| (i) Study Design and Quality |
Considerations:
Type of microbe/diet/outcome interaction
A relatively “simple” interaction with a single strain of bacteria, measuring the outcome (e.g., glucose response) over a number of weeks can give more confidence of “cause and effect.” A “complex” study may involve a prebiotic + several strains administered over several weeks or months to assess the weight management response and is likely to have higher inter-individual variation and therefore may be harder to establish cause/effect: is it the overall diet, or the microbes, or both, having an improved effect? The type of interaction also determines the confidence and the numbers of times a study should be repeated in order to have a level of confidence. However, there are pros and cons, and all types of studies are required. A simple or “direct” interaction gives confidence but the overall health benefit (e.g., short-term glucose) will be limited. A “complex” interaction is harder to give confidence but comes with a better overall health benefit (e.g., long-term weight management).
Levels of Interaction
A ‘direct’ interaction could be administration of a bacterial strain affecting glucose response. An intermediate interaction: specific prebiotics, fibre etc. with any type of response thus harder to determine if it is the nutrients or microbe growth, or both. An indirect interaction is the case where a mechanistic interaction between the microbe variant and the dietary component on a health biomarker, including disease, is affected to some extent but is also influenced by many other possibly unknown processes, and it may take years for symptoms to manifest. This type of interaction may not be fully explained physiologically or may be only demonstrated statistically.
|
| (ii) Biological Mechanism and Plausibility |
Considerations: Biological plausibility is a judgement based on the collected evidence of a microbe x diet interaction on a phenotype. An example of high biological plausibility could be a single microbial strain known to have benefits regarding saturated fat metabolism that leads to lower triglycerides and cholesterol.
In this respect, Neville and colleagues recently proposed a variant of Koch’s postulates to provide a framework to establish causation in the case of a single strain in human microbiota research [269]. On the other hand, a vegan diet high in fibre affects the gut flora and over time the symptoms of metabolic syndrome improve—this type of interaction may not be fully explained physiologically or may only be demonstrated statistically. |
| (iii) Probability Term |
| Considerations: Assessing the validity of a putative microbe × diet interaction is generally complex, and as knowledge deepens, assessment of its validity will develop. |
| Probability terms based on subjective probability range |
| Probability term |
Subjective probability range (%) |
| A. Convincing |
> 90 |
| B. Probable |
66–90 |
| C. Possible |
33–66 |
| D. Insufficient |
< 33 |
