By Larry Gorkin, Ph.D., Gorkin & Cheddar Consulting
Payers in the U.S. are “cost-targeting” certain established blockbuster drugs via larger rebates and more generous pricing contracts with biopharmaceutical companies. Arguably, the “poster child” for cost-cutting is the curative oral drugs to treat hepatitis C, approved initially in late 2013. The total cost of a course of Gilead’s Sovaldi or Harvoni in the U.S. has been reduced to under $50,000 annually, cut approximately in half in 2016, as compared to the prices at launch. The U.S. pricing is comparable to the cost in key countries in Europe (e.g., France), which is an anomaly for a specialty drug. Although revenue for the two drugs combined was quite impressive in 2016 ($13.1 billion), this figure pales in comparison to the corresponding total in 2015 ($19.1 billion), due to a precipitous drop in U.S. 2016 pricing power.
Less evidence exists that payers are impacting the pricing of innovative pharmaceuticals at the time of launch, particularly in key therapeutic areas such as oncology and drugs to treat “orphan” indications. In 2016, the FDA approved Biogen/Ionis’ Spinraza, the first drug to treat spinal muscular atrophy, a rare neuromuscular disorder characterized by progressive muscle wasting and impairment, often leading to an early death. Biogen priced Spinraza at $125,000 per dose, yielding a $750,000 price (6 doses in Year 1) and $375,000 (3 doses) in subsequent years of treatment. This is not a “one-off”; the average pharmaceutical cost per orphan U.S. drug patient in 2016 was $140,443, and has risen from $116,216 in 2012, according to an analysis by EvaluatePharma. The 2016 cost is nearly three times the corresponding median annual household income in the U.S.
If the companies can establish a “beachhead” of high pricing at the time of launch, this mitigates the impact of marked discounts and rebates. The high price at launch also lessens the pressure to increase the price of established blockbuster drugs each year at rates that have drawn criticism. Examples include rapid rises in the costs of injected drugs to treat MS or rheumatoid arthritis. A 2017 Credit Suisse report alleges that all growth by branded pharmaceutical manufacturers in 2016 was due to price increases, not volume growth of established drugs or innovative launches.
This report will focus on the high price of acquiring biotech assets representing the next generation of innovative products. Initially, we explore how the industry has sustained high pricing, in part, by offering qualitative improvements in the design of Phase 3/4 clinical trials since 2014. Not often discussed, these changes include increasing use of: active treatment control groups, head-to-head trials against an appropriate competitive agent, and earlier presentation of clinical outcomes rather than surrogate measures.
Improving Research Design of Phase 3/4 Research Programs
There was a time when generating blockbuster drugs was relatively easy. Me-too drugs did not have to be compared to one another in head-to-head trials for either regulatory or commercial success. Instead, the placebo-only controlled study had a long tradition for clinical investigations of developmental drugs, including Phase 3. In the years from 1998 to 2004, the rate of experimental drugs becoming blockbusters in the U.S. was three times higher as compared to the corresponding period from 2005 to 2011, based on data supplied by drugs.com (accessed in 2012).
This downward shift is explained primarily by the finding that by 2005, the two-decade practice of developing undifferentiated, me-too drugs to grow into blockbuster drugs was no longer successful. This finding is due to the efforts of managed care in the U.S. and key countries internationally. Evaluating U.S. drug sales in 2004, for example, there were 33 blockbuster drugs listed with $1 billion or more in revenue, and 2/3 of these shared a mechanism of action with another drug on the list (multiple drugs for, e.g., schizophrenia, hypertension, depression, acid reflux). Since then, the number of blockbuster drugs that share a common mechanism of action has dropped markedly, according to an analysis by Light and Lexchin (2012), as a more disciplined managed care in the U.S. prevented me-too drugs from gaining traction commercially.
For an update, since drugs.com did not have revenue data past 2013, a database of 2016 global sales was utilized. Rather than a $1 billion criterion based on U.S. sales, required worldwide sales were set at $2 billion in 2016, doubling the criterion that defined blockbuster inclusion.1 This analysis revealed that 54 drugs achieved the global $2 billion criterion in 2016, and less than 1/3 of these drugs shared a common mechanism of action. Notwithstanding a significant increase in the sample size from the 33 U.S. blockbuster drugs in 2004, a shared mechanism of action was twice more common in the 2004 listing. Thus, the concerted efforts of payers have led to the increased innovation within the industry, which has been evident since 2010, according to an analysis I conducted previously.
One way managed care accomplished this shift away from me-too branded drugs was to funnel patients to the increasing array of former blockbusters that had lost patent protection. In 2004, the U.S. Health & Human Services noted that 57 percent of all scripts written in the U.S. were for generic drugs, but this has grown rapidly to about 90 percent by the end of 2016. As the increasing number of blockbuster drugs lose patent protection, patients are encouraged financially to consume these lower cost, first-line generic agents. Industry has adapted to this change, across many disease indications, by not competing with the high-efficacy, inexpensive generic agents. Instead, manufacturers develop and market a second-line branded drug for difficult-to-treat patients.
This has implications for drug development, as one may still gain regulatory approval in the U.S. with two placebo-only controlled trials. An example of this effect involves the treatment of type 2 diabetes, where American guidelines consistently specify efficacious and inexpensive, generic metformin as first-line therapy. This has not prevented the development of successful drug classes among patients who need additional therapeutics beyond metformin. These include: (1) the oral dipeptidyl peptidase 4 (DPP4), with $8.8 billion in global sales, led by Merck’s Januvia (sitagliptin) and (2) the oral sodium glucose co-transporter-2 (SGLT2) drugs, led in sales by JNJ’s Invokana.
Given this type of Phase 3 research design (e.g., Januvia versus placebo), having metformin as active background therapy for all patients randomized avoids the anathema placebo-only control group. An active treatment for all patients also removes the ethical concern regarding withholding treatment via a placebo-only design. Therefore, the results of these studies are helpful in establishing a therapeutic hierarchy of clinical treatment lines for payer and physician stakeholders.
Since 2014, this trend of combining innovative developmental drugs with inexpensive generic agents, and sometimes with another branded product, has expanded to several clinical indications. By avoiding placebo only designs, manufacturers have improved the quality of clinical trials conducted for regulatory and commercial success. Here are some examples: (1) Regeneron/Sanofi tested the PCSK9 biologic, Praluent (alirocumab), versus placebo, in combination with a high-potency statin (e.g., Lipitor) to reduce cholesterol; (2) Regeneron’s Dupixent (dupilumab) was tested in a placebo-controlled trial with a standard of care (SoC), topical corticosteroids applied to affected areas of the skin for all patients, in moderate-to-severe atopic dermatitis; (3) Novartis’ Kisqali (ribociclib) was approved first-line to treat post-menopausal women with hormone receptive-positive/human epidermal growth factor receptor 2 (HER2)-negative, advanced breast cancer. Kisqali was tested in combination with SoC letrozole versus placebo plus letrozole, an aromatase inhibitor established in this type of breast cancer; (4) Amgen demonstrated improved progression-free survival when Krypolis (carfilzomib), a proteasome inhibitor, was added to SoC regimen of Revlimid (lenalidomide) plus dexamethasone to treat relapsed multiple myeloma versus the SoC regimen only; (5) Newron Pharmaceutical’s Xadago (safinamide), an inhibitor of monoamine oxidase B, was approved as a treatment for inadequately controlled Parkinson’s disease (PD) among patients who received carbidopa/levodopa and other SoC PD medications, in placebo-controlled studies; (6) in asthma, GlaxoSmithKline (GSK) is testing a once daily triple therapy versus the more traditional double therapy. The control group is the once daily combination of an inhaled corticosteroid, fluticasone furoate, with a long-acting beta2-adrenergic agonist, vilanterol, known as GSK’s Relvar/Breo (i.e., a variation on GSK’s Advair, which has lost patent protection, along with its inhaler delivery system). The experimental drug is a long-acting muscarinic antagonist, unmeclidinium, which is combined with Relvar/Breo as background therapeutics; (7) in chronic obstructive pulmonary disease (COPD), GSK is testing the same triple therapy combination as in asthma, versus a similar double therapy combination, although using AstraZeneca’s Symbicort (an inhaled corticosteroid, budesonide, with a long-acting beta2-adrenergic agonist, formoterol) as the control group; (8) among patients with tophaceous gout, AstraZeneca investigated lesinurad, a renal urate transporter inhibitor, versus placebo, using xanthine oxidase inhibitor, febuxostat, a medication approved as cost-effective in the United Kingdom, as background therapy; (9) Merck’s biologic, bezlotoxumab, providing passive immunity to toxin B, was tested in a placebo-controlled trial to treat c difficile, combined with SoC antibiotic therapy; and (10) Roche’s Perjeta plus Herceptin (which bind to different regions of the HER2 receptor) and an aromatase inhibitor proved more effective in terms of disease-free progression in metastatic hormone receptor-positive/HER2-positive metastatic or locally advanced early breast cancer, as compared to Herceptin and aromatase inhibitor only.
Improving research design by avoiding the placebo only control group, though, doesn’t meet the standard of a head-to-head clinical trial, which stakeholders ideally want to compare an experimental treatment against an appropriately dosed, SoC drug in a randomized controlled trial in Phase 3/4. Manufacturers have been hesitant to conduct such clinical trials, and have been given a pass by regulators who accept placebo-only studies as sufficient for regulatory submission. Obviously, companies run the risk of losing when conducting an appropriately designed head-to-head clinical trial. In an analysis published in the Journal of the American Medical Association of 328 studies conducted over a 16-month period between June 2008 and September 2009, only 45 of these studies involved more than one active treatment. Thus, only 13.7 percent of the overall sample provided any hope of differentiating which available clinical trial treatment is best for a particular patient population. The percentage of head-to-head trials was significantly lower than this figure.
In a review of rheumatoid arthritis (RA) studies published up to 2015, Fleischmann, et al. (2016) note that only a handful of studies involved such double-blinded head-to-head trials of novel, mostly biologic, treatments. This conclusion was consistent with a 2012 review of RA studies that pointed to only 5 of 91 being head-to-head studies. None of the specialty drugs approved to treat RA, starting with the TNF-alpha blockers in the late 1990’s through 2015, utilized a head-to-head, Phase 3 clinical trial. This includes the three TNF-blockers, Humira, Enbrel, and Remicade, and other RA drugs that contribute to an estimated global pharmaceutical market approaching $20 billion annually in 2017.
Since 2014, though, the industry has been more willing to conduct appropriately dosed, head-to-head clinical trials at significantly higher rates than in prior years. The rationale for this began with the realization that most me-too drugs did not provide an adequate ROI. With the focus on innovation, the industry began to slowly learn that the quality of drug being developed could withstand competition from active control agents.
To address this claim empirically, ideally, one would generate a list of criteria representing what constitutes an appropriate head-to-head clinical trial related to a SoC comparative agent, at appropriate dosing. Patients would be randomly assigned to treatment conditions, either double-blinded, or if not, with clinical outcomes (e.g., overall survival, hospitalization) that are markedly less vulnerable to bias than are surrogate measures. In this report, however, limitations in time and funding, prevented such an analysis. Accordingly, a circumstantial argument is made, contrasting recent information on increasing use of head-to-head trials with previous results from systematic reviews noting the lack of head-to-head studies.
In less than a two-year period, involving 2016 and 2017, in contrast, five RA specialty drugs conducted head-to-head clinical trials against a SoC biologic. Three of these included a head-to-head trial as part of their regulatory submission package. Two others conducted a head-to-head Phase 4 trial years after FDA approval: (1) In a study of methotrexate-intolerant patients, JNJ’s sirukumab was tested versus appropriately dosed Humira (adalimumab), a leading tumor necrosis factor (TNF)-alpha blocker; (2) Lilly’s oral Janus kinase (JAK) inhibitor, baricitinib, awaiting FDA approval after a “complete letter,” conducted a 1,300-patient study versus Humira, with methotrexate as background therapy; (3) Pfizer randomized patients to the oral JAK inhibitor, Xaljanz, with or without methotrexate, or Humira plus methotrexate (the preferred clinical regimen); (4) Regeneron/Sanovi’s Kevzara (sarilumab), an interleukin-6 inhibitor, was approved in 2017 for patients with an inadequate response to one of more DMARDs, based, in part, on a Phase III trial conducted versus Humira (each monotherapy), among patients with less than optimal response to methotrexate; (5) UCB sponsored a trial randomizing patients who had an inadequate response to methotrexate to either Cimzia (certolizumab pegol-CZP) or Humira. These findings diverge markedly with the pattern regarding the lack of head-to-head trials observed for nearly two decades.
It is probably not a coincidence that all five trials utilized Humira as the comparative agent. Although some consider Amgen’s Enbrel the most effective TNF-alpha-blocking biologic to treat rheumatoid arthritis, in the absence of head-to-head clinical trials, Humira is the number one ranked pharmaceutical in global revenue ($4.12 billion in 1Q17).
Examples of recent head-to-head trials include: (1) Novartis’ Entresto (sacubitril/valsartan) was found to reduce re-hospitalization relative to SoC, generic enalapril, among patients with heart failure and low left ventricular ejection fraction levels; (2) Merck’s Keytruda (pembrolizumab), a PD-1 inhibitor, was approved as a frontline treatment of advanced melanoma based on overall survival results as compared to Bristol-Myers’ Yervoy (ipilimumab), a CTLA-4 inhibitor; (3) in pulmonary hypertension, GSK and Gilead sponsored the AMBITION study, which investigated a novel dual combination, Letairis (ambrisentan), a selective endothelin-A–receptor antagonist, and Lilly’s Cialis (tadalafil), a phosphodiesterase type 5 inhibitor, known primarily as a treatment for erectile dysfunction. The control groups consisted of the two branded drugs, each as monotherapy; (4a) Lilly’s Taltz (ixekizumab), which blocks IL-17, established higher efficacy when tested against JNJ’s Stelara (ustekinumab), which blocks IL-12/IL-23, in a randomized study among patients with moderate-to-severe plaque psoriasis, utilizing a stringent criterion of 90 percent change in the Psoriasis Area Severity Index (PASI); (4b) Valeant’s Siliq (brodalumab), an IL-17 blocker, was also approved for the treatment of plaque psoriasis based, in part, on head-to-head trial data versus JNJ’s Stelera; (5) Roche’s Ocrevus (ocrelizumab), an anti-CD20 antibody, was approved for the treatment of relapsing and primary progressive multiple sclerosis, including less disability and relapsing, based on double-blinded clinical trials against a SoC agent, Merck KGaA’s Rebif (interferon, IL beta1-a); (6) In a test of proteasome inhibitors, Amgen’s Krypolis (carfilzomib) significantly improved overall survival versus JNJ’s Velcade (bortezomib) to treat relapsed or refractory multiple myeloma; (7) AstraZeneca’s Imfinzi (durvalumab), a PD-L1 antibody, is being tested with and without tremelumumab, a CTLA-4 antibody, versus SoC regimen of Lilly’s Erbitux (cetuximab), a EGFR inhibitor, plus platinum-based chemotherapy to determine progression-free and overall survival among patients with recurrent/metastatic head and neck cancer; (8) Amgen/UBC’s Evenity (romosozumab) , an anti-sclerostin antibody, reportedly reduces the risk of vertebral fracture, and women in Phase 3 are being randomized to either subcutaneous injections of Evenity for twelve months, followed by oral alendronate (generic Fosamax, a bisphosphonate ) weekly for twelve months, versus a control group of 24 months of alendronate treatment; (9) Boehringer Ingelheim’s Gilotrif (afatinib),a pan-HER inhibitor, was tested open-label versus Roche’s Tarceva (erlotinib), an EGFR blocker, as a second-line treatment among 795 patients with advanced squamous cell carcinoma of the lung, with overall-survival a secondary outcome to progression-free survival; and (10) in second-line treatment for advanced or metastatic renal cell carcinoma (N=803), Bristol-Myers’ Opdivo (nivolumab), a PD-1 antibody, demonstrated superior overall survival versus SoC Afinitor (everolimus), which inhibits mTOR kinase activity.
A different parameter associated with improved research design is the increasing choice of clinical endpoints rather than surrogate measures in Phase 3/4 trials. A surrogate endpoint is a biomarker (e.g., blood pressure [mmHg], glycemic control [A1c]) that substitutes for a clinical endpoint (e.g., death, myocardial infarction). Greater use of the preferred clinical endpoints, since 2010, is evident, notwithstanding that to gain FDA approval, in many disease indications, a pharmaceutical company has only to meet criteria related to surrogate measures.
Clinical outcomes in studies are being encouraged by regulators, payers, and as we argue here, by the companies themselves, as part of the armamentarium for sustained higher pricing. A few examples will make this point:
The approval of Repatha was based on the surrogate measure of LDL-c reduction. Among 901 patients with high LDL-c and a range of cardiovascular risk, Repatha 420-mg subcutaneous monthly, in addition to a potent statin for each patient, reduced mean LDL-C by 57 percent from baseline at week 52, as compared to placebo. The Phase 3 design is a real contrast to that of each statin drug. That is, Merck’s Mevacor (lovastatin), was approved by the U.S. in 1987, based on placebo-only trials. Thus, Repatha met a significantly higher hurdle for regulatory approval, given that the LDL-c efficacy occurred in the context of a prescribed statin such as Lipitor or Zocor, and either statin lowers LDL-c significantly more than Mevacor.
Both PCSK9 drugs and Mevacor were approved without clinical cardiac endpoint results available. Indeed, the initial statin clinical outcome trial was the Scandinavian Simvastatin Survival Study (4S), which was released in 1994. This was seven years after the first approval of a statin drug. In the absence of clinical data regarding mortality and non-fatal MI rates, the expensive PCSK9 drugs are subject to rigorous prior authorization criteria for reimbursement by payers (e,g., evidence of intolerance to maximal dosing of potent statin), and a co-pay of ~$250/month. The uptake of these drugs has been modest, to date, with year 2016 worldwide totals of $141 million for Repatha and $116 million for Praluent.
The first of the PCSK9 clinical cardiac event trial reported was Repatha versus placebo (with statin background therapy; N=27,564), published in the New England Journal of Medicine in 2017. The primary efficacy endpoint was the composite of cardiovascular death, myocardial infarction, stroke, hospitalization for unstable angina, or coronary revascularization. The relevant point here is that this trial was completed within two years of the date of FDA approval of the drug. This is markedly shorter than the seven-year hiatus noted between the initial statin clinical cardiac trial and the initial approval of a statin drug. With CNN reporting that over one in five Americans between the ages of 40 and 75 already take a statin to prevent an initial heart attack or stroke, the potential for a huge blockbuster drug remains in the balance for the PCSK9 biologic agents.
Return on Investment for Recent Overpriced Biotech Assets Requires Sustained High Pricing of Launched Drugs
Companies are being pressured regarding drug pricing from multiple U.S. shareholders, including private and governmental payers, physicians, and patients. Resistance to a return to value-based pricing based on cost-effectiveness analysis remains strong, as pharma and large biotech rely increasingly on the assets of smaller companies for drug development. This implies that less drug development is occurring in-house, and more often at nimble and creative research sites that are partnered with or purchased outright. The recent trend is for the larger companies to lock-on biotechs with either late-stage promising assets or with marketed products. This strategy evolved from large pharma having been burnt repeatedly by purchasing successful early-stage assets that eventually prove useless in Phase 3 studies. Accordingly, a restricted sample of smaller companies is being pursued by the larger biopharma companies.
In 2014, the median value of an acquisition was eight times revenue, as compared with 19 times revenue in 2015. By 2016, the ratio had ballooned to 39 times its revenue, as analyzed by the consulting firm Novasecta. At the same time, the billions spent in R&D by biopharmaceutical firms has resulted in lower ROI, from 10.1 percent in 2010 to 3.7 percent in 2016, according to a recent Deloitte study. This finding is not surprising, given the increasing valuations on developmental assets paid by the larger companies.
The blockbuster drug is the lifeline of the large biopharma industry. Companies need to boost pipelines in the hunt for an innovative drug that delivers sufficient revenue to cover the enduring failure inherent in R&D. The constant pressure is a reliable cause of over-payment for assets in development. This, in return, leaves companies fighting to preserve high prices for the next generation of innovative agents, given that ROI is inexorably more difficult to achieve.
Some examples will help make this point:
The aim is to provide a moderated Dickensian approach, neither “the best nor worst of times” for the U.S. prescription pharmaceutical industry and its $323 billion in revenue, after discounts and rebates. Instead, a balanced perspective is provided, including the less discussed, significant reforms which have improved the design of Phase 3/4 clinical trials, while intransigent high pricing of drugs at launch remains salient. Payers have seen much greater use of appropriate head-to-head trials, clinical outcomes, and now even attempts to shift from fee-for-service to a fee-for-performance paradigm, which is beyond the scope of this report. These marked changes have demonstrated greater value for drugs, far from the mundane regulatory package of placebo-only studies with surrogate measures as outcomes. To help explain the inexorable rise in pricing the focus here is on the deals that have become commonplace by large pharma to feed the endless pipeline needs. Large pharma has decreased internal resources for drug development and become more dependent on smaller companies to discover the next generation of innovative products. But as the supply of rare assets decrease, the deals tend to favor sellers, which ensures that buyers must fight for sustained high pricing of innovative launched drugs.
Larry Gorkin, Ph.D., is a clinical psychologist and grant writer at Brown University, Providence, who then joined Pfizer/Health Economics, at Manhattan world headquarters.
1Although U.S. sales of prescription drugs is estimated at approximately 40 percent of total global sales in 2017, this figure is higher for the specialty drugs that are the focus of any blockbuster analysis (e.g., presumably at 50 percent or higher, and consistent with the decision to double the $1 billion U.S. sales criterion).
2Among the top 20 companies in terms of global sales of oncology agents in 2014, Roche was ranked #1, with 32.7 percent of the total market