In recent months, news came out about several distressing cases that have underscored the risks of gene therapy, which delivers DNA to a patient’s cells to treat disease. All of these cases have one thing in common that’s not apparent at first glance.

A child in China, enrolled in a clinical trial for gene therapy, died of acute respiratory distress syndrome, in which fluid builds up in the lungs and oxygen cannot enter. In a separate clinical trial in China, a child died after clots formed in their small blood vessels, likely disrupting the delivery of oxygen to organs such as the kidneys and brain. And earlier this year, researchers reported that a 5-year-old boy treated in Philadelphia had a brain tumor that was likely caused by gene therapy. (He’s doing okay now.) In each of these cases, the gene therapies relied on a versatile but unpredictable type of virus: adeno-associated virus, or AAV for short.

AAVs are the workhorses of many of the most successful gene therapies in the world. Scientists modify them so that they can deliver a working copy of particular genes to the body, but cannot replicate and aren’t infectious. These treatments have transformed a form of spinal muscular atrophy—once the genetic disorder most responsible for infant mortality—into a survivable condition, and have been approved for a devastating enzyme deficiency, muscle-wasting disease, and hemophilia, to name a few.

The terrible outcomes AAVs can cause are rare, but they’re a real risk. So far, the treatments have been used in small patient groups—the most commonly given one has been received by about 5,000 children since 2019—and focused on diseases that dramatically affect or shorten people’s lives. But AAV-based gene therapies are being explored for prevalent ailments, including heart disease, high blood pressure, and perhaps even Alzheimer’s. As the use of gene therapy increases, so might the instances in which it goes dramatically wrong.

The modified AAVs involved in gene therapy each contain a single strand of DNA. One portion of that strand drives the virus’s ability to get inside cells, and another is the helpful human gene that’s being delivered. The viral components are more likely than the therapeutic gene to be unpredictable and can cause side effects. On occasion, for instance, instead of hanging around within cells, AAVs insert themselves into the patient’s genetic code. This seems to have been what happened to the 5-year-old who developed a brain tumor—the AAV put itself into a part of his DNA that then produced the cancer.

Hurler syndrome, the condition the boy was being treated for when he received the gene therapy at 13 months old, is progressively debilitating. The syndrome can cause problems for the skeleton, heart, lungs, and liver, and affect cognitive development; life expectancy is short. Many gene therapies are geared toward children for whom a working copy of a vital gene would promise something closer to a healthy life. “We should remember what diseases we are talking about,” Albert Rizvanov, a scientist at Kazan Federal University in Russia who studies gene therapy, told me.

For such patients, gene therapy may seem like an obvious choice. But parents are still left weighing the potential benefits of treatment against the real risks, including fatal complications.

Jo Kaur and her husband, Richie DiGeorge, faced this question not too long ago. Their son Riaan was born in 2019 weighing less than six pounds, and by the time he was three months of age, they were concerned about his eyesight. Riaan’s pediatrician suspected it was just near-sightedness, Kaur, who has written extensively online about her family’s story, told me, but her “mom radar went off.”She secured an appointment with an ophthalmologist who found that her son had cataracts in both his eyes. Riaan’s overall growth also wasn’t on track, and he was missing milestones. At 15 months old, he was diagnosed with a severe form of Cockayne syndrome, a genetic condition that causes progressive muscle weakness and nerve damage.

This diagnosis meant not just that Riaan had problems with his eyes or was growing a little slowly, but that his life would likely be short—perhaps just five to seven years. He could roll and pull his family in for cuddles, but by the time he was 6 years old, at the end of 2025, he was still unable to sit independently for more than a few minutes at a time, or stand or walk independently, or feed himself. He was so small that he was still wearing clothes for 18-month-olds.

By then, Riaan had lost a lot of nervous tissue. His brain size was about 50 percent of what would be typical for a child his age. But Kaur and DiGeorge had worked closely with scientists for years to develop a gene therapy for Cockayne syndrome. The hope was to stop or at least slow Riaan’s loss of nerve cells. “Because Cockayne syndrome is neurodegenerative, even an extended period in which Riaan simply remains where he is now would be enormously meaningful to us,” Kaur said.

The researchers developing a Cockayne-syndrome treatment for Riaan had done mouse experiments to determine whether the medication was safe and whether it worked. Some of the animals had signs of liver damage, but they also lived longer. Kaur and her husband had been involved in countless meetings and stayed up-to-date on all of the developments. Knowing the risks so intimately increased the weight of the decision to allow Riaan to receive the treatment. “We’d read all the papers, we knew the science,” she said, “but you’re still making a huge decision for your child.”She and her husband were primarily concerned about the treatment’s safety.

Ultimately, this past April, they decided to go ahead with the therapy. Riaan would be the first child in the world to receive this treatment for Cockayne syndrome. Surgeons created a small hole in his skull, inserted a catheter, and within minutes delivered the gene therapy to him. When night came, as Kaur wrote in her own account of Riaan’s procedure, she fought the urge to sneeze as he slept on her arm in the hospital. Everything seemed very delicate, and she dared not move.

His head had to stay elevated and he “had so many wires,” she told me. “I didn’t want to wake him.”The first successful gene therapy, tested back in 1990, used a nonreplicating version of a mouse virus. But that virus doesn’t readily get into all cells, so scientists kept looking for others to try. One was adenovirus—an attractive option in part because it could deliver working genes to a high number of cells.

But in 1999, an 18-year-old named Jesse Gelsinger died shortly after having a severe immunological response to a therapy that used adenovirus to treat his rare metabolic disorder. Gelsinger is believed to have experienced the same acute respiratory distress syndrome that killed one of the patients in China.

Currently, Riaan is being tapered off his immunosuppressants and monitored closely. In the meantime, Kaur is observing signs that the therapy might be helping. He can sit by himself on the edge of the couch. He still needs support to stand or walk, but he wants to try more often. Kaur knows these observations are anecdotal, but Riaan’s physical therapist has also noticed the changes, she said. He has become more expressive, too—an outcome that surprised Kaur. “He would rarely say ‘mama,’ and now he says it all day long … ‘ma-ma-ma,’” she says. “It’s beautiful to hear.”