The FDA problems encountered with rivaroxaban are a significant, but not crushing, negative for the investment thesis of Johnson & Johnson (JNJ). I would still be buyer of the stock as JNJ is a very large company, not overly dependent on any one current or future product, that still has an excellent new drug product outlook even without rivaroxaban.
I have been on the fence in deciding whether to upgrade Bristol-Myers (BMY) to a buy. The issues now facing rivaroxaban are all significantly to the benefit of BMY’s apixaban. Pfizer (PFE) is co-marketing apixaban and this is clearly a positive for the company, but it is not sufficient to cause me to change my neutral position on the stock.
This note is being written on the day before the FDA Cardiovascular and Renal Advisory Committee will meet to consider whether to recommend Johnson & Johnson’s (JNJ) new anti-coagulant drug rivaroxaban for approval for the prevention of strokes caused by atrial fibrillation.
In the FDA’s briefing document to the committee, the reviewer recommended that rivaroxaban be given a Complete Response Letter, which would mean that it would not be approved at this time. It left open, the possibility of approving rivaroxaban as a second line agent to be used when warfarin fails, or to be used as a third line agent when the patient does not respond to warfarin or the recently approved new anti-coagulant drug dabigatran, would have limited commercial potential.
In reading the briefing letter and talking with physicians, I think that the primary issue for the FDA is that in the 15,000 patient ROCKET trial, which compared rivaroxaban and warfarin, warfarin was not dosed correctly. Warfarin is very difficult to dose and the FDA believes that, with skillful administration, the patient can be kept in the therapeutic range perhaps 60% to 65% of the time. In ROCKET, the patients were only in the therapeutic range about 55% of the time.
If the FDA were to only judge rivaroxaban on the basis of the results demonstrated in ROCKET, it would almost certainly be considered to be non-inferior to warfarin and possibly superior. However, the FDA believes that warfarin was under-dosed and is not willing to accept the results of the trial.
What went wrong? I don’t know. JNJ has suggested that the patients treated in ROCKET were a sicker population than those studied in other trials and this made it harder to dose warfarin. It is doubtful that the FDA will listen to this argument. Key opinion leaders have been concerned for some time that Bayer and JNJ made a mistake in giving rivaroxaban as a once a day dose instead of two. Simply put, physicians want to keep peak levels of the drug in the blood below a level at which there is excess bleeding and above a level at which the drug is not effective. This is much harder to achieve with a once a day dose than with twice a day. Many key opinion leaders feel that with a twice a day dose, rivaroxaban would be superior to warfarin and equivalent to apixaban. This is speculative, but I tend to agree.
Another major issue for rivaroxaban is that a competitive anti-coagulant drug dabigatran has been approved with a non-inferiority claim to warfarin and Bristol-Myers Squibb’s apixiban has been shown to be superior to warfarin in the recently reported ARISTOTLE trial. Incidentally, apixiban, with a longer half-life than rivaroxaban, was given as twice a day in this trial. There is one approved drug that is now available as an alternative to warfarin (dabigatran), and another (apixaban) that is likely to be approved in the next year and a half. If rivaroxaban were the only new anti-coagulant drug, there is a good chance that it would be approved even with the flaws in the trial.
Investment Consequences for Johnson & Johnson
There is a small chance that rivaroxaban will be approved as a second or third line therapy. However, it is much more likely to receive a Complete Response Letter. There is a possibility that JNJ and Bayer might identify a sub-group of patients in whom it can be clearly demonstrated that rivaroxaban is superior to warfarin, but this is a long shot. The most likely course of action is that another large (15,000 patient) trial will have to be conducted, which could take three years or more to complete. Designing this trial would have the uncertainty as to whether rivaroxaban should be compared to warfarin or dabigatran or apixaban.
As I mentioned above, the loss of rivaroxaban is a significant, but not crushing, negative for the investment thesis of JNJ. I would still be buyer of the stock as JNJ is a large company, not overly dependent on any one current and future product, that still has an excellent new product outlook even without rivaroxaban.
Whether JNJ will conduct another large trial for rivaroxaban is not clear given that the company is now likely to be three years behind apixiban in coming to the market. I suspect that it still might, as this is a huge market and probably needs more than just three effective drugs. In any event, the commercial promise has been damaged severely.
Investment Consequences for Bristol-MyersSquibb and Its Partner Pfizer
The ARISTOTLE trial topline data for BMY/PFE’s apixaban has been presented and, in the view of the academics who conducted the trial, it was superior to warfarin. The trial did not have a problem with warfarin dosing, so this is not likely to be a stumbling block. I think that there is a very high probability that apixiban will be approved, and a good chance that it could eventually become the standard of care in preventing strokes caused by atrial fibrillation. It will benefit greatly from not having to compete head to head with rivaroxaban, if that is the case, as now seems likely.
There is a good chance that the FDA will not grant the claim that apixaban is superior to warfarin. The FDA will do its own statistical analysis of the data and it may be the case that its statistical analysis will challenge the superiority claim. I don’t think that the FDA is inclined to give a superiority claim that would cause many patients to be switched off warfarin. From a commercial standpoint, I don’t think this matters.