Carrier Frequencies of HLA-DRB1~DQB1 Haplotypes Used in Environmental Susceptibility Testing: Estimates from Registry, Biobank, Sequencing and Family-Phased Datasets
Year: 2026
Type: preprint
Abstract: Background. Patients evaluated for illness associated with water-damaged buildings are frequently offered HLA-DR/DQ typing, with results interpreted against a published table that labels certain haplotypes "mold susceptible" or "multisusceptible." That table belongs to the chronic inflammatory response syndrome (CIRS) framework, a proposed construct rather than an established diagnostic entity. Materials describing the framework state that approximately 24% of the population carries a susceptible type; no derivation of that figure was identified in the sources examined.Methods. Haplotypes were read from the publicly posted classification table and mapped to HLA-DRB1~DQB1 two-locus definitions. Because the table's DQ notation is internally inconsistent, two explicit readings were constructed. Carrier frequencies were calculated as 1 − (1 − Σp)² using National Marrow Donor Program frequencies, and tested against an independent British cohort, a United States organ-donor cohort, 1000 Genomes genotypes, family-phased haplotypes from the 17th International HLA and Immunogenetics Workshop, and 64 further world populations.Results. Under the broad reading, an estimated 88.4% of European Americans carry at least one listed haplotype; under the strict reading, 59.1%. Across four independent European-ancestry datasets the broad estimates range from 86.5% to 89.5%, while the strict estimates vary more widely, from 53.2% to 63.2%. Of 64 world populations with two-locus data, 28 met pre-specified data-completeness criteria; all 28 exceed 24% under the broad reading. The Hardy-Weinberg step was tested by direct enumeration of 2,690 individuals, where it overestimated carriage of the relevant DRB1 allele groups by 0.56 percentage points.Conclusions. Under both operational definitions, estimated carrier frequencies exceed 50% nationally, though they may overstate carriage of the original classifications where historical serological distinctions impose restrictions that two-locus data cannot resolve. The frequently cited 24% figure could not be reproduced from any dataset analyzed here; the originator attributes it to an unpublished clinical series. The historical literature reviewed describes a 111-person clinic-based comparison group, but its relationship to the population estimate remains uncertain. A positive result is informative only insofar as it is uncommon among people who do not have the condition. Taking the sensitivity values reported for the classification at face value, the carrier frequencies estimated here give positive likelihood ratios of 0.85 to 1.07 under the broad reading — a range containing 1.0, the value at which a positive result carries no information — and 1.3 to 1.6 under the strict reading.
Source: Preprints.org
Author: Diana Stafford
Language: English
Subfield: Genetics
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7224527309