Bad Medicine

This week’s Economist has a report on the disastrous clinical trial of TGN1412, an antibody designed to combat leukaemia and arthritis. The drug passed tests on animals, including primates, and was administered correctly, as the MHRA (British health agency) just confirmed. Yet during the clinical trial all six test subjects became seriously ill, some requiring intensive care. The story with references is on Wikipedia:

Phase I clinical trials conducted by PAREXEL at Northwick Park Hospital, London, resulted in hospitalization of six volunteers in March 2006, at least four of whom suffered major organ failure. Healthy volunteers were recruited to the study with a £2000 fee, reportedly much higher than the ‘few hundred quid’ offered for other medical tests in the region. The most severely affected volunteer, Mohammed Abdalla, a 28-year old who had hoped to set his brother up in business in Egypt, was said to have suffered a ballooned head similar to the “Elephant Man”.

The trials were carried out by PAREXEL International, a company that carries out drug trials on behalf of pharmaceutical companies, at their facilities at Northwick Park Hospital, in Harrow in Greater London. Eight young, healthy men participated in the trial, with two receiving a placebo and the remaining six receiving TGN1412. The doses were given by injection with an interval of around two minutes between each patient. Roughly five minutes after the last participant had received his dose, the participant receiving the first dose started complaining of a severe headache, fever and pain. He took his shirt off, complaining that he felt like he was burning. Shortly after, the remaining participants who received the actual drug became ill as well, vomiting and complaining of severe pain. Within twelve hours all six had collapsed. At least one participant begged the doctors to “put [him] to sleep” because of the suffering.

As of 31 March 2006, four of the six patients had returned to their homes. One of the other two has been transferred out of Critical Care, and the worst-affected patient “is now fully conscious and we are encouraged by his progress,” according to a press release on the North West London Hospitals NHS Trust website.

The high fee sounds suspicious but if the company knew of any dangers it did not communicate nobody could prove it so far. If this was indeed an unforeseeable accident the news may be worrisome for modern drug development in general. New drugs like these custom-built antibodies are supposed to target diseases much more accurately than the “chemical bombs” of old – but does that mean we can no longer get any accurate idea of their effect without using human test subjects? Is this a general danger or just a freak accident?

Read the article.

If we cannot find some technobabble fix (magical computer model of the immune system, cloned tissue systems for tests, etc) then it seems one would need an additional protocol for high-risk tests.

Do the desert-island-i-found-some-berries-who-eats-first test. You eat the red ones, I’ll take the blue ones. Tomorrow we find out which is poisonous. Stagger the tests on an even smaller pool of subjects. Problems I can foresee are additional expense and length on a study. Also, who do you pick to take the poison first?

You’d also have to identify what you consider a test dangerous enough to introduce an additional safety layer. If it starts going beyond say 3% of tests then research expenses will be even higher.

How do you gauge how much danger to test subjects is worth it for possible benefits for thousands or even millions of humans (in case of say, malaria. Of course, we could also give people mosquito nets and spray DDT.)

From reading the wiki page, it looks like this stuff had the potential to set off a positive feedback loop in the immune system. I’m guessing that the standard ‘tradecraft’ of drug studies doesn’t deal well with highly non-linear responses like this. So I’m thinking that they designed a drug that operates in a very diffrent way, but used the same ‘tried and true’ trial apparatus.

So in my totally uninformed opinion, people are probably looking into the assumptions built in to the trail design and figuring out which ones this drug violated.

This is something pharma companies have always struggled with, and an illustration of why it costs so much to develop and market new drugs. The company probably spent close to a billion dollars getting a drug of that type this far through testing, which means it also had its best resources working on this rather than something else, and now it’s pretty clear that that investment is in the trash bin. There’s absolutely no incentive to pay people more than normal to volunteer to test a drug that you “know” will have adverse effects, because no matter what you pay them the data will be the data. You spend all of your money trying to avoid precisely what appears to have happened here.

Lots of people working on computer models and every other methodology to avoid human testing and minimize animal testing, but in the end we’re just not that smart yet. Look at the drugs that have been on the markets for years and declared safe that have recently been pulled (with subsequent lawsuits) due to adverse effects. It’s a tough, complex, and risky business.

I am thoroughly disappointed that this is not a thread about Bon Jovi.

Shoot! You stole my line!

The high fee sounds suspicious but if the company knew of any dangers it did not communicate nobody could prove it so far.

Was hot news over here a few weeks back. The fees about right apparently for a first human trial of something. They’ve wheeled a few people out who are suspicious about the number of people involved in the trial but other than that I don’t think anything concrete has come out of it yet regarding what went wrong.

This would not have occurred had the drugs been tested in the flesh vats of Merck.