J&J’s Invokana receives CV indication for reducing three-point MACE in US – October 31, 2018

In exciting news, J&J today announced that its SGLT-2 inhibitor Invokana (canagliflozin) has received an FDA-approved CV indication in the US for reducing the risk of MACE events (heart attack, stroke, and cardiovascular death) in patients with type 2 diabetes and established cardiovascular disease. The label will now include information from the CANVAS program reflecting Invokana’s effects on three-point MACE in patients with type 2 diabetes and CVD or high CV risk. While the CANVAS program included one-third of patients without established CVD (“primary prevention”), our understanding is that this indication will only apply to secondary prevention.

The update also applies to fixed-dose combinations Invokamet (canagliflozin/metformin) and Invokamet XR (canagliflozin/metformin extended-release). Invokana received a CV indication from EMA in September, following a positive CHMP opinion in August. Notably, today was the last day on which an FDA decision was expected to be released: The agency originally delayed its decision on J&J’s sNDA by three months in July 2018, requesting additional analyses from the company.

With this announcement, Invokana becomes the second SGLT-2 inhibitor with a CV indication in the US, and the first with one for all components of three-point MACE. Lilly/BI’s Jardiance (empagliflozin) received the first-ever CV indication for a diabetes drug in December 2016, based on the EMPA-REG OUTCOME CVOT, but that indication applies only to the reduction of risk for CV death, also in secondary prevention. Notably, that approval followed a 12-11 vote (in favor of the indication) at a June 2016 Advisory Committee – clearly, the issue of CV indications for diabetes drugs has become far less contentious, as no Ad Comm was held for Invokana.

Additionally, Novo Nordisk’s GLP-1 agonist Victoza carries a CV indication equivalent to Invokana’s, based on LEADER CVOT data.

Commercial Implications

• Commercially, we’ll be intrigued to see whether Invokana’s new CV indications in both the EU and US translate to a tailwind for the product’s steadily declining revenue. To be sure, Invokana is in dire need of further investment as well as an increase in sales, seeing as the franchise posted its seventh consecutive quarter of double-digit YOY decline in 3Q18, falling 28% YOY to $190 million, a dynamic widely attributed to concern over the amputation signal identified in CANVAS. The vast majority of this loss is due to falling US sales, where Invokana is the only SGLT-2 inhibitor with a regulatory warning for amputations; in Europe, the same warning applies to the entire SGLT-2 class. Notwithstanding, Invokana posted its first-ever YOY decline (11%) outside of the US in 3Q18, as well as its third-ever (and most sizable) sequential drop (13%) to $40 million. Altogether, from 2Q17, just before the presentation of CANVAS results at ADA, to 2Q18, Invokana’s share of the SGLT-2 inhibitor market has fallen from 34% to 21%. J&J itself has cited higher rebates to PBMs/payers, increasing patient discounts, and competitive pressures in explaining this share loss, and our sense is that the amputation signal from CANVAS has weakened J&J’s negotiating position, requiring the company to offer higher rebates and discounts and reducing Invokana’s profitability. Perhaps now, with a first-ever MACE indication among SGLT-2s, J&J’s negotiating power and Invokana’s profitability will see an uptick.

Class Implications

• We remain optimistic that expanded recognition of the CV benefits associated with SGLT-2 inhibitors can drive growth for the class. Across the board for SGLT-2 manufacturers, company management has cited slowed underlying class growth to ~5% annually in explaining the somewhat lackluster performance of the class in 2Q18. While year-over-year growth of 21% (in 2Q18) is still encouraging, the class rose only 3% sequentially from 1Q18 to 2Q18, when it fell 2% sequentially. That said, SGLT-2 inhibitors altogether still sold over $1 billion in the second quarter of the year, which is certainly impressive. Overall, it’s disappointing to see such a great class of molecules, offering solid glucose-lowering efficacy, weight loss, oral administration, and increasingly likely class-wide CV benefit suffer at the hands of a single safety signal (cost and genital infections are certainly also significant barriers). Without downplaying the importance of amputations as an outcome for both patients and providers, the overwhelming commentary we’ve heard from thought leaders is that amputation risk (i) is mostly isolated to patients who are already at very high risk; and (ii) can be managed with proper patient selection, education, and monitoring. In fact, we believe that the focus on amputations ultimately has put more focus on patient safety, definitely a patient positive. Further down the line, we hope that potential CV indications for AZ’s Farxiga (more on this one below) and Merck/Pfizer’s Steglatro can help solidify the CV and even renal benefits of this class and drive SGLT-2s into more patient hands. So far, Merck and Pfizer seem not to have invested meaningfully in Steglatro, as it generated only an estimated $16 million in revenue in 3Q18 – a 57% sequential drop in the product’s second full quarter on the market – and was not mentioned during prepared remarks or Q&A. Hopefully, Invokana’s CV indication will serve as inspiration to Merck’s management for greater investment in their SGLT-2 inhibitor specifically and diabetes broadly.

• Another critical boon for Invokana and the SGLT-2 class should come from J&J’s renal outcomes trial for Invokana, CREDENCE, and other renal outcomes trials for the class. In some of the most positive news we’ve heard in diabetes this year, CREDENCE was stopped early (July 2018) for meeting its primary endpoint ~ one year ahead of schedule, indicating very impressive efficacy on preventing the progression of CKD (primary endpoint components were end-stage renal disease, doubling of serum creatinine, and renal or CV death). While we still await the first topline data, we certainly expect another set of regulatory submissions for a renal indication from J&J – which would be an incredibly welcome addition to the treatment arsenal for diabetic kidney disease, given the depth of unmet need patients and providers face in treating CKD. Results from Dapa-CKD and EMPA-KIDNEY, both of which are enrolling people with and without diabetes, are poised to solidify this benefit as a class effect (expected completion in November 2020 and June 2022, respectively).

Comparison to Jardiance

• Invokana’s full three-point MACE indication raises questions about how the FDA is now approach CV indications, especially considering Jardiance’s lone indication for reducing risk of CV death among those with type 2 diabetes and existing CVD. There are some key differences between the EMPA-REG OUTCOME results for Jardiance and the CANVAS results for Invokana. While both found a nearly-equivalent 14% risk reduction on three-point MACE (95% CI: 0.74-0.99 for Jardiance, 0.75-0.97 for Invokana), only Jardiance gave a significant improvement on the critical secondary endpoint of CV death and all-cause mortality. In EMPA-REG OUTCOME, Jardiance gave a 38% risk reduction on CV death (HR=0.62, 95% CI: 0.49-0.77) and a 32% reduction on all-cause mortality (HR=0.68, 95% CI: 0.57-0.82). Our understanding is that Jardiance received its narrower indication from FDA because CV death drove the overall reduction in three-point MACE, and the EMA label reflects a similar benefit. However, given that no individual component of three-point MACE met significance in CANVAS, we can’t help but wonder whether Jardiance would have received a full three-point MACE indication if the sNDA were submitted today. This raises interesting questions about how the agencies will approach the specifics of CV indications moving forward. Have EMA and/or FDA become more open to granting full MACE indications, like the one granted to GLP-1 agonist Victoza, since the Jardiance submission was considered? From this result – and also considering that the FDA elected not to have an advisory committee for the first time in the history of CV s indications for diabetes drugs – it seems this may be the case.

  • Some have suggested the evidence of Invokana’s CV benefit isn’t as robust as the evidence supporting Jardiance’s benefit. In a provocative presentation at CMHC West, Dr. Sanjay Kaul re-analyzed CANVAS data using Bayesian statistics and argued that there is insufficient evidence of Invokana’s CV benefits, going so far as to say he “wouldn’t be surprised” if FDA denied the indication. Moreover, as we understand it, FDA uses a different and often more rigorous statistical methodology: At the Ad Comm for Jardiance’s CV indication (June 2016), FDA presented slightly different findings for EMPA-REG OUTCOME than those reported by Lilly/BI. On the other hand, our overwhelming sense from diabetes thought leaders is that cardioprotection is an SGLT-2 inhibitor class effect – of course, the two opinions aren’t mutually exclusive: A weakness in the evidence certainly doesn’t mean the benefit isn’t there. In our observation, the fields of both endocrinology and cardiology continue to see CANVAS as a very positive trial on cardio- and renal-protection, and we (and now the FDA) vehemently agree.

  • Given that AZ’s DECLARE CVOT for dapagliflozin was neutral on the three-point MACE endpoint, we wonder how FDA’s decision on Invokana might inform a potential CV indication application from AZ for Farxiga. Of note, DECLARE enrolled a very sizable (59%) primary prevention population. Certainly, it seems likely that Farxiga could receive an indication for the first of its co-primary endpoints, the composite of CV death and hospitalization for heart failure, for which Farxiga was found to be superior vs. placebo. However, we wonder if Farxiga could also receive an indication for another component of the three-point MACE if it was shown to be superior vs. placebo. Moreover, might the requirement of a p-value <0.025 for superiority due to the co-primary endpoints in DECLARE at all affect FDA’s decision?

Comparison of Relevant Results from CANVAS (for Invokana) vs. EMPA-REG OUTCOME (for Jardiance)

Close Concerns’ Questions

Q: Why was Invokana granted an indication for all three components of MACE by the FDA, whereas Jardiance was only granted one for CV death? Would FDA consider revising Jardiance’s indication? We find this especially pertinent given that none of Invokana’s MACE components achieved superiority, and results on three-point MACE are extremely comparable between the two CVOTs.

Q: How has FDA’s approach to CV indications changed since Jardiance and Victoza were reviewed?

Q: How will the CV indication impact Invokana’s sales specifically and the SGLT-2 class’ sales broadly? How will it affect Jardiance sales specifically?

Q: How might FDA’s decision on Invokana inform a potential CV indication application from AZ for Farxiga based on results from DECLARE? Could Farxiga receive an indication for a superior component of three-point MACE (results have not been released, of course) despite overall neutrality on the composite? Might the DECLARE-specific requirement for a p-value <0.025 for superiority, due to the co-primary endpoints, effect the FDA’s decision?

-- by Peter Rentzepis, Ann Carracher, and Kelly Close