Patient 12 had 1 positive sputum sample, but 5 blood cultures with centrifugation lysis, 12 sputa, the bronchoalveolar lavage, and the bone marrow aspiration and biopsy specimen did not reveal M gordonae. All these diagnostic procedures may expose the patient to potential complications. On the other hand, empirical treatment may induce unwanted side effects such as gastrointestinal symptoms, liver dysfunction and failure, anaphylaxis, renal failure, bone marrow suppression, Clostridium difficile enterocolitis, systemic fungal disease, and drug interactions in these patients who already receive multiple medications. The decision has to be individualized, egy patient 13 had a suggestive history, abnormalities on his CXR, and did not respond to treatment with multiple antibiotics. With a bone marrow biopsy specimen, a presumptive diagnosis was achieved and his condition improved significantly with antituberculosis therapy.
Definitive diagnosis is clouded by the fact that patients might have been treated with antibiotics that are not commonly regarded as anti tuberculosis therapy but might still prevent dissemination such as penicillin derivatives with P-lactamase inhibitors, imi-penem, amikacin, trimethoprim/sulfamethoxazole (see patient 20 in Table 2), new macrolides like azithromycin, and especially fluoroquinolones. Five of our 36 patients received treatment with medication potentially active against M gordonae, which might explain why repeated cultures did not reveal this organism.