December 14-20 2020
The pandemic in Brazil: could herd-immunity be a myth?
Infection by SARS-CoV-2 does not seem to confer herd-immunity when the virus spreads to large cities.
Brazil is one of the countries worst affected by COVID-19 today. Manaus, the capital and largest metropolis of the Amazonian region with more than 2 million inhabitants and a population density of 158 inhabitants / km2, is one of the worst affected cities.
If we suppose an R0 = 2,5 à 3,0 for the Amazonian region (this figure represents the expected number of cases resulting directly from one infected case) then the infection rate should reach 89% to 94%. So, when this percentage is greater than the figure generally accepted as bringing about herd-immunity (from 60% to 67%) then each infection should generate less than one new infection.
But do these statistics hold up in relation to this large city? Recent research studied the rate of infection by the detection of IgG specific antibodies in a SARS-CoV-2 protein (the N protein) by testing samples of blood from the blood bank in Manaus.
It was reported that the sensitivity of the test carried out on these blood bags was imperfect. So the test was initially carried out on different cases of COVID-19, about twenty days after symptoms first appeared. In patients made seriously ill by SARS-CoV-2, the test proved sensitive in 91,8% of cases, suggesting therefore that about 8% of severe cases do not generate a detectable concentration of IgG anti-N antibodies. But from 50 to 121 days after infection, in the same patients made moderately ill by the virus, the percentage changed to 80,4%, indicating a declining production of antibodies.
In addition, blood from the blood banks in Manaus and São Paulo was tested between February and October 2020: in February and March, less than 1% of samples were positive, a coherent result since the first case in Manaus was announced on the 13th of March and the first case in São Paulo on the 12th of February. In April, 44,5% of samples from Manaus were positive and in June, 52,5%. In São Paulo, this percentage also increased, reaching 13,6% in June. Between June and October, seroreversion (the return to normal levels of antibody production), in combination with the effects of measures such as mask-wearing, reduced levels, with only 25,8% of samples positive for IgG anti-N in October in Manaus. Taking into account the depletion of IgG anti-N during the months following infection, the corrected percentage level of positive samples may have reached 66,2% in July and up to 76% in October, as the following diagram illustrates:
Estimation of the proportion of the Manaus population infected by SARS-CoV-2 (anti-N seropositivity)
Contrary to what might have been expected, although the city of Manaus had more cases of infections, more deaths were reported by the city of São Paulo: up to the 1st of October 2020, 2 642 deaths per million of inhabitants were reported in Manaus whereas there were 12 988 deaths per million of inhabitants in São Paulo, including high numbers of elderly residents.
This study shows the impact the epidemic can have, without significant government intervention, in cases where population is very dense or where the socio-economic context is unfavourable. In Manaus there was no evidence of herd-immunity against SARS-CoV-2, normally operational when between 60% and 67% of the population is infected by a virus. Infection by SARS-CoV-2 does not seem to confer long term immunity, and this may explain the results. We are far from any concrete evidence for herd-immunity.