Paradoxical Reaction during Tuberculosis Treatment in HIV-Seronegative Patients

Sir—We read with interest the article by Hawkey et al. [1] describing the occurrence of paradoxical reaction (PR) during antituberculous treatment among HIV-seronegative patients with lymph node tuberculosis (TB). In another recent issue of Clinical Infectious Diseases, Vidal et al. [2] described 4 patients who developed TB after treatment with infliximab and in whom the site of PR was mainly the lymph nodes. In our experience, among HIV-seronegative patients with TB, the CNS is a more frequent site of PR than the lymph nodes. The importance of the CNS as a site of PR has been recognized in previous reports [3, 4]. From January 2003 to December 2004, at the outpatient TB clinic of the University of Brescia (Brescia, Italy), we studied the frequency of cases of TB with PR, as well as the clinical characteristics of HIV-seronegative patients with TB who had PRs. PR was defined as a clinical or radiological worsening of previous TB lesions or development of new lesions after at least 1 month of TB treatment in a patient who initially responded to antituberculous therapy (figure 1). PR was detected in 11 (8%) of 137 HIV-seronegative patients with TB. Eight patients (73%) were men, 10 (90%) were immigrants, the median age was 37 years (range, 26–68 years), 4 (36%) had isolated extrapulmonary TB (3 had lymph node TB and 1 had spinal TB), and 7 (64%) had disseminated TB. The median time to PR onset was 107 days (range, 31–443 days). Insurgence of cerebral tuberculomas (6 patients [50%]) and lymph node enlargement (5 patients [42%]) were the main manifestations of PR. Nine (82%) of 11 patients had an initial positive culture result for Mycobacterium tuberculosis, which was sensitive to all first-line TB drugs. At the moment of PR occurrence, new bacteriological tests were performed for 6 patients (54%), and needle aspiration was performed for all 5 patients with lymph node enlargement. These test results were negative for M. tuberculosis for all but 1 patient, a woman from Pakistan who had a strain of M. tuberculosis with the same pattern of sensitivity as the initial isolate from the inguinal lymph node. PR was significantly associated with disseminated and extrapulmonary TB (29% and 11% of patients, respectively, had PR, compared with 0% of patients with pulmonary TB who had PR; P = .002), but no association between PR and sex, age, or ethnicity was found. TB therapy was changed for 2 patients (18%) with cerebral tuberculomas; pyrazinamide was reintroduced for both patients, and levofloxacin was added for 1 of them. Patients with CNS tuberculomas had antituberculous therapy extended for a median duration of 522 days (range, 315–772 days). Steroid therapy was used to control edema in 4 patients (36%) with cerebral tuberculoma. Top, A normal cranial CT scan of a 26-year-old man with an initial diagnosis of pulmonary tuberculosis and tuberculous meningitis. Bottom, CT scan showing an insurgence of cerebral tuberculomas with surrounding edema after 4 months of antituberculous therapy. The same patient also developed cervical lymph node enlargement in the 12th month of antituberculous therapy. PR was not uncommon among HIV-seronegative patients at our TB clinic, and it was significantly more frequent among patients with extrapulmonary or disseminated TB. To distinguish PR from TB treatment failure may represent a clinical dilemma. The initial isolation of drug-sensitive M. tuberculosis and the favorable clinical response to TB treatment experienced by the majority of our patients allowed us to wait for culture results before deciding whether to change antimycobacterial therapy. Steroid therapy was an important tool for controlling cerebral edema, and the need for and benefit of extended antituberculous therapy remained questionable. Repeated needle aspiration of lymph nodes was implemented to prevent fistula formation and to avoid the need for symptomatic steroid therapy or surgical excision. The bacillary load, but more likely the hypersensitivity response to mycobacterial antigens, seems to be the basis of PR pathogenesis; this has been described in studies of HIV-seropositive patients with TB after introduction of antiretroviral therapy [4, 5]. Disseminated TB and associated conditions, such as HIV infection or infliximab treatment, could represent risk factors for PR, because such conditions allow a large load of bacilli in granulomatous lesions. However, further studies are necessary to understand the immunopathogenesis of PR; such studies should help clinicians identify patients who are at higher risk for paradoxical deterioration during TB treatment and better control its clinical manifestations. Potential conflicts of interest. All authors: no conflicts.