•Immunogenicity of BNT162b2 vaccine were evaluated using two immunoassays.•Elderly Japanese residents showed lower antibody responses than staff.•Aging, steroid and NSAID usage are independent risk factors for lower responses. To evaluate antibody responses against the primary series of vaccination of severe acute respiratory syndrome coronavirus-2 SARS-CoV-2 vaccines in the staff and residents of Japanese geriatric intermediate care facilities. All subjects (159 staff and 96 residents) received two doses of the BNT162b2 mRNA vaccine 3 weeks apart. Baseline data of subject were collected using a structured form. Serum samples were collected three times: before vaccination, 3 weeks after the first dose, and 4 weeks after the second dose, and anti-receptor binding domain of the spike protein of SARS-CoV-2 anti-RBD IgG was measured using two immunoassays. After the second dose, geometric mean titers GMT of anti-RBD with both the Abbott and Roche assay were significantly lower in residents than staff (2282 AU/mL vs. 8505 AU/mL, and 258 U/mL vs. 948 U/mL, respectively). Multivariate analysis of characteristics affecting antibody responses (≥1280 AU/mL for Abbott and > 210 U/mL for Roche) showed lower odds ratios ORs for older age (adjusted OR per 10 year increase aOR = 0.62, 95 % confidence interval 95 %CI; 0.38–1.02), steroid usage (aOR = 0.09, 95 %CI; 0.01–0.60) and regular nonsteroidal anti-inflammatory drugs NSAIDs usage (aOR = 0.16, 95 %CI; 0.03–0.88). Elderly people and steroid and NSAID users had lower antibody responses following the second vaccine dose.