An experiment has been performed that fits with the ideas in the previous slide. In a small clinical trial (about 60 HIV-positive tuberculosis patients) treatment was with the standard therapy for 2 months to suppress disease. The patients were then separated into two groups to test the idea that rifapentine, a long-lived derivative of rifampicin, could be used less frequently than rifampicin. In each group the number of relapses was similar at about 10%. But the number of rifampicin-resistant cases was quite different: 0% with rifampicin, 80% with rifapentine (16).