INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Transbronchial needle aspiration (TBNA) of mediastinal and hilar lymph nodes through the flexible bronchoscope was primarily designed to obviate mediastinoscopy in the diagnosis and staging of primary lung cancer (1). Since the recent introduction by Wang of a larger caliber needle capable of obtaining specimens for pathologic examination (4, 5), recognition of the usefulness of TBNA in the diagnosis of benign diseases has begun to emerge. For example, TBNA for histology specimens has been shown to aid in the diagnosis of sarcoidosis (6, 7). Additionally, four cases of Mycobacterium tuberculosis (8, 9), and one case each of Mycobacterium avium-intracellulare (MAI) (10), and Cryptococcus neoformans (11) infection diagnosed by TBNA via flexible bronchoscopy have been reported; in each of these cases a small-gauge "cytology" needle was used.

At Bellevue Hospital Center, a large inner city public hospital, many of the patients who present with intrathoracic adenopathy have benign etiologies, particularly mycobacterial infection. The majority of this group of patients are coinfected with human immunodeficiency virus (HIV). TBNA through flexible bronchoscopy for histologic specimens was introduced at our institution in February 1992. We report the results of the first 44 TBNA procedures we performed using a histology needle in patients with HIV infection and intrathoracic adenopathy.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A total of 44 procedures were performed in 41 patients who were seen on admission to the Bellevue Chest Service or in consultation. All patients had mediastinal or hilar adenopathy identified on standard chest radiographs or on computed tomography (CT) of the chest. Bronchoscopy was performed because of undiagnosed intrathoracic adenopathy, and when present, parenchymal abnormalities. All patients underwent CT of the chest (GE 9800, Hispeed Advantage; GE Medical Systems, Milwaukee, WI) to evaluate the precise location and nature of the adenopathy. Intravenous contrast was administered at the discretion of the radiologist. All patients had at least three negative examinations of sputum smears for acid-fast bacilli (AFB) prior to bronchoscopy.

Infection with HIV was demonstrated by a positive ELISA for HIV confirmed by Western blot, or by a documented opportunistic infection in a patient with risk factors for HIV infection.

All patients underwent bronchoscopy performed by a single attending pulmonologist (T.J.H.) and pulmonary fellows in training. Using the chest CT as a guide, TBNA of enlarged lymph nodes was performed using a 19-gauge needle capable of obtaining core biopsy specimens (MW-319; Mill-Rose Laboratories, Mentor, OH), employing the technique previously described by Wang (12, 13). Core specimens were fixed in formalin and sent for histologic review. Aspirated fluid was flushed from the needle with sterile saline and sent for AFB smear and culture, and in some cases, for cytologic examination. Adequate lymph node specimens were defined as those demonstrating lymphoid tissue on cytologic or histologic specimens or yielding positive mycobacterial cultures. Bronchoalveolar lavage (BAL), transbronchial biopsy (TBBx), and endobronchial biopsy (EBBx) were performed as clinically indicated.

TBNA was considered to be diagnostic for mycobacterial disease if specimens revealed characteristic granulomata, if acid-fast organisms were identified, or if aspirates grew mycobacterial species in culture. Immediate diagnosis of mycobacterial disease was defined as the finding of granulomata or acid-fast organisms. If granulomata or acid-fast organisms were seen and no organism was cultured, M. tuberculosis was considered to be the causative organism if the adenopathy responded to standard anti-tuberculous chemotherapy. Nontuberculous mycobacterial (NTM) disease was diagnosed by culture of TBNA specimens.

TBNA was considered diagnostic of sarcoidosis if characteristic noncaseating, epithelioid granulomata were found on histologic examination, all AFB smear and culture specimens were negative, the adenopathy was stable radiographically during empiric antituberculous therapy, and either the patient had a previous history of cutaneous sarcoidosis (one patient), or transbronchial and endobronchial biopsies also revealed characteristic noncaseating granulomata (one patient).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

TBNA established a diagnosis in 23 of 44 (52%) procedures (Table 1). Adequate specimens were obtained by TBNA in 35 of 44 procedures (80%). TBNA provided the only diagnostic specimen in 13 of 41 patients (32%). No complications of TBNA occurred, and no bronchoscope was damaged by the technique. Peripheral CD4 counts were available in 28 patients, with a range of 2 to 525 cells/ml (median 65). Twenty-three patients had CD4 counts below 200 cells/ml.

Mycobacterial Disease

Of the 44 procedures, 23 (52%) were performed in patients with mycobacterial disease. TBNA diagnosed mycobacterial disease in 20 of 23 (87%) procedures (11 M. tuberculosis, seven MAI, one combined M. tuberculosis and MAI, one M. fortuitum). AFB smears of aspirated material were positive in 11 of 23 (48%), and histologic specimens were positive in 15 of 23 (65%) (Figure 1), yielding an immediate diagnosis in 17 of 23 (74%). TBNA mycobacterial cultures were positive in 14 of 23 (61%). TBNA provided the only specimen diagnostic of mycobacterial disease in 11 of 23 (48%). TBNA diagnosed M. tuberculosis in 12 of 15 (80%) of procedures. The specimen which grew M. fortuitum showed histologic evidence of infection with a Nocardia species as well, although this organism did not grow in culture. Treatment for both infections led to regression of the adenopathy. Three patients with tuberculosis diagnosed by other bronchoscopic specimens had a nondiagnostic TBNA. This group included one patient who underwent a second TBNA 1 mo after an initial procedure did not yield an immediate diagnosis; shortly after the second TBNA (which ultimately proved to be nondiagnostic), the culture from the first grew M. tuberculosis. Review of all positive AFB cultures in the hospital subsequent to the beginning of the study did not reveal any additional mycobacterial infections among the 41 patients.

View larger version (71K): [in this window] [in a new window]   View larger version (145K): [in this window] [in a new window]   Figure 1.   (Top) Contrast-enhanced CT section at the level of the right middle lobe bronchus showing markedly enlarged low-density "sub-subcarinal" (arrow) and left hilar lymph nodes. No parenchymal abnormalities were found (images not shown). (Bottom) Hematoxylin-eosin-stained section of TBNA biopsy showing neutrophils, nuclear debris, and epithelioid cells consistent with ill-defined suppurative granulomatous inflammation. Culture of TBNA aspirate grew MAI.

Other Causes of Intrathoracic Adenopathy

TBNA diagnosed sarcoidosis in 2 of 4 (50%) patients, and provided the only diagnostic sample in 1 of 4 (25%). TBNA did not reveal granulomata in two patients with sarcoidosis and diagnostic transbronchial biopsies.

Of two patients with non-small cell carcinoma, TBNA was positive in one (Figure 2). In this patient TBNA provided the only diagnostic specimen. Although TBNA specimens revealed fibrotic lymph node consistent with treated lymphoma in one patient, this was not considered to be a diagnostic procedure. Other pulmonary diagnoses established by BAL, TBBx, or EBBx in patients with nondiagnostic TBNA included Kaposi's sarcoma (one), cryptococcal pneumonia (two), non-Hodgkin's lymphoma (one), nocardiosis (one), bacterial pneumonia (one), viral pneumonia (one), and Pneumocystis carinii pneumonia (PCP) (two). No specific pulmonary diagnosis was identified in five patients.

View larger version (133K): [in this window] [in a new window]   View larger version (64K): [in this window] [in a new window]   Figure 2.   (Top) Contrast-enhanced CT section at the level of the aortic arch revealing markedly enlarged homogeneous right paratracheal lymph node (arrow), with a normal-appearing trachea. No parenchymal abnormalities were found (images not shown). (Bottom) Hematoxylin-eosin-stained section of TBNA biopsy, revealing a cluster of non-small cell carcinoma cells admixed with red blood cells and lymphocytes, indicative of malignancy within the lymph node.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

TBNA has proved to be a valuable tool in the diagnosis and nonsurgical staging of bronchogenic carcinoma. Extensive experience in the use of TBNA in malignant disease has been reported (1, 14). Initial reports described the use of 21- or 22-gauge needles to obtain cytology specimens, but more recent articles have addressed the utility of larger caliber needles (18- or 19-gauge) capable of retrieving core samples of lymph nodes suitable for histologic examination (18, 19, 21). In the setting of lung cancer diagnosis and staging, this capability provides the advantage of decreasing reliance on the cytologic specimen; a histologic specimen overcomes the concern of a false-positive finding caused by contamination of a cytology specimen with airway secretions containing malignant cells, although the false-positive rate of cytologic specimens by TBNA is very low (3, 17). More importantly, the addition of histologic specimens to cytologic specimens may increase the sensitivity of TBNA in staging lung cancer. In a direct comparison of 19- and 22-gauge needles, Schenk and colleagues have shown that the combination of histologic and cytologic specimens provided by the larger needle significantly increased the yield of TBNA over cytologic specimens obtained with the 22-g needle (86% versus 53%) (21).

An added benefit of a histology needle is the enhancement of the capability of TBNA to diagnose nonmalignant causes of intrathoracic adenopathy, particularly granulomatous disorders. The widest experience of this nature is in the use of 18-g needles to diagnose sarcoidosis. Wang and coworkers found that TBNA had a yield of 90% in the diagnosis of sarcoidosis, compared with a yield of 60% for TBBx (6). Morales and coworkers reported a yield for TBNA of 51% in sarcoidosis, and found that the addition of TBNA to TBBx increased the overall yield of bronchoscopy by 18% (7).

Due largely to the current HIV epidemic, Bellevue Hospital has a large population of patients with intrathoracic adenopathy unrelated to lung cancer. The majority of these patients have tuberculosis, which is the most common cause of mediastinal adenopathy in patients with HIV. Other causes of intrathoracic adenopathy in HIV include Kaposi's sarcoma, nontuberculous mycobacteria, fungal infections, bacterial pneumonia with reactive nodes, lymphoma, and HIV itself (22). We felt that TBNA offered an attractive means to obtain diagnoses in this group of patients, obviating more invasive surgical procedures such as mediastinoscopy or thoracotomy.

Tuberculosis generally can be easily diagnosed by sputum smear and culture techniques. However, patients with tuberculosis causing intrathoracic adenopathy without radiographic evidence of pulmonary parenchymal involvement will frequently have negative sputum smears. Mediastinal or hilar adenopathy without cavitary lung disease is a common radiographic finding in patients with AIDS and tuberculosis, and is usually believed to represent "primary" disease (23). Although in some of these patients tuberculosis can be diagnosed by sputum culture, the danger presented by untreated tuberculosis to both the HIV-infected patient and to those who come in contact with the patient necessitates an aggressive attempt at an early diagnosis (24). A previous study from our institution reported the utility of CT in the diagnosis of intrathoracic adenopathy due to tuberculosis in HIV patients; the finding of adenopathy with low-density centers, and with peripheral rim enhancement after intravenous contrast administration, is characteristic of tuberculosis (22). However, this CT appearance can also be seen in nontuberculous mycobacterial infections, fungal infections, and in lymph nodes involved with malignancy. Furthermore, the epidemic of multidrug-resistant tuberculosis (MDR-TB) has made the procurement of adequate specimens for drug susceptibility testing imperative in order to provide appropriate therapy. In one recent report, HIV-infected patients with MDR-TB who were not on effective therapy had a median survival of only 2 mo (24). For these reasons, we investigated TBNA as a means to establish a diagnosis in HIV patients with intrathoracic adenopathy.

About one-half of the patients in this study had intrathoracic adenopathy due to mycobacterial infection, and TBNA proved to be of greatest value in this group of patients, with a yield of 87%. Rapid diagnosis of tuberculosis permits the clinician to initiate appropriate therapy at the earliest opportunity, providing therapeutic benefit to the patient as well as decreasing the risk of transmission of infection to others. TBNA yielded a rapid diagnosis of mycobacterial infection in almost three-quarters of patients, and provided the only diagnostic specimen in almost one-half. Although 6 of the 17 procedures that gave a rapid diagnosis ultimately proved to involve nontuberculous mycobacterial disease, meaning that the patients had received standard antituberculous therapy until culture results were available, this is a problem inherent in any biopsy specimen from a normally sterile site (e.g., liver or bone marrow) in HIV-infected individuals. Furthermore, in four of these six individuals, the only specimen that revealed mycobacterial disease was obtained by TBNA. Tissue diagnosis of nontuberculous mycobacterial infection is preferable in the setting of HIV infection because of the common finding of apparently nonpathogenic MAI in sputum cultures (23).

While we found TBNA to be slightly less useful in diagnosing sarcoidosis in this study, the yield of 50% is comparable to that in a previous report (7). Although two of the four patients with sarcoidosis had a positive TBNA, one of these also had diagnostic TBBx and EBBx. The remaining patient had cutaneous sarcoidosis diagnosed 4 yr earlier, prior to his HIV diagnosis. He presented to us with new intrathoracic adenopathy and a CD4 count of 130 cells/ml. Although TBNA in both patients revealed noncaseating granulomata typical of sarcoidosis, these specimens were initially presumed to represent tuberculosis because of the presence of HIV infection. Therefore both patients were treated with standard antituberculous therapy. When all AFB cultures were finalized as negative and the adenopathy had remained stable radiographically, the medications were stopped, and the adenopathy was attributed to sarcoidosis. Although one patient had previously been diagnosed with sarcoidosis, TBNA contributed the only bronchoscopic specimen diagnostic of sarcoidosis during the more recent presentation. In both patients, TBNA made an important contribution by providing a diagnosis other than tuberculosis. The coexistence of sarcoidosis and HIV has been previously reported (25).

TBNA was not positive in any of the 16 procedures performed in patients who did not have either mycobacterial disease, sarcoidosis, or non-small cell lung cancer. Some of these results may represent "true negative" procedures, as the lymphadenopathy may represent inflammatory reaction to the primary pulmonary process (or to HIV infection) rather than actual lymph node involvement. This is most likely in the patients with PCP or nocardiosis, in which recovery of the organism from lymph nodes in uncommon. Although the potential of at least some true negative TBNA is suggested by the finding of adequate lymph node sampling in 8 of the 16 procedures, we do not have confirmatory data such as a mediastinoscopic biopsy or postmortem examination in this group of patients.

The most important contribution made by TBNA was in the 13 patients for whom it provided the only diagnostic material, i.e., those in whom other bronchoscopic specimens and sputum smears and cultures were negative. TBNA was the sole means of diagnosis in 5 of 15 (33%) of procedures in patients with M. tuberculosis, in 6 of 8 (75%) with nontuberculous mycobacteria, in 1 of 2 (50%) with bronchogenic carcinoma, and in 1 of 4 (25%) with sarcoidosis. These patients would have required additional diagnostic tests had TBNA not been performed.

TBNA to diagnose intrathoracic lymphadenopathy can clearly circumvent the need for a more invasive procedure such as mediastinoscopy or thoracotomy, with its attendant risks of general anesthesia. An alternative, minimally invasive procedure, CT-guided needle aspiration of mediastinal nodes, has also been successfully used in the diagnosis of mediastinal tuberculous lymphadenitis. A recent study by Khan and colleagues reported positive results of CT-guided fine needle aspiration in 8 of 12 (66%) patients with mediastinal tuberculosis (26). Not only did TBNA have a higher yield in our series, but it provided the added advantages of not occupying a CT scanner for the length of a procedure and not exposing the patient to additional radiation. However, a choice between these two procedures should probably depend on available expertise.

We conclude that TBNA is a useful and safe means of diagnosing the etiology of intrathoracic adenopathy in patients with HIV infection, particularly in those with mycobacterial disease. TBNA can provide a rapid diagnosis of tuberculosis, and frequently provides the only diagnostic specimen in patients with HIV infection and mediastinal or hilar adenopathy.