Cytomegalovirus in the Bronchoalveolar Lavage Fluid of Patients with AIDS* lbtricia R. Miles, M.D.; Robert P. Baughman, M.D., F.C.C.P.; and Calvin C. Linnemann, Jr., M.D.
This study investigated the significance of detecting cytomegalovirus in the bronchoalveolar lavage Ruid of patients with human immunode&ciency virus infection. Bronchoscopy with BAL was performed on all patients. Lavage was examined for CMV by cytology, culture, and immunoftuorescence. The lavage results were compared to clinical status at the time of bronchoscopy and the outcome of the respiratory event. Cytomegalovirus was detected in 51 percent of the BALs in the patients with HIV infection and 25 percent of the immunosuppressed patients without HIV. No association was found in the HIV infected patients between CMV and hypoxemia, abnormal chest roentgeno-
has long been recognized as a C ytomegalovirus cause of pneumonia in the immunocompromised host, 1-4 and it is an important cause of increased morbidity and mortality in the transplant patient. 5 •6 However, CMV has also been detected in asymptomatic immunocompromised patients, as well as in patients simultaneously infected with other pathogenic agents. 7 In some of these cases, the pathogenic role of CMV has been questioned. 1•8 The significance of detecting CMV in the patient with the acquired immunodeficiency syndrome is unclear.&- 11 Cytomegalovirus is frequently cultured from bronchoalveolar lavage specimens, detected by typical cytologic changes in BAL cells and transbronchial biopsy specimens, and detected by indirect immunofluorescence assay to viral early antigens. 1a- 14 Often, CMV is found in association with other pathogens such as Pneumocystis carinii. Several authors have stated that patents with coexistent P carinii and CMV seem to have a poorer outcome than patients with P carinii alone. 14 •15 Others have questioned the role ofCMV as a pathogen in the lung disease of AIDS patients. "' 11 In the present study, BALs performed in immunocompromised patients over a three-year period were reviewed, and patients evaluated in regard to the presence or absence of human immunodeficiency virus, P carinii, or CMV. The alveolar-arterial oxygen (A-a) gradients, peripheral white blood cell counts, *From the DepartJpent oflnternal Medicine, University of Cincinnati Medical Center, Cincinnati. Manuscript received August 21; revision at:cepted October 9. Reprint requests: Dr. Baughman, Pulmonary/Critical Care, 7511 Medical Science Bldg, 231 Bethesda Avenue, Cincinnati 45267-0564
1072
gram, leukopenia, and increased mortality. As indicated by mortality, CMV did not significantly increase the severity of Pneumocyma carinii pneumonia. The study also suggested that CMV in BAL Ruid reftected bronchopulmonary replication of the virus, and not contamination by virus in the blood. Cytomegalovirus does not appear to contribute directly to the pulmonary disease found in most patients with HIV infection. (Cheat 1990; 97:1072-76) CMV =cytomegalovirus; BAL = bronchoalveolar lavage; IFA = immuno8uorescence assay
chest roentgenograms, mortality, and BAL differentials of the different groups of patients were compared. In some patients, blood cultures for CMV were performed within two weeks of the bronchoscopy. METHODS
lbtient lbpulation
Between January 1985 and December 1987, 244 BALs were performed in 19.3 immunocompromised patients with pulmonary signs or symptoms. There were 120 BALs in 79 individuals (77 men and two women) with HIV infections (HIV +) who were at risk for AIDS because of homosexuality, intravenous drug abuse, prostitution, or hemophilia. There were 124 BALs in 116 patients (74 men and 42 women) who were immunocompromised for other reasons (HIV-). These included the following: 39 transplant patients (27 renal, seven cardiac, four bone marrow, and one liver), 25leukemic patients, 17lymphoma patients, two patients with myeloproliferative disorders, 21 cancer patients, and 12 other patients receiving immunosuppressive therapy for various reasons (one rheumatoid arthritis, one Wegener's granulomatosis, two temporal arteritis, three systemic lupus erythematosis, and one each sarcoidosis, vasculitis, interstitial pulmonary fibrosis, renal failure, and idiopathic neutropenia).
Bronchoalveolar Lavage The BAL was performed through a fiberoptic bronchoscope as previously described.•• BrieRy, the bronchoscope was advanced and wedged into a distal airway in the area involved by the infiltrate as seen on chest roentgenogram, or into a subsegment of the right middle lobe or lingula if no infiltrate was noted or diffuse infiltrates were present. After wedging, lavage was done by segmental instillation and immediate withdrawal of two to four 60-ml aliquots of sterile, nonbacteriostatic saline solution (0.9 percent NaCI). The aliquots were pooled. lToceBBing
of Specimens
Aliquots of the pooled BAL were sent for semiquantitative bacteriology culture as previously described, 17 Legionella pneumophilta culture, acid-fast stain, mycobacterial culture, fungal culture, Cylomegalovlrus in BAL Auld of AIDS Pallentll (Miles, Baughman, Unnetmmn)
Table 1-CMV Positive BALs in Patients with and without HIVandPcP
BAL procedures Patients CMV Present: Positive eytology Positive IFA Positive culture PC present: CMV and Pc present:
HIV Infected
HIV Not Infected
120 79 62 (51.6%) 5 34/44* 60 65 (54.2%) 40 (33%)
124 116 31 (256%) 9 12/15* 31 13 (10.5%) 8 (6.4%)
*Number positive/number tested which were culture positive. Two HIV( +)positive patients were IFA( +)but culture negative.
CMV and herpes viral culh1re, and in 141 cases, a rapid IFA to CMV. In the latter, CMV was detected by inoculating human embryonic lung fibroblast cells (MRC-5) with the BAL fluid and then performing an indirect immunofluorescent assay with monoclonal murine antibody to early viral antigen approximately 16 hours later. 11 Aliquots were sent to cytology for Papanicolaou stain and Grocott-methenamine silver stain. Two hundred-microliter samples of the unspun BAL fluid were spun onto glass slides by a cytocentrifuge. Slides were stained with a modified Wright-Giemsa stain for differential counts of the nucleated cells in the BAL Ruid. Differential counts were performed using standard morphologic criteria as previously described in our laboratory. •• Most patients had a WBC count, room air blood gas studies, and chest roentgenogram done prior to the bronchoscopy. The chest roentgenogram was read by the staff radiologist unaware of the results of the BAL. The roentgenogram was categorized as normal or abnormal with either diffuse bilateral infiltrates or local infiltrates involving less than three-lobe disease. The patient had leukopenia ifthe WBC count was less than 3,000 cells/cu mm. The CMV blood cultures were done by inoculating a white blood cell suspension from heparinized blood specimens into MRC-5 cell cultures and observing for typical cytologic changes. •• FoUow-up Evaluation
Patients were followed throughout hospitalization. The HIV( +) patients with a negative BAL were followed for at least four weeks without being given empiric therapy. Patients who died during the hospitalization, which included the bronchoscopy, were classified as deaths associated with respiratory diseases. Analysis of Data
Means and standard deviations were determined for continuous variables. The values between groups were compared using oneway analysis of variance (ANOVA). For discontinuous variables, calculations were performed using chi-square analysis. A p value of less than 0.05 was
This study investigated the significance of detecting cytomegalovirus in the bronchoalveolar lavage Ruid of patients with human immunode&ciency virus infection. Bronchoscopy with BAL was performed on all patients. Lavage was examined for CMV by cytology, culture, and immunoftuorescence. The lavage results were compared to clinical status at the time of bronchoscopy and the outcome of the respiratory event. Cytomegalovirus was detected in 51 percent of the BALs in the patients with HIV infection and 25 percent of the immunosuppressed patients without HIV. No association was found in the HIV infected patients between CMV and hypoxemia, abnormal chest roentgeno-
has long been recognized as a C ytomegalovirus cause of pneumonia in the immunocompromised host, 1-4 and it is an important cause of increased morbidity and mortality in the transplant patient. 5 •6 However, CMV has also been detected in asymptomatic immunocompromised patients, as well as in patients simultaneously infected with other pathogenic agents. 7 In some of these cases, the pathogenic role of CMV has been questioned. 1•8 The significance of detecting CMV in the patient with the acquired immunodeficiency syndrome is unclear.&- 11 Cytomegalovirus is frequently cultured from bronchoalveolar lavage specimens, detected by typical cytologic changes in BAL cells and transbronchial biopsy specimens, and detected by indirect immunofluorescence assay to viral early antigens. 1a- 14 Often, CMV is found in association with other pathogens such as Pneumocystis carinii. Several authors have stated that patents with coexistent P carinii and CMV seem to have a poorer outcome than patients with P carinii alone. 14 •15 Others have questioned the role ofCMV as a pathogen in the lung disease of AIDS patients. "' 11 In the present study, BALs performed in immunocompromised patients over a three-year period were reviewed, and patients evaluated in regard to the presence or absence of human immunodeficiency virus, P carinii, or CMV. The alveolar-arterial oxygen (A-a) gradients, peripheral white blood cell counts, *From the DepartJpent oflnternal Medicine, University of Cincinnati Medical Center, Cincinnati. Manuscript received August 21; revision at:cepted October 9. Reprint requests: Dr. Baughman, Pulmonary/Critical Care, 7511 Medical Science Bldg, 231 Bethesda Avenue, Cincinnati 45267-0564
1072
gram, leukopenia, and increased mortality. As indicated by mortality, CMV did not significantly increase the severity of Pneumocyma carinii pneumonia. The study also suggested that CMV in BAL Ruid reftected bronchopulmonary replication of the virus, and not contamination by virus in the blood. Cytomegalovirus does not appear to contribute directly to the pulmonary disease found in most patients with HIV infection. (Cheat 1990; 97:1072-76) CMV =cytomegalovirus; BAL = bronchoalveolar lavage; IFA = immuno8uorescence assay
chest roentgenograms, mortality, and BAL differentials of the different groups of patients were compared. In some patients, blood cultures for CMV were performed within two weeks of the bronchoscopy. METHODS
lbtient lbpulation
Between January 1985 and December 1987, 244 BALs were performed in 19.3 immunocompromised patients with pulmonary signs or symptoms. There were 120 BALs in 79 individuals (77 men and two women) with HIV infections (HIV +) who were at risk for AIDS because of homosexuality, intravenous drug abuse, prostitution, or hemophilia. There were 124 BALs in 116 patients (74 men and 42 women) who were immunocompromised for other reasons (HIV-). These included the following: 39 transplant patients (27 renal, seven cardiac, four bone marrow, and one liver), 25leukemic patients, 17lymphoma patients, two patients with myeloproliferative disorders, 21 cancer patients, and 12 other patients receiving immunosuppressive therapy for various reasons (one rheumatoid arthritis, one Wegener's granulomatosis, two temporal arteritis, three systemic lupus erythematosis, and one each sarcoidosis, vasculitis, interstitial pulmonary fibrosis, renal failure, and idiopathic neutropenia).
Bronchoalveolar Lavage The BAL was performed through a fiberoptic bronchoscope as previously described.•• BrieRy, the bronchoscope was advanced and wedged into a distal airway in the area involved by the infiltrate as seen on chest roentgenogram, or into a subsegment of the right middle lobe or lingula if no infiltrate was noted or diffuse infiltrates were present. After wedging, lavage was done by segmental instillation and immediate withdrawal of two to four 60-ml aliquots of sterile, nonbacteriostatic saline solution (0.9 percent NaCI). The aliquots were pooled. lToceBBing
of Specimens
Aliquots of the pooled BAL were sent for semiquantitative bacteriology culture as previously described, 17 Legionella pneumophilta culture, acid-fast stain, mycobacterial culture, fungal culture, Cylomegalovlrus in BAL Auld of AIDS Pallentll (Miles, Baughman, Unnetmmn)
Table 1-CMV Positive BALs in Patients with and without HIVandPcP
BAL procedures Patients CMV Present: Positive eytology Positive IFA Positive culture PC present: CMV and Pc present:
HIV Infected
HIV Not Infected
120 79 62 (51.6%) 5 34/44* 60 65 (54.2%) 40 (33%)
124 116 31 (256%) 9 12/15* 31 13 (10.5%) 8 (6.4%)
*Number positive/number tested which were culture positive. Two HIV( +)positive patients were IFA( +)but culture negative.
CMV and herpes viral culh1re, and in 141 cases, a rapid IFA to CMV. In the latter, CMV was detected by inoculating human embryonic lung fibroblast cells (MRC-5) with the BAL fluid and then performing an indirect immunofluorescent assay with monoclonal murine antibody to early viral antigen approximately 16 hours later. 11 Aliquots were sent to cytology for Papanicolaou stain and Grocott-methenamine silver stain. Two hundred-microliter samples of the unspun BAL fluid were spun onto glass slides by a cytocentrifuge. Slides were stained with a modified Wright-Giemsa stain for differential counts of the nucleated cells in the BAL Ruid. Differential counts were performed using standard morphologic criteria as previously described in our laboratory. •• Most patients had a WBC count, room air blood gas studies, and chest roentgenogram done prior to the bronchoscopy. The chest roentgenogram was read by the staff radiologist unaware of the results of the BAL. The roentgenogram was categorized as normal or abnormal with either diffuse bilateral infiltrates or local infiltrates involving less than three-lobe disease. The patient had leukopenia ifthe WBC count was less than 3,000 cells/cu mm. The CMV blood cultures were done by inoculating a white blood cell suspension from heparinized blood specimens into MRC-5 cell cultures and observing for typical cytologic changes. •• FoUow-up Evaluation
Patients were followed throughout hospitalization. The HIV( +) patients with a negative BAL were followed for at least four weeks without being given empiric therapy. Patients who died during the hospitalization, which included the bronchoscopy, were classified as deaths associated with respiratory diseases. Analysis of Data
Means and standard deviations were determined for continuous variables. The values between groups were compared using oneway analysis of variance (ANOVA). For discontinuous variables, calculations were performed using chi-square analysis. A p value of less than 0.05 was
