Kenneth Gunasekera, PhD Student


GHES U.S. Fellow 2019-2020

FELLOWSHIP SITE: Edendale Hospital, south Africa

Project Title: Diagnostic yield of non-sputum-based tests in the clinical evaluation of tuberculosis among children with high risk of HIV-associated TB

In 2017, there were an estimated 1 million incident cases of pediatric tuberculosis (TB), about 10% of all global incident TB cases. There are significant limitations in diagnostics for pediatric TB, especially among children co-infected with HIV. Obtaining viable sputum specimens for microbiological-confirmation in children requires induction of sputum or gastric lavage, both of which are challenging to perform in the POC setting. Additionally, children have low bacterial load, potentially rendering these methods inconclusive, and such tests may be further compromised in the context of HIV.

In high TB/HIV co-burden settings, symptom-based screening may fail to identify potential cases of TB, as HIV co-infection is associated with subclinical presentation of TB without symptoms of TB and with atypical radiographic findings. Diagnosis of TB in patients infected with HIV is clinical indication for initiation of antiretroviral therapy; thus, there is a critical need for non-invasive, non-sputum-based tools to improve diagnosis of pediatric TB in point-of-care (POC) and high TB/HIV co-burden settings.

This study aims to assess the diagnostic yield of non-invasive, non-sputum-based tools against the microbiological and clinical reference standards. The study site is Edendale Hospital, a regional hospital with a catchment area of over 1 million in KwaZulu-Natal, South Africa. This site serves a population with a high TB case notification rate and persistent HIV vertical transmission rate leading to a large vulnerable population for pediatric TB and TB/HIV co-infection.

Secondary aims of this study include better estimation of the local epidemiology of TB in children and preliminary evaluations of diagnostic tool performance and cost-effectiveness analysis. An additional component of this project will be a study of methods to evaluate diagnostic tests using accurate measures of patient outcomes.