LAUSR

Nature MediciNe | VOL 25 | NOVEMBER 2019 | 1640–1643 | www.nature.com/naturemedicine systematic review concluded that although the overall microbiota is altered by various dietary interventions, more research is needed to support the assumption that A. muciniphila can be targeted with dietary intervention because the results from this approach have not been consistent between studies3. As a matter of fact, we demonstrated in two human studies that either supplementation with isolated inulin and/or oligofructose (16 g/day)4 or increasing the consumption of selected vegetables to naturally boost the intake of prebiotic fibers (15 g/day) induces a marked shift in the gut microbiota composition (e.g., a consistent increase in Bifidobacterium spp. and Faecalibacterium prausnitzii and a decrease in other specific taxa)4,5. However, although both types of regimens are associated with a strong impact on gut microbiota composition, none of these dietary approached induced changes in A. muciniphila levels4,5. Because the intake of all the nutrients mentioned by Janket et al.1 were listed among our exclusion criteria, these confounding factors were excluded in our study2. We believe that our study cannot be viewed as a phase 1 trial, as such studies are designed differently and are used to find the best compromise in regard to the safety, tolerability, and pharmacodynamics (PD) and pharmacokinetics (PK) of an absorbed compound, and also normally include dose ranging to determine at what dose adverse effects begin to appear. Although we designed our nutritional study with all best efforts to achieve high quality standards, such as are used in testing pharmaceutical compounds, our study obviously did not include any escalation of doses to determine the minimal toxic dose, and it would have been impossible to assess PK and PD with this sort of nutritional approach. Janket et al.1 also point out that in our previous study in mice6, we found that “the health benefits of the A. muciniphila membrane protein Amuc_1100 were more pronounced than those of live A. muciniphila.” They suggest that supplementation with Amuc_1100 rather than live bacteria will be more appropriate in future research. The protein indeed replicated part of the effects of the live bacterium; however, we wish to highlight that the effects of the protein matched those observed with the pasteurized bacteria, which had greater efficacy than the live bacteria. Thus, using pasteurized A. muciniphila probably has different effects than simply increasing the levels of endogenous live bacteria or supplementing with a high dose of live bacteria. In conclusion, the fact that many potential confounding factors were indeed taken into account and monitored clearly imply that Simpson’s paradox may not apply to our study. ❐