SARS‐CoV‐2 infections in households in a peri‐urban community of Lima, Peru: A prospective cohort study

Abstract Background We assessed the prevalence and incidence of SARS‐CoV‐2 infections in a prospective study of households in Lima, Peru. Methods Households with a child, a young adult 18–50 years, and an adult age >50 years in peri‐urban Lima were followed with twice‐a‐week household visits during a 2‐month period. Nasal swabs and saliva specimens were collected twice weekly, and nasopharyngeal swabs were collected weekly from each participant, regardless of symptoms. Laboratory‐confirmed SARS‐CoV‐2 infection was defined by two RT‐PCR tests from any of the collected specimens within a week. Blood samples collected at enrollment and end of follow‐up were tested with rapid serological tests. We calculated the prevalence and incidence of laboratory‐confirmed SARS‐CoV‐2 infections. Results We enrolled 132 participants from 44 households: 44 children, 44 young adults, and 44 older adults. A total of 13 SARS‐CoV‐2 infections were detected in eight households, for an overall period prevalence of 9.85% (95% confidence interval [CI]: 5.35–16.25). Most (61.54%) infections were symptomatic. Eight of 11 (72.73%) SARS‐CoV‐2 detections corresponded to the Lambda variant. During 218.79 person‐months at risk of follow‐up, there were six new SARS‐CoV‐2 infections detected (2.74 per 100 person‐month, 95% CI: 1.25–6.04). At enrollment, 59 of 128 participants tested had positive SARS‐CoV‐2 IgG serology (46.09%, 95% CI: 37.25–55.12). Five of six new infections occurred among participants with negative baseline serology. Conclusions We demonstrated high incidence of SARS‐CoV‐2 infections in households, especially among subjects without evidence of prior infection, most of them not detected by the Ministry of Health system.


| INTRODUCTION
The ongoing SARS-CoV-2 pandemic has disproportionally affected individuals with limited access to diagnostic and healthcare services. Yet, most available estimates of disease incidence come from populations with good access to those services, primarily high-income countries. [1][2][3][4] Moreover, a majority of incidence estimates are based on passive reporting of laboratory results from testing centers or healthcare locations. 4,5 Reliable estimates of incidence from the community are limited, and data from low-to middle-income countries (LMIC) are lacking.
Households, where individuals share common areas and usually interact in close proximity, provide an optimal venue for viral transmission. Whereas several prior studies have examined household transmission of influenza and other respiratory viruses, 6 few studies have examined the incidence of SARS-CoV-2 infections in households. 1,4 7 We conducted a prospective cohort study in households from a densely populated peri-urban area in Lima, Peru. We sought to investigate the frequency of SARS-CoV-2 infections, independent of respiratory symptoms, and explore the patterns of SARS-CoV-2 infections in households.

| METHODS
This was a prospective cohort study of households conducted from December 9, 2020, through March 19, 2021. Enrolled participants were followed through twice-a-week household visits, with systematic data collection on respiratory symptoms and healthcare use.
Respiratory specimens were systematically collected during household visits, regardless of the presence of symptoms, for laboratory detection of viral infections. Baseline and follow-up blood samples were obtained at home to determine serologic evidence of prior SARS-CoV-2 infections. No COVID-19 vaccines were available to the study population during the study period.

| Study population and selection criteria
Households were eligible for participation if they included at least three consenting members-one person <18 years, one 18-50 years, and one >50 years old-who were available during weekday working hours to be visited at home or in a nearby working area and had no plans to moving out of the area within the planned study follow-up period. Additional household members could not be enrolled due to budget constraints; sociodemographic information of these was obtained from consenting members; however, samples and follow-up information were obtained only from the three consenting members.

| Enrollment and follow-up
Identification of eligible households was performed through a house-to-house screening census within the study area in the San Juan de Lurigancho District of Lima. Households identified as eligible were approached for enrollment.

| Ethical approvals
The study was approved by the ethics review boards (ERBs) of the Instituto de Investigacion Nutricional and Vanderbilt University. All adults signed their own ERB-approved written informed consent form (ICF). For individuals under 18 years of age, either the mother or the father signed an ICF. Additionally, children 8 years and older signed a written informed assent form. All study participants were encouraged to use their regular health providers for any illness suspected to be related to COVID-19, including their own diagnostic methods and treatment protocols. Laboratory tests were conducted after the surveillance was terminated to maintain blinding of participants and study members on viral infection status. After study completion, a final home visit was done to inform study participants about the study results and their own samples' results.

| Collection of study specimens
During follow-up household visits, trained fieldworkers collected twice-weekly nasal swabs and weekly nasopharyngeal (NP) swab specimens. Nasal swabs were collected using Polyester Swabs (Puritan ® ) in viral transport media (VTM Remel ® ). NPs were collected using rayon swabs in STGG media. After collection, samples were maintained in cold boxes and transported within 4 h of collection to the field laboratory. Study participants were trained to self-collect saliva samples on the same days nasal swabs were taken. For saliva sample collections, study participants were given a non-sterile capped plastic container and instructed to fill it with about 5 ml of saliva passively collected either before breakfast or at least 3 h after eating, close the container tightly, and keep the samples in refrigerators (if available) until samples were picked up by fieldworkers.
In addition, we obtained venous blood samples (3 ml) from participants at enrollment and at the end of the 2-month surveillance period.
Blood samples were transported at room temperature to the field laboratory where serum was separated, and two aliquots were initially frozen at À20 C. At the end of each week, blood samples were transported to the central laboratory for long-term storage at À80 C.

| Laboratory procedures
The saliva and nasal swab specimens were distributed in a maximum of four aliquots in cryovials as soon as they were received by the laboratory. Each aliquot was frozen at À80 C for long-term storage. To increase the efficiency of RT-PCR testing, we created separate pools of nasal and saliva samples combining sample aliquots (100 μl) collected on the same week from all members of the same household (pools sizes ranged from 2 to 12 samples). 8 NP samples were tested separately and on an individual basis (i.e., without pooling). All respiratory and saliva samples were tested for SARS-CoV-2 using RT-PCR. If  Prevention-CDC's operational definition, 10,11 that is, when participants reported at least 1 day with: cough, difficulty breathing, or loss of smell or flavors, or the combination of two or more of fever (reported or measured), throat sore, nausea or vomiting, diarrhea, or nasal congestion during the episode. Infection episodes without any of the aforementioned COVID-like symptoms were defined as asymptomatic.

| Data analysis
In-house data entry and management software were used to enter data into a study database, with internal (within study forms) and external (between study forms) data quality checks. For prevalence estimates, we used infections detected at any time of the study period. Participants with positive samples at enrollment were included in the prevalence estimate but were not considered for incidence calculation (i.e., were not at risk). Because the follow-up period was short, we considered that a participant was no longer at risk once infected and did not attempt to assess recurrent infections. Incidence rates were first calculated considering all person-days observed during the surveillance period. We also estimated incidence rates selecting all person-days observed prior to the start of a new infection and including the day when the new infection definition was met, as period at risk. All person-days observed after an initial infection were considered not at risk. We then applied the incidence (person-time at risk) estimated to the census population of the sectors of Metropolitan Lima classified as peri-urban to estimate the number of SARS-CoV-2 infections per day and compare that number with the daily number of COVID-19 cases reported by the Peruvian Ministry of Health for that sector of Lima, as well as the daily number of COVID-19 attributable deaths reported by the national system of death certificates. 12 3 | RESULTS

| Enrollment
In the screening census of the study area, 631 potential eligible households were initially identified. After approaching 114 eligible households by trained fieldworkers, 44 households were enrolled in the study, with one individual for each study age group providing written consent/assent ( Figure 1).

| Study population
A total of 132 subjects from 44 households were enrolled in the study. Figure 2 shows the GPS location of the 44 households enrolled, which were homogeneously distributed within the study area and among those eligible households. The mean age of the 44 study participants <18 years old was 3.66 years AE 2.50, range 1-17 (59.09% female); for the 44 participants 18-50 years old, the mean age was 32.93 years AE 7.08, range 19-49 (95.45% female); and for the 44 participants >50 years old, the mean age was 63.75 years AE 6.79, range 51-85 (75.00% female). The sociodemographic characteristics of the 40 enrolled households that completed the socioeconomic survey are described in Table 1. Most basic services (e.g., water and sewage) were available in enrolled households. Few enrolled members were students, due to school closures due to the pandemic and being at summertime when schools are closed. Groceries purchasing were done either daily or one to three times per week in the local food market, and most household members utilized different types of public transportation. From the 44 households initially enrolled, 38 completed follow-ups, encompassing a total of 114 household members for an 86.36% retention through the end of the study (Figure 1).

| Serological evidence of prior infection
We obtained a blood sample from 128 study members at baseline and 111 at end of the surveillance period. A baseline positive IgG test was

| Prevalence and incidence of SARS-CoV-2 infections
There were 13 SARS-CoV-2 infections detected among the 132 persons observed (9.85% period prevalence, 95% CI: 5.35-16.25) (  The age-adjusted incidence rate of SARS-CoV-2 infections detected by RT-PCR was 5.8 (95% CI, 3.1-8.4) per 1000 person-weeks at risk, which is equivalent to 2.52 infections per 100 person-months, like the overall incidence rate observed in our study. 7 A higher incidence rate   15 suggesting a high intensity of prior infections in the study area.
During this study, Peru started introducing the Sinopharm vaccine to the healthcare providers working in the first line of pandemic control, but no vaccination occurred in our study population. As shown in

ACKNOWLEDGEMENTS
The authors would like to express their appreciation to the participating household members that were able to provide us with the high number

ETHICS STATEMENT
The study was approved by the ethics review boards (ERBs) of the Instituto de Investigacion Nutricional and Vanderbilt University. All participants (or one parent for participants <18 years of age) signed an ERB-approved written informed consent form. Children 8-17 years of age signed their own assent form.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/irv.12952.

DATA AVAILABILITY STATEMENT
Data without any participant's identifying information, could be provided after all final analysis are completed, if requested to the correspondence author. After approval, and the signature of a data transfer agreement form, data could only be used for the purpose requested, unless a new request is made.