Morbidity and Mortality of Patients with AIDS and First-Episode Pneumocystis carinii Pneumonia Unaffected by Concomitant Pulmonary Cytomegalovirus Infection 1 •2

MARK A. JACOBSON, JOHN MILLS, JOANNE RUSH, LAURENCE PEIPERL, VINITA SERU, PRASANNA K. MOHANTY, PHILIP C. HOPEWELL, W. KEITH HADLEY, V. COURTNEY BROADDUS, GIFFORD LEOUNG, and DAVID W. FEIGAL Introduction SUMMARY To determine the significance of cytomegalovlru8 (CMV) pulmonary colnfectlon with The pathogenic importance of cytomegPneumocystls ca,'nll pneumonia In AIDS, we examined the _oclatlon of long- and short-term alovirus (CMV) infection in pneumoni8urvlvaland morbidity (as defined by length of h08pltai atIIy) with recovery of CMVfrom bronchoscotis associated with the acquired immupy 8peclmens and an Indirect me..ure of vlru8 titer In bronchoalveolar II1V8g8 fluid (the time to nodeficiency syndrome (AIDS) is condevelop CMV cytopathology In culture) In 111 patients dlegnoud with a first episode of If carlnll troversial (1). Since CMV pulmonary pneumonia. Compared with 57 Individuals from whom CMV WIIS not Isolated, the 54 Individuals infection is rarely documented in the abfrom whom CMV were Isolated did not differ In buellne characteristics, long-term 8urvlval (213 versus 275 days, P = 0.97), acute death rete (19% In both, p = 1.0), or length of hospital 8t8y (19.7 sence of other pulmonary opportunistic versus 21.1 days, P =0.88). Also, the time to develop CMVcytopathology In culture did not correlate pathogens, it is difficult to determine if with acute or long-term survival. Our observetlons thus do not 8Upport the use of CMY-apeclflc CMV truly causes pneumonitis in paantiviral therapy In AIDS patlent8 with R carlnll pneumonia who also heve evidence of pulmonary tients with AIDS. CMV has been reportAll REV RESPIR DIS 1111; 144:8-1 CMV Infection. ed in 37 to 53070 of respiratory specimens from patients with Pneumocystiscarinii pneumonia (2-5), the most common cause of pneumonitis in AIDS, and in were decanted and fresh medium (Eagle Methods 19 to 43070 of patients undergoing bronminimum essential medium with 2070 fetal calf Patients choscopy for diagnosis of opportunistic serum) was added. Positive cultures were recWe reviewed the results of all specimens obinfection (6, 7). ognized by typical cytopathic effect and Several small retrospective studies in tained by either bronchoalveolar lavage(HAL) confirmed by reaction with fluoresceinAIDS patients with R carinii pneumo- or transbronchial biopsy (TBB) and submit- conjugated monoclonal CMV-specific antinia reported a poorer clinical outcome ted to the Virology and Parasitology Labora- bodies (SyvaCorporation, Palo Alto, CA) and tories of San Francisco General Hospital bewhen CMV coinfection was present (3, tween October 1, 1983 and September 30, failure to react with polyclonal adenovirus re8, 9). On the other hand, data from two 1984. This time interval was chosen because agents (Microbiologic Associates, Bethesda, larger, more recent studies suggested that it permitted complete long-term survival anal- MD). A specimen was considered positive if . CMV presence in the lung of patients ysis and it preceded the period during which either bronchoalveolar lavage fluid or a transwith R carinii pneumonia did not in- zidovudine and ganciclovir therapy became bronchial biopsy specimen yielded CMV. During the study period, all cultures submitfluence survival (7, 10). In addition, pre- available (which could impact on patient sur- ted to San Francisco General Hospital Virolliminary data from a third recent vival and recovery of CMV from specimens). ogy Laboratory routinely were examined retrospective study suggested that pa- Hospital charts were reviewed for all patients weekly or more frequently for cytopathic eftients with R carinii pneumonia appeared who had either a BAL or TBB specimen sub- fect and held for 30 days, with weeklychanges to have a better outcome (better short- mitted for cytomegalovirusculture and a posi- of medium, before being discarded as tive result from the same specimen for R negative. term survival and less frequent acute in- carinii. Those individuals with a risk factor To compare the titer of CMV in specimens tubation for mechanical ventilation) for human immunodeficiency virus (HIV) inwhen CMV was isolated from bronchoal- fection and no other cause for immunosup- with the time required for cytopathic effect to be detectable in vitro, 30 specimens from veolar lavage fluid (11). pression whose specimen was diagnostic for To determine the significance of CMV their first episode of R carinii pneumonia other patients shedding CMV weretitered (10pulmonary coinfection in a first episode made up the study sample. Dates of death of R carinii pneumonia, weexamined the weredetermined by reviewingpermanent hos- (Received in originalform August 14, 1990 and in association of short-term morbidity and pital charts and records of the AIDS Clinic revised form November 7, 1990) short- and long-term survival with recov- of San Francisco General Hospital, the San 1 From the Departments of Medicine and Laboery of CMV from bronchoscopy speci- Francisco County Department of Public ratory Medicine, University of California, San FranHealth, and the California Statewide Death mens (obtained at the time of diagnosis Index. cisco, and the Medical Service and Virology Laboratory, San Francisco General Hospital, San Franof first-episode R carinii pneumonia), as cisco, California. well as with an indirect measure of virus CMV Culture Technique 1 Correspondence and requests for reprints titer in the bronchoalveolar lavage fluid Specimens obtained by BAL or TBB were in- should be addressed to Mark A. Jacobson, M.D., (i.e., the time to develop CMV cytopa- oculated into human foreskin fibroblast cell Ward 84, Building 80, San Francisco General Hosthology in culture). lines. After I.S h, nonadherent cells and fluids pital, 995 Potrero, San Francisco, CA 94110.

6

7

PNEUMOCVSTIS AND CVlOMEGALOVIRUS

fold dilutions) simultaneously with conventional culture. There was a strong inverse correlation between the titer of CMV in the specimen (greatest dilution from which CMV could be isolated) and the time required for the development of cytopathic effect in culture, r = - 0.57, p < 0.0001 (data not shown). Bronchoalveolar Lavage and Transbronchial Biopsy Method After inspection 0 f the tracheobronchial tree to the subsegmental level, the bronchoscope was wedged into a subsegmental bronchus in the right lower or middle lobe, and 100 to 120 ml sterile, nonbacteriostatic, roomtemperature saline was instilled in 20-ml boluses. Suctioning after each bolus yielded a total of 40 to 80 ml fluid. After lavage was completed, an average of six transbronchial lung biopsy samples weretaken from the right lower lobe. Four drops of lavage fluid or minced biopsy specimen (after mixing in 1.5 ml minimum essential medium) werecultured for virus as described previously.

Statistical Analysis Survival following first-episode R cannu pneumonia and length of hospital stay for this episode were analyzed by the Kaplan-Meier product-limit method. Survival times and lengths of hospital stay were compared by the log rank test. For those individuals for whom date of death could not be determined, survival was censored at the date of last medical chart entry. Differences in baseline characteristics and acute mortality were compared by analysis of variance, the chi-square test, or Fisher's exact test. Survival and length of hospital stay were compared to in vitro time to CMV culture positivity by the Pearson product moment correlation test.

Results

A total of 146individual patients had 170 bronchoscopy samples that both contained R carinii trophozoites or cysts and were cultured for CMV. Of these samples, 111 were from patients having their first episode of R carinii pneumonia (PCP). These 111 patients made up the study sample. BAL was performed on all 111 patients and TBB on 77 (69070). Length of hospital stay for treatment of the first PCP episode could be determined for 110(99070) and date of death for 105(95070) of these 111 patients. Mean time between the date of bronchoscopy and hospital admission for treatment of the first PCP episode was 3.3 days (range o to 26 days). Mean age at the time of diagnosis was 36 yr (range 21 to 59). For 74 (67070) of these 111 individuals, this PCP episode was their index AIDS diagnosis; 28 (25070) had Kaposi's sarcoma and only 9 (8070) had other opportunistic infections or neoplasms as their index AIDS diagnosis. One patient had

concomitant cryptococcal pneumonia at the time of PCP. Fifty-four (49070) of these 111 patients had cytomegalovirus (CMV) isolated from the bronchoscopy specimen (CMV culture-positive group), and 57 (51070) had negative bronchoscopy viral cultures (CMV culture-negative group). BAL and TBB were both performed in 41 (76070) of 54 CMV culture-positive and 36 (63070) of 57 culture-negative individuals (p = 0.16). Among the 41 CMV culturepositive individuals who had both BAL and TBB performed, both specimens were positive for CMV in 18(44070), BAL only was positive in 18 (44070), and TBB only was positive in 5 (12070). Mean age, mean hemoglobin, lactate dehydrogenase (LDH), and absolute lymphocyte count at the time of hospital admission, median time between diagnostic bronchoscopy and hospital admission, and the number of patients with a prior index AIDS diagnosis (such as Kaposi's sarcoma) did not differ between the CMV-positive and CMV-negative groups (table 1). Median survival after first PCP diagnosis, median length of hospital stay for this episode, and acute PCP-associated mortality (defined as death occurring either ~ 28 days following diagnosis or ~ 7 days beyond the length of hospitalization for this episode, whichever was shorter), are summarized in table 2 for the CMVpositive and CMV-negative groups.

We also correlated the time required for specimens to produce CMV cytopathology in culture (previously shown to have to strong inverse correlation with virus titer, see METHODS) to survival. There was data on time to cytopathology in culture on 51 of the 54 individuals from whom CMV was isolated. Median time to cytopathology in culture was 12 days (range 3 to 25 days). Twelve individuals who died acutely during their first PCP episode had a median time to cytopathology in culture of 11.9days compared to 13.5 days for the 39 individuals who survived their first PCP episode (p = 0.39, t test). Survival did not correlate with time to CMV cytopathology in culture (r = 0.10, p = 0.45, Pearson product moment correlation test), even in patients with cytopathology in culture observed within 7 days. Discussion

The significance of CMV pulmonary coinfection in R carinii pneumonia is not a trivial question as 70,000 new AIDS cases are expected in the United States in the next year alone, with R carinii pneumonia occurring in the majority, and CMV pulmonary coinfection likely present in up to half of these new R carinii pneumonia cases. Our study showed that recovery of CMV from the lung had no impact on either short-term

TABLE 1 COMPARISON OF BASELINE CHARACTERISTICS OF AIDS PATIENTS WITH FIRST-EPISODE PCP AT THE TIME OF HOSPITAL ADMISSION

Age, yr Hemoglobin, g/dl LDH, lUlL ALC, cell sllJ. I Bronchoscopy-hospital admission interval, days Prior index AIDS diagnosis PCP KS Other

CMV+ (n)

CMV- (n)

36.7 12.3 381 1184

35.0 12.3 401 951

(54) (45) (16) (46)

p Value

(57) (55) (24) (53)

0.25 0.94 0.78 0.10

2.8 (54)

3.8 (57)

0.26 0.28

39 (54) 10 (54) 5 (54)

35 (57) 18 (57) 4 (57)

DefInition of abbreviations: CMV - cytomegalovirus; PCP - Pneumocystls carlnll pneumonia; LDH lactic dehydrogenase; ALC • absolute lymphocytecount; KS • Kaposi's sarcoma.

TABLE 2 OUTCOME FOR PATIENTS WITH FIRST-EPISODE PCP AS A FUNCTION OF PULMONARY CMV INFECTION

CMV Culture

N

Positive Negative P value

54 57

Median Survival (days)

Median LOS (days)

Acute Mortality

213 275 0.97

19.7 21.1 0.68

19 19 1.0

(0/0)

Definitionof abbreviations: LOS • hospital length of stay; PCP - Pneumocystlccar/nil pneumonia; CMV - cytomegalovirus.

JACOBSON, MILLS, RUSH, ET AL.

8

1.0

0.8 a»

.;c .~

0.6

:::I

UJ C

o ;:; .. 04 . o

.

G-

o

A.

0.2

0.0

o

6

18

12

24

30

Months Fig. 1. Probabilityof survivalfor patientswith or without CMV isolatedfrom bronchoscopy specimensat the time of diagnosisof first episodeof PCRClosedtriangles = CMVculturenegative;opencircles = CMVculturepositive.

morbidity or short- or long-term survival following first episode R carinii pneumonia. Although survival separated at the median value between CMV-positive and CMV-negative patients (213 versus 275 days; see figure 1),this difference was insignificant when the entire two survival curves werecompared (p = 0.97, log rank test). We also examined, in relation to outcome, a more specific indicator of in vivo CMV activity: the time to CMV cytopathology in culture (inversely related to virus titer in the specimen cultured). Survival also could not be related to this variable. Previously published retrospective studies addressing the significance of CMV pulmonary coinfection in R carinii pneumonia have arrived at conflicting conclusions (3, 7, 10). However, none have indexed survival to date of first R carinii pneumonia episode (a powerful predictor of long-term survival [12, 13]), which could have biased the analysis if unequal proportions of second or later R carinii pneumonia episodes were included in the case and control groups. Also, none of the published studies have correlated pulmonary quantitative virology to long-term survival, and none have

examined the association of CMV pulmonary infection with acute morbidity, another clinical measure of pathogenicity. Finally, a negative conclusion regarding the association of CMV and R carinii pneumonia was limited by the large Perror of these studies, which all involved 26 or fewer cases of R carinii pneumonia with CMV pulmonary coinfection. With a sample size of 26 or fewer cases having both R carinii pneumonia and CMV pulmonary coinfection for comparison with control subjects having R carinii pneumonia alone, only a reduction in long-term survival to 105 days or less or an increase in acute mortality to 55070 or more could have been detected with an a error (one-tailed) = 0.05 and Perror = 0.20. In contrast, we have correlated CMV culture results and quantitative virus titer to short- and long-term mortality and to acute morbidity after first-episode R carinii pneumonia. Also, with 54 cases ofR carinii pneumonia and CMV coinfection and 57 control subjects, we could have detected a substantially smaller reduction in long-term survival (to ~ 162 days) and a smaller increase in acute mortality (to ~ 38070) with the same a and P errors.

Pneumonitis frequently has been attributed to CMV in AIDS patients, but there are few reported cases in whom a clinical syndrome of dyspnea, hypoxemia, and diffuse interstitial infiltrates has been reported to correlate with histopathologic evidence of pulmonary CMV cytopathic effect in the absence of other opportunistic pathogens. In fact, although careful autopsy studies of AIDS patients have reported an association between a high density of CMV inclusions and severe lung damage, it has been difficult to precisely define the role of CMV in causing clinically significant pulmonary disease (14). This is in marked contrast to CMV pneumonitis in immunocompromised patients without AIDS, especially renal and bone marrow transplant patients. In these patients, CMV pneumonitis appears to be an immunopathogenic disease: animal and human data support the hypothesis that host inflammatory responses to CMV antigens in the lung can cause clinically significant parenchymal lung damage (15). There is also no evidence that specific antiviral therapy, such as ganciclovir or foscarnet, is beneficial for AIDSassociated CMV pulmonary infection. Preliminary data from a randomized, placebo-controlled prospective study of foscarnet therapy for AIDS-associated pneumonitis showed no difference in death or speed of recovery between the active and placebo groups (16). Thus, there appears to be little evidence supporting CMV as an important opportunistic lung pathogen in AIDS and there seems to be no rationale for instituting CMV-specific antiviral therapy in AIDS patients with R carinii pneumonia who also have evidence of pulmonary CMV infection. Such therapy probably should be reserved for patients with clinical evidence of progressive pneumonitis, hypoxemia, and CMV histopathology in the absence of any other pathogens. Clinicians should not be surprised if even in the latter situation.there is no response to antiviral therapy. Reference. 1. Jacobson MA, Mill~; J. Pulmonary effects of AIDS: cytomegalovirus infection. Clin Chest Med 1988; 9:443-8. 2. Murray JF, Felton CP, Garay SM, et ale Pulmonary complications of the acquired immunodeficiency syndrome: report of a National Heart, Lung, and Blood Institute workshop. N Eng! J Med 1984; 310:1682-8. 3. Stover DE, White DA, Romano PA, et ale Spectrum of pulmonary diseases associated with the acquired immunodeficiency syndrome. Am J Med

9

PNEUMOCVSnS AND CYTOMEGALOVIRUS

1985-; 78:429-37. 4. Miles PR, Baughman RP, Linnemann CC. Cytomegalovirus in the bronchoalveolar lavage fluid of patients with AIDS. Chest 1990; 97:1072-6. 5. Spech TJ, Rehm SJ, Longworth DL, KeysTF, McHenry MC. Significanceof cytomegalovirus isolation from bronchoalveolar lavage in AIDS patients. (abstract 7137). IV International Conference on AIDS. Stockholm, 1988. 6. Broaddus C, Dake MD, Stulbarg MS, et al. Bronchoalveolar lavage and transbronchial biopsy for the diagnosis of pulmonary infections in the acquired immunodeficiency syndrome. Ann Intern Med 1985; 102:747-52. 7. Millar AB, Patou G, Miller RF, et al. Cytomegalovirus in the lungs of patients with AIDS: respiratory pathogen or passenger? Am Rev Respir Dis 1990; 141:1474-7. 8. Minozzi C. Cytomegalovirus in bronchoalveolar lavage of AIDS patients with Pneumocystis carinii

pneumonia (abstract 835). XXVII InterscienceConference on Antimicrobial Agents and Chemotherapy. New York, 1987. 9. Feleke G, Agins B, Lipson S, Forlenza S. The role of cytomegalovirus pulmonary isolates in Pneumocystiscarinii pneumonia in patients with the acquired immunodeficiency syndrome (abstract 250). XXIX Interscience Conference on Antimicrobial Agents and Chemotherapy. Houston, 1989. 10. Bower M, Barton SE, Nelson MR, et al. The significance of the detection of cytomegalovirus in the bronchoalveolar lavage fluid in AIDS patients with pneumonia. AIDS 1990; 4:317-20. 11. Arcia JM, Bozzette SA, Bartok A, McCutchan JA, Rachman D, Spector S. Bronchoalveolar lavageCMV culture positivity is associated with better outcome of Pneumocystis carinii pneumonia in men with AIDS (abstract T.B.P.8). V International Conference on AIDS. Montreal, 1989. 12. Bacchetti P, Osmond D, Chaisson RE, et at.

Survival patterns of the first 500 patients with AIDS in San Francisco. J Infect Dis 1988; 157:1044-7. 13. Harris IE. Improved short-term survival of AIDS patients initially diagnosed with Pneumocystis carinii pneumonia, 1984through 1987. JAMA 1990; 263:397-401. 14. Wallace JM, Hannah J. Cytomegalovirus pneumonitis in patients with AIDS: findings in an autopsy series. Chest 1987; 82:198-203. 15. Grundy JE, Shanley JD, Griffiths PD. Is cytomegalovirus interstitial pneumonitis in transplant recipients an immunopathological condition? Lancet 1987; 2:996-9. 16. Youle M, Gazzard B, Chanas A, Lernstedt J. Treatment with foscarnet of presumed CMV pneumonitis in patients with AIDS: a double-blind placebo controlled study (abstract 3587). IV International Conference on AIDS. Stockholm, 1988.

Morbidity and mortality of patients with AIDS and first-episode Pneumocystis carinii pneumonia unaffected by concomitant pulmonary cytomegalovirus infection.

To determine the significance of cytomegalovirus (CMV) pulmonary coinfection with Pneumocystis carinii pneumonia in AIDS, we examined the association ...
422KB Sizes 0 Downloads 0 Views