[Ed note: Here is our latest piece from “Doc Gumshoe”, who is a medical writer (not a doctor) who shares his wisdom with us at Stock Gumshoe a couple times a month. He does not typically write specifically about investments, but does write about publicly traded companies and their drugs, and has agreed to our trading rules. As always, Michael’s words and opinions are his own. Michael’s previous columns can be found here.]
The cardiology community might have an excuse to do a bit of resting on their laurels, but they’re showing no signs of doing so. In spite of the real and meaningful progress that has taken place in reducing the impact of most forms of heart disease, the focus of the clinical community, as well as that of the pharmaceutical industry, has been on the half-empty part of the glass.
The half-full part of the glass includes the information that if heart disease rates had remained at their 1996 peak, there would have been about 10 million more deaths attributable to heart disease since then, and also the decline in stroke mortality, from 88.8 per 100,000 population in 1950 to about 26.5 per 100,000 in 1996. But the half-empty part of the glass tells us that in the United States there were still 614,358 deaths from heart disease in 2015. It’s the leading cause of death in the US as well as worldwide. According to the World Health Organization, about 17.5 million people died from cardiovascular disease (CVD) in 2012. This accounted for about 31% of all deaths on planet earth.
So the health-care community is well-advised to focus on the half-empty part of the glass.
On the agenda for this installment of Doc Gumshoe speculations and assertions are a couple of studies that attempt to plant some fairly important stakes in the ground. One was a study aimed at determining whether lowering low-density lipoprotein cholesterol (LDL-C) by means of adding treatment with PCSK9 inhibitors in persons who had were already on statin treatment would have any benefit in terms of specific health outcomes. The other was a study investigating whether there would be any improvement in outcomes when a patient population received antihypertensive treatment aimed at lowering their systolic blood pressure to 120 mg/hg rather than the usual target of 140 mg/hg in that specific population. Both of those studies were targeting real clinical outcomes rather than easily observable measures such as cholesterol levels and blood pressure.
Before we delve into the details of those two studies – and, what’s perhaps more important, their repercussions – let me clue you in on a surprising little fact about clinical studies in the cardiovascular disease area. It is this: studies based on real clinical outcomes are not the usual thing. Most studies of drugs intended to lower cholesterol focus almost entirely on just that – does the drug lower cholesterol. Similarly with antihypertensives, the focus is on the blood pressure readings. Those are known as surrogate endpoints, and they are widely accepted as being near-guarantees of meaningful clinical benefit. After all, there are oceans of data linking elevated cholesterol, particularly the low-density lipoprotein kind, as well as elevated blood pressure, with adverse clinical events, and oceans of data linking favorable levels in those two measures with clinical benefit. Individuals with blood pressure and cholesterol in “normal” ranges tend to have lower incidences of heart disease. But data demonstrating that, in persons who are not at elevated risk for heart disease, instituting treatment to lower blood pressure and cholesterol has clear clinical benefits is not so common.
There are considerable difficulties in designing studies requiring real meaningful clinical results as measures of success or failure. Some patients in the study need to experience genuine specific events within a specified time period, in order to be compared with the patients who did not experience any of those specific events. So, for practical purposes, the patient population in general has to be at a higher risk of experiencing those specific clinical events than the population at large; otherwise it would be necessary to recruit enormous numbers of patients in order to have a statistically significant number of those meaningful clinical events. And also, for practical purposes, the time period has to be limited – the study can’t just go on and on and on. Finally, the control group has to be managed very carefully. You can’t form a control group consisting of a bunch of bums who pay no attention to their health and life styles. The control group needs to conform to the then-current standard of care. Not only that, there has to be some benefit for the control group to take part in the study. Patients in the control group get monitored, and the clinicians conducting the study can recommend or even prescribe treatment if a need for treatment is identified. And, remember, the clinicians do not know whether or not any individual patient is getting the treatment that’s being studied.
Having gotten that under our belts, let’s take a look at the PCSK9 study.
First, what are PCSK9 inhibitors, and how do they work?
The short answer is that PCSK9 inhibitors are a new class of drugs that have the capacity to lower low-density-lipoprotein cholesterol (LDL-C) by huge percentages. PCSK9 stands for proprotein convertase subtilisin/kexin type 9, which plays a role in regulating cholesterol levels in the blood by affecting the number of receptors for LDL-C on cell surfaces. When this protein is overactive, it breaks down these receptors. LDL-C receptors in the liver absorb LDL-C particles from the bloodstream for eventual excretion. As a result, overactive PCSK9 blocks one of the pathways through which our bodies get rid of excess cholesterol. And PCSK9 inhibitors, as the name implies, counter the activity of PCSK9 and facilitate that cholesterol-removal pathway.
The discovery of PCSK9 came about accidentally, as do many discoveries in science. It started with the finding that some individuals have what might be considered freakishly low levels of LDL-C – somewhere in the range of a quarter or a fifth of the LDL-C levels in healthy individuals in the general population. In other words, if a normal healthy person has an LDL-C level around 100 milligrams per deciliter, there are a few – very, very few – persons whose LDL-C levels are down around 25 or 20 mg/dL.
After testing and analyzing every conceivable factor in these individuals, it was found that some had inherited genetic mutations in the production of PCSK9 from both parents, so that the protein that diminished the activity of the LDL-C receptors that took that form of cholesterol out of the circulation was not present in those persons. In short, no PCSK9 → undiminished LDL-C receptor activity → highly effective clearing of LDL-C from the circulation → extremely low LDL-C levels.
Doc Gumshoe did a piece about PCSK9 inhibitors just about three years ago, entitled “Will PCSK9 Inhibitors Be the Next Blockbusters in Cardiovascular Disease?” At that time there were three of these in the running – Amgen’s evolocumab, Regeneron/Sanofi’s alirocumab, and Pfizer’s bococizumab. Evolocumab and alirocumab have both been FDA-approved, as Repatha and Praulent respectively, while Pfizer’s entry got scrapped in November of 2016.
My conclusion at that time was that while there would certainly be a market for these new agents, there would have to overcome big obstacles to attain “blockbuster” status. The patients who would be good candidates for PCSK9 inhibitors would be those who failed to attain desirable LDL-C levels with statins as well as those with demonstrated intolerance to statins. A third fairly small group might be patients with an uncommon condition called heterozygous familial hypercholesterolemia, which is due to a genetic condition occurring in about 1 in 500 persons. Persons with this genetic disorder are prone to develop heart disease at an earlier age than the general population, i.e., in their 40s or even in their 30s. Even though this form of elevated LDL-C is not related to life-style, many of these individuals do respond fairly well to statin treatment, so it’s by no means obvious that even most of this group would be natural candidates for PCSK9 inhibitors.
The heterozygous population inherited the culprit gene from one parent only. Unfortunate individuals who got the bad gene from both parents – i.e., homozygous – are at even higher risk for cardiovascular disease and may have serious CV events much earlier in life, even in childhood. Homozygous familial hypercholesterolemia is extremely rare, affecting a tiny fraction of the total population, perhaps 1 in 25,000. It is doubtful whether clinical trials will be mounted in these patients, although they would surely be eligible for treatment with PCSK9 agents in an effort to do whatever possible to lower their LDL-cholesterol.
As Doc Gumshoe predicted in that piece three years ago, those patient groups that would clearly benefit from treatment with PCSK9 inhibitors were not enough to catapult those agents into the blockbuster category. One group in particular has been the subject of a good deal of contentious back-and-forth: that is the numerous and vocal cohort of patients who claim that they are not able to tolerate statins for a range of reasons, the most common of which are the muscle aches that have been attributed to statins by many patients as well as by some health workers.
Health insurers in particular are highly skeptical of assertions by patients that they are unable to tolerate statins. Statins are cheap, and if the alternative is a PCSK9 inhibitor, the difference in cost is emphatically non-trivial. PCSK9 inhibitors are given by injection, and the monthly cost for the two available agents – Amgen’s Repatha and Regeneron/Sanofi’s Praulent – is just about $1,200 per month, $14,000 per year. And that’s for life, of course. So health insurers were not eager to jump on the PCSK9 inhibitor bandwagon.
Amgen presented data at the recent American College of Cardiology meeting reporting that more than 70% of PCSK9 prescription claims presented to commercial payers were ultimately rejected. Patients covered by government plans did somewhat better – only 40% of their claims were rejected. The data presented by Amgen did not specifically look at the reasons for denial of claims, but unsubstantiated assertions by patients that they are not able to tolerate statins probably carries little weight with insurers.
Health insurers do not question that there are some persons who have significant adverse effects in response to statins. Their skepticism, which is widely shared in the clinical community, is whether the muscle aches and pains reported by many statin users are really due to statins in most cases. This skepticism, by the way, does not extend to rhabdomyolysis, a genuinely serious adverse effect which is caused, albeit rarely, by statins. Rhabdomyolysis occurs when muscle tissue is severely damaged or destroyed due to any of a number of causes, including accidents, prolonged immobility, toxins, and others. The dead muscle tissue is removed and eliminated in the urine, which typically turns dark brown. The process places great strain on the kidneys, and kidney failure can be a result of rhabdomyolysis. The incidence of rhabdomyolysis due to statins is very small, about 0.3 to 1.5 cases per 1 million statin prescriptions.
However, the reported incidence of myopathy in patients taking statins varies widely and is difficult to verify. In clinical trials, myopathy has been reported in between 1.5% and 5% of patients taking statins; however, these are usually based on patient reports and are not generally verified according to NIH standards, which define myopathy as a condition in which an enzyme called creatine kinase (CK) exceeds 10 times the upper limit of normal. CK’s function is to aid in the removal and elimination of dead tissue; since we are constantly getting rid of dead muscle tissue and replacing it with new tissue, there is always some CK present in the bloodstream. Levels of CK tend to rise after exercise, muscle strain, and muscle injury, and thus it is a useful indicator of myopathy. (One variant, CK-MB, is present in the bloodstream shortly after damage to heart tissue and is widely used in the rapid assessment of patients with suspected heart attacks.)
A recent article in Lancet (Collins R et al, Lancet 2016;388:2532-2561) goes a long way towards confirming the insurers’ skepticism regarding the link between statins and myopathy. This was a retrospective analysis of a very large trial conducted in 1998 – 2002 in more than 10,000 patients with several cardiovascular risk factors. In the first double-blind randomized phase, patients were assigned to take atorvastatin 10 mg or placebo, and were followed for three years. In this phase, the incidence of muscle-related symptoms was virtually identical whether the patients received a statin or placebo. In the second non-blinded phase of the trial, patients were offered the choice of taking a statin or not. Two thirds chose to take the statin, and in that cohort, the muscle-related symptoms were 41% more common than in the cohort not taking statins.
The authors suggest that a fundamental cause of this large difference is what they termed the “nocebo” effect. When patients were unaware that they were taking statins there was no reported increase in myopathy. But, when patients knew they were taking statins, they were more likely to report symptoms. By no means do the authors suggest that the symptoms were imaginary. The individuals experiencing those symptoms really did feel pain. They simply attributed their symptoms to the new element in their treatment protocol. As in the placebo effect, where patients feel an improvement in symptoms and attribute that improvement to the medication they are taking, even if the medication is a placebo, in this case, the patients felt an adverse effect and attributed it to their medication.
The authors also point out that the study was conducted long before the current widely disseminated view that statins do more harm than good. In the present day, the nocebo effect might well have been considerably larger.
It was probably clear from the start that persuading patients – and more important, their doctors and health insurers – to switch from statins to PCSK9 inhibitors would be a steeply uphill battle. But how about demonstrating that patients who were already on statins could derive substantial benefit by further lowering their LDL-C by adding a PCSK9 inhibitor to their regimen? That was the basis of a trial launched by Amgen, starting in 2013.
The FOURIER trial.
The full name of the trial is Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk (FOURIER). The results of this trial were announced in March of this year at the annual meeting of the American College of Cardiology (ACC) and simultaneously published online in the New England Journal of Medicine. (Sabatine MS, N Engl J Med 2017;376:1713-1722) (Why they named it after the French mathematician Jean-Baptiste Joseph Fourier, or, alternatively, after the French utopian theorist Charles Fourier, is a small mystery. Were they trying to make the trial look mathematically accurate, or perhaps utopian?)
The trial enrolled 27,564 patients with established heart disease at 1,242 clinical sites in 49 countries. Most of these patients, about 81%, had sustained heart attacks, 19% had had strokes, and 13% had peripheral arterial disease. All patients in the study were on statin therapy optimized for their clinical status, which is to say, aiming to lower their LDL-C as close as possible to the target of 70 mg/dL. Patients were randomly assigned to receive either evolocumab (Amgen’s Repatha) by subcutaneous injection, either 140 mg every two weeks or 420 mg monthly, or placebo.
The median LDL-C level in the study population at entry was 92 mg/dL, considered a desirable level in the population at large, although not meeting the 70 mg/dL target in an at-risk population. Treatment with evolocumab lowered this marker by 59% to 30 mg/dL by the 48-week observation point. Once LDL-C levels had reached the 30 mg/dL level, they remained constant for the duration of the study.
The primary endpoint in this trial was a composite of heart attack, stroke, hospitalization for angina, coronary revascularization, or cardiovascular death. This primary endpoint was observed in 9.8% of patients receiving evolocumab, and 11.3% of patients in the placebo group – a difference of 1.5% in absolute risk, translating into a 15% reduction in relative risk. In terms of individual outcomes, there was no reduction in the risk of cardiovascular mortality, nor any observed effect on the rates of hospitalization for unstable angina, hospitalization for worsening heart failure, or death from any cause. However, there was a 27% relative reduction in the risk of sustaining a heart attack and a 21% relative reduction in stroke risk.
The trial also defined a secondary endpoint, consisting of heart attack, stroke, and cardiovascular death, but omitting hospitalization for angina and coronary revascularization. Evolocumab patients experienced a relative risk reduction with regard to this endpoint that grew to 25% by the second year of treatment. Note that since there was no difference in cardiovascular deaths between the two patient groups, the reduction in MIs and strokes accounted for the entire risk reduction.
There were no differences in the incidence of adverse events between evolocumab and placebo patients, with the sole exception of injection site reactions, which were somewhat more common in evolocumab patients than in placebo patients – 2.1% versus 1.6%.
There’s no denying that the results of the FOURIER trial amount to good news; whether pretty good or just fairly good remains to be seen. Reactions to the trial results were mixed. On the positive side, no less a personage than Eugene Braunwald compared the FOURIER results to the dawn of the age of statins, suggesting that it would be truly transformative. There’s no challenging Braunwald’s eminence – he is, among other things, the editor of the standard text, Heart Disease. My well-thumbed edition runs 1,996 pages plus a 55 page index.
But there were cavils. Several clinicians pointed out that the expected relative risk reduction for the primary endpoint was in the range of 22% to as high as 30%, but in the trial it fell well short of that mark. There was also concern over the absence of any benefit in terms of either cardiovascular mortality or all-cause mortality. And the relatively small absolute risk reduction figure (1.5%) combined with the high cost of the drug ($14,000 per year) means that one would have to treat about 65 patients to prevent a single cardiovascular event at a cost of about a million dollars.
Finally, FOURIER does not address primary prevention. It does not attempt to get an answer to the question, “Will treatment with a PCSK9 inhibitor such as evolocumab reduce the incidence of cardiovascular disease in a person with cardiovascular risk factors but without any evidence of established disease?” With statins, that question has been answered in the affirmative. With PCSK9 inhibitors, not yet.
A few other PCSK9 items
Pfizer’s bococizumab: As I mentioned earlier, Pfizer’s entry in the PCSK9 sweepstakes (bococizumab) got scrapped a few months ago. However, a couple of clinical trials with bococizumab produced some interesting results. The studies, SPIRE-1 and SPIRE-2, enrolled subjects with LDL-C greater than 70 mg/dL and greater than 100 mg/dL respectively. The study demonstrated that bococizumab treatment effectively reduced the risk of cardiovascular events, but only in subjects with baseline LDL-C levels greater than 150 mg/dL. In that cohort, bococizumab reduced the risk of cardiovascular events by 25% compared with placebo. On average, bococizumab lowered LDL-C by about 55% after about 12 weeks of treatment. An unexpected development, however, is that it lost its effectiveness in about 10% to 15% of patients during the course of the trial, and also showed a wide variation in the level of LDL reduction achieved, with some patients showing only a 15% reduction and others showing an 80% reduction.
Alnylam/The Medicines Company’s inclisiram: Another PCSK9 inhibitor in the early stages of development is inclisiram from Alnylam Pharmaceuticals and The Medicines Company. The big advantage for this candidate is that it may only need to be given as seldom as twice a year. The projected cost is as yet unknown, but it is likely to be considerably cheaper than Amgen’s and Regeneron/Sanofi’s agents. A Phase 2 trial presented at the ACC meeting reported that a 300 mg dose given every three months lowered LDL-C by about 45% at the nine month point. Inclisiram may well give the other two PCSK9 agents now on the market a good run for the money.
What should the blood pressure treatment targets be?
These have always been contentious, and recent attempts to dig up some data justifying lower target blood pressure won’t end the uncertainty. The most recent guidelines, put forward in 2013 by the Joint National Committee Eight (JNC8), were as follows:
First, normal blood pressure was defined as systolic pressure under 120 mmHg and diastolic pressure under 80 mmHg. (“Systolic” BP is measured at the moment when the heart muscle contracts and pushes blood into the circulation, and “diastolic” BP is measured in between contractions; the mmHg part refers to millimeters of mercury in the glass column that the clinician scrutinizes while pumping air into the cuff.) High blood pressure or hypertension (HTN) was defined as a systolic BP greater than 140 mmHg.
JNC8 recommended the following targets:
- Adults age 30 – 59 with HTN: target BP is under 140/90 mmHg
- Adults age 60 or older with HTN: target BP is under 150/90 mmHg
- Adults with diabetes or chronic kidney disease target: BP is under 140/90 mmHg
The JNC8 recommendations permit somewhat higher BP levels than the JNC7 version, issued in 2003, which called for a maximum systolic target of 140 mmHg regardless of age, and a systolic target of 130 mmHg for adults with diabetes or chronic kidney disease. The new looser targets were met with some criticism by many clinicians who supported the position that many persons with HTN would benefit from being treated to considerably lower targets.
The SPRINT trial, a large NIH-funded clinical trial, published in November 2015 in the New England Journal of Medicine (SPRINT Research Group, N Engl J Med 2015;363:2103-2116) did indeed report data supporting the view that treating persons at elevated risk for cardiovascular disease to lower targets than those set in JNC8 would result in significant benefits. SPRINT enrolled 9,361 persons with systolic BP of 130 mmHg or higher, but who did not have diabetes. The population was randomly assigned to intensive treatment or standard treatment. Intensive treatment had a systolic BP target of 120 mmHg, while standard treatment kept to the JNC7 recommended target of 140 mmHg. The primary endpoint of the SPRINT trial was a composite of myocardial infarction, other acute coronary syndromes, stroke, heart failure, or death from cardiovascular causes.
After one year, mean systolic BP levels were 121.4 mmHg in the intensive treatment cohort compared with 136.2 mmHg in the standard treatment cohort. The trial was stopped after about 39 months when it was demonstrated that a smaller proportion of the intensive treatment group – 1.65% per year – had experienced primary endpoint events than did the standard treatment group, in whom the incidence of these events was 2.19% per year. Although the difference between the two groups was small, it was judged to be statistically significant, with a P value of less than 0.001. (The P value is the statistical probability that the event could have take place by chance. P less than 0.001 put that probability at less than one in a thousand. The cut point for attaining statistical significance is P = 0.05, or 1 in 20 that the event could have taken place by chance.) Also, lower mortality was seen in the intensive treatment cohort than in the standard treatment cohort – 155 versus 210 deaths.
Reaction to the SPRINT results was mixed. Aram Chobanian, who was the lead author of the JNC7 guidelines, agrees that a target of less than 150 mmHg is appropriate for most people with HTN, but is dubious about the 120 mmHg target, suggesting that a systolic BP target of 130 mmHg may be appropriate for hypertensives over the age of 50 who do not have diabetes or a previous stroke. Other clinicians note that the important question raised by SPRINT is how to apply the findings of the study to real patients who differ from the patient profile in SPRINT and who are not participating in a clinical trial.
An important point that many clinicians fastened on was that the subjects in the intensive treatment arm of the trial required on average 2.8 different drugs, while the standard treatment group needed 1.8 different drugs. Hypertension is routinely managed with more than one drug. In general, it works better to use drugs with different mechanisms of action and differing side effects profile. For example, in the hugely influential ALLHAT trial, many patients were treated with a diuretic, which reduces fluid volume and therefore pressure in the circulatory system, plus a beta blocker, which reduces cardiac output. Other agents lower blood pressure by blocking the calcium ion channel, which triggers contraction in the vascular system, and by inhibiting agents that directly produce contraction.
However, the potential side effects resulting from a three-drug regimen cannot be underestimated. The intensive treatment group in SPRINT experienced more serious side effects, including episodes low blood pressure, fainting (syncope), electrolyte abnormalities, and acute kidney injury, than did the standard treatment group. As distinct from the adverse effects that are commonly seen with many drug regimens, the side effects mentioned above are specifically related to the blood pressure lowering action of the antihypertensive drugs. Low blood pressure is a common cause of syncope, and diuresis can lead to electrolyte imbalance. Thus, the potential benefits of lowering BP to the desired 120 mmHg level or less must be weighed against the potential harms of an intensive drug regimen.
An issue not addressed by SPRINT was the treatment of HTN in individuals with diabetes. Diabetics have a much greater cardiovascular risk than non-diabetics, and HTN greatly increases that risk. However, at this point, there are no data showing that driving BP to less than 135 mmHg results in much benefit.
Where do we come out?
Lifting our eyes from this welter of tedious detail, what do we make of those two studies, FOURIER and SPRINT? My answer it that they are important, but far from earth-shaking. FOURIER will not catapult Amgen’s Repatha into the billion dollar category. 2016 sales for Repatha were about $141 million, so it has a way to go. Amgen is, at least for now, winning a legal battle with Regeneron and Sanofi which may boost Repatha sales in the near term.
SPRINT will not bring about epic changes in how hypertension is treated. The media response largely ignored the large effect that patient behaviors have on HTN treatment in particular. The fact is that most patients with hypertension feel just dandy. If taking the blood pressure medication is not too much of a nuisance and doesn’t lead to disagreeable or dangerous side effects, most people – once diagnosed – will go along with taking a couple of pills a day. But if to bring the blood pressure from the previously acceptable 140 mmHg level to the newly-desirable 120 mmHg level, people have to increase the number of bills they take and put up with bothersome side effects, there will be push-back and discontinuation.
The bottom line is that good doctors will tailor their treatment protocols to their individual patients. The guidelines are useful information, but not the final determination.
* * * * * * *
There’s been a little Doc Gumshoe hiatus, mostly due to my having just about a month ago undergone my second total knee replacement. Lots of you have sent me your good wishes, for which I am most grateful, and which no doubt have been helpful in speeding my recuperation. I have been in the hands of Lorenzo the Magnificent and the Cruel Christina, and they have once again worked their (painful!) magic on me. Thanks all, best, MJ (aka Doc Gumshoe)