Monday, August 06, 2007

More Unfortunate Details About Targeted Genetics' Fatal tgAAC94 Trial

Rick Weiss in the Washington Post just published a follow-up report on the death of a study participant in an early phase trial of gene therapy for (apparently inflammatory, probably rheumatoid) arthritis. We had first posted about this here.

In summary, the trial was of tgAAC94, a transgene encoding the receptor for tumor necrosis factor (TNF)-alpha, a cytokine that causes joint swelling in arthritis patients, made by Targeted Genetics. The trial used a viral vector injected directly into an arthritic joint. An NIH review panel that first considered the trial was worried that the new treatment was to be given to not very sick patients, and that the virus could escape from the joint into which it was injected and cause systemic problems.

Oddly enough, Targeted Genetics had bought out Genovo, the company whose gene therapy was associated with the death of a patient in Pennsylvania in 1999.

The new Washington Post report revealed some detail about the death in the tgAAC94 trial.

[The patient] had been feeling fine just a few weeks earlier, save for occasional stiffness from her arthritis.

[The] first shot, administered on Feb. 26, had no noticeable effect, and she wondered whether she got the placebo. But she was excited that the next one would be the real thing....

That happened on July 2, a Monday. She was tired and out of sorts after a weekend of boating with her husband and daughter....

[Her physician] recorded her temperature at 99.6, then gave her the shot.

'The next day she woke up and didn't feel good at all. By afternoon she started vomiting,' [her husband] ... recalled. 'By evening her temperature had shot up to 101.'

She spent July Fourth feverish and vomiting. Her family physician told her it was probably just a virus.

When her symptoms worsened and her temperature hit 104.1 on Saturday, she went to the emergency room. Tests indicated a possible infection and signs of liver damage, but she was sent home for more care from her family doctor....

But [she] ... only got worse, and on Thursday she was admitted to the hospital.

Things went downhill fast, with [her] ... body showing signs of being ravaged by an infection. But tests for standard bacteria and viruses came up negative. With breathing problems and the possibility she might need a liver transplant, she was transferred to the University of Chicago hospital.

[Soon she] was breathing with the help of a ventilator in a Chicago intensive care unit -- her body bloated from internal bleeding, her liver failing -- and no one could figure out what was wrong with her.

Three weeks after ... the injection that she had hoped would cure her, she was unconscious and beyond hope of recovery. With family and friends gathered around, her life support was removed.

Furthermore, the Washington Post article documented a series of problems with how the trial of tgAAC94 was done.

1) The study was presented to the patient as if it possibly could help her clinical condition, even though it was only a Phase I study designed to assess toxicity.

'It was presented to her like this is going to make her knee better,' said [her husband.]

Further in [the study's consent form], after long descriptions of how the product may help, a single sentence states: 'We do not expect you to receive any direct medical benefits from participation in this study.'

2) The patient was persuaded to enter the study by her rheumatologist, Robert Trapp, who was paid by Targeted Genetics to recruit patients, and who administered the consent form to her without giving her an opportunity to take it home and think it over.

Her rheumatologist, [was] Robert Trapp, whose Springfield clinic got payments from Targeted Genetics for each subject he recruited....

Trapp, one of 20 U.S. doctors testing tgAAC94, invited her to join the study on Feb. 12.

There would be two injections, months apart, he explained. The first might be real, or it might be a placebo, but the second would definitely be the test product. She signed up immediately, and Trapp drew several tubes of blood to get the study going.

Two fundamental rules of clinical research were violated that day, experts said. First, consent forms are to be taken home and considered, not signed on first sight. Second, when a patient's own doctor is a principal investigator in a study, someone else is supposed to make the proposal.

3) A lay-person would have found the consent form difficult to understand.

Jonathan Moreno, an expert in the ethics of medical experiments at the University of Pennsylvania, said the consent form used by Targeted Genetics to outline the drug's potential dangers was thick with technical descriptions and thin on explaining 'what's really going to happen.'

'Even a smart person would have a very hard time figuring out what they're talking about,' said Moreno, who examined the form at the request of The Washington Post.

4) The consent form and apparently the whole trial were approved by a not yet identified for-profit institutional review board.

The form was approved by one of the growing number of for-hire review boards that contract with biotechnology companies to ensure studies meet patient-protection standards. Targeted Genetics noted that the review firm it used is accredited and accepted by the FDA. But the use of private boards, as opposed to those run by universities or government agencies, has raised alarms among some medical ethicists since a for-profit review board risks losing repeat business if it is too tough on its clients.

Note that it is not yet clear whether Targeted Genetics directly administered the trial, or whether it was run by an academic institution, or a contract research organization.

5) Targeted Genetics delayed reporting of the patient's illness to the US Food and Drug Administration (FDA).

But Targeted Genetics and Trapp had at first classified the problems as not serious, and later classified them as serious but unrelated to the treatment. So no FDA report was made, and the study went on, with other volunteers unaware of the problems.

Finally, on July 20, a day after Mohr's emergency transport to Chicago and four days before she died, the company sent a serious adverse event' report to the FDA and suspended the study, conceding that her life-threatening symptoms were "possibly" due to the treatment.

In summary, it is too soon to tell if the patient died as a direct result of the gene therapy. It is also not clear whether any of the problems with how the research was designed and conducted were at all related to her death.

However, this tragic case does once again illustrate that commercially funded research is often not designed and carried out as well as it should be.

Furthermore, problems in research design and operation tend to favor the vested interests of the research sponsor. For example, regarding Targeted Genetics delayed reporting of a research participant's serious illness,

That reflects a widespread problem in clinical trials, said Adil Shamoo, a professor at the University of Maryland School of Medicine and editor in chief of the journal Accountability in Research.

'There are no uniform standards for 'adverse events' reporting,' Shamoo said. 'And there is no motivation to report them. . . . No one wants to show their dirty linen.'

Dirty linen can drag down a company's bottom line, and Targeted Genetics, like all companies, puts a lot of work into keeping that line afloat. An interim report on tgAAC94, for example, spoke in June of 'very encouraging results' and evidence of 'clinical benefit,' although, by one measure the company considered key, patients who got high, medium or low doses of the drug did the same as those who got placebos.

'The company was talking about lucrative markets and a promising product much too soon,' said Marcy Darnovsky, associate executive director of the Center for Genetics and Society, an Oakland, Calif.-based public interest group that focuses on genetic technologies.
This is another vivid and very unfortunate case that raises questions whether commercial firms should sponsor, run, or even just influence clinical research designed to test their own products.

Hat tip to the WSJ Health Blog.


S. Isbill said...

Re: This is another vivid and very unfortunate case that raises questions whether commercial firms should sponsor, run, or even just influence clinical research designed to test their own products."

Do you have an alternative proposal? After something goes wrong, all the critics come out of the woodwork. Rarely do they propose a workable alternative. I seriously doubt that taking the consent form home would have resulted in a decision to not enter the study. These patients want relief; they are likely to follow their doctor's recommendations and possibly not even read the multipage, technically worded consent form.

Roy M. Poses MD said...


One alternative would have been to simply follow good research practice. Write a clear consent form that explained the trial was not meant to provide ANY BENEFITS to the patient, that the therapy was experimental and potentially risky, etc. Have the patient approached with the request to enter the study by someone not and independent of her own physician. When the patient became severly ill, assume pessimistically that the illness was related to the study, stop the study, and notify the FDA, etc, etc

Another alternative would be to have all clinical studies (that is, studies involving live patients) done only by groups and organizations who have no vested financial interest in their outcomes.

Chgo PharmD said...

I agree with Dr. Poses completely. Those of us in clinical research became cognizant of good practice in study protocol and medical ethics in our first weeks of training. Violations of good study design and implementation are precisely where surreptitious influences can come into play.

Anonymous said...

Given S. Isbill's position (GCP Solutions,) I, a layman, find her comments deeply alarming.

She says, essentially, to hell with the regs, what else can we do but proceed willy-nilly? And she's claiming to be an advisor to companies wishing to enter into clinical trials?

S. Isbill said...

Fungus Infected Woman Who Died After Gene Therapy
By Rick Weiss
Washington Post Staff Writer
Friday, August 17, 2007; A10

The 36-year-old Illinois woman who died last month after being treated with
an experimental gene therapy was infected with a fungus that usually causes
only a mild illness. But the infection spun out of control and ravaged her
organs, suggesting that her immune system was seriously impaired, said a
doctor who is part of the medical investigation.
The woman's body was also teeming with a cold-sore virus that the body
normally keeps in check, another indication of a faltering immune system.
And because of a tear inside her abdomen -- perhaps caused by infection,
perhaps by injury -- she had an internal blood clot the size of a

In view of this information, assuming that it is accurate, I still doubt that taking the consent form home would have prevented the problem. "Anonymous" is anonymous because "anonymous" does not wish to have an intelligent discussion regarding this study. I still don't see any alternative suggestion for how the study should have been handled; I'm not suprised; laymen usually just want to rant and rave without getting into a real scientific discussion. They don't have the training and experience to understand what's going on. By the way, I'm a he, not a she.

Roy M. Poses MD said...

This case has not been adequately investigated yet, so we still don't know all relevant details about how this trial was designed and conducted.

Nonetheless, my original post above suggests a number of ways, in retrospect, it could have been improved:
1 - The consent form should have clearly explained that no patient would be likely to gain any direct clinical benefit from enrolling in the trial, but would be exposed to as yet unknown risks.
2 - No patients should have been enrolled in the trial by their own personal physicians. Enrollment should have been done by trial personnel, e.g., research assistants.
3 - The patient should have been given the consent form to read at home, not asked to sign it immediately.
4 - The consent form should not have been "technically worded." It should have been written in clear, plain language that a lay-person could have uncerstood.

None of these suggestions are rocket science. They all are based on standard principles of research ethics. They should be familiar to anyone experienced in designing and conducting controlled trials.

In retrospect, one cannot prove that these improvements would have prevented this patient from enrolling, or lead to a better outcome for her.

But they would have made for a more ethically conducted trial.