The clinical, radiological and laboratory features of tuberculosis in children at Mulago Hospital.

Abstract

The incidence of tuberculosis (TB) In Uganda is 150-200/100,000 of the population and an estimated incidence of smear positive adult cases of > 110/100,000 of the population. About a half of the ugandan population was aged less than 15 years. Children form a very large vulnerable age group. Because of inability of children to expectorate sputum there is difficulty in confirming the diagnosis childhood tuberculosis. In this study, children aged 2 months to 5 years who presented to mulago hospital were evaluated for suspected TB using WHO criteria, they were further evaluated using the same criteria for probable or confirmed tuberculosis. OBJECTIVES: To describe the clinical, laboratory and radiological features used in the diagnosis of childhood tuberculosis and to determine the prevalence HIV infection among children with tuberculosis who presented to mulago hospital during the 3 months study period (17th september 2001 to 18th December 2001). DESIGN: Cross sectional descriptive study METHOD: Children (750) who presented to mulago hopital and suspected of having tuberculosis were evaluated for close contact with an adult with pulmonary TB, Weight loss, symptom duration, respiratory signs, lymphadenopathy and hepatosplenomegaly, chest radiography and tuberculin test (Mantoux). The probable cases were subjected to sputum induction or lymphnode biopsy. Smears and cultures of these materials were onbtained. HIV ELISA and RNA PCR tests were performed on all study cases. DATA ANALYSIS: Precoded data was obtained and entered using EPI-INFO software package (version 6). Analysis was done using SPSS software package (version 10). Chi-square tests, bivariate, BMDPLR analysis was used to test association between various clinical, laboratory and radiological variables in reference to confirmed TB cases of RHIV infection. Sensitivities, specifities and positive predictive values of various variables and WHO clinical scoring system were calculated in reference to culture positive cases (gold standard). RESULTS: Children (126) were enrolled in the study as probable or confirmed TB cases in a period of 12 weeks The most common signs and symptoms included; cough for more than 2 weeks, history of weight loss, history of TB contact and excessive sweating. Other symptoms included history of measles (in the last 3 months), complaints od painless swelling (Lymphadenopathy), Persistent wheezing and swelling of the spine. Hepatomegaly was found in 55/126 (44%) and splenomegally in 21/126 (17%) of the cases. The most common respiratory signs were; tachypnoea, chest indrawing, crackles and bronchial breathing. Wasting was found in 67 (53%), oedematous malnutrition in 14 (11%), and stunting in 48 (38%) of the cases. Twenty one (17%) had digital clubbing. There was a high prevalence of HIV infection among children with TB (49%). Wasting (p=0.020), combined stunting and wasting (p=0.026) and digital clubbing (p=0.013) were associated with the presence of of HIV infection. The mantoux test was positive in 55/121 (45%) of the patients had a low sentivity to mantoux test was associated with HIV infection (p=0.000). The mantoux test had a low sensitivity (47%), specificity (60%) and a positive predictive value (36%). Fifteen (12%) of the children had a normal ESR. Out of the 126 children sputum induction was done in 101 cases, lymphnode biopsies were done in 8 cases, cerebrospinal fluid was obtained in 6 cases and 2 pleural fluid samples. Samples for culture were not obtained in 14 patients. Induced sputum yielded 12 (12%) positive smears and 30 (30%) positive cultures. Six out of the 8-lymphnode biopsies were culture positive and the histology of 7 samples were supportive of TB. Overall positive cultures were associated with a history of excessive sweating (p=0.049) and paratracheal adenopathy on chest x-rays (p=0.024). Excessive sweating (p=0.019) and chest X-rays supportive of TB (positive chest x-rays, p=0.043) were found to be independently associated with positive cultures on stepwise logistic regression. Six children had tuberculous meningitis and 5 cases had TB of the spine. Chest X-rays showed various abnormal patterns (singly or mixed). The commonest findings included; Hilar and paratracheal adenopathy, segmental and labar consolidation and other pulmonary infiltrates. Other radiological findings were military TB 5 (4%), pleural effusion 9 (7.2%) and cavity formation in 3 (2.4%) of cases. Chest x-rays supportive of TB had relatively a low sensitivity (72%), specificity (54%) had a positive predictive value (43%) in reference to positive cultures. Most of the children with probable and confirmed TB presented with similar signs and symptoms. The sensitivites, specifities and positive predictive values of clinical, laboratory and radiological features were low and hence use of a constellation of all features is required for the diagnosis of childhood TB. The WHO clinical scoring system had a sensitivity of 86%, specificity of 22% and a positive predictive value of 35%. CONCLUSION: Clinical criteria alone is insufficient in the diagnosis of TB in children. A continued evaluation of patients using other investigations is necessary . Chest X-rays, Mantoux test and induced sputum cultures were clinically helpful adjuncts in the diagnosis of TB in children although they were statistically not significant in this study. HIV infection rate was high among children with TB (49%).