Avastin and the power of hope
July 1, 2011 3:49 AM   Subscribe

Based on a lack of evidence for efficacy, "an FDA panel unanimously voted to revoke its approval of Avastin (bevacizumab) for breast cancer. The decision evoked cheers from some groups and jeers from others. At least one group derided the decision as the work of a 'death panel'". An interesting article on the withdrawal of a "miracle" drug from a section of the market, explaining the reasoning behind the decision and discussing the reaction from patients, many of whom believe -- despite the evidence -- that the drug was actually helping them.

I thought that the New England Journal of Medicine "Perspectives" article linked from the post is particularly interesting, although I'm not sure how clear it'll be for non-biomed people.
posted by metaBugs (11 comments total) 1 user marked this as a favorite
 
explaining the reasoning behind the decision and discussing the reaction from patients, many of whom believe -- despite the evidence -- that the drug was actually helping them.
Despite the evidence? It's entirely possible that a drug might work for some people, but not the general population. It's also possible that other drugs might not work as well. So I don't see how the evidence suggests that it's not helping these particular women.
posted by delmoi at 3:56 AM on July 1, 2011 [1 favorite]


(The article seems to also suggest that the drug improves quality of life, even if it doesn't actually extend life)
posted by delmoi at 3:58 AM on July 1, 2011 [1 favorite]


Given the number of people who embrace "cures" that haven't shown clinical efficacy for anything, nothing about the public response to this surprises me.

The problem here is that if you do a full survival study for a drug that really works against something that kills a person slowly, you'll never make back your investment, much less a profit, because it will be off patent by time your have your phase III clinical data. Meanwhile people with the disease will be doing that long, painful death thing. A system that self-corrects as is goes along, rather than depending on being right 100% of the time, is a good thing, but I'm sure someone out there is going to see this as proof positive that the FDA is hopelessly corrupt.
posted by Kid Charlemagne at 4:25 AM on July 1, 2011 [3 favorites]


We've had this row in the UK at least twice. Avastin is not currently available on the NHS. People confuse anecdotes for evidence and journalists like to present those anecdotes as proof of efficacy. Cancer is scary and I understand how people in that situation want to believe in magic bullets, but we're not doing them any favours by ignoring or misrepresenting the evidence.

But this isn't really about cancer, in the same way that the climate change debate isn't about the weather. There is a growing number of people who are either ignorant of the scientific method or, worse, who reject it outright. I don't like where we're headed.
posted by londonmark at 4:37 AM on July 1, 2011 [4 favorites]


In actuality, I think that this is really just unknown territory for the moment (not only Avastin, but many other oncology therapies have been approved based on improvements in progression-free survival (PFS) endpoints). There have been some studies designed with PFS as an endpoint rather than overall survival (OS) with rationale that collecting data to demonstrate significantly improved PFS may take months whereas collecting data to demonstrate significantly improved OS may take ~a decade, as is the case with breast cancer (so if it demonstrates superiority on grounds of PFS, sometimes all patients in both study groups may be given access to a drug with the belief that it is ethical and PFS was supposedly a surrogate endpoint for survival). Whether PFS is a surrogate endpoint for OS has been debated for many years.

Also, I think that delmoi does make a valid point: "It's also possible that other drugs might not work as well. So I don't see how the evidence suggests that it's not helping these particular women."

Although there is nothing in these studies at the moment that suggests it does or does not benefit a particular group of patients, some of the new oncology drugs have been found to provide benefit for a particular group of patients many years later. A good example of this is the drug gefinitib (Iressa). This drug was initially approved in the US for non-small cell lung cancer but it was later removed and only given to a smaller group of patients, or per the PI, only to be given to patients who are currently or who have benefited from gefitinib (and how could you if the drug is not approved for the general public). Anyways, part of this story is that it was not until years later that a specific population that benefited from this therapy was identified, namely, patients with epidermal growth factor receptor (EGFR) mutations, which have been found in nonsmoking females in Asia.
posted by Wolfster at 6:55 AM on July 1, 2011 [1 favorite]


Although there is nothing in these studies at the moment that suggests it does or does not benefit a particular group of patients, some of the new oncology drugs have been found to provide benefit for a particular group of patients many years later.

That's very interesting, but also surely besides the point? It hardly seems like best practice to develop and release a drug first, then scramble about to try and prove it's useful to someone.
posted by londonmark at 7:12 AM on July 1, 2011


As I understand it, the drug is still available for other oncology treatments, and can be black labeled...but the wrench in the works is that this decision from the FDA means that insurance won't cover it for breast cancer patients.
posted by dejah420 at 7:13 AM on July 1, 2011


This is really tricky, because as Wolfster says, some new anticancer drugs will only help a small %age of people with certain metabolic conditions and/or genetic profiles who have cancer. It's quite difficult (and expensive) science to find out what those conditions/profiles are and why they work in aiding the chemo.

So, people facing death hear a story about a cure which may well be true for an individual patient or two, but because these conditions/markers are not prevalent, the cure is not statisitically valid - ie. it won't affect people in a manner that can be measure above the noise given the small sample size....a small %age of some cancers simply regress or slow down of their own accord and this confounds analyses like these.

It's a real conundrum: doing hard, personalized medicine like this ACTUALLY costs a fortune, and we as a culture have difficulty delivering simple diagnostic tests and cures without massive cost overruns, rentseeking, administrative bloat and error.

My heart breaks for cancer patients and their loved ones; hope is the cruelest of emotions sometimes.
posted by lalochezia at 7:37 AM on July 1, 2011


Also for intelligent disucssion about the drug industry and studies like these, the blog In the Pipeline is very good reading.
posted by lalochezia at 7:40 AM on July 1, 2011 [2 favorites]


I always thought "Avastin" sounded like a remedy for Pirate Talk Syndrome.
posted by fourcheesemac at 12:06 PM on July 1, 2011


It hardly seems like best practice to develop and release a drug first, then scramble about to try and prove it's useful to someone.

Typically (i.e. everywhere I'm familiar with) you start out with the knowledge of some off kilter biological pathway that causes bad things to happen (or a pathway you can throw off kilter to prevent bad things from happening) and try to figure out what you can do to de/re-kilterify things. If you know that A causes B and B leads to an excess of C anc C causes people to die. Logically, you're going to try to block the A -> B step, or maybe the B -> C step right?

Unfortunately, your patients are an assembly of complex systems, which do not always respond in a predictable manner. When you get to the clinic and test this thing out, all those people who were dying of some horrible disease die anyway because there is also some Θ link in a chain of events that goes A -> Θ -> C that really only happens in people with minimal B which is pretty much your patients and no one else.

Humanity’s total knowledge of Θ might be a predicted amino acid sequence for some protein that's involved, so even if you went looking for whatever was happening in your patient that didn't line up with your model, there's no guarantee you'd ever even find this hypothetical Θ factor.

It makes it hard to get out of bed in the morning.

Unless you’re a systems biologist, then it’s why you get out of bed in the morning.
posted by Kid Charlemagne at 1:03 PM on July 1, 2011


« Older Can I Get Some Pockets Or Something?   |   Simple Symbols Newer »


This thread has been archived and is closed to new comments