Once the decision to treat has been made, it is critical to define adequate therapy. Seven of our HIV-positive patients who received antituberculosis therapy survived, and none of those who died received antituberculosis therapy. Nevertheless, we cannot definitively prove that they died because of M gordonae disease, and not because of an additional infection (see Table 1 and reference 4).
Our data do not provide enough information regarding the optimal regimen, although patient 1 and 11 responded to 6 drugs, patient 8 and 10 responded to 5 drugs, patient 13 responded to 4 drugs, and patient 2 responded to a “new” regimen of ciprofloxacin and amikacin.
With in vitro testing resistance to isoniazid and pyrazinamide, and susceptibility to ethambutol, rifampin (variable) and aminoglycosides were the rule. In a review of the HIV-negative patients in the literature, all responders had isolates sensitive to at least rifampin or ethambutol and were treated with those drugs with good response. Immunocompetent hosts seem to respond even to tuberculostatic therapy. Whether HIV-positive patients with decreased immune response might need two tuber-culocidal drugs remains unknown. The value of a positive anergy panel as a good indicator of sufficient host response could not be evaluated with our data.
The HIV-positive patients with positive M gordonae cultures need further evaluation to differentiate colonization from genuine infection. The workup may include blood cultures, CXRs, and even invasive procedures like bone marrow aspiration and biopsy, fiberoptic bronchoscopy (with bronchoalveolar lavage, transbronchial biopsies, protected specimen brush, or lavage), and liver biopsy may be necessary. There are not enough data to help in deciding which patients need isolation to protect other AIDS patients, especially those with a low CD4+ cell count. Whether it is reasonable to isolate those with repeatedly high colony counts in respiratory secretions has yet to be clarified, since infectivity among HIV-positive patients remains to be investigated.
We conclude that M gordonae usually is a nonpatho-genic organism in HIV-negative patients. In HIV-positive patients with advanced immunosuppression or low CD4+ cell count, it can be an opportunistic pathogen. Antituberculous medication improves symptoms, clinical disease, and may prolong survival.