Skip to Content

PCSK9: Real World Data Arrives, Unfortunately

So, PCSK9. A wonderful story of genetics-based drug discovery, and a huge commercial opportunity. People with loss-of-function PCSK9 genes have very low LDL cholesterol, with no other ill effects, and people with gain-of-function mutations have chronically high cholesterol. That’s about as good as the validation gets, so a number of drug companies have gone charging after the target over the last ten to fifteen years.

The two big players, with antibody-based drugs on the market, are Regeneron/Sanofi and Amgen (and they’re all currently fighting it out in court). But the drugs themselves are not the gigantic blockbusters that some had predicted by this point, mainly because everyone’s been waiting for real cardiovascular outcomes data. Insurance companies are turning down prescriptions in favor of much cheaper and far better proven statin therapy, which also lowers LDL, until they can be sure that PCSK9 is worth the steep price. Sure, you can lower LDL by taking the PCSK9  drugs – but does that really translate into lower death rates from cardiovascular events? It should, according to  what most people think about LDL and heart disease, but does it really?

This morning we have three-year data from Amgen and their drug Repatha (evolocumab), an announcement that has been eagerly awaited. And it’s honestly not all that impressive. There’s a 15% relative reduction in cardiovascular risk (heart attack, stroke, etc.) relative to placebo, but investors were looking for something more over 20%. Insurance companies were probably looking for that, too, and given the price they’d have been happier to see something more like 25%. Amgen is defending the data (as quotes in this Adam Feuerstein piece show), but I don’t think that’s going to do the job. The numbers shouldn’t have to be interpreted and spun; in a three-year study with over 13,000 patients in each arm, the numbers should be able to speak for themselves, and they don’t.

Keep in mind that these are numbers relative to placebo, too – ;what would a head-to-head trial versus a generic statin show? Update: the control arm was indeed on high-dise statin therapy. Is there enough difference to justify the $14,000 list price for Repatha (or for the similarly priced Regeneron/Sanofi drug?) I have to doubt it. Look for everyone in this space to take a hit in the stock market today, and for insurance companies to continue turning down prescriptions.

Update: in a follow-up post, I’ll have more thoughts on what this means for the gene-to-drug philosophy that PCSK9 has been a key example for. . .

47 comments on “PCSK9: Real World Data Arrives, Unfortunately”

  1. Michael says:

    Hi Derek,

    You comment, “Keep in mind that these are numbers relative to placebo, too – what would a head-to-head trial versus a generic statin show?”

    Remember that these patients are, for the most part, all on statins (in both arms). Per the NEJM paper:

    “9.3% of the patients were taking high-intensity statin therapy … 30.4% were taking moderate-intensity statin therapy … 5.2% of the patients were also taking ezetimibe”

    1. Michael says:

      Missed a digit while copy/pasting – that should read “69.3% of the patients were taking high-intensity statin therapy…”, not “9.3%”

  2. David Borhani says:

    Recent analysis of statins, etc.

    Association Between Lowering LDL-C and Cardiovascular Risk Reduction Among Different Therapeutic Interventions. A Systematic Review and Meta-analysis.

  3. David Borhani says:

    Date on article on website suggests 2013, but when you open the PDF it is today’s NEJM article.

    From Table 2, “Death from any cause”, 444 (evolocumab [+statin]), 426 (placebo [+statin]), with 13,780 patient in each arm. Not good.

    Yes, on various other end points evolocumab has a better hazard ratio, e.g. myocardial infarction (haz. ratio 0.73 [0.65-0.82, 95% CI]), but the case for adding evolocumab to existing statin therapy just got a lot weaker in my opinion.

    Perhaps there truly are adverse consequences of reducing LDL-C so dramatically? (Number of genetic knockout humans is very small…).

  4. Ed says:

    Does this mean that the PCSK9 antibodies still might have some utility for very high-risk patients, but not enough utility relative to cost for the majority of lower-risk patients? IOW, that they could still be approved but will never be prescribed at blockbuster rates.

    1. Todd says:

      That seems to be the story. I could see this for people with multiple heart attacks, heart disease at <40 without obesity issues, people with strong family histories of heart disease and the like. But no one is going to become a millionaire off of these drugs.

    2. Anonymous says:

      Statins do produce side effects, such as muscle pain, in some patients. A rare side effect is rhabdomyolysis which can be lethal. Some people suffer liver damage, kidney failure, etc.. The “side effect” class might also be candidates for the Ab therapies.

      I wonder if that’s enough for a more limited trial? Only study patients who can’t tolerate conventional statins.

  5. SirWired says:

    If we were to make the bold assumption that the new drug is free from significant side-effects over just statin therapy, this might become a real “test case” for how much we are willing to pay for treatment outcomes.

    A 15% risk reduction for $14k a year is indeed a metric *bleep!*-ton of money, and there are a lot better places that $14k of healthcare spending could go. (Cue the normal song-and-dance from the manufacturers about how “nobody” pays the list prices, they’ll be offering financial assistance to those without insurance and for the copays, and insurance should cover the rest [and it will be conveniently not mentioned that insurance money does not fall from the sky.])

    At some point, some real line-drawing for cost vs. benefit is going to have to be done, and that’s a debate we’ve been avoiding in the US for far too long. The fact of the matter is that as unpalatable as it may seem, society cannot afford nearly unchecked spending increases indefinitely for marginal gains in life-years.

    This drug provides some nice, solid, numbers for those calculations, and the math isn’t going to be pretty. We might even have to do things we haven’t tried yet, like stating that this drug is going to be covered for those with uncontrolled LDL under “X” age, but won’t support the cost past that point.

    1. tangent says:

      What’s the production cost for a year’s worth of a mab like this, if we have all the possible efficiencies of scale? If we decide the drug has value but isn’t worth the price for use in the general population, then the question arises, what could the price be?

      I’m not attacking the fact of Amgen pricing the drug to cover all their fixed costs, but we ought to have the landscape of what is economically possible. (In general, even if this particular drug doesn’t really motivate it.)

      — You know, it might be an interesting point about biologicals versus small molecules. In our system, you should research whatever works (which is probably tilting towards biologicals), because production cost hardly matters. In a different system where we tried to mass-produce for mass use, we might tilt differently.

  6. Dionysius Rex says:

    Would the payers not be better off just paying the patients directly to walk for 30 minutes every day? How does $7.5k sound? Everyone’s a winner and. Except Amgen. And Sanofi.

    1. WayneTracker says:

      amen – an ounce of prevention is definitely worth $14K/yr of dubious cure.

      Derek – looking forward to your follow-up commentary on what this means for the gene-to-drug paradigm.

    2. Pennpenn says:

      Ah, gotta love the holier-than-thou attitude of the “you just need to exercise, damn you!” crowd, who whole-heartedly embrace the reductionist notion that all human ills can apparently be managed by a “brisk walk”, and that all the people these trials were just gorging layabouts and there’s no such thing as genetic prediposition (oh wait there is).

      And by “love” I mean “wish you’d shove it up whatever orifice would make you the most uncomfortable”.

      1. Dionysius Rex says:

        (Furious) straw man alert!

      2. fajensen says:

        Dispensing with Emotion and getting on with The Science: An experiment has been performed on 600 actual people over one year in Denmark.

        Quite a lot more effort is needed than “a brisk walk”. On the other hand, it does seem that following changes in lifestyle, there was quite an effect on the health and well-being of the patients.

        Diabetes 2 appears to be reversible in many of the patients participating in the experiment, as is elevated blood pressure and the ratio between “Evil/Good” cholesterol.

        These results are not bad news at all, it means that one can be cured and it also means that one can probably choose to avoid some of the nasty disabilities and afflictions that may come with ageing.

        Ried-Larsen M, Christensen R, Hansen KB, et al
        Head-to-head comparison of intensive lifestyle intervention (U-TURN) versus conventional multifactorial care in patients with type 2 diabetes: protocol and rationale for an assessor-blinded, parallel group and randomised trial
        BMJ Open 2015;5:e009764. doi: 10.1136/bmjopen-2015-009764
        Probably paywalled, I have the privilege to be inside the paywall.

        1. fajensen says:

          Head-to-head comparison of intensive lifestyle intervention (U-TURN) versus conventional multifactorial care in patients with type 2 diabetes: protocol and rationale for an assessor-blinded, parallel group and randomised trial (article)
          Ried-Larsen, Mathias and Christensen, Robin and Hansen, Katrine B and Johansen, Mette Y and Pedersen, Maria and Zacho, Morten and Hansen, Louise S and Kofoed, Katja and Thomsen, Katja and Jensen, Mette S and Nielsen, Rasmus O and MacDonald, Chris and Langberg, Henning and Vaag, Allan A and Pedersen, Bente K and Karstoft, Kristian
          BMJ Open
          2017-03-20 12:04:49 +0000
          2017-03-20 12:04:49 +0000
          BMJ Open
          Introduction Current pharmacological therapies in patients with type 2 diabetes (T2D) are challenged by lack of sustainability and borderline firm evidence of real long-term health benefits. Accordingly, lifestyle intervention remains the corner stone in the management of T2D. However, there is a lack of knowledge regarding the optimal intervention programmes in T2D ensuring both compliance as well as long-term health outcomes. Our objective is to assess the effects of an intensive lifestyle intervention (the U-TURN intervention) on glycaemic control in patients with T2D. Our hypothesis is that intensive lifestyle changes are equally effective as standard diabetes care, including pharmacological treatment in maintaining glycaemic control (ie, glycated haemoglobin (HbA1c)) in patients with T2D. Furthermore, we expect that intensive lifestyle changes will decrease the need for antidiabetic medications.Methods and analysis The study is an assessor-blinded, parallel group and a 1-year randomised trial. The primary outcome is change in glycaemic control (HbA1c), with the key secondary outcome being reductions in antidiabetic medication. Participants will be patients with T2D (T2D duration \<10 years) without complications who are randomised into an intensive lifestyle intervention (U-TURN) or a standard care intervention in a 2:1 fashion. Both groups will be exposed to the same standardised, blinded, target-driven pharmacological treatment and can thus maintain, increase, reduce or discontinue the pharmacological treatment. The decision is based on the standardised algorithm. The U-TURN intervention consists of increased training and basal physical activity level, and an antidiabetic diet including an intended weight loss. The standard care group as well as the U-TURN group is offered individual diabetes management counselling on top of the pharmacological treatment.Ethics and dissemination This study has been approved by the Scientific Ethical Committee at the Capital Region of Denmark (H-1–2014–114). Positive, negative or inconclusive findings will be disseminated in peer-reviewed journals, at national and international conferences.Trial registration number NCT02417012.

          Local Files
          Remote URLs

    3. Jen says:

      Hi there, I just happened to stumble upon this article and I have never commented on something before but I felt compelled to educate the public this time. I have Familial Hypercholesterolemia (FH…the genetic high LDL they talked about in the article). I am 44 years old and was diagnosed with atherosclerosis (plaque build up) in my 30’s. I weigh 108 and I’m 5’2”. I was born with this condition and found out I had it when I donated blood at 17 yrs old. My total cholesterol without medication hovers around 450 (LDL 330…normal should be below 100). My dad had the condition and had a heart attack at age 29. My son who is 12 also has FH with levels around mine (he is a super slender and active boy). I eat a plant based diet in order to give my body the best nutrition and exercise daily. Even with all that, my cholesterol will not budge. I take a high dose statin and am currently taking Repatha. My LDL is now 30. I have to say, it is a miracle drug and I agree, is very expensive. The drug is actually not intended for people who have high cholesterol due to diet and lack of exercise but let’s not assume that is the case for everyone. I could walk all day until I pass out (would love to get paid for that!) but my cholesterol would not lower one bit. FH occurs in 1 out of 250 people and only 90% of people that have it have not been diagnosed yet! My point is, there are expensive injections for people with MS yet no one seems to complain about those. FH is life-threatening and can be debilitating if you have a major heart attack or a stroke. I am not attacking your comment in any way, I would just like to educate the public and raise some awareness. Feel free to learn more at

  7. Chrispy says:

    The failure was not in the drug but in our understanding of the disease. Repatha was highly effective in reducing LDL cholesterol — very impressive! Unfortunately it appears that slamming down cholesterol levels is not as helpful as was hoped in this population. I have two observations:

    First, this shows the great advantage of antibodies in advancing our understanding of disease pathways. The very high specificity of antibodies means that the effects and side effects are all likely to be “on target,” in stark contrast to small molecules. Recall torcetrapib, a small molecule drug that acted on CTEP to lower LDL and raise HDL cholesterol (plus whatever side effects) — the clinical trials ended abruptly after it caused more deaths in the treated arm. No one really knows what happened there, though; it raised blood pressure a bit, so maybe that caused the deaths. Had it been an antibody to CETP we’d know pretty unequivocally that the target itself had issues.

    Secondly, as we debate about the marginal value-to-cost ratio of Repatha, it is striking that in the United States the decisions about whether to treat are left to private, profit-driven insurers. These are the same people who balked at CURING hepatitis due to the cost of Sovaldi. I’m not going to wade into the value proposition here, but what I will suggest is that these decisions should not be made by private companies.

    1. Barry says:

      As Chrispy notes, mAbs are valuable biological tools. In as much as we trust them to be specific and where the pharmacodynamics are clear, they can invalidate an hypothesis and thereby move our understanding forward in ways that small-molecules often don’t. Notably, e.g. the failure of Torcetrapib (despite stomping Pharmacodynamics) did not invalidate CETP as a target in Merck’s eyes (Merck has gone on to spend $billion on a small-molecule CETP blocker that does not increase blood pressure as torcetrapib does; it remains to be seen if it has any positive effect on morbidity and/or mortality).
      But Alzheimer’s disease offers abundant evidence that the specificity of a mAb is not enough to falsify an hypothesis. Questions as to whether the mAbs could reach plaques in the brain (or if enough of the mAb could reach the plaques) or if enough plaque was cleared leave the amyloid plaque hypothesis undisproven.

  8. Ralic says:

    I find this quite funny:

    Contrast the BBC headline to the article on Science (and of course, Derek’s commentary):

  9. Tocrat says:

    Despite these results which your learned analysis shows are hardly that encouraging, the U.K. science reportage (BBC in this case) has gone massively overboard with the positive message again. Giving the impression of another major breakthrough

    1. TX raven says:

      headline from the NYT: A drug prevents heart attacks in many with no other options left. “It’s a new ballgame,” one expert said. It’s also $14,523 a year.

      1. ajp says:

        It doesn’t matter what the BBC says because NICE won’t reimburse for it at these price levels.

  10. Noni Mausa says:

    Dear chemists: a question from a clueless layperson.

    The people in this study were taking statins as well as the PCSK9 antibodies. Is it possible that the two somehow interfered with each other? Two racehorses harnessed together are not twice as fast as one…

    1. Mira says:

      As I understand it no group was taking both statins and PCSK9 antibodies unless there were three groups. We know one group had the antibodies and one used statins as would be normal for a study like this (control group receives “best current” treatment because it would be unethical to use placebo for a deadly condition). There *may* have been a group that received both, but that’s beside the point that PCSK9 Ab wasn’t all that much better than statins.

  11. cardiosci says:

    As a cardiovascular scientist I don’t look at this result negatively. It’s actually a huge step forward for the field, just not so good for supporting expensive antibody therapy in a large patient population. There are a few key take aways that are important.

    1) The 15% was combined risk (including hospitalization for chest pain and other things), while it was 20% less likely to die from heart disease, have a heart attack or stroke. And as time goes on that number goes up with a 25% reduction being seen in the second year. These numbers are big for cardiovascular drugs, compared to something like statin+NPCL1 inhibitor (IMPROVE-IT trial), which reduced combined risk by only about 2%.

    2) PCSK9 inhibitor did not induce diabetes or cause cognitive issues compared to control group. This is a very important result as mutations in the HMG COA reductase and PCSK9 genes have been associated with increased diabetes risk (a recent NEJM paper). As statins increase diabetes risk, you really don’t want something that further increases that.

    3) Cost is a huge issue at $14K/year, or almost $1 million just to prevent a single incident in a patient, insurance companies will not like this. However, these results strongly support PCSK9 inhibition as a target, should something like a small molecule, or RNA based therapy become available at some point that can bring cost way down. Obviously not an easy task as many have tried, failed, or are still trying to achieve this, but it is likely not insurmountable either.

    1. provocateur says:

      watch for the sub-groups post-hoc/prspective analysis?Clinical results are rarely unambigous

    2. Imaging guy says:

      The antibody does not reduce cardiovascular death or death from all other causes. You have to look at table 2 in the original article. Both primary and secondary endpoints in this trial are composite endpoints. The primary endpoint consists of five individual endpoints; cardiovascular death, myocardial infarction, stroke, hospitalization for unstable angina, or coronary revascularization. It does significantly (both in clinical and statistical sense) reduce myocardial infarction, stroke and coronary revascularization but not cardiovascular death or hospitalization for unstable angina. This information is confusingly conveyed in NYT.

  12. DLIB says:

    One has to remember the need to disentangle the effect that loss of function and gain of function phenotypes to complex issues like cardiovascular with respect to the fact that these individuals have a lifetime of these genetic phenotypes that may add up over time to a larger advantage than would starting an intervention after 60 years of damage. Not trivial.

  13. Glue says:

    PCSK9 costs 14K per year or 28K for 2 years, would need to treat 66 patients to prevent one event at cost of $1,848,000/ 2 yrs, $924,000/yr

  14. Sitaroundallday says:

    Professor Hillard Kaplan, from the University of New Mexico, who led the study, said: ‘Our study shows that the Tsimane indigenous South Americans have the lowest prevalence of coronary atherosclerosis (hardening and narrowing of the arteries) of any population yet studied.
    ‘Their lifestyle suggests that a diet low in saturated fats and high in non-processed fibre-rich carbohydrates, along with wild game and fish, not smoking and being active throughout the day could help prevent hardening in the arteries of the heart.

    Remote tribe in the Amazon have the healthiest arteries ever studied

    1. Morten G says:

      Yay, Daily Mail here to save the day.

      So it’s either their food or exercise. But it’s a small, genetically homogeneous population so I’m gonna put some money on genetics? Or – hear me out – a flawed study.

    2. fajensen says:

      Maybe not working 8-10 hours per day in some god-forsaken office full of morons and then commuting with all those morons has something to do with it too?

  15. dearieme says:

    Nicola Harley in this morning’s Telegraph related the usual “Scientists say …” guff, and quoted the boasts of a Professor Sever of Imperial College. Happily she also wrote “the findings … revealed that the drug had no impact on the rate of cardiovascular mortality.”

    And yet it reduced cholesterol by gazillions of percents. You might almost wonder whether the lipid hypothesis is, to some degree or another, false.

  16. ajp says:

    No one is going to sign up for monthly injections for something preventative. It needs to mitigate bothersome symptoms and changing numbers on a blood test isn’t going to cut it. Nice work guys, now make a small molecule version that can be put in pills. Oh yes, and be ready to price it reasonably.

  17. Anon says:

    I think all this proves is that the glass is half full when it’s half empty, and vice versa. And yet people still debate the issue!

    1. JAD says:

      The glass is too large

      1. DrOcto says:

        the glass is full, half liquid, half gas

  18. Andrew Bamji says:

    The hype about PCSK9 inhibition is based on false data and false (or at least misleading) statistics. First, the entire cholesterol/heart disease hypothesis is based on a selective analysis of data from a large study. Ancel Keys chose the six countries from 22 which best fitted his theory. Had he chosen others, the theory would have been demolished at the outset. Second, this study claims a 15% reduction in morbidity from stroke/heart attack, but this is a relative reduction; the absolute reduction is nearer 1%. Third, the study shows no reduction in mortality. Fourth, it appears to have been terminated early, or at least too early for proper analysis of long term benefit, but more particularly long-term risk. Fifth, there is growing acceptance that cholesterol deposition in arterial plaque is not a primary phenomenon, but that vessel wall damage is caused by an inflammatory process that results in cholesterol deposition as an attempted healing process. Statins, according to this theory, work through their anti-inflammatory effect rather than a cholesterol-lowering effect which is essentially an epiphenomenon. PCSK9 inhibitors do not appear to have such an effect.

    As for Amazonian tribespeople having a low incidence of heart disease because of a low fat, high carb diet, that theory is confounded by their low body weight, high exercise and non-smoking status. There are several populations of high fat intake who have low heart disease risk. There is also the question of whether cholesterol-lowering drugs increase the long-term risk of diabetes and interfere with neuronal myelination, quite apart from musculoskeletal side-effects. The answer to both is – probably. The noted increase in recent years in cardiovascular disease has a closer relationship to trans-fat intake than to “bad” cholesterol levels.

    Thus – the theory is flawed, the evidence is overstated, the relationships are association and not causation. To spend £14k annually on a drug under these circumstances is neither logical nor economically justifiable.

  19. Some interesting details
    – all cause death and CVD death was non-significantly increased in the treatment arm; was this the real reason the trial was stopped early?
    – there was no benefit at all from treatment over placebo for the large European part of the trial cohorts
    – many of the composite endpoint events are unblinded diagnoses, eg LDL level is one factor you consider when advising stenting or diagnosing angina

  20. Andrew Bamji,

    that is a good comment with which I largely concur – however, the Amazonian indians had high markers of inflammation due to parasites. They also appear to have very low iron status, which is linked to CHD protection.
    I propose a different aetiology of CHD – that small defects in the vascular intima are the product of haemodynamic stress, that lack of resilience due to deficiencies in silicon, selenium, vitamin C and vitamin E (note that I am not recommending supplements here but a varied diet of unrefined food preventing deficiencies) is a factor, as is high blood pressure which increases haemodynamic stress.
    The repair of these defects involves the atherosclerotic process, but this is dysregulated by excessive post-prandial insulin (the reason TG/HDL ratio correlates so well with CHD risk), which activates the HMG-R pathway in the epithelia and increases the sdLDL fraction in lipids.
    I could add to this, thrombosis is important hence DHA, EPA, aspirin, and alcohol are protective, but you have a scenario in which atherosclerosis looks like inflammation, because it is a biological repair job gone wrong.

  21. John says:

    I’ve always been suspicious of the ldl blood pressure correlation. I really wonder if, in most cases, all we are doing is turning down a thermostat so that the number looks better, by overriding the wiring. We aren’t doing anything to the furnace. I really wonder if there is true causation here.

  22. dave w says:

    Hmmm… does the upstream correlation pan out in the first place? Do individuals with unusual LDL levels, due to natural “loss-of-function” or “gain of function” PSCK9 mutations, actually show corresponding differences in vascular health in the first place?

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.