Scenario #1: The companion diagnostic as part of the clinical trial
In the simplest case, if the companion diagnostic is part of the clinical trial process and the drug cannot be approved unless the diagnostic is used to segment patient populations, the diagnostic should be launched. This is the case even if the diagnostic’s precision needs improvement—because without the diagnostic, the drug cannot be marketed—and there will be no sales. The Herceptin example clearly demonstrates that launching a drug without the ideal diagnostic is not an impediment to later diagnostic upgrades that improve predictive power.
Scenario #2: Drug category facing generic competition
When diagnostic testing is recommended for a marketed drug that faces generic competition, companion diagnostic testing leads to declines in branded drug sales while boosting sales of corresponding generics. This was clearly the case with carbamazepine, which is used to treat epilepsy, mania/bipolar disorder, and neuropathic pain.
The branded drug (Tegretol) was launched in the U.S. in 1996 for epilepsy and had significant generic presence by 2006. When the FDA approved language at the end of 2007 warning of adverse reactions in patients with a specific genetic variation, branded carbamezepine products (Tegretol, Carbatrol, and Equetrol) experienced sales declines while generic carbamazepine had revenue gains from approximately $11 million to nearly $20 million in 2008. Volume gains also nearly doubled with generic standard units growing from 225 million to nearly 400 million.
Scenario #3: Fixed-dose combination drugs in a competitive market
In the case of drugs that contain two or more active ingredients, imposing diagnostic testing has led to a decline in sales. For example, Abacavir, a synthetic nucleoside analogue indicated for the treatment of HIV-1 infection, was part of two fixed-dose combinations, Trizivir and Epzicom. When the label for all drugs containing Abacavir was changed in July 2008 to highlight adverse reactions in patients with a specific genetic variation, both of the marketed fixed-dose combinations experienced sales declines in the succeeding years.
Trizivir, which already had declining sales from its peak in 2003, saw the annual rate of decline increase after diagnostics testing was recommended from -10% to -16%. Epzicom, a more recent drug, also saw a 10% dip in U.S. sales from $332 million and as of 2010, drug sales had not recovered to pretesting levels. Ziagen, the single active ingredient version of Abacavir, and the oldest drug on the market, actually saw improved sales performance after the testing recommendation. Ziagen’s sales improved from an annual sales decline of nearly 15% from its peak in 2002 to 2007 to a more modest sales decline of 2% from 2008 to 2009.
Scenario #4: Dosing optimization diagnostics
When diagnostic testing is used to optimize the dosing of a drug, sales tend to improve. The commonly prescribed anticoagulant Warfarin, for example, has a narrow therapeutic index. Historically, physicians had been challenged in establishing individualized doses that would protect against formation of dangerous clots without inducing excessive bleeding.
In August 2007, the FDA required that the Warfarin label be changed to highlight appropriate doses based on patients’ individual differences in metabolizing the drug. The first genetic test kit for Warfarin was approved the following month. Overall prescriptions of Warfarin (including both branded and generics) subsequently increased by 8% in 2008. While cheaper generics continued to replace the more expensive brands, the branded products also gained after testing was introduced as shown by an improvement in sales compared to historical trends.
These scenarios, gleaned from examples of FDA-mandated label changes recommending or requiring diagnostics testing, provide a useful insight into the likely impact companion diagnostics will have in various situations. The scenarios also clearly demonstrate that companion diagnostics are neither a universal panacea nor an impending disaster for pharmaceutical sales. Rather they confirm that in deciding whether to launch a companion diagnostic, companies must take multiple factors into account.