About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« Things I Won't Work With: Chlorine Azide | Main | Use Avastin! Don't Use Avastin! »

February 28, 2011

Down In Phase III. Again.

Email This Entry

Posted by Derek

Past performance (Phase II results) are no guarantee of future success (Phase III). That warning has been proven over and over in this business, and an awful lot of time, effort, and money have gone down the waste chute in the process. To give you an idea, though, of how hard it is to break out of that cycle, consider Renovo.

As the InVivoBlog details, Renovo was founded to try out ideas to reduce scar tissue formation. And their whole strategy was to go into humans as quickly as possible, to firm up the clinical relevance of their candidate therapies. That's a bit easier to do with something like scarring, if you can find patients willing to have small cuts made in their skin. That's just how one of the Phase II trials was run for the company's Juvista (recombinant TGF beta 3) - two cuts, one treated with the drug and one without. And the results looked quite good.

But not in Phase III. Earlier this month, the company announced that Juvista has completely, utterly missed its endpoints in the larger trial, and no one seems to know why. According to the InVivoBlog, investors were reduced on the conference call to asking if somehow the data collection had been messed up - surely some of the placebo group and the treatment group had been, uh, switched somehow? But no.

It's worth remembering, though, that not all the Phase II data were so convincing. In retrospect, the earlier trials that looked bad were predictive, while the impressive numbers appear to have been artifacts. But how do you figure that out in advance? And how do you run only the trials that will be predictive, and how do you know to trust them? I'm tempted to ask Francis Collins to get on this for all of us, but that would be unfair. I think.

Comments (17) + TrackBacks (0) | Category: Clinical Trials | Drug Development


1. David Formerly Known as a Chemist on February 28, 2011 11:04 AM writes...

This is why there are three phases to clinical development, and isn't surprising. Plenty of therapeutics fail in phase 3. Humans are a complex organism, and outcomes are difficult to predict.

Permalink to Comment

2. std on February 28, 2011 1:02 PM writes...

Even if some drug has no toxicity and excellent results lawyers advice to stop supplying patients in Phase I with that drug once the trial is complete. The reason is simple, no positive results can be reported to FDA, but the negative ones have to be reported, and that is not good for attracting investors for Phase II. At the end though, the potential negative results cost more.

Permalink to Comment

3. MTK on February 28, 2011 1:20 PM writes...

Uh STD (which is a curious choice of screen name),

Aren't Phase I studies usually with healthy volunteers to test toxicity, PK/PD, tolerability, but not necesssarily effectiveness? So why would you continue supplying the drug? Even if you did, what good would any data outside the study be?

And aren't all clinical trials now required to be reported?

Maybe I'm missing something here, but I honestly can't make any sense of what you wrote.

Permalink to Comment

4. a-non on February 28, 2011 5:10 PM writes...

if many drugs show good phase II results and fall over at phase III, then doesn't it follow that many drugs that show bad phase II results will sail through phase III to the clinic

Permalink to Comment

5. H on February 28, 2011 5:29 PM writes...

Phase II results sure allowed a lot of scam outfits to get rich. Given that The House of Fraud (aka Wall Street) believes there is an endless supply of suckers, I don't suppose the notion of "Humans are a complex organism" will ever be taken seriously in industry. Therefore, Academia to the rescue!

Permalink to Comment

6. Still Scared of Dinosaurs on February 28, 2011 5:54 PM writes...

It's always worthwhile to consider whether you are getting a convincing answer to what appears a rather simple question. Continuing to give drug after the primary needs to be justified by clearcut descriptions of what you are going to learn or overwhelming ethical considerations. Either should, um, be pretty easy to understand but most of the time you are probably getting BS.

When looking at Phase 2 data in which some studies look good and others don't the question of which ones to believe is secondary. The primary question is what differentiates them other than the results. Listen carefully to the answer and then clean your shoes 'cuz you can't avoid stepping in it when that much is shoveled into the room.

Permalink to Comment

7. r.pal on February 28, 2011 10:50 PM writes...

Today, the National Institutes of Health (NIH) estimates that a mere 10% of the cells that comprise Homo sapiens are human cells. The remaining 90% are bacterial
in origin. The number of Escherichia coli in a single human is comparable to the entire human global population – approximately six billion people (Staley, Biodiversity: are microbial species threatened? Curr Opin Biotechnol 8(3): 340–3451997).
The makeup of a person’s microbiota is unique: humans may share as little as 1% of the same species (Eckburg et al., 2005).
Eckburg PB, Bik EM, Bernstein CN et al (2005) Diversity of the human intestinal microbial flora. Science 308(5728):1635–1638

In pharma we are focussed on 10% of human genes. And of this we are focused on 1 target gene ( one of 23000)
It is like finding a needle in a hay stack
Naturally the chances of success are small

Permalink to Comment

8. std on February 28, 2011 11:22 PM writes...

You are missing something. I was referring to cancer treatment and situation when phase I is finished, no (low) toxicity is observed, tumor growth stopped, patients feel good and want to continue the treatment. They will be refused.

Permalink to Comment

9. Mr. Obvious on March 1, 2011 7:49 AM writes...

Phase II is self-reported efficacy in a narrowly defined, well-controlled patient populationoften with surrogate endpoints.

Phase III is pre-specified clincial endpoints and a statistical plan. The trial must be conducted in a "real-life" setting that will support the labeling.

It's the real life stiuation that is the issue. Patient compliance, SEs that are less common, DDI, dodgy early development, etc., all come home to roost.

Permalink to Comment

10. Still Scared of Dinosaurs on March 1, 2011 9:03 AM writes...

Mr. O,

Not sure what you mean by "self-reported" which I take to mean "patient reported" as opposed to MD or lab test. I infer that you mean that sponsors make up their endpoints which can happen. In my experience, the same instruments are used in Phase 2 and Phase 3. We just often pick different ones or redefine the endpoints in Phase 3 based on what we saw in Phase 2. Sometimes with good reason, sometimes out of desperation.

Expanding the target population in Phase 3 can be very dangerous, especially when protocols are amended to increase enrolment (which is always one of my most concerning red flags). It's useful to remember, though, that it is never really that close to a real life setting. In real life if your doctor does not know what treatment he/she is giving you, run away.

I think that of all the risks inherent in going into Phase 3 the two that break drugs the most are sloppiness caused by moving too fast and failing to sustain Phase 2 numbers that were in part the result of good luck.

Permalink to Comment

11. Celbio on March 1, 2011 9:56 AM writes...

In my experience, Ph2 and Ph3 endpoints have been predetermined and the same. What has varied is number of patients in total and number of arms (doses, routes, schedules), and duration of exposure which tests durability of response. This can lead to low power in Ph2, with a greater chance of hitting the endpoint with statistical significance, maybe real, maybe chance variation such as control arm had worse disease or lower placebo effect. Also have seen several times a selective reasoning that allows for Ph3 surprises, usually based upon a magnitude of response in the treatment arm, that would be impressive if only the placebo rate was lower and in line with historical rates. Much of this thinking is solved by looking at the data from a biological significance perspective rather than statistical. One can show statistically significant weight loss, for instance, with a tight study and good power. But if the effect size is only 5% of body weight, should you spend 50MM or more in Ph3? Many try in these types of scenarios.

Permalink to Comment

12. MTK on March 1, 2011 12:33 PM writes...


Can you provide a reference or link to an actual example of this?

That would seem highly unethical as you described it.

And yes, i'm skeptical, but willing to listen

Permalink to Comment

13. std on March 1, 2011 9:41 PM writes...


Sure there is no link (no records once the trial is complete). I heard that at the seminar by Mark Davis (he is a prof at cal tech) about half a year ago. And he maid it sound like it is common practice.

Permalink to Comment

14. Ed on March 2, 2011 4:08 AM writes...

A shame for Renovo - I know a couple of people who work for them, it would have been an excellent thing if the company had been successful in getting this drug to market.

Having said that the writing was pretty much on the wall a year or so ago when they layed off 2/3 of their staff, and the wife of the CEO (who was herself on the board) upped sticks to become an industry consultant in drug development.

Would like to know who it was who sold around 3 million shares about a week before the announcement of the trial failure.

Permalink to Comment

15. MTK on March 2, 2011 12:27 PM writes...


I'd take that with a grain of salt then.

I've never heard of terminally ill patients responding well to an experimental therapy showing little in terms of adverse effects, but not being allowed to continue that therapy just because the study ended.

In fact, I highly doubt that happens, much less is common.

Permalink to Comment

16. To-a-non on March 2, 2011 12:57 PM writes...

In reply to a-non's question: no, it doesn't follow, it's quite unlikely. The Phase III trials merely raise the bar for success; they don't radically change what's being tested--both are in humans. An analogy would be high jumping (hence the word choice above). Just because after raising the bar many athletes you'd thought could have succeeded do not, does not imply that those who failed at lower heights would now magically succeed. Especially if multiple tries are allowed.

Where you might have read that argument, or where it might be better applied, is in regard to animal models, in particular screening done before any clinical trial. It is likely that animal models for some diseases reject good candidates and favor bad ones. Especially, as in Alzheimers, where it is uncertain whether they model the essential feature or a side-effect of the disease.

Permalink to Comment

17. Still Scared of Dinosaurs on March 3, 2011 10:53 AM writes...

There's a medical term for the experience where "no (low) toxicity is observed, tumor growth stopped, patients feel good".

It's called "spontaneous remission".

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry