Epidemic model shows how one thing leads to another

18 December 2001 16:30 EST

http://news.bmn.com/news/story?day=011219&story=2

by Apoorva Mandavilli, BioMedNet News

                        A mathematical model that predicts disease outbreaks suggests that the AIDS epidemic can double the number and frequency of tuberculosis (TB) outbreaks, report researchers from the University of California in Los Angeles. The results renew a controversial debate on the World Health Organization's targets for treating TB in developing countries.

While several empirical studies have collected statistics of HIV's impact on TB, the probabilistic model uses 1.3 million computer simulations to predict and quantify outbreak sizes, at the level of a population, over a 2-year period. "We modeled how HIV speeds up TB's progression from infection to full-blown disease," said Sally Blower, principal investigator and professor of biomathematics at the UCLA AIDS Institute. "This telescoping effect is very important, because only people in the disease stage of TB can transmit it to others."

Currently almost a third of the world is infected with TB, according to the World Health Organization (WHO). Under normal circumstances, only 10% of those would advance to the contagious stage of the disease, but an immune system weakened by HIV permits Mycobacterium tuberculosis to flourish and overwhelm the lungs. On an individual level, therefore, HIV infection speeds up TB progression from a latent stage to an active stage, in which it becomes contagious. In a population, Blower says, "this affects the transmission rates and you can get more and more severe outbreaks."

The mathematical model's predictions are consistent with what public health officials observed during TB's resurgence in the US between 1985-92, says Kenneth Castro, director of TB elimination at the US Centers for Disease Control and Prevention. HIV causes a very rapid progression of TB from latent to active, Castro says: "From that biological perspective, it makes sense and has been observed in a real setting."                      

If tuberculosis treatment rates are only mild or moderate, as is the case in developing countries, Blower says, HIV can nearly double TB's impact. In such cases, the TB epidemic becomes "a fire raging out of control," Blower told BioMedNet News. In contrast, in cities like San Francisco, where 95% of TB cases are cured, outbreaks are like small, isolated fires contained by firefighters. The results are published today in the Journal of AIDS.                      

                       The effects of HIV on TB transmission can be "very, very dramatic," agreed Richard Hafner, medical officer at the US National Institute of Allergy and Infectious Diseases. TB was somewhat in control in some developing countries, but "when you superimpose HIV, it's now out of control in many of those countries," he said.

                       World health professionals predict that more than one-quarter of the AIDS patients in the world will die from TB this year. But the WHO's calculations do not reflect HIV's amplifying effect on TB, the researchers contend. Instead, the WHO grossly underestimates its target treatment levels for TB.

                       The WHO should probably raise its targets, agrees Hafner. But in countries lacking the appropriate infrastructure, higher rates simply may not be feasible, he says

                      Setting target rates for TB treatment is a "hotly debated topic," Castro told BioMedNet News, adding that he agrees the target rates may be too low. Recently, Castro says, the WHO has begun to consider the importance of HIV in managing TB infections worldwide (WHO officials could not be reached for comment.)

                       Treating latent infections in countries lacking infrastructure may only add to the problem of drug-resistant TB, Castro warns. In contrast to the authors, who advocate aggressively treating both HIV and TB, Castro favors a more "stepwise approach" which places a higher priority on active infections than on latent ones.