IPAB hearing on the Nexavar compulsory license, part 1, R&D costs

On January 16, the India Intellectual Property Appeals Board (IPAB) began a hearing on the merits of Bayer’s appeal of India’s first compulsory license.

The outcome of this trial, which focuses on the cancer drug Nexavar, is a matter of first impression for the IPAB, and is expected to set precedents on a wide range of issues, including the permissible grounds for granting compulsory licenses, the relationship between the India patent law and the TRIPS Agreement, and the setting of terms and conditions for the compulsory license, including the royalty rates.

The two companies facing off in the IPAB hearing are a large German and a smaller Indian firm. Bayer AG has a market cap of US$117 billion on sales of US$ 47.2 billion, and 111,000 employees. Natco Pharma Ltd has a market cap of US$ 270 million on sales of US$ 99 million.

Bayer is appealing a March 12, 2012 ruling by the India Controller General Controller General of Patents, Designs & Trade Marks that granted NATCO a compulsory license to Bayer’s patents on Nexavar. A copy of the March 12, 2012 ruling is available here, as well as the February 13, 2012 KEI affidavit (/node/1359) and KEI’s earlier statement (/node/1384) on the granting of the compulsory licenses.

From January 16 to 18, the IPAB heard three days of arguments from lawyers for Bayer and NATCO. The hearing is expected to conclude on Wed the 23rd.

Bayer was represented by Sanjay Kumar, from the law firm Perfexio Legal, and Natco was represented by Rajeshwari Hariharan.

The first two days were largely taken up with Bayer’s presentation, which covered dozens of issues, including many procedural and technical issues that did not go the most important issues.

In this blog, I will focus on just one of the issues in the case – R&D costs

Bayer claimed it had spent huge sums on the development of Nexavar, and that the Controller did not provide an opportunity to Bayer to recoup those costs.

To make this argument, Bayer presented a January 9, 2013 affidavit from Harold Dinter which made the claim that from 1999 to 2005 Bayer had spent “2 billion euros (approximately US$ 2.5 billion) in the identification and development of anti-cancer molecules leading to the successful approval of Nexavar in 2005.” Dinter did not provide detailed support for the numbers, but said they were based upon Bayer’s general R&D outlays for anti-cancer drugs, including but not limited to Nexavar, and that the estimate was supported by a new December 2012 study by Jorge Mestre-Ferrandiz, Jon Sussex and Adrian Towse, published by the Office of Health Economics (OHE). Despite its name, the OHE is not part of the government, but rather a largely industry funded private consulting firm. The study itself was paid for by AztraZeneca. Dinter and Bayer’s lawyer also made extensive reference to the work of Joseph DiMasi, an academic who is also a drug company consultant.

Bayer had declined to present its actual outlays that were specific to Nexavar development, even though Onyx Pharmacetuicals, Bayer’s partner in the development of Nexavar, has made extensive disclosures of R&D costs to the US Securities and Exchange Commission (SEC) in its annual 10-K reports.

The $2.5 billion figure Bayer presented was more than a billion higher than the December 2012 Mestre-Ferrandiz, Jon Sussex and Adrian Towse estimate of the average cost of developing a new drug, and about $1.7 billion higher than DiMasi’s 2003 paper.

There are several reasons why the court is likely disregard Bayer’s extremely aggressive assertions of R&D costs, as did the patent controller. First, there are all sorts of inconsistencies between DiMasi estimates and the facts in the Nexavar case.

Nexavar is a Licensed-In drug.
On page 156 of DiMasi’s widely cited 2003 paper, he says “Licensed-in compounds were excluded because non-survey firms would have conducted portions of the R&D.” DiMasi further notes on page 179 that:

“while PhRMA has traditionally disaggregated its reported R&D expenditure data into expenditures on new drugs and expenditures on improvements to existing drugs, it has not gathered information on how expenditures on new drugs can be further decomposed into expenditures on self-originated and on licensed-in new drugs. Our R&D cost estimates are for self-originated drugs, and a substantial portion of the R&D expenditures on licensed-in drugs are likely missing from the PhRMA data.”

In their discussion of “Self-Originated versus Licensed-In” compounds, the OHE authors note that DiMasi’s subsequent research on licensed-in compounds suggests that risks are lower, and that DiMasi’s 2003 estimate of $802 million for self-originated compounds would fall to $640 million in 2003, based upon on the higher success rate alone, when the analysis was applied to licensed-in compounds.

“DiMasi et al (2010) . . . estimate a higher overall clinical approval success rate for licensed-in versus self-originated drugs: 27% versus 16%. However, the two groups of medicines had identical success rates for both Phase III (64%) and regulatory review (93%). Higher overall clinical approval success rate implies, other things constant, lower total R&D costs per approved drug; for instance, and using the results from DiMasi et al (2003), the R&D cost per approved drug with the higher 27% success rate would fall to US$640m (2003 prices), assuming other things stay constant, as compared with the original estimate of US$802m . . .” page 53, OHE paper

Note that the DiMasi and OHE methodologies are “fully loaded” as regards overheads, pre-clinical research, all risks of failures, and an 11 percent real cost of capital, and also that the $640 million number is a 2003 estimate expressed in 2003 dollars. Since this comes out to 25.6 percent of the Bayer number of $2.5 billion, one has to wonder why Bayer chose to put this into the record, but it is highly likely they don’t have anyone on the staff that actually thinks these things through. And, once you take the $640 million and adjust for some of the other issues discussed here, the numbers shrink even more.

Nexavar’s development time was quicker

In DiMasi’s 2003 paper, he estimates the average time between the beginning of clinical testing and FDA approval is 90.3 months. Nexavar began clinical testing in 2000 and was approved by the FDA December 2005. The difference in the period of time for testing, and the fact that most of the spending on Nexavar took place in 2004 and 2005, is important, because DiMasi and the OHE both add an 11 percent (or more) after inflation compound real cost of capital to adjust the out-of-pocket costs upwards.

Nexavar benefited from Orphan Drug Status
DiMasi avoided orphan products in his 2003 paper, and said on page 175, “For development as a whole, it is highly likely therefore that the share of R&D expenditures for which the orphan drug credit was applicable for traditional large multinational pharmaceutical firms is quite low.”

By avoiding orphan products, which have lower regulatory requirements, smaller trials and 50 percent tax credit available for qualifying clinical trials, DiMasi was able to provide higher estimates of drug development costs – which was one of the goals of the drug companies that sponsor much of DiMasi’s work.

But Nexavar is not only a licensed-in product from a small company, it has also received four orphan drug designations, including one for renal cell carcinoma.

When Nexavar receives an orphan drug designation, the drug developer can directly reduce US federal income tax liability by 50 percent of the costs of the qualifying clinical trials, for that indication, via a special and very generous tax credit, and still deduct the remaining 50 percent of costs from federal income taxes. The Orphan Drug Tax Credit is equivalent to receiving a grant from the US government, to cover half of the cost of the qualifying trials. This is particularly interesting in this case, because Bayer was receiving payments for 50 percent of its R&D costs on Nexavar from Onyx, and therefore, it is likely that Bayer incurred no cost of its own for the trials that qualified both for the 50 percent tax credit and the 50 percent funding from Onyx.

As an Orphan Drug, the size of the trials submitted to the FDA to establish the efficacy of the product were relatively small. For the 2005 approval, the two cite phase II trials involved 173 patients, and the Phase III trial cited involved between 769 and 900 patients, depending upon how it was reported. Even when adding Phase I trials, the numbers are far smaller than the 5303 average number of patients for the trials related to approval of a NME compound that were cited by DiMasi in 2003.

Using prevailing average costs for oncology trials at the time, Dr. Ruth Lopert has made an initial estimate that Bayer would not have spent more than $33 million for the trials that the FDA relied upon for its 2005 approval, before even taking into account the benefits of the 50 percent orphan drug tax credit.

Additional pre-approval Nexavar R&D projects

Bayer and Oynx also undertook research on liver cancer, metastatic melanoma, or advanced skin cancer, non-small cell lung cancer, breast, prostate, ovarian and other cancers.

Apparently after receiving its 2004 Orphan Drug designation, Bayer and Onyx made Nexavar available to 2000 patients in an expanded access program for kidney cancer, a factor that increased Bayer and Onyx’s R&D costs, but may also have qualified for the 50 percent Orphan Drug tax credit. Note that Bayer has consistently refused to provide details of the orphan drug tax credit subsidies, even though the tax credit has certainly reduced its net costs.

In looking at R&D costs, the expanded access trials present a dilemma. Yes, they involve research protocols, because the drug is not yet approved for sale, and they cost money. But not only do the trials have a much lower scientific value that the trials used to support a drug approval, but the expanded access program has no benefits to cancer patients outside of the USA, and one could easily argue they should be excluded when considering costs in an India proceeding.

Onyx’s 10-K reports of R&D outlays for Nexavar.

As noted elsewhere, Bayer’s partner in the development of Nexavar is Onyx Pharmaceuticals. Onyx published annual estimates of its R&D spending on Nexavar.

Beginning in 1994, Bayer entered into a drug development agreement with Onyx Pharmaceuticals. According to the Onyx 2002 10-K annual report filed with the US Securities and Exchange Commission:

“Effective February 1994, we established a research and development collaboration agreement with Bayer to discover, develop and market compounds that inhibit the function, or modulate the activity, of the Ras signaling pathway or that appropriately modulate the activity of this pathway to treat cancer and other diseases. Bayer and we concluded collaborative research under this agreement in 1999, and based on this research, a development candidate, BAY 43-9006, was identified.”

From 1994 to 1999, Onyx reports that Bayer provided Onyx with $26.1 million for the Onyx work on the Ras signaling pathway.

Beginning in 2000, Onyx and Bayer began another collaboration, aimed at the clinical testing and development of BAY 43-9006 and other products that had been identified in their joint research program. BAY 43-9006 was later given the generic name sorafenib, and the brand name Nexavar.

Onyx’s accounting for R&D outlays is quite inclusive, and the primary components beginning in 2000 were “clinical manufacturing costs, clinical trial expenses, consulting and other third-party costs, salaries and employee benefits, supplies and materials and allocations of various overhead and occupancy costs.”

After Nexavar was approved by the FDA in December 2005, Onyx provided an account of the cumulative R&D outlays on Nexavar from 2000 to the end of 2005.

The Onyx 10K reports include not only the trials used to support the 2005 approval of Nexavar for the treatment of renal cell carcinoma, but also the outlays related to approval of Nexavar for any and all other uses, and the 2000 patients in the expanded access program for kidney cancer.

According to the 2005 Onyx 10-K annual report: “Aggregate research and development costs-to-date through December 31, 2005 incurred by Onyx since fiscal year 2000 for the Nexavar project is $134.8 million.”

To summarize, note that Bayer paid for all research from 1994 to 1999 ($26.1 million), and this included research on several compounds in addition to the one now marketed as sorafenib/Nexavar. From 2000 onward, Bayer and Onyx split the R&D costs 50:50, and Onyx’s share of the R&D costs were $134.8 million. The outlays on the entire R&D program that lead to the 2005 approval of Nexavar for Kidney cancer were $26.1 + (134.8 x 2) = $295.7 million. Of the $295.7 million, only a fraction was spent on the development of Nexavar for kidney cancer, and some of that benefited from a 50 percent tax credit under the US Orphan Drug Act.

To the put the entire $295.7 million into perspective, ignoring the tax credits, that represents a little more than one quarter of the current global sales for sorafenib/Nexavar, a product that will maintain its monopoly in most markets through 2020.

$295.7 is also just 11.8 percent of the $2.5 billion estimate that Bayer wants the IPAB to accept as its R&D costs.

Finally, it is difficult to understand how Bayer expects data on drug development costs to fit into the case. If you believed the $2.5 billion number, should Bayer be rewarded for being a high cost supplier of R&D? And, will Bayer promise to start giving away Nexavar for free now that revenues have vastly exceeded even the most creative ways to blowing up the drug development cost figures?

Onyx Pharmaceuticals, Inc.
DECEMBER 31, 2005.

Page 35

Research and Development Expenses.
Research and development expenses were $63.1 million in 2005, a net increase of $27.3 million, or 76 percent, from 2004. In 2005, the increase in research and development expenses were primarily driven by a $28.7 million increase in Onyx’s share of codevelopment costs for the Nexavar program, principally for the clinical trial program which included the expanded access program in the Phase III kidney cancer trial initiated in the second quarter of 2005. In addition, 2005 Nexavar development costs reflect the ongoing pivotal Phase III kidney cancer trial, a Phase III trial in liver cancer initiated in the first quarter of 2005 and a Phase III trial in metastatic melanoma initiated in May 2005, as well as several Phase Ib and II clinical trials. This increase was partially offset by a decrease of $1.4 million from the therapeutic virus program, which was terminated in 2003.

Research and development expenses were $35.8 million in 2004, a net increase of $3.8 million, or 12 percent, from 2003. The increase in 2004 was primarily due to a $14.0 million increase in Onyx’s share of codevelopment costs for the Nexavar program, which expanded into the Phase III kidney cancer trial in the fourth quarter of 2003. This increase was partially offset by a decrease of $10.2 million of expenses from the therapeutic virus program. It is anticipated that research and development expenditures will continue at 2005 levels or increase as we continue with our clinical trials of Nexavar and as we add additional Phase III clinical trials of Nexavar, including a pivotal trial in lung cancer announced in the first quarter of 2006. However, the presentation of our Statement of Operations will change in future periods as we reflect Nexavar’s commercial status and present our share of the profits or losses from Nexavar. The major components of research and development costs include clinical manufacturing costs, clinical trial expenses, consulting and other third-party costs, salaries and employee benefits, supplies and materials and allocations of various overhead and occupancy costs.


Page 36

. . .

Aggregate research and development costs-to-date through December 31, 2005 incurred by Onyx since fiscal year 2000 for the Nexavar project is $134.8 million.