**3. Discussion**

The gu<sup>t</sup> microbiome has a highly complex structure, involving thousands of microorganisms belonging to very di fferent taxonomic classifications with millions of relationships between them, making its study a grea<sup>t</sup> challenge. In IBD, no specific pathogen has been definitively identified, but the gu<sup>t</sup> microbiota as a whole has been shown to be pathogenic, contributing to the development of a deregulated inflammatory response in susceptible hosts. Several authors defend the suggestion that there is a reduction in biodiversity and abundance of *Bacteroidetes* and *Firmicutes*, as well as an increase in *Proteobacteria*, characteristic of patients compared to healthy individuals [6,7].

Treatment for CD is selected according to the severity of the disease and the response to previous therapies. Among the drugs used are biological treatments based on TNF α monoclonal antibodies such as infliximab and adalimumab, developed for the induction and maintenance of remission, allowing the control of symptoms and an improvement in the quality of life of responders, as well as changes in the natural course of the disease [38–40]. Nonetheless, although approximately one third of patients do not respond to these inhibitory therapies, we currently do not have a non-invasive biomarker that serves as a tool to predict this response, and invasive methods such as colonoscopy continue to be the gold standard for assessing therapeutic response.

In this prospective observational study, a new tool, massive sequencing of bacterial DNA, will be applied to the study of a clinical problem that a ffects an important number of patients. This will enable both the identification of early biomarkers of response in patients with CD after anti-TNF treatment as well as the prediction of therapeutic response from the start to thus improve the clinical managemen<sup>t</sup> of these patients, reduce morbidity rates and increase e fficiency. Since this is a longitudinal study, the patients will be analyzed before and after exposure to anti-TNF treatment, and the data will be paired, thereby diminishing the e ffect of the high variability of gu<sup>t</sup> microbiota.

We know that the composition of the microbiota varies due to multiple external factors, particularly diet, and that a dietary intervention of just three days can cause a change in enterotype [41,42]. Nevertheless, after ten days, the enterotypes stabilized in one study suggesting a tendency to return to the original state [41]. Even so, we added to the protocol a 3-day dietary record prior to stool collection for microbiota analysis and will repeat this dietary record prior to the second assessment after six months of anti-TNF treatment to minimize significant diet-induced changes. An additional

measure to objectively determine whether significant changes in the eating behavior of patients occur during the six months of the study will be undertaken through completion of a validated survey on adherence to the Mediterranean diet, both at inclusion and completion of the study. Given that long-term disturbances have a more profound effect, with a one-year diet modification having a strong impact on the *Bacteroidetes*/*Firmicutes ratio* [41,43], this may lead to changes in enterotype.

The proposed design takes diet into consideration and is therefore novel with respect to similar studies published to date, which did not evaluate this factor known to modify the composition of the microbiota. This study may provide additional evidence regarding potential non-invasive tools such as biomarkers of the response before and after anti-TNF therapy in CD as a starting point for future clinical trials. These trials could determine the most effective treatment among not only these therapies but all therapies used in the managemen<sup>t</sup> of CD based on patient microbiota and provide more appropriate, inexpensive and non-invasive tools for predicting clinical response to treatment.
