Effects of diets containing synbiotics on the gut microbiota of critically ill septic patients: a pilot randomized controlled trial

2.1 Study design

We conducted a pilot randomized controlled trial between February 2019 and March 2021 at the intensive care unit (ICU), cardiac care unit (CCU), respiratory care unit (RCU) and other general medical wards of the Songklanagarind Hospital (Hat Yai, Thailand), an 800-bed hospital and the largest university-based tertiary care center in southern Thailand.

2.2 Participants

The study inclusion criteria were: (1) aged 18 years or older, (2) received invasive mechanical ventilation, (3) were fed via a nasogastric tube, (4) diagnosed with sepsis, severe sepsis or septic shock, as defined by the sepsis diagnostic criteria from the American College of Chest Physicians/Society of Critical Care Medicine consensus conference 1991 [14], and (5) after admission, they received antibiotics less than 48 hours prior to inclusion, and provided informed consent. Those in end-stage or palliative care; had lower gastrointestinal bleeding or diarrhea; were pregnant or lactating; had a history of bowel surgery, medical diet allergy, synbiotic or probiotic use in the last month, and; were immunocompromised or had undergone prior steroid, chemotherapy, targeted therapy or immunotherapy within one month were excluded.

2.3 Randomization and blinding

The investigators evaluated patients for eligibility, obtained informed consent, and enrolled the participants. After inclusion, the patients were randomly assigned without restriction in a block of 4 at a 1:1 ratio stratified by sepsis, severe sepsis or septic shock according to a computer-generated randomization table derived from www.randomization.com, by a research nurse assistant who had no role in patient management. The research nurse assistants not otherwise involved in the study administered both study diets. The attending physicians, nursing care teams, research investigators, participants, and their family members were blinded to treatment allocation. The diets were prepared by a nutritionist who had no other role in the trial. The diets were packaged in identical bottles labeled with sequential numbers that contained a specific diet volume depending on the attending physician’s orders. The patients who received the symbiotic diet were assigned to the synbiotics group, and those who received the non-synbiotic diet were assigned to the non-synbiotics group.

2.4 Study procedure

The synbiotics group received the commercial polymeric formula Boost optimum® (Nestle Japan Co. Ltd., Tokyo, Japan; 1 kcal/mL; 49:16:35 ratio of carbohydrate, protein, and fat; 310 mOsm/L), which contained probiotics, such as 1 billion colony forming units (CFUs) of Lactobacillus paracasei NCC 2461 (ST11) per 500 kcal and 12.3 gm/L of prebiotics consisting of 70% fructooligosaccharide (FOS) and 30% inulin [15]. The non-synbiotics group received the commercial polymeric formula Ensure® (Abbott Japan Co. Ltd., Tokyo, Japan; 1 kcal/mL; 55:16:29 ratio of carbohydrate, protein, and fat; 433 mOsm/L). Enteral feeding was initiated as soon as possible through a nasogastric tube, which did not determine the volume or rate of feeding depending on the attending physician’s decision. The feeding protocol strictly adhered to the Songklanagarind Hospital feeding protocol (Supplementary Fig. 1). Enteral feeding with the study diet was continued for 7 days or until the patient asked for another diet, such as soft food or a regular diet.

Fecal samples were acquired from the subjects before or within 1 day after starting the study enteral feeding and before day 3. The samples were collected using a specific stool collection tube with a defined quantity of feces, stored in a 4 ℃ sample collection box, preserved in a DNA/RNA Shield (209501, Zymo Research, Irvine, CA, USA) and transferred to a −80 ℃ refrigerator within 1 hour of collection. Bacterial DNA was extracted using the ZymoBIOMICS^TM DNA Miniprep kit (211447, Zymo Research, Irvine, CA, USA) following the manufacturer’s instructions. The DNA samples were sequenced using the Illumina® MiSeq^TM platform at V3–V4 of 16s rRNA with the ZymoBIOMICS® Targeted Sequencing Service for Microbiome Analysis (Zymo Research, Irvine, CA, USA), with international quality control. After sequencing, all sequence reads were analyzed using the Quantitative Insights Into Microbial Ecology (QIIME) pipeline (KnightLab, La Jolla, CA, USA) to generate data on alpha diversity, beta diversity and bacterial composition.

2.5 Outcomes

The primary endpoint of the study was fecal microbial diversity, defined by the alpha (within-sample) and beta (between-sample) diversity.

The alpha diversity was represented by the number of different taxa (richness) and their distribution (evenness) in fecal samples. Here is an illustration of alpha diversity using two samples (groups 1 and 2). The samples from groups 1 and 2 contained four distinct taxa each, suggesting that the two samples’ richness levels were equivalent. The distribution of taxa in the sample of group 2 was 94% for taxa one, 1% for taxa two, 2% for taxa three, and 3% for taxa four, compared to 25% for each of the four taxa in the sample of group 1. Thus, compared with group 2, group 1 exhibited greater diversity and was regarded as having higher alpha diversity. Richness and evenness within fecal samples were assessed using Shannon’s index. Faith’s phylogenetic diversity was used to qualitatively measure richness using phylogenetic relationships within the fecal samples. Pielou’s evenness measured the evenness of taxa distribution [16].

Beta diversity was used to measure the differences in the community composition between the samples. It was calculated and analyzed using principle coordinated analyses, with the derivation of Bray-Curtis distance, Jaccard’s distance, unweighted UniFrac distance matrices and weighted UniFrac distance matrices. Bray-Curtis distance measured the dissimilarity in microbial composition among the samples by taking abundance into account. Jaccard’s distance measured the number of members in a community that only contained the shared genera number, unrelated to their abundance. Unweighted UniFrac was used to detect the presence or absence of lineages and was calculated based on the relative abundances of lineages within communities. The beta diversity values varied from zero to one, of which a higher value denoted lesser similarity between two groups, while a lower value denoted greater similarity [16]. Based on these calculations, we could determine the association of the given diet with a change in gut microbiota diversity by comparing the alpha and beta diversity indices in two groups of fecal specimens.

The secondary endpoints included VAP, nosocomial diarrhea, ventilator days, length of hospital stay, and mortality. VAP was defined as pneumonia occurring 48 hours after intubation according to the 2005 American Thoracic Society guidelines, which included new infiltration based on the chest film and clinical worsening of secretions, respiration, or fever [17]. Nosocomial diarrhea was defined as acute diarrhea 3 days after admission [18].

Baseline data, including age, gender, comorbidities, the sequential organ failure assessment (SOFA) score, lactate level, sepsis characteristics, including the source of infection and organism, and the use of antibiotics, were collected. For adverse events related to feeding, we monitored feeding intolerance and other feeding issues, including bowel ischemia and gut obstruction. Feeding intolerance was defined as either gastric residual volume (GRV) ≥250 mL or feeding symptoms, including vomiting, ileus, abdominal pain and abdominal distension resulting in feeding interruption.

All patients were followed up until discharge or death, whichever occurred first.

2.6 Statistical analyses

No previous studies have investigated the effects of a commercial diet containing synbiotics on gut microbial diversity. For this pilot study, the minimum sample size of the patients required to show the optimal effect size was determined to be 24 [19]. No planned interim analysis was conducted in this trial.

The study was conducted on an intention-to-treat basis. No imputation was performed. The Shapiro-Wilk test was used to assess the normal distribution of the continuous variables. Continuous data are described as mean and SD or median and interquartile range (IQR), depending on the data distribution. Numbers and percentages are used to describe the categorical variables. The differences in patient characteristics and outcomes between the two groups were compared using the Wilcoxon-rank sum test, Fisher exact test, and χ² test, as appropriate.

For the statistical analysis of alpha diversity, the Kruskal-Wallis test was used to calculate statistically significant differences using the QIIME2 plugin. For beta diversity, the permutational multivariate analysis of variance (PERMANOVA) test was used to determine significant differences using the QIIME2 plugin based on 999 permutations.

All statistical analyses were performed with Stata version 16 (StataCorp, College Station, TX, USA), and a p-value < 0.05 was used to indicate statistical significance for all comparisons.

3.1 Primary outcome

We found no significant difference in alpha diversity between the synbiotics and non-synbiotics groups after feeding for three days, as measured by Shannon’s index, Faith’s phylogenetic diversity and Pielou’s evenness. However, a significant difference was observed in beta diversity between the two groups, whereby the non-synbiotics group was found to have a greater beta diversity than the synbiotics groups (Bray Curtis distance, p = 0.001; Jaccard’s distance, p = 0.001; unweighted UniFrac, p = 0.001; weighted UniFrac, p = 0.029) (Table 2, Figs. 2,3, Supplementary Fig. 2).

Fig. 2.Alpha diversity in the synbiotics and non-synbiotics groups on day 3 of feeding. 2A: Shannon’s index. 2B: Faith’s phylogenetic diversity. 2C: Pielou’s evenness. Abbreviation: PD, phylogenetic diversity.

Fig. 3.Beta diversity in the synbiotics and non-synbiotics groups on day 3 of feeding. 3A: Bray Curtis distance. 3B: Jaccard’s distance. 3C: Unweighted-UniFrac. 3D: Weighted-UniFrac.

The fecal microbial diversity compared between day 0 (before starting feeding) and day 3 (after feeding) in both groups showed decreased alpha and beta diversity at day 3. Alpha diversity was significantly affected in terms of richness based on Faith’s physiologic diversity. In addition, beta diversity was significantly affected, as indicated by the Bray Curtis distance and Jaccard’s distance (Table 3 and Figs. 4,5).

Fig. 4.Alpha diversity on day 0 (pre-treatment) and day 3 (post-treatment) of feeding. 4A: Shannon’s index. 4B: Faith’s phylogenetic diversity. 4C: Pielou’s evenness. Abbreviation: PD, phylogenetic diversity.

Fig. 5.Beta diversity between day 0 (pre-treatment) and day 3 (post-treatment) of feeding. 5A: Bray Curtis distance. 5B: Jaccard’s distance. 5C: unweighted-UniFrac. 5D: weighted-UniFrac.

3.2 Secondary outcomes

The comparison of secondary outcomes between the two groups is shown in Table 4. Although the synbiotics group showed a numerically lower incidence in VAP and ventilator days than the non-synbiotics group, the difference was not statistically significant. Further, no significant differences were observed between the two groups in terms of the incidence of nosocomial diarrhea and other outcomes, including ICU length of stay, hospital stay, and hospital mortality.

3.3 Adverse events

There were no reports of major adverse events related to the administration of synbiotics (Table 5). Only three patients in the synbiotics group had feeding intolerance during the study compared to two in the non-synbiotics group, but the difference was not statistically significant. No incidence of ischemic bowel of bowel obstruction was observed in this study.