Additional analysis for AZD1222 from an ongoing UK phase II/III trial (COV002) - preprint available
here - with a focus on variants.
Participants were a mixture of recipients of half/full and full/full two dose regimens. Weekly self-administered PCR swabs were sequenced to facilitate determining efficacy with respect to variants.
Efficacy (measured at 14 days post second dose) seen against symptomatic non-B.1.1.7 disease of 84.1% (95CI: 70.7%—91.4%) compared to 74.6% (95CI: 41.6%—88.9%) efficacy against symptomatic B.1.1.7 - so a small reduction in efficacy against the new "UK"/"Kent" variant (minus E484K). For asymptomatic infection the efficacies were determined to be 75.4% (95CI: 39.9%—89.9%) for non-B.1.1.7 versus 26.5% (95CI: -112.0%—74.5%) for B.1.1.7 (note the paucity of data in the later case leading to huge confidence intervals).
So overall efficacy against the B.1.1.7 variant from all cases was
66.5% (95CI: 37.1%—82.1%) compared to 80.7% (95CI: 69.2%—87.9%) against other variants (neutralising titres from vaccine recipients were 9-fold lower against the B.1.1.7 lineage than against an original lineage, which, not unsurprisingly, hints at a more complex relationship to immunity than direct measurement from sera).
Finally a paper that refers to cycle threshold values. The weekly swabs from the vaccinated were lower than the control group which
might point towards reduced transmission. Notes of caution:
(i) bear in mind poor time resolution,
(ii) self-administered sample collection,
(iii) those in the category of "non/failed sequenced swabs" had such widely varying efficacies (3% v -29%) from sequenced (75%) that there might be a not insignificant number of false-positive PCR results skewing the calculations,
(iv) the numbers of sequenced data points are fairly small and there is also still quite a significant spread in viral loads in both non-sequenced arms (vaccinated and control) of the study,
(v) the paper didn't detail the age profile of the participants in this sub-study (only 15% of participants enrolled in the overall COV002 study are reported to be over 55).
(vi) B.1.1.7 was on the increase during the sub-study period but was not yet dominant and didn't represent the majority of cases.