Monday, November 19, 2007

Merck's failed HIV vaccine trial

This has nothing to do with biological warfare, but I feel compelled to make a few points about this trial anyway.

1. The volunteer subjects who received vaccine were apparently harmed. But did they receive sufficient warning that this could happen? Merck/NIAID will no doubt say it could not have been foreseen (though one article suggests the consent form stated that participation might decrease, increase, or have no effect on your risk of AIDS). Does this mean the lawyers knew the vaccine might be harmful but the doctors didn't?

A number of animal vaccines were developed, some even licensed, sold and used on pets and livestock -- when it turned out they increased susceptibility to the illness they were trying to protect against. This potential adverse effect should be well known to vaccine developers.

Not to mention the obvious -- small pilot trials of the vaccine should precede large-scale trials to rule out this dread complication. This was a Phase 3 trial in 3000 people.

2. The vaccine trial that took place in South Africa started at a later date (2007) than trials in the other countries started (2004). Yet it was the South African trial that blew the whistle on the increased rates of HIV in vaccine recipients. At that point, in late October, investigators from the earlier trials claimed they would have to evaluate their data to see if it bore out the South African findings. Now they admit their data confirms what the South African investigators identified: more HIV infections in those who received vaccine compared to placebo.

Guess what? Each trial of this vaccine was required to have a Data Safety Monitoring Board whose job is to monitor interim trial results to end a trial early if things are going wrong, as in this case -- and warn subjects if needed.

Clearly, this board was not doing its job in a timely fashion. Even worse, it is a good possibility that had the South Africans not blown the whistle, the other investigators might have attempted to sweep these data under the rug. Merck may have even tried to sell this vaccine, if it had been able to analyze the data in such a way this negative effect was not identified.

Why did it take the South Africans only months to see what the other centers only acknowledged last week, 3 years into the trial?

Although it sounds far-fetched, many pharmaceutical companies require medical investigators, even those at prestigious medical schools, to agree they will not get access to their own data, as a condition of being hired to conduct a clinical trial. Only the company gets to see the raw data.

The whitewashing has already begun: In Australia, investigator "Kelleher defended the protocol that halted the trial halfway on the advice of the Data Safety Monitoring Board and praised the transparency of the communication to participants."

But look back to the beginning of this fiasco becoming pubic: from the October 25, 2007 Washington Post:

Warning Is Sent to AIDS Vaccine Volunteers

...The Merck vaccine trials took place in 15 cities in the United States, including Boston, Los Angeles and New York, and three in Canada. There also were sites in Peru, Brazil, Australia, Haiti, the Dominican Republic and Jamaica. Those trials began in December 2004 and included 3,000 participants, mostly gay men.

In South Africa -- where an estimated 5.5 million people are infected with HIV, more than in any other country -- the study used the same vaccine but was administered separately. The trial here started later, with the first injections this year, and had its own ethics oversight board. Most of the subjects were heterosexual.

The ethics oversight board in the United States, which monitored the trial everywhere but in South Africa, has not decided whether to tell participants if they received the placebo or the vaccine, said Mark Feinberg, vice president for medical affairs and policy for Merck..."

Think of this article next time you considering volunteering for a scientific experiment

No comments: