The Role of Bronchoscopy in the Diagnosis of Pulmonary Tuberculosis in Patients at Risk for HIV Infection

The Role of Bronchoscopy in the Diagnosis of Pulmonary Tuberculosis in Patients at Risk for HIV Infection

The Role of Bronchoscopy in the Diagnosis of Pulmonary llIberculosis in Patients at Risk for HIV Infection* Steve H. Salzman, M.D., F.C.C.E; Michael L...

580KB Sizes 0 Downloads 6 Views

The Role of Bronchoscopy in the Diagnosis of Pulmonary llIberculosis in Patients at Risk for HIV Infection* Steve H. Salzman, M.D., F.C.C.E; Michael L. Schindel, M.D.;t Conrado E Aranda, M.D.;i Robert L. Smith, M.D., F.C.C.E;§ and Milena L. Lewis, M.D.i

The present study was undertaken to clarify the role of

bronchoalveolar lavage (BAL) and transbronchial biopsy (TBB) in tbe diagnosis ofpulmonary tuberculosis in patients at risk for buman immunode8ciency virus (HIV) infection. We retrospectively identi&ed 31 patients at risk Cor HIV who proved to have MycobtJcterium tuIJer:culoN on culture of at least one pulmonary specimen. All bad pulmonary symptoms but initial sputum smears negative Cor acid-fast bacilli(AFB). All underwent 8beroptfc broncboscopy (FOB), including BAL and TBB; postbroncboscopy sputum was also collected in 19 patients. A specimen was considered to yield an immediate diagnosis when positive Cor AFB either on smear or histologic study; granulomas alone were considered positive wben DO other causes were identi6ed. Overall, an immediate diagnosis was made by broncboscopic specimens in 15 (48 percent) of 31 cases. TBB was the sole positive specimen in seven patients (23 percent).

studies have demonstrated that paE Pidemiologic tients infected with human immunodeficiency

virus (HIV) are at increased risk of developing tuberculosis; matching tuberculosis and acquired immunodeficiency syndrome (AIDS) registries in San Francisco have shown that is percent of patients with tuberculosis also have AIDS.I The sensitivity of sputum acid-fast bacilli (AFB) smears in HIV-infected tuberculosis patients ranges from 31 to 83 percent in different studies. 2,3 However, because of atypical appearance on chest roentgenograms, the diagnosis of tuberculosis may not be suspected in HIV-infected patients, and sputum may not be submitted for acid-

fast staining. 4

Previous reports in nonimmunocompromised patients with sputum smear-negative tuberculosis indicate that bronchoscopic specimens and postbronchoscopy sputum are helpful in establishing a diagnosis.5,6 We wished to clarify the specific contributions of bronchoalveolar lavage (BAL) and transbronchial bi*From the Department of Medicine, New York University and the Bellevue-DVAMC Chest Service, New York. fFellow in Pulmonary Disease. *Clinical Associate Professor of Medicine. §Clinical Assistant Professor of Medicine. Manuscript received April 25; revision accepted July 3. &"nnt requem: Dr. R. L. Smith, VA MediCal Center, First Avenue at East 24th Street; New York City 10010

For comparison, similar specimens from 40 patients in whom M ooium compla (MAC) grew on culture were also evaluated. An immediate identmcation of AFB was made in only four patients (10 percent). We conclude the 6nding of AFB on staining of any pulnionary specimen is highly suggestive of tuberculosis, rather than MAC, and warrants institution or antituberculosis therapy. aU broncboscopic specimens, TBB provides the bighest yield Cor an immediate diagnosis of tuberculosis. (Cheat 1992; 102:143-46)

that

or

APB = acid-fast bacilli; AIDS = acquired immunode6ciency syndrome; BAL bronchoalveolar 1.lavase; CDC Centen for Disease Control; FOB = 6beroptic broncboseopy; IDV = human immunode8ciency virus; MAC= M~m tmium complex; MAl Mycobacterium """m-inlracelWlare; MTb= Mycobt.Jcterium tuberculoaia; pcp PneulllOCfl3lia C4finii pneumonia; TBB = traosbronchial biopsy

=

=

=

=

opsy (TBB) to the diagnosis of tubetculosis in HIVinfected patients. Given the high sensitivity of BAL for diagnosis of Pneumocustis carinii pneumonia (PCP), BAL may be the sole diagnostic procedure performed in HIV-infected patients with pulmonary symptomatology? Therefore, it is important to determine the likelihood that the diagnosis of tuberculosis may be missed by BAL. We also wished to establish the yield of bronchoscopic specimens for immediate diagnosis of AFB infection, by microscopic examination, as distinct from delayed diagnosis by culture. Since tuberculosis in HIV-infected patients usually responds well to conventional therapy early diagnosis is important.' Finally, because Mycobacterium avium complex (MAC) is also commonly found in HIVinfected patients and might confound the immediate diagnosis of tuberculosis, 8 we wished to assess the yield ofbronchoscopic procedures for MAC. .

MATERIALS AND METHODS

Criteria for inclusion in the study were as follows: (I) Mycobacterium tuberculosis (MTb) or MAC was isolated from a specimen of sputum, BAL fluid, or TBB; (2) diagnosis of AIDS or HIV infection was established, or the patient was a member of a group at risk for HIV infection; (3) a diagnostic bronchoscopy; which included both BAL and TBB, was performed; (4) when obtained prior to bronchoscopy, sputum stain was negative for AFB. The medical records of patients hospitalized between April 1984 CHEST I 102 I 1 I JUL'(, 1992

143

established subsequently in six patients. Four patients had PCP and pulmonary tuberculosis concurrently. The findings on the 19 available chest roentgenograms were as follows: focal infiltrates were present in 53 percent, diffuse infiltrates were present in 36 percent, and no infiltrates were present in 11 percent. Mediastinal or hilar adenopathy, or both, was present in 47 percent. Focal infiltrates were in the upper lobes in three (16 percent), midlung fields in one (5 percent), lower lobes in four (21 percent), and multiple areas in two (11 percent). In only two patients was the chest roentgenogram typical of reactivation tuberculosis, ie, upper lobe cavitary infiltrate without adenopathy Sputum smears for AFB were obtained prior to bronchoscopy in 23 of 31 tuberculosis patients; study design dictated that prebronchoscopy sputum was smear negative. Postbronchoscopy sputum was collected in 19 patients. The yield of bronchoscopy-related specimens for an immediate diagnosis of mycobacterial disease is presented in Table 1. Considering all types of bronchoscopy-related specimens (BAL, TBB, sputum), an immediate diagnosis was established in 15 (48 percent) of 31 patients with the diagnosis of tuberculosis. Transbronchial biopsy resulted in an immediate diagnosis in 12 patients (39 percent) and was more sensitive than BAL (10 percent). Postbronchoscopy sputum was smear positive in six (32 percent) of 19 specimens obtained. TBB was the exclusive source in seven (47 percent) of the 15 patients in whom an immediate diagnosis was made; in three of these, diagnosis was based on granuloma in the absence of AFB. In five patients, an immediate diagnosis was obtained from two or more bronchoscopy-related specimens. Diagnostic yield on culture is also presented in Table 1. Although both prebronchoscopy (78 percent) and postbronchoscopy (79 percent) sputum had a high yield on culture, these were not always obtained. The three types of bronchoscopy-related specimens were the only source of positive culture in 13 (42 percent) of 31 patients; in five of these 13, prebronchoscopy sputum was culture negative, and in eight it was not

and December 1988 at Bellevue Hospital Center, a municipal hospital in New York City, and at the Department of Veterans Affairs Medical Center, New York, were reviewed. Seventy-one patients met the above criteria; M1D was isolated in 31 patients and MAC was isolated in 40 patients. In the two institutions, an average of 378 patients at risk for HIV infection underwent diagnostic bronchoscopy in each of the years covered by the study. Fiberoptic bronchoscopy (FOB) was performed transnasally, after premedication with sedatives, opiates, and atropine. Topical lidocaine (Xylocaine) (1 to 2 percent) was used for local anesthesia. After inspection of the airways, the bronchoscope was wedged in a subsegmental bronchus of either the right middle lobe or lingula, when diffuse parenchymal disease was present. In patients with focal parenchymal involvement, the bronchoscope was wedged in areas of maximal involvement, as evident on chest roentgenogram. The BAL consisted of Bve or six 3O-ml aliquots of normal saline solution; Buid was aspirated using mechanical suction. Subsequently, a series of TBBs was performed after instillation of 2 to 5 ml of epinephrine (1:1,000). Aliquots of BAL fluid were sent for staining of the cell pellet in the CYtology laboratory, for AFB smear and culture, and for potassium hydroxide preparation and fungal culture. Four to six TBB specimens were obtained and submitted for acid-fast staining of touch preparations, for histologic examination, and for AFB and fungal culture. In evaluation of diagnostic yield, sputum collected within Gve days after bronchoscopy was considered "postbronchoseopy sputum." Stained specimens of sputum (auramine-rbodamine technique) and BAL cell pellet were considered diagnostic of mycobacterial infection if AFB were identiGed. The TBB specimens were considered diagnostic if AFB were seen on touch preparation or identifted on histologic study. Granulomas alone were presumed to be due to mycobacterial disease when no other causes were identiGed, ie, Pneumocystis, Histoplasma, or foreign bodies. Chest roentgenograms for 19 of the 31 patients with MTb were available for review RESULTS

All 31 patients with MTb infection were male; mean age was 38.3 years (range, 25 to 56 years). Risk factor for HIV infection was intravenous drug use in 24 patients, homosexuality in six, and blood transfusion in one. "IV infection was established by serologic tests in 18 patients who consented to testing. An additional three patients, who were not tested, met Centers for Disease Control (CDC) criteria for AIDS. In total, 12 patients met the CDC criteria for AIDS. The diagnosis of AIDS preceded that of tuberculosis in one patient, was coincident in five, and was

Table 1- Tuberculoaia DGta* Culture Diagnosis

Immediate Diagnosist

Patients Prebronch sputum Bronch spec Sputum BAL TBB

No.

No. Pos

%

Excl

No. Pos

%

Excl

31 23*

15

48

10

31 18

100 78

12 5

6

32 10

2 1 7

15 16 13

79 52

2 3 2

19* 31* 31*

Of

3 12

39

42

*Excl = number of specimens yielding exclusive identiGcation; prebronch = prebronchoscopy; bronch spec = bronchoscopy-related specimens. t AFB on sputum smear, BAL cell pellet, TBB touch preparation or histology,and/or granuloma. *Number of patients in whom at least one such specimen was obtained. tBy study design.

144

Bronchoscopy in Diagnosis of Pulmonary Tuberculosis (salzman et 81)

obtained. Postbronchoscopy sputum was the sole source of positive culture in two patients, BAL in three, and TBB in two. We separately analyzed the subgroup of 21 patients who had known HIV infection. Bronchoscopic specimens yielded an immediate diagnosis in 48 percent, of which TBB was the sole source in 24 percent. We also reviewed the yield of bronchoscopy specimens from 40 patients at risk for HIV infection in whom MAC grew out of sputum specimens, BAL, or TBB. Prebronchoscopy sputum was obtained in 26, and postbronchoscopy sputum was obtained in 19. Specimens of blood were submitted for AFB culture in 20 patients. In contrast to findings in the tuberculosis patients, an immediate diagnosis of possible AFB infection was made in only four patients (10 percent); in two patients, both postbronchoscopy sputum and TBB stains were positive for AFB; in both of these patients, blood also proved culture positive for MAC. In the third patient, postbronchoscopy sputum was smear positive, and in the fourth, TBB was positive for granuloma, but no AFB were identified in tissue. At the time of bronchoscopy, P carinii was found in 11 of the 28 patients from whom pulmonary specimens were ultimately culture positive for MAC. DISCUSSION

The diagnosis of tuberculosis is often unsuspected in HIV-infected patients. Symptoms are nonspecific, upper lobe cavitary infiltrates are usually not present, and the PPD skin test is often nonreactive. 4 A previous study from our institution found that 5 (31 percent) of 16 HIV-infected patients with positive sputum cultures had AFB seen on smear." A more recent study found 19 (83 percent) of 23 patients with HIV infection and pulmonary tuberculosis had positive sputum smears." The role of FOB in the diagnosis of pulmonary tuberculosis in patients with HIV infection has also been evaluated. Broaddus et al? reported on 276 bronchoscopic procedures in 171 AIDS patients. BAL was positive in four (80 percent) of five patients with pulmonary tuberculosis, and 22 (76 percent) of29 with Mycobacterium avium-intracellulare (MAl) pulmonary infection; TBB was positive in 2 of 4 with tuberculosis, and 10 (43 percent) of 23 with MAl. These authors did not distinguish between immediate diagnosis by smear or histopathologic finding from diagnosis on culture; the distinction may be important in HIVinfected patients, given their generally favorable response to antituberculosis therapy Modilevsky et al3 reported that AFB were seen in smears of bronchoscopic specimens in 7 (47 percent) of 15 patients with HIV and tuberculosis; cultures were positive in 12{80 percent)." These authors did not distinguish the contribution of BAL, TBB, bronchial brushing, or open biopsy to this diagnosis.

The present study was designed to clarify the role of BAL and TBB in the diagnosis of tuberculosis in patients at risk for HIV infection, in whom sputum is AFB smear negative or unobtainable. An immediate identification of AFB was established by BAL or TBB in 48 percent of patients with tuberculosis. TBB was the source of immediate diagnosis in 39 percent, whereas BAL had a low sensitivity (10 percent). The importance ofTBB is emphasized by the fact that this procedure was the exclusive source of an immediate diagnosis in 7 (47 percent) of 15 patients in whom an immediate diagnosis was made. Postbronchoscopy sputum provided an immediate diagnosis in 6 (32 percent) of 19 patients in whom it was available, but was the sole source of immediate diagnosis in only 10 percent of these. Bronchoscopy-related specimens also contributed significantly to overall yield on culture. Prebronchoscopy sputa were culture-positive in 18 of 23 patients in whom specimens were submitted, but 42 percent of tuberculosis patients in this study had a diagnosis made exclusively from bronchoscopy-related specimens. Unlike the findings in patients with tuberculosis, acid-fast organisms or granuloma were rarely observed in bronchoscopy-related specimens from patients in whom MAC was grown on culture. These were identified in only five specimens from 4 of 40 patients; only 2 of these patients had positive blood cultures for MAC, confirming disseminated infection. This is in contrast to 21 positive specimens in 15 of 31 patients with tuberculosis. Although the identity of an acidfast organism cannot be defined from a stained specimen, the important finding of our study is that the presence of AFB in smears of bronchoscopic specimens or the finding of granuloma in TBB was strongly associated with culture-proven tuberculosis. These findings agree with those of Modilevsky et al3 who found only one positive AFB smear among bronchoscopic specimens from 17 patients with MAC infection. It is of interest that P carinii was simultaneously identified in 15 patients in whom MAC was cultured, whereas simultaneous PCP was observed in only four patients with tuberculosis, confirming previous observations that patients with MAC infection have more profound immunosuppression. Overall, BAL and TBB contributed to treatment in 20 (65 percent) of 31 patients with tuberculosis, providing an immediate diagnosis in 15, and an exclusive source of positive culture in five in whom prebronchoscopy sputum was culture-negative or not submitted. This finding is consistent with studies in nonimmunocompromised tuberculosis patients by Wallace et al." An immediate diagnosis was established in 48 percent, 26 percent exclusively by TBB. A limitation of the present study is that prebronCHEST I 102 I 1 I JUL~

1992

145

choscopy and postbronchoscopy sputum specimens were not collected in all patients. However, only half of the patients who ultimately proved to have MTb infection had focal infiltrates with adenopathy on chest roentgenograms. Thus, tuberculosis was not suspected, and the diagnosis was Dot specifically sought. Our experience with Hlv-mfeeted patients reflects a common clinical situation reported by others," Another limitation of our study is that only 21 of 31 patients had known "IV infection. A separate analysis of this subgroup yielded results similar to the overall group of31. In summary, FOB contributed to an immediate diagnosis of tuberculosis in 48 percent of patients at risk for "IV infection who proved to have pulmonary tuberculosis. TBB was more sensitive than BAL for establishing an immediate diagnosis, but TBB and BAL had nearly equal sensitivity for ultimate culture diagnosis. Both prebronchoscopy and postbronehoscopy sputum had a higher yield oil culture, but they were not always obtained. The presence of acid-fast organisms in stained smears or histologic specimens is strongly suggestive of infection with MTb rather than MAC.

148

REFERENCES

1 Chaisson BE, Schecter GA, Theuer C~ Rutherford G~ Echenberg OF, Hopewell PC. Tuberculosis in patients with the acquired immunodeficiency syndrome. Am Rev Respir Dis 1987; 136:57074 2 Louie E, Rice LB, Holzman RS. Tuberculosis in non-Haitian patients with acquired immunodeficiency syndrome. Chest 1986; 90:542-45 3 Modilevsky T, Sattler FR, Barnes PF. Mycobacterial disease in patients with human immunodeficiency virus infection. Arch Intern Med 1989; 149:2201-05 4 Flora GS, Modilevsky T, Antoniskis D, Barnes PF. Undiagnosed tuberculosis in patients with human immunodeficiency virus infection. Chest 1990; 98:1056-59 5 Jett JR, Cortese DA, Dines DE. The value of bronchoscopy in the diagnosis of mycobacterial disease. Chest 1981;80:575-78 6 Danek SJ" Bower JS. Diagnosis of pulmonary tuberculosis by ftexible fiberoptic bronchoscopy. Am Rev Respir Dis 1979; 119:677-79 7 Broaddus C, Dake MD, Stulbarg MS, Blumenfeld Hadley K, Golden JA, 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 8 Tenholder MF, Moser RJ, Tellis CJ. Mycobacteria other than tuberculosis: pulmonary involvement in patients with acquired immunodeficiency syndrome. Arch Intern Med 1988; 148:953-55 9 WallaceJM, Deutsch AL, Harrell JH, Moser KM. Bronchoscopy and transbronchial biopsy in evaluation of patients with suspected active tuberculosis. Am J Med 1981;70:1189-94

w

Bronchoecopy in Diagnosisof PulmonaryTuberculosis (S8Izman sf 8/)