ABSTRACT
BackgroundCurrent testing guidelines for COVID-19 substantially underestimates the spread of SARS-CoV-2 in dense urban populations. Granular estimates of infection are important for understanding population-level immunity. We examined seroprevalence of anti-SARS-CoV-2 antibodies in Pune city in India and its implication for protective immunity. MethodsSeroprevalence was estimated during July 20-August 5, 2020 from 1659 randomly selected individuals recruited from five administrative Pune sub-wards (combined population 366,984). Prevalence of anti-SARS-CoV-2 spike protein antibodies were estimated and along with correlates of virus neutralisation. FindingsSeropositivity was extensive (51{middle dot}3%; 95%CI 39{middle dot}9-62{middle dot}4) but varied widely in the five localities tested, ranging from 35{middle dot}8% to 66{middle dot}4%. Seropositivity was higher in crowded living conditions in the slums (OR 1{middle dot}91), and was lower in those 65 years or older (OR 0{middle dot}59). The infection-fatality ratio was estimated at 0.28%. Post survey, COVID-19 incidence was lower in areas noted to have higher seroprevalence. Substantial virus-neutralising activity was observed in seropositive individuals, but with considerable heterogeneity in the immune response and possible age-dependent diversity in the antibody repertoire. InterpretationDespite crowded living conditions having facilitated widespread transmission, the variability in seroprevalence in localities that are in geographical proximity indicates a heterogenous spread of infection. Declining infection rates in areas with high seropositivity suggest population-level protection and is supported by substantial neutralising activity in asymptomatically infected individuals. The heterogeneity in antibody levels and neutralisation capacity indicates the existence of immunological sub-groups of functional interest. FundingPersistent Foundation, Pune, India RESEARCH IN CONTEXTO_ST_ABSEvidence before this studyC_ST_ABSWe searched the literature (upto 2 Nov 2020), using the terms "seroprevalence", "serosurveillance", "seroepidemiology", "immune response", "seroconversion" and "SARS-CoV-2," without any article type restrictions, and selected only population- or community-level seroprevalence studies for collecting background information. The survey of literature indicated that community serosurveys for SARS-CoV-2 in LICs and LMICs have been limited and have largely reported correlations of seroprevalence with demographic factors. There are no reports of protective immunity-associated characteristics in community surveillance settings from LMIC/LICs. In fact, such studies from the global North are also limited. The existing evidence thus lacks granular details critical to understand community-level heterogeneities, and provides limited epidemiological data without meaningful immunobiological correlates. Added value of this studyThis is the first systematic study (at the time of submission) from a LMIC reporting community SARS-CoV-2 sero-surveillance of high granularity alongside estimation of correlates of immune protection. We estimated seroprevalence as well as serological correlates of protection in a cross-sectional cohort of 1659 asymptomatic participants from five small urban localities in the metropolitan city of Pune, India. IgG seroprevalence was determined against the receptor-binding-domain (RBD) of the SARS-CoV-2 spike protein, to aid correlation with immune protection since RBD is the predominant target for neutralising antibodies. Large subsets of the sera were also tested for surrogate neutralisation as well as live SARS-CoV-2 virus neutralisation, data not so far reported in community sero-surveillance studies. We identified substantial locality-specific variations in seropositivity levels and infection fatality rates (IFRs), highlighting heterogeneities of infection behaviour even in dense, urban populations often lost in more global analyses. Notably, the incidence of new infections after the sero-sampling period revealed a strong negative association with seropositivity, indicating potential modification of transmission by community immunity. While RBD-specific antibody levels expectedly showed broad correlation with neutralisation capacities, 30% of individuals showed significant departures from this correlation, again underlining significant immune response heterogeneities. Implications of all the available evidenceHigh seroprevalence in the dense urban localities of the study site, despite a protracted and stringent lockdown, provides a realistic account on transmission dynamics crucial for public health policies in LMICs. Micro-geographic variability within locales, dominated by sub-optimal living conditions, needs to be acknowledged and used to develop measures designed for people in such socio-economic contexts. The heterogeneity of correlation between RBD seropositivity and neutralising capacity, as well as the complex association with age encountered in this study open up a plethora of research questions into epitope dominance and affinity variations in anti-viral antibodies in asymptomatic infection.