Tuberculosis is the major cause of death in HIV-infected people in sub-Saharan Africa. TB in HIV-infected patients (HIV-TB) is characterized by high rates of smear negative pulmonary TB and an increase in disseminated and extrapulmonary disease, which makes TB difficult to diagnose. Delayed TB diagnosis is associated with an increased risk of death. The WHO devised clinical algorithms in an attempt to accelerate diagnosis. One of the two WHO algorithms is for seriously ill patients with cough, who are the group of patients at highest risk of death. However, the algorithm is based largely on expert opinion. Entry to the algorithm is based on cough of >2 weeks duration, yet recently published evidence suggests that cough of any duration is a better discriminator of TB than cough >2 weeks.
The aim of this project is to develop an evidence-based algorithm for diagnosing TB in seriously ill patients. We will assess the duration of cough, the role of other TB symptoms, and the role of abdominal ultrasound to improve the diagnostic accuracy of the algorithm with a multivariable risk prediction model. An important recent advance in TB diagnostics is the development of rapid genotypic tests. One of these, Gene Xpert MTB/Rif test, can be used on-demand and on site with minimal training. It has demonstrated high sensitivity and specificity in ambulatory patients, including those with smear negative disease. We plan to investigate the performance of the Gene Xpert MTB/Rif test in sputum in seriously ill inpatients and will also evaluate the assay in blood following lysis-centrifugation. Furthermore, by using a battery of tests for other respiratory pathogens (including Pneumocystis jirovecii, Cryptococcus neoformans, typical and 'atypical'bacteria) we will investigate the contribution of these organisms to this patient group both as primary and co-infecting pathogens. This will help to rationalize future antibiotic prescribing.

Public Health Relevance

Tuberculosis is the leading cause of death in HIV-infected patients in high HIV prevalent countries. A major contributor to death from TB in HIV-infected people is diagnostic delay due to increases in sputum smear negative pulmonary TB and extrapulmonary TB. This project aims to improve the current WHO management algorithm for seriously ill patients admitted to hospital with cough, with the aim of shortening time to diagnosis, increasing diagnostic accuracy and reducing mortality.

National Institute of Health (NIH)
National Institute of Allergy and Infectious Diseases (NIAID)
Research Project (R01)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1-IDM-R (50))
Program Officer
Decarlo, Ellen S
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Cape Town
South Africa
Zip Code
Griesel, Rulan; Stewart, Annemie; van der Plas, Helen et al. (2017) Optimizing Tuberculosis Diagnosis in HIV-Infected Inpatients Meeting the Criteria of Seriously Ill in the WHO Algorithm. Clin Infect Dis :