Novavax: >90% efficacy in UK but 60% in South Africa, hints of lower effect in HIV positive participants

On 28 January 2021, Novavax released preliminary top-line results from two studies using it’s NVX-CoV2373 spike protein-based vaccine candidate. [1]

This was at the same time as concern about the B.1.351 variant first started to dominate news in the UK. [1]

The phase 3 study run in the UK enrolled more than 15,000 participants between 18-84 years of age, including 27% over the age of 65. It reported 89.3% efficacy (95% CI: 75.2 to 95.4), based on the first occurrence of PCR-confirmed symptomatic (mild, moderate or severe) COVID-19, at least 7 days after the second study vaccination.

This analysis was based on 62 cases, 56 vs 6 in the placebo vs active group, respectively. Of the 62 cases, 61 were mild or moderate, and 1 was severe (in placebo group). Approximately half the cases were the B.1.1.7 variant and a post hoc analysis calculated 95.6% efficacy against the original COVID-19 strain and 85.6% against B.1.1.7.

Results from a phase 2b study in 4,400 participants in South Africa reported 60% efficacy (95% CI: 19.9 to 80.1) in the 94% of participants who were HIV negative. This was based on 44 cases in 2,536 participants: 29 vs 15 in the placebo vs active groups, respectively. Data available for 27/44 showed 92% to be the B.1.3.5.1 variant.

However, when results from the 148 (6%) of HIV positive participants were added, overall efficacy dropped to 49.4% (95% CI: 6.1 to 72.8). This was due to worse outcomes in the vaccine group: 2.6% vs 5.5% had symptoms in the placebo vs active group. Although these figures are in the wrong direction, the wide confidence intervals means this might not be a real effect.

Importantly. roughly one-third of the 4,400 participants (not included in the analysis) were seropositive at baseline indicating previous COVID-19 infection. This was likely the pre-variant strain, but most infections during the study were with the B.1.351 variant. So although the vaccine provided sginificant although not complete protection against the B.1.351 variant.