Table 1.
Summary of recent trials reporting on the association between baseline gut microbiota composition and association with clinical response in obesity.
| Study | Study Design | No. of Subjects | Clinical Condition | Dietary Intervention | Control | Key Findings | Ref. |
|---|---|---|---|---|---|---|---|
| Controlled trials – weight response | |||||||
| Cotillard et al., 2013 | Non-randomised clinical trial | 49 (f = 41, m = 8) | Obesity (n = 38, and n = 11 overweight) | 6-week energy-restricted high-protein diet followed by a 6-week weight-maintenance diet | - | Responders: Higher gene richness (where responders were those with marked improvement of adipose tissue and systemic inflammation). | [33] |
| Dao et al., 2016 | Non-randomised clinical trial | 49 (f = 41, m = 8) | Obesity and overweight | 3-week calorie restriction | - | Responders: Higher gene richness and Akkermansia muciniphila abundance was associated with most improved body fat distribution, fasting plasma glucose, plasma triglycerides, improvement in insulin sensitivity. | [37] |
| Modelling studies – weight response | |||||||
| Kong et al., 2013 | Network modelling | 50 (f = 42, m = 8) | Obesity and overweight | 6-week energy-restricted, high-protein diet followed by maintenance phase | - | Responders: Baseline microbiota not identified as a predictor. Non-responders: High Lactobacillus/Leuconostoc/Pediococcus. | [38] |
| Korpela et al., 2014 | Predictive modelling | 78 (f = 40, m = 38) | 3 cohorts with obesity | 3 different types of dietary interventions varying in carbohydrate quality and quantity | - | Responders: High abundance of Firmicutes, where the microbiota composition was associated with change in serum cholesterol levels. | [39] |
| Controlled trials – glycaemic response | |||||||
| Kovatcheva-Datchary et al., 2015 | RCT, crossover | 39 (f = 33, m = 6) | Healthy | 3-day barley kernel-based bread | 3-day white wheat flour bread | Responders: Higher Prevotella/Bacteroides ratio and increased Dorea that could predict PPGR to barley kernel-based bread. | [40] |
| Korem et al., 2017 | RCT, crossover | 20 (f = 11, m = 9) | Healthy | 1 week of 3× 145 g whole-grain sourdough/day | 1 week of 3× 110 g refined white bread/day | Responders: Specific microbial signature (especially abundances of Coprobacter fastidiosus and Lachnospiraceae bacterium) could predict PPGR to either bread. | [41] |
| Modelling studies – glycaemic response | |||||||
| Zeevi et al., 2015 | Machine learning algorithm | 800 (f = 480, m = 320) | Healthy (assessing glycaemic response) | 1-week usual diet with one standardised meal with 50 g available carbohydrate/day | - | Responders: Proteobacteria, Enterobacteriaceae and Actinobacteria were associated with elevated PPGRs. Non-responders: Clostridia and Prevotellaceae associated with lower PPGRs. | [6] |
| Mendes-Soares et al., 2019 | Same modelling framework as Zeevi et al., 2015 [6] | 327 (f = 255, m = 72) | Healthy (assessing glycaemic response) | 6-day usual diet including four standardised meals with 50 g available carbohydrate | - | Baseline microbiota combined with other physiological characteristics was more predictive of PPGR than using only calorie or carbohydrate content of foods. | [42] |
PPGR, postprandial glycaemic response; RCT, randomised clinical trial.