Thread by @lgraca

Proud to present BA.5-infection risk drops after BA.1/BA.2 infection – data from the national Portuguese registry, one of the countries first hit with BA.5. Important, as adapted vaccines in development are BA.1-based. Follow the twitorial! 1/9 @lgraca medrxiv.org/content/10.110…

First, we used the SARS-CoV-2 genetic diversity surveillance data (great work from @irj_pt and J.P. Gomes) to define periods where each variant was dominant (>90% of national sample isolates). 2/9 #COVID19PT #DGS #INSA #InstitutoRicardoJorge

Next, we extracted from the Portuguese COVID-19 registry (SINAVE) all infections from periods of dominance by each variant. SINAVE is very comprehensive, and until recently, there was a large testing effort in the country. Kudos to @DGSaude for maintaining the registry. 3/9

We were then in a position to calculate the risk for BA.5 infection (dominance after June 1st) in people initially infected in a period of dominance of Wuhan, Alpha, Delta, BA.1/BA.2, as well as individuals without a documented infection. 4/9

It is almost a prospective study, but in the past: groups of people with infection with different variants (and without two or more infections) were selected. Then, we followed what happened to them during the period between June 1st - july 4th, all under similar conditions! 5/9

Compared with people without a documented infection, one past infection with any variant reduces the risk of BA.5 infection, especially for people infected with BA.1/BA.2 that have a risk about 5 times lower than the uninfected. 6/9

Note that these are results from a highly vaccinated population: the Portuguese population aged 12 and older (in the study) had 98% coverage with the primary vaccination series by the end of 2021 (the total in the graph includes ages 0-11). 7/9
#CTVC #COVID19vaccines @DGSaude

What about unreported cases among the “uninfected”? We tested that. Assuming an additional 29.2% of unreported infections (data from the @irj_pt national serologic survey), results don’t change much. Even if unreported cases were 20% or 40%. (graph: assumes 29.2% unreported) 8/9

This work was a collaboration with M Carmo Gomes, Ruy Ribeiro, @VlterRFonseca1, the team from DSIA of @DGSaude The last word goes for the PhD student that was dealing with the tables and scripts @jtmalato 9/9
@IMMolecular @medicina_lisboa @FC_UL