FormalPara Key Summary Points

This is the first systematic review of sequelae of RSV disease among adults.

RSV infection can cause long-term complications and chronic sequalae in adults, yet, these implications has not been fully described in this population.

High-morbidity sequelae include bronchiolitis obliterans syndrome, decline in lung function, sustained function loss, cardiovascular events, and graft dysfunction.

Following RSV-related hospitalization, patients experience decline in functional status, including discharge to a skilled nursing facility and reduction in functional status scores.

The burden of RSV sequela supports the need for effective RSV vaccination programs in adults.

Introduction

RSV is an enveloped single-stranded RNA virus that can cause mild clinical manifestations among adults, including rhinitis, cough, and fever, or progress to complications such as pneumonia and chronic heart and lung disease exacerbations, which may require intensive care unit admission and mechanical ventilation [1, 2]. In 2019, an estimated 787,000 RSV-associated hospitalizations occurred among adults ≥ 65 years in high-income countries [3]. Although two RSV vaccines (Arexvy, Abrysvo) have been licensed for use in adults in some countries [4], and additional vaccines are in clinical development [5, 6], the burden of RSV infection has not been fully estimated in this population [7], as the long-term effects of RSV should also be considered.

Respiratory viral infections can present complications and chronic sequelae. Beyond respiratory sequelae, influenza infections are associated with cardiovascular, neurologic, hepatic, and renal manifestations [8]. Similarly, persistent symptoms have been reported in COVID-19 survivors, such as neuropsychiatric (e.g., brain fog, anxiety) and cardiac (e.g., palpitations, tachycardia) manifestations [9]. While there is information on RSV-related intrahospital complications, information on longer-term RSV sequelae in the adult population is limited.

Assessment of the full burden of disease associated with RSV infection, including any RSV-related sequelae, is required for fully informed policy recommendations among adults. No review of RSV sequelae among adults has been published. Therefore, we conducted a systematic literature review to describe the frequencies, duration, and severity of adult RSV sequelae that occur after hospitalization or resolution of acute infection and the risk of sequelae in adult patients with RSV versus those without RSV.

Methods

Literature Search Strategy

This systematic literature review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [10] (Table S1) and was registered with PROSPERO (CRD42022327317). An experienced systematic reviewer developed the search strategy and performed the literature search in MEDLINE (accessed via PubMed), Embase, LILACS, and SciELO databases and in the following grey literature sources: GreyLit, OpenGrey, Centers for Disease Control and Prevention, European Centre for Disease Prevention and Control, World Health Organization, European Society of Clinical Microbiology and Infectious Disease, IDWeek, International Society for Influenza and other Respiratory Virus Diseases, Respiratory Syncytial virus Consortium in Europe, and Respiratory Syncytial Virus Network.

Searches in MEDLINE and Embase applied the search strings listed in Table S2, whereas the searches in LILACS, SciELO, and grey literature sources used the following keywords: respiratory syncytial virus, RSV, complications, sequelae, consequences, and morbidity, which were combined using the Boolean operators “OR”, to broaden the population or outcomes search, and “AND”, to narrow the search to the relevant publications of interest. Peer and non-peer-reviewed publications from 2000 to 2021 in English, French, Portuguese, and Spanish were considered for eligibility. Reference lists of relevant studies were screened to identify additional articles.

Definition of Sequelae

We categorized sequelae into two groups as follows: clinical sequelae defined as any pathological conditions, functional deficits, or worsening of underlying medical conditions that occur after RSV hospitalization or among non-hospitalized adults after the resolution of acute illness. Utilization sequelae include readmission, placement in a skilled nursing facility, or use of home services after RSV hospitalization or among non-hospitalized adults. The patient categories of RSV sequelae are defined in Table S3.

Study Selection Process

All publications retrieved were de-duplicated, after which the unique publications were screened using a two-phase process. First, the title and abstract of the publications were independently screened by a single reviewer against the population, intervention, comparator, outcomes, and time (PICOT) eligibility criteria outlined in Table S4. Thereafter, a second reviewer independently screened 30% of the titles and abstracts. Disagreement between the reviewers was 3%. Given that the discordance rate was less than 10% of the abstracts screened, no further screening was performed. In the second phase, the full text of each publication included in the first phase was assessed by two independent reviewers using the same eligibility criteria. At both phases, any discrepancies between the reviewers were discussed, and if not resolved, a third reviewer was consulted. As part of the selection process, studies were excluded if the entire study population was children (i.e., aged < 18 years), the study had a mix of children and adults and the results were not reported separately by age, the sequelae were not related to RSV, the study was conducted in low-to-medium-income countries, or the study design was not a randomized control trial (RCT) or observational study (Table S4). We limited to high-income countries to allow for greater homogeneity in medical diagnostic practices, which vary on the basis of resources available for medical care since this would impact the frequency of sequela detection.

Data Extraction and Risk of Bias Assessment

A data extraction table was designed to collect all relevant data. Full-text journal articles underwent a risk of bias (RoB) assessment using the National Institutes of Health (NIH) tool for cohort and cross-sectional, case–control, and case-series studies (Table S5). Since each arm of an RCT was treated as a unique cohort, the NIH tool used for cohort studies was used for RCTs. Following the assessment of each publication, a rating score classified studies as good quality with low RoB (> 70% scoring), fair quality with moderate RoB (50–70%), and poor quality with high RoB (< 50%). Both RoB assessment and data extraction were conducted by a single reviewer, and a second reviewer verified the accuracy of all assessments and extracted data.

Data Analysis

We used the Wilson ‘score’ method with asymptotic variance without continuity correction to calculate 95% confidence interval (CI) for proportions [11]. Random-effects meta-analyses using restricted maximum likelihood were employed to pool proportions and relative risks from different studies. Cochran’s Q test statistic was computed to test for heterogeneity examining the null hypothesis that all studies produce the same effect. The between-study heterogeneity was quantified using the I2 statistic, as the power of Cochran’s Q test is low in analyses with a small number of studies [12]. A meta-analysis was conducted for each sequela if at least two studies presented data or one study reported data on different non-RSV populations. All analyses were performed in RStudio 1.4 [13]. When data were sufficient, we stratified the analysis by follow-up period (≤ 1, 2–5, and ≥ 6 months), subpopulation (lung transplant, hematopoietic transplant, and general population), and treatment setting (inpatient, outpatient, mixed, and not reported). Details on the data stratification are summarized in Table S6.

Results

Search Results and Study Characteristics

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).

Fig. 1
figure 1

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

Full size image
Table 1 Number of studies reporting each sequela
Full size table

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.

Frequency of RSV-Associated Sequelae

General Population

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).

Fig. 2
figure 2

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

Full size image

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).

Transplant Patients

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.

Fig. 3
figure 3

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

Full size image

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.

Risk of RSV-Associated Sequelae

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.

General Population

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.

Fig. 4
figure 4

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

Full size image

Transplant Patients

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.

Discussion

To our knowledge, this is the first systematic review and meta-analysis that evaluates the frequency and risk of longer-term sequelae after hospitalization or resolution of acute RSV infection in adults. Published studies were highly heterogeneous and have primarily assessed high-morbidity sequelae among transplant patients such as bronchiolitis obliterans syndrome, decline in lung function, acute cellular rejection, chronic lung allograft dysfunction, and graft dysfunction. Fewer studies have focused on sequelae among adults in the general population hospitalized with lower respiratory tract infections (LRTI) due to RSV, but those that exist report sustained function loss, cardiovascular events, the decline in lung function, and high frequency of all-cause readmission and admission to a skilled nursing facility. These sequelae in turn are consistent with those identified among adults hospitalized for other respiratory infections (e.g., influenza, COVID-19) as well as acute LRTI overall [35,36,37]. Where comparative data were available for an outcome (i.e., all-cause readmission and cardiovascular events), sequela risk was approximately 20% higher among adults with RSV infection than for controls (largely persons with influenza).

Sequelae mostly occurred among older adults, indicating the burden likely increases with age and adds a disease impact beyond the high mortality rate seen among older adults hospitalized with RSV disease. Older adults are more susceptible to serious RSV infection and complications. This may be due to changes in their immune system and lung function, which impair their ability to respond to infection and prolong inflammation even after the clearance of an infection [38]. Furthermore, it may take several months for respiratory function and the ability to perform activities of daily living to return to the state before RSV diagnosis [15]. Additionally, a significant proportion may be discharged to assisted care facilities (skilled nursing homes, rehabilitation centers) not needed prior to RSV infection [14, 30].

Using results from four controlled studies, we calculated a pooled estimate of the increase in all-cause readmission risk for patients with RSV compared to patients without RSV (largely patients with influenza or ILI) of 20%, which is likely due in part to cardiovascular and pulmonary sequelae as well as general functional decline following RSV infection as documented in other more specific studies. As the comparison group (influenza-related hospitalizations) is known to have multiple potential complications and sequelae and increased risk of readmission, this relative risk increase is likely conservative [39]. Our estimated pooled prevalence of readmission (23.9% [95% CI 16.8–32.9]) was higher than that reported in a meta-analysis of COVID-19 survivors using studies of the general population (7.5% [95% CI 5.4–9.6]) [40]. Similarly, in a cohort study of 106,543 patients in the USA, 9% of patients were readmitted 2 months after discharge from COVID-19 hospitalization [35]. Differences across studies for several etiologies may reflect differences in populations (e.g., younger patients in studies of COVID-19) and study methodologies. Regardless, we found no evidence that readmission occurs less frequently following RSV LRTI than following other viral etiologies of LRTI.

Several studies documented a reduction in functional status following an RSV-related hospitalization, including requiring discharge to a skilled nursing facility and a reduction in functional status scores. The pooled frequency of skilled nursing facility placement at hospital discharge or within 1 month after RSV illness was 13% (95% CI 8.4–19.7). Branche et al. found 33%, 32%, and 25% of patients with RSV experienced decreased ability to perform activities of daily living (e.g., dressing and toileting), instrumental activities of daily living (e.g., taking medication and preparing meals), and increased breathlessness, respectively, at 6 months compared to pre-hospitalization baseline [15]. Furthermore, in Ackerson et al., older adults (≥ 60 years) infected with RSV were more likely to require home health service after discharge (OR 1.3; 95% CI 1.0–1.6) than those infected with influenza [14]. Reduced functional status following RSV illness is consistent with the negative impact seen on cognitive function following respiratory infections. For example, Girard et al. documented increase in cognitive impairment following community-acquired pneumonia with approximately a third of subjects experiencing cognitive impairment at 2 and 6 months after hospitalization versus 12% before hospitalization. Deficits were most often noted in visuospatial function, attention, and memory [36].

We identified a large body of literature reporting on medically significant sequelae of RSV infection among transplant patients, particularly lung transplant recipients. Multiple studies reported on the high frequency of respiratory sequelae, such as decline in lung function and bronchiolitis obliterans syndrome (a form of chronic allograft rejection), among lung transplant patients. Our pooled estimate of decline in lung function after RSV in lung transplant patients was 49.8% (95% CI 32.7–66.9), which is similar to that reported for lung transplant recipients following COVID-19 (40.9%) with a follow-up period up to 1 year [37]. Moreover, our pooled estimate of the frequency of bronchiolitis obliterans syndrome (31.3% [95% CI 26.0–37.1]) ≥ 6 months following RSV infection was in the range of those reported for other viral infections. For example, bronchiolitis obliterans syndrome at any stage was reported after respiratory viral infections, including RSV, parainfluenza, influenza, and adenovirus, in 30–53% of patients that received lung transplants between 1996 and 2000 [22]. In another cohort of lung transplant recipients, 12% developed bronchiolitis obliterans syndrome within 1 year after respiratory infections, including patients with rhinovirus, coronavirus, and influenza A infections [41]. Other related medically significant sequelae identified were acute cellular rejection and graft dysfunction in lung transplant patients and pulmonary impairment in hematopoietic transplant patients.

We found two studies reporting on the frequency of longer-term cardiovascular sequelae in the general population [18, 20], with an elevated relative risk following RSV- versus influenza-associated hospitalization. Influenza is a useful comparator because influenza infection has been shown to increase the risk of myocardial infarction in the first 3–7 days after infection, which can be alleviated by vaccination [42, 43]. Moreover, vaccination can reduce the risk of mortality in adult patients with cardiovascular diseases [44, 45]. Respiratory infections have been associated with the epigenetic activation of the nuclear factor ‘kappa-light-chain-enhancer’ of activated B cells and the transforming growth factor-beta signaling pathway, which induces airway inflammation and vascular remodeling processes [46]. Other etiologies of respiratory infection have been associated with cardiovascular sequelae. For example, a meta-analysis of COVID-19 survivors with an approximately 2-month follow-up period estimated an overall prevalence of any abnormal cardiovascular magnetic resonance imaging of 46.4% (95% CI 43.2–49.7%) and a pooled prevalence of myocarditis of 14.0% (95% CI 11.6–16.8%) [47]. Pneumococcal pneumonia is also associated with an increased risk of cardiovascular and thromboembolic events, with relative risks ranging from 2- to 48-fold higher than controls [48].

The burden of infections such as influenza, pneumococcal disease, and COVID-19 and their long-term sequelae has been estimated in adults, and vaccination programs and other prevention strategies have been established to alleviate the public health burden [49]. Available literature suggests similar risks exist following RSV infection, but ideally, larger controlled studies would be conducted to fully document this trend. Self-controlled studies where a pre-infection follow-up period is used to compare the incidence of events before and after RSV infection are one efficient approach that could be employed.

This systematic review has several important strengths, including an extensive search in multiple databases (MEDLINE, Embase, LILACS, and SciELO), as well as grey literature sources, resulting in a high likelihood that eligible studies were captured. Secondly, we included data published over a 30-year period, from 1990 to 2019. In addition, these studies contained data from a substantial number of patients and nine different countries. Furthermore, our broad inclusion criteria enabled us to scope the literature on sequelae and not limit ourselves to a predetermined set of outcomes observed within specific settings.

There are several limitations to note as well. Only five studies were identified with a low risk of bias. The retrospective nature of most studies limited the identification of a comparison group and the ability to identify potential confounding factors. We identified relatively few studies, hampering the possibility of performing relevant subgroup analyses (e.g., by age), and data originated only from high-income countries, mainly the USA, where clinical practices might differ from other countries. We restricted the study to the four languages because this was all that was available among authors, which may have resulted in missed relevant publications. Most studies focused on hospitalized patients, and only two studies reported outpatient data. The follow-up period was not harmonized among studies, limiting the comparison of the proportions of patients with sequelae on follow-up visits. Few studies reported the onset time, duration, and severity of sequelae. No studies stratified the risk of RSV sequelae by preexisting conditions besides lung and hematopoietic transplants. The heterogeneity and scarcity of data impeded our assessment of whether RSV infection was an independent contributor to the reported sequelae. Lastly, the source articles did not consistently report the approach used to diagnose the RSV cases. RSV can only be diagnosed by laboratory testing in adults because of its clinical presentation being very similar to other respiratory viruses but different laboratory approaches have different sensitivities [50, 51]. If low sensitivity approaches were used or RSV testing was incomplete in the source population, RSV events could have been classified as controls and this would have likely reduced the association seen in the analyses (i.e., bias results to the null). It is also possible that low levels of RSV testing among adults resulted in only more clinically severe cases being identified [52], and as such the observed risk of sequela could be higher than if all persons were consistently tested.

Economic analyses for etiologies preventable with licensed vaccines, such as PCV and influenza vaccine, have included the value of preventing sequelae [53,54,55,56,57,58]. As the first RSV vaccines (Arexvy, Abrysvo) are now available in older adults and many others are under development for the prevention of RSV LRTI in adult populations [4], economic evaluations should similarly include the value of preventing sequelae. This recommendation would be facilitated by more robust studies that comprehensively enroll and provide RSV testing for all patients hospitalized with lower respiratory tract disease, conduct follow-up for at least 6 months, and enroll enough patients to provide data on important subgroups such as the oldest adults and those with underlying conditions. Additionally, innovative methods may have a role, such as using a self-control approach with the pre-admission period constituting the control period.

Conclusions

Consistent with literature on other respiratory pathogens, our findings suggest the potential for an elevated risk of RSV sequelae in adults, encompassing respiratory manifestations, function loss, readmission to hospital, cardiovascular events, and transplant complications. Where data were available, sequelae following RSV infections were at least as common as those following influenza infections. More studies are needed to determine the most frequent and severe adult RSV sequelae.