Skip to main content

Business and Markets

A Bit About Biopharma Investing

Well, as predicted here (and many other places), Merck announced yesterday that they are indeed not going to try to seek approval for their CETP inhibitor, anacetrapib. That’s despite an ostensibly positive Phase III trial, but it apparently wasn’t positive enough to take the risk of bringing this drug before the FDA (or to the market).

That’s an understandable decision (see that post linked above), but it should also make investors in this sector sit and think for a minute. Let’s talk about that: here’s one of the largest and most respected drug developers in the world. They announce that one of their biggest programs (for a huge market) met its primary endpoint in its most advanced clinical trial. Is that good news? You’d think so, wouldn’t you, from a distance? But that distance is what is going to lose you all your money if you keep investing from that perspective. Because (as this example shows) you need to know at least enough about the history of the CETP field and about the way that press release was worded to understand what you’re being told, beyond the phrase “met its primary endpoints”. And this is the situation that obtains with a big, respectable, well-run drug company that does not go around trying to fool people. If you want to invest in the really exciting parts of the biopharma market, you’re going to have to have your wits around you to a much greater degree.

Not to pile on these folks, but look at the recent Axovant story. How many of the bullish investors in that company realized (or really understood) that the same exact compound had failed a well-run Phase II trial in Alzheimer’s, and that since then two other drugs with the exact same mechanism of action had done the same? Given the record of Alzheimer’s clinical trials, this should have been enough to make anyone run the other way. Now, I know that there were some intelligent people who put some money down on that one, but (from what I can see) the only intelligent case to be made was that this trial was much more statistically powered than the last ones, and that there was a small chance that it might show a sliver of statistical efficacy somehow that didn’t manifest itself in the earlier work, and that this might (in the current regulatory environment) be enough to get the drug on the market – and a new Alzheimer’s drug, whether it really helps anyone in the real world or not, would most certainly sell. That was the story, and if you wanted to take some wild cash and place a bet on that possibility, you could. But it’s for sure that a lot of the money that went into raising Axovant’s stock price in the week or so before the announcement was not being invested in that manner. There were people who thought that this was a great idea, a hot stock that was going to make them serious profits, because. . .well, because they’d heard it was. And they hadn’t done their homework, not enough to realize how very heavily the odds were against that happening. Axovant’s stock, in a more rational world, would not have taken the beating it did, because in such a world it never would have risen so high in the first place.

Almost invariably, the smaller the company (and the more desperate the disease) the more you have to scrutinize the data and announcements being made. Merck certainly did not try to deceive anyone with their announcement earlier this summer about anacetrapib, but remember, mundus vult decipi: people want to deceive themselves, and they will line up and pay money to be told the stories that they want to hear. That works out pretty well at the movies or at the bookstore, but not so great in the stock market. To lapse into Latin some more, when you read the public statements made about a drug discovery project, you will encounter at times both suppressio veri, where the truth is incompletely revealed, and suggestio falsi, where things that aren’t true are made to sound as if they are. Blatant examples of either one can land people in trouble, of course, but that’s the tricky part about biopharma investing: they don’t have to look so blatant to be effective. Touting positive results in a trial that’s so underpowered that they don’t mean much, sliding over the fact that this was an open-label study rather than a double-blinded one, not going into the previous times this compound or therapy has been in the clinic, not placing the results in a context with clinical or regulatory relevance – there are a lot of ways to accomplish the same goal, and if you stick with the field you’ll see them all. Again, and again, and again.

What to do about that? Well, the sovereign remedy is to become better informed. Get smarter, learn more about drug development and about the particular therapeutic area you’re thinking about investing in. This obvious step, though, seems to be an unpleasant prospect for many small investors, so they try to go the next best route and listen to people who have put in that effort. The problem then is that if you don’t know enough to evaluate the company or its drug(s), how do you know enough to evaluate the people who are evaluating them for you? Quis custodiet ipsos custodes? Now, breaking out three or four Latin tags in the space of a few minutes will often get things thrown at you (I speak from experience here), but one thing to keep in mind is that if a mistake is old enough and common enough to have its own Latin or Greek name attached to it, it’s worth looking into. Depending on other people to tell you about biopharma stocks just kicks the fundamental problem down the dirt road a bit – there’s no substitute for learning something yourself. If that’s too time-consuming or difficult, you should go invest in something else that you understand better. Otherwise, you are inviting people to rip you off, and there is, has been, always will be a ready crowd of people to help you with the task of removing money from your pockets.

28 comments on “A Bit About Biopharma Investing”

  1. Drug Developer says:

    1. How did you miss using “Caveat emptor”?
    2. I strained my shoulder patting myself on the back for investing in several of the gene editing companies over the past year, because it’s the hot new thing, and the stocks have all shot up, but…

  2. RM says:

    you should go invest in something else that you understand better

    Then cash under the mattress it is!

  3. Emjeff says:

    This post should be required reading for Pharma investors.

  4. Hap says:

    1) Is it possible to understand enough to actually be comfortable investing in biopharma (assuming you aren’t risk-tolerant enough to jump into forest fires wearing only gasoline-soaked underwear)?

    2) Technical fields seem like hard areas to understand without lots of work, at which point you usually work in them and are already overinvested in the field or at risk for conflicts of interest. What do you do?

    3) This advice seems relevant and useful.

  5. Emjeff says:

    I went back to look at the REVEAL results, and no wonder they are not filing it. The CV event rates were 10.8% for anacetrapib and 11.8% for placebo. That is a microscopic effect size. True, the p-value was 0.004, but that is probably due to the “black hole” effect common to many outcome trials – small difference trend toward significance when the sample size becomes large. No payer would reimburse for this level of efficacy, p-value or not.

    However, Merck should not be praised too much – they basically did a large, expensive trial which will not result in a marketable product. They could have done a pre-specified look to see if there was efficacy worth pursuing, but they didn’t for some reason. I bet they wish they did now. We need to stop running large outcome trials to the end like this – it is a huge waste of resources.

    1. Ian Malone says:

      Statisticians will quite happily tell you this if you ask. Statistically significant difference means ‘we have good enough data to show a difference exists’ (formally, at pre-specified probability null hypothesis was false), but not the size of that difference. While a p value makes a simple headline figure, what are usually needed are confidence intervals. Given your trial data, what are the likely ranges for the effect size? Too small sample, wide range of possible effects in either direction. Big sample, narrow down your estimate of the mean effect.

      1. Emjeff says:

        Well, maybe. There has been an awful lot of criticism aimed at large cardiovascular outcome trials which shows similar results (i.e., small effect low p-value). I am sure that there was an army of statisticians working on REVEAL – did any of them warn the team about this well-known effect? Did anyone suggest any design features which would guard against this? Hard to say…

      2. stats1 says:

        The p value says the difference is probably significant, you don’t need to explicitly calculate confidence intervals to know this. The p-value tells you that the two means DON’T overlap with a > 99% CI (as the p value is below 0.01). The outcome is significant, its just meaningless. That is what statisticians will tell you, you can’t equate significance to being useful.

        1. Ian Malone says:

          That’s exactly my point really, you don’t need the confidence intervals to get the p value, though they correspond to doing the test at your request confidence, but what they do tell you is the likely range of effect sizes, which is after all the important bit.

          I’m sure the people working on this study were aware of the effect, but the thing is a small study is going to give you large uncertainty in what the true effect is (related to, but not the same as, low power). And even the intervals you estimate from a small study are going to be variable, which is what you do the big study for if you think there’s really a chance the thing works. Bringing us back to Derek’s article, having done that they know the effect isn’t worthwhile.

  6. Lane Simonian says:

    The problem is that you only have to convince investors that your drug is somehow different from previous ones that have failed: you have the right dose, your drug is a stronger antagonist or agonist, your drug hits additional targets, your clinical trials are better run, etc. In hindsight, the failed drug was seen to have some recently discovered fatal flaw, while the drug pursuing the same or a similar mechanism of action is viewed as being top of its class. Such was the case presented for why Axovant’s intepiridine was destined to succeed even after Lundbeck’s idalopiridine had failed.

    I expect this to be repeated time and time again in the coming years for anti-amyloid and anti-tau strategies in Alzheimer’s disease.

    There are thus many ways to explain away failures. Unfortunately, there are equally as many ways to explain away apparent successes: there was no placebo, the trial was not double-blinded, the clinical trial was underpowered. And herein lies the trap–failure with a certain approach cannot be entirely ruled out as solely due to the wrong mechanism of action and success cannot be entirely ruled in as the correct mechanism of action. But if one mechanism of action always produces failure and another mechanism of action always produces some success than that should be telling you something.

  7. Barry says:

    Pfizer was a $billion into torcetrapib when they concluded that the toxicity was inseparable from the target. Did Merck have reason to believe that the torcetrapib’s adverse effects were off-target? Are the tissue distributions different for the two cmpds?
    I’m baffled that Merck jumped late onto a target whose inhibition had already been shown not to reduce mortality.

    1. Anon says:

      Torcetrapib does induce blood pressure elevation and increase circulating levels of aldosterone (, while anacetrapib does not (

      That being said, I do not think taking anacetrapib into a Ph3 study is a good idea.

  8. Anonymous says:

    I’m not surprised that Merck jumped late into something; that was their primary strategy in the last decade.

    1. Barry says:

      Sure, Merck did very well jumping second onto something someone else had validated (although they were first in statins). But this was a target Pfizer had just spent $billion invalidating; boosting HDL by CETP blockade doesn’t reduce morbidity and mortality. I.e. it’s not a great (or even a good) surrogate marker for delta-death.

      1. Anonymous says:

        Merck has had a lot of success jumping in late and beating other people to the best medicine. One downside of this track record is the risk of convincing yourself that your medicine is somehow different and better, when it is the mechanism itself that could be the problem. These projections are hard to make, and it is super easy to criticize in hindsight.

        All that being said, I’m not surprised that Merck is making some bold and incorrect decisions based on internal belief in the strong quality of it’s programs. Keep in mind this is the same company that took Vioxx off the shelves when a lot of people suspected the issues were a class effect. It was a class effect, and that decision hurt both the company and patients.

  9. tlp says:

    i feel like the business strategy of “…vant” empire was described perfectly by xkcd

  10. Chrispy says:

    The problem with investing in biotech (smaller companies are where the real snake oil can be found) is that the market can stay nuts longer than you can stay solvent. Many, many times I have watched as companies like Sirtris and Ariad (just to name a couple) managed to sell their crappy selves to chumps, even when the science was sorely lacking. I have realized two things: first, investors in biotech tend to invest in people, not science. So if you are a big name you can sell stuff like stapled peptides (sorry, Aileron, but I don’t believe it). Second, biotech really relies upon the “bigger fool” strategy — yeah, you may be an idiot to invest in a biotech, but there may be an even bigger fool to make you look smart.

  11. You Know Who I Am says:

    Some small biotech company will license anacetrapib from Merck for $25 million, put together a pitch book and go out and raise money to conduct another P3 trial in some narrow(er) indication where they believe the drug will work and be safe. Some low-level I-Bank will bring them public via a reverse merger to the OTC and then some IR firm will promote the heck out of them. Zacks and Edison will be paid to write nice things about them in research reports and investors will be duped into buying the stock. The drug will fail and everyone will lose money except for the biopharma CEO that paid him/herself huge money while the trial was ongoing, stock options that the promo-scheme turned into money, and the I-bank that facilitated the entire thing.

    1. tangent says:

      This is what I wonder. Are they going to sell it off? Because for sure someone would buy it, and proceed something as you describe.

      If they don’t sell it, how do they justify to their stockholders that they passed up free money?

  12. anon says:

    “you should go invest in something else that you understand better”

    Not really. The stock (MRK) was $45 5 years ago and now it is $63. That’s a really nice return.

    1. anon the II says:

      That’s less than 7% annually. That’s not a nice return, especially in something as risky as pharma. If you’re gonna go cherry picking data from the past, you should at least beat a broad market index mutual fund.

      1. anon jr. says:

        To be fair, that doesn’t count dividends. When you add those in, MRK returned about 10% annually over the last five years. It’s not a stunning rate of return, but it’s not bad during a period of very low interest rates and modest inflation.

        That said, they’re still below the five-year average return for major drug manufacturers as a group (12.8%) or the S&P 500 (14.7%). A stock that doesn’t perform at least as well as the market (or even their market sector) isn’t anything to write home about.

        On the third hand, I wouldn’t call Merck a “risky” buy within the pharma world. They’re not a freshly-IPO’d small-cap with a single Phase II compound in their pipeline; they have actual products and assets that aren’t going to evaporate overnight.

    2. Derek Lowe says:

      As others have pointed out, not really. Just on a price basis, MRK is up 40% over that 5-year span, while the S&P500 is up 75%. I know which one I’d rather have owned – and the risk of owning the S&P index is a lot lower than the risk of owning one drug company. I know that Merck pays dividends, but for that matter, so does the S&P, and that’s not going to make up that big a difference in return.

  13. invest says:

    I would have thought good investment advice for anyone in the pharma biz is not to invest in the pharma biz (being that your job is already enough of an investment in this one sector). This leaves you to invest in areas outside of pharma, where you have no domain knowledge. And so you best invest in the S&P500 and leave it at that.

  14. Overtime says:

    Hmm, I think this entry could be summarized as “apply Bayesian statistics when investing in biopharma”. Simply put, the priors are so bad on CETP that one needs more than a single positive result to have confidence in it.

    That aside, should the last Latin phrase be “Quis custodiet ipsos custodes” and not “Quid …”? I didn’t take Latin in college but that is what the great Google tells me …

    1. Derek Lowe says:

      You’re absolutely right – my typos are bad enough in English, but somehow even more annoying in Latin.

  15. cytirps says:

    Anacetrapib accumulate in fat should be expected (it has 10 damn F). Merck is just too desperate to beat Pfizer. I am glad that I was retired by Merck 4 years ago. How many scientists have to be laid off in order to cover for the clinical trial cost?

    1. Anonymous says:

      Unfortunately, we will find out next year.

Comments are closed.