Validity of symptoms of Tuberculosis and factors Predictive of Culture-positive Pulmonary tuberculosis among adults with presumptive symptoms of Tuberculosis in a Ugandan Clinic

Abstract

Abstract Background: The WHO TB symptoms screening algorithm including cough, unexplained weight loss, night sweats, fevers, chest pain and hemoptysis has been found to have a validity (sensitivity and specificity) of 72% and 76% respectively for TB screening. This validity may vary among the individual TB symptoms. Although several factors have been suggested as a possible predictors of TB, it is still unknown the predictors of culture-positive pulmonary TB (PTB) among patients with a presumptive TB diagnosis. This study determined the validity of individual TB symptoms as well as the factors most predictive of culture-positive PTB among patients with presumptive TB. Objective: To determine the validity of individual TB symptoms in predicting PTB as well as to determine the factors predictive of culture-positive PTB among presumptive TB patients at health facilities in Kampala, Uganda. Methods: We conducted a validation study utilizing 310 TB records for adult patients in a TB prevalence study among people aged ≥15 years in an urban Ugandan community for the period May 2018 to July 2019. We used the TB symptoms that participants recorded singly in combination with other factors to predict sputum culture-positive PTB. We calculated the sensitivity, specificity, positive predictive values, negative predictive values. Modified Poisson regression was used to estimate prevalence ratios (PR) and 95% confidence intervals of predictors of a positive TB sputum culture in this population. Result: When individual TB symptoms were considered as a tool for identifying culture-positive pulmonary TB, Sensitivity was highest for chronic cough at 86% (95%CI: 76, 92) and lowest for hemoptysis at 8%(95%CI: 3,16). The corresponding specificity was 45%(95%CI:39,52) and 97%(95%CI:94, 99) for cough and hemoptysis respectively. The PPV was highest for hemoptysis at 35% (95%CI:16, 60) and lowest for chest pain at 23%(95%CI:18, 28). Both chronic cough and weight loss had highest NPV at 94%(95%CI: 90, 96 and 91, 96 respectively) whereas hemoptysis had lowest NPV of 84%(95%CI:83, 85).. Predictors of positive culture PTB among presumptive participants included Body Mass Index (BMI) less than 18.5 kg/m2(with adjusted PR of 2.08(95%CI: 1.39, 3.12) and active cigarette smoking, (adjusted PR:1.84, 95%CI:1.15, 2.93). Conclusion:. These findings call for wider public health use of cough as screening tool for PTB and use of protective measures to prevent transmission especially among patients with hemoptysis. Public health awareness on cigarette smoking and risk of PTB should be emphasized.