Omicron, the SARS-CoV-2 B. 1.1. 529 variant of concern (VOC), was first detected in southern Africa in November, 2021, and its BA. 1 sublineage is now dominant in the UK. Omicron BA. 1 contains 32 coding changes in its spike protein (appendix p 2), and it is unclear to what extent its spread is driven by an intrinsic increase in transmissibility or escape from previous infection-induced and vaccine-induced immunity. In the UK, the BNT162b2 (Pfizer–BioNTech) and AZD1222 (ChAdOx1 nCoV-19, Oxford–AstraZeneca) COVID-19 vaccines were administered as part of a primary twodose course. A subsequent third dose of either BNT162b2 or mRNA1273 (Moderna) vaccine has been administered since September, 2021. To determine the ability of vaccineinduced antibodies to neutralise omicron, and to compare this to our previous measurements of VOC neutralisation by BNT162b2 and AZD1222, 1, 2 we carried out a third analysis of the Legacy study cohort (NCT04750356). The Legacy study was established in January, 2021, by University College London Hospitals and the Francis Crick Institute in London, UK, to track serological responses to vaccination during the national COVID-19 vaccination programme in healthy staff volunteers recruited prospectively, or following a positive COVID-19 test, after vaccination. Details of the methods and clinical cohort are available in the appendix. The Legacy study was approved by London Camden and Kings Cross Health Research Authority Research and Ethics committee (IRAS number 286469) and is sponsored by University College London Hospitals. Using a high-throughput live SARS-CoV-2 neutralisation assay, we determined neutralising antibody (NAb) titres (NAbTs) against omicron in 620 serum samples from 364 unique participants (appendix p 3) and compared these to NAbTs against alpha and delta VOCs, for which there is significant vaccine efficacy data correlated with NAbTs. 3 In participants sampled 2–6 weeks after two-dose vaccination with BNT162b2, most (166 [83%] of 199) had a quantifiable NAbT against omicron (median 50% inhibitory concentration [IC50] 122 [IQR 46–173]), which was seven-fold lower [95% CI 6· 3–7· 4] than NAbTs against alpha (median IC50 600 [IQR 384–1141]) and three-fold [95% CI 2· 8–3· 3] lower than NAbTs against delta (median IC50 301 [IQR 171–572]; appendix p 2). However, when sampled 12–16 weeks after two-dose vaccination with BNT162b2, only around half of participants (69 [51%] of 136) had a quantifiable NAbT against omicron, whereas nearly all still had a quantifiable NAbT against alpha (131 [96%] of 136) and delta (132 [97%] of 136). The drop in omicron NAbT in the 10 weeks after the second dose was significant (χ2 p< 0· 0001).

The same analysis of participants following two-dose vaccination with AZD1222 showed that less than half had a quantifiable NAbT against omicron 2–6 weeks after second dose (25 [37%] of 68), dropping further (five [19%] of 26) 12–16 weeks after second dose, whereas most participants had a quantifiable NAbT against alpha (59 [87%] of 68) and delta (52 [76%] of 68) 2–6 weeks after second dose of AZD1222 (appendix p 2). Notably, NAbTs after AZD1222 vaccination differed significantly according to whether participants reported experiencing COVID-19 symptoms (χ2 p< 0· 0001): those who had received two doses of AZD1222 and had not experienced COVID-19 symptoms before their second vaccine dose largely had no detectable