Prognostic survival models and adaptive trial designs in tuberculous meningitis

Project Overview

The whole project is focused on patients with tuberculous meningitis. The first aim is to develop a model that predicts mortality within the first nine months after diagnosis, combining patient data with clinical and laboratory examinations. Results from a model that only uses information collected at baseline (i.e. diagnosis) have been published. A paper that compares different methods of model construction has been submitted. Currently, we investigate the added value of time-updated information on Glasgow Coma Score and sodium.

The second aim is to use and develop statistical methodology of adaptive clinical trial designs. The methodology is used in a randomized double blind placebo controlled non-inferiority trial of adjunctive dexamethasone for the treatment of HIV-uninfected adults with tuberculous meningitis stratified by Leukotriene A4 hydrolase genotype. 

Training Opportunities

The DPhil student will be trained in modern methods for the analysis of time-to-event data, especially in relation to prediction based on time-updated covariables. Other training is in adaptive trial design and scientific writing.

Theme

Clinical Trials & Epidemiology and Immunology & Infectious Disease

Admissions

Project reference number: 978

Funding and admissions information

Supervisors

Name Department Institution Country Email
Professor Ronald B. Geskus Tropical Medicine Oxford University, Ho Chi Minh City VNM rgeskus@oucru.org
Professor Guy Thwaites Tropical Medicine Oxford University, Ho Chi Minh City VNM gthwaites@oucru.org

Thao LTP, Heemskerk AD, Geskus RB, Mai NTH, Ha DTM, Chau TTH, Phu NH, Chau NVV, Caws M, Lan NH, Thu DDA, Thuong NTT, Day J, Farrar JJ, Torok ME, Bang ND, Thwaites GE, Wolbers M. 2017. Prognostic models for 9 month mortality in tuberculous meningitis. Clin. Infect. Dis., Read abstract | Read more

Background: Tuberculous meningitis (TBM) is the most severe form of extra-pulmonary tuberculosis. We developed and validated prognostic models for 9-month mortality in HIV-uninfected and HIV-infected adults with TBM. Methods: We included 1699 subjects from four randomized clinical trials and one prospective observational study conducted at two major referral hospitals in Southern Vietnam from 2001-2015. Modelling was based on multivariable Cox proportional hazards regression. The final prognostic models were validated internally and temporally, and displayed using nomograms and a web-based app (https://thaole.shinyapps.io/tbmapp/). Results: A total of 951 HIV-uninfected and 748 HIV-infected subjects with TBM were included, of whom 219/951 (23.0%) and 384/748 (51.3%) died during 9-month follow-up. Common predictors for increased mortality in both populations were higher Medical Research Council (MRC) disease severity grade and lower cerebrospinal fluid lymphocyte cells count. In HIV-uninfected subjects, older age, previous tuberculosis, not receiving adjunctive dexamethasone, and focal neurological signs were additional risk factors; in HIV-infected subjects, lower weight, lower peripheral blood CD4 cell count, and abnormal plasma sodium were additional risk factors. The areas under the receiver operating characteristic curves (AUCs) for the final prognostic models were 0.77 (HIV-uninfected population) and 0.78 (HIV-infected population), demonstrating markedly better discrimination than the MRC grade (AUC 0.66 and 0.70) or the Glasgow Coma Score (AUC 0.68 and 0.71) alone. Conclusions: The developed models showed good performance and could be used in clinical practice to assist doctors in identifying TBM patients at high risk of death and at increased need of supportive care. Hide abstract