I appreciate the pertinent and correct observations of doctors Mejía-León and Calderón de la Barca, who indeed are the first to publish the frequency of risk haplotypes for celiac disease (CD) in children from the Mexican State of Sonora, in the Revista de Gastroenterología de México. The original idea of the work of Cerda et al. was to evaluate the frequency of HLA DQ8/DQ2 haplotypes in a highly selected population, such as patients with chronic diarrhea. The results compared between the groups are shown in Table 1 (data not published in the article).

We only found significant differences in relation to more frequent expression of the DQ8 haplotype in celiacs. Both the DQ8 haplotype and the DQ2 haplotype were more common in persons with CD than in Mexican Mestizos, according to the data reported by Barquera et al.1 (80 vs. 24% and 50 vs. 16%), confirming their importance in the physiopathogenesis of the disease, as Mejía-León and Calderón de la Barca accurately pointed out.2 Likewise, the high prevalence of those haplotypes in the population with no CD, reduces the specificity of said typing, making it inadequate for diagnosis. I agree with the two authors that DQ8 and DQ2 typing should only be carried out in selected cases, as proposed in the clinical guidelines of the AMG on celiac disease that are soon to be published.

I underline that the fact that in the present work, the highest frequency of HLA DQ-2 was in subjects with chronic diarrhea and no CD (63 vs. 16% of healthy Mestizos). The most common diagnosis in that group was bacterial overgrowth syndrome (BOS), a relatively common entity seen at our hospital center and one that shares clinical and histologic characteristics with CD. We found it interesting that, despite having common comorbidities (diabetes mellitus and connective tissue disease alterations, among others), only some patients developed BOS, suggesting that a biomarker for characterizing those patients could be risk haplotype typing.

Finally, as was demonstrated in the study by Mejía-León et al., we also found that the DQ-8/DQ-2 combination had the highest odds ratio (10.4; 1.2-89.1) for CD.

On behalf of all the authors that participated in the study by Cerda et al., I sincerely apologize to doctors Mejía-León and Calderón de la Barca and to the scientific community that reads our journal, for the involuntary omission of their excellent article published in 2015 in the Revista Mexicana de Gastroenterología.2 Our study confirms their findings in a different population.