BACKGROUND: Tuberculous pericardial effusion is most often due to the spread of tuberculosis from the mediastinal lymph glands; however, no attempt has yet been made to study these glands. We studied the mediastinal glands in proven tuberculous pericardial effusion patients and hypothesized that the findings may be of use in the etiological diagnosis of pericardial effusion. METHODS AND RESULTS: We studied 45 patients with large pericardial effusion or tamponade. All underwent chest computed tomographic studies that were reviewed by radiologists blinded to the diagnosis. Of these 45 patients, 27 had tuberculosis and 18 had viral or idiopathic effusion. Pericardial biopsy was done in 25/27 and tuberculin skin test in 22/27 patients with tuberculosis, and all received specific treatment. In patients with tuberculosis the skin test measured 17+/-3.3 mm. All 27 had mediastinal lymph glands > or = 10 mm in size. The mean size of the mediastinal glands was 19.5+/-8.6 mm and the mean number was 2.5+/-1.2. The aortopulmonary glands were the most frequently enlarged (63%), and hilar the least often (14.8%). The glands showed a hypodense center in 52% of the patients. On follow-up of 15.8+/-10.4 months, glands were not seen in 80.9%, and were smaller in size in 19%; none had a hypodense center. Marked lymphadenopathy was not seen in any patient with viral/idiopathic pericardial effusion. Two had glands < or = 5 mm in size. CONCLUSIONS: Only patients with tuberculosis had substantial mediastinal lymph gland enlargement and not those with viral or idiopathic pericardial effusion. Such glands disappeared or regressed on treatment. In the appropriate clinical context, marked nonhilar mediastinal lymphadenopathy on chest computed tomographic studies along with a strongly positive tuberculin skin test could be of value in the noninvasive diagnosis of pericardial effusion due to tuberculosis.
BACKGROUND:Tuberculous pericardial effusion is most often due to the spread of tuberculosis from the mediastinal lymph glands; however, no attempt has yet been made to study these glands. We studied the mediastinal glands in proven tuberculous pericardial effusionpatients and hypothesized that the findings may be of use in the etiological diagnosis of pericardial effusion. METHODS AND RESULTS: We studied 45 patients with large pericardial effusion or tamponade. All underwent chest computed tomographic studies that were reviewed by radiologists blinded to the diagnosis. Of these 45 patients, 27 had tuberculosis and 18 had viral or idiopathic effusion. Pericardial biopsy was done in 25/27 and tuberculin skin test in 22/27 patients with tuberculosis, and all received specific treatment. In patients with tuberculosis the skin test measured 17+/-3.3 mm. All 27 had mediastinal lymph glands > or = 10 mm in size. The mean size of the mediastinal glands was 19.5+/-8.6 mm and the mean number was 2.5+/-1.2. The aortopulmonary glands were the most frequently enlarged (63%), and hilar the least often (14.8%). The glands showed a hypodense center in 52% of the patients. On follow-up of 15.8+/-10.4 months, glands were not seen in 80.9%, and were smaller in size in 19%; none had a hypodense center. Marked lymphadenopathy was not seen in any patient with viral/idiopathic pericardial effusion. Two had glands < or = 5 mm in size. CONCLUSIONS: Only patients with tuberculosis had substantial mediastinal lymph gland enlargement and not those with viral or idiopathic pericardial effusion. Such glands disappeared or regressed on treatment. In the appropriate clinical context, marked nonhilar mediastinal lymphadenopathy on chest computed tomographic studies along with a strongly positive tuberculin skin test could be of value in the noninvasive diagnosis of pericardial effusion due to tuberculosis.