Is the FDA responsible for hundreds of thousands of deaths?

No. But the WSJ would like you to believe so.

One libertarian talking point I hear a lot (Cato of course loves this story), and is repeatedly pushed by the WSJ, is that the market and consumers should decide the safety and efficacy of drugs - not dirty gov'mint bureaucrats who want nothing but death and suffering for cancer patients. The latest is this commentary from Ronald Trowbridge and Steven Walker which has some fun with math to suggest the delay in approval of cancer drugs has led not to dozens, or hundreds, or thousands, but hundreds of thousands of premature deaths.

Is there any basis in fact for these accusations? Is the FDA somehow worse than Hitler? Hmmm.

Let's take a look at their argument.

Gleevec set a tragic standard for loss of life at the hands of FDA bureaucrats. Coming out of Phase I testing in 1998, it was known beyond any reasonable doubt to be safe and effective. The Alliance started requesting access to the drug for chronic myelogenous leukemia (CML) patients in June 2001. By the time FDA approved Gleevec in March 2003, approximately 3,600 patients had been denied access to the drug. Many died waiting. More than 80% of the small number of patients who got Gleevec in clinical trials before the drug was approved are alive today.

Gleevac truly was a revolutionary drug, exceptionally so. It is a drug that specifically inhibits the product of the bcr-abl fusion gene formed by chronic myelogenous leukemia (CML), and surprised everybody when it actually forced CML into remission in a majority of patients - it wasn't absolutely clear that inhibiting the fusion gene would cure the cancer but it does. While it is unfortunate that people died while the drug was being tested, the problem here is that hindsight is 20/20. How were they to know if the tyrosine kinase inhibitor would only affect the bcr-abl protein? There are lots of tyrosine kinases in the body, some of them are critical and if inhibited, even in a minority of patients due to some mutation, people would have dropped dead (insulin receptor TK anyone?). It is simply impossible to know ahead of time if a drug is going to be safe, and they have to be tested. While the goal of speeding up promising drugs is a noble one, the idea of throwing any cancer drug in a clinical trial at patients simply because they have terminal illness strikes me as reckless and dangerous. Especially considering there were existing treatments for CML that had survival rates of about 50% after 5-10 years (interferon alpha and bone transplantation both had success at treating AML). Gleevec raised the progression-free survival rate to 84% which was excellent, and caused a trial comparing it to interferon alpha to be stopped so all patients would get the drug. But there was simply no way to know if the long-term survival, relapse rates etc., would outperform alpha interferon or bone marrow transplantation. In the window that was studied, alpha interferon had a 81% 3-year survival rate vs Gleevec at 96%, a clear benefit over interferon - but it's not like other treatments at the time didn't exist or didn't need to be carefully compared to the new drug.

It's arguable that Gleevec could have gone through faster, it's safer than most cancer drugs, but how about some of their other examples?

Eloxatin, for advanced colorectal cancer, was summarily rejected by the FDA in March 2000 despite its being approved in at least 29 other countries. In January 2002, we started to ask the FDA to allow patients access. The agency delayed approval until August. In between, about 40,000 Americans died without ever getting the drug.

Erbitux, for the treatment of colorectal and head and neck cancers, was rejected by FDA in December 2001 when the agency refused to review the sponsor's application. The Alliance had begun asking the FDA to allow patient access to the drug six months earlier. The FDA delayed approval until February 2004. Almost 179,000 people with colorectal and head and neck cancer died waiting.

Ah, the appeal to "other countries". I'm not sure I want the regulatory bodies in Mexico or China and even many European countries to be cited as authorities, and it's funny that this is the one time you'll see appeals to the policies of other countries in the WSJ. Usually they prefer to use the evil socialist medical care as an example of how valued treatments are delayed due to their communistic universal coverage. But I digress.

This is where their argument gets weaker. Few drugs are as effective as Gleevec was for CML. How does Eloxatin(Oxaliplatin) and Erbitux (Cetuximab - a monoclonal antibody to EGF receptro) compare to standard treatments? Did people die because of lack of access? Well not really. While these drugs have increased the median survival from that of 5-FU alone, about 11 months, to about 21 months they by no means are revolutionary cures like Gleevec. While such an increase in survival is excellent news, and these drugs are a valued addition to the medical armentarium to say that people died due to lack of access is somewhat inaccurate. Some deaths would have been delayed, but to say 40,000 or 179,000 people are dead due to lack of access is gross hyperbole.

The Alliance began working for access to Revlimid, for multiple myeloma and myelodysplastic syndrome, in June 2002. Patients had to wait until December 2005 for FDA approval. Nearly 74,000 patients with these terminal cancers died without ever getting Revlimid.

And what would Revlimid (Lenalidomide - a drug similar to thalidomide) have done for multiple myeloma and myelodysplastic syndromes? For myelodyplastic syndroms about 60% of patients given the drug have to stop or lower the dose because of severe hematologic side effects. About 58% that can tolerate it have a some hematologic response and decreased need of transfusion - mostly in patients with a specific cytogenetic anomaly.

It's also used in therapies for Multiple Myeloma where it's been shown to be as good (or possibly better) than thalidomide. However Lenalidomide has significant safety issues including risk of thromboembolism similar to thalidomide.

Have 74,000 patients with Multiple Myeloma or MDS died because of lack of access to this drug? Decidedly not. It's not clear that it is superior to current combined therapies using thalidomide for MM although it has a clear benefit in reducing transfusions in a specific subgroup of MDS patients (identified in the 2005 study) for whom it has been approved.

They go on to list example after example, and sure enough, even though none of these drugs are absolute cures, and there are already existing therapies, they make it sound like nothing was done at all for hundreds of thousands of people while studies of drug safety and efficacy are pursued.

One more example because it's just too much:

In June 2002 we began our efforts to gain access to Alimta for lung cancer patients. FDA didn't approve it until February 2004. In the interim, approximately 249,000 lung cancer patients died without the chance of trying this drug to see if it would control their disease or extend their life.

Now here they're really abusing the numbers. Alimta (Pemetrexed) is approved for mesothelioma and small cell lung cancer because it provides similar response rates in refractory cancers as other drugs. The suggestion is a quarter of a million people are dead because they didn't try this drug - but no drugs currently available for these types of cancers are remotely curative, and other drugs already existed with similar efficacy to Alimta.

Their summary:

In sum, these 12 drugs -- had they been available to people denied entry to clinical trials -- might have helped more than one million mothers, fathers, sons and daughters live longer, better lives. We have actually underestimated the number of "life-years" lost at more than 520,000, because we have not included other safe and effective uses of these drugs that the FDA has yet to approve.

Now Orac has written before about this conflict between drug safety testing and access to lifesaving drugs. While I similarly understand the desire of people facing life-threatening illness to do anything to try and extend their lives, these attacks on the FDA and drug safety regulation are misleading, inappropriate and exaggerated. Many of these drugs represent potential increases in life, or maybe palliation, but in each case they're representing the unapproved drugs as the only treatment modality - or as a potentially curative one. Most of these drugs represent incremental advances or equivalence to existing treatments. They also are universally poisons (possibly excepting Gleevec) - chemotherapeutics tend to be very dangerous and potentially deadly drugs with huge side-effects and safety issues. To suggest that doctors should be able to prescribe these drugs without adequate knowledge of their side-effects, issues in specific patient populations, and knowledge of which organ systems to monitor is simply nuts. Not every doctor can treat their patients as if they're enrolled in a clinical trial - monitoring every organ system, and every available biomarker. Trials don't just tell us that drugs work or don't work. They tell you which patients are most likely to benefit, what kinds of problems to expect, what to watch, what to ignore, etc.

I feel a great deal of sympathy for these advocates, but the FDA is not their enemy. And undermining the precautions in place to monitor and enforce drug safety for the goal of promoting reckless experimentation on patients is not admirable.


More like this

[Note: Part I is here.] I tell ya, I stay up all night putting the finishing touches on a grant, and what happens? Mark Hoofnagle over at finds a real hum-dinger of stupidity published in the editorial pages of the Wall Street Journal. Unfortunately (or fortunately, given the rampant…
When Gleevec hit the market in 2001 for chronic myelogenous leukemia , it was hailed as a major breakthrough in cancer treatment. Gleevec, which inhibits bcr-abl kinase, was the harbinger of targeted chemotherapy and represented a departure from the cytotoxics which, although effective, possess a…
The 101st Annual Meeting of my primary professional society, the American Association for Cancer Research (AACR), convened in Washington, DC, on Saturday and will run through Wednesday, April 21. The theme for this year's meeting is "Conquering Cancer Through Discovery Research," and focuses…
One of my favorite shows right now is True Detective, an HBO show in which two cops pursue a serial killer over the course of over 16 years. Starring Woody Harrelson and Matthew McConaughey, it's an amazingly creepy show, and McConaughey is amazing at playing his character, Rustin Cohle. I'm sad…

Ugh. It's stuff like this that would (and often does) let lethal quackery run free. Back to the days of caveat emptor and zero accountability.

Is their reasoning about Gleevec even correct? They said "coming out of phase I testing in 1998, [Gleevec] was known beyond any reasonable doubt to be safe and effective."

Is this true? Phase I testing merely tests for safety, not efficacy. In some drugs you accidentally find out efficacy, but that doesn't happen very often. Are they out to lunch here, or was Gleevec so remarkable that they saw efficacy in the Phase I trials?

Great post. It's a good reminder that even the blockbuster durgs are only incrementally better than what has come before. A couple more points to add:
1) the WSJ doesn't point out that the number of drugs that look promsing (even after phase II) and fail phase 3 is acknowledged by Pharma to be greater than the number of drugs that end up looking safe and effective after phase 3/approval. By this simple math it can be seen that drugs that come out of phase 1 or even phase 2 looking promising means very little in terms of real safety and efficacy.

2) Most phase 1 trials say nothing about efficacy, only a bit about safety of expanding the trials. Phase 2 is generally about finding out who benefits under what schedule. There's a reason phase 3 trials are so big. They need to be to show safety and efficacy. Okay so that's sort of point 1b.

3) What the Abigail Alliance and others seem to forget is that the key question is not what happens to those who are willing to take the risk, but all the other people who would be affected by a dismanteling of the current system. I wrote a little more about this here:…

By angrytoxicologist (not verified) on 15 Aug 2007 #permalink

On Gleevec, they did have a pretty good idea. In patients with chronic CML who had tried aIFN, Gleevec showed a complete response in 53 of 54 patients @ 300 mg. Amazing. I don't know of any other anti-cancer drug that has had that kind of activity.

Druker et al. 2001. Efficacy and safety of a specific inhibitor of the BRC-ABL tyrosine kinase in chronic myeloid leukemia. NEJM. 334:1031-1037.

By angrytoxicologist (not verified) on 15 Aug 2007 #permalink

Kerry Howley and others at Reason have been harping on this lately, explicitly siding with the Abigail Alliance. It's good to see you do a post on this.

Yeah, "reason". What a bunch of idiots. Only libertarians would write idiocy contradicted by every bit of historical knowledge we have about unregulated drug markets and call it "reason". I often wonder, if they love this completely unregulated and government-free zone, why don't they just move to Mexico already and be done with it.

I for one prefer drugs that work. And what did we have before the FDA? Tonics, patent medications, and various poisons sold as cure-alls. No thanks.

Cocaine in my sodas, heroin in my cough syrup, amphetamines sold as diet pills... mmm... I sure do miss those carefree days before the FDA!

FWIW, the Washington Legal Foundation, the group that litigated the Abigail Alliance case, is in bed with every big-money interest in Washington. They're crazy-eyed conservative, and even write briefs in favor of the death penalty!

By anonymous (not verified) on 15 Aug 2007 #permalink

I agree with your comments regarding the WSJ's exaggerations. It is quite possible that the net effect of providing free access to experimental drugs would be negative, particularly if there is a "fad" on a drug that turns out to have bad side effects.

Nevertheless, I do support allowing terminal patients free access to experimental treatments. I think that people who are looking at a probable death sentence should be permitted to make their own decisions as to what chances they are willing to take with their remaining years.

Even Gleevec isn't safe "beyond any reasonable doubt." My late wife was on Gleevec for several months near the end of her life, and was forced to stop taking it due to liver toxicity. She had metastatic endometrial stromal sarcoma, not CML, and was on Gleevec because of its promise for treating GIST, which is similar in some respects to ESS (both tend to show elevated expression of the c-kit tyrosine kinase). I don't know how the dosage for solid tumors compares to the dosage for CML, but I do know that the dosage necessary to keep Julia's cancer in check was higher than her liver could tolerate. Lower doses were easier on the liver but allowed the tumors to resume growing, so she resumed a more traditional chemotherapy regimen, which ultimately failed.

The goal of these organizations is to kill the FDA altogether. They see terminal patients' rights as the thin edge of the wedge in getting the government out of drug safety. Cato wouldn't be making all this noise if they weren't carrying water for the drug companies.

Nevertheless, I do support allowing terminal patients free access to experimental treatments. I think that people who are looking at a probable death sentence should be permitted to make their own decisions as to what chances they are willing to take with their remaining years.

But this is bad medicine. Without the clinical trials, what information will doctors have to go on in terms of dosage, side effect profiles, what to monitor, who will benefit, etc.? Too many unknowns makes for bad medicine. While it seems like there is no harm in letting people do whatever they want if they're terminal, it's going to be dangerous for doctors (I doubt many would cooperate with such a desire), isn't evidence-based medicine, and is really just a can of worms.

Then you ad the potential for abuse, drug companies advertising to desperate patients...


I was a resident when Gleevic was being studied, and had a chance to have lunch with Drucker, and worked closely with a number of hematologists involved in the clinical trials.

It was an unusual drug from the beginning, more promising than most new drugs. Applications were made for compassionate use. But even so, not all patients responded, and it wasn't clear what the safety would be.

Allowing patients to evaluated drugs will only encourage woo-meisters.

I'd just ask, "How's China doing without all that pesky regulation of business?"

Great post, Mark. The WSJ conveniently forgets that we come to know certain drugs as "safe" after the fact, not before. While it's easy to take on the role of Monday morning quarterbacks, we have to remember that the whole process of drug testing naturally involves a time when the disease existed but the treatment did not. The cautious process we have now has evolved for a very good reason in our litigious culture -- if the industry erred instead on the side of releasing drugs earlier, how would it withstand the lawsuits on behalf of the drugs' additional casualties? Serving on a research hospital's IRB, I'm reminded of the ways in which we do make some drugs and devices available for the earliest use possible, even prior to FDA approval, but only with considerable caution and accountability.

"If the industry erred instead on the side of releasing drugs earlier, how would it withstand the lawsuits on behalf of the drugs' additional casualties?"

That's easy. The so-called libertarians would outlaw the lawsuits.

By Michael Glenn (not verified) on 15 Aug 2007 #permalink

"That's easy. The so-called libertarians would outlaw the lawsuits."

No way. All the libertarians are lawyers!

Correct me if I'm wrong, but didn't the initial push for fast-tracking dangerously experimental drugs prior to FDA approval come out of ACT UP and the movement to release AIDS drugs for use by HIV-positive patients?

You know, I can't help but note that one of the drugs they mention is a thalidomide derivative. What do you suppose the current WSJ editorial board would have said when that one inspector blocked the approval process for thalidomide back in the 60s? (I mean, it's nice that they found a use for something so dangerous, but it never would have come to the market in the first place if the original company that created it in Germany had done their preclinical testing properly...)

The c-kit protein may be expressed in some other tumors, however, lots of other tumors can express c-kit. But this expression is only correlated with response to Gleevec in GIST tumors. It is meaningless in other tumors. Gleevec will kill virtually anything if used at a sufficiently high concentration. But at what cost, killing the patient?

There are lots of tumors which are c-kit positive which don't respond to Gleevec. Just like there are lots of tumors which are estrogen receptor positive which don't respond to tamoxifen (e.g. melanoma). And lots of tumors Her2/neu positive which don't respond to Herceptin.

Oncologists prescribe patients one standard empiric chemotherapy regimen after another, until they find one that works. This often can expose patients to the side effects of chemotherapy, without showing any cancer-killing results.

Unfortunately, the introduction of Gleevec and the other new "targeted" drugs has not been accompanied by specific predictive tests allowing for a rational and economical use of the drugs.

By gpawelski (not verified) on 15 Aug 2007 #permalink

Nevertheless, I do support allowing terminal patients free access to experimental treatments. I think that people who are looking at a probable death sentence should be permitted to make their own decisions as to what chances they are willing to take with their remaining years.

A friend of mine worked at a small biotech company that had several "compassionate use" applicants for their anti-cancer treatment. Keep in mind, they still don't know if this treatment works. They are producing such small amounts of their drug that they can barely afford to produce enough for use in drug trials. If they were forced to divert resources into a production facility prior to knowing if their drug worked, the would likely go bankrupt (as it is, the company has barely enough money to keep their trials going for a year or two).

Allowing patient access to unproven drugs makes no sense on any level.

But this is bad medicine. Without the clinical trials, what information will doctors have to go on in terms of dosage, side effect profiles, what to monitor, who will benefit, etc.? Too many unknowns makes for bad medicine. While it seems like there is no harm in letting people do whatever they want if they're terminal, it's going to be dangerous for doctors (I doubt many would cooperate with such a desire), isn't evidence-based medicine, and is really just a can of worms.

No, allowing people to make their own choices is not EBM, but these are people that EBM doesn't have much to offer. The "don't allow terminal patients to choose, because then we won't be able to get them to sign up for randomized study" argument does not strike me an adequate reason for denying human beings a choice. There was a time when it was considered ethical to use prisoners as research subjects for new drugs. That is no longer the case because it is recognized that a person who is imprisoned cannot truly give consent. I feel the same way about telling somebody, "Join a study, because that is your only chance of getting access to a drug that may save your life."

My suspicion is that there will still be people who will consent to participate in randomized research studies, understanding (a) that the chance that a new, untested drug will really cure them is very, very slim, and it could even make them worse, and (b) that participating in a research study is a way that they can use their remaining time to help other people who may be facing a similar disease in the future.

And as for those who choose to insist on a particular untested drug, so what? No, it's not as good as a randomized trial, but there are other study designs available. It is still possible to follow them to see how they do, comparing them to patients with similar disease who choose other treatment modalities.

Then you ad the potential for abuse, drug companies advertising to desperate patients.

Who said anything about exempting drug companies from FDA restrictions regarding advertising?

What would be more advantageous is to sort out what's the best "profile" in terms of which patients benefit from this drug or that drug. Can they be combined? What's the proper way to work with all the new drugs? If a drug works extremely well for a certain percentage of cancer patients, identify which ones. If one drug or another is working for some patients then obviously there are others who would also benefit. What's good for the group (population) may not be good for the individual.

So, all the validation of this gene or that protein providing us with a variety of sophisticated techniques to provide new insights into the tumorigenic process, if the "targeted" drug either won't "get in" in the first place or if it gets pumped out/extruded or if it gets immediately metabolized inside the cell, it just isn't going to work.

Dr. Len Lichtenfeld, deputy chief medical officer for the American Cancer Society, in a response about the Cancer Genome Project, said "We're going to be able to take a cancer specimen, analyze it, and follow those genetic changes that influence particular pathways, then we'll use one, two, three of more "targeted" therapies, perhaps simultaneously, and be able to completely interrupt the flow of the cancer process."

I agree! Upgrading clinical therapy by using drug sensitivity assays measuring "cell death" of three dimensional microclusters of "live" fresh tumor cell, can improve the situation by allowing more drugs to be considered. The more drug types there are in the selective arsenal, the more likely the system is to prove beneficial.

By gpawelski (not verified) on 14 Sep 2007 #permalink