Methodological considerations
A major strength is that this study investigated the association betweenFLG genotype and a large panel of B and T cells in the general population for the first time. We had detailed and extensive information on immune cell numbers from 11‐color flow cytometry and obtained objective information on genetic ancestry. However, the following limitations need to be addressed. First, the AD population for the sensitivity analyses was relatively small which could have limited the power in the statistical analyses. Nevertheless, in comparison to previous studies, only including a maximum of 6 AD patients withFLG mutations, this is the largest study on FLG mutations in both the general population and AD patients. Second, our AD population was defined by ever having physician diagnosed AD before or at the age of 10 years. Therefore, it is likely that a subset of the children has outgrown AD at the age of 10 and this might affect their immunophenotype. In addition, the use of questionnaires for AD diagnosis could have introduced recall bias. Third, as mentioned previously, our study included the four most common FLG mutations in the European population. To prevent misclassification, we selected children with genetic European ancestry for the current study. Although the choice for including the most common FLG mutations in European populations is in line with previous studies 11,12, other less frequent FLG mutations could exist in low numbers since up to 113FLG mutations resulting in premature protein termination have been described. A recent study including patients with AD and Ichthyosis Vulgaris (IV), showed that screening the entire encoding region ofFLG for mutations led to an improvement of the diagnostic yield.30 Furthermore, biallelic genotypes were only present in 3 children of the study population and therefore not studied separately. Since homozygous or compound heterozygous mutations lead to complete absence of filaggrin expression, these mutations might have different effects on immune cell numbers compared to heterozygousFLG mutations which lead to 50% reduction in filaggrin expression.1 Therefore, studies investigating biallelic mutations compared to heterozygous FLG mutations and wild type FLG are needed. Furthermore, as this is the first study in a general cohort addressing the association between FLGmutation and immune cell numbers, future studies are needed for validation of our results.
In conclusion, within the general population, school-aged children withFLG mutations have higher Th22 cell numbers. Furthermore, children with and without FLG haploinsufficiency did not differ in other B and T cell subsets. This suggests that with age, other factors such as environmental exposure contribute to risk thatFLG mutations poses to the development of allergy.