Model-based estimation of transmissibility and reinfection of SARS-CoV-2 P.1 variant

Abstract

Background: The SARS-CoV-2 variant of concern (VOC) P.1 (Gamma variant) emerged in the Amazonas State, Brazil, in November 2020. The epidemiological consequences of its mutations have not been widely studied, despite detection of P.1 in 36 countries, with local transmission in at least 5 countries. A range of mutations are seen in P.1, ten of them in the spike protein. It shares mutations with VOCs previously detected in the United Kingdom (B.1.1.7, Alpha variant) and South Africa (B.1.351, Beta variant).

Methods: We estimated the transmissibility and reinfection of P.1 using a model-based approach, fitting data from the national health surveillance of hospitalized individuals and frequency of the P.1 variant in Manaus from December-2020 to February-2021.

Results: Here we estimate that the new variant is about 2.6 times more transmissible (95% Confidence Interval: 2.4-2.8) than previous circulating variant(s). Manaus already had a high prevalence of individuals previously affected by the SARS-CoV-2 virus and our fitted model attributed 28% of Manaus cases in the period to reinfections by P.1, confirming the importance of reinfection by this variant. This value is in line with estimates from blood donors samples in Manaus city.

Conclusions: Our estimates rank P.1 as one of the most transmissible among the SARS-CoV-2 VOCs currently identified, and potentially as transmissible as the posteriorly detected VOC B.1.617.2 (Delta variant), posing a serious threat and requiring measures to control its global spread.

Keywords: Dynamical systems; Viral infection.

Fig. 1. Diagram of the extended deterministic compartmental model (SEAIHRD).

The model compartments and the respective connections between them are summarized in this diagram, and they are named as S: Susceptible, E: Exposed (pre-symptomatic), H: Hospitalized (severe infected individuals), I: Infected (symptomatic individuals, not hospitalized), A: Asymptomatic. D: Deceased, R: Recovered. Compartments are subdivided into three age categories, not represented here for simplicity. Compartments with subindex 1 represent the wild-type variant, subindex 2 refers to the VOC P.1. Continuous lines represent flux between each compartment; dashed lines, infection pathways. Small arrows indicate force of reinfection and transmissibility. λ = force of infection. β = relative transmission rate. p_r = relative force of reinfection. γ = average time between being infectious and presenting symptoms. σ = proportion of severe cases that require hospitalization. α = proportion of asymptomatic cases. ν_s = average time between being infectious and recovering for severe cases. ν_i = average time between being infectious and recovering for mild/asymptomatic cases. μ = in-hospital mortality ratio.

Fig. 2. Hospitalization cases and frequency of the P.1 variant in Manaus city.

a Weekly new hospitalized COVID-19 cases in Manaus city. Grey line represents the fitted values of total cases (all variants) by maximum likelihood estimation (MLE) of the parameters. Red and blue lines represent the predicted values of cases due to P.1 and wild-type variants, respectively. Black dots are nowcasted observed data of hospitalizations. Panels b, c show the fittings to the time-series frequency of P.1 on datasets provided by previous works, which were aggregated monthly and weekly, respectively. The area around the lines indicate the 95% CI of the expected values. Dots and lines are the sample proportions of P.1 in sequenced genomes, and their 95% sample CI. The fitted values of the model parameters are presented in Table 2.