Effects of a multi-strain probiotic supplement for 12 weeks in circulating endotoxin levels and cardiometabolic profiles of medication naïve T2DM patients: a randomized clinical trial

Background The present randomized clinical trial characterized the beneficial effects of a multi-strain probiotics supplementation on improving circulating endotoxin levels (primary endpoint) and other cardiometabolic biomarkers (secondary endpoint) in patients with T2DM. Methods A total of 78 adult Saudi T2DM patients (naïve and without co-morbidities) participated in this clinical trial and were randomized to receive twice daily placebo or probiotics [(2.5 × 109 cfu/g) containing the following bacterial strains: Bifidobacterium bifidum W23, Bifidobacterium lactis W52, Lactobacillus acidophilus W37, Lactobacillus brevis W63, Lactobacillus casei W56, Lactobacillus salivarius W24, Lactococcus lactis W19 and Lactococcus lactis W58 (Ecologic®Barrier)] in a double-blind manner for 12 weeks. Anthropometrics and cardiometabolic profiles were obtained at baseline and after 12/13 weeks of treatment. Results After 12/13 weeks of intervention and using intention-to-treat analysis, no difference was noted in endotoxin levels between groups [Placebo − 9.5% vs. Probiotics − 52.2%; (CI − 0.05 to 0.36; p = 0.15)]. Compared with the placebo group however, participants in the probiotics groups had a significant but modest improvement in WHR [Placebo 0.0% vs. Probiotics 1.11%; (CI − 0.12 to − 0.01; p = 0.02)] as well as a clinically significant improvement in HOMA-IR [Placebo − 12.2% vs. Probiotics − 60.4%; (CI − 0.34 to − 0.01; p = 0.04)]. Conclusion Using a multi-strain probiotic supplement daily for 12/13 weeks significantly improved HOMA-IR and modestly reduced abdominal adiposity among medication naïve T2DM patients. Trial Registration: ClinicalTrials.gov Identifier: NCT01765517, Registered January 10, 2013 Electronic supplementary material The online version of this article (10.1186/s12967-017-1354-x) contains supplementary material, which is available to authorized users.


Background
In the last few years the gut microbiome has gained considerable interest due to its ability to coexist with its human host and complement several key physiologic processes peacefully maintaining homeostasis and overall human health [1]. One of the accepted theories that explain the contribution of gut microbes in the development of diseases is sub-chronic inflammation secondary to endotoxemia. This state occurs when fragments of gut-derived Gram negative bacteria (lipopolysaccharides or endotoxin) traverse the intestinal mucosa to enter the Sabico et al. J Transl Med (2017) 15:249 circulation, and may represent an important mediator of low-grade systemic inflammation influenced by the host's own gut microbiota and metabolic state [2]. Previous studies have also shown that endotoxin can stimulate an innate immune response from adipose, liver and skeletal muscle tissues, leading to increased production of proinflammatory cytokines [3].
There has been accumulating evidence pointing to the manipulation of the gut microbiome in the prevention and reversal of several chronic non-communicable diseases such as obesity, type 2 diabetes mellitus (T2DM) and the metabolic syndrome [4]. It is now established that dietary intake and nutrition management are significant and effective external factors in modifying the gut ecosystem [5]. Specifically, probiotics, or live bacteria naturally found in the human body that confer health benefits, have shown great potential as adjuvant therapies for a number of insulin-resistant diseases. Currently, randomized clinical trials are limited to strengthen this case. The present investigation aims to fill this gap.
In this study, we determine whether a 12/13 week supplementation of a multi-strain probiotic would induce favorable changes in circulating endotoxin levels (primary outcome) and cardiometabolic profile (secondary outcome) of medication naïve T2DM individuals.

Methods
The present study is a 12-week single-center, doubleblind, randomized, placebo-controlled study. The protocol has been published previously [6] and was registered at the US National Institute of Health (ClinicalTrials.gov Identifier: NCT01765517). Ethical approval was obtained from the Ethics Committee in the College of Science (Approval Code 8/25/16519), King Saud University in Riyadh, Saudi Arabia and written informed consents were obtained from all participants prior to inclusion.

Subjects
All participants were recruited in the outpatient department of King Salman Hospital, Riyadh, Saudi Arabia. A total of 150 adult Saudis (aged 30-60 years old) with newly diagnosed T2DM were initially invited to participate. Participants with T2DM complications and unstable glycemic control were excluded. Those who were already taking probiotics as well as antibiotics, 6 weeks before inclusion, women lactating or pregnant, participants who would anticipate changes in antidiabetic medications (if any) in the next 6 months, on insulin or its analogues and those with gastrointestinal diseases, such as irritable bowel syndrome, were excluded. Thirty participants failed to meet the inclusion criteria and 24 refused to participate. A total of 96 participants were eligible and gave informed consent. They were randomly assigned to either placebo or probiotic sachets unknown to both the principal investigator and the participants. Seventy-eight participants were able to complete the 12 week intervention. A flowchart is shown in Fig. 1.

Probiotic supplements and allocation
All eligible and consenting participants were given a unique code as identifier. They were allocated (1:1) to receive either probiotics or placebo. The randomization scheme was computer generated by Winclove using permuted blocks with block size equal to 4. It was impossible for research personnel involved with participants to adjust randomization or discern what product participants were receiving, ensuring true allocation concealment. The probiotic group received sachets with 2 g freeze-dried powder of the probiotic mixture Ecologic ® Barrier (Winclove probiotics, the Netherlands). Ecologic ® Barrier (2.5 × 10 9 cfu/g) contains the following bacterial strains: Bifidobacterium bifidum W23, Bifidobacterium lactis W52, Lactobacillus acidophilus W37, Lactobacillus brevis W63, Lactobacillus casei W56, Lactobacillus salivarius W24, Lactococcus lactis W19 and Lactococcus lactis W58. The placebo group received the same sachets consisting of the carrier of the probiotic product that is maize starch and maltodextrins. The placebo is indistinguishable in color, smell and taste from the probiotic sachets. All participants were asked to consume two sachets per day (dissolving contents in glass of water) once before breakfast and before going to bed.

Monitoring and blood sample collection
Anthropometry include height (cm), weight (kg), blood pressure (mmHg) waist and hip measurements (cm) were done at baseline and after 12 weeks of intervention. Fasting blood samples were also collected at baseline and after 12 weeks. Samples were immediately centrifuged and serum samples separated. All samples were put in ice and immediately delivered to Prince Mutaib Chair for Biomarkers of Osteoporosis in King Saud University for storage at -20 °C until further analysis. Participants were asked to return every 4 weeks to surrender unused sachets and to be given fresh refill to monitor compliance. Participants were also asked for any side effects.

Biochemical analyses
Fasting serum samples were analyzed for glucose and lipid profile (total cholesterol, HDL-cholesterol and triglycerides) using routine analyzer (Konelab, Espoo, Finland). LDL-cholesterol was calculated using the Friedwald equation. Serum insulin and c-peptide were measured using electrochemiluminescence assay (Roche Diagnostics, Germany). Endotoxin was measured using a limulus amebocyte lysate (LAL) quantitative kinetic assay (Lonza, MD, USA). A spike recovery was performed using a sample dilution of 1:40. The recovery spike was 60% and was within the acceptable range of 50-200%. All serum samples were analyzed at baseline and after 12 weeks of intervention.

Data analysis
Sample size calculation was previously done using endotoxin as primary endpoint [6] and repeated after obtaining actual results. The sample size in the present study (N = 39 per group) has an actual power of 83%. Statistical analysis was performed using Intention-to-treat (ITT) analysis, where missing data were dealt by using the last observation carried forward (LOCF) method. Per-protocol (PPA) analysis was also performed on participants who successfully completed the trial and is presented as a additional data (Additional file 1: Table S1). Data were analyzed using SPSS (version 16.5 Chicago, IL, USA). Mean and standard deviations were used to represent the data for the normal variables, while median and interquartile range were used to report non-normal variables. Furthermore, frequencies and percentages (%) were reported for the categorical data. Changes were also calculated as mean and as percentage. Correlations between endotoxin anthropometrics, glycemic and lipid profile were measured using Spearman correlation coefficient. Independent sample Student T-test and Mann-Whitney U test was used to determine the significant difference between placebo and probiotic groups at baseline. Mixed method analysis of covariance (ANCOVA) was used to determine within and between group differences after adjusting for baseline observations and covariates including WHR, MAP, Glucose (mmol/l), TC/HDL and Endotoxin (IU/ml). All non-normal variables were transformed prior to parametric testing. Intervention effects were presented at 95% confidence interval (CI). p < 0.05 was considered statistically significant. Table 1 shows the demographic characteristics of participants assigned to placebo (N = 39) and probiotics (N = 39). No differences were found in age, weight, BMI and sex distribution. The placebo group had significantly higher waist-hip ratio than the probiotics group (p = 0.02). The probiotics group on the other hand, had significantly higher diastolic and mean arterial blood pressure (p values 0.04 and 0.024), respectively. With regards to glycemic and lipid parameters, the probiotics group had significantly higher median glucose levels, as well as circulating levels of total cholesterol and LDLcholesterol than the placebo group (p < 0.001, 0.001 and 0.05, respectively). The rest of the parameters were not significantly different from one another.

Endotoxin
Within and between group effects of participant characteristics are shown in Table 2. No difference was noted in endotoxin levels between groups [Placebo − 9.5% vs. Probiotics − 52.2%; (CI − 0.05 to 0.36; p = 0.15)]. Within group comparisons however showed a significant decrease in endotoxin levels in the probiotics group (p < 0.01). This was not observed in the placebo group. Mean differences between placebo and probiotics group are presented in Fig. 2.

Anthropometric and clinical measures
Compared with the placebo group, participants in the probiotics groups had a significant improvement in WHR [Placebo 0.0% vs. Probiotics 1.11%; (CI − 0.12 to − 0.01; p = 0.02)]. This significant reduction was also observed using PPA [Placebo 0.0% vs. Probiotics 1.11%; (CI − 0.14 to − 0.03; p = 0.01)] (Additional file 1: Table S1). . Within group comparisons showed a significant decrease in systolic and diastolic as well as mean arterial blood pressure in the probiotics group (p < 0.01, 0.03 and < 0.01, respectively) post intervention. These differences were not observed in the placebo group. Furthermore, no significant changes in weight, BMI and WHR were observed in either group after 3 months (Table 2).

Glycemic profile
A clinically significant improvement in HOMA-IR was observed in the probiotics group compared to placebo

Table 2 Anthropometrics, glycemic and lipid profile characteristics before and after supplementation with placebo or probiotics using intention-to-treat analysis
Data presented as Mean ± SD for normal variables while non-normal variables are presented as Median (inter-quartile range) All non-normal variables were transformed prior to parametric testing BMI body mass index, WHR waist-hip ratio, SBP systolic blood pressure, DBP diastolic blood pressure, MAP mean arterial pressure, Glu glucose, Ins insulin,    15:249 post-intervention levels of glucose, insulin, C-peptide and insulin resistance were significantly lower than baseline in the probiotics group (p < 0.01, < 0.01, 0.01 and < 0.01, respectively). Within group comparisons showed that both placebo and probiotics group had significantly lower levels of total cholesterol after intervention (p < 0.01). Only the probiotics group, however, showed significantly lower circulating triglycerides and LDL-cholesterol after intervention (p values 0.04 and 0.02, respectively). Both groups had no significant changes in the total/HDL cholesterol ratios post-intervention. Table 3 shows the bivariate associations between endotoxin and parameters measured. In all participants, endotoxin was significantly associated with diastolic BP (R = 0.27; p = 0.03) and MAP (r = 0.26; p = 0.04). HDLcholesterol was inversely and significantly associated with endotoxin levels in all participants (R = − 0.25; p = 0.04) and in the probiotics group (R = − 0.35; p = 0.05). In the probiotics group, there were also significant associations between endotoxin and triglycerides (R = 0.37; p = 0.04) and total/HDL cholesterol ratio (R = 0.42; p = 0.02). The latter was also significant in all participants (R = 0.32; p = 0.01).

Lipid profile
Lastly, none of the participants complained of any serious side effects from the clinical trial. The most common complaint were minor gastrointestinal discomfort (feeling bloated and increased flatulence during the first week of treatment) (N = 5, 1 in the placebo group and 4 in the probiotics group) which is common for first time probiotics users. This symptom gradually faded during the first weeks of treatment.

Discussion
Individuals with T2DM and those with insulin-resistance in general exhibit higher metabolic endotoxemia than their non-diabetic counterparts [7]. In animal studies it was demonstrated that increased levels of circulating insulin alters intestinal permeability, allowing gut endotoxins to leak in the circulation, which, in turn, initiates a cascade of inflammatory reactions via the innate immune pathway, thus explaining the subclinical inflammation in obesity and insulin-resistant states [8]. Furthermore, a widely accepted theory is that probiotics supplementation can restore a weakened intestinal barrier, preventing endotoxin influx in the circulation and ultimately reducing subclinical inflammation [9]. By manipulating endotoxin levels through the introduction of probiotics in the digestive tract, it is believed that many endotoxin-induced metabolic disorders can be reversed, if not controlled. We conducted our 3-month randomized, double-blind, placebo-controlled clinical trial on the endotoxin-lowering effects of an 8-strain probiotics supplement among participants with T2DM. We found that while circulating endotoxin levels in the probiotics group were no different than placebo after 3 months of intervention, significant improvements in WHR and HOMA-IR were observed.
A recent meta-analysis of RCTs done thus far on probiotics and T2DM revealed that multiple species of  15:249 probiotics and interventions longer than 8 weeks had stronger metabolic improvements in terms of improved glucose control and lipid profiles [10]. With 8-strains of probiotics supplements used over 3 month duration, we confirm these beneficial effects in reducing abdominal adiposity (measured as WHR) and insulin resistance (HOMA-IR). The lack of improvement in lipid profile however and other indices assessed in the present study do not supersede previous findings. Our results are in agreement with a recent double-blind, randomized trial involving 43 participants (Placebo N = 22 and Probiotic mix N = 21) who were given 8 weeks supplementation of probiotic mix (Lactobacillus acidophilus and casei; Lactococcus lactis; Bifidobacterium bifidum and lactis; 2 × 10 10 cfu/day) and found significant reduction in abdominal adiposity with no concomitant decrease in endotoxin levels [11]. Three probiotic species used in the former and the present study, namely, Bifidobacterium bifidum, Lactobacillus acidophilus and Lactobacillus casei, have been demonstrated to significantly improve glycemic, inflammatory and lipid profiles of patients with gestational diabetes mellitus after 6 weeks of supplementation [12,13]. Lactococcus lactis, another potent probiotic species used in our study, was recently reported to cure diabetes in non-obese diabetic mice, in combination with low-dose anti-CD3, through a series of actions including decline in insulin autoantibody positivity and stable reversal of hyperglycemia [14]. Lactobacillus salivarius was also shown to reverse diabetes-induced intestinal defense impairment through reversal of enteric dysbiosis and decreased endotoxin levels in streptozotocin-induced diabetic mice [15]. Studies using Lactobacillus salivarius as a stand-alone probiotic supplement for 4-6 weeks in women with gestational diabetes, however, was not associated with any improvement in metabolic health and pregnancy outcome [16]. In this study, most likely, the cumulative potency of the 8 species employed may have contributed to the significant improvements in the HOMA-IR and WHR of the probiotics group. A recent randomized clinical trial involving 136 Malaysians with T2DM supplemented with either placebo or probiotics (Bifidobacterium and Lactobacillus) for 12 weeks also showed improvement in terms of glycemic control [17].
We were not able to elicit significant changes in BMI and body weight in both groups studied. This confirms several studies, including the recent meta-analysis by Park and Bae, who concluded limited efficacy of probiotics in weight management [18]. However, clinical trials overall are still very limited and therefore current evidence on probiotics, as weight loss agents are at most, suggestive. We also found no significant improvements in blood pressure although a recent study in animal models showed remarkable improvements in blood pressure