Paradoxical deterioration during antituberculosis therapy, defined as the clinical or radiological worsening of pre-existing tuberculous lesions or the development of new lesions in a patient who initially improves, remains a diagnostic dilemma. Although different clinical presentations of paradoxical response have been described, a systematic analysis of the entity in non-HIV-infected patients is lacking. Reported here are two cases of paradoxical deterioration in which sequential changes in lymphocyte counts and tuberculin skin test results are emphasized. In addition, 120 episodes of paradoxical response after antituberculosis treatment were reviewed. Of the total 122 episodes, 101 (82.8%) were associated with extrapulmonary tuberculosis. The median time from commencement of treatment to paradoxical deterioration was 60 days. The median time to onset of central nervous system manifestations (63 days) was longer than the time to onset of manifestations at other sites (56 days) ( P=0.02). Development of new lesions in anatomical sites other than those observed at initial presentation was observed in 31 (25.4%) episodes. A surge in the lymphocyte count, accompanied by an exaggerated tuberculin skin reaction, was observed in our patients during the paradoxical deterioration, analogous to the findings in HIV-positive patients. Treatment of the paradoxical response included surgical intervention (60.7%) and administration of steroids (39.3%). The use of steroids appeared to be safe in this series, as 95% of the Mycobacterium tuberculosis isolates were susceptible to first-line antituberculosis therapy.

The results of a consecutive series of 1,349 fine needle aspiration (FNA) biopsies from the head and neck region of 1,193 patients has been reviewed in order to evaluate the efficacy of this method in the diagnosis of tuberculous lymphadenopathy (TBLN). Of the 108 patients whose fine needle aspiration cytology (FNAC) showed granulomatous changes, 68 had subsequent surgery and histological confirmation of the cytological appearance. Sixty-three had TBLN, thus the specificity of FNAC was 93 per cent in diagnosing tuberculous related granulomatous lymphadenopathy. One false positive FNAC was reported histologically to be metastatic mucoepidermal carcinoma.Of the 1,193 patients, 90 patients had subsequently TBLN confirmed histologically. Of these 90 patients, FNA from 69 showed granulomatous changes or acid fast bacilli (AFB), thus the sensitivity of FNAC in detecting tuberculous lymphadenopathy was 77 per cent. Fifty-two cytological smears were stained for acid fast bacilli. Nineteen (37 per cent) contained AFB. It is evident from this review that FNAC is an efficient way to detect cervical tuberculous lymphadenopathy.

No abstract