TB clusters show where HIV treatment is missing in South Africa

Clusters of new cases of tuberculosis (TB) provide strong evidence of areas that need to be prioritised for improved antiretroviral therapy (ART) coverage, and improved ART coverage is associated with a reduction in new TB cases, according to research carried out in KwaZulu-Natal province, South Africa, published in July in Nature Scientific Reports.

TB is the single largest cause of death in South Africa and people living with HIV are at especially high risk of developing active TB. Antiretroviral treatment reduces the risk of developing TB as it restores the immune system.

The HIV epidemic in southern Africa has greatly increased the burden of TB in the region since the early 1990s, and although antiretroviral treatment availability is correlated with a reduction in TB cases, information is still lacking on how best to target TB control efforts.


mathematical models

A range of complex mathematical techniques which aim to simulate a sequence of likely future events, in order to estimate the impact of a health intervention or the spread of an infection.


Affecting a specific body site, organ or system.

active TB

Active disease caused by Mycobacterium tuberculosis, as evidenced by a confirmatory culture, or, in the absence of culture, suggestive clinical symptoms.

immune system

The body's mechanisms for fighting infections and eradicating dysfunctional cells.

multivariate analysis

An extension of multivariable analysis that is used to model two or more outcomes at the same time.

Although a recent modelling study found a strong correlation between antiretroviral coverage and reductions in TB incidence, TB does not appear to be evenly distributed in the community. TB case reporting from South Africa suggests that the areas of highest TB incidence do not always match the areas of highest HIV prevalence.

Researchers at the Africa Health Research Institute in KwaZulu-Natal investigated the relationship between the geographical distribution of TB and antiretroviral coverage in a northern rural district of the province. Using data on the health of the population gathered systematically since 2000, they mapped the distribution of self-reported cases of TB over time to identify clustering of cases in time and space.

Mapping these clusters allowed the researchers to identify likely focal points of transmission in the district, to see if these locations changed over time and to assess the geographical and demographic factors – including HIV and ART coverage – associated with the distribution of TB in the district.

The district had an extremely high TB case notification rate. In 2008, case notifications peaked at 1773 cases per 100,000 inhabitants and declined to 756 cases per 100,000 in 2014. In comparison, the highest TB notification rate in the United Kingdom in 2018 was 21 cases per 100,000 in London.

The study population comprised all adults aged 15 and over who answered at least one individual health questionnaire between 2009 and 2015 (41,812 persons, approximately 60% women). TB incidence was assessed by self-report and the location of the respondent was tagged by geolocation.

The researchers found that an average of 3.2% of respondents per year had recently diagnosed TB, although TB prevalence began to decline after 2011. In comparison, South African national research found a self-reported TB rate for the entire country of 0.6% in 2012.

Cases were clustered in time and space into nine clusters, and the four largest clusters were grouped in a periurban area adjacent to a major highway in the south-east of the district. Three clusters persisted throughout the study period, indicating that a high number of active cases were present in this area.

HIV prevalence varied from 22% in 2009 to 25% in 2012. Just over half (52%) of people with HIV were on antiretroviral treatment in 2012.

Multivariate modelling that controlled for household income, age and other demographic factors, and antiretroviral treatment showed that for every 1% increase in ART coverage in communities within the district, the odds of being recently diagnosed with TB fell by 2% (adjusted odds ratio 0.98, 95% CI 0.97-0.99).

Although the study relied on self-report of TB diagnosis, the researchers say that the clustering of TB cases is consistent with the pattern of drug-resistant TB cases in the district, which are reliant on laboratory confirmation.

The researchers conclude that their findings support the development of “precision public health strategies” which use localised data to identify communities that may benefit from intensified efforts to diagnose HIV and TB, initiate people on treatment and increase the proportion of people with HIV who have fully suppressed viral load.


Tomita A et al. Space-time clustering of recently-diagnosed tuberculosis and impact of ART scale-up: evidence from a hyper-endemic rural South African population. Nature Scientific Reports, 9: 10724, 2019. https://doi.org/10.1038/s41598-019-46455-7