to better understand their wider implication for the overall wellbeing and productivity of the working population. Through our meta-analysis, which included 3 trials and 1024 healthy working adults, we suggest that, based on the available evidence, probiotics do not influence CRP levels in a clinically meaningful way. Consistent with our findings, a meta-analysis by Samah et al. of four RCTs evaluating probiotic interventions in type 2 diabetes mellitus reported no statistically significant effect of probiotics on C-reactive protein levels (standardised mean difference = −0.31, 95% CI: −0.85 to 0.23) [94].

In contrast to our results and the above findings, a recent meta-analysis by Mazidi et al., which included 12 RCTs involving over 1,000 healthy adults, found a statistically significant reduction in serum C-reactive protein levels following probiotic supplementation. The Mazidi meta-analysis examined the effects of different probiotic strains and suggested that the CRP-lowering effect was not specific to any particular probiotic species or strain [93]. In addition, another meta-analysis by Milajerdi et al. [95] examining the effect of probiotic consumption on inflammation in adults also found a significant reduction in serum hs-CRP in both healthy and unhealthy adults [95].

Although our meta-analysis and several other studies suggest that probiotic supplementation may not significantly affect CRP levels, it’s important to recognise that the relationship between probiotics, inflammation and CRP is complex and likely influenced by multiple factors [93]. One possible explanation for the observed lack of effect on CRP could be the dose of probiotics used. Mazidi et al. pointed out that the effects of probiotics on inflammation may be highly dependent on the dosage used [93].

In fact, some in vitro studies have suggested beneficial anti-inflammatory effects at lower doses of probiotics, and state that higher doses could potentially have the opposite effect and even exacerbate inflammation [96]. Firstly, they attribute the reduction in the enzymatic synthesis of hepatic C-reactive protein by the beneficial bacteria to the production of short-chain fatty acids in the colon [97]. Secondly, they suggest that excessive doses of probiotics could potentially promote the induction of inflammatory cytokines such as interleukin-6, which could counteract the anti-inflammatory properties seen at lower doses [98]. In addition to dosage, the specific probiotic strains and combinations used may also be a relevant parameter in determining their effect on CRP and other inflammatory markers [99]. Certain probiotic strains may be more effective than others in modulating inflammation, with Lactobacillus and Bifidobacterium species often showing the most encouraging results [100]. Furthermore, a review by Zhang et al. suggested that the viability of probiotic strains is also an important factor, with heat-killed beneficial bacteria potentially having better anti-inflammatory effects than live cultures [101].

Interestingly, a record by Zhao et al. [102] found that probiotic interventions with multiple strains of Lactobacillus and Bifidobacterium showed the greatest reductions in CRP levels compared to single-strain probiotics or those with other probiotic species. In fact, while some strains have shown beneficial effects in modulating inflammation, others may have limited or even harmful effects, potentially exacerbating inflammatory responses in certain contexts [55]. Thus, the specific probiotic strains used in interventions may have a critical impact on the observed outcomes [103]. The baseline inflammatory status of the study population is another important factor to consider when analysing the effects of probiotics on CRP levels. In healthy individuals, the potential for probiotics to induce a substantial reduction in CRP may be limited as the baseline levels in this population are already within a healthy range reflecting their balanced inflammatory response [104].

Individuals with mental or physical health conditions, such as mood disorders or chronic diseases, often have elevated baseline CRP levels due to underlying inflammation, which may lead to more noticeable reductions in CRP following probiotic interventions [105].

Probiotics and cortisol level

A sensitive indicator of the body’s stress response is cortisol, an important hormone produced by the adrenal glands [106]. Cortisol levels naturally peak in the morning and decline throughout the day, but chronic stress, especially in high-pressure work environments, can disrupt this rhythm, leading to persistently elevated levels [107]. The high cortisol levels negatively impact brain regions like the hippocampus, amygdala, and prefrontal cortex, thereby impairing cognition, emotional regulation, and increasing the risk of anxiety and depression [108]. Our meta-analysis, which included five randomised controlled trials and 437 participants, provided compelling evidence that probiotic supplementation may contribute to stress reduction by significantly reducing cortisol concentrations in healthy working adults. We observed a statistically significant reduction in cortisol levels following probiotic intervention, with a pooled standardised mean difference of −0.26 (95% confidence interval [CI] [−0.45, −0.08]), p = 0.005. This small to moderate effect size observed suggests that the magnitude of cortisol reduction, although statistically significant, should be interpreted with caution. Notably, the lack of heterogeneity among the included studies strengthens the robustness of our findings, suggesting a consistent effect of probiotics on cortisol modulation. Although the mechanisms underlying probiotic effects on cortisol are not fully defined, evidence indicates a key interaction between the gut microbiome, the brain, and the HPA axis [106]. Strains such as Lactobacillus rhamnosus and Bifidobacterium longum can enhance gamma-aminobutyric acid (GABA) synthesis [109], which may modulate HPA axis activity and reduce cortisol output [110]. Additionally, probiotics may attenuate HPA activation through anti-inflammatory actions that lower pro-inflammatory cytokine signalling [111].

Using a rat model of early life stress, Gareau et al. [112], demonstrated that probiotic treatment normalised corticosterone release, the rodent equivalent of cortisol, by regulating HPA axis activity, even in the context of early adversity.

Further supporting our findings, Messaoudi et al. [113], investigated the psychotropic-like effects of a probiotic formulation containing Lactobacillus helveticus R0052 and Bifidobacterium longum R0175 in both rats and humans. Their preclinical work demonstrated anxiolytic-like effects in rats, while the human study showed a significant reduction in perceived psychological distress as measured by the Hopkins Symptom Checklist. Notably, this same trial also revealed a reduction in 24-hour urinary free cortisol levels in the probiotic group, directly supporting our findings of cortisol reduction. Nevertheless, we must acknowledge that not all studies reported the same finding. For example, Kazemi and colleagues [114], in a study of patients with major depressive disorder (MDD), found that probiotics helped relieve depressive symptoms but had no significant effect on urinary cortisol levels.

The observed discrepancies in the effects of probiotics on cortisol levels between studies may be due to a number of factors. First, Probiotic efficacy appears dose- and strain-dependent, as higher doses do not necessarily enhance cortisol reduction [115]. Additionally, individuals with elevated baseline cortisol or HPA axis dysregulation may show greater reductions following intervention. Additionally, variability in cortisol assessment methods and diurnal timing underscores the need for standardized measurement and reporting protocols [116].

Probiotic and oxidative stress markers

Elevated oxidative stress is linked to cognitive decline, inflammation, and neurodegeneration, all contributing to mental health disorders [117]. The modern workplace, characterised by demanding workloads, strict deadlines and interpersonal conflicts, may amplify oxidative stress levels, particularly for individuals in high-stress occupations [118]. Chronic exposure to such occupational stressors can perturb the balance between ROSs’ production and antioxidant defences, leading to increased oxidative stress, which can adversely affect mental well-being and lead to symptoms of anxiety, depression and burnout [8]. Emerging evidence suggests that probiotics may have the ability to alleviate oxidative stress, which presents them as a potential avenue for investigation [119].

We investigated the effects of probiotic supplementation on established oxidative stress biomarkers in healthy workers; however, a quantitative analysis was not possible due to the limited number of studies directly investigating this relationship. The meta-analysis of Musazadeh et al. [120] included studies that assessed a range of antioxidant indicators, including total antioxidant capacity, malondialdehyde (a marker of lipid peroxidation), and antioxidant enzymes such as superoxide dismutase and glutathione peroxidase [120]. Their analysis showed consistent improvements in these markers following probiotic supplementation, suggesting an antioxidant effect, although the analysis of strain-specific effects was limited due to the heterogeneity of the included studies. Similarly, Wang et al. [121] reported consistent improvements in antioxidant status across diverse populations, suggesting that probiotics may modulate systemic oxidative stress pathways. Their review also emphasized the strain-specific nature of these effects: Lactobacillus rhamnosus and certain Bifidobacterium longum strains enhance antioxidant activity through glutathione synthesis and enzyme upregulation, while Streptococcus thermophilus produces exopolysaccharides with notable antioxidant properties [121]. These exopolysaccharides may help reduce oxidative stress by directly neutralising reactive oxygen species [122]. In contrast, Ding et al. [123] reported no significant effects on oxidative stress markers in their systematic review and meta-analysis, which evaluated probiotic supplementation in adults with type 2 diabetes.

Certainly, there is a growing body of evidence in the literature elucidating the strain-specific mechanisms by which probiotics may mitigate oxidative stress [124]. Certain probiotic strains possess their own enzymatic systems capable of directly neutralising reactive oxygen species, such as superoxide dismutase and catalase; Other probiotic strains have the potential to enhance the host’s natural defence mechanisms against oxidative damage by stimulating the expression of endogenous antioxidant enzymes, such as glutathione peroxidase and superoxide dismutase, thereby effectively reducing the burden of free radicals [121]. Additionally, probiotics can chelate metal ions, such as iron, which can form highly reactive hydroxyl radicals through the Fenton reaction [121, 125]. These hydroxyl radicals are a potent type of reactive oxygen species that can contribute to oxidative stress and cell damage [121]. By chelating iron and limiting its availability for the Fenton reaction, probiotics can effectively inhibit the generation of these harmful reactive oxygen species, thereby reducing overall oxidative stress [121, 126–128] Probiotic supplementation can also displace deleterious bacteria in the gut, limiting their ability to generate reactive oxygen species [121]. At the same time, it can shape the gut microbiome towards a healthier profile, favouring antioxidant-producing species that may reduce oxidative stress [121, 128, 129].

Strengths and limitations

We conducted an extensive literature search across multiple databases without language restrictions, and a rigorous screening process was implemented by experienced researchers to ensure the inclusion of studies, even when mental health outcomes were not the primary endpoint. Nonetheless, drawing definitive conclusions about the efficacy of probiotic supplementation remains challenging. First, the potential for publication bias cannot be fully excluded, particularly due to the omission of grey literature sources. Future meta-analyses with larger datasets should incorporate grey literature and employ analytical tools such as funnel plots and Egger’s tests to better assess and mitigate this bias.

Second, the limited number of eligible studies (many with small sample sizes) reduces the statistical power and generalizability of our findings. Participant heterogeneity and the lack data of adherence in some studies, despite strict inclusion criteria, further complicates interpretation. Variations in baseline gut microbiome composition, which may influence responsiveness to specific probiotic strains, remain insufficiently characterized. The absence of long-term follow-up data and the unclear clinical significance of small SMDs should also be noted. Future research should prioritize larger, well-characterized cohorts and standardized reporting of participant attributes to enable subgroup analyses and clarify interindividual variability in probiotic effects. Third, substantial variability in intervention protocols, including probiotic strains, dosages, durations, and delivery methods, restricted cross-study comparability and the strength of pooled conclusions. These inconsistencies underscore the need for standardized supplementation protocols to enhance reproducibility and facilitate robust meta-analytic synthesis.

Fourth, an additional limitation concerns the gender imbalance across included studies. This overrepresentation may limit the generalizability of our findings to broader working populations, particularly male-dominated sectors. It may also partially explain discrepancies between our results and prior systematic reviews that examined more gender-balanced cohorts or included larger proportions of male participants. Future trials should include more gender-representative samples to ensure wider applicability of outcomes.

Finally, potential conflicts of interest warrant attention. Three included studies [31, 33, 36] involved authors affiliated with probiotic manufacturers, raising concerns regarding possible bias in study design, data interpretation, and reporting. Ensuring independent funding and transparent disclosure in future trials will be essential to strengthen the evidence base and uphold scientific integrity in this field.

Conclusion

This systematic review and meta-analysis demonstrated that probiotic supplementation is associated with a statistically significant, though modest, improvement in subclinical psychological symptoms, corresponding to small-to-medium standardized mean differences, specifically for depression, anxiety, and perceived stress, among healthy working adults. This finding is supported by the observed, statistically significant reduction in circulating cortisol levels, indicating a robust biological mechanism for improved mental well-being, even with a relatively marginal psychological effect. However, no significant impact was found on C-reactive protein. Given the substantial heterogeneity across the included studies regarding probiotic strains, dosages, and intervention durations, establishing definitive clinical guidelines remains challenging. Future research must prioritize well-controlled, long-term trials to definitively elucidate the sustained efficacy and dose-response relationship of probiotic interventions on mental health and occupational well-being in non-clinical populations, clarifying the optimal daily administration (e.g., between 10⁸ and 10¹¹ CFUs per day, or higher) for specific strains.

Abbreviations

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

PROSPERO

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

MeSH

Medical Subject Headings

BMI

Body mass index

ID

Identification

SMD

Standardised Mean Difference

ITT

Intention To Treat

RoB

Risk of Bias

DASS-42

Depression Anxiety and Stress Scale

PSQI

Pittsburgh Sleep Quality Index

GABA

Gamma-aminobutyric Acid

ACTH

Adrenocorticotropic Hormone

CRP

C-Reactive Protein

HPA axis

Hypothalamic Pituitary Adrenal axis

MDD

Major Depressive Disorder

ROS

Reactive Oxygen Species

CFU

Colony Forming Unit

MTD

Maximum Tolerable Dose

ISAPP

International Scientific Association for Probiotics and Prebiotics

IDF

International Dairy Federation

WGOGG

World Gastroenterology Organisation Global Guidelines

Authors’ contributions

SBF, HK, JM, and AC conceptualized the study, developed the research question and search strategy, conducted the analysis, and drafted the initial manuscript. ND, RG, NZ, and AS refined the search strategy, developed the data extraction form, and contributed to the editing and final approval of the manuscript. IA, IH, MM, and OE further revised the data extraction form and participated in the editing and final approval of the article. All authors made substantial contributions to this work.

Funding

No funding was used in this study.

Data availability

The data are contained within the article.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.