Volume 119 Number 5
Clinical and laboratory observations
HHV-6 occurred during the convalescent phase of their illness, indicating that HHV-6 infection had recently been acquired. In one of four infants who had no antibodies when the outbreak started, seroconversion occurred without rash 2 months later, indicating atypical roseola. The frequency of virus isolation from six infants with roseola was only 50%. We believe that the low frequency of virus isolation was related to the time of blood sampling, because it is difficult to isolate virus from the blood collected from patients with rash. The restriction endonuclease cleavage profiles of four individual HHV-6 strains isolated from infants who had no contact with each other showed obvious differences. These findings proved that differentiation among HHV-6 strains by using these endonucleases is feasible. We therefore used this technique to look for differences among three strains isolated from infants in the outbreak; identical cleavage profiles were detected among these strains. This result suggests that these three strains were probably identical, although the possibility remains that differences might have been found if other endonucleases had been used. These infants thus probably acquired HHV-6 from the same source. These infants had no possibility of artificial transmission through blood (for instance, by way of a needle). In addition, they had no contact/w~h their mothers; nurses had always taken care of them~Because the outbreak occurred in a short period, the i n f a n ~ l y acquired HHV-6 at around the same time from a single nurse or infant. The route of transmission of HHV-6 infection to infants /
76 1
is not known; the possibilities of both vertical transmission from infected mothers during pregnancy and horizontal transmission remain. However, the vertical route of transmission seems unlikely in this outbreak, because all isolated viruses were identical and the infants were 6 to 11 months of age when they had symptoms. HHV-6 may occasionally be secreted by previously infected individuals; saliva is considered the most likely source of virus. Some reports have appeared on the isolation of HHV-6 from the saliva of healthy adults. 4, 5 We thank Dr. Stuart Starr, Division of Infectious Diseases and Immunology, Children's Hospital of Philadelphia, for stimulating discussion and critical reading of this article. REFERENCES
1. Yamanishi K, Okuno T, Shiraki K, et al. Identification of human herpesvirus-6 as a causal agent for exanthem subitum. Lancet 1988;1:1065-7. 2. Hayakawa Y, Torigoe S, Shiraki K, Yamanishi K, Takahashi M. Biologic and biophysical markers of a live varicella vaccine strain (Oka): identification of clinical isolates from vaccine recipients. J Infect Dis 1984;149:956-63. 3. Okuno T, Takahashi K, Balachandra K, et al. Seroepidemiology of human herpesvirus 6 infection in normal children and adults. J Clin Microbiol 1989;27:651-3. 4. Pietroboni GR, Harnett GB, Bucens MR, Honess RW. Isolation of human herpesvirus 6 from saliva [Letter]. Lancet 1988;1:1059. 5. Levy JA, Ferro F, Greenspan D, Lennette ET. Frequent isolation of HHV-6 from saliva and high seroprevalence of the virus in the population. Lancet 1990;335:1047-50.
Value of bronchoalveolar lavage in diagnosing severe respiratory syncytial virus infections in infants M e l i n d a T. Derish, MD," Julie A. Kulhanjian, MD, Lorry R. Frankel, MD, a n d David W, Smith, MB,ChB From the Department of Pediatrics, the Divisions of Infectious Diseases and Intensive Care, Stanford University School of Medicine, Stanford, California
Methods for the rapid diagnosis of respiratory syncytial virus infections are available in most medical centers. The direct fluorescent antibody assay for RSV is a noninvasive, sensitive method for making a rapid diagnosis of RSV
Submitted for publication Dec. 18, 1990; accepted May 30, 1991. Reprint requests: Melinda T. Derish, MD, Department of Pediatrics, Division of Pediatric Intensive Care, Children's Hospital of San Francisco, 3700 California St., San Francisco, CA 94149. *Now at Children's Hospital of San Francisco, San Francisco, Calif. 9/22/31385
infection. 1 This test is commonly performed on specimens taken from the nasopharynx. Confirmation of nasopharyngeal DFA results by simultaneous viral culture is recommended. 2 The diagnosis of RSV infection by bronchoalveolar layage has not been described. Between November 1987 and March 1990, we diagnosed RSV lower respiratory tract infection in four infants by bronchoalveolar lavage. All four infants had lower respiratory tract infections seemingly classic for RSV disease and were initially examined with NP swab specimens. These specimens were negative for RSV by both D F A and viral culture. All four infants had
762
Clinical and laboratory observations
DFA NP RSV
Direct fluorescent antibody Nasopharyngeal Respiratory syncytial virus
severe progressive pulmonary disease requiring endotracheal intubation and mechanical ventilation. Influenza A and B viruses, RSV, and parainfluenza virus were not detected on routine bacterial cultures and NP DFA assay. When the disease states persisted without known cause, the infants eventually underwent diagnostic bronchoalveolar lavage. METHODS Nasopharyngeal specimens were obtained by pediatric house staff physicians with calcium alginate swabs, which were inserted through the anterior nares into the nasopharynx and rotated. On removal, swabs were immediately inoculated at the bedside into 1 ml of viral transport medium consisting of veal infusion broth. In the diagnostic virology laboratory, specimens were mixed by vortex for 10 seconds and treated with gentamicin, 50 mg/ml, and fungizone, 0.05 mg/ml. Before inoculation, the medium was decanted from tissue culture tubes to be inoculated. A volume of 0.1 to 0.2 ml of treated sample was then added to each tissue culture tube within 2 hours after the samples were obtained. The cell lines inoculated were primary rhesus monkey kidney, human foreskin fibroblasts, HEp-2, and A549. After incubation at 37 ~ C for a minimum of 2 hours, the tubes were fed with 1 ml of minimal essential medium containing 10% fetal bovine serum. The tubes were placed on a rotator and incubated at 37 ~ C. Cultures were observed daily for 14 days. Respiratory syncytial virus isolates were identified by characteristic cytopathic changes in tissue cultures and confirmed with immunofluorescent staining with the use of monoclonal antibody reagents. Direct fluorescent antibody testing was performed by collecting a second swab from the nasopharynx and preparing a slide at the bedside. Before staining, slides were fixed in acetone for 10 minutes and then air dried. The RSV antigens were detected by a commercially available direct immunofluorescence technique using fluoreseein isothiocyanate-labeled mouse monoclonal antibodies directed against RSV internal capsid protein and envelope surface glyeoprotein as directed by the manufacturer (Bartels Diagnostics, Bellevue, Wash.). Before staining, slides were screened by phase microscopy for the presence of adequate numbers of citiated respiratory epithelial cells (> 10 cells/ field at x 16 magnification). Specimens were processed and stained within 2 hours of collection. The sensitivity of testing for RSV by DFA assay compared with the sensitivity of testing for viral cultures from NP specimens is 79% in our laboratory with this technique (unpublished data). Bronchoalveolar lavage specimens were obtained from
The Journal of Pediatrics November t991
the first two infants with a flexible fiberoptic bronchoscope, model BF3C4 (Olympus Corp., Lake Success, N.Y.). The bronchoscope was wedged into a distal bronchus and 5 to 10 ml of sterile, nonbacteriostatic saline solution was introduced via the suction channel of the bronchoscope; suction was then applied. Lavage was repeated three or four times until approximately 20 ml of fluid was obtained. Bronchoalveolar lavage for the second two patients was obtained with a 24-inch, 16-gauge Intracath catheter (Delmed, Inc., Concord, Calif.) via the endotracheal tube. This technique, previously described by Barzilay et al., 3 was modified slightly. The needle was disengaged from the catheter and then, after disconnection of the ventilator, the catheter was advanced quickly into the endotracheal tube. While this was done, the catheter remained in its sterile sleeve. The patient's head was turned to the left to facilitate passage of the catheter into the right main-stem bronchus. The wire stylet was then withdrawn and the catheter advanced out of the sleeve and into the airway until it felt "wedged." Then lavage was performed with 5 ml of sterile, nonbacteriostatic saline solution, and suction was applied with the syringe so that the fluid returned. This procedure was repeated three or four times to obtain approximately 10 ml of bronchoalveolar fluid. Bronchoalveolar lavage performed by either technique was well tolerated by our four patients. Adequate oxyhemoglobin saturation was maintained by preoxygenation with 100% oxygen and completion of the procedure in 30 to 60 seconds. Bronchoalveolar lavage fluid obtained from both techniques was processed as described previously. Samples were not centrifuged before processing. Adsorption and inoculum techniques were identical to those used for NP cultures. Smears for DFA assay were prepared from specimens by the diagnostic virology laboratory, and the slides were stained as described above. RESULTS All specimens were positive for RSV on DFA and viral culture. One of the specimens also yielded influenza A virus by DFA assay and viral culture. All bronchoalveolar lavage specimens yielded negative bacterial cultures. DISCUSSION We believe that these four cases illustrate that the diagnosis of severe lower respiratory tract infection caused by RSV can be missed if only NP specimens are evaluated. Theoretical causes for the failure of NP specimens to yield positive DFA results or culture include problems encountered while collecting the clinical specimen, problems with processing the specimens, and differences in the quantity of viral shedding at different sites of the respiratory tract. Some reports suggest that NP swab specimens are inferior to specimens obtained by nasal washing or aspiration for the detection of RSV. 2, 4 However, many authors have reported
Volume 119 Number 5
sensitivities of 72% to 98% for NP swab D F A tests. 5"8 The sensitivity of RSV DFA tests on N P swab specimens in our hospital is 79%, which is comparable to that found in these reports. Other series have identified technical reasons for failure to isolate the virus by DFA assay or culture, including suboptimal epithelial cell collection or transport delaysfl 2 It is doubtful that either of these reasons would account for negative D F A results in our four patients, because slides were prepared directly at the bedside and the specimens stained only after adequate numbers of cells were seen. In all cases, viral culture specimens were collected in appropriate transport media and inoculated into cell lines within 2 hours. We have reviewed the English-language literature for adult and pediatric medicine and could not find studies comparing NP or endotracheal aspirate samples with bronehoalveolar lavage samples for evaluation of viral pneumonias. From the studies of bacterial pneumonias, it seemed possible that bronchoscopy could prove to be a useful diagnostic tool for evaluating viral pneumonia as well. Many studies during the past 3 years have found bronchoalveolar lavage to be more helpful than endotracheal aspirates for examining patients with either nosocomial or communityacquired bacterial pneumonias. The sensitivity of bronchoalveolar lavage in diagnosing such infections has been reported to be as high as 86% to 88%, 9,10 and the specificity as high as 100% 9-l I These studies have predominantly been concerned with immunocompromised adults with pneumonia. A few studies have also mentioned the incidental finding of viral pathogens. For patients with mild to moderate RSV disease, one would be hard-pressed to say that the benefit of diagnosis outweighs the risk of flexible fiberoptic bronchoscopy. However, the patient who requires mechanical ventilation is not in this category. As with any diagnostic test, the risks and benefits of bronchoscopy should be closely evaluated before the procedure is widely applied to patients with an endotracheal tube in place. In skilled hands, flexible fiberoptic bronchoscopy has been reported to have a very low complication rate. 12 In addition, it is possible that the catheter technique may provide the diagnosis at lower cost than flexible fiberoptic bronchoscopy. The diagnosis of RSV by bronchoalveolar lavage in our patients preempted more invasive and costly diagnostic procedures such as open lung biopsy. The positive results allowed us to discontinue broad-spectrum antibiotic therapy and to institute antiviral treatment. Infants with a history of prematurity, congenital heart disease, chronic pulmonary disease, and hematologic malignancies are at risk for severe or fatal RSV disease and thus may be candidates for antiviral therapy. One of our patients had cyanotic heart disease and another patient had bronchopulmonary dysplasia. After detecting RSV infection by bronchoalveolar lavage, we elected to treat these high-risk patients with rib-
Clinical and laboratory observations
763
avirin. Although the efficacy of ribavirin in such cases remains controversial, an informed decision regarding the use of ribavirin or other future antiviral agents could not be made without confirmation of the RSV infection. At the present time, specimens collected from NP swabs or nasal aspirates for DFA assay and viral culture remain the "gold standard" for the diagnosis of RSV respiratory tract infections. From the results of our small series we suspect that there may be a difference in viral shedding at different sites in the respiratory tract, especially for those patients with severe lower respiratory tract disease. For these patients a bronchoalveolar lavage may provide the diagnosis when the results of conventional tests are negative but RSV infection is still suspected. We thank Stanford Diagnostic Virology Laboratory for clinical contributions and discussion. REFERENCES
1. Lauer BA. Comparison of virus culturing and immunofluorescence for rapid detection of respiratory syncytial virus in nasopharyngeal secretions: sensitivity and specificity. J Clin Microbiol 1982;16:411-2. 2. Truehaft MW, Soukup JM, Sullivan BJ. Practical recommendations for the detection of pediatric respiratory syncytial virus infections. J Clin Microbiol 1985;22:270-3. 3. Barzilay Z, Mandel M, Keren G, Davidson S. Nosocomial bacterial pneumonia in ventilated children: clinical significance of culture-positive peripheral bronchial aspirates. J PEOIATR 1988;112:421-4. 4. Ahluwalia G, Embree J, McNicol P, Law B, Hammond GW. Comparison of nasopharyngeal aspirates and nasopharyngeal swab specimens for respiratory syncytial virus diagnosis by cell culture, indirect immunofluorescence assay, and enzymelinked immunosorbent assay. J Clin Microbiol 1987;25:763-7. 5. Arens MQ, Swierkosz EM, Schmidt RR, Armstrong T, Rivetna KA. Strategy for efficient detection of respiratory viruses in pediatric clinical specimens. Diagn Microbiol Infect Dis 1986;5:307-12. 6. Bell DM, Walsh EE, Hruska JF, Schnabel KC, Hall CB. Rapid detection of respiratory syncytial virus with a monoclonal antibody. J Clin Microbiol 1983;17:1099-101. 7. Kim HW, Wyatt RG, Fernie BF, Brandt CD, Arrobio JO, Jeffries BC. Respiratory syncytial virus detection by immunoftuorescenee in nasal secretions with monoclonal antibodies against selected surface and internal proteins. J Clin Microbiol 1983;18:1399-404. 8. Kumar ML, Super DM, Lembo RM, Thomas FC, Prokay SL. Diagnostic efficacy of two rapid tests for detection of respiratory syncytial virus antigen. J Clin Microbiol 1987;25:873-5. 9. Thorpe JE, Baughman RP, Frame PR, Wessler TA, Staneck JL. Bronehoalveolar lavage for diagnosing acute bacterial pneumonia. J Infect Dis 1987;155:855-61. 10. Kahn FW, Jones JM. Diagnosing bacterial respiratory infection by bronchoalveolar lavage. J Infect Dis 1987;155:862-9. 11. Martin W J, Smith TF, Bruntinel WM, Cockerill FR, Douglas WW. Role of bronchoalveolar lavage in the assessment of opportunistic pulmonary infections: utility and complications. Mayo Clin Proe 1987;62:549-57. 12. Wood, RE, Postma D. Endoscopy of the airway in infants and children. J PEDIATR1988;112:1-6.
Clinical and laboratory observations
HHV-6 occurred during the convalescent phase of their illness, indicating that HHV-6 infection had recently been acquired. In one of four infants who had no antibodies when the outbreak started, seroconversion occurred without rash 2 months later, indicating atypical roseola. The frequency of virus isolation from six infants with roseola was only 50%. We believe that the low frequency of virus isolation was related to the time of blood sampling, because it is difficult to isolate virus from the blood collected from patients with rash. The restriction endonuclease cleavage profiles of four individual HHV-6 strains isolated from infants who had no contact with each other showed obvious differences. These findings proved that differentiation among HHV-6 strains by using these endonucleases is feasible. We therefore used this technique to look for differences among three strains isolated from infants in the outbreak; identical cleavage profiles were detected among these strains. This result suggests that these three strains were probably identical, although the possibility remains that differences might have been found if other endonucleases had been used. These infants thus probably acquired HHV-6 from the same source. These infants had no possibility of artificial transmission through blood (for instance, by way of a needle). In addition, they had no contact/w~h their mothers; nurses had always taken care of them~Because the outbreak occurred in a short period, the i n f a n ~ l y acquired HHV-6 at around the same time from a single nurse or infant. The route of transmission of HHV-6 infection to infants /
76 1
is not known; the possibilities of both vertical transmission from infected mothers during pregnancy and horizontal transmission remain. However, the vertical route of transmission seems unlikely in this outbreak, because all isolated viruses were identical and the infants were 6 to 11 months of age when they had symptoms. HHV-6 may occasionally be secreted by previously infected individuals; saliva is considered the most likely source of virus. Some reports have appeared on the isolation of HHV-6 from the saliva of healthy adults. 4, 5 We thank Dr. Stuart Starr, Division of Infectious Diseases and Immunology, Children's Hospital of Philadelphia, for stimulating discussion and critical reading of this article. REFERENCES
1. Yamanishi K, Okuno T, Shiraki K, et al. Identification of human herpesvirus-6 as a causal agent for exanthem subitum. Lancet 1988;1:1065-7. 2. Hayakawa Y, Torigoe S, Shiraki K, Yamanishi K, Takahashi M. Biologic and biophysical markers of a live varicella vaccine strain (Oka): identification of clinical isolates from vaccine recipients. J Infect Dis 1984;149:956-63. 3. Okuno T, Takahashi K, Balachandra K, et al. Seroepidemiology of human herpesvirus 6 infection in normal children and adults. J Clin Microbiol 1989;27:651-3. 4. Pietroboni GR, Harnett GB, Bucens MR, Honess RW. Isolation of human herpesvirus 6 from saliva [Letter]. Lancet 1988;1:1059. 5. Levy JA, Ferro F, Greenspan D, Lennette ET. Frequent isolation of HHV-6 from saliva and high seroprevalence of the virus in the population. Lancet 1990;335:1047-50.
Value of bronchoalveolar lavage in diagnosing severe respiratory syncytial virus infections in infants M e l i n d a T. Derish, MD," Julie A. Kulhanjian, MD, Lorry R. Frankel, MD, a n d David W, Smith, MB,ChB From the Department of Pediatrics, the Divisions of Infectious Diseases and Intensive Care, Stanford University School of Medicine, Stanford, California
Methods for the rapid diagnosis of respiratory syncytial virus infections are available in most medical centers. The direct fluorescent antibody assay for RSV is a noninvasive, sensitive method for making a rapid diagnosis of RSV
Submitted for publication Dec. 18, 1990; accepted May 30, 1991. Reprint requests: Melinda T. Derish, MD, Department of Pediatrics, Division of Pediatric Intensive Care, Children's Hospital of San Francisco, 3700 California St., San Francisco, CA 94149. *Now at Children's Hospital of San Francisco, San Francisco, Calif. 9/22/31385
infection. 1 This test is commonly performed on specimens taken from the nasopharynx. Confirmation of nasopharyngeal DFA results by simultaneous viral culture is recommended. 2 The diagnosis of RSV infection by bronchoalveolar layage has not been described. Between November 1987 and March 1990, we diagnosed RSV lower respiratory tract infection in four infants by bronchoalveolar lavage. All four infants had lower respiratory tract infections seemingly classic for RSV disease and were initially examined with NP swab specimens. These specimens were negative for RSV by both D F A and viral culture. All four infants had
762
Clinical and laboratory observations
DFA NP RSV
Direct fluorescent antibody Nasopharyngeal Respiratory syncytial virus
severe progressive pulmonary disease requiring endotracheal intubation and mechanical ventilation. Influenza A and B viruses, RSV, and parainfluenza virus were not detected on routine bacterial cultures and NP DFA assay. When the disease states persisted without known cause, the infants eventually underwent diagnostic bronchoalveolar lavage. METHODS Nasopharyngeal specimens were obtained by pediatric house staff physicians with calcium alginate swabs, which were inserted through the anterior nares into the nasopharynx and rotated. On removal, swabs were immediately inoculated at the bedside into 1 ml of viral transport medium consisting of veal infusion broth. In the diagnostic virology laboratory, specimens were mixed by vortex for 10 seconds and treated with gentamicin, 50 mg/ml, and fungizone, 0.05 mg/ml. Before inoculation, the medium was decanted from tissue culture tubes to be inoculated. A volume of 0.1 to 0.2 ml of treated sample was then added to each tissue culture tube within 2 hours after the samples were obtained. The cell lines inoculated were primary rhesus monkey kidney, human foreskin fibroblasts, HEp-2, and A549. After incubation at 37 ~ C for a minimum of 2 hours, the tubes were fed with 1 ml of minimal essential medium containing 10% fetal bovine serum. The tubes were placed on a rotator and incubated at 37 ~ C. Cultures were observed daily for 14 days. Respiratory syncytial virus isolates were identified by characteristic cytopathic changes in tissue cultures and confirmed with immunofluorescent staining with the use of monoclonal antibody reagents. Direct fluorescent antibody testing was performed by collecting a second swab from the nasopharynx and preparing a slide at the bedside. Before staining, slides were fixed in acetone for 10 minutes and then air dried. The RSV antigens were detected by a commercially available direct immunofluorescence technique using fluoreseein isothiocyanate-labeled mouse monoclonal antibodies directed against RSV internal capsid protein and envelope surface glyeoprotein as directed by the manufacturer (Bartels Diagnostics, Bellevue, Wash.). Before staining, slides were screened by phase microscopy for the presence of adequate numbers of citiated respiratory epithelial cells (> 10 cells/ field at x 16 magnification). Specimens were processed and stained within 2 hours of collection. The sensitivity of testing for RSV by DFA assay compared with the sensitivity of testing for viral cultures from NP specimens is 79% in our laboratory with this technique (unpublished data). Bronchoalveolar lavage specimens were obtained from
The Journal of Pediatrics November t991
the first two infants with a flexible fiberoptic bronchoscope, model BF3C4 (Olympus Corp., Lake Success, N.Y.). The bronchoscope was wedged into a distal bronchus and 5 to 10 ml of sterile, nonbacteriostatic saline solution was introduced via the suction channel of the bronchoscope; suction was then applied. Lavage was repeated three or four times until approximately 20 ml of fluid was obtained. Bronchoalveolar lavage for the second two patients was obtained with a 24-inch, 16-gauge Intracath catheter (Delmed, Inc., Concord, Calif.) via the endotracheal tube. This technique, previously described by Barzilay et al., 3 was modified slightly. The needle was disengaged from the catheter and then, after disconnection of the ventilator, the catheter was advanced quickly into the endotracheal tube. While this was done, the catheter remained in its sterile sleeve. The patient's head was turned to the left to facilitate passage of the catheter into the right main-stem bronchus. The wire stylet was then withdrawn and the catheter advanced out of the sleeve and into the airway until it felt "wedged." Then lavage was performed with 5 ml of sterile, nonbacteriostatic saline solution, and suction was applied with the syringe so that the fluid returned. This procedure was repeated three or four times to obtain approximately 10 ml of bronchoalveolar fluid. Bronchoalveolar lavage performed by either technique was well tolerated by our four patients. Adequate oxyhemoglobin saturation was maintained by preoxygenation with 100% oxygen and completion of the procedure in 30 to 60 seconds. Bronchoalveolar lavage fluid obtained from both techniques was processed as described previously. Samples were not centrifuged before processing. Adsorption and inoculum techniques were identical to those used for NP cultures. Smears for DFA assay were prepared from specimens by the diagnostic virology laboratory, and the slides were stained as described above. RESULTS All specimens were positive for RSV on DFA and viral culture. One of the specimens also yielded influenza A virus by DFA assay and viral culture. All bronchoalveolar lavage specimens yielded negative bacterial cultures. DISCUSSION We believe that these four cases illustrate that the diagnosis of severe lower respiratory tract infection caused by RSV can be missed if only NP specimens are evaluated. Theoretical causes for the failure of NP specimens to yield positive DFA results or culture include problems encountered while collecting the clinical specimen, problems with processing the specimens, and differences in the quantity of viral shedding at different sites of the respiratory tract. Some reports suggest that NP swab specimens are inferior to specimens obtained by nasal washing or aspiration for the detection of RSV. 2, 4 However, many authors have reported
Volume 119 Number 5
sensitivities of 72% to 98% for NP swab D F A tests. 5"8 The sensitivity of RSV DFA tests on N P swab specimens in our hospital is 79%, which is comparable to that found in these reports. Other series have identified technical reasons for failure to isolate the virus by DFA assay or culture, including suboptimal epithelial cell collection or transport delaysfl 2 It is doubtful that either of these reasons would account for negative D F A results in our four patients, because slides were prepared directly at the bedside and the specimens stained only after adequate numbers of cells were seen. In all cases, viral culture specimens were collected in appropriate transport media and inoculated into cell lines within 2 hours. We have reviewed the English-language literature for adult and pediatric medicine and could not find studies comparing NP or endotracheal aspirate samples with bronehoalveolar lavage samples for evaluation of viral pneumonias. From the studies of bacterial pneumonias, it seemed possible that bronchoscopy could prove to be a useful diagnostic tool for evaluating viral pneumonia as well. Many studies during the past 3 years have found bronchoalveolar lavage to be more helpful than endotracheal aspirates for examining patients with either nosocomial or communityacquired bacterial pneumonias. The sensitivity of bronchoalveolar lavage in diagnosing such infections has been reported to be as high as 86% to 88%, 9,10 and the specificity as high as 100% 9-l I These studies have predominantly been concerned with immunocompromised adults with pneumonia. A few studies have also mentioned the incidental finding of viral pathogens. For patients with mild to moderate RSV disease, one would be hard-pressed to say that the benefit of diagnosis outweighs the risk of flexible fiberoptic bronchoscopy. However, the patient who requires mechanical ventilation is not in this category. As with any diagnostic test, the risks and benefits of bronchoscopy should be closely evaluated before the procedure is widely applied to patients with an endotracheal tube in place. In skilled hands, flexible fiberoptic bronchoscopy has been reported to have a very low complication rate. 12 In addition, it is possible that the catheter technique may provide the diagnosis at lower cost than flexible fiberoptic bronchoscopy. The diagnosis of RSV by bronchoalveolar lavage in our patients preempted more invasive and costly diagnostic procedures such as open lung biopsy. The positive results allowed us to discontinue broad-spectrum antibiotic therapy and to institute antiviral treatment. Infants with a history of prematurity, congenital heart disease, chronic pulmonary disease, and hematologic malignancies are at risk for severe or fatal RSV disease and thus may be candidates for antiviral therapy. One of our patients had cyanotic heart disease and another patient had bronchopulmonary dysplasia. After detecting RSV infection by bronchoalveolar lavage, we elected to treat these high-risk patients with rib-
Clinical and laboratory observations
763
avirin. Although the efficacy of ribavirin in such cases remains controversial, an informed decision regarding the use of ribavirin or other future antiviral agents could not be made without confirmation of the RSV infection. At the present time, specimens collected from NP swabs or nasal aspirates for DFA assay and viral culture remain the "gold standard" for the diagnosis of RSV respiratory tract infections. From the results of our small series we suspect that there may be a difference in viral shedding at different sites in the respiratory tract, especially for those patients with severe lower respiratory tract disease. For these patients a bronchoalveolar lavage may provide the diagnosis when the results of conventional tests are negative but RSV infection is still suspected. We thank Stanford Diagnostic Virology Laboratory for clinical contributions and discussion. REFERENCES
1. Lauer BA. Comparison of virus culturing and immunofluorescence for rapid detection of respiratory syncytial virus in nasopharyngeal secretions: sensitivity and specificity. J Clin Microbiol 1982;16:411-2. 2. Truehaft MW, Soukup JM, Sullivan BJ. Practical recommendations for the detection of pediatric respiratory syncytial virus infections. J Clin Microbiol 1985;22:270-3. 3. Barzilay Z, Mandel M, Keren G, Davidson S. Nosocomial bacterial pneumonia in ventilated children: clinical significance of culture-positive peripheral bronchial aspirates. J PEOIATR 1988;112:421-4. 4. Ahluwalia G, Embree J, McNicol P, Law B, Hammond GW. Comparison of nasopharyngeal aspirates and nasopharyngeal swab specimens for respiratory syncytial virus diagnosis by cell culture, indirect immunofluorescence assay, and enzymelinked immunosorbent assay. J Clin Microbiol 1987;25:763-7. 5. Arens MQ, Swierkosz EM, Schmidt RR, Armstrong T, Rivetna KA. Strategy for efficient detection of respiratory viruses in pediatric clinical specimens. Diagn Microbiol Infect Dis 1986;5:307-12. 6. Bell DM, Walsh EE, Hruska JF, Schnabel KC, Hall CB. Rapid detection of respiratory syncytial virus with a monoclonal antibody. J Clin Microbiol 1983;17:1099-101. 7. Kim HW, Wyatt RG, Fernie BF, Brandt CD, Arrobio JO, Jeffries BC. Respiratory syncytial virus detection by immunoftuorescenee in nasal secretions with monoclonal antibodies against selected surface and internal proteins. J Clin Microbiol 1983;18:1399-404. 8. Kumar ML, Super DM, Lembo RM, Thomas FC, Prokay SL. Diagnostic efficacy of two rapid tests for detection of respiratory syncytial virus antigen. J Clin Microbiol 1987;25:873-5. 9. Thorpe JE, Baughman RP, Frame PR, Wessler TA, Staneck JL. Bronehoalveolar lavage for diagnosing acute bacterial pneumonia. J Infect Dis 1987;155:855-61. 10. Kahn FW, Jones JM. Diagnosing bacterial respiratory infection by bronchoalveolar lavage. J Infect Dis 1987;155:862-9. 11. Martin W J, Smith TF, Bruntinel WM, Cockerill FR, Douglas WW. Role of bronchoalveolar lavage in the assessment of opportunistic pulmonary infections: utility and complications. Mayo Clin Proe 1987;62:549-57. 12. Wood, RE, Postma D. Endoscopy of the airway in infants and children. J PEDIATR1988;112:1-6.
