Respiratory Syncytial Virus vs Influenza Virus Infection: Mortality and Morbidity Comparison Over 7 Epidemic Seasons in an Elderly Population

Abstract

Background

Respiratory syncytial virus (RSV) infection is gaining interest due to the recent development of vaccines but is still misdiagnosed in the elderly. The primary objective was to compare all-cause mortality at day 30. Secondary objectives were to compare clinical presentation and rates of consolidative pneumonia, hospitalization, and intensive care unit (ICU) admission.

Methods

A single-center retrospective study was conducted in a French university hospital during 7 epidemic seasons including 558 patients aged ≥75 years: 125 with RSV and 433 with influenza (median age, 84.8 years).

Results

Patients with RSV had more respiratory symptoms (wheezing, dyspnea) whereas patients with influenza had more general symptoms (fever, asthenia, myalgia). The following were higher in the RSV group: consolidative pneumonia (28.8% vs 17.2%, P = .004), hospitalization (83.2% vs 70%, P = .003), ICU admission (7.2% vs 3.0%, P = .034), and length of stay (median [IQR], 9 days [2–16] vs 5 days [0–12]; P = .002). Mortality rates at day 30 were comparable (9.6% vs 9.7%, P = .973).

Conclusions

This study included the largest cohort of patients infected with RSV aged >75 years documented in-depth thus far. RSV shares a comparable mortality rate with influenza but is associated with higher rates of consolidative pneumonia, hospitalization, ICU admissions, and extended hospital stays.

Respiratory syncytial virus (RSV) is the most common respiratory pathogen in infants and young children worldwide [1]. The clinical presentations and potential complications associated with RSV infection in the elderly population have garnered increased attention since more and more diagnosis technologies are available and affordable in the context of advancements in novel prophylactic treatments and vaccine development [2–5].

Some studies have evidenced severity of RSV infections. However, the majority of these studies have focused on general populations of all ages [6, 7] or specific vulnerable populations, such as hospitalized patients [8] or those with chronic comorbidities (chronic obstructive pulmonary disease, heart failure) [9, 10]. The few studies addressing the elderly populations were often conducted in outpatient settings, among nursing home residents [11] or independently living adults aged ≥60 years [12, 13].

A recent meta-analysis showed that RSV leads to clinical manifestations at least as severe as those induced by influenza [14]. However, these studies focused almost exclusively on populations aged <75 years, and the clinical phenotype of patients has not been extensively documented.

The number of in-hospital admissions during the RSV epidemic season seems to be comparable to that of the flu [12, 15]. In a study conducted in the United Kingdom [16], RSV contributed to an estimated 71 hospitalizations annually per 100 000 for acute respiratory tract infections among patients aged 65 to 74 years and at least 3 times more for adults aged ≥75 years, confirming the heavy burden of RSV on health care systems, which will undoubtedly increase as the world's population ages [17].

The aim of this study was to compare the day 30 mortality rates of RSV infections with those of influenza in a general population aged >75 years attending emergency departments. Secondary objectives were to compare symptoms, clinical signs, and biological parameters at baseline, as well as complications during hospitalization, including lower respiratory tract infection (LRTI), consolidative pneumonia, in-hospital admission, intensive care unit (ICU) admission, and hospital length of stay.

METHODS

Design

This single-center retrospective cohort study was conducted between 31 March 2016 and 5 May 2022 at the Henri Mondor University Hospital. The trial was designed and conducted following the International Conference on Harmonization guidelines (ICH E6) for good clinical practice and was approved by the local ethics committee (IRB 00011558). Due to the retrospective nature of the study involving deidentified data, informed consent was not required. All authors had access to the study data.

Testing Practices

During the winter seasons of the study period, nasopharyngeal swab testing for RSV, influenza A, and influenza B was performed on all patients who exhibited respiratory and/or infectious symptoms (including fever as the only sign) and were at high risk for complications (eg, elderly or immunocompromised).

Patients

All patients aged ≥75 years consulting the emergency department with a positive result on viral reverse transcription polymerase chain reaction (RT-PCR) for RSV or influenza were included. Patients with a positive RT-PCR result for both viruses and those who had already consulted with a positive RT-PCR result in the previous month were excluded.

Samples

Nasopharyngeal swabs in transport media were held at 4 °C (or −80 °C if >12 hours) prior to testing on the Cepheid Xpert Xpress Flu/RSV (January 2016–June 2020) or Xpert Xpress CoV2/Flu/RSV (after June 2020).

Data Collection

Data collection was carried out by 2 physicians following the same protocol and using a standardized data collection tool. All data were collected in the Assistance Publique–Hopitaux de Paris's integrated electronic patient record (ORBIS; Agfa HealthCare). Mortality data were verified at the earliest 1 year after each patient's inclusion date via the same Orbis software, itself fed by French public health data. To reduce the bias inherent in retrospective studies, all study variables were defined a priori. The patient's independence was assessed with the Katz Activities of Daily Living Scale and the Lawton Instrumental Activities of Daily Living Scale. The Charlson Comorbidity Index was used to assess patients’ clinical frailty. Immunosuppression was defined as patients consuming at least the equivalent of 7.5 mg/d of prednisone for ≥2 weeks [18], taking any other immunosuppressive drug, or undergoing chemotherapy. LRTI was defined by the presence of at least 1 of the following criteria: dyspnea, wheezing, oxygen saturation <90%, or consolidative pneumonia. Consolidative pneumonia was defined by the presence of lung consolidation images on chest imaging (radiology or computed tomography scan) interpreted by a board-certified radiologist. The category “ICU admission criteria” included patients admitted to the ICU and those rejected due to their functional status.

Statistical Analysis

The study population was described in terms of sociodemographic, clinical, and biochemical features. Univariate analysis was based on a Student t test or Mann-Whitney nonparametric test for continuous variables and a chi-square test or Fisher exact test for categorical variables. Time-to-event analyses were estimated with Kaplan-Meier analyses, and differences in survival were analyzed with a log rank test. All tests were 2-tailed and P ≤ .05 was considered significant. Multivariate associations were computed with binary logistic regression models. For all models, variables with nominal 2-tailed P values <.1 were entered into the multivariate model, except for variables with obvious collinearity.

RESULTS

Patients

During the study period, 21 989 patients exhibiting respiratory and/or infectious symptoms were tested by RT-PCR in the Henri Mondor Emergency Department. Among them, 1484 patients tested positive for RSV or influenza. Of these, 563 were >75 years of age. Five patients had to be excluded: 2 for lack of medical records, 2 for positive RT-PCR results for both viruses, and 1 for a passage with a positive RT-PCR result <1 month previously. In the end, 558 patients were included in our study: 125 in the RSV group and 433 in the influenza group (Figure 1). The distribution of patients within each epidemic season is shown in Supplementary Figure 1.

There was no significant difference between the RSV and influenza groups in terms of age (median [IQR], 85.5 years [81.3–90.8] vs 84.7 [80.3–89.2], respectively; P = .297) and sex (men, 39.2% vs 46.0%; P = .180; Table 1). Charlson Comorbidity Index was similar between the groups (median [IQR], 7 [5–9] for RSV vs 6 [5–8] for influenza; P = .337). Chronic congestive heart failure was more prevalent in the RSV group (48.8% vs 36%, P = .010). There were no differences in other comorbidities—in particular, chronic obstructive pulmonary disease, diabetes, or oncohematologic pathologies (Supplementary Table 1). The RSV and influenza groups were comparable in terms of immunosuppressive treatments (12.8% vs 9.2%, P = .243) and annual influenza vaccination (54.3% vs 55.9%, P = .851), but the latter information was available for only 198 of the 558 included patients (36.8% of RSV and 35.1% of influenza). There was no difference between the RSV and influenza groups regarding the rates of independently living patients (37.5% vs 32.9%, P = .352) or institutionalized patients (16.8% vs 15.0%, P = .626).

Clinical and Biological Characteristics at Admission in the Emergency Department

The time from symptom onset to emergency department consultation was longer in the RSV group at a median 3 days (IQR, 1–5) vs 2 days (1–4) for the influenza group (P = .024; Table 2).

At baseline, patients with influenza were more likely to present with fever (48.4% vs 33.6%, P = .004), myalgia (19.9% vs 6.4%, P < .001), and asthenia (54.3% vs 36.8%, P = .001). Patients with RSV had lower transcutaneous oxygen saturation measurements (median [IQR], 92% [88%–96%] vs 94% [90%–96%]; P = .019) and more often presented with dyspnea (64.0% vs 44.1%, P < .001) and wheezing on auscultation (32.8% vs 15.9%, P < .001).

On initial laboratory tests, the RSV group had higher levels of white blood cells (median [IQR], 9.7 × 10⁹/L [7.0–12.9] vs 7.4 × 10⁹/L [5.7–9.8]; P < .001) and neutrophils (8.0 × 10⁹/L [5.1–10.5] vs 5.6 × 10⁹/L [3.9–7.7], P < .001) than the influenza group (Table 2). There was no difference between the groups in terms of lymphocytes (median [IQR], 0.9 × 10⁹/L [0.6–1.3] vs 0.8 × 10⁹/L [0.5–1.2]; P = .162) and C-reactive protein levels (56.2 mg/L [19.2–112.9] vs 45.3 mg/L [19.4–100.3], P = .686).

Emergency Department Care and Patient Outcomes

Acute LRTIs were more common in patients with RSV (83.2% vs 61.5%, P < .001), as was consolidative pneumonia (28.8% vs 17.2%, P = .004; Table 3). Antibiotics were administered to 47.2% of the RSV group as compared with 35.8% of the influenza group (P = .021). In addition, 67.7% of the influenza group received oseltamivir.

Patients in the RSV group were associated with a higher rate of hospitalization (83.2% vs 70.0%, P = .003). In multivariate analysis, the adjusted odds ratio for hospitalization in patients with RSV infection was 1.95 (95% CI, 1.10–3.45; Table 4). Furthermore, the RSV group exhibited a higher rate of ICU admission (7.2% vs 3.0%, P = .034) and a longer length of hospital stay at a median 9 days (IQR, 2–16) as compared with 5 days (0–12) in the influenza group (P = .002).

Mortality rates were similar between the RSV and influenza groups at day 30 (9.6% vs 9.7%, P = .973; Figure 2). No significant difference was confirmed in the multivariate analysis (Supplementary Table 3). Mortality was slightly higher in the RSV group than in the influenza group, with no significant difference on day 10 (7.2% vs 6.0%, P = .627), day 90 (16.0% vs 12.7%, P = .341), and day 365 (24.0% vs 21.2%, P = .512; Supplementary Figure 1 and Supplementary Table 2).

DISCUSSION

In our study of 7 consecutive epidemic seasons, RSV was associated with similar mortality but a higher risk of respiratory symptoms (dyspnea, oxygen desaturation), LRTI, consolidative pneumonia, hospitalization, ICU admission, and prolonged hospital stay as compared with influenza in a general population of elderly persons aged ≥75 years.

General symptoms (fever, myalgia, asthenia) were less frequent in the RSV group. These findings are in agreement with previous studies [19–22] and could explain the longer delay between symptom onset and first visit to the emergency department in patients infected with RSV over those infected with influenza, as observed in the current study and several previous studies [23–25].

Conversely, respiratory signs and symptoms (dyspnea, oxygen desaturation, wheezing) were more frequent in our RSV group than our influenza group. These results may be explained by the distal airway obstruction characteristic of the pathophysiology of RSV infection. They could account for the higher rate of LRTI (83.2% vs 61.5%, P < .001) and consolidative pneumonia observed in the RSV group (28.8% vs 17.2%; P = .004), as supported by the high rate of chest imaging in our study (555/558 patients), all of which was interpreted by a board-certified radiologist. Moreover, the initial laboratory tests showed a higher total white blood cell count and higher neutrophil rate among patients infected with RSV. These results, coupled with the higher rates of radiologic pneumonia found among the patients with RSV, might suggest that RSV infections are more frequently associated with bacterial coinfections. While this hypothesis cannot be confirmed by the present study because no systematic bacterial samples were taken, it aligns with the findings of Hedberg et al [26], who reported higher rates of bacterial coinfection at admission when comparing patients infected with RSV, influenza, and SARS-CoV-2. Ultimately, the suspicion of pulmonary superinfection and the need for oxygen may have influenced physicians’ management decisions and may explain the higher rates of hospitalization in patients infected with RSV (83.2% vs 70.0%, P = .003).

As with hospitalization rates, our study found higher ICU admission rates in the RSV group than the influenza group (7.2% vs 3.0%, P = .034). Although mortality rates were slightly higher in the RSV group, there were no statistical differences. These findings are consistent with multiple previous studies examining RSV severity [19, 20, 23, 27–29]. They align with recent research highlighting the burden of RSV that increases health care costs and leads to hospital overcrowding [12, 15–17], thereby emphasizing the need for the use of upcoming effective vaccines and treatments.

Regarding influenza vaccination, our study lacked sufficient data for robust conclusions, as information was available for only 198 of the 558 patients (36.8% of RSV and 35.1% of influenza). However, it is noteworthy that the rates of missing data and vaccinated patients were similar between the groups, thereby mitigating the potential impact of any bias. Furthermore, in contrast to previous studies comparing RSV and influenza [12, 15], our population appeared to be relatively undervaccinated against influenza (55.9% of vaccinated patients in the influenza-infected group), which may highlight the high pathogenicity of RSV.

The value of seasonal influenza vaccination is widely acknowledged [30], and a recent review and meta-analysis demonstrated a 29% reduction in the risk of hospitalization and an 18% reduction in all-cause mortality in the general population [31]. For RSV, the discussion of vaccination strategy is more recent and ongoing. Early trials on RSV-preventive vaccines appear promising, showing a reduction in LRTIs without significant adverse effects [2, 32], and these vaccines have received approval from the US Food and Drug Administration and the European Medicines Agency [30]. A theoretical study conducted in the Netherlands suggested that RSV vaccination would be cost-effective, particularly in high-risk populations aged ≥60 years, in terms of preventing general practitioner visits, antibiotic prescriptions, hospitalizations, and deaths [31]. However, a recent report from the UK Health Security Agency highlights the burden of RSV in adults, particularly in the population aged >75 years [33], and the Joint Committee on Vaccination and Immunisation advises a program for older adults aged ≥75 years [34].

In addition to vaccination, our study appears to underscore the importance of developing RSV-targeted treatments. Indeed, antibiotics were overprescribed in the flu- and RSV-infected groups. A majority of these patients had an LRTI or upper respiratory tract infection of viral origin, with no evidence of bacterial coinfection; of note, approximately 20% of patients who received antibiotics in the emergency department had no radiologic evidence of pneumonia. Overprescription of antibiotics could be explained by early initiation of antibiotic treatment before RT-PCR results were available or suspicion of bacterial superinfection, especially in patients who were highly febrile. This finding highlights the importance of rapid reporting of viral RT-PCR results and the development of rapid diagnostic tests that can reliably differentiate viral from bacterial infections to reduce inappropriate antibiotic use. In patients infected with RSV, the frequent use of antibiotics without clear justification could also be explained by the lack of approved antiviral treatments. This underscores the importance of pursuing the development of RSV-specific antiviral treatments, such as neuraminidase inhibitors and polymerase inhibitors for influenza, which should have the dual benefit of reducing not only symptom severity if initiated early [35] but also antibiotic misuse [4, 36].

Screening practices at our hospital remained consistent throughout the study period, before and after the COVID-19 pandemic. During the winter, all patients with respiratory and/or infectious symptoms who were considered at high risk for complications were tested for RSV, influenza A, and influenza B. While screening practices may vary among physicians, individuals aged >75 years are almost always considered at risk for complications, suggesting that the majority were tested. After the pandemic, the test was expanded to detect SARS-CoV-2 with the same sample.

The COVID-19 pandemic had a notable impact on the circulation of seasonal respiratory viruses, including RSV and influenza, globally and in France [37]. Some authors attributed the decline in positive test results to personal protection measures implemented during this period, such as hand hygiene campaigns, physical distancing, and the reduction of travel and circulation due to lockdown measures [38, 39].

The retrospective nature of our study is a nonnegligible bias, mainly because we were not able to obtain clinical presentation characteristics for all patients included. Respiratory frequency was missing in a high number of charts. Six patients had no laboratory tests, either because there were no clinical signs of severity or because the patient's prognosis was poor, so no blood tests were done. Among those whose blood samples were analyzed, C-reactive protein was not systematically measured in all patients (289 missing data), nor was differential white blood cell count in some patients (78 missing data). However, the influenza and RSV groups were identical on arrival at the emergency department, making comparative analyses reliable.

One other potential bias is the monocentric nature of our study. Our study population is probably not representative of the general geriatric population because the university hospital where the study was conducted houses several transplant centers and other national reference centers for rare diseases, suggesting that patients consulting this emergency department are probably more often carriers of serious comorbidities than the general population. Our study population might be at greater risk of complications, and so hospitalization and mortality rates might be overestimated.

Although most developed countries widely acknowledge the chronologic age of 65 years as the definition of “elderly” [40] and RSV vaccination is recommended from the age of 60 years in the United States [30], we opted to set the inclusion age threshold at 75 years for our study, motivated by several reasons. First, the RSV burden is mainly linked to patients aged ≥75 years, as shown by a previous study [16] and confirmed by a recent report [33]. Furthermore, our university hospital, similar to the majority of French medical facilities, imposes an age limit of 75 years for admission to geriatric service. In daily practice, patients aged >75 years are often considered at high risk of complications and are thus almost systematically tested for RSV and influenza when they present to our emergency department with respiratory or infectious symptoms. Including younger patients would have posed a significant risk of introducing selection bias, as many patients aged 60 to 75 years might not have undergone testing during their visits to the emergency department. But above all, the 75-year threshold was chosen to distinguish our work from prior research on RSV in the elderly, which typically set inclusion limits at 60 or 65 years. With a median age of 84.8 years and 77.4% of patients aged ≥80 years, our study constitutes the largest cohort of patients aged >75 years with RSV infection, documented in-depth to the best of our knowledge.

Some of the other strengths of this study are the large sample size, the case ascertainment method by RT-PCR, and the data collection by 2 physicians with a rigorous and standardized collection form. Another strength of our study is the long inclusion period. As RSV and influenza subtypes alternate from year to year, the inclusion period of 7 consecutive winters may have mitigated potential differences in severity between RSV subtypes in the elderly.

CONCLUSION

This study, conducted at Henri Mondor University Hospital over 7 consecutive winters, included the largest cohort of patients aged >75 years with RSV infection, documented in-depth thus far. Among patients aged >75 years presenting at the emergency department with respiratory or infectious symptoms, RSV shares a comparable mortality rate with influenza but is associated with higher rates of LRTI, pneumonia, hospitalization, ICU admissions, and extended hospital stays. These elements and the constant aging of the global population should encourage the development of effective molecules and vaccines against RSV.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Financial support. No financial support was received for this study.

References

Author notes