According to a study based on a detailed mathematical model, tuberculosis treatment options that would allow for shorter courses of treatment would lead to significant survival gains and decreases in tuberculosis incidence in areas with high HIV prevalence. Benefits would be increased still further with increased tuberculosis case detection and cure rates. The study was reported in the May 11th issue of AIDS.
Tuberculosis (TB) is an enormous health problem worldwide, with over 9 million new cases in 2006, and current figures suggest that progress against the disease may be slowing. Current treatment options are one of the major challenges in fighting the disease, as treatment durations are six months or longer, with a high risk of treatment failure and drug resistance if not completed. HIV has also contributed to the TB epidemic, as HIV-positive individuals are more likely to become infected with TB, to progress to active disease, relapse, and die if not adequately treated.
Due to "considerable interest in the development of shorter drug regimens," an international research team recently developed a mathematical model to estimate the effects that such shorter treatment regimens would have on the TB epidemic. They found that the introduction of two-month treatment regimens could lead to dramatic reductions in TB incidence and mortality, doubling or tripling current rates of decline if shorter regimens were accompanied by enhanced case detection (Salomon 2006).
Now, in a study reported in AIDS, the same research team extended the previous model into a "detailed mathematical and computational framework" to analyse the potential impact of three factors – reduced treatment duration, enhanced TB case detection, and cure rate – on TB epidemiology.
This model also adjusted for HIV prevalence levels and five clinical categories (WHO stages I–IV, and HIV-uninfected). Benefits were assessed in terms of absolute and relative reductions in new TB cases and deaths over the long term, to the year 2030. The model was calibrated using long-term TB and HIV epidemiological data from the years 1980–2004 in Kenya – a sub-Saharan African country with high prevalence rates of both infections, and a strong disease surveillance program.
The baseline scenario followed present-day Kenyan conditions, with TB treatment under directly observed treatment, short-term (DOTS) programmes lasting six months, and under non-DOTS programs lasting eight months. Rates of treatment discontinuation were set at observed default rates. An HIV prevalence rate of 7% was assumed to continue throughout the period from 2006 to 2030. To generalise the findings to other settings, the simulation was also run at HIV prevalence rates of 3%, 15%, 20%, and 35%. (Prevalence rates were assumed constant over the projected period.)
In the first alternate scenario, treatment duration was assumed to be reduced to two months in both DOTS and non-DOTS programs for all TB cases. Discontinuation rates were assumed to be steady over the course of the two-month duration. Probability of cure was assumed equal to the probability of cure under six- or eight-month current standard treatment.
The second alternate scenario was the same as the first alternate, plus a doubled rate of TB case detection across all HIV categories.
A third scenario was the same as the first, plus a substantial increase in cure rate: 80% of patients susceptible to relapse under the baseline scenario were now considered cured. A fourth and final scenario included all three improvements – shortened treatment, increased case detection, and increased cure rate.
The simulations of present-day conditions (assuming a continued HIV prevalence of 7%) showed decreasing TB incidence and mortality rates throughout the projected 25-year period. The simulated reduction in TB treatment duration from six to two months reduced TB incidence and deaths by an additional 15–20% over the 25 years. When combined with increased detection and cure rates, there was an 80% reduction in incidence and mortality.
HIV prevalence rates and clinical stage distribution had a strong effect on the modelled projections: as expected, the number of new TB cases and deaths increased substantially at higher HIV prevalence levels.
While the benefits in terms of absolute numbers were therefore correspondingly higher at higher HIV prevalence rates, the relative benefits from shortened duration alone became progressively smaller with rising HIV prevalence rates.
At extreme HIV prevalence levels (35%), the benefits of reducing treatment duration became unclear and might even be reversed.
Overall, compared with present-day strategies, the model showed a 6% to 20% decrease in TB incidence and mortality in 25 years by reducing TB treatment duration (without other improvements) in areas with HIV prevalence ranging from 3–20%.
As with any mathematical model, these interpretations rest on many assumptions and may neglect other significant factors. However, the researchers believe that "reducing tuberculosis treatment duration, alone or in combination with other control strategies, [has the potential to] provide enormous benefits [in areas with] high HIV prevalence." They also note that "tuberculosis control policies need to account for HIV levels because the efficacy of different interventions varies substantially with HIV prevalence."
References:
Sánchez MS et al. Impact of HIV on novel therapies for tuberculosis control. AIDS 22: 963-972, 2008.
Salomon JA et al. Prospects for advancing tuberculosis control efforts through novel therapies. PLoS Medicine 3: e273, 2008.