The Bull Case for PCSK9 (AMGN, SNY, REGN, PFE): Why Estimates Are Too Low

Print Friendly, PDF & Email

Richard Evans / Scott Hinds / Ryan Baum


203.901.1631 /.1632 / .1627 richard@ / hinds@ /


April 2, 2014

The Bull Case for PCSK9 (AMGN, SNY, REGN, PFE): Why Estimates Are Too Low

  • Please see the SSR Health YouTube Channel for a detailed video walk-through of this research note
  • PCSK9 inhibitors are injectables (biweekly to monthly) that lower LDL (aka ‘bad’) cholesterol by ≥ 50%, on par with max dose statins. AMGN and SNY/REGN are closest to market, followed by PFE
  • Statin-intolerant patients generally are restricted to the use of bile acid sequestrants or Zetia. At best the sequestrants lower LDL by 25% with substantial GI side effects; Zetia lowers LDL by just 15%. PCSK9’s offer 2x – 3x efficacy to sequestrant patients, and a chance to trade GI side effects for self-injection – in many cases a very fair trade. PCSK9’s offer Zetia patients 3x the efficacy at the cost of self-injection; this is a fair trade to patients at higher near-term risk of atherosclerotic events, but less attractive to many of the rest. The US statin-intolerant market accounts for roughly 19M months of treatment annually; we believe PCSK9’s can capture 9M months of treatment in this sub segment
  • Further PCSK9 demand can be found among the 12M current statin patients who – despite already being on statin therapy – still meet the new criteria for needing high intensity (≥ 50%) LDL-C lowering. These patients’ options boil down to: 1) maxing statin dose if they haven’t already (20% add’l LDL-C reduction at most); 2) adding ezetimibe to their statin (15% add’l LDL reduction); or 3) adding a PCSK9 to their stain (≥ 50% add’l LDL reduction)
  • Of these high-risk current statin patients, 5M have LDL’s above 100 mg/dL; despite the new guidelines’ de-emphasis of specific LDL targets, we believe physicians will see the PCSK9’s as the most attractive option for these patients, even in the absence of outcomes data. If and when outcomes evidence warrants, we would expect the use of PCSK9’s to extend to the remaining high-risk statin patients whose LDLs are at or below 100 md/dl
  • ‘Pre-outcomes’, we expect 9M months of treatment demand from statin intolerant patients, combined with an additional 9M to 12M months of treatment demand from high risk statin patients with LDL’s > 100 mg/dL. At $300 – $400 / month ‘list’ ($245 – $330 ‘net’), this indicates an early PCSK9 US market of $4.4B to $6.9B
  • ‘Post-outcomes’, we expect a further 6M to 8M months of therapy, raising the total US market estimate to between $5.9B and $9.6B
  • Current consensus reflects a median 2019 global sales expectation of just $3.2B, as compared to our lower bound US expectation of $4.4B

Background – PCSK9 Inhibitors and New Cholesterol Treatment Guidelines

Amgen and Sanofi / Regeneron presented phase III LDL-C (‘bad cholesterol’) lowering data for their respective PCSK9 inhibitors (evolocumab and alirocumab, respectively) at last weekend’s American College of Cardiology Scientific Sessions. Efficacy appears comparable with a possible edge going to Amgen’s evolocumab, which reduced LDL-C by 55-57% as compared to roughly 47% for Sanofi / Regeneron’s alirocumab, though Sanofi / Regeneron’s product has shown reductions on par with Amgen’s product in smaller trials. Amgen also may have an edge with respect to dosing; its product can be given monthly (subcutaneous injection), as compared to biweekly for Sanofi / Regeneron’s product (also injected sub-q). The volume of Amgen’s monthly injection (as much as 6ml) limits the advantage of less frequent dosing; Sanofi / Regeneron may be able to limit biweekly injection volumes to a far more comfortable 1ml. Importantly, neither company has yet completed outcomes trials demonstrating whether reducing LDL-C via PCSK9 inhibition actually reduces rates of atherosclerotic vascular events. These two leading PCSK9 inhibitors are followed by numerous other PCSK9 inhibitors in earlier stages of clinical development, including Pfizer’s bococizumab (RN-316), Lilly’s LY3015014, Alnylam’s ALN-PCS02, and Novartis’ LFT-209

As the PCSK9 inhibitors are approaching the market, the consensus guidelines on how to treat hypercholesterolemia are changing significantly. Most important for the PCSK9’s is that the new guidelines throw out the ‘lower is better’ approach to LDL-C generally, and specifically throw out target LDL-C levels for various patient risk groups. This is unfortunate for the PCSK9’s; many actively treated patients were well short of their LDL-C targets under the prior guidelines, and PCSK9’s are plainly the most capable option for further LDL-C reductions on top of what a patient may already have achieved on statins. In place of the prior guidelines’ LDL-C targets, the new guidelines call for varying intensities (low, moderate, and high) of LDL-C lowering. ‘High intensity’ LDL-C lowering calls for LDL-C reductions of ≥ 50%. This level of LDL-C reduction is on par with the efficacy of these two early PCSK9 inhibitors; however because the new guidelines are anchored tightly to outcomes benefits proven in randomized controlled trials (RCTs), the guidelines call for LDL-C reductions to be achieved, wherever possible, by statins. Thus with the exception of statin-intolerant patients, the new guidelines – to the extent they guide actual practice – leave prescribers uninformed as to whether and how to use the PCSK9 inhibitors

Where PCSK9’s Fit In Part I: Statin Intolerant Patients

The most obvious immediate role for the PCSK9 inhibitors would be in statin intolerant patients; these patients currently rely largely on bile acid sequestrants[1] and Zetia[2]. At reasonably well tolerated doses the sequestrants produce very modest (10-15%) LDL-C reductions; only at higher and far less tolerable doses do the sequestrants produce LDL-C reductions in the range of 25%. Zetia, available in only one strength (10mg) produces average LDL-C reductions of only 15%. On a rolling 12 months basis, combined US demand for Zetia and the bile acid sequestrants equals roughly 19M months of therapy at an average WAC price of $171. Zetia accounts for nearly 13M of these months of therapy, and Zetia’s $185 monthly price drives the average; the less used sequestrants carry monthly prices in the $60 to $80 range (Exhibit 1). All in, including rough estimates of various discounts and rebates, we believe the US statin-intolerant market is roughly $2B annually, with the majority attributable to Zetia. The PCSK9 inhibitors offer these statin-intolerant patients substantially (roughly 3x) greater LDL-C lowering, albeit at the cost of having to self-inject. Self-injection is unlikely to be a major hurdle to the bile acid sequestrant patients who endure considerable GI discomfort; Zetia patients, who face relatively few significant adverse events, presumably will be more reluctant to self-inject

Where PCSK9’s Fit In Part II: High Risk Patients Insufficiently Treated by Statins

The new guidelines call for high intensity (≥ 50%) LDL-C lowering in several patient groups, specifically patients that: have atherosclerotic cardiovascular disease (ASCVD) and are less than 75 years of age (‘Group A’); adult patients with LDL-C ≥ 190 mg/dL (‘Group B’); and, patients aged 40-75 with LDL-C’s of 70 to 189 mg/dL, diabetes, and 10-year risk of an ASCVD event ≥ 7.5% (‘Group C’). The guidelines call for ‘moderate to high’ intensity therapy for non-diabetic patients aged 40-75 with LDL-C’s of 70 to 189 and 10-year ASCVD risk ≥ 7.5% (‘Group D’) – so Group D patients are eligible for high intensity, but the strength of the high intensity recommendation is less than for Groups A through C

Using data from the National Health and Nutrition Evaluation Survey (NHANES), we estimated how many patients are subject to the high intensity LDL-C lowering recommendation in each of these risk groups, net of patient types that were excluded from the evolocumab phase 3 trials (renal insufficiency, triglycerides > 400 mg/dL, type 1 diabetes, and uncontrolled type 2 diabetes). There are 212M non-pregnant adults aged ≥ 20, 32M of whom are currently on statins. Of these, 19M would have been excluded from the evolocumab trial because of renal insufficiency, 4M for triglycerides in excess of 400mg/dL, 1M for diabetes (Type 1), and 2M for uncontrolled diabetes (Type 2). This leaves 186M ‘non-excluded’ persons, of whom 24M are currently on statins (Exhibit 2). We then estimated the number of persons (still net of the evolocumab exclusion criteria) falling into any of the four (Groups A through D) risk groups for whom high intensity LDL-C reductions are recommended; this leaves 37M persons eligible for high intensity LDL-C lowering, 12M of whom are already on statin therapy (Exhibit 2, again)

We see these 12M persons as a key target market for PCSK9’s – they are diagnosed, and despite being under active (statin) therapy, still appear to be candidates for aggressive LDL-C lowering. Under the old guidelines, a majority of the statin-treated patients in the four high-intensity risk groups are above LDL-C targets – 42 percent (5M) are at or above the least aggressive ‘old goal’ of 100 mg/dL LDL-C; 89 percent (10M) are at or above the most aggressive ‘old goal’ of 70 mg/dL LDL-C (Exhibit 3). Confusing matters is that under the new guidelines, if these patients have achieved a 50% LDL-C reduction from whatever LDL-C levels they carried pre-treatment, then the ‘new guideline’ goals of therapy technically have been met – regardless of the absolute LDL-C level achieved

In practice, we doubt physicians will ignore a safe, effective, and reasonably well-tolerated option that produces large incremental LDL-C reductions in statin-treated high risk patients, particularly if those patients are above the higher 100 mg/dL ‘old goal’ – a category that currently holds 5M patients (Exhibit 3, again). Over time, assuming the PCSK9 inhibitors can demonstrate the outcomes benefit of pushing LDL-C’s below 70 mg/dL, they could become a standard recommendation for all 10M high risk persons who, despite being on statin therapy, still have LDL-C levels above this threshold


Product Positioning, Pricing and Sales Potential

PCSK9’s cannot compete head-to-head with high-dose statins and statin-ezetimibe combinations – the statins (and to a lesser degree the statin-ezetimibe combos) have the benefit of outcomes data, oral dosing, reasonable tolerability, and affordability (80mg atorvastatin is widely available for +/- $13 per month; Crestor 40mg and the ezetimibe combos will be available at this price or below for much of the PCSK9 products’ lifecycles). Thus to our minds there’s no question of positioning / pricing against statins of similar LDL-C lowering efficacy. Instead, we believe the best option is to bring premium efficacy and premium pricing to statin intolerant patients, and to higher risk statin-recipients who are likely to benefit from further LDL-C reductions

The US statin intolerant market currently consumes roughly 19M months of therapy (Exhibit 1, again). If we assume three-quarters of the bile acid sequestrant patients[3] and one-third of the Zetia patients[4] can be converted to PCSK9’s, we would expect 8.9M months of annual PCSK9 demand from this patient group

If we also assume high-risk statin recipients with LDL-C’s > 100 mg/dL are early (i.e. ‘pre-outcomes’ data) candidates for PCSK9’s, and that only one-quarter to one-third of these 4.9M patients receive PCSK9’s and exhibit a 60 percent compliance rate, we would expect an additional 8.8M (1/4 penetration) to 11.6M (1/3rd penetration) months of therapy in high risk statin patients

Taken together, these two markets represent 18M to 21M months of ‘pre-outcomes’ PCSK9 demand; and, provided the class is approved without significant safety restrictions, we believe this represents a reasonable basecase demand estimate. Post outcomes, we would expect the PCSK9’s to move into the high-risk statin-recipients with LDL-C’s between 70 and 100 mg/dL, which would expand the high-risk statin-recipient potential patient pool by 175%. This in turn would raise months of therapy in this population to 15.4M (1/4 penetration) to 20.3M (1/3rd penetration), and combined months of therapy across all patient types to 24M to 29M

Statin, Zetia, and ezetimibe combo US pricing (around $200 / month at ‘list’) is an extreme lower pricing bound; these products’ prices reflect their broader positioning in primary prevention markets in which the immediate risk of atherosclerotic events is relatively low – meaning demand is relatively elastic. Statin intolerant patients and high-risk statin patients with still-high LDL-C’s will be less elastic – statin intolerant patients have no options of comparable efficacy and tolerability to the PCSK9’s, and high-risk statin patients by definition face much higher risks of near term atherosclerotic events. DPP-4 inhibitors for type 2 diabetics (e.g. Januvia and Onglyza) at roughly $300 / month are a more representative lower pricing bound; the patient populations differ, but the near- to mid-term health consequences of poor blood glucose control are a reasonable proxy for the near- to mid-term risks of higher LDL-C, especially in high-risk patients. A similar case can be made for Xarelto (blood clot prevention in high risk patients, $300 / month)

Unfortunately there are no clear precedents for higher pricing of the PCSK9’s; above the $300 / month level generally requires either very tight links between drug consumption and perceived clinical symptoms (e.g. Celebrex at $400 / month, Abilify at $800 / month), and/or between drug consumption and morbidity / mortality (e.g. Atripla at $2,000 / month). We doubt the PCSK9’s can be priced significantly above $500 / month

Summarizing: Before outcomes data are available, we expect 18M to 21M months of PCSK9 demand (US only), at prices ranging from $300 to $400 / month ‘list’, which translates to roughly $245 to $330 / month ‘net’. This indicates ‘pre-outcomes’ US market potential of between $4.4B and $6.9B. ‘Post outcomes’, we expect 24M to 29M months of US demand, at the same pricing this indicates a potential US market of $5.9B to $9.6B

Consensus Timing and Sales Expectations

Consensus appears divided over launch timing, with the majority anticipating a 2016 launch. This reflects concerns either with the FDA’s recent requests for data related to potential neurocognitive adverse events; or, that the agency will refuse to approve the class in the absence of outcomes data. There is little if any public evidence of neurocognitive events in the class, so we have no way of knowing the basis of FDA’s concern – nor can we handicap whether this turns into a significant delay. Many new drug classes are reviewed and approved in the context of similar uncertainties. With respect to outcomes, we believe it is exceedingly unlikely that FDA would deny approval for lack of outcomes data, particularly given statin intolerant patients’ need for an effective and tolerable option

Median WW sales expectations for the 3 leading PCSK9’s (Amgen, Sanofi / Regeneron, Pfizer) are only $3.2B by 2019; the summed maximum PCSK9 estimate across all 3 companies for 2019 is just $5.7B. Even if PCSK9’s have no supporting outcomes evidence (the first major evolocumab outcomes trial should complete in early 2018) we would expect sales of $4.4B to $6.9B, in the US alone, with total global sales potential being roughly 1.4x to 1.6x this US figure ($6.2B to $11B). Our low-end US sales estimate matches the maximum WW consensus estimate, thus we believe consensus undervalues the PCSK9 class

  1. Cholestyramine, colestipol, colesevalam
  2. In the interest of a more conservative estimate, we’re excluding niacin and fibrates from our estimate of demand from statin-intolerant patients. We recognize that both are used by statin intolerant patients; however each is often used for primary purposes other than lowering LDL-C. Niacin often is used to raise HDL-C, and fibrates often are used to lower triglycerides
  3. Bile acid sequestrant side effects are high and efficacy is low, thus self-injecting PCSK9 should in most cases be the better option
  4. PCSK9’s offer 3x the efficacy, but require self-injection, and Zetia is tolerable – thus we assume significantly lower penetration
Print Friendly, PDF & Email