Burden and epidemiology of influenza‐ and respiratory syncytial virus‐associated severe acute respiratory illness hospitalization in Madagascar, 2011‐2016

Background Influenza and respiratory syncytial virus (RSV) infections are responsible for substantial global morbidity and mortality in young children and elderly individuals. Estimates of the burden of influenza‐ and RSV‐associated hospitalization are limited in Africa. Methods We conducted hospital‐based surveillance for laboratory‐confirmed influenza‐ and RSV‐associated severe acute respiratory illness (SARI) among patients of any age at one hospital and a retrospective review of SARI hospitalizations in five hospitals situated in Antananarivo during 2011‐2016. We estimated age‐specific rates (per 100 000 population) of influenza‐ and RSV‐associated SARI hospitalizations for the Antananarivo region and then extrapolated these rates to the national level. Results Overall, the mean annual national number of influenza‐associated SARI hospitalizations for all age groups was 6609 (95% CI: 5381‐7835‐rate: 30.0; 95% CI: 24.4‐35.6), 4468 (95% CI: 3796‐5102‐rate: 127.6; 95% CI: 108.4‐145.7), 2141 (95% CI: 1585‐2734‐rate: 11.6; 95% CI: 8.6‐14.8), and 339 (95% CI: 224‐459‐rate: 50.0; 95% CI: 36.3‐74.4) among individuals aged <5, ≥5, and ≥65 years, respectively. For these same age groups, the mean annual number of RSV‐associated SARI hospitalizations was 11 768 (95% CI: 10 553‐12 997‐rate: 53.4; 95% CI: 47.9‐59.0), 11 299 (95% CI: 10 350‐12 214‐rate: 322.7; 95% CI: 295.6‐348.8), 469 (95% CI: 203‐783‐rate: 2.5;95% CI: 1.1‐4.2), and 36 (95% CI: 0‐84‐rate: 5.8; 0.0‐13.5), respectively. Conclusion The burden of influenza‐ and RSV‐associated SARI hospitalization was high among children aged <5 years. These first estimates for Madagascar will enable government to make informed evidence‐based decisions when allocating scarce resources and planning intervention strategies to limit the impact and spread of these viruses.


| INTRODUC TI ON
Influenza and respiratory syncytial virus (RSV) infections are responsible for substantial global morbidity and mortality annually, with the highest burden experienced by young children and older adults. [1][2][3][4][5][6][7][8] In addition, a higher burden of influenza-and RSV-associated hospitalization has been reported among African children compared with children in other Regions. 2,5,8,9 Nonetheless, studies were limited to a pediatric population and estimates from Africa were obtained from a very limited number of participating countries. Estimates of the national burden of influenza-and RSV-associated hospitalization across age groups are severely limited in many African countries, having been described only in five countries for influenza, 3,10-13 and two countries for RSV 4,7,10 on the continent.
The World Health Organization (WHO) has highlighted the need for influenza disease burden estimates especially from low-and middle-income countries. These estimates would enable governments to make informed evidence-based decisions when allocating scarce resources and planning intervention strategies to limit the impact and spread of the disease. 14 The effectiveness of RSV candidate vaccines is being evaluated, 15

| DS-1: Number of SARI hospitalizations in Antananarivo Renivohitra District
Antananarivo Renivohitra District is situated in the Analamanga Region within Antananarivo Province where the capital city of the country is located. We obtained the total number of patients hospitalized with SARI in Antananarivo Renivohitra District during January 2011 through December 2016 through an anonymized retrospective record review implemented in three public university hospitals (Joseph Raseta Befelatanana hospital, Tsaralalana mother and child hospital, and Ambohimiandra mother and child hospital), one military hospital (Centre Hospitalier Universitaire de Soavinandrina (CENHOSOA)), and one private hospital (Saint-François d'Assise Ankadifotsy hospital). These hospitals were selected after a presurvey conducted to identify all health care facilities within the study district with the capacity of admitting patients with SARI. Six specialty hospitals such as eye or surgical hospitals were excluded from the pre-survey. Hospital admission books and medical records were reviewed to identify patients hospitalized with signs and symptoms consistent with the WHO SARI case definition reported below.
For each identified patient admitted at the selected hospitals, age, gender, dates of admission and discharge, and location of residence were recorded.

| DS-2: Influenza and RSV surveillance among patients hospitalized with SARI
We conducted active, prospective, hospital-based surveillance for SARI at the CENHOSOA from January 2011 through December 2016. Surveillance activities were conducted at the pediatric and adult pneumology wards. A SARI patient was defined as a hospitalized person of any age presenting with (a) either a recorded temperature ≥38°C or history of fever and (b) cough of duration of ≤10 days. 14 Trained surveillance nurses completed case report forms that included demographic, clinical, and epidemiological information for all SARI cases. In addition, respiratory specimens (nasopharyngeal and oropharyngeal swabs) were collected from all consenting patients. Specimens were placed in vials containing universal transport medium, stored at 4-8°C, and transported to the National Influenza Center (located at the Institute Pasteur de Madagascar) within 24 hours of collection for testing. Specimens were tested within 48 hours from collection for influenza A and B viruses and RSV using a real-time reverse transcription polymerase chain reaction assay. 16,17 Influenza A-positive samples were further subtyped. All patients meeting the SARI case definition were eligible for enrollment. Verbal informed consent was obtained from all patients prior to data and specimen collection. For children <15 years, verbal consent was obtained from a parent or legal guardian.

| DS-3: Prevalence of risk factors for pneumonia and health care seeking behavior for acute respiratory infection
We obtained the regional-level prevalence (

| Estimation of the national number and rate of influenza-and RSV-associated SARI hospitalization
To estimate the national number and rate of influenza-and RSVassociated SARI hospitalization, we used a four-step approach. In Step 1, we estimated the SARI hospitalization rate in Analamanga Region considered to be the base region in our estimation approach (where we conducted the retrospective record review and the prospective hospital-based surveillance for SARI). In Step 2, we estimated the SARI hospitalizations rates for the other regions using estimates from the base region. In Step 3, we estimated the influenza-and RSV-associated SARI hospitalizations rates using available virologic surveillance data for influenza and RSV. In Step 4, we obtained the number of influenza-and RSV-associated SARI hospitalizations using the estimated rates and the population at risk in each region. 13,20,21 The description of the estimation approach for each step is provided below and in Figure 1. The equations used for the estimations are provided in the Methods section of the Supporting Information. All estimates were obtained overall and within the following age categories: <5, 5-24, 25-44, 45-64, ≥65, and ≥5 years of F I G U R E 1 Method used to estimate the numbers and rates of influenza-and respiratory syncytial virus-associated severe acute respiratory illness hospitalization in Madagascar, 2011-2016. Data inputs steps are in light gray boxes, and data outputs are in dark gray boxes age. Rates were expressed per 100 000 population. All estimates were reported as mean annual estimates over the study period.
Annual estimates were also provided overall and among individuals aged <5 and ≥5 years.

| Step 1: Estimation of SARI hospitalizations rates in Analamanga region
To estimate the SARI hospitalizations rates in Analamanga Region, we followed the WHO Manual for Estimating the Disease Burden Associated with Seasonal Influenza. 14 First, we obtained the SARI hospitalization rates for the Antananarivo Renivohitra District by dividing the total number of SARI hospitalizations that occurred in the district (DS-1) by the district population (DS-4). Thereafter, we used the Antananarivo Renivohitra District SARI hospitalization rates as a proxy for Analamanga Region as previously described. 12,13,20,21

| Step 2: Estimation of SARI hospitalizations rates in the other regions
Estimates of SARI hospitalization rates for the other 21 regions in Madagascar were derived by adjusting the Analamanga Region rate (base region-obtained in Step 1) for the regional-level prevalence of known risk factors for pneumonia obtained from the DHS (DS-3) as previously described (Step 2.a). 13,20,21 Risk factors included exposure to indoor air pollution, crowding, malnutrition, low birthweight, and non-exclusive breastfeeding. 13,20,21 The last three were only included for children aged <5 years. HIV infection prevalence, which was used as an adjustment factor in other similar studies, was excluded in our analysis because of the very low prevalence of HIV infection in Madagascar (<0.2% in the general population) 22 and the lack of region-specific prevalence estimates. The relative risk of SARI associated with each risk factor was determined from the published literature. 13,20,21,23 In addition, we adjusted the regional rates by the proportion of ARI cases seeking care in the given region to the proportion of ARI cases seeking care in the base region using data from the DHS (DS-3) as previously described (Step 2.b). 13,20,21 The differential health care seeking behavior between regions among patients with ARI was used as a proxy for the differential health care seeking behavior among patients with SARI.
An adjustment factor >1 resulted in a greater SARI hospitalization rate in the given region relative to the base region and vice versa.
The equations used for the regional adjustments are provided in the Supporting Information.

| Step 3: Estimation of influenza-and RSVassociated SARI hospitalizations rates in all regions
We estimated the regional rates of influenza-and RSV-associated SARI hospitalization by multiplying the estimated regional SARI hospitalization rates (obtained in Steps 1 and 2) by the influenza and RSV positivity proportion obtained from influenza and RSV sentinel surveillance implemented among inpatients with SARI (DS-2). 13,20,21 The influenza and RSV positivity proportion was the number of positives cases divided by the total number of SARI cases tested.

| Step 4: Estimation of the number of influenza-and RSV-associated SARI hospitalizations in all regions
We estimated the regional number of influenza-and RSV-associated SARI hospitalizations by multiplying the regional influenza-and RSVassociated SARI hospitalization rates (obtained in Step 3) by the midyear population at risk in each region over the study period. 13,20,21 We obtained the 95% confidence intervals (CI) using bootstrap resampling over 1000 replications for all parameters included in the calculations. 13,20,21 This included (a) the age-and year-specific SARI hospitalization rates in the base region; (b) the regional prevalence of the risk factors for pneumonia; (c) the regional proportion of ARI cases seeking care; and (d) the age-specific influenza and RSV positivity proportion among SARI cases tested. The lower and upper limits of the 95% CI were the 2.5th and 97.5th percentiles of the estimated values obtained from the 1000 resampled datasets, respectively. Finally, we aggregated the regional estimates by the 6 provinces of Madagascar and nationally.

| Ethics
The influenza and RSV virologic data (DS-2) and the collection of

| National number and rates of influenza-and RSV-associated SARI hospitalization
In individuals of all ages, the estimated mean annual number of influenza-associated SARI hospitalization was 6609 (95% CI: A U-shaped trend of the magnitude of the mean annual influenza-and RSV-associated SARI hospitalizations rates was observed across age groups (Table 1). Compared to RSV, mean annual influenza-associated SARI hospitalization rates were lower (with non-overlapping CIs) among children aged <5 years and higher among individuals aged ≥5 years. No substantial differences (with overlapping CIs) of the influenza-and RSV-associated SARI hospitalizations rates were observed between provinces (Table 1); however, variations were observed between years ( Table 2). The provincial mean annual number and rates of influenza-and RSV-associated SARI hospitalizations by age group are provided in Table S1. The influenza-and RSV-associated SARI hospitalization rates by region are provided in Figure S1.

| Comparison of the demographic and clinical characteristics of influenza-and RSV-positive patients hospitalized with SARI
During 2011-2016, there were 239 patients hospitalized with SARI that tested positive for influenza virus only, 344 that tested positive for RSV only, and 69 that tested positive for both viruses; the latter were excluded for this analysis. On multivariable analysis, compared to individuals that tested positive for influenza virus, those TA B L E 1 Estimated mean annual number and rate of influenza-and respiratory syncytial virus-associated severe acute respiratory illness hospitalization, Madagascar, 2011-2016

| D ISCUSS I ON
We reported national and provincial estimates of influenza-and  12 and Zambia (187 per 100 000 population) 13 and global estimates for Africa (174 per 100 000 population), 5 and such estimates were higher compared to those of other Regions. 5 Among individuals aged ≥5 years, the Madagascar estimates (11.6 per 100 000 population) were generally consistent with the TA B L E 2 Estimated annual number and rate of influenza-and respiratory syncytial virus-associated severe acute respiratory illness hospitalization, Madagascar, 2011-2016  in these settings may explain some of the observed variation in the influenza-associated SARI hospitalization rates in this age group.

Influenza-associated SARI hospitalization RSV-associated SARI hospitalization
Cultural differences and differential access to health care across countries can also play a role in health care seeking behavior potentially also contributing to variability in hospitalization rates. of inpatients with respiratory illnesses that did not meet the SARI case definition tested positive for influenza. 36 In addition, ecological studies have suggested that influenza and RSV are responsible for hospitalizations and deaths also among patients presenting with circulatory illnesses or other non-respiratory and non-circulatory syndromes. 24,26,27 Lastly, individuals that may have developed influenza-or RSV-associated severe illness, but did not seek care, would have been missed in our study; hence, our estimates should be considered minimum estimates.
In conclusion, we reported a substantial hospitalization burden associated with influenza virus and RSV infection especially in children aged <5 years. The Madagascar Ministry of Health has not yet implemented a national influenza vaccination program. The information presented here could be used by policymakers to consider vaccine introduction. Should an influenza vaccination program be introduced in Madagascar young children, and older adults may benefit most from annual influenza immunization. No influenza vaccine is licensed for children aged <6 months, but this group may be protected through the vaccination of their mothers during pregnancy. 37,38 The burden of RSV-associated SARI hospitalization was higher than those of influenza. Should a RSV vaccine become available it would have the potential to prevent a substantial number of severe illnesses, especially in children.

ACK N OWLED G EM ENTS
We thank all members involved in SARI surveillance and the collection of SARI hospitalization data at the selected hospitals.

CO N FLI C T O F I NTE R E S T S
All authors declare that they have no commercial or other associations that may pose a conflict of interest.

AUTH O R CO NTR I B UTI O N S
All authors take responsibility for the integrity of the data and the