TSAKANE, South Africa — When Tsholofelo Msimango joined a clinical trial of new tuberculosis drugs, she was 20 years old and weighed just 57 pounds.
Stricken with a deadly strain of tuberculosis, Ms. Msimango was terrified. She was to be transferred to Sizwe Tropical Diseases Hospital in Johannesburg. The nurses in the township here told her that it was far away and infested with vervet monkeys.
“I cried the whole way in the ambulance,” she recently recalled. “They said I would live with monkeys and the sisters there were not nice and the food was bad and there was no way I would come back. They told my parents to fix the insurance because I would die.”
Five years later, Ms. Msimango, who became the trial’s fifth enrolled patient, is now tuberculosis-free. She is healthy at 103 pounds, and has a young son.
The trial she joined, nicknamed Nix-TB, is small — it has enrolled only 109 patients. But experts are already calling the preliminary results groundbreaking. The three-drug regimen tested on Ms. Msimango has shown a 90 percent success rate against a deadly plague: extensively drug-resistant tuberculosis, better known as XDR-TB.
On Wednesday, the Food and Drug Administration effectively endorsed the Nix-TB regimen by approving the last of its three drugs, pretomanid, for XDR-TB patients as long as it is used with the other two previously approved ones, bedaquiline and linezolid.
Usually, the World Health Organization eventually adopts approvals made by the F.D.A. or its European counterpart, meaning the regimen could soon come into use worldwide.
Tuberculosis has now surpassed AIDS as the world’s leading infectious cause of death, and the XDR strain is the ultimate in lethality. It is resistant to all four families of antibiotics typically used to fight the disease.
Only a tiny fraction of the 10 million people infected by TB each year get the XDR strain, but very few of those who do survive it.
There are about 30,000 cases in over 100 countries. Three-quarters of those patients die before they even receive a diagnosis, experts believe, and among those who get typical treatment, the cure rate is only 34 percent.
The treatment itself is extraordinarily difficult. A typical regimen in South Africa requires up to 40 daily pills, taken for up to two years. Other countries rely on even older regimens that include daily injections of antibiotics that can have devastating side effects, including deafness, kidney failure and psychosis.
But the new regimen tested in Nix-TB requires only five pills a day for a mere six months. Its three drugs are not without side effects, but most are manageable, according to the doctors running the trial.
Linezolid was approved in 2000 and bedaquiline in 2012. Some drug-access groups had opposed pretomanid’s approval, saying it needed further testing. Other TB experts argued that the situation is still so desperate that risks had to be taken.
Dr. Gerald Friedland, one of the discoverers of XDR-TB and now an emeritus professor at Yale’s medical school, called Nix “a wonderful trial” that could revolutionize treatment: “If this works as well as it seems to, we need to do this now.”
A killer appears
News that tuberculosis had evolved a terrifying new strain first broke in 2006, when doctors at a global AIDS conference learned of a doomed group of tuberculosis patients in Tugela Ferry, a rural South African town.
Of the 53 patients in whom the strain had been detected, 52 were dead — most within a month of diagnosis. They were relatively young: The median age was 35.
Many of them had never been treated for TB before, meaning they had caught the drug-resistant strain from others and had not developed it by failing to take their drugs. Several were health workers who were assumed to have caught it from patients.
Within months, South Africa realized it had cases of the deadly infection in 40 hospitals. Alarmed, World Health Organization officials called for worldwide testing.
The results showed that 28 countries, including the United States, had the deadly strain, XDR-TB, and that two-thirds of the cases were in China, India and Russia. It took far longer to determine how widespread it was in Africa, because most countries there could not do the sophisticated testing.
H.I.V., the virus that causes AIDS, helped drive the epidemic. Anyone infected with it is 25 times as likely to get TB, according to the W.H.O. But many victims, including Ms. Msimango, catch this type of TB without ever having H.I.V.
In the early years, XDR-TB was a death sentence. Doctors tried every drug they could think of, from those used to treat leprosy to those for urinary tract infections.
“From 2007 to 2014, we threw the kitchen sink at it,’’ said Dr. Francesca Conradie, a researcher at the University of the Witwatersrand, Johannesburg, and director of the Nix trial.
The death rate was about 80 percent. Sometimes the drugs killed patients. In other cases, patients died because they could not tolerate the drugs and stopped taking them.
Tuberculosis germs burrow deep into the lungs and barricade themselves inside clumps of dead cells. Breaking those nodules apart and killing all the bacteria inside requires taking drugs for months.
Nearly all antibiotics cause nausea and diarrhea. But some, especially the injections, are far tougher. “I had a patient who tried to commit suicide because of the tinnitus — ringing in his ears,” said Dr. Pauline Howell, a tuberculosis researcher who runs the Nix trial at Sizwe.
“Some get hallucinations,” Dr. Howell said. “I had one patient who tried to cut open his skin because he thought bugs were crawling under it. I’ve had patients who became wheelchair-bound because of the vertigo.”
Patients can go deaf over a weekend, she said, or the nerves in their feet and hands can wither until they can no longer walk or cook. Some develop disrupted cardiac rhythms, which can trigger heart attacks.
Ms. Msimango, too, veered close to death because the drugs were too much for her.
When she was 19, she said, she caught drug-resistant TB from another young woman — the temporarily homeless daughter of a friend of her mother.
Her mother had generously taken in the young woman and had told her daughter to share her bed, a common arrangement in the tiny houses of townships like Tsakane. “A few weeks after she left, I started coughing,” Ms. Msimango said.
“She had not told us that she had drug-resistant TB and had defaulted,” she added, using a common term for dropping out of treatment.
At first, Ms. Msimango got her injections at a hospital and took her pills under her mother’s watchful eye. But they made her feel so awful that she secretly spat them out, stuffing them between the sofa cushions when her mother wasn’t looking.
After she defaulted twice herself, she was transferred to Sizwe, terrified that she would die alone.
No masks, no doctors
Although it lies in South Africa’s largest city, Sizwe does host monkeys, along with feral peacocks and the occasional mongoose. It has long been on the front lines of South Africa’s protracted battle against tuberculosis.
The hospital sits on an isolated hilltop 10 miles from downtown. The British built it in 1895 as Rietfontein Hospital to house victims of contagious diseases like leprosy, smallpox and syphilis.
Although Sizwe still has its 19th-century style — the grounds are a jigsaw puzzle of tin-roofed, one-story barracks — the TB explosion of the last 25 years has forced changes.
In 1996, with the AIDS epidemic raging, the W.H.O. announced that South Africa had the world’s worst tuberculosis epidemic: 350 cases per 100,000 citizens.
That finding shocked the government, which had been using outdated approaches against the disease. Diagnoses were made by X-ray, which are less accurate than sputum tests, and doctors hospitalized every patient.
I visited shortly after the 1996 announcement, and the situation inside Rietfontein was unnerving. In the men’s ward, dozens of TB patients lay on cots only two feet apart; those with drug-resistant strains slept next to patients with more common strains. No one wore a mask, and no doctor was on duty.
During the day, the transom windows were open to let in fresh air. But at night, the patients shut them and turned up the heat. (Johannesburg sits at a higher elevation than Denver, so nights get chilly.) The room became a deadly incubator.
A return visit this month showed that far more than the name had changed (Sizwe means “nation” in Zulu).
The former men’s ward is now a mostly empty meeting hall. Patients with TB that is not drug-resistant are treated at home, and even those with partially drug-resistant strains are usually hospitalized only briefly.
The XDR-TB patients rest in a ward atop the hill, and golf carts transport those too weak to walk. Each patient has a separate room and bathroom, hookups for oxygen and lung suction, a TV and big windows and a door to the lawn outside.
The building has a negative-pressure air system, but it keeps breaking down, so the policy is to keep all the doors and windows open as much as possible, said Dr. Rianna Louw, the hospital’s chief executive.
Patients can work in the garden, play pool or foosball, and take classes in sewing, beading or other crafts that might help them earn a living when they get out.
But the months of isolation needed for treatment can be tough. “Our children are scattered, they are falling apart!” a patient who gave her name only as Samantha shouted at a group-therapy session that turned into an airing of grievances.
“The father of my kids is in prison,” she said. “My firstborn son is arrested for robbing people in the street. That would not happen if I was home!”
The counselor interrupted to say: “We understand your frustration. But if we discharge you, we are taking a risk. You are not healthy. You can still expose people to your disease. That’s why you will stay a minimum of four months.”
A rush to approval?
The regimen successfully tested at Sizwe is called BPaL, shorthand for the three drugs it comprises: bedaquiline, pretomanid and linezolid.
The BPaL regimen is “bold, because it’s three killer drugs instead of two killers plus some supportive ones,” Dr. Howell said.
Most regimens, she explained, rely on two harsh drugs that can destroy bacterial walls and include others that have fewer side effects but only stop TB bacteria from multiplying.
But even the new treatment poses hazards. Short-term use of linezolid against severe hospital infections causes few problems, but use for many weeks against TB can kill nerves in the feet, making it hard to walk, or can suppress the bone marrow where blood cells are made. (To find the ideal linezolid dose, the Nix investigators have started a new trial, ZeNix.)
The F.D.A. approved bedaquiline in 2012 for use against multi-drug resistant TB (the XDR strain is an even deadlier subset), and in 2015 the W.H.O. followed suit.
Until Wednesday, pretomanid was in dispute, although in June an F.D.A. advisory committee voted 14 to 4 to approve it.
Some advocacy groups argued at the time that the drug had been too little tested.
“Pretomanid looks like a promising drug, but it’s being rushed forward, and we don’t want to see the F.D.A. lower the bar for approval,” Lindsay McKenna, co-director of the tuberculosis project at the Treatment Action Group, an advocacy organization, said in July.
Her organization and others had asked the F.D.A. to first demand two of the most rigorous tests: a head-to-head comparison of pretomanid against delamanid, a related drug, and a randomized clinical trial in thousands of patients of the Nix regimen versus another used by the W.H.O.
Delamanid also works against XDR-TB but has serious disadvantages, Dr. Howell said. It requires two daily doses, and they must be taken an hour apart from any other drug.
Pretomanid — which is named for Pretoria — is not owned by a drug company but by the TB Alliance, a nonprofit based in New York that is seeking new treatments.
Dr. Mel Spigelman, the alliance’s president, had argued that a full clinical trial would be both impractical and unethical.
“Put yourself in a patient’s position,” he said. “Offered a choice between three drugs with a 90 percent cure rate, and 20 or more with less chance of cure — who would consent to be randomized?”
Such a trial would cost $30 million and take five more years, he added: “That’s a very poor use of scarce resources.”
Dr. Mario C. Raviglione, who ran the W.H.O.’s tuberculosis programs for 14 years and is on the TB Alliance’s board, said he had found the opposition from activist groups “disturbing.”
“I’m used to public calls by activists for rapid approval to provide new options,” he said. “I’m not used to calls against a new drug. It’s like we’re forgetting that we’re facing a lethal disease.”
‘There is no survival here’
Innocent Makamu, 32, was facing two years in the hospital when he chose to join the Nix trial in 2017.
Like Ms. Msimango, he also had caught drug-resistant TB from a roommate. A plumber, he had shared a room at a distant construction site with a carpenter.
“He was too much on the bottle,” Mr. Makamu said. “He kept defaulting.”
Soon afterward, he began feeling tired and lost his appetite. Doctors at the hospital near his home diagnosed tuberculosis, and put him on 29 daily pills and a daily injection.
“It was deep in my bum,” he said. “I couldn’t sit properly. It hurt every day.”
At the hospital, he watched two other inpatients wither and die because they could not stick to the regimen. “I thought, ‘Oh, there is no survival here.’”
Then further tests showed that he had full-blown XDR-TB. He was transferred to Sizwe and offered a spot in the Nix trial.
Some patients there who were on the standard 40-pill regimens discouraged him. “They said, ‘They are using you as guinea pigs,’” he said. “Even the nurses thought that.”
But he found the possibility of taking only five pills for six months very tempting, and so he volunteered. Within a month, he could tell it was working.
“Then the patients who called us ‘guinea pigs’ — they wished they had taken the research pills,” he said.