When Science Goes Wrong (21 page)
Three days before Christmas, Paul arrived home to find Jesse vomiting in the living room. He was admitted to the hospital, where tests revealed that his blood ammonia levels were six times higher than normal. His metabolism was falling into a vicious cycle whereby, having consumed all the fat in his body, it was now digesting his proteins, thus liberating even more ammonia. Six days later, after a rocky course, he became delirious. Thinking that he was near death, Jesse’s doctors asked Paul whether he wanted a ‘do not resuscitate’ order placed on his son. Paul vehemently refused. The doctors then decided to move Jesse into intensive care, but before they could do so Jesse stopped breathing. Luckily, Paul was present at his bedside: he summoned a doctor who called a code blue. Jesse was intubated and put on a respirator.
A new and more powerful medicine – sodium phenylbutyrate – was fed to Jesse via a stomach tube, and it eventually lowered his ammonia level to the point where he regained consciousness and began a complete recovery. He returned to school with ammonia levels near normal for the first time in his life and with a newfound resolve to follow his doctors’ orders. A few months later, he graduated from high school.
On June 18, 1999 – Jesse’s 18th birthday – the entire family flew east. After a few days’ visit with relatives in New Jersey, they drove down to Philadelphia so that Jesse could sign up for the OTC trial. The person who explained the trial and walked Jesse and Paul through the ‘informed consent form’ was Steven Raper, the liver surgeon.
Raper said that the adenoviral vector would be infused directly into Jesse’s hepatic artery – the artery that supplies the liver. The idea behind this was that most or all of the viral particles would be taken up by the liver, where the new gene was needed, and the rest of the body would be spared any ill-effects of infection by the virus. To reach the hepatic artery, Raper would have to insert a flexible cannula into the femoral artery in Jesse’s groin and thread it backward up the aorta, in a similar fashion to what is done for coronary angiography. The infusion of the vector would take a few minutes, he said, and Jesse would have to lie still for several hours afterward. Over the following days, blood tests would be done to check whether there was any effect on Jesse’s ability to metabolise ammonia. Then, a week after the infusion, Raper would remove a small sample of Jesse’s liver by means of a needle stuck through the front of his abdomen. This biopsy sample would be studied to test whether the vector had been taken up by the liver cells, whether the new gene was working and whether the vector had caused any damage to the liver. Raper emphasised that, although the infusion might cause some brief improvement in Jesse’s ability to excrete ammonia, it would not offer him any long-term benefit. The benefit might come eventually to others, especially to OTC-deficient babies: the vector might help tide them over their first neonatal crisis and save them from immediate death or brain damage. In the long run, the hope was to develop other vectors that would implant the OTC gene more permanently in the children’s genomes.
The informed consent form mentioned a laundry list of possible ill-effects that Jesse might suffer. It was quite possible that he would experience mild flu-like symptoms over the day or so after the infusion. Blood clots might break loose. The vector might cause hepatitis. The needle biopsy might cause a haemorrhage. There might be unforeseen harmful consequences. Jesse signed the consent form, and he had blood drawn to measure his ammonia level. Then he underwent a several-hour test in which he had to drink a sample of ammonia labelled with a non-radioactive isotope of nitrogen (
15
N). The fate of the nitrogen in this test would reveal how efficient Jesse’s metabolism was at getting rid of ammonia.
Once out of the hospital, the family did a bit of sightseeing: they went over to the Spectrum Arena to see the famous statue of Sylvester Stallone as Rocky Balboa. A photo of Jesse standing in front of Rocky, his arms raised high in triumph, later accompanied many news stories about him, and it still can readily be found on the internet. It aptly illustrated the new and positive focus that participation in the OTC trial had injected into his life – a life that otherwise was drifting rather aimlessly. After another day of sightseeing, this time in New York, the family returned to Tucson, where Jesse waited to hear more from the Penn team.
‘Informed consent’ is really a figure of speech. No layperson can truly evaluate the potential risks and benefits of participating in a clinical trial, least of all the trial of a genetically engineered virus. To some extent, signing a consent form is a confession of faith – faith that the researchers have done their homework and that the experimental protocol has been adequately reviewed by experts. In the case of Wilson’s OTC trial, it looked like the review process had been extraordinarily thorough. Wilson’s protocol for the study had been reviewed by the University of Pennsylvania’s Institutional Review Board (IRB), the US National Institutes of Health and the FDA. It had also undergone a special review by the FDA’s Recombinant DNA Advisory Committee, or RAC – a group that vetted protocols involving genetically engineered viruses and other biological therapies.
Yet all had not gone smoothly during the approval process. For one thing, there had been setbacks during the animal testing that preceded the clinical trial. Three monkeys who had received very high doses of the vector developed severe liver failure combined with a blood-clotting disorder, and they had to be killed.
In reaction to these deaths, Wilson prepared a second version of the vector that he claimed was safer. But was it? In 2006, I put this question to Inder Verma, a leading virologist and gene-therapy expert at San Diego’s Salk Institute. ‘It’s possible,’ Verma said, ‘but he had no proof of that. And in fact it’s ironic, because we proved later on that every batch of adenovirus had that problem; it didn’t matter whether it was the first, second or third version. The viral proteins are going into cells and [causing them to be] recognised as foreign by cytotoxic T lymphocytes, which destroy them.’
Some of the initial reviewers had serious reservations about the study. Among them was Robert Erickson, a paediatric geneticist at the University of Arizona who was a member of the RAC. In a 2006 interview, Erickson told me that he had been concerned by the adverse events in the animal studies and also by Wilson’s plan to infuse the vector into the hepatic artery, which Erickson viewed as a risky procedure. He changed his mind when Wilson described changes to the vector and also agreed to infuse it into a peripheral vein. (That change got reversed by a subsequent FDA panel.)
In the final plan for the trial, the vector would be administered to 18 adult volunteers. The volunteers were to be in good health, with their OTC deficiency under reasonable control, which meant their plasma ammonia concentrations could be no higher than 70 micromoles per litre (µM/L) – about twice the maximum level seen in healthy people. The volunteers would be divided into six cohorts, with three volunteers in each cohort. The volunteers in the first cohort would receive a tiny amount of the vector, and – assuming there were no ill-effects – the dose would be increased stepwise until the sixth cohort received the maximum dose. If there were serious adverse effects, they would have to be reported to the FDA before the trial could proceed further.
Another safety consideration had to do with the sex of the volunteers. Because women, with their two X chromosomes, are generally less severely affected by OTC deficiency than are men, it was decided that the first two volunteers in each cohort would be women. Men, if they participated at all, could only be the third and last in a cohort. In that way, the doctors would already have some experience with that dose level before they treated a man. Jesse would therefore be the last of his cohort to receive the vector.
A few weeks after the visit to Philadelphia, Jesse and Paul got a letter from Mark Batshaw, the OTC deficiency specialist on the Penn trial. Batshaw wrote that Jesse’s test results made him an acceptable subject for the trial. His blood ammonia was below the cut-off level of 70 µM/L, and his efficiency at excreting ammonia, as measured by the
15
N test, was six per cent of normal. Because of this very low efficiency – the lowest of anyone in the trial – any increase in ammonia excretion caused by the viral OTC gene would be readily apparent.
Soon after, Paul Gelsinger had a phone conversation with Batshaw in which, according to Paul, Batshaw mentioned the good results they’d had to date. In experiments on OTC-deficient mice, he said, the gene transfer had worked so well that the mice had been protected from what would otherwise have been a lethal dose of ammonia. And because the human trial had now been under way for more than a year, Batshaw was able to give Paul some idea of the early results. The most recent of the volunteers, he said, had experienced a 50 per cent increase in the efficiency of her ammonia excretion after the infusion of the vector. He did not mention any adverse events in the human subjects, according to Gelsinger. In a brief 2007 email exchange, Batshaw confirmed to me the general content of the conversation but said that he remembered having told Gelsinger of adverse effects, including fever and short-term liver abnormalities. (Batshaw – like Wilson and Raper – declined my request for an interview.)
Both Paul and Jesse reacted with enthusiasm to this news, and Jesse was doubly excited about participating. Because the trial was now nearing its end, Jesse would be in the final cohort and he would therefore receive the highest dose of the vector. But he would have to cool his heels for a while because, as a male, he would be the last of the three volunteers in that group. Thus, he would be the final subject in the entire trial. His infusion was scheduled for October.
Jesse spent the intervening time working as a supermarket clerk and, in his free hours, riding a motorbike that his father and stepmother had given him as a graduation present. He seemed as upbeat and full of life as Paul had ever known him.
In mid-August, the co-ordinator of the clinical trial called to say that their next patient (the second patient in the last cohort) had a scheduling conflict, and that they would like Jesse to take her place. This would mean that the infusion would take place in September. Jesse agreed to the change of date.
Putting Jesse – a male – into the second position in a cohort was a clear-cut violation of the protocol that Wilson and his colleagues had agreed to. Wilson has never denied this, although public statements put out by the Institute of Gene Therapy have maintained that the FDA OK’d a similar switch in an earlier cohort, so Wilson felt entitled to make the switch in this cohort too, even without express permission.
According to Paul Gelsinger, the reason for the switch was not that the other patient had a problem with the September date, but that she backed out of the trial altogether. (I was not able to ask any of the investigators directly about this, however.) If the other patient did drop out, it might have seemed best to infuse Jesse right away and wrap up the trial, rather than endure the delays involved in recruiting another volunteer. ‘They just wanted to finish,’ Inder Verma speculated, ‘because this was the last dose and they wanted to stop and get it over with.’
Unknown to Jesse or his father, a number of untoward events occurred at Penn prior to Jesse’s visit. Several of the earlier volunteers experienced significant liver damage from the adenovirus infusions. The damage was assessed by measuring enzymes, such as transaminase, that were released from dying liver cells into the bloodstream. Even as early as June of 1998, one of the volunteers who had just been infused with the adenovirus vector experienced a surge in her serum transaminase levels to nearly eight times the upper limit of normal.
This finding indicated very significant damage to the woman’s liver, even at a dose of the vector far below that which Jesse was scheduled to receive. It was a ‘grade-III adverse event’ on the scale of severity established by the FDA, but Wilson reported it to the FDA as a milder ‘grade-II’ event. Another grade-III event (a high fever) was also reported as grade-II. Numerous other volunteers experienced grade-II adverse events – in fact, there were grade-II or grade-III events in every cohort from the second one onward. According to Paul Gelsinger, all four of the volunteers who immediately preceded Jesse experienced grade-III liver toxicity, but I could not find independent documentation of this. Nevertheless, an official letter of reprimand later sent to Batshaw by the FDA listed five grade-III adverse events, two involving liver toxicity and three involving high fever.
The protocol called for halting the trial for regulatory review if there was a single grade-III event or at least two grade-II events, but the Penn researchers did not halt the trial. They reported the adverse events many months after they occurred, if they reported them at all, and often the information was hidden in the back pages of their reports, while the summaries at the front were much more positive. In the cover letter to his IRB report of August 9, 1999, for example, Wilson wrote as follows: ‘No serious adverse effects have occurred as a result of this study. There have been no significant treatment-related toxicities or procedure-related toxicities, and all participants have remained well.’ Much later, the usually circumspect FDA described this statement with a simple and damning adjective: ‘False’.
