Abstract

In order to evaluate procedures leading to the diagnosis of tuberculous lymphadenitis, a prospective clinical study was carried out of patients with lymphadenopathy admitted to the medical wards of a referral hospital in Tanzania. The yield of diagnostic procedures (direct auramine/Ziehl-Neelsen (ZN) stained smears, Löwenstein-Jensen (LJ) cultures, cytology and histological examinations of fine needle aspirations (FNA) and biopsy material of lymph nodes, respectively, was compared. We also tried to identify clinical diagnostic markers. One hundred and twenty-eight (99 HIV-seropositive) patients were included. In 89 (67 HIV-positive) patients TB lymphadenitis could be proven. Histology and LJ culture of a lymph node biopsy had the highest diagnostic yield, 85% and 88% respectively, followed by detection of acid-fast bacilli (AFB) in biopsy smear (53%) and in fine-needle aspirations (35%). The diagnostic yield of the several procedures was not affected by associated HIV infection. Macroscopic caseation was 100% predictive for TB with a sensitivity of 69%. Firm and matted lymph nodes, ESR > 100 mm/hr, a positive PPD skin test and pleural opacity on a chest x-ray proved to be independent predictors for TB. Retrospective testing of a stepwise diagnostic approach based on direct smears of FNA, macroscopic visible caseation and direct smear of biopsy tissue, suggests that in 93% of the patients a definite diagnosis of TB lymphadenitis could have been made. Our data suggest that in HIV/TB epidemic areas most of the cases of TB lymphadenitis can be diagnosed correctly by simple and cheap methods which are generally available at district hospitals. Our findings need further prospective validation, however.

PIP:

In order to evaluate procedures leading to the diagnosis of tuberculous lymphadenitis, a prospective clinical study was carried out on patients with lymphadenopathy admitted to four medical wards of the Muhibili Medical Center, Dar es Salaam, Tanzania, from January to August 1991. The yield of diagnostic procedures (direct auramine/Ziehl-Neelsen stained smears, Lowenstein-Jensen [LJ] cultures, cytology and histological examinations of fine needle aspirations [FNA], and biopsy material of lymph nodes) was compared. Clinical diagnostic markers were also identified. 128 (99 HIV-seropositive) patients with a mean age of 30 years were included. 41% were male. In 89 (67 HIV-positive) patients, TB lymphadenitis could be demonstrated. 46 (30 HIV-positive) had TB lymphadenitis only and 43 (37 HIV-positive) had disseminated TB. In 10 patients TB was found in specimens other than the lymph node, making the total of TB patients 99. Histology and LJ culture of lymph node biopsy had the highest diagnostic yield, 85% and 88%, respectively, followed by detection of acid-fast bacilli in biopsy smears (53%) and in FNAs (35%). The diagnostic yield of the procedures was not affected by associated HIV infection. Macroscopic caseation was 100% predictive of TB with a sensitivity of 69%. Multivariate logistic regression analysis demonstrated four independent predictors of TB: 1) firm and matted lymph nodes (odds ratio [OR] 11.8); 2) erythrocyte sedimentation rate (ESR) 100 mm/hr (OR 4.6); 3) a positive purified protein derivative (PPD) skin test (OR 10.2); and 4) pleural opacity on a chest X-ray (OR 9.5). Retrospective testing of a stepwise diagnostic approach based on direct smears of FNA, macroscopic visible caseation, and direct smears of biopsy tissue suggests that in 93% of the patients a definite diagnosis of TB lymphadenitis could have been made. These data suggest that in HIV/TB epidemic areas most of the cases of TB lymphadenitis can be diagnosed correctly by simple and cheap methods which are generally available at district hospitals.