Incidence and seasonality of respiratory viruses among medically attended children with acute respiratory infections in an Ecuador birth cohort, 2011–2014

Abstract Background Ecuador annually has handwashing and respiratory hygiene campaigns and seasonal influenza vaccination to prevent respiratory virus illnesses but has yet to quantify disease burden and determine epidemic timing. Methods To identify respiratory virus burden and assess months with epidemic activity, we followed a birth cohort in northwest Ecuador during 2011–2014. Mothers brought children to the study clinic for routine checkups at ages 1, 2, 3, 5, and 8 years or if children experienced any acute respiratory illness symptoms (e.g., cough, fever, or difficulty breathing); clinical care was provided free of charge. Those with medically attended acute respiratory infections (MAARIs) were tested for common respiratory viruses via real‐time reverse‐transcription polymerase chain reaction (rRT‐PCR). Results In 2011, 2376 children aged 1–4 years (median 35 months) were enrolled in the respiratory cohort and monitored for 7017.5 child‐years (cy). The incidence of respiratory syncytial virus (RSV) was 23.9 (95% CI 17.3–30.5), influenza 10.6 (2.4–18.8), adenoviruses 6.7 (4.6–28.0), parainfluenzas 5.0 (2.3–10.5), and rhinoviruses, bocaviruses, human metapneumoviruses, seasonal coronaviruses, and enteroviruses <3/100 cy among children aged 12–23 months and declined with age. Most (75%) influenza detections occurred April–September. Conclusion Cohort children frequently had MAARIs, and while the incidence decreased rapidly among older children, more than one in five children aged 12–23 months tested positive for RSV, and one in 10 tested positive for influenza. Our findings suggest this substantial burden of influenza occurred more commonly during the winter Southern Hemisphere influenza season.


| BACKGROUND
There is insufficient information about the burden and timing of respiratory virus epidemics among young children in tropical low-and middle-income countries (LMICs) to adequately guide mitigation. [1][2][3] Acute respiratory illnesses, typically caused by respiratory viruses, are common causes of morbidity among young children and are frequently associated with pneumonia, hospitalization, and death. Investment in their prevention and control depends, in part, on understanding their incidence and seasonality. Much is known about respiratory viruses in temperate high-income countries; comparatively little is known from tropical LMICs where most of the mortality occurs.
For example, in 2018, influenza caused >10 million acute respiratory infections among children aged <5 years. 1 Such annual epidemics disproportionately affect tropical LMICs. Epidemics in more northern latitudes typically occur October-March; those in southern latitudes occur April-September, but it is often unclear when epidemics occur in tropical latitudes. 4 Understanding when epidemics occur helps inform disease burden modeling, vaccine formulation selection, and vaccination timing. In the tropics, however, influenza epidemics can seem unpredictable and their associated disease burden poorly characterized.
Like influenza, respiratory syncytial virus (RSV) causes substantial morbidity and mortality. For example, in 2015, RSV caused >30 million acute respiratory infections and an estimated 59600 in-hospital deaths in children aged <5 years. 4 While there are currently no RSV vaccines in use to mitigate this burden, there are several in development. Countries can use monoclonal antibodies during RSV epidemics to prevent RSV in high-risk children, but these are typically costly and infrequently used in LMICs. Like influenza, RSV epidemics in temperate countries typically occur during cool weather; RSV seasonality in humid, tropical climates, however, is not well characterized. Ecuador, with three geoclimatic zones (coastal, Andean, and Amazonian), has complex respiratory virus epidemics 5 and limited burden information to justify sustained investments in mitigation. 6 In 2006, Ecuador's Ministry of Public Health introduced influenza vaccination among pregnant women and infants aged 6 months through 5 years 7 and subsequently recommended the use of antivirals for the treatment of persons with severe or progressive community acquired pneumonia during influenza epidemics. 8 Ecuador is the only South American country, however, to use a Northern rather than Southern Hemisphere vaccine formulation against influenza illness 9 in part because of uncertainty about its influenza seasonality. While Ecuador does not currently use monoclonal antibodies within the public health system to prevent severe RSV illness, it does launch annual handwashing and respiratory hygiene campaigns to prevent contagion with respiratory viruses.
To substantiate the need for investments in respiratory virus prevention and optimize the timing of interventions, we used systematically collected birth cohort data from Ecuador to (i) estimate the incidence of medically attended influenza, RSV, and other respiratory viruses; (ii) assess whether influenza incidence was highest during the April-September Southern Hemisphere season versus the October-March Northern Hemisphere season; and (iii) explore peak epidemic weeks of influenza and RSV activity.

| METHODS
A description of the Ecuador Life (ECUAVIDA) study has been previously published. 10  On January 1, 2011, children still participating in ECUAVIDA were consented into the respiratory cohort and followed prospectively to identify medically attended acute respiratory infections We calculated the incidence of MAARI and MAARI-associated respiratory viruses using previously described methods. 11 We accounted for under-ascertainment of influenza, RSV, adenovirus, and parainfluenza virus 1-3 among children with MAARIs who were not swabbed by multiplying their frequency by the proportion of samples testing positive for these viruses at the NRL surveillance in the same age group and epidemic week. We calculated participating children's person time as the number of days from study enrollment (January 1, 2011) to study conclusion (June 30, 2014), last completed checkup visit, or eighth birthday and excluded days when parents missed routine checkups and/or 14 days after a MAARI event. We also restricted the risk period to weeks when specific viruses were circulating in Ecuador according to NRL. Resulting rates incorporated the variance in the age-adjusted proportion of samples testing positive per epidemic week and the sensitivity of the immunofluorescence assay used by NRL to test surveillance samples for non-influenza viruses. 12 To assess influenza and RSV epidemic timing, 9 we calculated the average proportion of samples testing positive for these viruses each month; we did not assess the epidemic timing of other respiratory viruses because we did not anticipate sufficient detections for a meaningful analysis. The start of the epidemic was defined as the first month when the proportion of positive samples was greater than the annual mean for ≥2 months, and the end was defined as the first month when the proportion of influenza positive samples remained below the annual mean for ≥2 months. 13 We also calculated the influenza rate ratio during April-September Southern Hemisphere season versus October-March Northern Hemisphere season.
The study protocol was approved by the Bioethics Committee of the Universidad San Francisco de Quito, Ecuador (Protocol 6-11-2010). Informed written consent for participation in the study was obtained from parents or legal guardians of children.
Half (49%) of the children were female, and most (74%) were from mothers who self-identified as Mestizo (i.e., of mixed Spanish and indigenous descent, Table 1). Twenty-four percent had received at least one influenza vaccine between ages 6 and 24 months, and 17% had received two doses. The 1984 (84%) children who completed the study had similar demographics to those who did not; 221 (9%) moved from the study area, and 102 (4%) withdrew; 67 (3%) missed their last checkup, and two (0.1%) died because of unknown reasons.