INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Respiratory syncytial virus (RSV) is an RNA virus in the Paramyxovirus family and has long been recognized as the most important cause of respiratory tract infection in infants (1). Recently, RSV has also been identified as a cause of serious illness in certain adult populations, including the elderly and those with underlying chronic cardiopulmonary disorders (2- 5). In a recent study of persons over age 65, RSV infection was associated with 10% of winter hospitalizations for cardiopulmonary disease or flulike illness compared with 13% due to influenza A and B (2). In another study of adults hospitalized with pneumonia, RSV was the third most common pathogen identified, only slightly less frequent than pneumococcus and influenza A (4). In both of these hospital-based studies, underlying chronic obstructive pulmonary disease (COPD) or congestive heart failure (CHF) was present in approximately half of the patients. The role of viral infection, including RSV, in exacerbations of chronic bronchitis has been noted in a number of studies, although previous studies were generally small and diagnostic tests for RSV were relatively insensitive (6). Importantly, none of these reports studied chronic cardiac patients or described in detail the clinical outcome of RSV infection.

In this report we describe the incidence and outcome of RSV infection in 134 community-dwelling adults with underlying COPD and/or CHF prospectively followed during two consecutive winters. The clinical impact of RSV infections is contrasted to other common respiratory viruses.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Adults with a diagnosis of New York State Heart Association Class III or IV CHF or chronic lung disease were eligible to participate in the study. COPD was defined as a physician's diagnosis of obstructive airway disease and a history of smoking. Patients with asthma were excluded. Subjects were recruited from the pulmonary and cardiac rehabilitation programs at Rochester General Hospital (Rochester, NY) or from the Department of Medicine cardiology practice. All subjects were required to sign informed consent according to the guidelines of the University of Rochester and the Rochester General Hospital Human Subjects review boards.

Enrollment and Study Period

Recruitment took place in the month prior to the study period, which was mid November to mid April in two consecutive winters, 1996- 1997 and 1997-1998. All subjects underwent a standard history and physical examination of the cardiorespiratory systems. Demographic data, living situation including frequency of contact with children, medications, functional status as measured by Katz Activities of Daily Living (ADL) and the Lawton Scale of instrumental activities of daily living (IADL) were recorded (15, 16). Arterial oxygen saturation (Sa_O2) was measured on room air unless the subject required constant supplemental oxygen. Blood samples for baseline viral serology were collected at enrollment. Cardiac ejection fraction, as determined by echocardiography, and/or percent predicted FEV₁ were recorded from medical records.

Surveillance for Respiratory Illnesses

Respiratory illnesses were monitored during the study period by a combination of active surveillance and passive reporting. Subjects were asked to contact study personnel if they had any symptoms consistent with a respiratory illness, such as nasal congestion, sore throat, new or increased cough or sputum production or dyspnea, and fever. Subjects were also to call if they experienced worsening of their baseline cardiopulmonary disease. Illness evaluation included a directed cardiopulmonary history and physical exam, Sa_O2 measurement, nasopharyngeal swab (NPS) for virus culture, and acute and convalescent blood (4 to 6 wk) for virus serology. Bacterial sputum cultures were not done as part of the study. Changes in medication requirements, functional status, and utilization of medical services were recorded until symptoms resolved. Participants who were unavailable for direct evaluation were asked to fill out an illness diary card until symptoms returned to baseline. The medical record of all hospitalized subjects was reviewed and data extracted. At the end of the study period, subjects were reevaluated and a final blood sample was obtained. The relationship between a virus infection and an illness was categorized as either definite, probable, possible, or asymptomatic. An illness was definitely attributed to a virus if the virus was isolated during the illness or if pre- and postillness serum showed a  4-fold virus-specific IgG response and the sera bracketed the illness closely (< 8 wk). An illness was probably attributable to a virus if the pre- and postsera were > 8 wk apart and there were no other illnesses in the time period. The relationship was considered possible if more than one illness occurred in the interval of pre- and postsera, or if more than one virus was identified by seroconversion in an interval with a single illness.

Virus Culture

NPS were collected, placed on ice for transport to the laboratory, and cultured on HEp-2, WI-38, and Rhesus monkey kidney (RhMK) cells using standard methods (5). Cell cultures were examined for cytopathic effect for 10 d and viral infection confirmed with virus-specific monoclonal antibodies (RSV, parainfluenza viruses, influenza A and B viruses, and adenovirus; Baxter Healthcare Corp., Issaquah, WA). RhMK cultures were tested for hemagglutination on Day 10. Acid lability was used to differentiate rhinoviruses from enteroviruses.

Viral Serology

Serologic evidence of RSV, influenza A and B, parainfluenza virus (PIV)-2 and PIV-1/3, coronaviruses 229E and OC43 infection was determined by comparing virus-specific IgG titers in pre- and postillness serum by enzyme immunoassay using published methods (5).

Statistics

Incidence rates and 95% confidence intervals (CI) were calculated for each virus. Proportions were compared using Fisher exact test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Sixty-seven subjects were recruited and followed during the winter of 1996-1997, one of whom died before the last blood draw in the spring. During the 1997-1998 winter season, 67 new subjects were recruited and 51 subjects reenrolled from the first year bringing the second season total to 118 subjects. During the second winter, six subjects died and four withdrew before the last blood draw. The demographic characteristics of subjects are shown in Table 1. The primary underlying disease was COPD or CHF in 66% and 34%, respectively, and 20% carried both diagnoses. Subjects ranged in age from 28 to 87, but were generally elderly, with a mean age of 69 yr. The severity of the underlying medical condition is reflected in a mean ejection fraction of 35% and percent predicted FEV₁ of 44% in those with CHF and COPD, respectively. The functional capacity of the subjects was limited; 25% used home oxygen and 30% had an IADL score of  2 which reflects a need for assistance with routine daily activities such as housework and shopping. Twenty-two percent of the subjects lived alone; only 11 persons (8%) lived with young children. However, 59% had contact with children at least twice per month. Vaccination rates for influenza virus and pneumococcal disease were high at 90% and 76%, respectively.

Respiratory Illnesses

During the first winter, 40 of the 67 subjects reported a total of 59 illnesses (Table 2). During the second winter season, 67 of the 118 subjects reported 96 illnesses. Three-quarters of the illnesses were evaluated by the study nurse and one-quarter were reported by an illness diary card. Overall, 42 virus infections were identified. A viral etiology was definitely or probably related to an illness in 36 of the 155 illnesses (23.2%) and possibly related in five others. In addition, one person had an asymptomatic seroconversion to coronavirus OC43. Infection with influenza A virus was identified in 13 subjects, RSV in eight, rhinovirus in seven, coronavirus OC43 in five, coronavirus 229E in four, influenza B virus in three, and PIV in two. The incidence of documented viral infection was significantly higher in the CHF group compared with COPD subjects (31 versus 16 infections per 100 person-winters; p = 0.02, Fisher exact test). Exposure to children was no more frequent among virus-infected subjects (53%) than in the entire group (59%).

RSV was the second most frequent virus infection and was identified in eight subjects (two culture-positive) (Table 2). The mean age of the RSV-infected subjects was similar to the population studied (67 versus 69 yr). The incidence of RSV during the two winters was 4.3 per 100 person-winters (95% CI = 1.4 to 7.2), which was not statistically different from the rate of influenza A (7.0; CI = 3.3 to 10.7) or rhinovirus infections (3.8; CI = 1.0 to 6.5). No subject had two RSV infections. Symptoms referable to both upper and lower respiratory tract were seen in the majority of subjects with definite or probable RSV-associated illnesses (Table 3). In contrast to influenza A infections, constitutional symptoms were less frequent (4/7 versus 10/11; p = 0.24 by Fisher exact test), although only fever was significantly less frequent (0/5 versus 10/11; p = 0.001 by Fisher exact test). Four of five subjects had a significant drop in oxygenation, defined as  3% fall in room air Sa_O2 or an increase in oxygen requirement of 2 L/min, a frequency similar to influenza A infection.

The clinical impact of RSV infection was significant (Table 4). The seven subjects in whom the virus was definitely or probably associated with an illness were symptomatic for a mean of 15 d with a range of 7 to 28 d. Five of seven subjects sought medical care with a total of five office visits and two hospitalizations. An additional subject had a seroconversion to RSV and coronavirus 229E seroconversion associated with an illness requiring hospitalization. This illness was clinically diagnosed as CHF. The two other hospitalized subjects had underlying COPD and were ill for 6 and 14 d prior to admission. Both had wheezing and significant falls in Sa_O2 despite negative chest radiographs and the clinical diagnosis was exacerbation of COPD. One subject required admission to the intensive care unit and had not returned to baseline functional status 3 mo after discharge. Notably, viral cultures during hospitalization were repeatedly negative, although RSV had been isolated 2 d before admission.

Influenza A was the sole virus definitely or probably associated with a respiratory illness in 12 cases (two culture-positive). In addition, a 59-year-old woman with severe COPD and CHF admitted to the hospital for respiratory failure requiring ventilatory support seroconverted to both influenza A and coronavirus OC43. The majority of the 12 subjects with illness solely associated with influenza A virus had nasal congestion, rhinorrhea, cough, and sputum production (Table 2). Fever  99° F was present in 90% of the illnesses and 82% had a significant decrease in oxygen saturation. The illnesses lasted a mean of 13 d, and 11 subjects utilized medical services (Table 4). There were six office visits, one emergency room visit, and five hospitalizations. Eight of the 12 (66%) subjects had received influenza vaccine, however during the second winter when the majority of cases occurred the antigenic match between the circulating virus and the vaccine was poor.

Seven subjects had rhinovirus recovered during a respiratory illness. Most had both upper and lower respiratory tract symptoms, and similar to RSV, fever was absent in all. Lower respiratory tract signs were slightly less common than influenza A infection. Two of the rhinovirus-infected subjects had a significant decrease in Sa_O2, however none was treated with steroids. In contrast to RSV- and influenza-infected subjects, none of the rhinovirus infections was associated with hospitalization.

Nine subjects had evidence of coronavirus infection, five with the OC43 strain and four with the 229E strain. Symptoms were generally indistinguishable from RSV or influenza A, although fever and lower respiratory tract signs were less frequent (data not shown). None of the coronavirus infections was related to hospitalizations, except for the two subjects in whom RSV and influenza A infection were also identified.

Overall, the frequency of physician office visits was similar for illnesses associated with or without a viral pathogen (42% versus 41%). Corticosteroid and antibiotic use was slightly more frequent in viral-associated illnesses (29% versus 21% and 66% and 57%, respectively), although neither difference was statistically significant. However, steroid use was significantly more frequent in influenza A- and RSV-associated illness than in illnesses with other viruses or those without an associated virus (10/19 versus 0/17 or 24/114; p = 0.008, Fisher exact test). Hospitalization was also more likely during an illness associated with a virus (nine hospitalizations in 38 illnesses) than during an illness without a defined pathogen (13 hospitalizations in 115 illnesses; p = 0.07, Fisher exact test). Furthermore, subjects with an influenza A or RSV infection were more likely to be hospitalized (nine of 21) than subjects infected with other viral pathogens (0 of 17) (p = 0.003, Fisher exact test). This difference remains significant even when the two subjects with two viruses identified during an illness are eliminated.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Each winter acute respiratory infection is associated with increased morbidity and mortality among the elderly and those with underlying cardiopulmonary diseases (17). Although much of this excess has been attributed to annual epidemics caused by influenza virus, other viral pathogens may also contribute significantly. Two recent epidemiologic studies from the United Kingdom suggested that the excess morbidity from RSV may be equivalent to that of nonpandemic influenza (18, 19). Furthermore, two studies of lower respiratory tract disease in hospitalized adults noted that RSV was nearly as frequent as influenza A infection and that many had underlying chronic cardiopulmonary conditions (2, 4).

During the two winter seasons we found that 23% of the respiratory illnesses in high-risk adults were associated with a viral pathogen, a finding consistent with the rate noted by Fagon and Chastre in a review of respiratory infections in persons with COPD (9). This is lower than the 43% reported in a recent analysis of respiratory illnesses in community-dwelling adults in England (20). Our lower rate may reflect that all respiratory complaints were investigated in subjects in whom respiratory problems are frequent, some of which may be related to environmental allergens, and dietary or medication changes. Although we were dependent upon self-reporting of illnesses, it is unlikely that many illnesses were missed because only one asymptomatic seroconversion was detected. Influenza A, RSV, rhinovirus, and coronaviruses were the most common pathogens identified using a combination of serology and virus culture. It is likely that the incidence of rhinovirus was underestimated because we only used viral culture for identification (20). However, the incidence of influenza A and RSV infection in our study should be accurate because serologic diagnosis is highly sensitive with the methods described (21).

RSV was associated with eight of the 155 illnesses (5.2%) over the two winters. Eight of the 134 subjects (6.0%) developed RSV infection during the study periods for a rate of 4.3 per 100 subject-winters. In published series the percentage of illnesses caused by RSV in community-dwelling persons with COPD ranges widely from 0.8 to 17.4% and generally is slightly lower than rates for influenza A virus (6, 12). Carilli and coworkers noted eight serologically confirmed RSV infections in 46 illnesses in 30 subjects for a rate of 17% of all illnesses (7). The high frequency of RSV in that study may have been due to the fact that subjects were only followed for a single season, the RSV attack rate may have been unusually high, and the majority of subjects were between ages 25 and 45, an age likely to have close contact with young children. In contrast, Smith and coworkers evaluated over 1,000 exacerbations of COPD over a 6-yr period in 150 subjects and found only eight RSV infections for a rate of 0.8% of all illnesses (8). Sommerville identified 41 RSV infections in adults with COPD in a retrospective serologic analysis of respiratory illnesses at a hospital in Glasgow, Scotland (12). However, none of these reports describe the clinical outcome of RSV infection in detail and there are no studies in patients with CHF, a group in whom respiratory infection is an important cause of hospitalization (22). Interestingly, the rate of viral infection in the cardiac patients in this study was nearly twice that of the COPD group. The COPD subjects were more frequently on home oxygen (15% versus 42%; p = 0.002) which likely restricted their mobility. Frail cardiac patients may be better able to venture into the community, thus increasing their exposure to viral pathogens.

We found that the clinical findings caused by influenza A and RSV in high-risk subjects with COPD or CHF were similar, although the number of infections was relatively small. In this group which is highly vaccinated for influenza, hospitalization rates were similar for subjects with identified influenza A infection (six of 13) and RSV (three of eight), suggesting that the clinical impact may be comparable in this high-risk population. When combined, nine of 21 subjects with influenza A- and RSV-associated illnesses were hospitalized in contrast to zero of 17 for other viruses, or to 14 of 115 for illnesses not associated with a virus pathogen.

Thus when compared with nonpandemic influenza A, a pathogen for which vaccination in high-risk groups has been demonstrated as beneficial, RSV appears to cause similar morbidity. Although the number of subjects studied is relatively small, the results of this and other published series, coupled with those from hospital-based studies suggest that prevention of RSV infection in persons with COPD or CHF would also be of significant benefit. At this time, several candidate RSV vaccines, including purified F protein, a chimeric FG subunit protein, and a live attenuated virus vaccine are available for evaluation in adult populations (23, 24). In preliminary findings, we noted that high levels of circulating neutralizing antibody to RSV were correlated with protection in the frail elderly, and it is reasonable to conclude that this same relationship will be true for persons with COPD and CHF (25). If so, immunization with an RSV vaccine may be beneficial in these high-risk persons.