Effect of Psychobiotics on Psychometric Tests and Inflammatory Markers in Major Depressive Disorder: Meta-Analysis of Randomized Controlled Trials with Meta-Regression

Probiotics were shown to act positively on gut–brain axis signaling. We aimed to assess the effect of the administration of a new class of probiotics—psychobiotics—using data from individual psychometric scales, markers of the immune system and neuroactive metabolites. Medical databases were searched from database inception until 22 April 2021 for randomized clinical trials in clinically proven Major Depressive Disorder (MDD) patients treated with either probiotics or placebo reporting any psychometric score (PROSPERO registration number: CRD42021253024). Ten studies with 705 randomized participants and 603 analyzed were included. The mean age of individuals was 38.43 ± 12.1 years, predominantly women (n = 461, 76.45). The mean study duration was 48.8 ± 12.3 (range = 28–62) days. The dosage ranged between 1 × 109 to 2 × 1010 colony forming units (CFU)/day. We found that probiotics might alleviate symptoms of MDD; endpoint data (pooled scores): SMD = −0.292, 95%CI = −0.577 to −0.007, p < 0.044; change scores (BDI): SMD = −0.482, 95%CI = −0.854 to –0.109, p < 0.011; DM = −4.848, 95%CI = −8.559 to −1.137, p < 0.01. The therapy tended to be more effective with time of psychobiotic supplementation (coefficient = −0.12, SE = 0.06, Z = −1.84, p = 0.06) and in men (% of females: coefficient = 0.1, SE = 0.06, Z = 1.78, p = 0.07). Psychobiotics have great potential in the treatment of MDD. However, no specific strain/strains, dosage or duration of treatment can currently be recommended.


Introduction
Major Depressive Disorder (MDD) affects approximately 300 million people worldwide and is a common cause of disability and roughly 800,000 suicides per year [1]. Approximately 30% of patients with MDD do not respond to monoaminergic antidepressants [2], suggesting that other biological pathways are involved in MDD etiology. These mechanisms include, i.a., subclinical inflammation [3], hypothalamic-pituitary (HPA) axis dysregulation [4] and altered signaling of neurotrophic growth factors [5]. One very promising hypothesis of MDD pathogenesis is the gut-brain axis (GBA) dysfunction [6], with gut microbiota as a key player. The microbiota was shown to regulate different functions in the central nervous system (CNS), i.a., the promotion of neuropeptides synthesis, regulation of the HPA axis, production of neurotransmitters and tryptophan metabolism [7,8].
Only a few human and animal studies proved the association between gut microbiota and depression [9]. It is hypothesized that the bacterial taxonomic changes observed in patients with MDD are associated with their proinflammatory activity, reduced shortchain fatty acids production, impaired intestinal barrier integrity, skewed neurotransmitter

Search Results
The initial search yielded 386 hits. At first, we excluded 365 studies as for being duplicates and/or after evaluation on the title/abstract level. No additional articles were identified via hand search. Finally, 21 full-text articles were reviewed. Of those, 11 did not fit the inclusion criteria. Primary reasons for exclusion were: abstracts for full-text studies (N = 3), no clinically well diagnosis of MDD (N = 3) and other than a randomized controlled trial (RTC) design (N = 2). We excluded studies being review, with no outcome of interest and no intervention, one per each reason (N = 3). Finally, the search yielded 10 studies that were included in the meta-analysis ( Figure 1).
Bloating (5), Nausea (10).  Influence of probiotics on the expression of inflammation-related genes/Decreased expression of IL-6. $ -the same cohort * the same cohort;ˆ1 patients dropped out due to insufficient volume of collected serum, # -significant in comparison to placebo; PRO-psychobiotic; PBO-Placebo; ND-no data; HAM-D 17-item-Hamilton rating scale for depression; HOMA-IR-homeostasis model assessment of insulin resistance; hs-CRP-high sensitivity C-reactive protein; GSH-total glutathione; IL-interleukin.

Endpoint Data
Using random-effects weights, the standardized difference in means (SDM) for symptomatology of depression evaluated by pooled Beck Depression Inventory (BDI) and Hamilton Depression Rating Scale (HAMD) scores at the endpoint was −0.292 with a 95% confidence interval of −0.577 to −0.007 (z = −2.01, p < 0.044; Figure 2). Using random-effects weights, the standardized difference in means (SDM) for symptomatology of depression evaluated by pooled Beck Depression Inventory (BDI) and Hamilton Depression Rating Scale (HAMD) scores at the endpoint was −0.292 with a 95% confidence interval of −0.577 to −0.007 (z = −2.01, p < 0.044; Figure 2).       There were some covariates associated with study-level effects of probiotics on depression symptoms for SDM and DM effect sizes and psychometric score at the endpoint. The significant association between effect size (SDM) and probiotic strains (Formula R as the reference group) was found: probiotic strain "Formula 1" coefficient =  There were some covariates associated with study-level effects of probiotics on depression symptoms for SDM and DM effect sizes and psychometric score at the endpoint. The significant association between effect size (SDM) and probiotic strains (Formula R as the reference group) was found: probiotic strain "Formula 1" coefficient = −0.69, standard error (SE) = 0.27, Z = −2.55, p = 0.01; Probiotic "Formula 2" coefficient = −0.38, SE = 0.29, Z = −1.32, p = 0.18. This covariate explained 100% of the variance in the effect size ( Figure 6). Other covariates were not linked to the estimated effect size.  . Regression for SDM on strains used in trials. References [26,30], Formula 1 [23,28], Formula 2 [21]. Figure 6. Regression for SDM on strains used in trials. References [26,30], Formula 1 [23,28], Formula 2 [21].

Discussion
The present review included ten randomized clinical trials [21][22][23][24][25][26][27][28][29][30] that evaluated the effectiveness of probiotics in MDD treatment measured by psychometric scales. In five of them, the mechanism of probiotic action was also examined. The antidepressant efficacy of probiotics in MDD was demonstrated in three studies [21,23,25] and the improvement of cognition in two papers [28,30]. Saccarello et al. [29] reported improvements in symptoms of depression, anxiety and cognition along with somatic components, whilst Reininghaus et al. [26] did not report on any beneficial impact of probiotics in the course

Discussion
The present review included ten randomized clinical trials [21][22][23][24][25][26][27][28][29][30] that evaluated the effectiveness of probiotics in MDD treatment measured by psychometric scales. In five of them, the mechanism of probiotic action was also examined. The antidepressant efficacy of probiotics in MDD was demonstrated in three studies [21,23,25] and the improvement of cognition in two papers [28,30]. Saccarello et al. [29] reported improvements in symptoms of depression, anxiety and cognition along with somatic components, whilst Reininghaus et al. [26] did not report on any beneficial impact of probiotics in the course of MDD. The meta-analysis showed that particular combinations of probiotics or specific species and strains appear to be beneficial in MDD in terms of their effect on the BDI and HAMD pooled psychometric scales, but it is not possible to draw definitive and conclusive conclusions about their effectiveness. The beneficial effect of probiotics is evident in the analysis using change scores in the subjective BDI scale, the magnitude of which is at the middle level. Additionally, the effect size, DM = 4.85, may be of small clinical importance. The results on the BDI and HAMD psychometric scales measured at the endpoint do not confirm the high effectiveness of probiotics due to the small effect size, small statistical significance and weak clinical effect. In addition, serious heterogeneity was observed between the studies in which the HAMD scale was assessed. From among the meta-analyzes published so far, only two works [16,17] analyzed the effectiveness of probiotics in patients with MDD without comorbidities. A recently published updated meta-analysis [17] found a beneficial effect of probiotics in patients with MDD receiving antidepressants but not with probiotic monotherapy. In our study, it was not possible to distinguish such a subgroup (as we subgrouped studied by the psychometric scales), but it should be emphasized that only one study included in the meta-analysis described the population of treatment naïve patients [30]. Moreover, the inference is made more difficult by the fact that half of the analyzed studies took place in Iran, and five of them were carried out on two patient cohorts, which does not ensure adequate representativeness of the analyzed studies.
In this meta-analysis, by means of meta-regression, we failed to demonstrate the efficacy of a particular strain or a combination of different strains. It should be emphasized, however, that probiotic cocktails containing the strains that were used in the works of Chahwan et al. [30], Reininghaus et al. [26] and Reiter et al. [27] differed from each other and their pooling in the metaregression may be a source of error, which, however, does not affect the fact that there is no evidence so far that any particular probiotic strain or combinations of such strains can be recommended in patients with MDD based on the results of the meta-analysis instead of individual RCTs.
The results of the meta-regression suggest that the duration of use might be associated with their greater effectiveness in patients with MDD. Such observation could be explained by the time necessary for changes in the intestinal microbiota along with the administration of probiotics. However, it is difficult to prove such a thesis because only two studies analyzed gut microbiota [26,30], with one study reporting microbiota changes [26]. There is, however, no current contention on whether probiotic treatments could/should successfully alter microbiota composition [31,32]. Nevertheless, probiotics were reported to influence bacterial gene expression and cause anti-inflammatory effects regardless of the influence on microbiota composition [33]. This could explain the cognitive function improvement after probiotic administration. Alternatively, multi-strain probiotics (Ecologic Barrier, Winclove Probiotics, Amsterdam, The Netherlands) can improve gut barrier function in vitro [34] and in humans [35], which can be also related to a decrease in systemic inflammation, which can improve symptoms of MDD.
We observed that the effectiveness of probiotics has a tendency to be inversely proportional to the percentage of women participating in the study. Epidemiological studies have shown that depressive disorders occur approximately two or three times more frequently in women than men [36,37]. Moreover, the clinical characteristics and treatment outcomes of depressive disorders in women are different from those in men [37,38]. Interestingly, the composition of the microbiota in men and women suffering from MDD [39] differ from one another, which might stand for different probiotic efficacy.
In mechanistic studies, Rudzki et al. and Kazemi et al., 2019a [21,28] observed a decreased blood kynurenine concentration. Some kynurenine catabolites may have a role in patients with MDD due to its neurotoxic and neurodegenerative effects [40]. Metagenomic analysis showed a disorder of tryptophan synthesis in patients with MDD [41][42][43][44]. It seems, therefore, that probiotic administration may have influenced tryptophan metabolism. Other analysis [22] found a clinically significant decrease in urine cortisol concentration, Akasheh et al. [25] and Reiter et al. [27] observed anti-inflammatory effects of probiotics which, however, have not been confirmed in other studies [28,45]. Heidarzadeh-Rad et al. [24] reported an increased brain-derived neurotrophic factor (BDNF) level, which was shown to correlate with antidepressant response in MDD.
Based on experimental studies and results mentioned above, it can be concluded that probiotics have the potential to influence various mechanisms of etiopathology of depression involving inflammation, neurotransmitters and the hypothalamic-pituitaryadrenal (HPA) axis. However, the meta-analysis did not confirm the effectiveness of the use of probiotics in regulating the parameters associated with the HPA axis (cortisol level), inflammation (interleukins, TNF) and tryptophan degradation pathway (kynurenine). In another meta-analysis, Amirani et al. [19] reported that taking probiotics by patients with different psychiatric disorders (not only MDD) had beneficial effects on C-reactive protein (CRP), IL-10 and malondialdehyde (MDA) levels, but it did not affect other markers of inflammation (TNF-alpha, IL-1B) and oxidative stress. However, due to the ambiguous results of the research and the high heterogeneity of the studied populations, it can be said that the clinical mechanism of the action of probiotics in improving the symptoms of MDD remains the subject of speculation. Finally, it should be stated that taking probiotics is well tolerated, safe and poses no risk in patients with MDD. Additionally, the intervention is not associated with significant side effects.
Our systematic review and meta-analysis has many strengths. We applied a rigorous and repeatable methodology. Our search strategy was described in detail; moreover, unlike other studies, we qualified the data obtained in RCT in patients with medically confirmed MDD, and we did not take into account patients with comorbidities and healthy patients who were assessed for depressive symptoms. Finally, we conducted subgroup analyses to assess the treatment effect depending on the psychometric scale used, and we also conducted a risk of bias analysis, reasons for study discontinuations and a very detailed analysis of adverse effects. We also performed the Egger's test and meta-regression analysis.
It should also be emphasized that our systematic review and meta-analysis also have significant limitations. The number of studies and the size of the study groups are both small. The methodological heterogeneity is significant; the five analyzed studies concern two small cohorts; finally, Iranian patients are over-represented in these studies. Mechanistic research is quite sparse. Researchers have rarely analyzed the gut microbiota, and the immunome or metabolome has not been analyzed at all.

Search Strategy and Inclusion Criteria
There were two independent authors (KSZ and PL) who searched PubMed/Embase/ Cinahl and Web of Science from database inception until 22/04/2021 without language restriction for RCTs that compared adjunctive probiotics with placebo to counteract depression symptomatology.
Cinahl/Web of Science-(MDD OR depression) AND (probiotic OR psychobiotic OR bifidobacterium OR lactobacillus OR synbiotic) AND (RCT OR random OR placebo).