“We cannot win the battle against AIDS if we do not also fight TB”, Nelson Mandela has famously said.  His insistence on the importance of TB is understandable, when it is remembered that 82% of all cases of TB-HIV co-infection are concentrated in sub-Saharan Africa, and many Africans infected with HIV die from TB (STI/ Kaireh & Dao; STI/ Ogbuagu).  On an individual level, the treatment of co-infection has seen impressive results (STI/ Ogbuagu).   Against the contemporary background of recent demonstrations that ART can also have a preventative effect at population level, Mandela’s remarks raise the question – given the independent but interlinked nature of the two epidemics – whether the timely treatment of TB could not also have a population-level effect, thus allowing us to tackle the co-infection epidemic from the TB as well as the HIV end.

The recently reported Thibela randomized control study of isoniazid therapy among South African gold miners (Churchyard & Grant) has investigated whether comprehensive screening, followed by treatment of either active or latent TB within a community could serve, by itself, to disrupt TB transmission, given the high level in HIV in this population.  There has been evidence of the limited durability of the efficacy of isoniazid preventative therapy, which is not encouraging (Johnson & Whalen).  However, a 1960s trial in Alaska appeared to demonstrate a 55% decline in infection over five years (Comstock & Snider).  Besides, Churchyard & Grant’s population of 78,744 South African gold miners in 15 settlements  seem ideally suited for the intervention, given they live in hostels and mix predominantly with those from the same mine.

The study implemented screening and therapy over nine months in the intervention clusters, then measured relative TB incidence as compared with the control clusters over the 12 months after the intervention; it also measured TB prevalence at study completion.  In addition to the effectiveness of the intervention in disrupting TB transmission over the entire communities, the researchers analysed the efficacy of the isoniazid therapy in relation to a group of 10,909 miners surveyed at baseline.

It was found that the intervention had no significant population-level effect on incidence (3.02 cases per 1000 person years in the intervention, as against 2.95 in the control), or prevalence (2.35% vv. 2.14%).  This was despite the isoniazid therapy being demonstrated to have a substantial direct impact on TB incidence in the course of the intervention (1.10 cases per 100 person years in the intervention clusters, as against 2.91 cases in the control).  Though participation and retention was variable across the clusters, even where they were excellent, there was no significant improvement in these results.

The lack of epidemiological effectiveness could have been the result of post-treatment re-infection, but also of the very high prevalence of HIV (as well, perhaps, of silicosis) among the miners.  The latter factor represents the most obvious difference between this study population and that of 1960s Alaska.  As regards the potential effectiveness of community-wide isoniazid interventions to interrupt TB epidemics in sub-Saharan Africa, the study offers little grounds for optimism.  It recommends expanding strategies for controlling HIV and developing systems that minimize the time from receiving a positive TB result until treatment.  But also points to an interesting potential role for mathematical modelling to identify strategies for dealing with the TB-HIV co-epidemic.