Is Clinical Recognition of Respiratory Syncytial Virus Infection in Hospitalized Elderly and High-Risk Adults Possible?

Transcription

1 MAJOR ARTICLE Is Clinical Recognition of Respiratory Syncytial Virus Infection in Hospitalized Elderly and High-Risk Adults Possible? Edward E. Walsh, 1,3 Derick R. Peterson, 2 and Ann R. Falsey 1,3 Departments of 1 Medicine and 2 Biostatistics and Computational Biology, University of Rochester School of Medicine and Dentistry, and 3 Department of Medicine, Rochester General Hospital, Rochester, New York Background. The clinical and radiographic features of respiratory syncytial virus (RSV) infection in elderly hospitalized persons have not been described in detail, to our knowledge, despite its relative frequent occurrence. Methods. Clinical characteristics of 132 RSV infections were compared with 144 influenza A virus infections and with all non-rsv infections in elderly and high-risk persons admitted to the hospital with acute respiratory illness. Radiographic findings were categorized for RSV-infected persons. Results. RSV was more commonly associated with nasal congestion (68% vs. 55%; P p.03), wheezing by history (73% vs. 53%; P p.002) and on examination (82% vs. 68%; P p.02), and lower temperature ( P p.004) than influenza A virus. Compared with all non-rsv infected subjects, nasal congestion (odds ratio [OR], 2.0 [95% confidence interval {CI}, ]), wheezing on examination (OR, 1.8 [95% CI, ]), and temperature C (OR, 1.6 [95% CI, ]) were independent predictors of RSV infection, although their sensitivity and specificity were poor. New radiographic infiltrates were seen in approximately half of RSV-infected persons, and pneumonic opacities were typically small and unilateral. Conclusions. Although RSV causes a different clinical syndrome than influenza A virus, it is not readily differentiated on clinical grounds from influenza A nor from all non-rsv illnesses in elderly patients. Respiratory syncytial virus (RSV) and influenza A virus are 2 of the more common viral pathogens identified in elderly and high-risk persons hospitalized for acute respiratory symptoms during the winter months [1 3]. RSV infection in adults has been associated with a wide range of clinical syndromes, including upper respiratory illness, acute bronchitis, exacerbations of chronic obstructive lung disease (COPD), and pneumonia. Although several studies describe the clinical presentation of influenza virus infection in hospitalized older persons, only a few reports provide detailed information Received 17 August 2006; accepted 4 November 2006; electronically published 15 February Potential conflicts of interest: none reported. Presented in part: Respiratory Syncytial Virus 2005 Symposium, 16 September 2005, Oxford, United Kingdom. Financial support: National Institutes of Allergy and Infectious Diseases (grant AI 45969). Reprints or correspondence: Dr. Edward E. Walsh, Infectious Diseases Unit, Rochester General Hospital, 425 Portland Ave., Rochester, NY (edward.walsh@viahealth.org). The Journal of Infectious Diseases 2007; 195: by the Infectious Diseases Society of America. All rights reserved /2007/ $15.00 DOI: / on the clinical and radiographic manifestations of RSV in this population [1 8]. Because RSV and influenza epidemics frequently overlap temporally, and infection control measures differ for each, it would be useful if the clinical presentations of RSV infection and influenza could be distinguished from one another. In addition, differences between the 2 illnesses may provide insight into the pathophysiological features of RSV disease in elderly persons. Therefore, we compared the clinical manifestations of RSV infection to those of influenza A infection in a large cohort of hospitalized elderly and high-risk persons. We used a constellation of symptoms and signs to determine whether RSV disease could be differentiated from influenza A and from all non-rsv illnesses on clinical grounds. In addition, we describe the radiographic features of RSV infection in this population. PATIENTS AND METHODS Study population. Patients admitted to Rochester General Hospital with acute cardiopulmonary syndromes between 15 November and 15 April during JID 2007:195 (1 April) Walsh et al.

2 Table 1. Percentage of respiratory syncytial virus (RSV) infected and influenza A infected subjects with each clinical symptom and sign. Symptom or sign RSV (n p 118) Influenza A (n p 133) P Nasal congestion Sore throat NS Hoarseness NS Cough NS Dyspnea Wheezing by history Chest pain NS Feverishness Constitutional symptoms NS Duration of symptoms, mean SD, days Rhinorrhea 11 8 NS Wheezing on examination Rales NS Temperature 139 C Mean SD, C Respiratory rate, mean SD NS Pneumonia on chest radiograph NS consecutive winters from 1999 to 2003 were evaluated for RSV and influenza A infection as described elsewhere [1]. Persons 65 years of age or those 121 years of age with underlying cardiopulmonary disease were prospectively enrolled and evaluated. Patients with acute myocardial infarction, unstable angina, or documented pulmonary embolism were excluded. Demographic information, clinical parameters, and results of routine laboratory tests, including complete blood count, cultures of sputum and blood, and chest radiographs were also recorded. Only bacterial cultures collected within 48 h and chest radiographs taken within 72 h of admission were considered in the analysis. All subjects or their legal guardians provided written consent. The study was approved by the Rochester General Hospital Clinical Investigation Committee and by the University of Rochester Research Subjects Review Board. Viral diagnostic assays. Nasopharyngeal (NP) swab samples were collected from all subjects within 48 h of admission and transported to the laboratory on wet ice immediately. Samples were placed on HEp-2 and MDCK cells for virus isolation within 4 h and reviewed daily for 10 days for evidence of characteristic cytopathic effect. RNA was extracted from NP samples and subjected to reverse-transcription polymerase chain reaction (RT-PCR) for RSV and influenza A in accordance with previously published methods [5, 9]. Acute serum collected on admission and convalescent serum samples collected 4 6 weeks later were used to determine the titer of serum IgG to RSV and influenza A antigens by use of an established enzyme immunoassay [5, 10]. Subjects were considered infected with either RSV or influenza in they had a virus isolated, a positive result of RT-PCR, or a 4-fold increase in virus-specific serum IgG. Statistical analysis. Means of continuous variables were compared by group using the t test, whereas Pearson s x 2 test was used to compare proportions of dichotomous variables. For multivariate analysis, logistic regression was used to calculate odds ratios (ORs) with 95% confidence intervals (CIs; JMP; version 5.1.2; SAS Institute). In addition, a prediction model for RSV was developed using logistic regression on data from 3 of the 4 seasons. The model was then validated using the data from the fourth season. Specifically, for this last statistical analysis, all-subset logistic regression was used to develop multivariate classifiers for binary outcomes by dichotomizing the predicted probability, based on a linear combination of risk factors, at 50%. The optimal subset size was selected by minimizing the estimated misclassification rate based on 4-fold leave-1-season-out cross-validation. RESULTS Comparison of signs and symptoms in RSV and influenza A infections. A total of 1471 respiratory illnesses were evaluated in 1388 subjects admitted to the hospital with acute respiratory symptoms during the 4 winter seasons. From this group, 142 RSV and 154 influenza A infections were documented by culture, RT-PCR, serological analysis, or a combination of these tests. As expected, the occurrence of each virus overlapped temporally each winter. For the analysis of clinical signs and symptoms, the 10 subjects with evidence of dual virus infection were excluded, as were 14 RSV-infected and 11 influenza-infected subjects with evidence of concomitant bacterial infection. The latter were determined to have bacterial infection on the Table 2. Sensitivity and specificity of individual and combined signs and symptoms in respiratory syncytial virus (RSV) infected hospitalized subjects, compared with non-rsv infected subjects. Clinical characteristic Sensitivity Specificity Nasal congestion a Sputum production Wheezing by history Feverishness Rhinorrhea on examination Wheezing on examination a Temperature C a Combined b NOTE. Data are percentages of subjects with characteristic. a Identified as a statistically significant independent variable by multivariate analysis b Subjects with the presence of each of the independently significant variables (which are denoted by footnote a). Clinical Features of RSV in Elderly Persons JID 2007:195 (1 April) 1047

3 basis of a positive blood culture or a culture and Gram stain of a respiratory pathogen from an adequate sputum sample. Thus, the clinical manifestations of 118 RSV and 133 influenza A infections were available for comparative analysis. The demographic characteristics of the 2 virus-infected groups have previously been described in detail and were similar [1]. Both groups were primarily elderly (mean age, 76 years) and female ( 60%), with 49% 54% having underlying cardiac disease and 55% 58% having chronic lung disease. A comparison of symptoms and signs in the RSV-infected and influenza A infected subjects demonstrates substantial overlap in illness characteristics, although several significant differences were identified (table 1). RSV-infected persons were more likely to have nasal congestion (68% vs. 55%; P p.03), to have dyspnea (95% vs. 86%; P p.02), and to complain of wheezing (73% vs. 53%; P p.002). Although cough was nearly universal in both groups, RSV-infected persons were slightly more likely to produce sputum (78% vs. 70%). A sensation of fever was more common among influenza-infected persons (66% vs. 53%; P p.07), as was chest pain and constitutional symptoms, although none reached statistical significance. Despite the frequency of nasal congestion in the RSV group, observable rhinorrhea was uncommon. Consistent with the reported symptoms, audible wheezing was detected in a greater proportion of the RSV-infected persons (82% vs. 68%; P p.02), and the mean admission temperature was higher in influenza infection (38.1 C vs C; P p.004). Temperature 39 C was found in 23% of influenza infections compared with 13% of RSV-infected subjects ( P p.05). On average, the duration of symptoms before hospital admission was 1.3 days longer for RSV infection than for influenza infection, although this difference did not reach statistical significance. The percentage with documented pneumonia on chest radiographs was not significantly different between viruses. We also attempted to identify those signs and symptoms that might differentiate RSV from non-rsv illnesses in a population with a 9% incidence of RSV. Although nasal congestion, sputum production, wheezing by history or on examination, feverishness, and temperature C were all statistically more frequent in RSV-infected persons, none proved sufficiently sensitive and specific to accurately discriminate RSV from non-rsv illnesses (table 2). For example, the presence of nasal congestion increased the likelihood of RSV from the baseline incidence of 9% to 13% but was only 63% sensitive and 55% specific for RSV. Similarly, audible wheezing was 75% sensitive but was only 32% specific for RSV, whereas observable rhinorrhea was relatively specific (93%) but had very poor sensitivity (11%). We also determined whether a combination of signs and symptoms would be useful with logistic regression. Nasal congestion (OR, 2.0 [95% CI, ]), wheezing on examination (OR, 1.8 [95% CI, ]), and temperature Figure 1. Examples of radiographs of respiratory syncytial virus (RSV) infected subjects. A, Left lower lobe and lingular infiltrates in an RSV culture positive subject who was ill for 3 days with an oxygen saturation level (SaO 2 ) of 92%. B, Computerized tomography of RSV culture positive subject with a 3-day illness, a temperature of 38.5 C, and an SaO 2 of 88%. Performed to rule out a pulmonary embolus, the tomography shows a lacy infiltrate in the right upper lobe. The corresponding plain film showed a vague density in the same area (C) JID 2007:195 (1 April) Walsh et al.

4 Table 3. Demographic and clinical characteristics according to radiographic pattern. Characteristic Radiographic reading Pneumonia (n p 27) Normal or no acute disease (n p 54) Nasal congestion 17 (63) 37 (69) NS Sputum production 17 (63) 43 (80) NS Wheezing by history 10 (37) 35 (65).03 Constitutional symptoms 16 (59) 21 (39).10 Feverishness 19 (70) 26 (48).06 Wheezing on examination 12 (44) 45 (83).0006 Rales on examination 19 (70) 26 (48).06 Temperature, mean SD, C Inhaled corticosteroids 3 (11) 19 (35).03 Bronchodilators 6 (22) 33 (61).001 Cardiac disease 8 (30) 30 (56).03 Chronic lung disease 10 (37) 37 (69).009 Diabetes 5 (19) 12 (22) NS IADL score, a mean SD NOTE. Data are no. (%) of subjects with characteristic, unless otherwise indicated. IADL, independent activities of daily living. a Range, 0 12 (with 0 completely independent and 12 completely dependent). P C (OR, 1.6 [95% CI, ]) were independent predictors of RSV infection. However, the combined presence of all 3 clinical parameters improved sensitivity only to 13% with specificity at 91%. Chest radiographs in RSV infection. Admission chest radiographs were available for all but 2 of the RSV-infected subjects, including those with bacterial coinfection. When the 118 pure RSV infections were considered separately, the official radiologist s report described 42% of the radiographs as either normal or showing no acute disease. The latter description was often used in the setting of stable chronic pulmonary abnormalities such as COPD, pleural or parenchyma scarring, or fibrosis present on previous radiographs. Twenty-seven (20%) of the radiographs had new opacities consistent with pneumonia, and 8% were attributed to atelectasis (figure 1). An additional 13% had infiltrates compatible with either atelectasis or pneumonia. Edema due to congestive heart failure was noted on 9% of the radiographs, and 1 subject had diffuse edema with normal heart size. The remaining radiographs had other new findings such as small pleural effusions. Opacities read as pneumonia were generally unilateral and basilar in location, although all lobes were represented. They were remarkable for their relatively small size and were rarely dense, with the exception of a few subjects with concomitant bacterial infection. Only 1 had a diffuse interstitial pattern characteristic of classic viral pneumonia. Small ipsilateral pleural effusions were occasionally noted when either pneumonia or atelectasis was present. The demographic and clinical characteristics varied according to the radiographic patterns (table 3). Subjects with culturediagnosed RSV were less likely to have pneumonia on chest radiograph (14%) than were those diagnosed by serological analysis only (27%; P!.0001). When compared with those with normal or no acute disease on chest radiograph, those with radiographic pneumonia were less likely to have symptomatic wheezing (37% vs. 65%; P p.03) or objective wheezing (44% vs. 83%; P p.0006). However, patients with pneumonia on radiographs had higher temperature (38.4 vs C; P p.006) and more often had rales on examination. Subjects with pneumonia less frequently used chronic bronchodilators (22% vs. 61%; P p.001) or inhaled corticosteroids (11% vs 35%; P p.03) and had less underlying chronic cardiac (30% vs. 56%; P p.03) and lung disease (37% vs. 69%; P p.009). Five (36%) of the 14 subjects with evidence of concomitant bacterial infection had radiographs that differed substantially from those with RSV infection alone in that the infiltrates were described as large, with consolidation and air bronchograms in 2 subjects and multiple pulmonary lobes involved in 4. Notably, 3 of the 4 subjects with Streptococcus pneumoniae bacteremia had extensive pulmonary infiltrates. The exception had presumed acute sinusitis with meningitis and a normal chest radiograph. None of the patients with RSV infection alone were noted to have consolidation or large infiltrates. Figure 1 shows several examples of radiographs from RSV-positive subjects. DISCUSSION Despite its reputation as a pediatric illness, RSV has recently also been established as an important cause of respiratory tract infection in adults of all ages and can be particularly severe in older adults and those with underlying high-risk cardiopulmonary conditions. In the epidemiologic study from which these subjects were drawn, we found that RSV accounted for 9.6% and influenza A for 10.5% of winter hospitalizations for acute respiratory symptoms [1]. In the present analysis, we noted several significant differences in the clinical manifestations of RSV infection when compared with influenza A infection. Notwithstanding substantial clinical overlap, the 2 viruses appear to cause different clinical syndromes, even in a population of elderly persons with a high incidence of underlying cardiopulmonary disease. Most notably, RSV-infected persons were more likely to have lower temperature and more nasal congestion and wheezing than influenza-infected subjects. This observation is similar to that noted by Hall et al. in RSVand influenza-infected young adults and is consistent with the wheezing that is so characteristic of RSV infection in infants [11, 12]. It is noteworthy that RSV is associated with a greater frequency of bronchospasm than influenza A infection even in a population with a high incidence of COPD. Dowell et al. also observed a greater association between wheezing and RSV than for influenza and non-rsv illnesses in hospitalized adults with Clinical Features of RSV in Elderly Persons JID 2007:195 (1 April) 1049

5 pneumonia [3]. Various mechanisms may be involved in the development of wheezing. It is possible that RSV infection may spread to the lower airways because of specific tropism for small airway epithelial cells, with resultant edema and inflammatory changes causing airflow obstruction. Additionally, nasal or lower airway replication of RSV may induce a dominant Th2-type cellular immune response, with production of RSV-specific IgE and leukotriene secretion, resulting in bronchospasm [13, 14]. Despite the apparent differences in the clinical presentation of RSV and influenza A that were observed in this study and by others, recognition of individuals with RSV infection among all non-rsv infected patients is nearly impossible. We were unable to define any constellation of signs and symptoms that could discriminate RSV- from the non-rsv infected patients with sufficiently high sensitivity and specificity to be clinically reliable. Even when considering a combination of a history of nasal congestion, the presence of wheezing on examination, and temperature C, the sensitivity was only 13% and specificity was 91%. In contrast, we previously found that clinical symptoms in acute influenza A infection in hospitalized patients could be reasonably well distinguished from non influenza-infected patients if certain clinical parameters were applied during the relatively brief influenza epidemic period [5]. Because the epidemic period for RSV is considerably longer than that of influenza, use of similar time restrictions cannot be applied. Thus, the diagnosis of RSV in elderly adults must rely on the results of a highly sensitive and specific diagnostic laboratory test. Unfortunately, the only widely available tests (culture and antigen detection), which provide excellent results in pediatric patients who shed large quantities of virus, are insensitive for diagnosis in adults [15]. For older hospitalized adults, RT-PCR provides the only reliable diagnostic method for acute infection but is generally not available and is expensive [9]. We found that 20% of RSV-infected persons had radiographically defined pneumonia, a rate midway between the extremes of 5% and 55% that were reported by others [2, 3, 6, 10, 16]. Although this finding may reflect differences in populations studied and possibly virulence of circulating viral strains, the most likely explanation is that some investigators selected patients on the basis of a clinical diagnosis of pneumonia, which we did not. Of interest, the radiographic features of RSV in elderly and high-risk adults were relatively unimpressive despite the fact that patients were quite hypoxic and ill enough to require hospitalization. RSV-associated opacities were generally small, focal, and unilateral in the lower lobes and in marked contrast to larger, dense infiltrates seen in bacteremic pneumonias. Although the present study was not designed to accurately determine the rate of bacterial coinfection with RSV or to discriminate viral from bacterial pneumonia, it was notable that there was a significantly lower rate of radiographically diagnosed pneumonia in patients who were RSV culture positive compared with those who were RSV seropositive only. One interpretation of this finding is that patients presenting to the hospital late in their course after the virus is no longer detectable in nasal secretions by culture or RT-PCR may have developed bacterial pneumonia. However, this group represents the minority of patients hospitalized with RSV. In conclusion, although RSV infection produces a different clinical syndrome than influenza A infection in elderly persons that is best characterized by nasal congestion and wheezing, the ability to distinguish RSV infection in individual patients is not readily feasible, even when using a constellation of clinical signs and symptoms. Recognition of RSV infection in this population requires an accurate diagnostic laboratory test, such as RT-PCR. In the absence of proven effective antiviral therapy for RSV in adults, the major practical implication of diagnosing RSV infection in this population may be that antibacterial agents could be withheld or discontinued, as has been done in pediatric populations [17]. Although the present study was not specifically designed to identify bacterial coinfection, only 10% of the RSV-infected subjects had proven concomitant bacterial infection. Reference 1. Falsey AR, Hennessey PA, Formica MA, Cox C, Walsh EE. Respiratory syncytial virus infection in elderly and high-risk adults. N Engl J Med 2005; 352: Falsey AR, Cunningham CK, Barker WH, et al. Respiratory syncytial virus and influenza A infections in the hospitalized elderly. J Infect Dis 1995; 172: Dowell SF, Anderson LJ, Gary HE Jr, et al. Respiratory syncytial virus is an important cause of community-acquired lower respiratory infection among hospitalized adults. J Infect Dis 1996; 174: Mathur U, Bentley DW, Hall CB. Concurrent respiratory syncytial virus and influenza A infections in the institutionalized elderly and chronically ill. Ann Intern Med 1980; 93: Walsh EE, Cox C, Falsey AR. Clinical features of influenza A virus infection in elderly hospitalized persons. J Am Geriatr Soc 2002; 50: Fransen H, Sterner G, Forsgren M, et al. Acute lower respiratory illness in elderly patients with respiratory syncytial virus infection. Acta Med Scand 1967; 182: Fransen H, Heigl Z, Wolontis S, Forsgren M, Svedmyr A. Infections with viruses in patients hospitalized with acute respiratory illness, Stockholm Scand J Infect Dis 1969; 1: Wald TG, Miller BA, Shult P, Drinka P, Langer L, Gravenstein S. Can respiratory syncytial virus and influenza A be distinguished clinically in institutionalized older persons? J Am Geriatr Soc 1995; 43: Falsey AR, Formica MA, Walsh EE. Diagnosis of respiratory syncytial virus infection: comparison of reverse transcription-pcr to viral culture and serology in adults with respiratory illness. J Clin Microbiol 2002; 40: Falsey AR, Treanor JJ, Betts RF, Walsh EE. Viral respiratory infections 1050 JID 2007:195 (1 April) Walsh et al.

6 in the institutionalized elderly: clinical and epidemiologic findings. J Am Geriatr Soc 1992; 40: Hall CB, Long CE, Schnabel KC. Respiratory syncytial virus infections in previously healthy working adults. Clin Infect Dis 2001; 33: Hall CB, McCarthy CA. Respiratory syncytial virus. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and practices of infectious disease. 6th ed. Philadelphia: Elsevier Churchill Livingstone, 2004: Openshaw PJM. Immunity and immunopathology to respiratory syncytial virus. Am J Respir Crit Care Med 1995; 152:S Garofalo RP, Patti J, Hintz K, Hill V, Ogra P, Welliver R. Macrophage inflammatory protein 1a (not T helper type 2 cytokines) is associated with severe forms of respiratory syncytial virus bronchiolitis. J Infect Dis 2001; 184: Casiano-Colon AE, Hulbert BB, Mayer TK, Walsh EE, Falsey AR. Lack of sensitivity of rapid antigen tests for the diagnosis of respiratory syncytial infection in adults. J Clin Virol 2003; 28: Sorvillo FJ, Huie SF, Strassburg MA, Butsumyo A, Shandera WX, Fannin SL. An outbreak of respiratory syncytial virus pneumonia in a nursing home for the elderly. J Infect 1984; 9: Adcock PM, Stout GG, Hauck MA, Marchall GS. Effect of rapid viral diagnosis on the management of children hospitalized with lower respiratory tract infection. Pediatr Infect Dis J 1997; 16: Clinical Features of RSV in Elderly Persons JID 2007:195 (1 April) 1051