A Rhinovirus Outbreak among Residents of a Long-Term Care Facility

Methods

Study Sample

The study sample resided in a 685-bed, long-term care facility for wartime veterans and their spouses. The facility is in rural Wisconsin and includes four residential buildings connected by underground tunnels. The four buildings have a total of 13 residential floors, each housing 50 to 60 residents. Approximately one half of the beds are allocated for intermediate care and one half for skilled care. A 2-story recreational building allows residents from different buildings to congregate and is connected to the residential buildings by underground tunnels.

Surveillance

During the winter season, active surveillance for respiratory illness is done at the facility as part of an ongoing influenza-prevention study. During the rest of the year, including the period of this study, the nursing staff conducts passive surveillance for respiratory illness using a daily illness reporting system. This system identifies and tracks any resident with fever (oral temperature more than equals 100.0 °F) or other symptoms indicating possible viral infection; ill persons are identified by self-report or by the nursing staff during routine encounters. During the third week of August 1993, an increase in the number of respiratory illnesses was noted among persons residing in a building (building C) in which persons with pulmonary disease were housed. In response, the nursing staff were asked to report all new respiratory illnesses to the nurses participating in the influenza study. These nurses recorded the date of illness onset and obtained respiratory viral cultures.

Virus Culture

After verbal consent was obtained, viral cultures were taken from ill persons but not from well persons. Cultures were obtained by swabbing the nasopharynx (using rayon-tipped flexible aluminum swabs) and throat (using dacron-tipped plastic swabs) and placing the swabs into a virus-transport medium (veal broth with gentamicin, penicillin, streptomycin, and amphotericin). Specimens were cooled to 4 °C within 1 hour of sampling and were inoculated within 24 hours into primary rhesus monkey kidney, HEp-2, and Wi-38 tube-cell cultures and Madin Darby canine kidney 24-well plate cultures. This panel of cell cultures enables detection of all common cultivable respiratory viruses. Tube-cell cultures were incubated at 33 °C on a roller apparatus, and 24-well plate cultures were incubated at 35 °C. Cells were examined for cytopathic effect and hemadsorption. Rhinovirus was confirmed by growth with characteristic cytopathic effect in Wi-38 cells and by acid lability. Negative cultures were held for 2 weeks before they were discarded.

Epidemiology

The outbreak was defined as the interval between the onset of symptoms in the first and last persons to yield rhinovirus isolates (14 August to 2 September 1993). The daily illness reports for the entire facility were reviewed for the period during and surrounding the outbreak to identify all new respiratory illness. A respiratory illness was defined as the presence of at least one new respiratory symptom not accounted for by another disease process; the presence of fever was not required. A rhinovirus infection was defined as a respiratory illness associated with a culture isolate of rhinovirus. The incidence of respiratory illness was calculated for each residential building and floor. The room locations of persons with rhinovirus infection were determined so that the geography of the outbreak and the risk for developing illness among close contacts could be studied.

Rhinovirus was isolated from one person during three separate symptomatic episodes; the first episode occurred during the outbreak and the other two in the ensuing 3 months. Information from the first illness only is included in the discussion of the outbreak. Information from all three illnesses is included in the clinical results.

Medical Record Review

After the outbreak, the medical records of persons from whom rhinovirus was isolated were reviewed. Age, sex, level of nursing care, duration of residence in the facility, underlying medical conditions, scheduled medications, and pulmonary examinations done by the residents' physicians composed the “pre-illness” database. Underlying medical conditions were obtained from a clinical problem list generated by each resident's physician. Illness data were abstracted from the medical records and recorded on a standard form; these data included the type and duration of symptoms, concurrent pulmonary examinations, oral temperatures, physician telephone calls and visits, results of chest radiographs, laboratory and sputum test results, and the use of oxygen, medication for symptom relief, and antibiotics. Symptoms were determined to be illness-related if they were new or worse during the illness; symptoms included upper respiratory (nasal or sinus congestion, coryza, sore throat, or cough), lower respiratory (productive cough, hoarseness, or dyspnea), gastrointestinal (anorexia, nausea, vomiting, or diarrhea), and systemic symptoms (malaise, fatigue, myalgia, sweating, or chilling). The duration of illness was defined as the interval between the first recorded symptom and either the time that residents returned to their usual activities or the time that the medical staff determined that the illness had resolved, whichever came first. Lungs were examined at the discretion of the medical personnel caring for the residents; examinations were done in 29 illnesses (by both the physician and staff nurses in 9 illnesses, by the physician only in 4 illnesses, and by the nurses only in 16 illnesses). Lung examination findings were attributed to rhinovirus if they were not present on any of the three most recent pre-illness pulmonary examinations done by the resident's physician.

Data Analysis

Categorical data were analyzed using the chi-square test or the Fisher exact test. Continuous data were analyzed using the Wilcoxon rank-sum test. The significance level was P < 0.05 (two-tailed) for all tests.

Results

The Outbreak: Number of Illnesses and Rhinovirus Isolates

Figure 1 shows the outbreak, which includes the period between the onset of symptoms in the first and last persons whose cultures yielded rhinovirus (14 August to 2 September 1993). During the outbreak, each of 128 residents developed a new respiratory illness. Sixty-seven persons (52%) were cultured; 33 of the cultures (49%) yielded rhinovirus. No other respiratory virus was isolated from those persons cultured. The isolation of rhinovirus coincided with an increase in the number of respiratory illnesses facility-wide. After the outbreak, the number of respiratory illnesses decreased, and no further viruses were isolated.

The outbreak predominantly affected two of the four residential buildings (buildings C and D; see Table 1). The incidence of respiratory illness was significantly higher among persons residing in buildings C (32%) and D (29%) than in those residing in buildings A (9%) and B (9%; P < 0.001). Additionally, 32 of the 33 rhinovirus isolates were cultured from persons residing in buildings C and D.

Distribution of the Outbreak in Buildings C and D

Within the six residential floors of buildings C and D, the outbreak tended to localize to one floor in each building. In building C, the incidence of respiratory illness was 39% among those residing on one of two residential floors; 8 of the 10 rhinovirus isolates from residents of this building were from persons on this floor (Table 1). Similarly, in building D, the incidence of respiratory illness was 47% among those residing on one of four residential floors; 11 of the 22 rhinovirus isolates for the building were from persons on this floor (Table 1). For each of these two high-incidence floors, rhinovirus was isolated from 73% of the residents who were cultured.

Within four of the six residential floors of buildings C and D, the room locations of most persons with culture-confirmed illness were clustered. The remaining two floors (one in each building) each housed two persons with culture-confirmed rhinovirus infections; the rooms of these persons were distant from each other.

Among persons residing in buildings C and D, the risk for developing either a respiratory illness or a culture-confirmed rhinovirus infection was no higher for persons residing in double than in single rooms (data not shown). Additionally, sharing a room or bathroom with a person with culture-confirmed infection did not increase the risk for becoming ill or for developing rhinovirus infection (data not shown). The one documented instance of room-mate co-infection occurred in a husband and wife who had onset of symptoms and isolation of rhinovirus within 4 days of each other.

Demographic Information and Underlying Conditions

Rhinovirus infection was confirmed 35 times in 33 persons; 1 person had three separate culture-proven infections in a 3-month period. The age (mean, 73 years), sex (male, 76%), level of nursing care required (skilled nursing care, 52%), and duration of residence (mean, 4.3 years) of residents with culture-confirmed rhinovirus infection were similar to those of all other persons in the facility. Among those with rhinovirus infection, the most prevalent underlying conditions were heart disease (67%) and chronic obstructive pulmonary disease (52%). The underlying conditions among the rhinovirus-infected residents were similar to those of all other persons in the facility except that the infected residents had a higher incidence of chronic pulmonary disease (52% compared with 36%; P = 0.07) and lower incidences of dementia (9% compared with 31%; P = 0.007) and peripheral vascular disease (12% compared with 26%; P = 0.07).

Clinical Illnesses

Illnesses associated with rhinovirus infection varied in severity. The mildest infections consisted of brief episodes of upper respiratory symptoms. In contrast, several persons had more severe illness and developed overt lower respiratory tract involvement. All persons had one or more upper respiratory symptoms, most frequently coryza and cough. Dyspnea or hoarseness was present in 37% and productive cough in 54% of illnesses. Two thirds of persons had systemic and one third had gastrointestinal symptoms (Table 2).

Rhinovirus infection generally caused little change in body temperature. The mean maximum oral temperature was 99.3 °F (range, 97.6 °F to 101.0 °F). Only six persons had oral temperatures of 100.0 °F or more. The mean duration of illness was 6.1 days (range, 2 to 13 days).

New abnormalities on lung auscultation were identified in 52% of rhinovirus infections and included wheezing in 34%, rhonchi in 17%, and rales in 28% of infections. When only the physician's pre-illness and illness examinations were compared (to control for the effect of different examiners), the findings were similar; new abnormalities were noted in 54% of illnesses, including wheezing in 23%, rhonchi in 23%, and rales in 15% of infections. The frequency of new lung findings was also similar whether or not persons had underlying chronic obstructive pulmonary disease (50% and 55%, respectively). In particular, the onset of wheezing was noted in 33% of illnesses in persons with lung disease and 36% of illnesses in persons without lung disease.

Medication for symptom relief was taken an average of 7.6 times per illness (range, 0 to 26 times). In more than one half of illnesses, medication for symptom relief was taken for an average of 9.8 additional days (range, 1 to 24 days) after the acute illness had ended. In most instances, this was because of persistent cough, but various other lingering symptoms were also reported. Chest radiographs were done during 11 illnesses (31%), and antibiotics were prescribed for 21 illnesses (60%). Antibiotics were directed against isolated respiratory pathogens (Haemophilus influenzae) in only 3 instances; in the remaining 18, antibiotics were given empirically.

Comparison of Illness in Persons with and without Lung Disease

Although several persons with chronic obstructive pulmonary disease had mild illness, rhinovirus infection was generally more severe among those with underlying lung disease. Persons with lung disease were more likely than those without lung disease to report dyspnea (42% compared with 0%; P = 0.004), and they had a longer duration of illness (mean, 7.4 days [range, 2 to 13 days] compared with 4.6 days [range, 2 to 8 days]; P = 0.007). The duration of cough, which often continued after the acute illness had ended, was also longer in persons with lung disease (mean, 12.1 days; range, 1 to 29 days) than in persons without lung disease (mean, 5.9 days [range, 0 to 18 days]; P = 0.03). Among persons with lung disease, approximately 30% coughed for 3 weeks and 10% coughed for 1 month after rhinovirus infection. Although the frequency of symptomatic medication use was the same among persons with and without lung disease, persons with lung disease were more likely to be treated with antibiotics (74% compared with 44%; P = 0.07).

Clinically overt bronchospasm developed in 5 (29%) of the 17 persons with rhinovirus infection who had underlying chronic obstructive pulmonary disease. Among those with bronchospasm, 3 persons required oral or parenteral glucocorticoid therapy and 2 additional persons required the new use of inhaled or nebulized bronchodilators. Two of the 5 persons with bronchospasm required transfer out of the facility. In one instance, rhinovirus infection was associated with progressive wheezing and respiratory distress that necessitated transfer to the emergency department. In another instance, a resident with severe lung disease developed wheezing and respiratory distress that progressed over the ensuing 72 hours despite maximal therapy in the nursing home. This person required emergent mechanical ventilation and died of respiratory failure 8 days after intubation; rhinovirus was isolated from a culture taken 2 days after illness onset. Finally, 1 person with chronic obstructive lung disease had a persistent, productive cough after rhinovirus infection and had a new lower-lobe infiltrate on a chest radiograph taken 15 days after isolation of rhinovirus.

Discussion

In this report, we describe a large rhinovirus outbreak that occurred among elderly persons residing in a long-term care facility. The outbreak, as measured by the incidence of respiratory illness and the number of rhinovirus isolates, involved two of four residential buildings and was further clustered within these buildings. The illnesses varied in severity but were accompanied by clinically overt lower respiratory tract involvement in several persons with underlying lung disease. This is the first large-scale rhinovirus outbreak reported among nursing home residents, and it provides a unique opportunity for epidemiologic and clinical observations.

Few data describe the epidemiology of rhinovirus in long-term care facilities. Studies in which prospective surveillance for respiratory illness is done in nursing homes have isolated rhinovirus from a small proportion of ill persons, which suggests that this virus does not contribute substantially to respiratory morbidity among nursing home residents [12, 13]. In these studies, rhinovirus was recovered throughout the study periods and in conjunction with various other respiratory pathogens. In contrast, we show that rhinovirus may be associated with large, well-defined outbreaks of respiratory illness in nursing homes.

The clustering of rhinovirus-infected residents by floor and within floors suggests that rhinovirus was efficiently transmitted within portions of the facility. However, the lack of a higher incidence of illness among the roommates of infected residents was not surprising given previous data on rhinovirus transmission. Dick and coworkers [2, 4] showed that, among family members residing in student apartment buildings, patterns of rhinovirus spread were erratic. Although high attack rates were achieved in certain buildings, susceptible family members often remained uninfected. Subsequent studies of experimental rhinovirus infections in married couples [15] showed that only 38% of inoculated persons could transmit rhinovirus to their spouses and that successful rhinovirus transmitters had certain characteristics, including high nasopharyngeal virus titers, the presence of virus on the hand and anterior nares, and the presence of moderate illness. Furthermore, prolonged time spent in the same airspace was necessary for transmission. The conditions for successful transfer were often met for only about 2 days early in the course of infection, which may explain the erratic transmission patterns seen in epidemiologic studies.

We found that rhinovirus infection appeared to be capable of producing severe illness, particularly among persons with underlying lung disease. Even among persons without preexisting lung disease, new findings on lung auscultation and productive cough were noted in one half of infected persons. These findings contrast with the limited existing data that indicate that severe illness and lower respiratory tract involvement are uncommon among nursing home residents infected with rhinovirus [12, 13]. When compared with the 14 culture-documented cases described by Falsey and coworkers [13], persons in our series had a higher prevalence of systemic symptoms (71% compared with 43%), gastrointestinal symptoms (34% compared with not reported), productive cough (54% compared with 21%), wheezing (34% compared with 14%), antibiotic use (60% compared with 14%), and pneumonia and death (1 each compared with none). Different methods of surveillance and illness reporting may account for these findings. We used the existing passive surveillance system in effect at this facility. In contrast, Falsey and coworkers did active surveillance as part of a respiratory virus study, which may have facilitated the detection of milder illnesses. In our study, only two thirds of illnesses in buildings C and D were reported to the study nurses for culture. Those with more symptomatic illness may have been more likely to be reported, thus biasing our results towards the inclusion of residents with more severe illness.

Several limitations of our study deserve comment. Among the 128 persons in whom new respiratory illnesses were identified during the outbreak, 67 were cultured; 33 rhinovirus isolates were recovered. Could the 61 uncultured and 34 culture-negative illnesses have been caused by other viruses? This is unlikely for two reasons. First, specimen acquisition and culture techniques were done under optimal conditions, yet no other cultivable respiratory viruses were detected. Second, concurrent routine regional surveillance did not identify other respiratory viruses circulating elsewhere in the state.

It is more likely that most of the uncultured and culture-negative illnesses were caused by rhinovirus. Persons whose cultures did not yield rhinovirus may have been cultured late relative to the time of peak virus shedding. Studies of experimental rhinovirus infection [15, 16] have shown that peak nasopharyngeal viral titers typically occur 2 to 3 days after inoculation and then rapidly taper off. In our study, the median number of days between illness onset and culture was 2 days (range, 0 to 11 days) and was the same for persons whose cultures did and did not yield rhinovirus. Thus, approximately one half of cultures were obtained after the predicted peak of virus shedding. In addition, it is often technically difficult to culture rhinovirus; in family studies, 21% to 43% of rhinovirus infections were detected only by serologic testing [2, 3, 17]. Thus, because most uncultured and culture-negative persons were unlikely to have other viral infections to explain their respiratory illness and because rhinovirus culture is a technically difficult procedure, we propose that most of these persons were probably infected with rhinovirus.

We caution that our findings pertain to persons with culture-proven infection and may not be generalizable to all newly ill nursing home residents. Our method of surveillance may have biased the results toward the inclusion of more severe cases. Additionally, persons with dementia were underrepresented and persons with chronic obstructive pulmonary disease were overrepresented among those with culture-confirmed rhinovirus infections compared with all other residents in the facility. One possible explanation for this is that surveillance may have been insensitive among persons with dementia, who may have been less able to articulate symptoms [18], and it may have been more sensitive among persons with chronic lung disease, who have a high baseline prevalence of symptoms. We also caution that the finding of more severe illness among those with chronic lung disease should be viewed as preliminary, given the retrospective design of this study. We cannot exclude the possibility that the clinical findings among persons with lung disease were attributable to preexisting disease rather than to rhinovirus infection. However, because these persons were identified in the surveillance reports and medical records as having new, symptomatic respiratory illnesses, we think it unlikely that symptoms simply represented preexisting lung disease. Nevertheless, prospective study would be necessary to determine the precise degree to which underlying lung disease may have contributed to the illness description.

Future studies of respiratory illness in nursing homes should incorporate techniques for rhinovirus culture. Serotyping of recovered isolates may also help to define modes of transmission; we were not equipped to do this. Prospective studies may help to obtain further data on the incidence and nature of lower respiratory tract involvement, particularly in persons with chronic obstructive pulmonary disease.

In summary, we found that a large outbreak of respiratory illnesses among residents of a nursing home was caused by rhinovirus. The outbreak was sharply defined temporally and geographically. Rhinovirus infection was accompanied by clinically overt lower respiratory tract involvement in several persons with chronic obstructive pulmonary disease. The rhinoviruses should be included among the viruses recognized to cause outbreaks of clinically important respiratory illness in nursing home residents.

Dr. Shult: Wisconsin State Laboratory of Hygiene, 465 Henry Mall, Madison, WI 53706.

Ms. Krause and Dr. Drinka: Wisconsin Veterans Home, Department of Nursing, King, WI 54946.

Ms. Miller and Dr. Gravenstein: University of Wisconsin, Institute on Aging, 2245 Medical Sciences Center, 1300 University Avenue, Madison, WI 53706.