The journey to Dr. Moll’s terrifying discovery began in early 2005, when he noticed something peculiar. The staff at his hospital had become accustomed to the marvellous “Lazarus effect” of anti-retroviral treatment for AIDS: seeing desperately sick people quickly start gaining weight and return home or go back to work. But now, in his ward, he had two men in their 30s on ARVs whose HIV infections were suppressed to undetectable levels. Yet their TB, which would normally have cleared up in a matter of weeks, kept getting worse.

He suspected multidrug-resistant TB, or MDR, believed at the time to be as bad as the disease could get. So he collected sputum from 45 patients and sent it off to a lab in Durban for cell culturing. (The only way to tell if a TB strain is drug-resistant is to grow cultures from a patient sample, zap it with the different drugs and see which, if any, fail to kill it.) The process takes six to eight weeks. “In that time, we more or less forgot about it,” Dr. Moll said. One of his two young men died.

But the phone call from the lab, when it eventually came, slammed the issue to the top of their agenda: Of the 45 samples, 10 were indeed drug-resistant. But they weren’t resistant to just one or two of the drugs used against TB. They were resistant to all six medications available for use in Tugela Ferry. In other words, there was nothing to cure that TB at all.