The global prevalence of diabetes is set to increase within the next 10 years, affecting 439 million adults by 2030 [1
]. The overall prevalence of diabetes in New Zealand is 7%, with an additional 17% with prediabetes [2
]. Individuals with prediabetes are at an increased risk of cardiovascular disease [3
], and 5–10% of those with prediabetes transition to develop type 2 diabetes each year [4
]. Therefore, prevention interventions that are cost-effective and acceptable are crucial.
Recent research into very-low-calorie diets [VLCD; 400–800 kcal/day] for prevention or reversal of type 2 diabetes has shown success [5
]. Although VLCD is successful in achieving weight loss and subsequent improvement in glycemic control, this is largely dependent on the participants’ ability to adhere to dietary guidelines once the intensive meal-replacement phase is transitioned to the maintenance diet [8
]. A promising alternative to continuous energy restriction is intermittent fasting, which has been found to achieve weight loss goals with improved adherence [9
]. Intermittent fasting is an umbrella term for various diets that cycle between a period of fasting and non-fasting over a defined period [10
]. The intensity of fasting can vary from total absence of calories in a water-only fasting day to a reduction of calories of varying extent. Studies have found a significant reduction in glucoregulatory markers after implementing intermittent fasting [11
]. Furthermore, some intermittent fasting trials reported no change in self-reported hunger levels [12
]. In contrast, VLCD interventions have found an increase in ghrelin levels, an appetite-stimulating hormone [14
]. As hunger is a common precursor for weight regain, this suggests that intermittent fasting may achieve more sustained weight loss. The 5:2 diet consists of two days of reduced caloric intake to 600–650 kcal/day per week and five days of ad libitum eating. This form of intermittent fasting may suit individuals who struggle with continuous daily modest calorie restriction (i.e., 1200 kcal/day) or VLCD (400–800 kcal/day) and may therefore be a feasible weight loss approach with accompanying improvements in weight-related complications.
There is growing evidence linking gut dysbiosis with obesity and diabetes. Animal studies have suggested that probiotics can improve insulin sensitivity and beta-cell dysfunction by regulating the key signaling pathways [15
]. Specific strains from the Lactobacillus
genera have been named the most effective for glucose and weight control [17
]. Recent evidence from a New Zealand study showed that the probiotic Lacticaseibacillus rhamnosus
HN001 (previously known as Lactobacillus rhamnosus
HN001) can result in improved gestational diabetes prevalence despite no changes in body weight [18
]. Other research has linked probiotics to improvements in inflammation [19
] and other biomarkers of oxidative stress [20
]. Multiple studies have investigated the synergistic effect of a probiotic intervention with a hypocaloric diet and have reported augmented improvements in health markers [21
]. However, no studies have examined the effect of probiotics in combination with intermittent fasting.
Both intermittent fasting and probiotic supplementation show favorable health outcomes in their own right; however, combining the two may augment these benefits. The PRObiotics and intermittent FASTing to improve prediabetes (PROFAST) trial aimed to establish whether intermittent fasting and daily probiotic supplementation can improve HbA1c in individuals with prediabetes. To our knowledge, this is the first study to combine these two intervention methods. The results of this study will contribute to the growing body of evidence regarding intermittent fasting and probiotic supplementation for the prevention of type 2 diabetes.
2. Materials and Methods
2.1. Trial Design
The PROFAST trial was a 12-week, double-blinded, two-armed, randomized 1:1 study of daily probiotic vs. placebo in the presence of intermittent fasting. This study was a pilot, feasibility study; hence no power calculation was done. This study is registered on the Australian New Zealand Clinical Trials Registry, registration number: ACTRN12616001050448.
All participants provided informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Health and Disability Ethics Committee (16/STH/107) on 12 August 2016.
Participants were either referred to the PROFAST trial by primary care clinicians or self-selected to participate. Eligible participants were aged between 18 and 65 years, had prediabetes (defined as glycated hemoglobin (HbA1c) of 40–50 mmol/mol in the preceding 9 months), and a body mass index (BMI) of 30–40 kg/m2 (or 27–40 kg/m2 for Indian ethnicity). Participants were excluded if they had pre-existing diabetes, previous bariatric surgery, were currently taking glucose-lowering medications, had any conditions that may influence body weight regulation (e.g., thyroid disorders, malabsorption, eating disorders, use of systemic steroids, excess alcohol intake), or had any significant health issues (e.g., end-stage renal failure, congestive heart failure, thalassemia, psychiatric disease). Women were excluded if they were pregnant, breastfeeding, or intending to become pregnant. Participants were also excluded if they were planning any major changes in physical activity during the study to the extent that may interfere with the study outcomes, donated blood within 2 months before the study or experienced a weight change greater than 3 kg within 3 months of the first baseline visit.
The probiotic and placebo capsules were provided by Fonterra Cooperative Group, Ltd. (Auckland, New Zealand), who had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. Capsules had identical packaging and were labeled A or B. The probiotic capsule contained the strain L. rhamnosus HN001 at a dose of 6 × 109 colony forming units (CFU). The placebo capsule contained microcrystalline cellulose and dextrose anhydrate. Participants were given 90 capsules (12 weeks supply) of either the probiotic or placebo.
The intermittent fasting regimen consisted of eating ad libitum for five days and then restricting calories (‘fasting’) for two days per week, for a duration of 12 weeks. On the fasting days, participants were instructed to restrict their calories to a maximum of 600 kcal/day for females and 650 kcal/day for males. Before commencing the 3month intervention, participants were provided with dietary advice and education regarding this fasting regime by the study dietitian. This included education around label reading, calorie counting and how to manipulate recipes to increase satiety while restricting calories. Behavior change techniques were also utilized, such as problem-solving, goal setting and self-monitoring. Participants were encouraged to increase their intake of vegetables and protein on the fasting days and to eat like “normal” on the remaining five days. Physical activity was advised to be kept to the same duration, intensity and frequency as before the study; however, compliance was not measured. Participants received personalized advice and encouragement from the study dietitian on a fortnightly basis, either via phone call, text message or email. To provide additional support, participants were encouraged to contact the study dietitian when required and join a private Facebook group for online discussions.
2.4. Randomization and Blinding
Following the baseline visit, eligible participants were randomized to either intermittent fasting with daily probiotic or intermittent fasting with daily placebo in a 1:1 ratio. This was achieved using opaque sealed envelopes, given by the research assistant, which were labeled either treatment group A or B. The randomization process was determined by a computer-generated variable block size random enumeration, stratified for ethnicity (European, Indian, Māori, Pacific) and BMI (<35, ≥35 kg/m2).
Study participants and the research team were blinded to the probiotic/placebo allocation during the study. Those who analyzed the data were blinded to group assignments. This information was only released to the investigators when all data had been collected and finalized. Both placebo and probiotic-containing capsules were identical in appearance and packaging.
The primary outcome was change in HbA1c after 12 weeks. Secondary outcomes included changes in: anthropometric and body composition measurements, resting energy expenditure (REE), calorie and macronutrient intake, lipid measurements, liver function tests, gut hormones (ghrelin, GIP, active GLP-1, glucagon and PYY), inflammatory markers, glucoregulatory markers (glucose, insulin, C-peptide), mood and eating questionnaire scores, self-reported hunger and satiety and quality of life questionnaire (SF-12) scores.
2.6. Data Collection
At baseline (week 0) and conclusion (week 12), participants attended the Body Composition Laboratory (University of Auckland, New Zealand) after an overnight fast for assessment. This included blood collection for the primary outcome of HbA1c and a 6-point oral glucose tolerance test (OGTT), anthropometry, body composition measurements using dual-energy X-ray absorptiometry (model iDXA, GE-Lunar, Madison, WI) and resting energy expenditure by indirect calorimetry (TrueOne 2400, Parvo Medics, Salt Lake City, UT). Participants completed a visual analog scale (VAS) questionnaire to determine their perceived levels of hunger and fullness before ingesting the 75 g glucose drink for the OGTT. Four questions were included in the VAS questionnaire: “How hungry are you?”, “How full do you feel?”, How strong is your desire to eat?”, “How much food could you eat?” [24
]. The SF-12 questionnaire was used to assess quality of life and was completed by each participant at baseline and 12 weeks follow-up. The SF-12 consisted of 12 questions covering 8 health domains related to physical and mental health (physical function, role physical, bodily pain, general health, vitality, social functioning, role emotional and mental health) [25
All participants completed a mood and eating questionnaire adapted from three validated questionnaires: PHQ-9 [26
], generalized anxiety disorder (GAD) questionnaire [27
] and the Eating Disorders Questionnaire [28
]. Participants were required to bring a completed 7-day food diary to the baseline visit to ascertain their baseline calorie and macronutrient intake. Participants were instructed to complete a food diary for each of their fasting days throughout the 3-month intervention and to present a 7-day food diary at each clinic visit. The use of smartphone applications, such as MyFitness Pal and Easy Diet Diary, was encouraged to more accurately calculate calorie intake.
A subsample of participants (n = 14) was invited to complete a post-intervention interview with the study dietitian. Data were recorded on a standardized questionnaire and included topics such as dietary habits before, during and after the intervention; any changes in physical activity, adherence to the study protocol, food choices on fasting days, antibiotic use and experienced side effects. Qualitative data were also obtained on the acceptability of the intervention and whether participants would continue with the intermittent fasting regime post-intervention.
2.7. Statistical Analysis
Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA). All statistical tests were two-sided at a 5% significance level. Repeated measures analysis of variance (treatment group as a fixed factor, time as a repeated factor) was used to assess the effect of treatment for normally distributed data. When the group x time interaction was significant, post hoc analysis using paired t-tests was conducted to establish the significance of within-group changes. For non-normally-distributed data, the Mann–Whitney U test was used to compare changes between the groups and the Wilcoxon signed-ranks test for within cohort changes. Food diaries were analyzed using Foodworks®
Software), which used nutritional data from FOODfiles [29
The PROFAST trial showed that daily supplementation of Lacticaseibacillus rhamnosus
HN001 probiotic in individuals with prediabetes who practiced intermittent fasting did not result in additional glycemic improvement or weight loss when compared with placebo. However, those supplemented with HN001 showed greater improvements in mental health scores compared to those supplemented with placebo, and this between-group difference was statistically significant. The lack of significant differences between groups for weight loss and change in HbA1c suggests that the greater mental health improvements in the probiotic group were not attributed to these factors. Our results, therefore, offer new evidence that probiotics paired with intermittent fasting can improve mental health. Another New Zealand study investigated the effect of the same probiotic L. rhamnosus
HN001 versus placebo on postnatal mood in women with gestational diabetes and found that daily probiotic supplementation was associated with lower depression and anxiety scores in the postpartum period [30
]. Animal studies have also consistently shown that probiotics positively impact anxiety-and depressive-like behaviors [31
]. Although obesity is associated with depressive and anxiety disorders [32
], we did not compare our participants’ scores with those of a normal New Zealand population. Further research is needed to explore the effects of probiotic supplementation and intermittent fasting among participants with depressive and anxiety disorders.
This trial also adds to emerging evidence that intermittent fasting can reduce body weight and HbA1c. Weight loss in individuals with prediabetes, whether achieved via dietary or surgical methods, is consistently accompanied by reductions in HbA1c [33
]. The magnitude of HbA1c reduction is often proportional to the baseline HbA1c, as greater improvements are achievable from higher HbA1c. In our study, eligible participants had an HbA1c of 40–50 mmol/mol (prediabetes), so the observed 2 mmol/mol reduction may appear modest when compared to other trials that recruited individuals who had type 2 diabetes and thus a higher baseline HbA1c [12
]. Nevertheless, in view of other diabetes prevention trials, such a reduction in HbA1c represents considerable clinical success [35
] and is comparable to results of commercial weight management programs among people with prediabetes [36
]. A systematic review and meta-analysis of 36 studies assessing type 2 diabetes prevention in real-world settings suggest that modest reductions in weight (1.6 kg) and HbA1c (2 mmol/mol) can significantly reduce progression to type 2 diabetes compared to usual care by 26% [36
The 5% weight loss achieved through intermittent fasting is close to the recommendation of 7% weight loss for diabetes prevention by the American Diabetes Association [37
]. This was achieved in 12 weeks and required little time investment from the study team as participants were followed-up via text message, phone call or email on a fortnightly basis. There was no additional weight loss or HbA1c reduction with probiotics, which highlights that clinically significant weight loss can occur with two main components–dietary manipulation and regular virtual support. Unfortunately, there were insufficient resources to continue monitoring participants beyond the 12-week intervention, and thus we do not know whether weight loss and improvements in HbA1c were maintained once regular support was removed. We also cannot conclude that these results were due solely to intermittent fasting, as there was no comparison group using continuous energy restriction. Further research is also needed to establish whether this short-term intermittent fasting intervention can produce sustained weight loss and glycemic improvements.
As with all interventions, outcomes are dependent on participants’ adherence to nutritional guidelines. Participants reported a moderate change between their baseline diet and their diet post-intervention. Participants also reported an overall increased awareness of eating behavior. Although pretreatment binge eating disorder was not assessed, participants reported fewer episodes of overeating during the intervention. This hints that adopting an intermittent fasting regime may have more positive than negative impacts on eating behavior. A systematic review published in 2015 found inconsistent associations between severe dietary energy restriction (i.e., low or very low energy diets) and the onset of binge eating in individuals without pretreatment binge eating [38
]. This, therefore, affirms that calorie restriction may not necessarily trigger adverse eating behaviors in those without binge eating disorder; however, pretreatment screening is still advised.
The main limitations of this study are the small sample size, the lack of an intention-to-treat analysis and the absence of a comparison group using continuous energy restriction. Consequently, we are unable to recommend intermittent fasting over the traditional method of continuous energy restriction in clinical settings. Nevertheless, our results suggest that intermittent fasting may result in weight loss and diabetes prevention, which should be studied further. The results of the secondary analyses need to be interpreted cautiously owing to the multiplicity of tests conducted. Controlling for false discovery rate using Benjamini–Hochberg procedures [39
] yields the smallest p value of 0.392 (instead of 0.007 for mental health).
Several factors were not controlled for in this study, which may have influenced results. First, participants were instructed to maintain their habitual exercise patterns throughout the intervention period; however, this was not monitored to ensure compliance. If participants did alter their exercise levels, our observed results might not be due to intermittent fasting alone. Future studies could perhaps include specific exercise guidance and monitoring, which may also limit the reduction in fat-free mass and resting energy expenditure observed in the present study.
Given the emphasis on restricting calories ascribed to this intervention, the validity and accuracy of using self-reported food diaries to measure caloric intake are limited. Participants may have underreported caloric intake according to social desirability bias [40
] or reactively changed their eating behavior to appear more adherent to the PROFAST protocol. Emerging evidence also suggests that dietary fiber can improve the functionality of probiotics, owing to a synbiotic approach [41
]. Although participants were encouraged to increase their fiber intake to increase satiety on fasting days, this was not adequately controlled for in our study and may have influenced our results.