Panel A: A flowchart starts with 414,830 participants with WGS data, splits into CeD and Non-CeD groups, and applies various exclusion criteria and propensity matching to arrive at final cohorts of 2,899 CeD and 14,495 Non-CeD participants. Panel B: A table showing demographic characteristics of the matched cohorts, including gender, race, ethnicity, education, genetic sex, and genetic ancestry. Panel C: A Venn diagram showing the route of CeD diagnosis for the included participants, with categories for EHR, self-report, and seropositive CeD.
CONSORT flow diagram and cohort demographics for CeD analysis. (A) CONSORT-style flow diagram illustrating the steps of data processing and cohort selection. Propensity score matching was performed based on age, sex, and genetic ancestry. Confounding was reduced by excluding individuals with comorbidities commonly associated with CeD, including T1D, autoimmune thyroid disorders, selective IgA deficiency, IBD, CVID, cancer, and HIV infection, from the non-CeD control group. (B) Demographic characteristics of 5:1 propensity-matched non-CeD controls and CeD patients from All of Us. GED, general educational development. (C) Venn diagram showing the route of CeD diagnosis for the included participants. “Seropositive CeD” refers to individuals with tTG-IgA or DGP IgA or IgG levels >20 IU/ml documented in their EHRs. IBD, inflammatory bowel disease.