HIV and antiretroviral treatment as drivers of rifampicin-resistant tuberculosis in South Africa: insights from mathematical modelling
Johnson, L. F.; Fenner, L.; Naidoo, P.; Kubjane, M.; Evans, D.; Omar, S. V.; Zinyakatira, N.; Cox, H.
Show abstract
BackgroundHigh levels of multidrug and rifampicin resistant tuberculosis (MDR/RR-TB) are a global concern, although they have declined over the last decade. TB patients are at increased risk of acquired rifampicin resistance if they have HIV coinfection, especially at low CD4 counts, but this dynamic has not previously been modelled. MethodsWe extended a previously-developed model that simulates HIV and TB in South African adults, to include the acquisition and transmission of rifampicin resistance. In line with systematic reviews, the risk of acquiring RR with TB treatment is modelled as being negatively associated with patients CD4 counts. We allow for temporal changes in drug susceptibility testing, both before treatment initiation and at treatment failure, as well as other changes in TB prevention and treatment. The model is calibrated to data from national TB drug-resistance surveys, and recorded numbers of MDR/RR-TB laboratory diagnoses and patients initiating second-line TB treatment, using a Bayesian approach. ResultsThe model estimates that the proportion of South African TB patients with rifampicin resistance at diagnosis increased from 2.0% (95% CI: 1.7-2.3%) in 1986 to 5.9% (5.2-6.9%) in 2013, in line with survey data. In the absence of HIV, the prevalence of MDR/RR-TB would have increased to 4.1% (2.7-5.1%) in 2013, suggesting a third of rifampicin resistance in 2013 was attributable to HIV. In the absence of antiretroviral treatment (ART), the prevalence of rifampicin resistance would have been higher (6.5% [5.6-7.6%] in 2013, rising to 6.9% [5.7-8.2%] in 2019). ART reduced the prevalence of rifampicin resistance in 2019 by 17%. ConclusionsIn countries with high HIV prevalence, HIV may be a major driver of rifampicin resistance in people with TB. ART programmes have the potential to reduce the emergence of resistance substantially.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Treatment preferences among people at risk of developing tuberculosis: a discrete choice experiment 94%
- The Role of Modelling and Analytics in South African COVID-19 Planning and Budgeting 93%
- Evaluating the health impact, health-system costs and cost-effectiveness of using TrueNat on stool samples compared to usual care for the diagnosis of paediatric tuberculosis in primary care settings: a modelling analysis 93%
Similar papers in this journal
- Modelling the impact of tuberculosis preventive therapy: the importance of disease progression assumptions 95%
- Modelling testing and response strategies for COVID-19 outbreaks in remote Australian Aboriginal communities 93%
- Preventing a cluster from becoming a new wave in settings with zero community COVID-19 cases 92%
Similar papers in this journal
Similar papers in this journal
- Using genetic data to identify transmission risk factors: statistical assessment and application to tuberculosis transmission 95%
- Quantifying epidemiological drivers of gambiense human African Trypanosomiasis across the Democratic Republic of Congo 94%
- A model of tuberculosis clustering in low incidence countries reveals more on-going transmission in the United Kingdom than the Netherlands between 2010 and 2015. 94%
Similar papers in this journal
- Programmatic diagnostic accuracy and clinical utility of Xpert MTB/XDR in patients with rifampicin-resistant tuberculosis in Georgia 92%
- Understanding non-adherence to tuberculosis medications in India using urine drug metabolite testing: a cohort study 92%
- Potential Utility of C-reactive Protein for Tuberculosis Risk Stratification among Patients with Non-Meningitic Symptoms at HIV Diagnosis in Low- and Middle-Income Countries 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.