« Right Down the Alli |
Main
| Real Life, Which Costs Real Money »
June 20, 2007
Bigger, Tougher, Longer? Or Not?
Posted by Derek
Here's a question that was posed to me an an e-mail, which I thought I'd open up to everyone. Is the perception accurate that new clinical candidates (and new approved drugs) are getting more complex? And if so, are the processes used to make them getting longer and more complicated at the same rate?
I've seen the charts on the increasing molecular weight, etc., of candidates over the years, which is one surrogate for complexity. The relentless trend toward single enantiomers is probably a driver, too, so I'm certainly willing to credit the idea that the molecules are getting gradually woolier. What I'm wondering, though, is whether this is being reflected in the process work. Has anyone seen any statistics on "average number of chemical steps" or the like?
My guess is that it's been increasing, but more slowly. I think that modern synthetic methods are making up some of the difference, but I'd be interested in some actual, y'know, proof for this. Thoughts?
Comments (7)
+ TrackBacks (0) | Category: Drug Development | Drug Industry History
POST A COMMENT
- RELATED ENTRIES
- › Running Out of Helium?
- › Dealing with the Data
- › Merck's CALIBR Venture
- › Elsevier's RSS Feeds: Maybe a Not-So-Minor Complaint
- › Side Effects - Lots of Side Effects
- › Not Quite So Accelerated, Says PhRMA
- › The Blackian Demon of Drug Discovery
- › Nexavar Licensed by Force in India
1. TotallyMedicinal on June 20, 2007 9:37 AM writes...
Maybe it's a reflection on the increasing predominance of kinase research in pumping out candidates? Vieth's analysis in JMC (http://dx.doi.org/10.1021/jm0603825) indicates that the average kinase drug weighs 439, whereas the average ion channel drug only weighs 305. So more kinase candidates = more complex candidates in general.
Permalink to Comment2. jose on June 20, 2007 4:32 PM writes...
From the title, I was thinking the topic of the post would be PDE5 inhibitors...
Permalink to Comment3. totallyprocess on June 20, 2007 6:07 PM writes...
I don't think the historical data would indicate an overall increase in pharmaceutical target complexity over the last 20-30 years. Whether you measure step count, MW, # of chiral centers, or FTE per project, it does not trend towards higher complexity. Remember, you still have to contend with erythromycin (and other complex antibiotics, including the B-lactams, which are extremely difficult molecules to make), mevinolin (and other complex statins), crixivan (and other complex proteases). Some of these are semisynthetic; however, they still required multiple step manipulation of complicated, sensitive substrates with careful control and sophisticated analytical methods. I am not aware of a systematic, thorough investigation on this topic, but I have been involved in several discussions with people from various companies and I don't think that the data supports the assertion that "complexity", however one defines that, has significantly increased over the last 25-30 years. I would be interested in hearing alternate points of view.
Permalink to Comment4. CET on June 21, 2007 1:04 AM writes...
Permalink to CommentWasn't one of Lipinski's points that drug leads have gotten larger (among other things), but that the size of successful candidates hadn't changed much?
5. Spiro on June 21, 2007 7:55 PM writes...
There is a good article by a pool of authors from process groups at GSK, AZ and Pfizer about "Analysis of the reactions used for the preparation of drug candidate molecules"
OrgBiomolChem 4 (2006), 2337 dx.doi.org/10.1039/b602413k
Excerpt : " The 128 syntheses contained 1039 chemical transformations, an average of 8.1 steps (GSK 7.9, AZ 8.2, Pfizer 8.1). The distribution of stages per compound is shown in Fig. 1."
This article is a bonanza of data, but it doesn't tell us how things have evolved over time.
There is also this other article "Survey of GMP Bulk Reactions Run in a Research Facility between 1985 and 2002" at Pfizer, Groton. Again a lot of data, with a comparison of the 1985-1996 and 1997-2002 periods, but the main point is missing :-(
Org. Process Res. Dev., 2005, 9, 253–258 dx.doi.org/10.1021/op050021j
Derek, if you are interested in a copy of these articles, e-mail me.
Permalink to Comment6. Loon E. Toon on June 22, 2007 6:42 AM writes...
It seems to me that the clinical trials have, in some ways, both gotten simpler and gotten more complex.
Simpler, in that any drug candidate seems to be tested for just one possible indication, then later on tested for further, usually closely related, diseases. It almost seems as though drug companies want to do the bare minimum to be approved by the FDA, then show that it works for everything else.
More complex in that, post Vioxx (and even before) there is a concern about uncommon side effects and the statistical power to tease them out from the background. Following through for longer terms and using larger sample cohorts seems to be the norm, both of which are costly.
Permalink to Comment7. Great Molecular Crapshoot on June 23, 2007 4:01 PM writes...
I guess one could answer the question by defining some elements of molecular complexity and analysing changes in these with time for launched drugs. Number of chiral centers is an obvious measure of complexity but one could also look at extent of substitution, unusual substitution patterns, rare rings and anything else that makes sense. There are (at least) two approaches to analysis. First you can look at how the properties of launched drugs vary with launch date. A potentially more informative tactic is it to track the molecular complexity elements through the development path. Both of these approaches have been used to look at size and lipophilicity measures and I recently reviewed some of this literature in The Crapshoot in ‘The Rule of 5: Riding the wake’ (there’s a rule of 5 label). I didn’t find much convincing evidence in the reviewed literature to support the view launched drugs are getting bigger.
Permalink to Comment