Panel A: A flowchart starts with identifying 558 CVID individuals and 21,882 rituximab-treated patients. It filters these patients to those with at least 3 antibiotic-treated sinusitis encounters and matches them 1:10 with controls based on sex, age, and EHR time. Exclusion criteria include primary immunodeficiency, cystic fibrosis, structural airway abnormalities, autoimmune vasculitis, sarcoidosis, head and neck malignancy, and any measured IgG below 700 milligrams per deciliter. Encounter states (S1-S6) are defined by sinusitis codes and antibiotic type. Panel B: A timeline example showing simple counting versus windowing methods for quantifying episodes. Simple counting results in 11 episodes, while windowing with 28-day search and 14-day extension windows results in 6 episodes. The timeline includes orange circles for antibiotic-treated encounters, gray circles for excluded encounters, black lines for single-encounter episodes, and blue bars for extended episodes.
Study cohort assembly and episode quantification methodology. (A) CVID and RTX patients with hypogammaglobulinemia were filtered to ≥3 antibiotic-treated sinusitis encounters and matched 1:10 with controls on sex, age (±2 years), and EHR time (±180 days). Exclusion criteria included primary immunodeficiency (ICD-10 D80–D84), cystic fibrosis, structural airway abnormalities, autoimmune vasculitis, sarcoidosis, head and neck malignancy, and any measured IgG below 700 mg/dl. Encounter states (S1–S6) defined by sinusitis codes (J01, J32) and antibiotic type. (B) Example timeline: simple counting (11 episodes) versus windowing with 28-day search and 14-day extension windows (6 episodes). Orange circles, antibiotic-treated encounters (simple counting); gray circles, S3/S6 encounters excluded from counting; black lines, single-encounter episodes (windowing); blue bars, extended episodes (windowing).