High anti-SARS-CoV-2 antibody seroconversion rates before the second wave in Manaus, Brazil, and the protective effect of social behavior measures: Results from the DETECTCoV-19 cohort.

We report anti-SARS-CoV-2 nucleocapsid IgG seroconversion rates from a population-based cohort in Manaus, between August and November 2020, before the COVID-19 second wave in Brazil. Among seronegative and uninfected participants, we observed an overall incidence of 13.06% (95%CI, 11.52-14.79) and 6.78% (95%CI, 5.61-8.10) for symptomatic seroconversion. Risk factors for seroconversion were: having a COVID-19 case in the household, not wearing a mask during contact, relaxation of protective measures, and non-remote work. We observed an intense infection transmission period that preceded the second wave of COVID-19 in Manaus, and identified several modifiable behaviors that increased the risk of seroconversion. SARS-CoV-2 in Manaus.


Abstract
We report anti-SARS-CoV-2 nucleocapsid IgG seroconversion rates from a population-based cohort in Manaus, between August and November 2020, before the COVID-19 second wave in Brazil. Among seronegative and uninfected participants, we observed an overall incidence of 13.06% (95%CI, 11.52-14.79) and 6.78% (95%CI, 5.61-8.10) for symptomatic seroconversion. Risk factors for seroconversion were: having a COVID-19 case in the household, not wearing a mask during contact, relaxation of protective measures, and non-remote work. We observed an intense infection transmission period that preceded the second wave of COVID-19 in Manaus, and identified several modifiable behaviors that increased the risk of seroconversion.

Main
Brazil ranks second in worldwide COVID-19 cases and deaths since the start of the pandemic 1 . The healthcare system in Manaus has collapsed twice in less than eight months, and the current second wave has surpassed the number of deaths and cases of the first wave ( Figure 1a). The current high toll in morbidity and mortality is aggravating the precarious state in which the first wave left the region, with important consequences for the families and communities affected. In September 2020, several alerts were made regarding a possible second wave of infection if social distancing measures were relaxed 2 . However, social distancing rules were further relaxed, and their effects reflected on population mobility ( Figure 1a); additionally, a research report on achieving herd immunity in Manaus could have further contributed to this behavior 3 . In this report, we estimated the incidence and risk factors associated with SARS-CoV-2 seroconversion from the DETECTCoV-19 cohort before the start of the second wave of infections in Manaus. In our previous cross-sectional baseline analysis of the DETECTCoV-19 cohort, we reported a crude seroprevalence of 29.10% in August 2020 4 . From our initial recruitment cohort (n=3057), a total of 2496 participants (81.64%) returned for a second follow-up visit between 19/10/2020 and 27/11/2020 (Figure 1b)  Next, we calculated the incidence of SARS-CoV-2 seroconversion. From the total participants evaluated at follow-up, 204 reported having COVID-19 infection prior to the first visit, and 24 had no data regarding this. From the remaining 2268, 1709 were IgG seronegative at first visit. Of them, 1424 had a RI ≤1.5 at the second visit and deemed still negative. Meanwhile, 214 had a RI >1.5 with a RI ratio between the second and first visit >2.0, and were considered as seroconversion. Seventy-one participants had a second visit positive IgG, but did not meet the criteria of doubling the RI, thus were considered indeterminate and were not included in the analysis (Supplementary Figure 1). We found a high incidence of seroconversion at 13.06% (95%CI, 11.52-14.79) with a median follow-up duration of 57 days (IQR, 54-61 days); in other words, 1% of the sample seroconverted every 4.5 days. Of these cases, 48.1% were asymptomatic; therefore, the incidence of symptomatic seroconversion was 6.78% (95%CI 5.61-8.10, Figure 1d, Table 1). Overall, it is important to timely identify the ratio between these divergent incidences and their dynamics during the pandemic in order to do adequate planning of resources and mitigation strategies.
Our serology-based incidence rate, 5-10-fold higher than the official reports, indicates that the bulk of transmission prior to the second wave occurred mostly undetected by the current official survey methods.
We then analyzed the effect of sociodemographic, health-related, behavioral, case clustering, and COVID-19 testing on seroconversion (Supplementary Tables 2-4). In contrast with our previous report, which evaluated prevalence of seropositivity during August 2020 4 , this study looked at factors that affected incidence over a median twomonth period between August and November 2020. While in our first report, we found that prevalence was strongly associated with sociodemographic characteristics -as male sex, older age, lower income, occupation, and number of household members, these were no longer associated with the emergence of new cases during the study period. In contrast, we found that the main risk factors were related to the social behavior of the participants, such as not keeping social distance before August, relaxing social distancing after that, on-site working, and having contact with COVID-19 patients without a mask ( Figure 2, Supplementary Table 5). These results show that, between the first and second epidemic waves in Manaus, behavioral risk factors that increase exposure to SARS-CoV-2 were more important than biological or social characteristics as sex or poverty. Having a person with COVID-19 diagnosis in the household affected both the prevalence in August and the incidence in the following two-three months with similar magnitude. Crucially, independent of the moment of the pandemic, our findings confirm that having a COVID-19 contact in the household remains one of the more robust predictors for acquiring the disease 5 . Not surprisingly, having symptoms or having been diagnosed during the observation period were strongly associated with seroconversion, more so than in the cross-sectional study, underlying the correlation between these events.
Since the start of the pandemic effective surveillance, availability and ease of access to testing, well implemented non-pharmaceutical interventions (NPI) and the public health response measures linked to testing such as quarantine and isolation have been essential in controlling the virus transmission [6][7][8] . Our findings reveal that COVID-19 index cases may have likely driven seroconversion -a representative proxy for infection burden-within the household, as observed in other high transmission settings 9 . Our data suggests that household close contacts of COVID-19 surviving or deceased patients should be tested regardless of symptoms and advised voluntary isolation 10 ; strict follow-up of cases and contacts is essential to reduce virus transmission 6,11 . NPI including social distancing, mask use and hygiene have been the pillar to reduce community SARS-CoV-2 transmission worldwide 12 . Therefore, lifting NPI abruptly diminishes the gains accumulated by previously implemented policies 7,13 . In our study, individuals who relaxed social distancing measures 14 or had contact with COVID-19 individuals without mask 15,16 had the highest risk of acquiring

SARS-CoV-2 infection.
Our work has some limitations. A convenience sampling strategy based on on-line and university website advertising potentially excluded individuals who did not have access to this information and may not completely represent the general population. We also had shortcomings in performing active surveillance to identify symptomatic infections and positive RT-PCR tests results. Additionally, most symptomatic participants had a mild or moderate infection, but our surveillance method cannot exclude that severe illness cases and deaths occurred among the non-returning participants, underestimating total seroconversion events. Additionally, antibody response kinetics and their variability among populations limit our interpretations 17 . Despite these, our longitudinal serology approach to assess exposure and burden, and the size of our cohort, made possible in-depth statistical analysis to identify the risk factors associated with seroconversion in a setting of high transmission and low NPI containment measures. Unfortunately, we could not evaluate the role of the infecting SARS-CoV-2 strains. The cohort period between August and November 2020 theoretically predates the surge and dominance of the P.1 (B.1.1.248) strain in the region ; however, we cannot rule out that the high rate of seroconversion in our cohort could have been influenced partially by the emergence of a more infectious strain, like the P.1 lineage in Manaus 18,19 . We hypothesize that the role of P.1 could be ascribed to accelerating the transmission rate observed after November, but we consider that the high seroconversion incidence found in our cohort might be explained by case clustering and host-related behavioral factors, as observed in high transmission settings 5,9,20 .
Rampant spread of SARS-CoV-2 infection in settings with low compliance to behavioral and NPI containment measures is a cause of concern due to its high human costs, elevated burden imposed on healthcare systems, and a possible impact on the emergence of new variants -favored in high transmission settings-which can negatively impact the effectiveness of available and future countermeasures including diagnostics, vaccines and therapeutics. This study provides timely actionable evidence for policymakers to inform next steps to mitigate the pandemic in Manaus and elsewhere.     Table 5 for details). *Not included in the multivariate model due to collinearity with the Household contact variable.

Ethics and Study design
The  All samples with a RI value above 1.5 (assay cut-off) were considered positive 4 . were considered statistically significant.

Declaration of interests
The authors declare no conflict of interest.