Has the cost of the drug substance ever represented a make or break factor preventing the development and successful launch of a new pharmaceutical? Or the continuing sales of products already on the market? If not, will this change?

At first appearance this question can seem rhetorical and futile, and easily dismissible based on the modest cost of goods sold, profit margins and returns on capital invested reported by most pharmaceutical companies. Contrary to most other industrial sectors, topics like cost optimization and operational excellence are relatively new issues for the pharmaceutical industry.

However, a closer analysis indicates that the cost of the drug substance can be critical. This is illustrated by the problems encountered a few years ago by Roche with Fuzeon (enfuvirtide), a synthetic peptide containing thirty-six amino acid residues developed by Trimeris for the treatment of AIDS. This product had daily drug dosages of 180 mg (2 x 90 mg)—far more compared to most other peptide drug substances—creating major constraints on cost-of-goods sold in general and more particularly on the impact of the active ingredient in the overall cost structure. These constraints were exacerbated by the sales ceiling set on feasible prices when compared to alternative treatments and the high profile of AIDS as a disease amongst politicians and the general public. According to industry observers, Fuzeon yearly treatment costs of $25,000—at the time of the product launch this was a sky-high figure—were largely driven by the drug substance production costs. Besides the furor created amongst AIDS sufferers, these high prices are viewed as one of the main causes behind the limited success met by Fuzeon, which at the time was touted as a potential blockbuster.

The shortages of some mature generic drug products increasingly reported all around the world also provide a vivid demonstration of how drug substance costs have an impact. In countries where prices are constrained by controls imposed by the governmental authorities or by intense generic competition, the net back price achieved by pharmaceutical companies with some mature established drugs such as injectable ampicillin sometimes barely covers production costs, if not even those of the drug substance. Such situations hardly provide an incentive for producers to continue to market and supply these except if as loss leaders serving for pushing other more profitable items.

Similarly, multiple examples indicate that cost-of-goods considerations have contributed to the decision to discontinue the development of molecules providing limited therapeutic benefits compared to products already on the market and hence likely to see their price severely constrained, jeopardizing their ability to reach acceptable financial returns. Examples include iodinated non-ionic X-ray contrast media or anti-cholesterolemia products characterized by a crowded market where therapeutic needs are satisfactorily met by existing products.

The cost of the drug substance therefore matters. While predicting the future is always a risky exercise, looking at the developments impacting the pharmaceutical industry the most likely scenario is that drug substance cost related factors will represent an increasingly important consideration for pharmaceutical companies both for new and established pharmaceuticals. Such a prediction is based on the following:

Pricing pressures
The pricing pressures facing the industry as budget constraints become increasingly acute are forcing innovators to price new products, even if providing advantages compared to drugs already on the market, at a discount. This can be illustrated by Novo-Nordisk electing to offer Rybelsus, the oral form of semaglutide, at a lower price than the injectable form of Ozempic, even if it is more convenient for patients and involves 40 times higher dosages. Teva is following a comparable pricing strategy with Austedo, the deuterated analogue of terbenazine, marketed at a fraction of the price of the latter despite having a longer half-life allowing for more convenient drug administration regimens while being more expensive to produce. Clearly such approaches to pricing are aimed at having these products swiftly included in formularies for reimbursement. Obviously to achieve this requires that cost-of-goods sold in general and more particularly drug substance related costs do not represent a bottleneck in the process.

CDMO Consolidation
Consolidation of the CDMO supply structure is another major factor, particularly for the drug substance sector with the emergence of larger vendors like Evonik, Lonza, Siegfried and Thermo Fisher. Through their size, these CDMOs can be expected to be better able to exert pricing power on their customers compared to their small to medium-sized counterparts, the traditional backbone of the pharmaceutical fine chemicals industry. Parallel to this, over the last few years a growing number of suppliers have been acquired by private equity investors. This change of ownership suggests that these players will be more aggressive in their approaches to pricing in order to achieve the top and bottom-line growth required for meeting the financial returns expected by their more demanding new owners.

Given all these factors, properly assessing the drug substance cost is becoming imperative for pharmaceutical companies, regardless if they are innovators or generic marketers.

The impact of the drug substance cost on the end product price remains a highly controversial topic with estimates ranging widely. As an example, Solara, an Indian API producer, indicates on page 25 of its 2018/19 annual report,¹ a total drug substance value covering both the merchant market and captive production is valued in excess of $160 billion. Such a figure corresponds to approximatively 15% of total pharmaceutical sales at the retail level as reported by leading industry statistics and well above the 10% rule of thumb generally applied in the industry. Some observers even suggest that the impact of drug substance costs is less than 7% of pharmaceutical sales.

While it is difficult to comment on the accuracy of these various estimates as underlying assumptions are rarely provided, these divergences and discrepancies between the figures reported is hardly surprising. It can be explained by the multitude of potential sources of variability existing. Among various other factors, these include the level at which prices/sales are defined. By the manufacturer or pharmacy/wholesaler? What is the product status—branded innovative Rx or private label OTC/generic? What are the daily dosages? Is it a hospital/specialist or general practitioner product? First in class or me-too? What is the country of origin? These are just some of the questions and factors why drug prices vary widely across the world. Another critical consideration is the methodology applied for calculating the drug substance cost—namely does it assume market prices, internal transfer prices or activity-based costing?

Based on the analysis of dosage form sales, associated drug substance volumes, reported merchant drug substance market prices and assuming a distribution mark up of 30%, the average impact of the drug substance cost for the global pharmaceutical industry expressed as percent of ex-manufacturer sales can be estimated at around 7% to 8%. As illustrated in Figure 1 – this impact varies depending on the segment of the pharmaceutical industry and the region considered. It is comprised between a low of 4% in North America and 20 % outside the Triad—Japan, North America and Western Europe—reflecting differences in drug prices and sales mix.

Over the years the impact of the drug substance cost expressed as percent of pharmaceutical sales appears to have decreased. This trend can be explained by the introduction of high price products like anti-cancer and orphan drugs, the growing share of the U.S. market where pharmaceutical prices are typically higher and the price erosion noted for several drug substances prompted by “Chindia”-based suppliers.

At least from a dimensional perspective the estimated 7% to 8% average impact for the drug substance cost expressed as percent of ex-manufacturer prices is broadly congruent with a “back-of-the-envelope” bottom up calculation of the combined turnover of known industry players and estimated share of drug substance outsourced.

However, such a macro figure is of little practical significance as it represents an average varying between 5% for branded Rx products and more than 20% for generic drugs sold in developing and emerging countries—variations among single products are even more dramatic. Its relevance is therefore limited to high level strategic considerations, such as relating to the size of market potentially available to CDMOs. 

A more relevant and actionable figure for both pharmaceutical companies and their fine chemical suppliers is represented by the impact of the drug substance cost expressed as a percent of ex-manufacturer prices for single products. For generic marketers this percent is a critical indicator of achievable margins while for innovators engaged in the development of NCE it provides invaluable hints on the drug substance cost that cannot be exceeded to ensure the viability of the product. For suppliers like CDMOs the impact of the drug substance cost is one of the parameters used for calibrating their price offers when responding to requests for new product proposals or, in the case of established mature products, for assessing the room to maneuver and exert pricing power on customers.

As illustrated in Figure 2, the impact of the drug substance cost varies widely. In the sample analyzed it is comprised between less than one percent for brexanolone, a product recently launched for the treatment of severe post-partum depression; formoterol, a long acting beta agonist used as bronchodilator; and ursodesoxycholic acid, a product serving in the treatment of gallstones involving high dosages, and more than 20% drug substance cost, which is comparable to ibuprofen private label OTC formulations.


This variability in terms of impact of the drug substance cost is observed for both new and established products. It reflects several factors including the products therapeutic value and degree of novelty, competition, daily dosages, length of the treatment, molecular structure and synthesis complexity. By and large, the impact of the drug substance cost tends to be higher for pharmaceuticals involving substantial daily dosages and addressing large patient cohorts.

However, numerous exceptions do exist. A notable one is represented by mesalazine, also referred to as 5-amino-salicylic acid. It is a mature product introduced about thirty years ago and continues to be widely used as a first line treatment of IBS (inflammatory bowel syndrome) types of conditions and involves dosages in excess of 1 gram per day with often chronic intake where the impact of the drug substance is around 3% of ex-manufacturer sales. At the other extreme drugs addressing orphan types of diseases, or used for short periods of time and involve low dosages such as cytotoxics serving in cancer treatment or inhalable steroids for asthma, tend to be characterized by a below average impact of drug substance costs.

For products already on the market or having reached the generic status, multiple suppliers typically exist. Therefore, finding the price of the drug substance in the merchant market is rarely a problem greatly simplifying estimating the financial viability of generic products—merchant prices serve as proxy of cost. Unfortunately, the situation is totally different for products still in clinical Phase I or IIa development phase. For these no such external benchmark and proxy exist. An additional complication is the limited attention most often devoted before Phase IIb to the synthesis process for accessing the drug substance. In these early development stages, the drug substance is almost invariably synthetized applying the route originally designed during the discovery phase when time-to-market and speed outweigh considerations such as yields, environmental load or scalability. This creates an additional challenge for estimating production cost.

To overcome this problem an option for the pharmaceutical developer is to contact various CDMOs and them to provide an “open book” price offer for the drug substance envisaged where costs, margins and underlying assumptions are appropriately detailed. A more hands-on approach for the developer is to apply a spreadsheet model for calculating by himself variable, cash and full production costs.

Within this frame, variable production costs—raw materials, utilities and waste disposal—are calculated. This is assuming they are starting from intermediates available on the market, and taking into account the number and type of reaction steps, solvent requirements and yields being estimated based on those reported in the literature. Fixed costs including direct labor and plant overheads are estimated multiplying the estimated utilization time for each main piece of equipment required in the synthesis by its associated hourly costs – expressed in USD/gallon or m³ reactor capacity. For having the fully loaded production economics and a proxy of expected market price, non-cash costs including depreciation and a mark-up reflecting the required return on capital, must be added on top of cash production costs.

By using hourly costs that reflect investment outlays per unit of reactor capacity and unit labor costs in various parts of the world, adjusted to reflect productivity, it is possible to model the cost that the drug substance would have depending on production location. Obviously, the robustness of these estimates must be critically evaluated by applying a sensitivity analysis.

The applications of this type of analysis are multiple. It provides a semi-quantitative indication of the products financial viability, sets tangible cost reduction targets for process optimization and provides an objective base for assessing the price competitiveness of offers received from various vendors.

The suggested approach has been applied on elafibranor (Genset) and lanifibranor (Inventiva), two PPAR (Peroxisome Proliferation Activated Receptor) agonists in respectively Phase III and Phase IIB clinical development in the NASH (Non-Alcoholic Steato Hepatitis) indication. Based on the synthesis routes described in the literature, assuming market prices for starting materials, a ten-ton output scale and operations based in China, variable and fully loaded production costs can be estimated for elafibranor at respectively 210 and 450 USD/kg; the corresponding figures for lanifibrator standing at 1,100 and 1,500 USD/kg. The estimated 1,000-plus USD/kg differential between the cost of these two drug substances needs to be analyzed in light of their pharmacological profile and observed clinical properties.

While both elafibranor and lanifibranor are PPAR agonists and share structural similarities, their activity profile differs. Elafibranor is active against alpha and delta receptors involved in the control of lipid metabolism. In addition to alpha and delta agonist properties, lanifibrator has been demonstrated to have an activity also against gamma receptors, increasing sensitivity to insulin while reducing the production of collagen and fibronectin, two proteins associated with fibrosis and ballooning in late NASH stages. Another difference between the two drug substances is their expected daily dosages, about 80 to 120 mg/day/patient for elafibranor; the corresponding figure for lanifibranor being 1,000 mg/day/patient. Assuming chronic intake, these daily dosages correspond to a yearly intake of about 36 and 360 g/year/patient for elafibranor and lanifibranor respectively. Taking a cost of 450 USD/kg for efibranor and 1,500 USD/kg for lanifibranor, the impact the drug substance cost is in the order of 16 and 540 USD/patient/year respectively.

While at first, based on these figures lanifibranor appears to be very costly, these estimated drug substance costs must be compared to the net back revenues that the developer is realistically expected to generate by selling the product. In the specific NASH indication, yearly drug related outlays of 15,000 USD/patient/year are indicated by various industry analysts as a reference. Assuming there is a distribution mark up of 30%, this would translate in net back revenues for the pharmaceutical company in the order of 10,000 USD/patient/year. Drug substance cost for elafibranor and lanifibranor expressed as percent of sales are 0.16% and 5.4% respectively. Based on these estimates the drug substance cost should not represent a stumbling block for either product, particularly if clinical trials can demonstrate that the pan PPAR agonist properties of lanifibranor are also active against gamma receptors effectively. This translates into therapeutic benefits and a superior profile justifying a price premium compared to elafibranor (see Figure 3).


Obviously, the conclusions on the financial viability of these products would be very different should the expected drug price be substantially lower. This would result in a higher impact of the drug substance cost expressed as percent of ex-manufacturer prices, making financial returns more elusive. In such a scenario lanifibranor would be particularly handicapped due to the double whammy impact of its high daily dosages and substantial drug substance costs, making it a priority for the developer to set aggressive targets for significantly reducing these. An alternative course obviously would be to try to position the product at a premium price should clinical trial results allow for the support of superior efficacy claims.

Summing up, even if pharmaceutical prices most often reflect value in use or those of competing products rather than production costs—the cost of the drug substance expressed as percent of ex-manufacturer prices can be critical both for generic and innovative products. No product development decision should be taken without duly taking into consideration this parameter. This implies that adequate attention is devoted early in the product development cycle to the expected drug substance cost and stresses the need for startups and emerging pharmaceutical companies to gather adequate insights on their product process economics, possibly a challenge for this set of players typically applying a virtual supply chain model for accessing their drug substance requirements. 

References

1. https://webowizard.com/solara/wp-content/uploads/2019/10/Annual_Report_2018-19.pdf

---

Dr. Michele Jermini is the Managing Director of Alfa Chemicals (Suisse) SA; mjermini@alfachemicalsint.com
Dr. Enrico Polastro is a Vice-President of Arthur D.Little; polastro.enrico@adlittle.com