A total of 7434 references were retrieved from the search, of which 932 duplicates were removed, leaving 6502 references for title and abstract screening. Following this, 330 publications were selected for full-text screening, and 21 studies were eligible for this review (Fig. 1). Detailed characteristics of the 21 studies are summarized in Table 1. Seventeen studies were cohort studies (n = 15 retrospective, n = 2 prospective) [2, 14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30], two were case–control studies [31, 32], and one each was a case series [33] and an RCT [34]. Studies were conducted in Austria, Australia, Canada, France, Germany, Israel, Spain, and Switzerland, and approximately half were from the USA (n = 10). Included studies captured data from 1990 to 2019. Across studies, sample sizes ranged from 10 to 756 persons and included in- or outpatients (n = 11 and 2 studies, respectively), mixed (n = 3 studies), or not reported (n = 5 studies) treatment settings. Per the prespecified RoB assessment rating score, five studies had a low RoB [18, 21, 22, 30, 33], nine had a moderate RoB [2, 14, 19, 25, 27,28,29, 31, 34], and four had a high RoB [15, 17, 23, 32] (Table S7).

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram

We focused on the most assessed clinical (acute cellular rejection or graft dysfunction, bronchiolitis obliterans syndrome, cardiovascular events, decline in lung function, and sustained function loss) and utilization sequelae (readmission, placement in a skilled nursing facility, and use of home services) in the studies. Additionally, we stratified the patients into two groups: general population (n = 7 studies) and transplant recipients (lung [n = 11 studies] and hematopoietic [n = 3 studies]) because these were the most frequently evaluated.

In the general population, the clinical sequelae identified were sustained function loss (n = 1 study), cardiovascular events (n = 2 studies), and decline in lung function (n = 1 study). The estimated frequencies of these sequelae were 33.5% (95% CI 27.6–39.9), 20.6% (95% CI 14.8–27.9), and 18.0% (95% CI 8.9–32.7), respectively (Figs. 2 and S1). Other more specific clinical sequelae identified, with frequencies ranging from 24.6% to 32.5%, were decreased ability to perform complex activities and worsening respiratory and chronic disability 6 months after hospital discharge [15]. These sequelae were not included in the meta-analysis to avoid overestimation/representation of function loss from one study since sustained function loss was already covered (Table S8).

Proportions of RSV-associated cardiovascular events and function loss by a follow-up period and b treatment setting

Among the utilization sequelae, the proportion of hospitalized patients discharged to a skilled nursing facility ranged between 10.1% and 17.6%, producing a pooled estimate of 13.1% (95% CI 8.4–19.7, n = 2 studies) (Fig. 2). However, readmission was the most reported utilization sequela in the general population (n = 5 studies). The highest frequency of event estimated at 38.3% (95% CI 34.6–42.1) occurred ≥ 6 months after discharge, followed by 2–5 months (28.5% [95% CI 22.7–35.1]) and at ≤ 1 month (17.0% [95% CI 11.72–4.2]) (Fig. S2). Note that in the follow-up analyses, the estimate for the ≥ 6 months period was based on one study where we combined the results of the two groups (those with one hospitalization and those with two or more hospitalizations) [14]. This study also reported cases with one or more emergency department visits during this period, and the proportions ranged between 21.9% and 26.1% (Table S8).

By treatment setting, stratification was possible for readmission only, with pooled frequency of 26.3% (range 11.4–38.3%) for inpatients, while among the mixed group (in- and outpatients), the percentage of occurrence varied between 16.2% and 20.4% (Fig. S2B).

Additional sequelae such as use of home health services, higher level of care, and loss of previous independence were identified in the general population (Table S8).

In hematopoietic transplant patients, bronchiolitis obliterans syndrome and decline in lung function were reported in one study, with a proportion of 1.2% (95% CI 0.1–6.6) and 12.5% (95% CI 3.5–36), respectively (Figs. 3 and S1). Other identified clinical sequelae in hematopoietic transplant patients were supplementary oxygen therapy [19] and pulmonary impairment [28], with proportions of 11.1% and 26.7%, respectively (Table S9). Stratification by treatment setting was not possible because of the limited number of studies.

Proportions of RSV-associated bronchiolitis obliterans syndrome by a follow-up period and b treatment setting

Decline in lung function was the clinical sequelae with the highest pooled proportion in lung transplant patients (49.8% [95% CI 32.7–66.9], n = 4 studies). By follow-up period, the pooled estimate varied from 34.5% at ≥ 6 months to 83.3% at 2–5 months (Fig. S1A). Per treatment setting, the highest proportion was 60.5% (95% CI 16.4–92.2) for treatment setting not reported, followed by inpatients 46.7% (95% CI 27.3–67.2) (Fig. S1B). No data was available for outpatients and mixed treatment settings.

In lung transplant patients, bronchiolitis obliterans syndrome was the most assessed clinical sequela (n = 10 studies), ranging between 6.6% and 60%, with a pooled proportion of 29.5% (95% CI 21.8–38.4). Stratified by follow-up period, bronchiolitis obliterans syndrome was more frequent at ≥ 6 months (30.1% [95% CI 25.0–35.8]) than at 2–5 months (13.9% [95% CI 2.7–48.5]) (Fig. 3a). By treatment setting, a higher frequency of bronchiolitis obliterans syndrome was reported among inpatients (pooled estimate 33% [range 27.6–60%]; n = 4 studies) than among outpatients (38.5%, n = 1 study). For studies with unspecified (n = 3 studies) or mixed treatment settings (n = 2 studies), the aggregated proportion of bronchiolitis obliterans syndrome cases was 29.5% (range 23.5–46.7%) and 15.1% (range 6.6–50%), respectively (Fig. 3b).

Other clinical sequelae identified in lung transplant patients were acute cellular rejection, chronic lung allograft dysfunction, or graft dysfunction, with estimated proportions ranging between 4.8% and 72.2% (Fig. S3).

No data were available on the utilization sequelae of transplant patients.

Overall, six studies reported data for estimating the risk of sequelae in patients with RSV versus patients without RSV, which included patients with parainfluenza, influenza, or both as control group.

The estimated risk of cardiovascular events (clinical sequela) in RSV relative to influenza was 1.4 ([95% CI 1.0–2.0], n = 2 studies), while for readmission, the risk relative to non-RSV (influenza and RSV-negative patients with influenza-like illness [ILI]) was 1.2 ([95% CI 1.1–1.3], n = 4 studies) (Fig. 4a, b). Relative risks of readmission at ≤ 1 and ≥ 6 months after discharge were estimated with one study each and were comparable, 1.2 (95% CI 0.6–2.4) and 1.2 (95% CI 1.1–1.3), respectively. For function loss (utilization sequela), Ackerson et al. reported similar odds of skilled nursing facility placement between patients with RSV and patients with influenza, whereas for home services care, the odds were higher in patients with RSV than in patients with influenza [14]. The results of this study were not meta-analyzed because only one study reported comparative data for these sequelae.

Relative risks of RSV-associated a readmission, b cardiovascular events, and c pulmonary impairment by follow-up period

In hematopoietic transplant patients, the risk of pulmonary impairment (clinical sequela) in patients with RSV versus patients with parainfluenza and influenza combined was 1.4 ([95% CI 1.0–2.0], n = 1 studies). However, compared to parainfluenza only, the risk significantly reduced to 0.5 (95% CI 0.3–0.8) (Fig. 4c). No data were available on lung transplant patients or other sequelae.