Journal of Medical Virology 38111-116 (1992)
Shedding of Infectious Virus and Virus Antigen During Acute Infection With Respiratory Syncytial Virus Matti Waris, Olli Meurman, Maurice A. Mufson, Olli Ruuskanen, and Pekka Halonen Department of Virology ( M .W., O.M., P.H.) and Department ofpediatrics (O.R.), University of Turku, Turku, Finland; Marshall University School of Medicine and Veterans Administration Medical Center, Huntington, West Virginia (M.A.M.) Shedding of respiratory syncytial virus (RSV) in nasopharyngeal aspirates (NPA) of hospitalized children with acute respiratory infection was studied using direct antigen detection by timeresolved fluoroimmunoassay, rapid identification of infectious virus in centrifugally inoculated cell cultures by immunoperoxidase staining and conventional virus culture. Sequential NPAs, in which also local RSV-specific IgA response was measured, were collected from children with proven RSV infection. The shedding pattern was similar for both infectious virus and viral antigen. The overall agreement of the three methods was good (81%) in diagnostic specimens collected on admission, but markedly reduced (46%) in follow-up specimens. Secretory IgA was abundant in specimens giving discrepant or negative results only. The proportion of patients who shed RSV was high (287%) in the first week after onset of symptoms, and decreased sharply in the second week. An opposite temporal pattern was found in the proportion of patients with detectable RSV-IgA in their secretions. Sequentially isolated strains were antigenically stable as determined by their reactivity with a large panel of monoclonal antibodies. The findings suggest that RSV shedding should be monitored by using more than one method for virus detection. 0 1992 Wiley-Liss, Inc.
KEY WORDS: respiratory syncytial virus, virus shedding, virus culture, immunoperoxidase, time-resolved fluoroimmunoassay, secretory IgA
INTRODUCTION Respiratory syncytial virus (RSV)is the major cause of hospitalization of infants and young children with lower respiratory tract infections. Although the infection is self-limiting in normal patients, the treatment (0 1992 WILEY-LISS, INC.
being mainly supportive, rapid laboratory diagnosis of the agent is important for several reasons. It may avoid unnecessary antibiotic usage in uncomplicated cases of RSV infection [Hall et al., 19881 and in treatment failures, e.g., in RSV-associated acute otitis media [Arola e t al., 19903. Premature infants, children with severe underlying disease, and immunocompromised patients are at risk for life-threatening RSV infection [Hall e t al., 1986; Meert et al., 19891; ribavirin therapy is indicated in these high risk groups [Smith et al., 19911. Clinical trials with immunotherapy using intravenous gamma-globulins are also in progress [Hemming et al., 19881. Furthermore, nosocomial infection occurs in patients hospitalized with other diseases [Goldmann, 19891. In certain circumstances, the shedding of the virus should be monitored during the hospitalization of the patient. The response to specific treatment of RSV infection should be correlated with specific diagnostic procedures. Preventing nosocomial transmission of RSV demands isolation of the patients who are shedding the virus, which does not always correlate with the recovery from infection [Hall et al., 19751. The pattern of RSV shedding has been studied using conventional virus isolation in cell cultures [Frank e t al., 1981; Hall e t al., 1975, 1976, 19861, but this procedure is too slow to be of practical help in the management of acute clinical situations lArola et al., 1990; Hall et al., 1986, 1988; Meert et al., 19891. Immunofluorescence detection of infected cells in nasal specimens has been shown to give positive results in patients who have become negative for infectious virus [Gardner e t al., 19701. This technique has also been used by McIntosh et al. [1978, 19791 in studies of the secretory antibody response to RSV infection in infants. Due to the prerequisite of intact cells, the laboriousness of large series and the need for a highly skilled reader in immunofluorescence microscopy, solid phase immunoassays have been de-
Accepted for publication March 10, 1992. Address reprint requests to Matti Waris, Department of Virology, University of Turku, SF-20520 Turku, Finland.
Waris et al.
112
veloped and found useful for the diagnosis of RSV infections [Halonen et al., 1985; Kellog, 19911. However, i t is not clear how efficiently immunoassay of virus antigen can detect RSV during the course of acute RSV infection. In the present study, we compared antigen detection by time-resolved fluoroimmunoassay t TR-FIA) and rapid identification of virus in culture by immunoperoxidase staining (IPS)with conventional virus isolation for monitoring the shedding of RSV in nasopharyngeal aspirates (NPAs) during the acute phase of infection. The effect of nasal antibody on the detection of virus was evaluated by measuring RSV-specific IgA titers in nasopharyngeal secretions. In addition, the antigenic stability of sequentially isolated strains from several patients was characterized using a large panel of monoclonal antibodies. MATERIALS AND METHODS Patients and Specimens The study population consisted of children with acute respiratory disease hospitalized a t Turku University Central Hospital during one RSV season in Finland from March 1989 to February 1990. NPAs collected with a disposable mucus extractor were obtained for viral culture and for the detection of RSV antigens by TR-FIA as described earlier [Waris e t al., 19901. The first specimen from 194 children with unknown etiology of illness, after the admission to the hospital, was included in the comparison of methods for RSV diagnosis. One or more follow-up specimens were collected from 41 (including 8 of the 194) randomly selected patients with proven RSV infection determined by rapid laboratory diagnosis, during their stay in the hospital. The total number of these sequentially collected specimens was 116. The mean age of the children enrolled in the follow-up study was 8.6 months (range 2 weeks to 3 years 9 months, median 4.9 months) and 59% of them were males. The average duration of symptoms before hospitalization was 3.4 days (range 1to 7 days) and the average stay in the hospital lasted 5.9 days (range 2 to 14 days 1. Four patients had a compromising underlying disease (two had Down's syndrome, one had WolfHirschhorn syndrome, and one had bronchopulmonary dysplasia ). Viral C u l t u r e The methods of viral culture have been described previously [Waris et al., 19901. NPAs were inoculated onto 7-20 times passaged human diploid fibroblasts (from foreskins), in roller tubes and in 24-well plates. The tubes were incubated a t 36°C for 14 days and observed 3-5 times a week for typical cytopathic effect, whereupon questionably positive cultures were passaged once and negative cultures were subjected to assay for antigens. The plate cultures were centrifuged and incubated for two days at 36°C. The cultures were then fixed, stained with a peroxidase-labeled MAb against the fusion ( F )protein of RSV (RSlO1) and observed by light microscopy. These steps were otherwise
identical to the IPS procedure described earlier IWaris et al., 19901, except that drying of cell monolayers was omitted before fixing with cold 80% acetone. This modification intensified the color of stained cells. Antigen Detection Methods Primary detection of RSV nucleoprotein ( N P ) antigens in fresh NPAs by monoclonal one-step TR-FIA was performed as described earlier [Waris e t al., 19881. Briefly, specimens and a n europium chelate-labeled indicator MAb (NC4) were incubated simultaneously in wells of microtiter strips, previously coated with a capture MAb (RSV4) for 1 h r a t 37°C. The strips were washed, fluorescence enhancement solution was added, and the time-resolved fluorescence was measured as counts per second (cps).The AIB groups of RSV antigens in the aspirates, stored frozen at -2O"C, was determined by TR-FIA with group common (RSV4) and group A specific (RS102)monoclonal capture antibodies [Waris, 19911. Sequential isolates of RSV were passaged in Vero cells for antigenic characterization using enzyme immunoassay with a panel of 34 monoclonal antibodies generated to group A (Long) and group B (WV4843) strains IMufson et al., 1985, 19911. These included 14, 8, 5, 5, and 2 antibodies against different epitopes of the attachment (G), F, NP, matrix (MI and phosphorylated (P)proteins, respectively.
RSV Antibody M e a s u r e m e n t s Cell lysate antigen from RSV-infected and from uninfected Vero cells were prepared as previously described [Waris et al., 19881. Samples of NPAs, stored a t -2O"C, were titrated for IgA class antibody against RSV by EIA as described by Meurman et al. [19841. Briefly, microtiter plates were coated overnight with antigen. Specimens in fivefold serial dilutions were incubated in antigen-coated wells, and bound antibodies were detected with peroxidase-conjugated goat antibodies to human secretory IgA (alpha chain) (Cappel; CooperBiomedical, Malvern, PA). When paired sera were available, the specificity of virus detection was occasionally confirmed by serological diagnosis with fourfold or greater rise in serum RSV IgG titer as indication of acute infection [Meurman et al., 19841.
Statistical Analysis Linear regression analysis and Fisher's exact test were used. RESULTS Detection of RSV Shedding A total of 194 NPAs from children with acute respiratory symptoms were tested for initial laboratory diagnosis by three methods: viral culture with CPE, IPS, and TR-FIA. The results (Table I ) show a good overall agreement between the three methods. Of the total of 48 RSV positive specimens, 39 (81%)were positive by all methods and 41 (85%)were positive by both IPS and TR-FIA. Two specimens were contaminated in the IPS plate, but they were positive by the other two methods.
RSV Shedding
113
TABLE I. Diagnostic Detection of RSV in Nasopharyngeal Aspirates From 194 Children Admitted to the Hospital With Acute Respiratory Infection Method Culture-CPE Culture-IPS TR-FIA Total
No. of specimens Positive Negative 42 44 45 48
152 150 151 146
Percent Sensitivitv NPV" 88 92 94 100
96 97 97 100
"NPV, negative predictive value.
TABLE 11. Detection of RSV in Nasopharyngeal Aspirates Collected Subsequent to the Initial Aspirate From 41 Hnsoitalized Children With Proven RSV Infection
Method Culture-CPE Culture-IPS TR-FIA Total
No. of specimens (patients) Positive Negative 29 (18) 37 (23) 38 (24) 48 (28)
46 (23) 38 (18) 37 (17) 27 (13)
Percent Sensitivity NPV" 61 77 79 100
59 71 73 100
"NPV, negative predictive value.
Elimination of these specimens from the analysis would further increase the agreement between the methods. Four specimens were positive by one method only, two only by IPS and two by TR-FIA. All were considered as true positives because they were confirmed by serological diagnosis, specific reaction in A/B grouping assay, or specific IgA titer increase in sequential NPAs. One to 6 additional specimens (75 in total excluding the initial Specimens) were obtained from 41 children with a laboratory diagnosis of RSV 1 to 12 (mean 4.1) days after admission to the hospital. The results obtained from these follow-up specimens showed a dramatic decrease in the agreement between the methods (Table11).Of the total of 48 specimens positive by any of the methods, 22 (46%)were positive by all methods and 27 (56%)were positive by both IPS and TR-FIA. The result of one of two specimens which were negative by IPS, but which were positive by the other two methods, could not be interpreted because of abundant erythrocytes in the specimen. Fourteen specimens were positive by one method only, 5 only by IPS, and 9 by TR-FIA. Since all positive specimens were always positive either by IPS and/or TR-FIA, these two methods were compared further. A score of positives in IPS was easily obtained by counting the number (if
Shedding of Infectious Virus and Virus Antigen During Acute Infection With Respiratory Syncytial Virus Matti Waris, Olli Meurman, Maurice A. Mufson, Olli Ruuskanen, and Pekka Halonen Department of Virology ( M .W., O.M., P.H.) and Department ofpediatrics (O.R.), University of Turku, Turku, Finland; Marshall University School of Medicine and Veterans Administration Medical Center, Huntington, West Virginia (M.A.M.) Shedding of respiratory syncytial virus (RSV) in nasopharyngeal aspirates (NPA) of hospitalized children with acute respiratory infection was studied using direct antigen detection by timeresolved fluoroimmunoassay, rapid identification of infectious virus in centrifugally inoculated cell cultures by immunoperoxidase staining and conventional virus culture. Sequential NPAs, in which also local RSV-specific IgA response was measured, were collected from children with proven RSV infection. The shedding pattern was similar for both infectious virus and viral antigen. The overall agreement of the three methods was good (81%) in diagnostic specimens collected on admission, but markedly reduced (46%) in follow-up specimens. Secretory IgA was abundant in specimens giving discrepant or negative results only. The proportion of patients who shed RSV was high (287%) in the first week after onset of symptoms, and decreased sharply in the second week. An opposite temporal pattern was found in the proportion of patients with detectable RSV-IgA in their secretions. Sequentially isolated strains were antigenically stable as determined by their reactivity with a large panel of monoclonal antibodies. The findings suggest that RSV shedding should be monitored by using more than one method for virus detection. 0 1992 Wiley-Liss, Inc.
KEY WORDS: respiratory syncytial virus, virus shedding, virus culture, immunoperoxidase, time-resolved fluoroimmunoassay, secretory IgA
INTRODUCTION Respiratory syncytial virus (RSV)is the major cause of hospitalization of infants and young children with lower respiratory tract infections. Although the infection is self-limiting in normal patients, the treatment (0 1992 WILEY-LISS, INC.
being mainly supportive, rapid laboratory diagnosis of the agent is important for several reasons. It may avoid unnecessary antibiotic usage in uncomplicated cases of RSV infection [Hall et al., 19881 and in treatment failures, e.g., in RSV-associated acute otitis media [Arola e t al., 19903. Premature infants, children with severe underlying disease, and immunocompromised patients are at risk for life-threatening RSV infection [Hall e t al., 1986; Meert et al., 19891; ribavirin therapy is indicated in these high risk groups [Smith et al., 19911. Clinical trials with immunotherapy using intravenous gamma-globulins are also in progress [Hemming et al., 19881. Furthermore, nosocomial infection occurs in patients hospitalized with other diseases [Goldmann, 19891. In certain circumstances, the shedding of the virus should be monitored during the hospitalization of the patient. The response to specific treatment of RSV infection should be correlated with specific diagnostic procedures. Preventing nosocomial transmission of RSV demands isolation of the patients who are shedding the virus, which does not always correlate with the recovery from infection [Hall et al., 19751. The pattern of RSV shedding has been studied using conventional virus isolation in cell cultures [Frank e t al., 1981; Hall e t al., 1975, 1976, 19861, but this procedure is too slow to be of practical help in the management of acute clinical situations lArola et al., 1990; Hall et al., 1986, 1988; Meert et al., 19891. Immunofluorescence detection of infected cells in nasal specimens has been shown to give positive results in patients who have become negative for infectious virus [Gardner e t al., 19701. This technique has also been used by McIntosh et al. [1978, 19791 in studies of the secretory antibody response to RSV infection in infants. Due to the prerequisite of intact cells, the laboriousness of large series and the need for a highly skilled reader in immunofluorescence microscopy, solid phase immunoassays have been de-
Accepted for publication March 10, 1992. Address reprint requests to Matti Waris, Department of Virology, University of Turku, SF-20520 Turku, Finland.
Waris et al.
112
veloped and found useful for the diagnosis of RSV infections [Halonen et al., 1985; Kellog, 19911. However, i t is not clear how efficiently immunoassay of virus antigen can detect RSV during the course of acute RSV infection. In the present study, we compared antigen detection by time-resolved fluoroimmunoassay t TR-FIA) and rapid identification of virus in culture by immunoperoxidase staining (IPS)with conventional virus isolation for monitoring the shedding of RSV in nasopharyngeal aspirates (NPAs) during the acute phase of infection. The effect of nasal antibody on the detection of virus was evaluated by measuring RSV-specific IgA titers in nasopharyngeal secretions. In addition, the antigenic stability of sequentially isolated strains from several patients was characterized using a large panel of monoclonal antibodies. MATERIALS AND METHODS Patients and Specimens The study population consisted of children with acute respiratory disease hospitalized a t Turku University Central Hospital during one RSV season in Finland from March 1989 to February 1990. NPAs collected with a disposable mucus extractor were obtained for viral culture and for the detection of RSV antigens by TR-FIA as described earlier [Waris e t al., 19901. The first specimen from 194 children with unknown etiology of illness, after the admission to the hospital, was included in the comparison of methods for RSV diagnosis. One or more follow-up specimens were collected from 41 (including 8 of the 194) randomly selected patients with proven RSV infection determined by rapid laboratory diagnosis, during their stay in the hospital. The total number of these sequentially collected specimens was 116. The mean age of the children enrolled in the follow-up study was 8.6 months (range 2 weeks to 3 years 9 months, median 4.9 months) and 59% of them were males. The average duration of symptoms before hospitalization was 3.4 days (range 1to 7 days) and the average stay in the hospital lasted 5.9 days (range 2 to 14 days 1. Four patients had a compromising underlying disease (two had Down's syndrome, one had WolfHirschhorn syndrome, and one had bronchopulmonary dysplasia ). Viral C u l t u r e The methods of viral culture have been described previously [Waris et al., 19901. NPAs were inoculated onto 7-20 times passaged human diploid fibroblasts (from foreskins), in roller tubes and in 24-well plates. The tubes were incubated a t 36°C for 14 days and observed 3-5 times a week for typical cytopathic effect, whereupon questionably positive cultures were passaged once and negative cultures were subjected to assay for antigens. The plate cultures were centrifuged and incubated for two days at 36°C. The cultures were then fixed, stained with a peroxidase-labeled MAb against the fusion ( F )protein of RSV (RSlO1) and observed by light microscopy. These steps were otherwise
identical to the IPS procedure described earlier IWaris et al., 19901, except that drying of cell monolayers was omitted before fixing with cold 80% acetone. This modification intensified the color of stained cells. Antigen Detection Methods Primary detection of RSV nucleoprotein ( N P ) antigens in fresh NPAs by monoclonal one-step TR-FIA was performed as described earlier [Waris e t al., 19881. Briefly, specimens and a n europium chelate-labeled indicator MAb (NC4) were incubated simultaneously in wells of microtiter strips, previously coated with a capture MAb (RSV4) for 1 h r a t 37°C. The strips were washed, fluorescence enhancement solution was added, and the time-resolved fluorescence was measured as counts per second (cps).The AIB groups of RSV antigens in the aspirates, stored frozen at -2O"C, was determined by TR-FIA with group common (RSV4) and group A specific (RS102)monoclonal capture antibodies [Waris, 19911. Sequential isolates of RSV were passaged in Vero cells for antigenic characterization using enzyme immunoassay with a panel of 34 monoclonal antibodies generated to group A (Long) and group B (WV4843) strains IMufson et al., 1985, 19911. These included 14, 8, 5, 5, and 2 antibodies against different epitopes of the attachment (G), F, NP, matrix (MI and phosphorylated (P)proteins, respectively.
RSV Antibody M e a s u r e m e n t s Cell lysate antigen from RSV-infected and from uninfected Vero cells were prepared as previously described [Waris et al., 19881. Samples of NPAs, stored a t -2O"C, were titrated for IgA class antibody against RSV by EIA as described by Meurman et al. [19841. Briefly, microtiter plates were coated overnight with antigen. Specimens in fivefold serial dilutions were incubated in antigen-coated wells, and bound antibodies were detected with peroxidase-conjugated goat antibodies to human secretory IgA (alpha chain) (Cappel; CooperBiomedical, Malvern, PA). When paired sera were available, the specificity of virus detection was occasionally confirmed by serological diagnosis with fourfold or greater rise in serum RSV IgG titer as indication of acute infection [Meurman et al., 19841.
Statistical Analysis Linear regression analysis and Fisher's exact test were used. RESULTS Detection of RSV Shedding A total of 194 NPAs from children with acute respiratory symptoms were tested for initial laboratory diagnosis by three methods: viral culture with CPE, IPS, and TR-FIA. The results (Table I ) show a good overall agreement between the three methods. Of the total of 48 RSV positive specimens, 39 (81%)were positive by all methods and 41 (85%)were positive by both IPS and TR-FIA. Two specimens were contaminated in the IPS plate, but they were positive by the other two methods.
RSV Shedding
113
TABLE I. Diagnostic Detection of RSV in Nasopharyngeal Aspirates From 194 Children Admitted to the Hospital With Acute Respiratory Infection Method Culture-CPE Culture-IPS TR-FIA Total
No. of specimens Positive Negative 42 44 45 48
152 150 151 146
Percent Sensitivitv NPV" 88 92 94 100
96 97 97 100
"NPV, negative predictive value.
TABLE 11. Detection of RSV in Nasopharyngeal Aspirates Collected Subsequent to the Initial Aspirate From 41 Hnsoitalized Children With Proven RSV Infection
Method Culture-CPE Culture-IPS TR-FIA Total
No. of specimens (patients) Positive Negative 29 (18) 37 (23) 38 (24) 48 (28)
46 (23) 38 (18) 37 (17) 27 (13)
Percent Sensitivity NPV" 61 77 79 100
59 71 73 100
"NPV, negative predictive value.
Elimination of these specimens from the analysis would further increase the agreement between the methods. Four specimens were positive by one method only, two only by IPS and two by TR-FIA. All were considered as true positives because they were confirmed by serological diagnosis, specific reaction in A/B grouping assay, or specific IgA titer increase in sequential NPAs. One to 6 additional specimens (75 in total excluding the initial Specimens) were obtained from 41 children with a laboratory diagnosis of RSV 1 to 12 (mean 4.1) days after admission to the hospital. The results obtained from these follow-up specimens showed a dramatic decrease in the agreement between the methods (Table11).Of the total of 48 specimens positive by any of the methods, 22 (46%)were positive by all methods and 27 (56%)were positive by both IPS and TR-FIA. The result of one of two specimens which were negative by IPS, but which were positive by the other two methods, could not be interpreted because of abundant erythrocytes in the specimen. Fourteen specimens were positive by one method only, 5 only by IPS, and 9 by TR-FIA. Since all positive specimens were always positive either by IPS and/or TR-FIA, these two methods were compared further. A score of positives in IPS was easily obtained by counting the number (if
