The results of our study have demonstrated that in HIV-infected patients presenting with atypical chest X-ray and negative sputum smear, culture from bronchoalveolar lavage is the most sensitive method for detection of tuberculosis. The overall sensitivity of sputum culture was relatively low compared to BAL (15.3% versus 21.7%) and this is comparable to findings from other studies.[9
] However, in patients who could produce sputum, the sensitivity of sputum culture was comparable to BAL and post-bronchoscopy sputum. Of note is the fact that 38.5% of the ‘sputum culture’-verified cases were positive in sputum smears after concentration with sodium hypochlorite. This is an important observation because the technique of sputum concentration is simple and affordable in most developing countries. Besides, it will avoid delay in the diagnosis and allow immediate initiation of treatment for tuberculosis. In HIV-positive patients in whom sputum is likely to be negative and the chest X-ray is likely to be misinterpreted as pneumocystis pneumonia or bacterial infection, the value of such method cannot be underestimated. Sputum concentration technique has been shown to improve detection of AFB in both HIV-infected and non-infected patients in the past.[6
BAL smear was positive in one-third of the patients positive by BAL culture and was complementary to sputum smear by concentration technique. However, all the 6 patients that complemented sputum smear were from among patients who could not produce sputum. Results of both smear and culture suggest, to a significant level, that the major benefit of bronchoscopic aspirate is in patients who cannot produce sputum. The two quick methods of direct BAL smear and sputum AFS by concentration method allowed the detection of nearly half of the culture-proven cases. This is welcome news to the treating physician who, otherwise, had to wait 8 weeks before the diagnosis was made by sputum culture.
Post-bronchoscopy sputum culture is highly sensitive compared to BAL culture (sensitivity 88.5%). However, all cultures from post-bronchoscopy sputum were also positive in BAL culture and therefore did not add much to the overall diagnosis. It has been suggested in the past that the instillation of local anesthetic agents into the tracheo- bronchial tree during bronchoscopy will kill some of the mycobacteria and render the BAL less culture-positive compared to post-bronchoscopy sputum. Our finding, however, does not support this notion and BAL culture remains the most sensitive method for the diagnosis of tuberculosis.
Induced sputum has been shown to have an excellent yield compared to BAL and therefore can be a practical solution for those who cannot expectorate adequate sputum.[12
] Unfortunately, this technique has not been evaluated under routine clinical conditions in developing countries. However, it is likely to be an inexpensive and a more practical method compared to bronchoscopy.
Several studies have demonstrated that transbronchial biopsy (TBBX) provides incremental diagnostic information not available from evaluation of sputum or bronchoalveolar lavage.[15
] The procedure, however, is not without complications, particularly pneumothorax and this could mean a lot for an already seriously sick patient. Therefore, TBBX has to be done on selected patients who can tolerate the procedure and who preferably have a unilateral disease. We did not perform transbronchial biopsy in our patients.
As of recent, the polymerase chain reaction (PCR) has been shown to have a very high sensitivity and specificity in both sputum and BAL specimen and is increasingly being used for the diagnosis of tuberculosis.[19
] The clinical usefulness of this technology, however, will be questionable for several reasons, including cost in resource-constrained countries like Ethiopia.
In conclusion, our study demonstrated that sputum AFS using the concentration method and BAL smear for AFB are rapid methods for the diagnosis of PTB and bronchoalveolar lavage is a biologic specimen with the highest yield and is the most useful. It has also demonstrated that the added value of BAL culture over sputum culture is in patients who cannot produce sputum. It is therefore reasonable to conclude, that in HIV-infected patients with negative sputum smear and atypical chest X ray, bronchoscopy as a diagnostic tool for PTB should be spared for patients who cannot produce sputum. Such a recommendation, however, requires a larger and more detailed study before it is implemented. We also recommend that sputum induction, being relatively simple and inexpensive, should be evaluated as an alternative for bronchoscopy in resource-constrained settings.