Synbiotic Therapy Prevents Nosocomial Infection in Critically Ill Adult Patients: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials Based on a Bayesian Framework

Background: The efficacy of synbiotics, probiotics, prebiotics, enteral nutrition or adjuvant peripheral parenteral nutrition (EPN) and total parenteral nutrition (TPN) in preventing nosocomial infection (NI) in critically ill adults has been questioned. We conducted a systematic review and network meta-analysis (NMA) of randomized controlled trials (RCTs) to evaluate and rank the effectiveness of these therapies on NI amongst critically ill adults. Methods: Four electronic databases were systematically searched up to June 30, 2019 for RCTs comparing the administration of probiotics, prebiotics, synbiotics, EPN and TPN in critically ill adults. The primary outcome was NI. The relative efficacy of all outcomes was determined by a Bayesian framework with random effects NMA. We estimated the odds ratio (OR) and mean difference (MD) and ranked the comparative effects of all regimens with the surface under the cumulative ranking probabilities. The study has been registered on PROSPERO (CRD42019147032). Results: Fifty-five RCTs (7,119 patients) were identified. Primary outcome showed that synbiotics had the best effect in preventing NI than EPN (OR 0.37; 95% CrI 0.22–0.61), probiotics followed (OR 0.52; 95% CrI 0.34–0.77), whereas TPN significantly increased NI (OR 2.29; 95% CrI 1.48–3.67). Subgroup analysis showed that TPN significantly increased NI in intensive care unit (ICU) patients (OR 1.57; 95% CrI 1.01–2.56) and severe acute pancreatitis (SAP) patients (OR 3.93; 95% CrI 1.74–9.15). Secondary outcomes showed that synbiotics were more effective in preventing hospital-acquired pneumonia (HAP) (OR 0.34; 95% CrI 0.11–0.85), catheter-related bloodstream infection (OR 0.08; 95% CrI 0.01–0.80), urinary tract infection (OR 0.27; 95% CrI 0.08–0.71) and sepsis (OR 0.34; 95% CrI 0.16–0.70) than EPN. Amongst the treatments, probiotics were most effective for shortening the mechanical ventilation duration (MD −3.93; 95% CrI −7.98 to −0.02), prebiotics were most effective for preventing diarrhea (OR 0.24; 95% CrI 0.05–0.94) and TPN was the least effective in shortening hospital length of stay (MD 4.23; 95% CrI 0.97–7.33). Conclusions: Amongst the five therapies, synbiotics not only prevented NI in critically ill adults but also demonstrated the best treatment results. By contrast, TPN did not prevent NI and ranked last, especially in ICU and SAP patients. Take-Home Message: Nosocomial infection is a leading cause of mortality in critically ill patients in the ICU. However, the efficacy of synbiotics, probiotics, prebiotics, enteral nutrition or adjuvant peripheral parenteral nutrition and total parenteral nutrition in preventing nosocomial infection in critically ill adults has been questioned. The network meta-analysis provides evidence that amongst the five therapies, synbiotics not only prevented NI in critically ill adults but also demonstrated the best treatment results. By contrast, TPN did not prevent NI and ranked last, especially in ICU and SAP patients. The results of this study will provide a new scientific basis and a new idea for the debate on the efficacy of synbiotics and other treatments in the improvement of prognosis in critically ill adult patients. Tweet: Synbiotic prevents nosocomial infection in critically ill adults, while total parenteral nutrition has the adverse curative.

Background: The efficacy of synbiotics, probiotics, prebiotics, enteral nutrition or adjuvant peripheral parenteral nutrition (EPN) and total parenteral nutrition (TPN) in preventing nosocomial infection (NI) in critically ill adults has been questioned. We conducted a systematic review and network meta-analysis (NMA) of randomized controlled trials (RCTs) to evaluate and rank the effectiveness of these therapies on NI amongst critically ill adults.
Methods: Four electronic databases were systematically searched up to June 30, 2019 for RCTs comparing the administration of probiotics, prebiotics, synbiotics, EPN and TPN in critically ill adults. The primary outcome was NI. The relative efficacy of all outcomes was determined by a Bayesian framework with random effects NMA. We estimated the odds ratio (OR) and mean difference (MD) and ranked the comparative effects of all regimens with the surface under the cumulative ranking probabilities. The study has been registered on PROSPERO (CRD42019147032).
Conclusions: Amongst the five therapies, synbiotics not only prevented NI in critically ill adults but also demonstrated the best treatment results. By contrast, TPN did not prevent NI and ranked last, especially in ICU and SAP patients.
Take-Home Message: Nosocomial infection is a leading cause of mortality in critically ill patients in the ICU. However, the efficacy of synbiotics, probiotics, prebiotics, enteral nutrition or adjuvant peripheral parenteral nutrition and total parenteral nutrition in preventing nosocomial infection in critically ill adults has been questioned. The network meta-analysis provides evidence that amongst the five therapies, synbiotics not only prevented NI in critically ill adults but also demonstrated the best treatment results. By contrast, TPN did not prevent NI and ranked last, especially in ICU and SAP patients. The results of this study will provide a new scientific basis and a new idea for the debate on the efficacy of synbiotics and other treatments in the improvement of prognosis in critically ill adult patients.
Probiotics are live microorganisms that exert beneficial effects by protecting against pathogens, improving intestinal barrier function and inducing host immunomodulation (16). Prebiotics are a substrate that are selectively utilized by host microorganisms maintaining gut homeostasis and improving health outcomes (17)(18)(19)(20)(21)(22)(23). Enteral nutrition or adjuvant peripheral parenteral nutrition (EPN) and total parenteral nutrition (TPN) have the functions of protecting the intestinal barrier and providing adequate nutrient substrates, respectively (24). Therefore, all above therapies can partially improve intestinal microbiota dysbiosis, and are widely used in the treatment of NI in critically ill adults (17,25).
Many previous conventional meta-analyses have already examined the risks and benefits of probiotics or synbiotics compared with EPN in critically ill adults (38)(39)(40)(41)(42). However, all these meta-analyses were restricted to pairwise comparisons, and only the pooled risk ratio (RR) or odds ratio (OR) were calculated. There was heterogeneity between the included trials, and the relative merit of candidate therapies could not be informed through a direct comparison. Network meta-analyses (NMAs) can not only address this limitation but also improve precision by combining direct and indirect estimates (43). Therefore, this systematic review and NMA aimed to evaluate and rank probiotics, prebiotics, synbiotics, EPN and TPN to determine their effects on improving NI of critically ill adult patients. The results of this study will provide a new scientific basis for the debate on the efficacy of synbiotics and other treatments in the improvement of prognosis in critically ill adult patients.

Approval
This literature was written according to the Preferred Reporting Items for Systematic Review and Meta-analyses (PRISMA) Statement Extension Statement (44). This study was registered on the international prospective register of systematic reviews (PROSPERO CRD42019147032).

Exclusion Criteria
The trial did not report outcome variables. The trial was a duplicate publication.

Search Strategy and Study Selection
We conducted a systematic literature search for clinical trials in Pubmed, Embase, Cochrane (CENTRAL) and Web of Science electronic medical databases until June 30, 2019. There was no language restriction. The specific search terms were used for each database, and the details of the search strategy were modified with a combination of relevant terms as proposed by Cochrane for systematic reviews of RCTs (45). The following MeSH terms were used to search for relevant literature: "critically ill" OR "synbiotic" OR "probiotic" OR "prebiotic" OR "enteral nutrition" OR "parenteral nutrition" OR "nosocomial infection" combined with RCTs.
Five reviewers selected studies for inclusion by screening the titles and abstracts of the literature independently. Thereafter, they reviewed the full texts carefully according to the inclusion and exclusion criteria to determine the final inclusion of articles. Any discrepancies between reviewers were resolved by a consensus after a discussion with a sixth reviewer.

Definition of Interventions
Probiotics are live microorganisms that may confer health benefits on the host when administered in adequate amounts (16,17). Prebiotics are substrates that are selectively utilized by host microorganisms and confer a health benefit (16,18). By contrast, synbiotics are composed of probiotics and prebiotics (Supplementary File 3). The US Centers for Disease Control and Prevention (CDC) National Healthcare Safety Network (NHSN) criteria (46) were used to diagnose NI including HAP, VAP, BSIs, CRBSI, UTI, intraabdominal infection, gastroenteritis system infection and surgical site infection (Supplementary Table 2

.3).
We used definitions of diarrhea as defined by the authors in their original articles. From all trials, we combined hospital mortality where reported. If the mortality time frame was not specified as either ICU or hospital, it was presumed to be the latter.

Data Extraction
For duplicate studies, we included only the research with the most informative and complete data. Five investigators extracted independently all the available data from each study. These data included characteristics of study, details of patients enrolled, type and dose of intervention and details of primary and secondary outcomes. Disagreements among the three investigators were resolved by a consensus after discussing with a sixth reviewer.

Assessment of Risk of Bias (ROB) and Quality
We assessed each included studies' ROB in accordance with the Cochrane collaboration risk of bias tool (45). A summary of the ROB was documented as low, unclear or high. Studies were classified as having low ROB if none was rated as high ROB, and three or less were rated as unclear risk. Studies had moderate ROB if one was rated as high ROB or none was rated as high ROB but four or more were rated as unclear risk. All other cases were assumed to pertain to high ROB.
Additionally, we assessed the certainty of evidence contributing to network estimates with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system (high, moderate, low and very low) (49).

Quantitative Data Statistical Analysis
All data were conducted according to the Cochrane Handbook. In pairwise meta-analysis and NMA, dichotomous and continuous variables were analyzed using OR and mean differences (MD), respectively.
The study effect sizes were assessed using a Bayesian framework with a random effects NMA model (50,51). Dichotomous outcomes used the binomial likelihood, and continuous outcomes used the normal likelihood. Four Markov chains were adopted for initial value setting. The initial update iteration number of the model and the continuous update iteration number were set as 20,000 and 50,000, respectively. The first 20,000 annealing times were used to eliminate the influence of the initial value, and sampling was started from 20,001 times. The initial and continuous iteration numbers of the model increased if the convergence of models was not satisfactory. A potential scale reduction factor approaching 1 indicated that the model convergence was satisfactory (52).
The treatment for each outcome was ranked by using the surface under the cumulative ranking curve (SUCRA) (53).
Heterogeneity variance was considered to measure the extent of a cross-sectional study and within-comparison variability on treatment effects. I 2 < 25% and I 2 > 75% indicate low and high heterogeneity, respectively (54)(55)(56). Statistically significant heterogeneity was set at I 2 > 50%, and the sources of heterogeneity were discussed.
A statistical evaluation of inconsistency was assessed by the design-by-treatment test (55,57) and node splitting  Frontiers in Medicine | www.frontiersin.org (52). Inconsistencies were found between direct and indirect comparison evidence when P < 0.05.
The transitivity assumption underlying NMA was evaluated by comparing the distribution of clinical and methodological variables that could act as effect modifiers across treatment comparisons (53,58).
This study evaluated whether treatment effects for the primary outcome are robust in subgroup analyses by using ICU patients, MV patients, SAP patients, trauma patients, initial time of nutrition therapy, doses, study year, and quality. In view of the fact that European Society for Clinical Nutrition and Metabolism (ESPEN), Society of Critical Care Medicine (SCCM), and American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.) recommend that the initial time of early EN therapy is within 48 h (24, 25), we divided the subgroup of initial nutritional therapy into two groups: within 48 h and beyond 48 h. The average number of obligate anaerobes of normal people was around 10 [log10 colony-forming units (CFUs)/g of feces] (59)(60)(61). Therefore, we defined the dose of probiotics that was >2 × 10 10 CFU per day as high dose and the rest as moderate to low doses.
The sensitivity of our conclusions was evaluated by analyzing only datasets of studies with high quality.
All statistical analyses were performed with Review Manager 5.3, stata (version 14.0) and R software (version 3.6.1). Network plots and comparison-adjusted funnel plots of NMA were drawn by Stata. NMAs of all outcomes were duplicated using the Netmeta 1.1-0 package in R. Bayesian MCMC simulations were performed by means of JAGS software (gemtc 0.8-2 and rjags 4-10 package) in R. Graphs of SUCRA were obtained using the ggplot2 3.2.1 package in R.

Search Results and Characteristics of the Studies
The searches identified 7,468 articles, and 731 potentially eligible articles were retrieved in full text. Overall, 55 RCTs (comprising    7,119 patients) from 24 countries all over the world carried out between 1995 and 2019 were included (Figure 1). A total of 49 articles were published in English, 5 were in Chinese and 1 was in Spanish. Twenty-four (45%) of 55 trials recruited patients from Europe, 23 (42%) from Asia, 6 (15%) from the America and 2 (3%) from Oceania. Sample sizes varied greatly from 17 to 2410, with a mean of 60 participants (SD = 53). The mean age was 53 years old (SD = 12) for both men and women. Of these participants, 4,358 (61%) of 7,119 of the sample population were male. Eleven (20%) of 55 studies randomly assigned participants to three or more groups. Nine (16%) of 55 studies were multi-center studies, 32 (58%) of 55 studies were double-blind studies and 21 (38%) were open-label studies. Mixed diseases in ICU were the most included diseases, followed by MV support, patients with SAP, severe multiple trauma, victims of brain trauma alone and severe burns. Twenty seven (49%) of 55 studies were of high quality. Nineteen (35%) of 55 studies were of moderate quality (Figures 2, 3). A description of the included studies, interventions, and outcomes is presented in Tables 1-3. The details of the design, management description and antibiotics are shown in Supplementary File 2.

Primary Outcome
The primary analysis was based on the 43 studies comprising 6,215 patients. Figure 4 displays the network of eligible comparisons for NI. All treatment had at least one EPNcontrolled trial. Only synbiotic therapy was not directly compared with probiotic and TPN therapy in the network.  Figure 5 shows the SUCRA ranking curve of NI. Synbiotic treatment was the best choice in preventing NI, whereas TPN was the worst.

Secondary Outcomes
The network of eligible comparisons for secondary outcomes is presented in Supplementary Files 5, 6. Figure 6 presents the results of NMA for secondary outcomes. In terms of improving the efficacy of HAP, CRBIS, UTI and sepsis, synbiotic therapy was more effective than EPN, and the results of the network were OR 0.34; 95% CrI 0.

Direct Meta-Analysis
The forest plot of the pairwise and network effect estimate on NI is shown in Figure 5. The detailed results of all outcomes in pairwise meta-analysis are shown in Supplementary Files 5, 6.         Network Heterogeneity, Inconsistency, and Transitivity In the analysis of inconsistency (Supplementary File 8), there was no global inconsistency in all outcomes except diarrhea (p = 0.0018). Inconsistencies were found between direct and indirect comparisons of probiotic therapy and EPN for NI (p = 0.04143), synbiotic and prebiotic therapy for CRBIS (p = 0.03569), synbiotic therapy and EPN for CRBIS (p = 0.04404), prebiotic therapy and EPN for CRBIS (p = 0.02783), synbiotic and prebiotic therapy for UTI (p = 0.04033), synbiotic therapy and EPN for UTI (p = 0.03591), prebiotic therapy and EPN for UTI (p = 0.04071), probiotic and prebiotic therapy for diarrhea (p = 0.01030), probiotic therapy and EPN for diarrhea (p = 0.01008), prebiotic therapy and EPN for diarrhea (p = 0.01060), and probiotic therapy and TPN for hospital LOS (p = 0.04520).
In the assessment of transitivity (Supplementary File 9), most of the comparisons had similar mean age, but there were a few comparisons with relatively low or high age. Meta-regressions of mean age did not show that they affected the network estimates, although results from such analyses might suffer from ecological bias.

Subgroup and Sensitivity Analyses for Primary Outcome
Subgroup analysis of the diseases ( Table 5) revealed a significant effect on the therapeutic effect of synbiotic therapy except MV   CrI 0. 51-3.87). In addition, we found that the heterogeneity and consistency in different disease subgroups were not statistically significant. Amongst RCTs over the last 10 years, high-quality studies and doses were used in our NMA. They were found to have no material impact on the relative treatment effects (Supplementary File 13). The sensitivity analysis was evaluated based on highquality studies, and the results did not change substantially (Supplementary File 14).

Risk of Bias Assessments and Grade for the Primary Outcome
In summary (Supplementary File 4), 1 (2%) of 55 rials was rated as high risk of bias, 23 (42%) trials were deemed moderate and 31 (56%) were considered low. We did not find publication bias for the network of outcomes, except duration of MV, hospital and ICU LOS (Supplementary File 10).
GRADE judgments for primary outcome were assessed and reported in Table 6. The certainty of evidence for the relative treatment effects of NI varied. It was high and moderate for most of the comparisons involving synbiotics, probiotics and prebiotics and low for most comparisons involving EPN and TPN. When subgroup analysis was performed, the GRADE between all comparisons and ranking of treatment was raised to at least moderate. Details of GRADE for secondary outcomes are presented in Supplementary File 11.

DISCUSSION
This study was based on the analysis of 55 RCTs enrolling 7,119 patients. Results indicated that synbiotic therapy was the best regimen in preventing NI in critically ill patients, while TPN exerted adverse curative effects amongst all the studied treatments. The sensitivity analyses for NI were consistent with the previous conclusions. Subgroup analysis based on diseases did not show significant heterogeneity between the included trials, and GRADE was moderate or high. These results further confirmed that the model was relevant and robust, making it applicable for use in clinical practice. Moreover, this analysis found that synbiotic therapy was the best regimen in improving HAP, CRBIS, UTI and sepsis. Probiotic and prebiotic treatments were the best regimens in shortening the duration of MV and preventing diarrhea, respectively. TPN was the worst in prolonging the hospital LOS. Notably, this study differed from others in that it found no evidence that synbiotic therapies could reduce hospital and ICU mortality in critical patients (109). The mortality of critically ill patients was influenced by several complex risk factors (110). Probiotic and prebiotic therapy could not be fully absorbed by critically ill patients, so they may not have strong enough effects to reduce hospital and ICU mortality. Moreover, probiotic therapy did not significantly influence other clinical endpoints such as CRBIS, diarrhea and hospital LOS.
Results of subgroup analysis for the primary outcome were as follows. Firstly, subgroup analysis in different diseases showed that synbiotic therapy was the best treatment to improve NI in ICU patients. Conversely, TPN aggravated NI in ICU and SAP patients. These findings were consistent with the conclusions from NMA, thereby eliminating the effect of disease heterogeneity on the NMA outcome. Here, we focused on whether ICU patients can benefit from synbiotics. In addition, previous double-blind RCT and meta-analysis showed that TPN was associated with NI in ICU and SAP patients, which was consistent with the findings of this study. TPN therapy in ICU and SAP patients should be shortened as much as possible (25). Secondly, subgroup analysis in studies over the last 10 years and high quality showed that synbiotic therapy prevented NI, while TPN did not. These results were consistent with the standard analysis, including all studies in NMA, further confirming the robustness of the model and avoiding heterogeneity of the model. Thirdly, subgroup analysis in dosages of synbiotics showed no difference in the prevention of NI between the different doses. However, administered excessive synbiotic therapy not only failed to improve NI but also led to more infectious complications (16,17). Hence, administered synbiotics in accordance with physiological requirements should be advocated to reduce the incidence of infectious complications. Fourthly, the subgroup of MV patients analysis showed that probiotic therapy can prevent NI. Only 3 out of a total of 12 studies administer synbiotics as the main intervention, and the patients involved were <10% of the total patients in this subgroup. Therefore, the power did not suggest that synbiotics can prevent NI. Finally, by adjusting the risk of NI and mortality through the initial nutrition therapy time, we found that synbiotics were associated with a reduction in NI among patients who were administered nutrition therapy within 48 h, and TPN were not associated with a reduction in NI, regardless of the time of nutrition therapy. This result suggests that we should administer initial enteral nutrition therapy within 48 h for critically ill adult patients (24,25).
The primary finding of this study was inconsistent with results of previous studies. Many previous clinical trials, systematic reviews and meta-analysis efforts focused on whether symbiotics can improve NI in critically ill patients, and they rarely included probiotics. Moreover, those studies focused on the outcome of VAP (40,111). In spite of promising data for probiotic use in reducing overall infections, the role of probiotics as a strategy to prevent VAP has been controversial (112). Recently, the results of the largest and most updated systematic review and meta-analysis demonstrated that probiotics are associated with a significant reduction in ICU-acquired infections and in the incidence of VAP. In addition, probiotics appeared to be more effective in reducing NI in patients at high risk of death than in patients at low and medium risk. However, such findings were limited by clinical heterogeneity and potential publication bias (42). Although the mechanisms synbiotics were more effective than prebiotics and probiotics in preventing NI have not yet been clarified, the underlying mechanism areas discussed as follows: Firstly, synbiotics improve gut microbiota. Synbiotics not only increase the number of administered bacteria but also increase their genus groups and other microbiota, which could lead to the maintenance of gut microbiota (107). Secondly, synbiotics generate nutritional support for host epithelial cells. Synbiotic therapy had significantly increased levels of shortchain fatty acids are utilized mainly by intestinal epithelial cells as energy sources, The increased levels of short-chain fatty acids, especially acetate which might attenuate inflammation to reduce NI (60,113). Thirdly, synbiotics maintain gut epithelial barrier. Increased levels of acetate and lactate might inhibit intraluminal toxins and maintain tight junctions (109). Finally, synbiotics regulate immune system function. Synbiotics regulates the innate and adaptive immune systems to reduce systemic inflammation and promote extra-intestinal organ function (109). These changes indicated that synbiotic therapy could have beneficial effects on reduce the development of NI (114,115).
There were several strengths in this study. Firstly, this study was the first analysis using NMA to examine the effectiveness and determine the best choice of symbiotic regimen in improving NI in critically ill patients. This work helped us better assess the relative effects of treatment comparators in the absence of headto-head trials. Secondly, our study is the most updated evaluation of the overall effects of symbiotic therapy in critically ill patients. It contained new suitable trials published on this topic since 1995 by focusing on NI. Thirdly, our study is the largest assessment of symbiotic therapy that included 55 RCTs published in both English and non-English languages from 24 countries, enrolling 7,119 patients. Fourthly, this study examined several relevant clinical outcomes in a heterogenous ICU patient population, including mixed ICU patients, MV patients, trauma patients, SAP patients and postoperative patients. Therefore, the results of this study helped reduce heterogeneity and potential publication bias and could be applied to a broad group of critically ill patients. Overall, all these factors increased the validity and robustness of our results.
Several limitations were still present in drawing strong treatment inferences. Firstly, the definitions of some diarrhea included our study were inconsistent because they are based on criteria of frequency, consistency (116), weight, duration and a combination of frequency and consistency. Such variations are rather vague and subject to different interpretations. There are at least 14 different definitions (117). Making those different definitions consistent is difficult. We were also unable to perform further grouping analysis because of the limited number of studies. Analogously, the definition of prebiotics more or less overlapped with the definition of dietary fiber. In addition, some studies did not provide the accurate definitions of study outcomes. We acknowledge potential misclassification and inconsistency, which is one of the reasons why we downgraded the GRADE of those secondary outcomes. Moreover, the variety of synbiotic strains and length of administration of therapy amongst the different trials weakened any possible clinical conclusions and recommendations. Given the limited number of studies evaluating each endpoint, we were unable to perform subgroup analysis for all clinical outcomes. A further limitation is that the quality of many comparisons was assessed as low or very low level of evidence for hospital LOS, ICU LOS, and duration of MV. Hence, the inferences from current findings were weakened. Lastly, the generalizability of results was limited to other populations as nearly 90% of all studies came from Asia and Europe countries. In addition to the above limitations, we acknowledge potential heterogeneity among critically ill patients in different trials. We have conducted subgroup analysis from many aspects such as different diseases populations, initial time of nutrition therapy, and strive to minimize heterogeneity.
A multicentre, concealed, randomized, stratified, blinded, controlled trial (111) to evaluate the effect of probiotics on VAP and other ICU-acquired infections in 2,650 critically ill patients is ongoing in Canada, USA and Saudi Arabia (clinical trials. gov. registration NCT02462590). REVISE Trials are also ongoing in North America, Australia and Saudi Arabia. The results of these trials will provide further information about the curative effect on symbiotics in the ICU.

CONCLUSION
This systematic review and NMA provide evidence that synbiotic therapy ranked first over probiotics, prebiotics, EPN and TPN to prevent NI in critically ill adult patients. Conversely, TPN therapy significantly increased NI in the critically ill compared with other therapies. Physicians in critical care and related disciplines should consider the use of synbiotics as an adjunctive therapy to improve NI amongst critically ill adult patients. At the same time, the duration of TPN alone should be reduced to decrease NI, especially in ICU and SPA patients. However, on the basis of current data, there is not currently sufficient evidence to make a final strong recommendation for synbiotic therapy to be utilized in the improvement of NI in the critically ill. Numerous questions remain unanswered about a variety of synbiotic strains, wide range of daily doses and duration of therapy; such topics can be addressed in future work.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.