Objectives: To investigate the prevalence of spike-protein antibodies following at least 3 COVID-19 vaccine doses in immunocompromised individuals. Design Cross-sectional study using UK national disease registries of individuals with solid organ transplants (SOT), rare autoimmune rheumatic diseases (RAIRD) and lymphoid malignancies (LM). Setting: Participants were identified, invited and recruited at home by accessing the NHS Blood and Transplant Registry for those UK individuals who had received a SOT; and the National Disease Registration Service at NHS Digital for identifying individuals within England with RAIRD or LM. Participants: 101972 people were invited, 28411 recruited, and 23036 provided serological data, comprising 9927 SOT recipients, 6516 with RAIRD, and 6593 with LM. Interventions: Participants received a lateral flow immunoassay for spike-protein antibodies to perform at home together with an online questionnaire. Main outcome measures Odds of detectable IgG spike-protein antibodies in immunosuppressed cohorts following at least three COVID-19 vaccine doses by participant demographic, disease type, and treatment related characteristics Results: IgG spike-protein antibodies were undetectable in 23.3%, 14.1% and 20.7% of the SOT, RAIRD and LM cohorts, respectively. Participants had received three, four or ≥five vaccine doses at the time of testing in 28.5%, 61.8%, and 9.6%, respectively. In all groups, seropositivity was associated with younger age, higher number of vaccine doses and previous COVID-19 infection. Immunosuppressive medication reduced the likelihood of seropositivity: the lowest odds of seropositivity were found in SOT recipients receiving an anti-proliferative agent, calcineurin inhibitor and steroid concurrently, and those treated with anti-CD20 in the RAIRD and LM cohorts. Conclusions: Approximately one in five individuals with SOT, RAIRD and LM have no detectable IgG spike-protein antibodies despite three or more vaccines, but this proportion reduces with sequential booster doses. Choice of immunosuppressant and disease-type is strongly associated with serological response. Antibody testing could enable rapid identification of individuals who are most likely to benefit from additional COVID-19 interventions.