The attributable fraction of respiratory syncytial virus among patients of different age with influenza-like illness and severe acute respiratory illness in a high HIV prevalence setting, South Africa, 2012-2016

198 Main text 2640 Key point The attributable fraction of illness is key to accurate burden estimates, specifically in the presents of sensitive PCR testing. Burden and cost burden estimates are key to cost- effectiveness model for new prevention technologies in the pipeline for respiratory syncytial virus. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 4

Introduction The detection of respiratory syncytial virus (RSV) in upper airway samples does not necessarily infer causality of illness. Calculating the attributable fraction (AF) of RSV in clinical syndromes could refine disease burden estimates.

Using unconditional logistic regression models, we estimated the AF of RSV-associated influenza-like illness (ILI) and severe- acut e respirator y illnes s ( SARI) ca ses by com paring RSV- detection prevalence among ILI and SARI cases to those of healthy controls in South Africa, 2012-2016. The analysis, stratified by HIV serostatu s, was conducted in the age categories <1, 1-4, 5-24, 25-44, 45-64, ≥65 years.

We included 12,048 individuals: 2,687 controls, 5,449 ILI cases and 5,449 SARI cases. RSV- AFs for ILI were significant in <1, 1-4, 5-24, 25-44-year age groups: 84.9%(95% confidence interval (CI) 69.3%-92.6%), 74.6%(95%CI 53.6%-86.0%), 60.8%(95%CI 21.4%-80.5%) and 64.1%(95%CI 14.9%-84.9%), respectively. Similarly, significant RSV-AFs for SARI were 95.3%(95%CI 91.1%-97.5) and 83.4%(95%CI 70.9-90.5) in the <1 and 1-4-year age groups respectively. In HIV-infected persons, RSV was significantly associated with ILI cases versus controls in individuals aged 5-44 years.

. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 5 High RSV-AFs in young children confirm RSV detection is associated severe respiratory illness in South African children, specifically infants. These estimates will assist with refining burden estimates and cost effectiveness models. Key words: Attributable Fraction; Respiratory syncytial virus; Burden of disease . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 6

It is estimated that up to 33.1 million episodes of RSV-associated lower respiratory tract infection (LRTI) occur annually globally in children under-5 years of age, approximately 10% of which result in hospitalisation 1 . Of the nearly 120,000 estimated RSV-associated global deaths among children aged <5 years annually, 99% occur in low-and-middle income countries of Africa and Asia 2 . Data from South Africa confirm a substantial RSV burden in children, with estimated hospitalisation rates of 3,262/100,000 population in children aged <1 year 3 . HIV-exposed infants under 6 months of age are considered at increased risk of hospitalisation with RSV-associated LRTI compared to HIV-unexposed infants (incidence rate ratio (IRR) 1.4; 95% CI 1.3-1.6) 4 . In addition, there may be substantial burden of RSV-associated LRTI in adults, with estimated incidence of hospitalisation 30/100,000 in the general South African population aged ≥ 5 year s and from 106 to 390/100,000 population in individuals living with HIV 5 . HIV infection is a known risk factor for severe respiratory il lness (SRI) in all age groups 5 . Most recent burden estimates have relied on sensitive polymerase chain reaction (PCR) assays to detect RSV in respiratory samples 6 . Describing the correlation between the detection of RSV on these sensitive assays and the presence of RSV disease is important to refine burden estimates. Accurate burden estimates will be important to advocate for the

of new RSV prevention technologies, such as vaccines and new generation monoclonal antibodies 7,8 . In addition, estimates of the attributable fraction (AF) of RSV (RSV-AF) in LRTI and influenza like-illness (ILI) will assist in building accurate cost- effectiveness models for the RSV interventions in the pipeline. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 7 While there are some data on RSV-AF in children, there is very little information on the RSV- AF in HIV-infected and HIV-exposed uninfected (HEU) children 9,10 . The AF of RSV to disease in adults is also not well characterised. Prior to the introduction of interventions for RSV- associated illness, policy makers need accurate estimates of both in- and out-patient RSV disease burden. Burden estimates collected prior to introduction will assist with the assessment of vaccine impact. We aimed to evaluate the AF of RSV-detection in individuals of all ages presenting with non- hospitalized influenza-like illness (ILI) or hospitalised severe acute respiratory illness (SARI). Further we looked specifically at the following groups: HIV-infected and HIV-uninfected children <5 years of age, HEU infants (aged <6 months); and lastly in HIV-infected and HIV- uninfected adults. These analyses were further stratified by age-groups <1 year, 1-4 years, 5-24 years, 25-44 years, 45-64 years and ≥65 years. We also aimed to estimate the proportion of ILI and SARI cases attr ibutable to RSV infection after adjusting the observed detection rate of RSV for the RSV-AF.

Cases were enrolled as part of a prospective, hospital-based, sentinel surveillance programme for respiratory illness from June 2012 through May 2016 at two sites in North West (Klerksdorp-Tshepong Hospital Complex (KTHC), Klerksdorp), and KwaZulu-Natal (Edendale Hospital (EDH), Pietermaritzburg) provinces of South Africa and primary health care clinics in the hospital catchment population 11 . Specific descriptions of enrollment are provided below. Controls and influenza-like illness (ILI) cases . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 8 Study-specific controls were defined as individuals who sought medical care at one of the sentinel clinics for non-respiratory, non-gastrointestinal symptoms and who did not have a recorded temperature or history of fever within the last 14 days. Most of the controls were selected from non-urgent outpatient service visits including dental clinics, family planning, immunization and well-baby visits, voluntary HIV counselling and testing, or acute care for minor injuries. Two controls, one living with and one without HIV, were enrolled weekly in each clinic among age groups: 65 year s. ILI cases were enrolled at the same clinics. ILI was defined as illness in an outpatient of any age who had a r ecorded temperature > 38 °C or a history of fever and cough of < 10 days duration. Severe acute respiratory illness (SARI) surveillance A SARI case was defined as an illness in a hospitalised individual meeting the fol lowing age- specific criteria. A SARI case in children 3 months to <5 years of age was defined as any child with physician-diagnosed acute LRTI with symptom duration ≤10 days, including bronchopneumonia, pneumonia, bronchitis, bronchiolitis, and pleural effusion irrespective of signs and symptoms; we also included infants aged 2 days to <3 months with physician- diagnosed sepsis. A SARI case in individuals aged ≥5 years was defined as any individual with recorded temperature ≥38 °C or history of fever and cough. Study procedures Detailed procedures for the surveillance program have been described 11 . Briefly, dedicated surveillance staff systematically screened potentially eligible outpatients and hospitalized individuals and collected a clinical specimen and detailed history of symptoms and other . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 9 medical conditions from consenting persons meeting surveillance case definitions. Individuals with ILI who were referred to hospital on the day of enrolment were excluded from the analysis. Determination of HIV status HIV status was determined from one of following sources: (i) in hospital/clinic voluntary counselling and testing by rapid HIV test according to standard clinical protocols (ii) patient records or (iii) if necessary (with consent), a dried blood spot tested at the National Institute for Communicable Diseases (NICD), J ohannesburg. Testing for individuals aged > 18 months was done by an HIV ELISA, and PCR testing was conducted on samples from children aged <18 months. HIV-exposed-uninfected infants HIV exposed-uninfected (HEU) infants were those aged <6 months born to a mother who was living with HIV, and in whom the HIV infection status of the child was confirmed to be negative. These infants were enrolled at the same surveillance sites as the cases and controls above. Samples collection and virus detection Nasopharyngeal aspirates were collected from children aged <5 years and for individuals aged ≥5 years both oropharyngeal and nasopharyngeal swabs were collected. The specimens were placed in universal transport medium (Copan, Murruieta, California, USA) and stored at 4 °–8 °C at the surveillance site prior to being transported to NICD for testing. Testing was conducted within 72 hours of collection. All specimens were tested for RSV and nine other respiratory viruses (influenza virus A and B, parainfluenza types 1, 2 and 3, . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 10 adenovirus, human metapneumovirus, enteroviruses and rhinovirus). Testing was done on a multiplex real-time reverse transcription polymerase chain reaction assay during 2012- 2014 12 . From 2015 a combination of the commercial Fast Track Diagnostics (FTD) Flu/HRSV assay (Fast Track Diagnostics, Luxembourg) and Allplex Respiratory Assay, panel 2 and 3 (Seegene, Seoul, Korea) in 2015 were used to test for other respiratory viruses. Statistical analysis We used an unconditional logistic regression to estimate the AF of RSV-associated SARI and ILI by comparing the RSV detection rate observed in our surveillance among ILI and SARI to those of controls 13,14 . These estimates were adjusted for HIV infection and co-infections with the other respiratory viruses tested. The analysis was implemented overall (all age groups), and stratified by HIV serostatus in the following age categories: <1 year, 1-4 years, 5-24 years, 25-44 years, 45-64 year s, and≥65 years. HIV infection and age group were also included as covariates in the non HIV-stratified overall model and the models implemented among children aged <5 (<1 and 1-4 year s) and per son aged ≥5 years (5-24, 25-44, 45-64 ≥65 years). Significant associations were considered at p<0.05. The adjusted odds ratios obtained from these models were used to estimates the attributable fraction (AF) by applying the formula: 1 *100ORAF OR −= . We also estimated the RSV detection rate associated with illness among individuals with ILI and SRI ( Prev Illness ) from the observed detection rate ( Prev Observed ) as follows: Prev Illness = Pr ev Observed *AF. The statistical analysis was conducted using STATA version 14.1 (StataCorp, College Station, Texas, USA). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 11 Ethical considerations Approval for the hospital-based surveillance was obtained from the University of the Witwatersrand Human Research Ethics Committee (HREC) for the Klerksdorp site and form the University of KwaZulu-Natal Human Biomedical Research Ethics Committee (BREC) for the Pietermaritzburg site; protocol numbers M081042 and BF157/08, respectively. The ILI and control enrolment protocol was approved by HREC and BREC (protocol numbers M120133 and BF080/12, respectively). This surveillance was deemed non-research by the US Centers for Disease Control and Prevention (non-research determination number: 2012- 6197). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 12

Study population We enrolled 12,048 individuals of whom 99% (12,008) had RSV test results. This included 2,687 controls, 5,449 with ILI and 3,912 with SARI. Overall, the RSV detection positivity was 2% (52/2,687) in controls, 5% (303/5,449) in ILI cases and 14% (553/3,912) in SARI cases. The highest positivity was in children aged <1 year with SARI (31%; 413/1,328). Among ILI cases, detection was highest in children aged <1 year (13%; 78/621). RSV detection was similar among controls (3%; 12/389 and 3%; 17/552) in age groups <1 year and 1-4 years, respectively (Table 1). HIV results were available for 95% (11,445/12,048) of enrolled participants. Attributable fractions Influenza-like illness RSV-AFs in ILI cases were significantly different from zero in the following age groups: <1 year (84.9%, 95% CI 69.3%-92.6%), 1-4 years (74.6%, 95% CI 53.6%-86.0%), 5-24 year s (60.8%, 95% CI 21.4%-80.5%) and 25-44 year s (64.1%, 95% CI 14.9%-84.9%). In older age groups, the detection of RSV in ILI cases was not significantly attributable to disease (45-64 years, 69.6%, 95% CI 0%-90.8% and ≥65 years, 48.0%, 95% CI 0%-94.7%) (Table 1, Figure 1a- c). Severe acute respiratory illness The RSV-AFs for SARI were significantly different from zero in the age groups <1 year (95.2%, 95% CI 90.9%-97.5) and 1-4 years (82.9%, 95% CI 70.2-90.1) (Table 1, Figure 2a). This . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 13 association was also seen in HIV-uninfected individuals with similar RSV-AFs observed in HIV-uninfected infants aged <1 year with SARI (96.7%, 95% CI, 93.0%-98.5%) (Table 1, Figure 2a-c). The RSV-AF was not significantly different from zero in the age group ≥65year s irrespective of HIV status (80.6%, 95% CI, 0%-96.6%) and similarly in the age groups 5-24 years, 25-44 years and 45-64 years (25.6%, 95% CI 0%-73.8%; 26.1%, 95% CI 0%-72.6% and 52.6%, 95% CI, 0%-87.9%, respectively) (Table 1, Figure 2a-c) HIV-infected individuals Among HIV-infected individuals, the RSV-AFs for ILI were significantly different from zero in the 5-24 and 25-44-year age groups only 78.1%, 95% CI, 36.5%-92.5% and 76.6%, 95% CI 13.6%-93.7%, respectively). We were not able to calculate the RSV-AFs for ILI and SARI in the <1-year age groups due to no RSV observations occurring in HIV-infected infants with ILI and SARI (Table1, Figure 1b and 2b). HIV-exposed infants The RSV-AF, for HIV-exposed-uninfected infants (HEU) <6 months of age with SARI, was significantly greater than zero (89.9%, 95% CI 69.2%-96.7%). The RSV-AF for ILI in HEU <6 months of age was not significantly different from zero (64.2%, 95% CI 0%-90.1%). The RSV- AF for HIV-unexposed-uninfected (HUU) infants aged <6 months was significantly different from zero across the clinical syndromes (ILI 88.0%, 95% CI 54.2%-96.8% and SARI 97.0%, 95% CI 86.0%-99.3%) (Table 2).

. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 14 W e we r e a b l e t o de s c r i b e R S V - AF s t h a t w e r e s i g n i f i ca n t l y d i f f e r e n t f r o m ze r o i n I L I i n a l l a g e groups except for the older adul t age groups (45-64 years and ≥65 years). We confirmed the high RSV-AF in children with SARI with significant RSV-AF in children aged <1 year. In addition, we documented no RSV-AFs in adults with RSV-associated SARI which were significantly different from zero. These associations were similar for adults living with HIV. In HIV-uninfected individuals aged ≥5 years, the detection of RSV was not significantly associated with illness. The RSV-AFs were lower in HEU infants aged <6 months, as compared to the HUU, but higher than HIV-infected infants aged <6 months. In children the RSV-AF was high in all age groups, specifically in the children <1 year presenting with SARI confirming the findings in other studies 3,15 . Detection of RSV was significantly associated with illness in the children presenting with ILI aged <1 year and 1-4 years. Several studies in children confirm a strong association between the detection of RSV and disease that we describe in this analysis. Although there are many different study design s and case definitions of SARI the results are remarkable similar across studies 9,10,15,16,17,18 . An interim analysis of a subset of data presented in this study showed similar results to our analysis for those aged <5 years 14 . In the Drakenstein study, a case- control study, conducted in South Africa, of the association between the detection of respiratory viruses and pneumonia, RSV detection was associated with pneumonia (odds ratio (OR)) 8.5, 95% CI; 4.2-15.4 and very strongly associated with very severe illness OR 25.3, 95% CI 3.3-191.6 10 . A study in western Kenya described an association between RSV detection and disease in children aged <5 years with pneumonia (adjusted OR 3.0, 95% CI 1.1-8.2) 18 . The PERCH multicentre study reported that viruses account for almost two thirds of pneumonia cases (61.4%, 95% cr edible interval (CrI) 57.3-65.6. In addition, this study . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 15 reported RSV as having the greatest aetiological fraction 31.3 95% CrI, 28.4-34.2) of all pathogens, very similar to our results 17 . Additional evidence is presented in a systematic review and meta-analysis by Shi et al where the authors assign a causal attribution OR of RSV in children of 9.8, 95% CI 4. 9-19.3 and an AF of 90%, 95% CI 85%-95% 15 . In our assessment, among individuals living with HIV we describe that RSV detection is significantly associated with illness in individuals aged 5-24 years and 25-44 years for ILI. In HIV-infected children aged <5 years we did not see an association between the detection of RSV and illness, this is likely due to the very low prevalence of HIV in this age group, limiting our ability to detect significant associations. We did not detect an y cases of RSV-associated ILI or SARI in HIV-infected individuals <1 year. HEU infants are at increased risk of LRTI and hospitalisation with RSV-associated LRTI, the lower RSV-AF in this group compared to HIV- unexposed uninfected infants suggest that other pathogens are contributing proportionately more to the disease in this group of infants 4 . In individuals aged ≥5 years with SARI the point estimates of RSV-AF suggest that RSV may be asso ciated with illness however in many age group s the RSV-AFs were not significantly higher in cases compared to controls. Although older age (≥65 years) is considered a risk factor for RSV-associated SARI, we do not demonstrate a significant RSV-AF in this group of individuals 19 . This may be due to the low number of elderly individuals presenting at health care facilities in our setting or because there is not truly an association in this age group. When our RSV-AF data are displayed graphically by age group, U-shaped distributions are seen among SRI cases, with high RSV-AF in young children, dropping in younger adults and . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 16 rising again in the elderly (although point estimates of AF were not significant in some groups). Similar to the shape of the distribution for the incidence of hospitalisation with RSV-associated SRI 5,20,21 . By comparison the influenza-AF curve among individuals with SARI has lower estimates for children <1 year and higher estimates for the ≥65 year age group 22 . Conversely the RSV-AF in individuals with ILI forms a more bell-like shape with highest AF in the age groups 25-64 years. Data on the causal association of viral detection in LRTI in adults are sparse. In the EPIC study in the United States of America the incidence of RSV-associated pneumonia in adults was estimated at 7/100,000 population 95% CI, 5%-9%, lower than the estimate in our setting but without AF estimation, therefore the numbers are not comparable to our study 20,21 .

RSV is strongly associated with respiratory illness particularly in young South African children. While point estimates suggested an elevated AF for RSV-associated SARI among the elderly, we were not able to conclusively demonstrate an association between RSV detection and illness. More data are needed on the RSV-AF to illness among the elderly. These data are of interest to clinicians and policy makers. In addition, the calculation of the AF will improve the estimates of burden and cost-effectiveness models. These estimates of AF could be used to motivate for the introduction of vaccines and new generation monoclonal antibodies 8 . Disclaimer The findings and conclusions in this r eport are those of the author(s) and do not necessarily represent the official position of the funding agencies. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 17 Conflict of interest statement Prof Cheryl Cohen and Anne v on Gottberg has received institutional funding support from US CDC related to the current work and also received funding from Bill and Melinda Gates Foundation (BMGF), PATH, Sanofi, Wellcome Trust and South African MRC outside of the submitted work. Dr Moyes received institutional funding from Sanofi and PATH for work outside of this current work. Dr Dawood reports personal fees from Sanofi-South Africa. In addition, she reports workshop attendance sponsorship from Biomieruex-South Africa. Funding source The data collection for this manuscript was funded through a grant to the NICD from the Centers for Disease Control (CDC), Atlanta Georgia. Collaborative Research on Influenza, Coronavirus Diseases 2019 (COV ID-19), and Other Respiratory Pathogens in South Africa (Cooperative Agreement Number: 1U01IP001160) Contributions Jocelyn Moyes: Study design, implementing field work, data analysis and manuscript preparation. Stefano Tempia: Study design, statistical advise, manuscript preparation. Sibongile Walaza: Managed field work, data management and contribution to preparing the manuscript. Meredith L. McMorrow: Supported manuscript writing and data anlysis. Adam L. Cohen: Data management, analysis and writing. Florette Treurnicht : Sample testing, verifying laboratory results, contributed to manuscript writing. Orienka Hellferscee : Samp le testing, verifying laboratory results, contributed to manuscript writing. Nicole Wolter: Sample testing, verifying laboratory results, contributed to manuscript writing. Anne Von Gottberg: Verifying laboratory results, contributed to manuscript writing. Halima Dawood: . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 18 Assisted with field work set up, access to hospitals and writing the manuscript. Ebrahim Variava: Assisted with field work set up, access to hospitals and writing the manuscript. Kathleen Kahn: Assisted with field work set up , access to hospitals and writing the manuscript. Shabir A. Madhi: Study design, analysis and manuscript review. Cheryl Cohen: Study design, statistical support, data analysis and manuscript writing. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 19

1. Shi T, M cAllister DA, O’Brien KL, et al. Global, regional, and national disease burden estimates of acute lower respiratory i nfections due to respiratory syncytial virus in young children in 2015: a systematic review and modelling study. Lancet . 2017;390(10098):946-958. doi:10.1016/S0140-6736(17)30938-8 2. Nair H, Nokes DJ, Gessner BD, et al. Global burden of acute lower respiratory infections due to respiratory syncytial virus in young children: a systematic review and meta-analysis. Lancet . 2010;375(9725):1545-1555. doi:10.1016/S0140- 6736(10)60206-1 3. McMorrow ML, Tempia S, W alaza S, et al. The role of HIV in influenza- and respiratory syncytial virus-associated hospitalizations in South African children, 2011-2016. Clin Infect Dis . Published online June 29, 2018. doi:10.1093/cid/ciy532 4. Cohen C, Moyes J, Tempia S, et al. Epidemiology of Acute Lower Respiratory Tract Infection in HIV-Exposed Uninfected Infants. Pe di atrics . 2016;137(4):e20153272- e20153272. doi:10.1542/peds.2015-3272 5. Moyes J, W alaza S, Pretorius M, et al. Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV prevalence setting. J Infect . 2017;75(4):346-355. doi:10.1016/j.jinf.2017.06.007 6. Shi T, M cAllister DA, O’Brien KL, et al. Global, regional, and national disease burden estimates of acute lower respiratory i nfections due to respiratory syncytial virus in young children in 2015: a systematic review and modelling study. Lancet (London, England). 2017;390(10098):946-958. doi:10.1016/S0140-6736(17)30938-8 7. Vekemans J, Moorthy V, Giersing B, et al. Respiratory syncytial virus vaccin e research and development: World Health Organization technological roadmap and preferred product characteristics. Vaccine . Published online January 25, 2018. doi:10.1016/j.vaccine.2017.09.092 8. Mazur NI, Higgins D, Nunes MC, et al. The respiratory syncytial virus vaccine landscape: lessons from the graveyard and promising candidates. Lancet Infect Dis . 2018;18(10):e295-e311. doi:10.1016/S1473-3099(18)30292-5 9. Pretorius MA, Tempia S, W alaza S, et al. The role of influenza, RSV and other common respiratory viruses in severe acute respiratory infections and influenza-like illness in a population with a high HIV sero-prevalence, South Africa 2012-2015. J Clin Virol . 2016;75:21-26. doi:10.1016/j.jcv.2015.12.004 10. Zar HJ, Barnett W, Stadler A, Gardner-Lubbe S, Myer L, Nicol MP. Aetiology of childhood pneumonia in a well vaccinated South African birth cohort: a nested case- control study of the Drakenstein Child Health Study. Lancet Respir Med . 2016;4(6):463-472. doi:10.1016/S2 213- 2600(16)00096-5 11. Cohen C, Moyes J, Tempia S, et al. Severe influenza-associated respiratory infection in high HIV prevalence setting, South Africa, 2009-2011. Emerg Infect Dis. 2013;19(11):1766-1774. doi:10.3201/eid1911.130546 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 20 12. Pretorius MA, Madhi SA, Cohen C, et al. Respiratory viral coinfections identified by a 10-plex real-time reverse-transcription polymerase chain reaction assay in patients hospitalized with severe acute respiratory illness--South Africa, 2009-2010. J Infect Dis. 2012;206 Suppl 1(suppl_1):S159-65. doi:10.1093/infdis/jis538 13. Tempia S, Walaza S, Moyes J, et al. Attributable Fraction of Influenza Virus Detection to Mild and Severe Respiratory Illnesses in HIV-Infected and HIV-Uninfected Patients, South Africa, 2012-2016. Emerg Infect Dis . 2017;23(7):1124-1132. doi:10.3201/eid2307.161959 14. Pretorius MA, Tempia S, W alaza S, et al. The role of influenza, RSV and other common respiratory viruses in severe acute respiratory infections and influenza-like illness in a population with a high HIV sero-prevalence, South Africa 2012–2015. J Clin Virol. 2016;75:21-26. doi:10.1016/j.jcv.2015.12.004 15. Shi T, M cLean K, Campbell H, Nair H. Aetiological role of common respiratory viruses in acute lower respiratory infections in children under five years: A systematic review and meta–analysis. J Glob Health . 2015;5(1):010408. doi:10.7189/jogh.05.010408 16. Gilani Z, Kwong YD, Levine OS, et al. A Literature Review and Survey of Childhood Pneumonia Etiology Studies: 2000–2010. Clin Infect Dis . 2012;54(suppl_2):S102-S108. doi:10.1093/cid/cir1053 17. O’Brien KL, Baggett HC, Brooks WA, et al. Causes of severe pneumonia requiring hospital admission in children without HIV infection from Africa and Asia: the PERCH multi-country case-control study. Lancet . 2019;394(10200):757-779. doi:10.1016/S0140-6736(19)30721-4 18. Hammitt LL, Kazungu S, Morpeth SC, et al. A preliminary study of pneumonia etiology among hospitalized children in Kenya. Clin Infect Dis . 2012;54(SUPPL. 2). doi:10.1093/cid/cir1071 19. W alsh EE, Peterson DR, Falsey AR. Risk factors for severe respiratory syncyt ial virus infection in elderly persons. J Infect Dis . 2004;189(2):233-238. doi:10.1086/380907 20. Moyes J, W alaza S, Pretorius M, et al. Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV prevalence setting. J Infect . 2017;75(4):346-355. doi:10.1016/j.jinf.2017.06.007 21. Pastula ST, Hackett J, Coalson J, et al. Hospitalizations for Respiratory Syncytial Virus Among Adults in the United States, 1997-2012. Open forum Infect Dis . 2017;4(1):ofw270. doi:10.1093/ofid/ofw270 22. Tempia S, Walaza S, Moyes J, et al. The effects of the attributable fraction and the duration of symptoms on burden estimates of influenza-associated respiratory illnesses in a high HIV prevalence setting, South Africa, 2013-2015. Influenza Other Respi Viruses . 2018;12(3):360-373. doi:10.1111/irv.12529 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 21 Table 1: Att ribu t a b l e f raction ( A F ) and A F-adjus t ed prevalence of r espiratory syn c ytial v ir us ( RSV) among individ uals in a stud y of th e A F of RSV detection in mi ld and severe i llnes s, K lerksdorp and Eden dal e, Sout h Afr i ca, Jun e 20 12 t o May 2016 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 22 All enrolled participants Influenza-like Illness (ILI) Severe Acute Respiratory Illness (SARI) Obs erv e d R SV de t e c ti o n rat e Attributa b l e Fr acti on (A F ) % (9 5% Co n fi de n ce i n t e rv a l (C I)) AF adjust ed pre v a l enc e % R SV de te c t io n ra t e A F , % ( 95 % C I) A F - a dj u s te d p rev a len ce % <1 78/621 (13) 8 4.9 ( 69. 3- 92. 6) 11 413/1328 (31) 95. 3 ( 9 1. 1 -9 7. 5) 30 1-4 116/1101 (11) 74. 6 ( 5 3. 6 -8 6. 0) 8 101/786 (13) 83. 4 ( 7 0. 9 -9 0. 5) 11 5-24 43/1531 (3) 60. 8 ( 2 1. 4 -8 0. 5) 2 6/287 (2) 25.6 (0-73.8) 1 25-44 49/1532 (3) 64. 1 ( 1 4. 9 -8 4. 9) 2 16/842 (1) 26.1 (0-72.6) <1 45-64 15/561 (3) 69.6 (0-90.8) 2 11/492 (2) 52.6 (0-87.9) 1 65 + 2/103 (2) 48.0 (0-94.7) 1 6/177 (3) 80.6 (0-96.8) 2 Indi v idu a ls li v i ng w ith HI V Obs erv e d R SV de t e c ti o n rat e ILI Obs e rv e d R SV d et ecti on rat e SA R I AF , % (95 %C I) AF-a d j usted pre v a l enc e , % AF, % (9 5 %C I) AF -a dj u s ted pr e val en c e, % <1 0/8 (0) 0 (0-92.5) Not determined 3/20 (15) 0 (0-71.9) Not determined 1-4 1/19 (5) 75.8 (0-96.5) 4 9/91 (10) 72.1 (0-95.9) 7 5-24 12/230 (5) 78. 1 ( 3 6. 5 -9 2. 5) 4 5/114 (3) 55.4 (0-87.7) 3 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 23 25-44 29/865 (3) 76. 6 ( 1 3. 6 -9 3. 7) 2 16/722 (2) 61.8 (0-90.0) 1 45-64 6/227 (3) 75.7 (0-96.1) 2 8/299 (3) 80.4 (0-97.6) 2 65+ 0/10 (0) No RSV N A 3/33 (9) 87.4 (0-99.4) 8 HI V -uni n f ected i n di v idu a ls I L I S A R I Obs erv e d R SV de t e c ti o n rat e AF , % (9 5 %C I) AF-a d j ust e d pre val en c e, % Obs e rv e d R SV d et ecti on rat e AF , % (9 5 %C I) AF-adj usted pr e v a l enc e , % <1 68/574 (12) 87. 5 ( 7 2. 2 -9 4. 4) 1 1 371/1109 (33) 96. 7 ( 9 3. 0 -9 8. 5) 32 1-4 112/1038 (11) 77. 7 ( 5 6. 6 -8 8. 5) 9 91/656 (14) 84. 6 ( 7 0. 3 -9 2. 0) 12 5-24 27/1158 (2) 39.0 (0-74.1) 1 1/123 (1) 0 (0-78.7) No t de te r m i ned 25-44 18/588 (3) 30.9 (0-76.0) 1 0/93 (0) 0(0-53.5) No t de te r m i ned 45-64 7/301 (2) 60.8 (0-91.0) 1 0/162 (0) 0 (0-91.3) No t de te r m i ned 65+ 1/84 (1) 46.0 (0-94.5) 0.5 3/129 (2) 64.3 (0-94.8) 1 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 24 Table 2: A ttributable fr ac t ion ( AF) an d AF-adjusted prevalenc e of respiratory s yn cyt i al viru s ( RSV) among HIV exposed-uninf ec t e d (H EU ), HIV- infected and H IV-unexposed uninf ected ( HUU ) infants aged <6 mon ths, K l e r ks dor p and Edendale, June 2012 t o M ay 201 6 Influenza-like Illness (ILI) Severe Acute Respiratory Illness (SRI) Ob served detection rate of RSV Attributable Fraction (AF) % (95% Confidence interval (CI)) AF adjusted prevalenc e % Ob ser ve d detection rate of RSV AF, % (95% CI) AF adjusted prevalence % H IV Ex po s ed Uninfected 10/ 77 ( 13) 64.2 (0-90.1) 8 5/ 16 (3) 89.6 (50.1-97.8) 2 H IV infected 0/ 3 ( 0) 71.0 (0-99.2) Not determ i ne d 2/ 10 (20) 87.3 (0- 98.8) 17 H IV Unex po s ed Uninfected 20/ 82 ( 11) 88.0 (54.2-96.8) 10 9/ 27 (33) 97.0 (86.0-99.3) 32 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint Figure 1: A ttributab l e f raction of respirato ry s yn c ytial virus (RSV) and p roportion of RSV- pos it iv e cases at tribu t able to il lne ss and R ca ses not a t trib utab le t o illness in ind i viduals with influ enza-like illness (IL I) (1 a : Overall, 1b: HIV -infected and 1c: HIV uninfect e pr ovinc es South A frica 2012-2016. 25 R SV-positive e d), in two . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint Figure 2: A ttr ibu table f r ac t i o n of r espirato ry s yn c ytial vi r us ( RSV) and pr oport i o n o f RSV-positive cas es attr ibutable to i llne ss and RS V ca ses not at tribu t able t o illn es s in individuals with severe acute respirato ry ill n es s (SA RI) (2 a : Overall, 2b: H IV-inf ected an d 2c: H IV-u n 26 V -positive n infected), . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint 27 in two pr ov inces South Af r ica 20 12-2 016. ( F o o t n o te R S F - AF in age g ro up 5 -2 4 , 2 5 - 44 a n d 4 5 -5 4 we re n o t e s ti m a te d fo r H IV - un in fec te d in d iv i d u al s a s n o ca se s we r e d e tec te d i n t hes e a g e gr o u ps) . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted November 22, 2022. ; https://doi.org/10.1101/2022.11.22.22282618doi: medRxiv preprint