FOURIER Shows New Cholesterol Drugs Work, But Are They Worth It?

–Doctors and patients now must wrestle with a modestly effective but expensive drug.

As it turns out the PCSK9 inhibitor saga ends not with a bang but a whimper. The results of the highly anticipated FOURIER trial show that the drugs work, though not as powerfully as many had hoped and expected. The question now will be whether the modest efficacy of the drugs is worth their immodest cost, at least for the vast majority of patients who are not at extreme high risk for cardiovascular disease.

FOURIER is the first cardiovascular outcomes trial with one of the new cholesterol-lowering PCSK9 inhibitors, in this case Amgen’s Repatha (evolocumab). A second trial, Odyssey Outcomes, is about a year behind FOURIER and is testing Praluent (alirocumab), the PCSK9 inhibitor from Sanofi and Regeneron, in 18,600 patients. Both drugs have generated enormous controversy due to their high cost and, until now, uncertain efficacy. In the absence of clinical data wide adoption of the drug has been resisted by many clinicians and, even more crucially, insurance companies and benefits managers.

Now the FOURIER trial will allow much more precise calculations of the drugs’ benefits. In the trial, which was presented today at the American College of Cardiology meeting in Washington, DC and published simultaneously in the New England Journal of Medicine, 27,564 patients with CV disease and LDL levels above 70 mg/dl already receiving statins were randomized to evolocumab or placebo. At 48 weeks LDL fell by 59%, from 92 mg/dl at baseline to 30 mg/dl

The primary end point, which was the composite of CV death, MI, stroke, hospitalization for unstable angina, or coronary revascularization, was reduced by 15%, from 11.3% (1,563 patients) to 9.8% (1,344 patients) (hazard ratio 0.85, CI 0.79-0.92, p <0.001).

The key secondary endpoint, the more rigorous composite of CV death, MI, or stroke, was reduced by 20%, from 7.4% (1,013 patients) to 5.9% (816 patients) (HR 0.80, CI 0.73-0.88, p<0.001). The investigators reported that the findings were consistent across subgroups, including those with very low LDL levels at baseline. There were no significant differences in adverse events, including new onset of diabetes and neurocognitive events, though there were more injection site reactions in the active treatment group.

The investigators also reported an improvement over time. The risk reduction in the primary endpoint increased from 12% in the first year to 19% afterwards and for the secondary endpoint from 16% to 25%. Referring to this observation, the panel chair at the presentation in Washington, Valentin Fuster (Mt. Sinai) said that “the future is brighter than the present.”

There was no significant difference in cardiovascular death or all-cause mortality. The difference in the primary endpoint was achieved through the absolute reductions of 1.5% in coronary revascularization (reduced from 7% to 5.5%), 1.2% in MI (from 4.6% to 3.4%), and .4% in stroke (from 1.9% to 1.5%).

The authors calculated that 74 patients would need to be treated for 2 years to prevent a CV death, MI, or stroke. At the ACC Marc Sabatine, the PI of FOURIER, said that the NNT at 3 years was 50.

What does it mean?

Experts with whom I spoke agreed the trial successfully demonstrated clinical benefit with the drug, but they were less certain about the general importance of this benefit. And they were concerned by the modest size of the risk reduction and the lack of any signal of a mortality benefit. But all agreed that that trial represents a remarkable and positive example of the rapid translation of genetic research to clinical therapeutics.

Sek Kathiresan (Broad Institute) said that his take-home message was that “treating with a PCSK9 inhibitor to get LDL to ultra-low levels will prevent a recurrent heart attack or stroke. But,” he asked, “at what cost?”

James Stein (University of Wisconsin) said that he sees the result as “an extra base hit in the late innings of a tight baseball game.” The results validate the LDL hypothesis and the reduction in MI and strokes in particular represents welcome news to his patients, but, he acknowledged, “the bigger question is going to be the economics of it because these drugs are so darn expensive.”

Sanjay Kaul (Cedars Sinai) wondered whether FOURIER represents an important “therapeutic advance.” He said that the high NNT “doesn’t quite cut it for me,” especially in the absence of even a signal of benefit for CV death or all-cause mortality. He said it will be “interesting to see how the payers view the value of the CV outcome benefit.” Kaul also said it would be important to to learn whether the strokes and MIs in the trial were large and clinically significant.

Kaul and Kathiresan both wrestled with the observation that the trial results do not accord with the CTT meta-analysis, which would have predicted a 30-35% decrease in events based on the LDL reduction in the trial. “What we see actually is approximately half the predicted reduction in events,” said Kaul. “Does this refute or confirm the LDL hypothesis?”

Kathiresan said that “many hypotheses will be put forward” to explain this discordance “but I suspect we are seeing a plateauing of the benefit of LDL lowering as we get to lower and lower LDL levels.” Another possible explanation for the more modest benefit in FOURIER is that statins may produce additional effects unrelated to their LDL-lowering effect. Kaul also speculated that the beneficial effect might have been larger with longer followup, “but that is a risk one takes with time to event trials.” This also suggests that the longer followup in Odyssey Outcomes may produce a more robust result.

Stein offered a more practical view. “Whether it’s above or below the CTT line isn’t all that interesting to me because almost all landmark studies are above or below the line: there’s variability and since these are the lowest lipid levels that have been achieved I’m not surprised that it’s not a perfect predictive relationship.” Stein took a step back: “the big picture is that lower is better and that anyway you can lower LDL safely will reduce events. As we seem to learn each decade, even lower than we previously thought is better.”

“We now have another potent weapon in our arsenal to lower lipids,” said Stein. The trial confirms that PCSK9 inhibitors are “very effective and, in the intermediate term, very safe.”

One warning: although the two year experience in FOURIER “suggests no major side effects,” said Kathiresan, “I would still watch for type 2 diabetes, cataracts, and neurocognitive effects as the drugs get more widely used and for longer periods of time.”

Kathiresan said that “in patients with established atherosclerotic CVD, treating to get LDL levels to <50 is a good goal if it can be done without adverse effects. We now have three options to get to this level: statin, ezetimibe, and PCSK9 inhibitors. I will probably have patients take them in that order.” But in the end it will come down to economics and not science. “There will be a vigorous debate about cost-effectiveness of PCSK9i and this will be important for patients,” he concluded.

Ethan Weiss (UCSF) agreed that “the key discussion will be around the economics of it. Does the reduction of 1-2 MIs per 100 per year and without a mortality effect justify the cost, especially with ezetimibe being basically free. How payers deal with this will be the key and I predict they will make it hard. Again, how do you identify who should get this drug? Do you wait to treat people who fail max statin. Do you have to try Zetia first? Is there a way to identify prospectively those who will benefit and justify the cost? If so, how? I do not envy the guidelines writers…”

Previous Posts:



  1. jimmy_jimmy says

    Conventional statins ~ Given for 5 Years for Heart Disease Prevention (With Known Heart Disease)

    1 in 83 were helped (life saved)
    1 in 39 were helped (preventing non-fatal heart attack)
    1 in 125 were helped (preventing stroke)

    Really? Really? how long are we going to spin these wheels?
    Just think of the vast amount of time and efforts and monies expended via statinization instead of basic research into atherosclerotic heart disease.

  2. Interesting that there is significant heterogeneity by region in the key secondary end point with benefit restricted to North America (US) and Asia -Pacific – See supplementary file S5 – p for interaction significant – The majority of patients were recruited from Europe (63%) yet there was clear interaction of less benefit in that cohort – Interesting – How will the EMA view this? The cost effectiveness may be infinitely worse?

  3. After so many hopeful and over promised breakthroughts, it seem to me that the field of cardiology is at a sobering stagnation point well proven by the back to back failures of the Apo1 Milano gene research, the failure of the HDL boosters, the failure of the promised stem cell breakthroughs and now the modest improvements of the PCSK9 inhibitors. I feel a lot of this blame comes from the cardiology field and the physicians themselves who too often use phrases like “game changers” or “breakthroughs” to announce these ideas but very few seemed to have lived up to the hype and promises. I don’t have much hope for breakthroughs as I did 5 years ago.

  4. the lipid hypothesis is BS and to the extent that ldl matters it is the particle type and whether it is oxidized but there are many factors involved and the overly simplistic bad ldl theory is a joke. half of people w/ MI have ‘normal’ ldl. lpa,

    • Cardiovascular risk in patients achieving low-density lipoprotein cholesterol and particle targets Toth PP set al Atherosclerosis 235 (2014) 585e591

      Management, utilizing LDL-P (Apo-B) Resulted in a 24% further reduction in CV endpoints: FACT

      “Results: Among 15,569 patients with LDL-P measurements, the risk of a CHD event increased by 4% for each 100 nmol/L increase in LDL-P level (HR 1.04; 95% CI 1.02e1.05, p < .0001). The comparative analysis included 2,094 matched patients with !12 months of follow-up, 1,242 with !24 months and 705 with !36 months. At all time periods, patients undergoing LDL-P measurement were more likely to receive intensive lipid-lowering therapy and had a lower risk of CHD/stroke than those in the LDL-C cohort (HR: 0.76; 95% CI: 0.61e0.96; at 12 months).

      Conclusions: In this real-world sample of commercially insured patients, higher LDL-P levels were associated with increased CHD risk. Moreover, high-risk patients who achieved LDL-P <1000 nmol/L received more aggressive lipid-lowering therapy than patients achieving LDL-C <100 mg/dL, and these differences in lipids and therapeutic management were associated with a reduction in CHD/stroke events over 12, 24 and 36 months follow-up.”

      Drugs that reduce LDL-C are not as effective at reducing LDL-P, & vice-versa
      That is a fact not a theory.

      If LDL-C < 70,.. the MAJORITY of cases will STILL have an elevated LDL-P (Apo-B). The discordance is NOT acceptable. Using LDL-C As a surrogate does not work. No matter how low the LDL-C, Not measuring LDL-P (Apo-B) and follow up, leaves the assumption that the latter are well-controlled. NOT !!

      Also, NOT measuring Lp(a) results in missing the diagnosis all together. It’s important to recognize some patients may have their risk INCREASED, and certainly at the very least, They do not have the risk reduced with statin monotherapy.

      ASTRONOMER: rosuvastatin patients who developed aortic valve disease, did far worse than those on placebo. Another fact.

      4D atorvastatin doubled the risk of stroke. This affect was seeing in 2 other dialysis trials, 3 different statins. Silbernagel et al

      4S The hyperabsorbers, a.k.a. those with elevated plant sterols, received NO BENEFIT from simvastatin, and actually saw the risk for the combined CV endpoints, increased approximately 17%.

      If you’re not measuring LDL-P (Apo-B) at baseline, AND on statin, You really have no idea if the reduction in risk is optimized.

      It's long overdue that we focus more on individualize therapy.
      LDL-C Is a poor surrogate for risk, especially in certain subset of patients.

Speak Your Mind