Raising the Dose and Raising the Cost: The Case of Pembrolizumab in Lung Cancer

After adjusting for inflation, the price per life-year gained of cancer drugs rose from $54 100 to $207 000 over the 20-year time period between 1995 and 2013 (1). Higher launch prices are one way companies can generate more revenue from sales. Another is raising prices after launch. Gleevec launched in the United States at $4546 per month (in 2014 dollars), and it now costs $12 278. Prices typically rise when a cancer drug has gained a new indication or compendia listing, and even without these events prices rose around 5% per year; direct competitors entering the market cause prices to decline by only around 2% (2). An investigation by the Financial Times reported that prices for drugs such as Sprycel, Zytiga, and Revlimid have risen in price 60% to 100% over the past five years and that there are no substantial rebates being paid for them (3).

Manufacturers also have a subtler mechanism by which they can increase the revenue of their product: entering the market at one dose, establishing a per-milligram price, and then increasing the dose. In this issue of the Journal, Goldstein et al. (4) focus on this strategy, noting that without any clinical or biologic rationale, Merck, in adopting a flat dose of 200 mg, effectively increased the dose of pembrolizumab in a manner that in firstline non–small cell lung cancer (NSCLC) greatly increases the revenue to the company when compared with a scenario where pembrolizumab is used at a lower and presumably equally effective dose. Specifically, although the drug has been extensively studied and was initially registered at a dose of 2 mg/kg every three weeks (which implies an average dose of around 150 mg for a 75 kg patient), the company pursued studies in firstline NSCLC at a flat 200 mg dose every three weeks. In support of this, well before the US Food and Drug Administration (FDA) approved this 200 mg flat dose, Merck removed their 50 mg vial from sale in the United States (but not Europe) and only made the drug available in 100 mg vials. From an economic perspective, the impact is obvious. A milligram of pembrolizumab has an average sales price of around $47, which means for a 75 kg patient, the company earns an additional $2350 per dose just through this dosing change.

Although this math is straightforward, Goldstein et al.’s analysis does far more. By mixing epidemiologic data about the prevalence of patients with metastatic lung cancer who would be eligible for pembrolizumab, estimated duration of treatment data from clinical trials, real world data on patient weights who have lung cancer, and some economic and survival curve modeling, Goldstein et al. identify an additional $825 million in costs to the US health care system in a single year, spent on additional pembrolizumab that provides no additional clinical value to the patient. The article is a tutorial on how such an estimate can be generated plausibly, and the authors deserve credit for careful illustration of their methods and a readable set of sensitivity analyses where they varied their assumptions even more broadly than, in our view, was necessary to make their conclusion convincing.

If there is a problem with Goldstein et al.’s study, it is that the estimate they produced most likely understates the economic consequences for taxpayers, insurers, employers, and patients from Merck’s apparently clinically unjustifiable dose increase. It fails to note that, if anything, Merck’s dose increase creates an added incentive to use their product.

For instance, Goldstein et al. logically focus on pembrolizumab’s firstline indication in lung cancer as the FDA has approved the fixed 200 mg dosing. But as Goldstein et al. note, Merck had functionally achieved receiving revenue for 200 mg of pembrolizumab per dose for its earlier indications of melanoma and second-line NSCLC. They did so by removing the drug’s 50 mg vial from the US market and replacing it with a single dose of the 100 mg vial (the company still sells the 50 mg vial in Europe). Our group estimated that this shift from a 50 mg to 100 mg vial would add $1.2 billion to Merck’s revenues from 2016 to 2020 (5).

The increased dosing may also influence provider prescribing behavior. Medicare reimburses providers at 106% (or 104.3% under the sequester) of the average sales price of infused therapies; commercial insurers also mark up reimbursement in percentage terms when reimbursing physicians and hospitals for the cost of drugs like pembrolizumab. The percent-based profit margin has repeatedly been shown to incentivize physicians to use higher-cost drugs (6). With the shift from weight-based dosing to 200 mg flat dosing and the removal of the smaller vial size, the profit margin for providers is proportionately increased. For Medicare patients, this means an extra $100 per dose to the treating physician. This added profit may incentivize more doctors to use the drug, further magnifying the excess cost to the system.

In our view, Goldstein et al. may have underestimated the total cost to society from two perspectives. One is that the price of pembrolizumab will continue to rise if past patterns are perpetuated. Most medical benefit drugs see price increases of around 2% per year, and, consistent with this trend, the average sales price (ASP) of one milligram of pembrolizumab has risen from $45.70 to $47.30 over the past two years (ie, an annual inflation rate of 1.7%). Even such small price increases have large implications when spending on drugs is in the billions of dollars. As an example, Goldstein et al. used the 2016 fourth quarter ASP for pembrolizumab of $46.50 for their estimate; however, using the 2017 first quarter ASP would raise the estimate from $825 million in excess spending to $839 million. The second way in which Goldstein et al. may have underestimated the total cost to society is by their Medicare markup. We would modify the Medicare markup of 4.3% above the ASP across all modeled administrations of the treatment. Our group recently estimated that the average markup across providers and payers for patients with cancer led to a blended markup of approximately 37% (7). This markup figure increases Goldstein et al.’s estimated excess cost number from $825 million to $1.084 billion.

Goldstein et al. are to be commended for examining a nearly invisible change to pembrolizumab’s dosing schedule that yields large returns to the manufacturer at the expense of taxpayers, insurers, and patients. The implied safety and/or convenience issues of flat dosing appear to us to be a specious justification, given the decades of weight-based or body size–based dosing of oncology drugs. It is difficult to know if the FDA would invoke a dose change without the manufacturer’s approval as Goldstein et al. propose. Regardless, the packaging of pembrolizumab in 100 mg vials makes reducing doses a logistically difficult challenge.

Funding

Dr. Bach reports personal fees from the following: Association of Community Cancer Centers, America's Health Insurance Plans, AIM Specialty Health, American College of Chest Physicians, American Society of Clinical Oncology, Barclays, Defined Health, Express Scripts, Genentech, Goldman Sachs, McKinsey and Company, MPM Capital, National Comprehensive Cancer Network , Biotechnology Industry Organization, The American Journal of Managed Care, The Boston Consulting Group, Foundation Medicine, Anthem Inc., Novartis, Excellus Health Plan. Dr. Bach has received grants from the National Institutes of Health Core Grant P30 CA 008748, the Kaiser Foundation Health Plan, and from the Laura and John Arnold Foundation. Dr. Saltz reports research funding from Taiho Pharmaceuticals.

Note

The funders had no role in the writing of this editorial or the decision to submit it for publication. All funding is unrelated to the submitted work.

© The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.