Science-Based Medicine » Learning from Animals: Evolutionary Medicine with a Twist…

Okay, I can see that having a conversation with you is going to be difficult. You did not address my point about your curious dichotomy between “basic research” and some unspecified other form of research, where animal models are only predictive and useful in one of those and utterly useless in the other. Instead, you decided to blame me for misunderstanding your vague and meandering posts. It’s true — it did take three readings of your post for me to figure out that you were conceding that basic research can be done on animals, but attempting to draw a nebulous dichotomy between that and what one must conclude by context (since you did not specify it) is clinical research.

If it takes people three readings of your very long posts to even get a glimmer of your point, that is not due to their lack of interest or skill. It is because you are either a very bad writer, not terribly interested in people actually knowing what your point is. I understand you probably will not regard this as constructive criticism of your writing style, since you clearly feel quite comfortable with your prose. You are also clearly quite comfortable of your position (however badly stated it may be) on animal testing. Comfortable enough, on both counts, to completely close your mind.

Sadly, I think Dr Hall was correct. You were not coming into this thread with honest intentions. You were coming in to grind an axe. At least the axe is tangentially on topic, but it’s very unfortunate that the content of the original post in this thread has been largely lost now. The discussion on the ethics of veterinary medicine was fascinating. You stomped on it rather effectively. I will not humor you further.

You are not entirely wrong. That’s the pity. You have some good points to make, but you undermine them by going too far. I would be more disposed to take you seriously if you didn’t say so categorically that animal models are “never” predictive of human responses. That is demonstrably not true.

Indeed. Dr. Greek boxes himself in with his steadfast refusal to admit to the existence of even one good predictive animal model, as if doing so would completely invalidate his overall argument. Here’s a hint: It wouldn’t, necessarily, unless his argument is that all animal models are utterly useless and that there has never been a useful animal model to predict human responses to treatment. Clearly, he doesn’t have sufficient confidence in his own position to admit even one exception, not even transplantation, insulin, or the Blalock-Taussig shunt, to name a few. That’s some serious black-and-white, all-or-nothing thinking going on there, all gussied up with “reams of material” to make it appear nuanced. It is not. When asked to name a single good predictive animal model, Dr. Greek just can’t bring himself to do so, implying that he thinks that every “predictive” animal model ever utilized was useless. And, no, this isn’t a straw man position; it’s an inference from Dr. Greek’s arguments and his reaction to such questions. When presented with an example of an animal model that was reasonably predictive of human physiology, Dr. Greek tries very hard to redefine it as either not being a “predictive” animal model or as being “basic research,” because, clearly, to him no animal model is ever predictive of human responses.

The funny thing is, those of us sparring with Dr. Greek readily admit deficiencies and shortcomings in animal models and that some of them are actually not very good at all. However, we also point out the ones that do work. In this, we demonstrate considerably more nuanced thinking than Dr. Greek does in that he is unwilling or unable to admit even a single success from an animal model in predicting human responses. There’s the difference, and there’s the pity. In fact, I think I could, just by being a bit more flexible, argue Dr. Greek’s position better than he does and with 1/3 the verbiage.

Most people that have studied this topic, including me, say it takes quite a bit of time and effort before understanding it. I often likened learning about this topic to taking a course in organic chemistry. There is no way anyone is going to learn organic via blogs.

I wasn’t going to go there, but you decided to go there first, so now I feel justified in responding: Nor is anyone going to really learn about research, basic or translational, without actually doing basic or translational research. There is no way anyone is going to learn how to do research via blogs or just reading about it in textbooks or journals.

Neither is there anyway one is going to learn it without reams of material (the textbook). So criticizing the teacher/presenter for writing reams of material is irrational.

Nonsense. You are not being criticized for just “writing reams of material.” That is, of course, a straw man fallacy, and you’re too smart not to realize that that’s what it is. I write “reams of material” every week and only occasionally receive criticism for excessive length. No, you’re being criticized for writing reams of material of questionable relevance, much of which doesn’t actually support your arguments, and the interpretation of some of which is highly debatable. In other words, you’re being criticized for burying your “students” in material that obfuscates rather than enlightens.

Self-honesty via Dunning-Kruger demands that the readers of this exchange ask themselves if they are competent in everything listed in #1. If the reader is not, and based on the content comments, the commenters, at least, are not, then complaining about the amount of material presented is irrational. If you don’t want to learn organic chemistry, don’t take the course. But if you do, don’t go into the course assuming you know more than the teacher and then complain about how hard the course is or how much material you have to learn.

It is quite possible part way through a “course” (or from previous experience with the “teacher”) to realize that the teacher is himself a victim of the Dunning-Kruger effect and/or that he has an agenda to which he goes to great lengths to fit the material, rather than teaching the material reasonably objectively. Nobody expects complete objectivity. That is impossible in a human being. We do, however, expect that the other side will be represented accurately and fairly and that the strengths and weaknesses of one’s own argument will be recognized. You have failed at that.

One also needs critical thinking skills and an awareness of Dunning-Kruger. One needs self-honesty.

My irony meter just exploded again.

Finally, when I lecture to university classes on this topic I always state that the opposition I encounter is not primarily incompetence in general science but incompetence in critical thinking.

Dang! My backup irony meter just exploded, too!

@Dr Greek:
I appreciate that you’re willing to continue to expound upon the topic, and relatively objectively rather than take things personally. As someone who’s held extreme views, though, I’ll warn you that extremities are hard to defend (admittedly, a lesson I’ve also received in martial arts classes). If I understand the citations by others, you’re against animal usage in general. It’s something that you feel very strongly about, and it’s painful to not state that absolute in your arguments, even when you’re trying to find common ground with another person. You don’t want to compromise yourself even though you know that a minor level of compromise might help win people over to your point of you. I respect that. As someone who’s pro-life, I’ve been faced with that issue before. Frankly, there are rare cases where abortion is the right decision and there’s a minority of cases where it’s morally grey. If I take a hardline stance, I alienate the very people who’d agree with me in 95% of the cases and make them want to oppose me. I’ll admit that it’s still something I wrestle with, but as one principled person to another, take a good look at whether a change of approaches may do more good than harm.

Ray Greek MD said:
Second, most people make this exact argument and what they mean is either that animal model A did not predict human response for outcome P but did predict outcome Q and therefore was predictive for one but not the other.

I don’t think that is what they mean. You are looking at this as though everyone is just saying it happened to be right in this case, therefore it was predictive.

I will have to let your fellow doctors argue the rest in detail, but the way I have heard animal research mentioned in the past is in regards to whether or not a particular animal with a particular disease, in general, has a history of responding in similar ways to humans, and based on that assigning more or less confidence that a new animal study might be transferable to humans. I have not heard any of them refer to animal predictability with the kind of random success credit you are describing.

“if I am wrong”

You are not entirely wrong. That’s the pity. You have some good points to make, but you undermine them by going too far. I would be more disposed to take you seriously if you didn’t say so categorically that animal models are “never” predictive of human responses. That is demonstrably not true.

Tamoxifen is an excellent example of the concept that tumors in animals do not translate to tumors in humans. Hence the question regarding 90% of rats was answered. Penicillin kills 100% of certain strains of guinea pigs and hamsters. Again, this illustrates that toxicities simply do not translate in a predictive fashion. I could provide volumes on drugs that were toxic to animals but not humans. The length of the list is not important nor is the percentage of animals affected: 70%, 90%, or even 100%. The concept asked about in the question was lack of translation and I established that concept as well as answering the question very directly. Mechanisms simply do not translate in a predictive fashion. The Pharma literature supports this. Denying this concept does not change its reality.

The animal model community routinely uses the term “animal model” as a single class. When I say that animal models gave a PPV of such and such, that is what the articles says. Fault them for their terminology. I certainly do. Most papers actually commit the error of counting as a positive any animal exhibiting the toxicity in question but then do not count the negatives the same way. Few break the animal model into species or strains. But those that do, reveal PPVs in the same general area, which are far less than what is considered predictive in medical science. So when I say the animal model does not predict ADMET or efficacy or mechanisms or whatever, I mean the animal model as a paradigm and or individual species tested individually. There is no subterfuge here and as the literature conflates the term my use is actually more nuanced and hence more honest.

I do not believe PPV etc should apply to “experiments.” I believe the binomial classification table can be used to assess any modality that claims to be predictive. It would be silly to use it in judging basic research or anything not claiming to be predictive, as I have stated many times. Wikipedia use an example of drug sniffing dogs to catch drug smugglers in how to calculate PPV etc. Any modality can be assessed with this method

Water is incredibly toxic if given in high enough dose. Any physician should know this.

On to some general themes.
1. In order to understand our position a person needs to have a working knowledge of philosophy of science, the empirical evidence from many fields comparing animal and human results, the lab animal literature, complexity science, evolutionary biology with an emphasis on evo devo, the drug development literature, as well as general science principles including how to perform simple calculations from the binomial classification table. One also needs critical thinking skills and an awareness of Dunning-Kruger. One needs self-honesty.

2. Massimo Pigliucci, writing in his book “Nonsense on Stilts:”

A recurring theme of this book is that one cannot simply trust authority no matter how, well, authoritative it may appear to be. There is, unfortunately, no shortcut to using one’s brain and critical sense and doing some background research before taking a position. p90

Most people that have studied this topic, including me, say it takes quite a bit of time and effort before understanding it. I often likened learning about this topic to taking a course in organic chemistry. There is no way anyone is going to learn organic via blogs. Neither is there anyway one is going to learn it without reams of material (the textbook). So criticizing the teacher/presenter for writing reams of material is irrational. Self-honesty via Dunning-Kruger demands that the readers of this exchange ask themselves if they are competent in everything listed in #1. If the reader is not, and based on the content comments, the commenters, at least, are not, then complaining about the amount of material presented is irrational. If you don’t want to learn organic chemistry, don’t take the course. But if you do, don’t go into the course assuming you know more than the teacher and then complain about how hard the course is or how much material you have to learn.

Steven Novella wrote on 4-10-11 in his NeuroLogicaBlog the following regarding fringe beliefs:

Proponents of theories and ideologies are always looking for that knockout punch – the smoking-gun evidence that proves their beliefs in a single stroke. Most theories are too complex to be established by a single piece of evidence, and require multiple independent lines of evidence to establish them. But there are often cases in which a single solid piece of evidence can push a theory over the line to general acceptance. (Emphasis added.)

Available at http://theness.com/neurologicablog/?p=3089

Shanks and I developed a theory that explains empirical evidence from a variety of fields. Our theory (not hypothesis) draws from diverse fields and accordingly requires quite a bit of attention if one wants to understand it. Looking for a knockout punch before you understand the theory is unlikely to be fruitful unless you have an agenda.

3. Claiming someone is wrong is not the same as proving it. I stand by my offer to debate this topic under the conditions I listed in my blog. Videoing my presentation beforehand, vetting it with referees who are experts in the above-mentioned fields and having those same referees to rule on items of contention is about as fair as it gets, and would satisfy all of David’s supposed reservations about my debating techniques. The fact that he still refuses speaks volumes even if you do not have any clue what I have been writing about in this exchange.

4. If someone really wants to understand this topic a good beginning would be the books and article listed at
http://www.afma-curedisease.org/resources.html
under the Recommended Reading sections. None of these are by me. But if you do not understand the concepts in these sources there is no way you will understand our position. You cannot understand partial differential equations before taking calculus I. The sources listed are obviously minimal. Many people (myself included), will need (needed) more in order to understand the concept if it is outside your area of education.

After learning the very basics about the science, THEN read what I have written. I have not gone into the detail of say, evo devo and complexity science in those articles. They are articles that employ such concepts not articles meant to educate the reader on the concepts or fields.

5. Finally, when I lecture to university classes on this topic I always state that the opposition I encounter is not primarily incompetence in general science but incompetence in critical thinking. Very few scientists have the background in the various areas of science to address our position. That having been said, they could study the areas, become competent and then dissect our position. That is fair! But that is not what happens. No, what happens is a series of fallacies and accusations. Classic example of what we as skeptics try to combat. In his book “Hopeful Monsters,” Nicholas Mosley discusses the way the scientific community responded to the Lamarkian ideas of the biologist Paul Kammerer: “But what is striking about the objections to Kammerer on the part of mainstream biologists . . . was that they did not point out rationally, as they might so easily have done, the flaws in his arguments and procedures; they seemed intent on impugning emotionally his honesty and even his sanity; they claimed that he was ‘cooking’ his results – even those that were so obviously tentative.”

One thing is certain, if I am wrong, I am spectacularly wrong and deserve what I get. But if I am right, this exchange will be used in critical thinking courses.

So, thanks to everyone for reading all this! If anyone wants to pursue a debate in any venue or wants a copy of one of the books, please contact me. “Animal Models in Light of Evolution” was written for people with doctorates in science that don’t mind getting in to a certain amount of depth in areas of science they might not be familiar with. It is not an easy read. “FAQs About the Use of Animals in Science: A handbook for the scientifically perplexed” was written for people challenged by or uneducated in science or at least some areas of science. It is an easy read. I will probably send you one for free if you read this exchange.

Thanks to the editors of this website for allowing the discussion!
And! I promise to learn more about HTML before we do this again.

I always learn such interesting things on this blog.

http://www.nature.com/news/human-embryonic-stem-cells-restore-gerbil-hearing-1.11402

More later

Is that a promise or a threat?

My goodness. I think that I’ve figured out a new term for Dr. Greek’s “debate” style. It’s “argument by verbal tsunami.” And SBM readers occasionally criticize me for being too verbose! Talk about paragraph after paragraph after paragraph of tendentious, repetitive tsunamis of verbiage! Dr. Greek is clearly drowning out his opponents to the point where they throw up their hands and give up, concluding that countering every claim Dr. Greek tries to make—or even evaluating them to determine which are reasonable, which are not, and which are not supported by the references he cites and the arguments he makes—is just not worth the amount of work it would take. In blog discussions, it’s also a variant of threadjacking, given that Dr. Greek has, through sheer quantity of his verbiage, taken the thread in a direction that panders to his personal hobby horse. Unfortunately, Dr. Greek’s verbal waterfall (yes, I know I’m mixing metaphors) has worked with evilrobotxoxo (and probably others).

Of course, what Dr. Greek is doing is a variant of what those of us in the skeptic biz are unfortunately all too familiar with when dealing with creationists; i.e., the Gish Gallop. Sure, it’s a more sophisticated form of the Gish gallop in that Dr. Greek doesn’t just make stuff up the way Dwayne Gish used to. Also, not being able to take advantage of time limits that exist for in-person debates, he has to step up the verbal torrent a bit. So what Dr. Greek does do (and here is where he shares commonality with Dwayne Gish) is to drown his opponents in a torrent of claims, factoids, and—shall we say?—questionable interpretations of history in order to control the agenda and force his opponents to spend so much time countering them that they don’t have the time (or the energy) to make positive points of their own. Where he is more sophisticated is that, unlike the case for Gish, most of his references are legitimate scientific papers. However, his citations and examples might or might not support his point. (When examined more closely, often they do not, as those who have pointed out for specific examples.) The problem countering such an assault is that it takes far more effort to go and check to see which of Dr. Greek’s examples and citations actually do support his arguments (not very many), which ones don’t (quite a few), and which ones are spun (probably the majority). The Tamoxifen example is a perfect one. Tamoxifen is a drug whose history is best summed up by the words of the Grateful Dead, “What a long strange trip it’s been,” and by Greek’s own references it isn’t an example that refutes the point behind the question asked, although Dr. Greek does labor mightily to convince you that it does.

One also notes that Dr. Greek, when confronted with an example that he can’t counter (the development Blalock-Taussig shunt or transplantation surgery), retreats into the “let’s have a public debate” tactic so favored by those making his sorts of arguments:

After our prediction debate lets publically debate what you said about the BT shunt. For a surgeon, you are leaving out a few somewhat important details. But thats the problem, David. You will never put yourself in a position where you can be held accountable for spouting nonsense or misrepresenting facts, history, and positions.

Uh, no. Invitation declined. But thanks anyway.

Like evilrobotxoxo, whose criticism of Dr. Greek’s lapses of logic, science, and argumentation are so spot on that I wish he or she would continue to deconstruct them but fully understand why the desire isn’t there, I “have evaluated his [Greek's] claims, and I find that they simply do not merit a commitment of my time beyond writing this post.” In fact, I’m not sure that they merit even the time I’ve already spent and am spending now. However, if there’s one shortcoming I have, it’s that I sometimes can’t resist wrestling with the proverbial metaphorical pig in mud even though I just get dirty and the pig likes it. If that weren’t the case, I wouldn’t have been such a persistent blogger for the last seven and a half years. There are, I note, limits even to my penchant for getting dirty in this fashion, limits that I’ve pretty much reached right now.

In any case, I’ve written about my feelings with regard to debating people making such claims. I will paraphrase a historian whom I greatly admire: Debating people like Dr. Greek is “like trying to nail a blob of jelly to the wall.” The vast majority of the time, it’s a waste of time and effort. I trust that any science-based objective observer will see that this simile applies quite well to the prospect of debating Dr. Greek. If there’s one talent that he’s unequivocally demonstrated himself to have, it’s the talent to be the jelly that you can’t nail to the wall.

In any case, I need to leave for work; so no more fun (if any fun) until this evening.

Dr. Greek answered “Tamoxifen,” which clearly doesn’t meet that description.

And then said he would recommend human trials for a drug that killed 90% of the test animals.

Nuff said.

# Harriet Hall on 13 Sep 2012 at 6:33 pm
Tamoxifen? I don’t think so. Check your facts.

Tamoxifen, was developed as a birth control pill based on rat studies. Didn’t work in women. In fact it increased the chance for pregnancy.(1,2) No toxicities were observed in preclinical studies (1) when it was thought to be a birth control pill, but after tamoxifen was released as a cancer treatment, more studies were done and a strain of rat was shown to develop liver tumors. White writing in Current Drug Metabolism, 2003:

Tamoxifen, a non-steroidal antiestrogen, is the class representative of a group of drugs that include toremifene, droloxifene and idoxifene. Tamoxifen has been successfully used worldwide as adjuvant therapy in the treatment of women with breast cancer. However, such therapy results in a slightly increased risk of endometrial cancers. Lifetime exposure of rats to high closes of tamoxifen results in a high incidence of liver tumors. . . . The observation of Greaves et al. [28] that tamoxifen administration resulted in liver tumors in rats was quickly confirmed by other groups [29-31]. . . . At present, the rat is the only animal species where the formation of liver tumors in response to tamoxifen has been reported. No liver tumors are found in rats similarly treated with toremifene [33-35] or droloxifene [36], suggesting a mechanism of action that might be unique for tamoxifen [37,3 8].(3)

An editorial in Nature Reviews Drug Discovery 2003:

In Tamoxifen’s case, a drug first developed as a potential contraceptive languished for many years before its present application was found. Furthermore, its propensity to cause liver tumours in rats, a toxicity problem that thankfully does not carry over into humans, was not detected until after the drug had been on the market for many years. If it had been found in preclinical testing, the drug would almost certainly have been withdrawn from the pipeline.(4)

Animal studies played a minor role in the development of tamoxifen. According to D. N. Richardson of the Imperial Chemistries Industries PLC, the developer of the drug, writing in Drug Design and Delivery: “No laboratory tests for anti-tumour activity were carried out for Nolvadex [tamoxifen] until after the activity in human patients had been confirmed.” (5)

This is what I have been saying. Test enough species and strains and you will find one that gives whatever response: cancer, anticancer, hepatotoxicity, no toxicity. There is no way to know beforehand which species will mimic the human response. For that matter, there is no way to know which species will mimic which response for which humans as humans vary so much.

1. Tucker M, Adam H, Patterson J: Tamoxifen. In Safety Testing of New Drugs Laboratory Predictions and Clinical Performance. Edited by Laurence D, McLean A, Weatherall M. London: Academic Press; 1984: 125-162.
2. Jordan VC, Robinson SP: Species-specific pharmacology of antiestrogens: role of metabolism. Fed Proc 1987, 46:1870-1874.
3. White IN: Tamoxifen: is it safe? Comparison of activation and detoxication mechanisms in rodents and in humans. Curr Drug Metab 2003, 4:223-239.
4. Editorial. 2003. Follow the yellow brick road. Nat Rev Drug Discov 2 (3):167.
5. Richardson DN: The history of Nolvadex. Drug Des Deliv 1988, 3:1-14.

# Harriet Hall on 13 Sep 2012 at 6:33 pm
In fact, it undermines your argument. You are complaining about the requirement for animal testing for drug approval, saying that a drug might not be approved because it doesn’t work in the test animal even though it works in humans. Presumably tamoxifen was tested in animals and it passed the tests.

I never said the drug would not be approved I said it would not be developed. Many drugs are the market because they were developed despite the animal studies and more have had toxicities discovered after marketing. Some of the toxicities did not manifest in humans, eg tamoxifen, while others did but an animal model that mimicked the toxicity was found only in retrospect. We discuss some of the “toxic in animals but not humans” drugs in our first book. The point is that even the drug developers do not believe the animal models, which I have also mentioned. Obviously, companies vary and some will shelve a drug based on animal models quicker than others. But all agree the animal model is not predictive. Also, I am not the only one that said society has lost drugs. Pharma and the NCI have also said this.

# Harriet Hall on 13 Sep 2012 at 6:33 pm
In fact, doesn’t the approval process undermine your whole argument that animals are too different to serve as predictive models? Aren’t animal studies routinely used to predict the safety and effectiveness of drugs, allowing them to proceed to human testing?

No, my whole argument is that studies have been performed and conclusively demonstrated that animal models are not predictive for perturbations in humans, eg toxicities from the perturbation of administering drugs, and that everyone except basic researchers accept this. Consider the following.

Elias Zerhouni, former director of NIH and current head of R&D at Sanofi was quoted in the June 25, 2012 issue of Forbes as saying: “R&D in pharma has been isolating itself for 20 years, thinking that animal models would be enough and highly predictive, and I think I want to just bring back the discipline of outstanding translational science, which means understand the disease in humans before I even touch a patient.”

Sharp and Langer 2011:

The next challenge for biomedical research will be to solve problems of highly complex and integrated biological systems within the human body. Predictive models of these systems in either normal or disease states are beyond the capability of current knowledge and technology.(Sharp and Langer 2011)

A Reuters article on MSNBC discuses a computer-based method for predicting drug toxicity. The chip would test for activation of genes and proteins in various human tissues:

“If things are going to fail, you want them to fail early,” Dr. Francis Collins, the director of the National Institutes of Health (NIH), told Reuters on Friday. “Now you’ll be able to find out much quicker if something isn’t going to work.”
Collins said a drug’s toxicity is one of the most common reasons why promising compounds fail. But animal tests — the usual method of checking a drug before trying it on humans — can be misleading.
He said about half of drugs that work in animals may turn out to be toxic for people. And some drugs may in fact work in people even if they fail in animals, meaning potentially important medicines could be rejected.(Reuters 2011)

Alan Oliff, former executive director for cancer research at Merck Research Laboratories in West Point, Pennsylvania stated in 1997: “The fundamental problem in drug discovery for cancer is that the [animal] model systems are not predictive at all.” (Gura 1997)

Littman and Williams of Pfizer writing about using humans as models for other humans in Nature Reviews Drug Discovery 2005:

Experimental medicine is the use of innovative measurements, models and designs in studying human subjects for establishing proof of mechanism and concept of new drugs, for exploring the potential for market differentiation for successful drug candidates, and for efficiently terminating the development of unsuccessful ones. Humans are the ultimate ‘model’ because of the uncertain validity and efficacy of novel targets and drug candidates that emerge from genomics, combinatorial chemistry and highthroughput screening and the use of poorly predictive preclinical models . . . Experimental medicine in contemporary drug development is a business strategy that relies on experiments in humans for the purpose of demonstrating the mechanistic activity of new drugs at safe doses (exposures) and linking that activity to efficacy in patients. This strategy is very much a part of the ‘learning’ phase of early drug development that helps weed out drug failures early and precedes the ‘confirmation’ late phases of development in which high costs demand higher levels of success . . . In the new paradigm, studies in humans increase confidence in the relevance of novel drug targets and largely replace the animal efficacy models that are often poorly predictive of the efficacy of novel agents with unprecedented mechanisms of action (see below) . . . Until or unless a predictive preclinical model can represent each of these subtypes, humans will remain the ‘ultimate model organism’….
Why should experimental medicine become more involved in the validation of
novel drug targets? First, causes of attrition during the clinical phases of development have changed since the early 1990s. At that time, 40% of attrition was due to poor bioavailability and pharmacokinetics. By 2000, these factors caused the attrition of only about 10% of failed compounds. In fact, by that time 62% of compounds entering Phase II trials failed at some point prior to registration and 30% were due to lack of efficacy (Kola, I. & Landis, J. Can the pharmaceutical industry reduce attrition rates? Nature Rev. Drug Discov. 3, 711–715 (2004)). More recent data looking at the success rate of compounds between 1997 and mid-2004 suggest that about 30% of the failures occur during Phase III. For the large number of compounds with unprecedented mechanisms of action entering Phase II there are two reasons for failure due to lack of efficacy. These are inadequate pharmacology (not rigorously testing the drug target) and the lack of predictability of animal models, particularly in some therapeutic areas such as oncology and the neurosciences (Leaf, C. Why we’re losing the war on cancer — and how to win it. Fortune 9, March (2004)). The risk of expensive Phase II failure due to these two factors can be reduced in the future through appropriate human studies using modern biomarker technologies.
There are many examples of drugs that were effective in standard animal disease models but lacked efficacy in human disease. Xenograft models of cancer are a prime example . . . Even in the inflammation area, in which animal models tend to be more predictive of disease in humans, there are examples of failure due to lack of efficacy with adequate testing of the drug target. Leukotriene B4 (LTB4) antagonists, for example, were active in animal models of inflammatory arthritis, yet failed to achieve efficacy in rheumatoid arthritis (Polmar, S., Diaz-Gonzalez, F., Dougados, M., Ortiz, P. & de-Miguel, G. Limited clinical efficacy of a leukotriene B4 receptor (LTB4) antagonist in patients with active rheumatoid arthritis (RA). Arthritis Rheum. 50, S239 (2004)). In neuroscience, dopamine D4 antagonists showed activity in animal behaviour models that predicted efficacy of older dopamine D2 antagonists in schizophrenia, but these D4 antagonists were not efficacious in humans (Tarazi F. I., Zhang K. & Baldessarini R. J. Review: dopamine D4 receptors: beyond schizophrenia. J. Recept. Signal Transduct. Res. 24, 131–147 (2004)). N-methyl-d-aspartate (NMDA) antagonists seemed to be active in preclinical models of stroke but this did not translate convincingly into a therapeutic effect in humans (Ikonomidou, C. & Turski, L. Why did NMDA receptor antagonists fail clinical trials for stroke and traumatic brain injury? Lancet Neurol. 6, 383–486 (2002)).
Given these statistics and examples it makes sense to move the rationale for efficacy of novel drug targets to one based on evidence in humans. The key questions are whether this can this be done efficiently and how experimental medicine can contribute to this to provide benefit during clinical development. The hypothesis put forward here is that advances in molecular diagnostics and biomarker technologies will underpin the success of this new paradigm . . . Improved preclinical models have not materialized, and so human experimentation is still the ultimate model, although we would hope that preclinical models will in time improve. (Littman and Williams 2005)

Opar:

Drug-induced liver injury (DILI) is the primary reason why companies abandon compounds during development, and risks of idiopathic DILI even as low as 1 in 100,000 treated patients can lead companies to withdraw approved drugs. . . . “Animal models fall short, so a lot of people believe the solution is to humanize drug development,” says Paul Watkins, director of the Hamner-UNC Institute for Drug Safety Sciences and a member of the MIP-DILI external advisory committee.(Opar 2012)

Kevin Mullane of Profectus Pharma Consulting and Michael Williams of the Department of Molecular Pharmacology and Biological Chemistry, Feinberg School of Medicine, Northwestern University have written an article published in Drug Discovery Today.(Mullane and Williams 2012) The article acknowledges that advances have been made by using reductionism to study human disease and develop drugs to treat those diseases, just as I acknowledge this. However the main theme of the article is that Pharma and society are facing crisis in terms of developing new drugs that are safe, effective, and inexpensive. They cite the decrease in number of new chemical entities entering the market and late failure of many drugs in development, for example a success rate of only 5% for drugs that enter clinical trials (B. H. Munos and Chin 2011) and an 82% failure of drugs in Phase II proof of concept trials (Arrowsmith 2011), to illustrate the problem.(B. Munos 2009; Pammolli et al. 2011) They note that this is ironic since the total investment in biomedical research in the US reached $150 billion in 2010 (B. H. Munos and Chin 2011) and the amount of knowledge regarding knowledge of life has also increased substantially in the recent past. Mullane and Williams state: “The difficulties in predicting drug efficacy from preclinical models have been of concern for more than two decades . . . Thus, novel findings apparently related to the systems and targets involved in disease causality; the delineation of the efficacy, selectivity and safety of NCEs; and the predictive relevance of biomarkers and animal model data to the human disease state, even when there is evidence for target engagement in humans, all frequently fail to enhance the success rate for new drug applications (NDAs).”

They then question the notion that advances in biomedical science equate to potential advances in drug development. I too have discussed this notion that more knowledge of the material universe ipso facto means more knowledge that is useful for patients. They then discuss the very concepts I have been discussing in an attempt to understand why Pharma is in crisis:

As scientists active in TM [translational medicine] focus on defining appropriate criteria to improve decision-making and success at the preclinical and/or clinical interface, many of the basic paradigms in biomedical science that are key to these activities are being compromised. There is a failure to acknowledge the complexity of biology (Horrobin 2003; Walker 2011) to avoid confusing the simplistic, reductionist linearity of current approaches to biomedical research (Hogenesch and Ueda 2011; Kohl and Noble 2009). Added to this are significant concerns that the US research enterprise is now in crisis mode [31,32], coupled with a perception that the quality of scientific research has become ‘low’, approaching mediocrity (http://pipeline.corante.com/archives/2010/06/24/all_those_worthless_papers.php) with ‘any paper, however bad, . . .[get-]. . .published’ owing to the pressures on peer review (http://www.guardian.co.uk/science/2011/sep/05/publish-perish-peerreview-science/print). The venture capitalist Bruce Booth has commented (http://lifescivc.com/2011/03/academic-biasbiotech-failures/) that at least 50% of published studies from academic laboratories could not be repeated in an industrial setting. The prestige of the investigator or the journal did not appear to impact these numbers. An analysis (Prinz et al. 2011) by Bayer of their internal efforts to replicate published new drug target data indicated that 65% could not be reproduced to such an extent that projects had to be abandoned. A similar analysis of company-driven research programs and their reproducibility by independent third parties has not been performed, but it may not differ substantially since concerns of translation and robustness of data highlight several broad issues related to data generation, relevance, quality and transparency. These include:
(i) An over-reliance on animal models of diseases that are poorly validated in the manner they are applied. Such models are ‘validated’ either because they provide a phenotypic behavior in response to a ‘gold standard’ drug, or they represent some pathophysiological phenomenon thought to be associated with the human disease state. In the former situation, the models can only be relied upon with any assurance to identify NCEs with the same mechanism of action, whereas the latter often represents an oversimplification of the disease, where absolute belief in a mechanism often trumps any contrary data, however robust the latter. Examples of this include the T helper 2 (Th2)/eosinophil model of asthma [34], the Non-obese diabetic (NOD) mouse in diabetes [35] and the various animal models of stroke, that together, have led to over a 1000 failed compounds in the clinic [36]. Difficulties in interpreting results from animal models are far from new. However, they remain a key part of hypothesis testing provided that newer data are integrated hierarchically and taken in context with other datasets to inform broadly the validity of the hypothesis being tested [37,38].
(ii) The intrinsic reductionism of molecular biology, where engineered cell lines bearing little resemblance to native systems (or the human species) are used to define disease pathophysiology . . . (Mullane and Williams 2012)

Cook et al:

The primary purpose of preclinical therapeutic efficacy testing is to predict whether a particular compound will be successful in the clinic. Despite encouraging preclinical results, unfortunately most drugs are found to be ineffective late in their development, with only a small percentage (5%) of patients in Phase I clinical trials responding (Roberts et al. 2004). Apart front using inaccurate tumour models, there are many other reasons why preclinical studies fail to predict clinical activity. Species-specific PK, its addition to differences in drug delivery, and tumour heterogeneity might all contribute to discordant results. Such failures are costly to scientists and drug companies and of great consequence to the patients that optimistically enrol in experimental clinical trials. (Cook et al. 2012)

I can do this all day but I trust you get the point.
The FDA still requires animal testing for safety and efficacy but the federal government also requires/funds NCCAM so I would not put much stock in what the government requires.

# Harriet Hall on 13 Sep 2012 at 6:33 pm
A lot of drugs have different effects in some animals, but a lot of them have the same effects in some animals as in humans. Seems to me the lesson is that we must choose our test animals carefully, with the guidance of veterinarians who understand the similarities and differences. Cooperation with zoologists was the whole point of Zoobiquity.

The whole point here is that one cannot know which species will mimic humans until after the fact. There is some overlap when conserved processes are involved but even then, side effects vary and there are still clinically significant differences among species. See the conserved process paper. Cooperation with vets will not solve the problem. The problem is extrapolation between evolved complex systems.

More later

References for sections below tamoxifen
Arrowsmith, J. (2011). Trial watch: Phase II failures: 2008-2010. [10.1038/nrd3439]. Nat Rev Drug Discov, 10(5), 328-329.
Cook, N., Jodrell, D. I., & Tuveson, D. A. (2012). Predictive in vivo animal models and translation to clinical trials. Drug Discovery Today, 17(5/6), 253-260.
Gura, T. (1997). Cancer Models: Systems for identifying new drugs are often faulty. Science, 278(5340), 1041-1042.
Hogenesch, J. B., & Ueda, H. R. (2011). Understanding systems-level properties: timely stories from the study of clocks. [Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't
Review]. Nature Reviews. Genetics, 12(6), 407-416, doi:10.1038/nrg2972.
Horrobin, D. F. (2003). Modern biomedical research: an internally self-consistent universe with little contact with medical reality? Nat Rev Drug Discov, 2(2), 151-154, doi:10.1038/nrd1012
nrd1012 [pii].
Kohl, P., & Noble, D. (2009). Systems biology and the virtual physiological human. [Editorial
Research Support, Non-U.S. Gov't]. Molecular systems biology, 5, 292, doi:10.1038/msb.2009.51.
Littman, B. H., & Williams, S. A. (2005). The ultimate model organism: progress in experimental medicine. Nat Rev Drug Discov, 4(8), 631-638, doi:nrd1800 [pii]
10.1038/nrd1800.
Mullane, K., & Williams, M. (2012). Translational semantics and infrastructure: another search for the emperor’s new clothes? Drug Discovery Today, 17(9/10), 459-468.
Munos, B. (2009). Lessons from 60 years of pharmaceutical innovation. [Historical Article]. Nature reviews. Drug discovery, 8(12), 959-968, doi:10.1038/nrd2961.
Munos, B. H., & Chin, W. W. (2011). How to revive breakthrough innovation in the pharmaceutical industry. Science Translational Medicine, 3(89), 89cm16, doi:10.1126/scitranslmed.3002273.
Opar, A. (2012). Overtaking the DILI Model-T. [10.1038/nrd3818]. Nat Rev Drug Discov, 11(8), 585-586.
Pammolli, F., Magazzini, L., & Riccaboni, M. (2011). The productivity crisis in pharmaceutical R&D. [10.1038/nrd3405]. Nat Rev Drug Discov, 10(6), 428-438.
Prinz, F., Schlange, T., & Asadullah, K. (2011). Believe it or not: how much can we rely on published data on potential drug targets? [Comment
Letter]. Nature reviews. Drug discovery, 10(9), 712, doi:10.1038/nrd3439-c1.
Reuters (2011). U.S. to develop chip that tests if a drug is toxic. http://www.msnbc.msn.com/id/44554007/ns/health-health_care/ – .To5AMnPaixF. Accessed October 6 2011.
Roberts, T. G., Jr., Goulart, B. H., Squitieri, L., Stallings, S. C., Halpern, E. F., Chabner, B. A., et al. (2004). Trends in the risks and benefits to patients with cancer participating in phase 1 clinical trials. [Meta-Analysis
Research Support, Non-U.S. Gov't
Research Support, U.S. Gov't, P.H.S.]. JAMA : the journal of the American Medical Association, 292(17), 2130-2140, doi:10.1001/jama.292.17.2130.
Sharp, P. A., & Langer, R. (2011). Promoting Convergence in Biomedical Science. Science, 333(6042), 527, doi:10.1126/science.1205008.
Walker, M. J. (2011). The major impacts of James Black’s drug discoveries on medicine and pharmacology. [Biography
Historical Article]. Trends in Pharmacological Sciences, 32(4), 183-188, doi:10.1016/j.tips.2011.02.001.

I’ve read most of your posts here, and you fall prey to several errors in logic and/or understanding. First, you talk about the PPV and NPV of “animal models.” You explicitly deny to Dr. Gorski that you are lumping all animal models into a single class, but talking about the PPV and NPV of “animal models” is lumping all models into a class and using a statistic to describe them. People like me who work with animal models for a living spend a lot of time trying to determine the predictive validity of specific model systems. It’s not like your insight that animal models have variable predictive validity is new; the only new parts that you contribute to the discussion are incorrect, namely 1) your error in lumping them all together and 2) your mistaken belief that the criteria for diagnostic tests in a single patient should apply to scientific experiments.

Another critical error you make is in discussing quantitative estimates of the PPV and NPV in the context of drug testing. PPV and NPV are dependent on what set of compounds is being tested. You say that the PPV and NPV are around 0.5, but this is the estimate of those numbers derived from compounds that were tested in both humans and animals, i.e. the selected subset of compounds with known pharmacological activity that passed preliminary animal safety testing. Nobody knows what the PPV and NPV would be for the set of all drug candidates, or for the set of all possible compounds in the Sigma catalog. Human testing of all those compounds has never been performed and never will. However, your quoted PPV and NPV values are systematically biased to be lower than the relevant values (i.e. the PPV and NPV for all drug candidates) because drugs that are found to be overtly toxic or pharmacologically ineffective in animals are never tested in humans. The PPV and NPV for the set of all compounds in the sigma catalog would be even higher because the vast majority of those would either have no effect (like water) or be universally toxic in some obvious way (like sodium hydroxide powder), while all drug candidates simultaneously possess the possibility of toxicity (if they possess any biological activity at all) and the possibility of being nontoxic (if someone is bothering to test them).

Finally, you keep wanting to have a debate and bemoaning the fact that no one will debate you. I’m a physician-scientist who does laboratory work with animals on a daily basis, and I feel that I am qualified to debate you. However, I will not. I can’t speak for why Dr. Gorski, Dr. Hall, or anyone else would refuse, but I suspect their reasons are similar to mine. I am not intimidated by your profound intellect or afraid of being publicly shamed when you expose me as a mountebank toting a hollow idol. Less interestingly, perhaps, I have evaluated your claims, and I find that they simply do not merit a commitment of my time beyond writing this post. If there were a large subset of our society who shared your views, to the point where it threatened the integrity of the scientific enterprise (as with creationism, anti-vaxers, etc.), then I would feel an obligation to engage you. Fortunately, it appears that virtually no one cares what you think. Feel free to have the last word; I will not respond to you (or think about you) again.