Studies of any epidemiological design in Spanish, English, or Portuguese and surveillance reports published since January 1st, 2012, and January 4th, 2023, which included LAC patients of any age or risk group, were eligible for inclusion. We considered that previously published studies could make it difficult to estimate the current burden of RSV. We planned to include observational studies such as cohorts, case–control studies, and representative case series (involving at least 50 laboratory-confirmed RSV cases or at least 10 RSV patients with complications), the control arms of the randomized clinical trials (RCTs) and quasi-RCTs, controlled before-and-after studies (CBAs) and uncontrolled before-and-after studies (UBAs), interrupted time series (ITSs), controlled ITSs (CITSs) that meet the inclusion criteria of the Cochrane COPD group (16). We also included clinical practice guidelines and consensus documents that addressed immunoprophylaxis. We considered systematic reviews and meta-analyses only as sources of primary studies.

We searched records published during the study period in the following databases: PubMed, LILACS, Embase, CINAHL, Cochrane Library, Web of Science, Tufts Economic Database, and EconLIT. The detailed search strategy is described in the Supplementary material. For studies with multiple publications, the one with the largest sample size and/or the most recent publication was considered the primary reference; secondary references were used to complement the data. Reference lists of included articles were hand-searched for additional information. If necessary, authors of relevant articles were consulted for missing or clarifying information. The search for gray literature was performed in the following sources: databases of proceedings of regional and international congresses, websites of major medical regional and international societies and associations related to the topic, and the Virtual Health Library. We also contacted members of the PAHO, who gave us access to the database of RSV cases reported from LAC countries between 2017 and 2023. Generic internet search and metasearch engines (Google) were also searched. The complete search strategy is detailed in the Supplementary material.

Selection, data extraction, and risk of bias assessment of each article were performed independently by pairs of reviewers, and discrepancies were resolved by consensus of the entire team. All phases of study selection were carried out using COVIDENCE^®, a web-based platform designed for the systematic review process (17) and data extraction used a previously piloted data extraction form (based on five studies). From eligible articles, the research team extracted the following study information: publication and study characteristics (type of publication, year published, authors, geographic location, study design including domains for risk of bias assessment), study population characteristics (age, sex, sample size, population risk, inclusion, and exclusion criteria), and outcomes [incidence, prevalence, case fatality rate, hospitalization rate, intensive care unit (ICU) admission rate, disease complications of RSV, and RSV seasonality and outbreaks].

We evaluated the risk of bias of the included studies using a checklist developed by the U.S. National Heart, Lung, and Blood Institute, which classifies studies as high risk of bias (poor), uncertain (fair), and low risk of bias (good) (18). For cohort studies and cross-sectional studies assessment, the tool comprises 14 items, while case series studies are assessed based on nine items. The clinical practice guidelines were assessed by the AGREE-II instrument (19). The AGREE-II instrument consists of 23 items, grouped into six domains: (1) Scope and objectives, (2) Participation of decision-makers, (3) Methodological rigor, (4) Clarity of presentation, (5) Applicability, and (6) Editorial independence. A narrative and tabular synthesis were conducted of the available recommendations from clinical practice guidelines and consensus issued by scientific societies and health authorities at the national level.

We reported narrative and structured information alongside descriptive statistics to characterize results. To analyze our data, we conducted proportion meta-analyses using the R software (meta-package) for all analyses. The pooled proportion was calculated as the back-transformation of the weighted mean of the transformed proportions, using inverse arcsine variance weights for both the fixed and random effects models, and reported with a 95% confidence interval (95% CI) (20). We applied DerSimonian-Laird weights for the random effects model where heterogeneity between studies was found (21, 22). The I² statistic was calculated as a measure of the proportion of the overall variation attributable to between-study heterogeneity (23). Funnel plots were used to explore publication bias across at least 10 studies, although the usefulness for non-intervention studies is uncertain. Selective reporting within studies was assessed by comparing available protocols with the reports. We described the seasonality pattern by geographic region from 2017 to 2013 based on PAHO surveillance data (24).

Supplementary Table 1 summarizes the characteristics of the included studies. We included 156 studies: 125 full texts, 21 congress abstracts, and 6 theses. Among the studies included, there were 4 guidelines. Regarding the methodological design, there were 117 cross-sectional studies, 31 case series, 3 prospective cohorts, and 1 case–control study. The studies were carried out in 20 countries, and the most represented were Brazil (25), Colombia (22), Argentina (21), Chile (15), Peru (9), and Mexico (8). Nine studies included multiple countries from LAC. The articles were published between 2013 and 2022, and the inclusion period of participants was from 2000 to 2022, although we only included data from participants recruited from the year 2012 onwards. A total of 16 (10.5%) studies included participants during the years 2019–2020 of the COVID-19 pandemic. The reported duration of the studies ranged from one to 165 months.

Regarding the study population, 112 (73.7%) studies included pediatric patients, 22 (14.5%), both pediatric and adults, 10 (6.6%) only adult patients, and 8 (5.2%) studies did not report the age of the participants. Only 76 (50%) studies reported the clinical risk of the participants (e.g., prematurity, chronic conditions, immunocompromised), with 51 (67.1%) studies that included normal-risk participants, 18 (23.7%) high-risk, and 7 (9.2%) a mix of participants. Regarding the setting, 113 (85.6%) studies included inpatients, 17 (12.9%) included outpatients, and 2 (1.5%) included both.

A total of 1,445,198 samples, with 18% testing positive for RSV, along with epidemiological data from 69,891 patients, were included in the analysis. The diagnostic methods used were polymerase chain reaction (PCR) in 76 (50%) studies, Immunofluorescence (IF) in 36 (23.7%), direct (DIF) in 11 and indirect (IFI) in 17, immunochromatographic rapid test in nine (5.9%), mixed methods in eight (5.3%). Only 27 (17.8%) studies evaluated the type of RSV (A or B).

For cross-sectional and cohort studies, 57 (47.5%) were assessed as being at low risk (good quality), 56 (46.7%) were assessed as moderate risk (fair quality), and only seven (5.8%) as high risk (poor quality). The most frequent domains that did not meet appropriateness were related to sample size justification, the evaluation of exposures more than once or different levels of exposure, and the participation rate of eligible persons. It should be noted that many domains did not apply to the objectives of the included studies, such as blinding of the evaluators, assessment of exposure before the outcome, sufficiency of timeframe, percentage of loss to follow-up, and evaluation of potential confounding variables. For case series studies, 23 (74.2%) were rated as low risk, 6 (19.4%) as moderate risk, and only two (6.4%) as high risk. The domain that did not meet appropriateness more frequently was whether the cases were consecutive. Only one case–control study was included and rated as moderate risk due to a lack of information about the selection of participants, the use of concurrent controls, the blinding of the assessors, and because potential confounding variables were not measured. The risk of bias assessment by study design is presented in Supplementary Tables 3–5.

Table 1 describes the main findings of RSV burden of disease by age groups with separate estimations for periods 2012–2016 and 2017–2022, for low-risk of bias (RoB) studies, and for high-risk patients when available. Regarding the prevalence of RSV infections, the age group most represented among the included studies was 0–24 months (38 studies), mostly presenting a low RoB. The positivity prevalence among suspected cases in this age group was 42.3% (95%CI 34.1–51.0), with the highest prevalence of 57.6% (95%CI 38.7–74.5) among 0-12-month-old infants. The age group of ≥65 years was poorly represented (3 studies), with a prevalence of 10.7% (95%CI 6.7–17.3; Figure 2).

The incidence of LRTI due to RSV was 14.7 (95%CI 7.8–21.7) per 100 symptomatic 0–24 months-old infants and higher for the 0–6 months age group (21.6%; 95%CI 9.0–34.2). The proportion of LRTI among RSV cases among 0–24 months-old infants was 72.1 (95%CI 9.0 to 98.5) and was similar for children aged 0 to 18 years old. The RSV-LRTI lethality was 0.6% (95%CI 0.3–1.0) in the 0–24 months and 2.9% (95%CI 0.7–1.7) in the 0–5 years old group, respectively. The ≥65-year lethality was estimated to be 23.1% (7.6 to 52.2) based on a single study on a high-risk population (Figure 3). Among RSV-LRTI cases, studies reporting bronchiolitis showed 76.4% (95%CI 33.4 to 95.4) and 70.3% (95%CI 36.7 to 90.6) in the 0–24 months and 0–5 years old, respectively.

The highest proportion of ICU admissions among RSV-LRTI was 42.0% (95%CI 7.9–86.0) for the 0–24 months age group, with a mean length of stay at ICU of 3.2 days (0.1 to 6.2) and general ward of 6.5 days (4.7 to 8.3), respectively. Antibiotic use in children aged 0–24 months was 47.4, and 13.6% of this age group required invasive ventilation.

RSV type A presented a higher proportion than type B from 0 to 24 months (51.6% vs. 34.4, respectively), being more markedly at lower age. The most frequent co-infections were viral with similar percentages in younger and older patients, 20.8% between 0 and 24 months and 20.4% in ≥65 years, with more coinfections detected in the youngest during 2012–2016 compared to 2017–2022 (24.1% vs. 7.4%). The summary estimates reported in Table 1 are derived from meta-analyses, which forest plots are available in the Supplementary Figures 1–5. The heterogeneity was high for most meta-analyses.

Based on our literature search, we identified four guidelines for treating RSV in the LATAM region. The guidelines were developed in Argentina (26), Honduras (27), Brazil (28), and Chile (29). Three guidelines focused on the treatment and prevention of RSV-related bronchiolitis, while one focused on bronchiolitis of any viral cause (27). The assessed guidelines show notable methodological limitations. They are narrative reviews without systematic searches, clear methodologies, external expert reviews, and a proposal to update existing guidelines. They lack clear overall objectives (26, 27), and only one guideline specifies the aim of reducing RSV hospitalizations (28). Beneficiaries and outcomes are vaguely described (26, 27). Target populations are not mentioned consistently throughout the guidelines. While one outlines it briefly (26), another specifies it in detail (27). No guideline provides a specific health question. Treatment recommendations are included for premature infants. However, the infants’ age is presented ambiguously (26, 29) or not specified (27, 28). Author details are only fully provided in one guideline (29). No guidelines included views of the target population or provided information on literature searches. Three of the four guidelines do not mention the treatment benefits, side effects, and risks (27–29). Recommendations vary in specificity from general (27) to specific (26, 28, 29). In three guidelines, these recommendations are linked to clinical evidence from previously conducted studies (26, 27, 29). Facilitators, barriers, and stakeholders are not described. All guidelines lack implementation guidance and resource implications. Cost-effectiveness and economic data are missing, and monitoring criteria are not presented. Editorial independence is mentioned in only one guideline (27). The complete evaluation of the assessment of guidelines is in Supplementary Table 6.

These four guidelines provide recommendations for palivizumab use in vulnerable populations, offering insights into inclusion criteria, prescription procedures, post-administration serum levels, and epidemiological data related to RSV infection. There is consensus recommending palivizumab in premature babies less than 29 weeks, without chronic pulmonary disease in premature babies younger than 12 months old at the start of RSV station. Also, to infants under 12 months with hemodynamically significant heart disease or infants under 24 months of age undergoing heart transplant during RSV season.

The Figure 4 shows the seasonality pattern of RSV isolates from 2017 to 2023 across three geographical regions: South America, Central/North America, and the Caribbean. The time series, from PAHO surveillance, data show a clear cyclical pattern for each region. South America exhibits pronounced peaks in RSV isolates each year in the winter season, having the highest number of isolates; the intensity of these peaks seems to be diminishing slightly over the years. Central America’s RSV isolates also follow a seasonal trend. Still, the peaks are less pronounced than those in South America, relatively consistent over the years, with no notable increase or decrease in the number of isolates. The Caribbean shows the least pronounced peaks among the three regions and is more spread out, with a noticeable increase in recent years. Of note, there was a clear gap during the pandemic (2020–2021).