AHA (American Heart Association) 2018

November 10-12, 2018; Chicago, IL; Day #2-3 Highlights – Draft

Executive Highlights

  • Following a data-rich Day #1 at AHA 2018, our last two days in Chicago did not disappoint. During a Sunday morning session, Dr. Darren McGuire presented a CARMELINA post hoc demonstrating that Tradjenta’s reassuring HR on hospitalization for heart failure (HR=0.90, 95% CI: 0.74-1.08) was maintained across nearly all age, geographic, renal, and baseline HF and insulin use subgroups. The only two crossing unity were those on insulin at baseline (HR=1.00, 95% CI: 0.61-1.24) and the European cohort (HR=1.13, 95% CI: 0.85-1.51).

  • Late-breaking results from the mechanistic EMPA-HEART (n=97) study gave some of the best in-human data on cardioprotective mechanisms for SGLT-2s we’ve seen to date: Compared to placebo,  Jardiance significantly reduced left ventricular mass (adjusted mean difference -3.35 g/m2, 95% CI: -5.9, -0.81, p=0.01) in people with type 2 and well-treated CAD over six months. Dr. Subodh Verma explained that empagliflozin may promote reverse modeling of the heart, contributing to CV benefit; other mechanisms are also likely at play. Later in the day, Drs. Verma and John McMurray looked to the potential of SGLT-2s for treatment, not just prevention, of heart failure.

  • Computational modeling based on EMPA-REG OUTCOME data in patients with heart failure at baseline showed that, in this patient group, it’s cost effect at an ICER of ~$22,000 per QALY for commercial insurers to pay for Jardiance. While this model makes a number of assumptions, it was encouraging to see that cost effectiveness was maintained through a variety of adjustments. Jardiance gave ~1.2 additional years of life vs. placebo in this population.

  • While there was truly a robust focus on diabetes and CV outcomes at AHA, Dr. Mikhail Kosiborod lamented that SGLT-2s and GLP-1s remain devastatingly underutilized in clinical practice. In a recent analysis, only 5% of patients who met the inclusion criteria for EMPA-REG OUTCOME were on an SGLT-2 inhibitor and only 6% of those who met LEADER inclusion criteria were on a GLP-1, in the real world. Cardiologists, he says, have a critical role to play in pushing these numbers higher. We also note that Dr. Kosiborod was asked to comment on the recently-populated notion that SGLT-2s are little more than diuretics, explaining that there seem to be other mechanisms at play.

  • See below for more, including a poster linking glycemic variability, measured via CGM, to risk of recurrent major adverse cardiovascular and cerebrovascular events in post-ACS patients, both with and without diabetes. Also below, you’ll find coverage of a product theater for Amarin’s Vascepa and renal outcomes from the CAMELLIA CVOT for Arena/Eisai’s obesity therapy Belviq.

AHA 2018 just wrapped up Chicago, and this second installment of highlights includes our most exciting moments from the last two days of the meeting, plus coverage of eight companies in the exhibit hall. There was a notable focus on SGLT-2s, heart failure, and type 2 diabetes over these three days in the Windy City.

In case you missed it, check out our Day #1 Highlights, including readouts on DECLARE for AZ’s Farxiga (plus commentary from Drs. Butler + Verma) and REDUCE-IT for Amarin’s Vascepa.

Table of Contents 

Top Eight Highlights

1. CARMELINA’s Reassuring Trend on HHF with DPP-4 Tradjenta Holds Across Nearly All Subgroups, Save European Participants and Those on Insulin; First DPP-4 with Favorable Results

To a jam-packed room, UT Southwestern’s Dr. Darren McGuire expanded on heart failure results from CARMELINA, the CVOT for Lilly/BI’s DPP-4 Tradjenta (linagliptin), reinforcing the study’s reassuring result on hospitalization for heart failure (HHF) across subgroups. As a reminder, the HHF results in the full cohort trended solidly in favor of Tradjenta (HR=0.90, 95% CI: 0.74-1.08, p=0.2635 for superiority, 209 vs. 226 events), as presented at EASD 2018. The positive HHF trend held for all sensitivity analyses, HF-related outcomes (below), and almost all subgroups; subgroups included those based on age (above/below 65), region (North America, Latin America, Europe, Asia), insulin use (yes/ no), history of heart failure (yes/no), eGFR (above or below 60), UACR (less than 30, 30-300, >300). The only groups in which the HR hit 1.00 or above were (i) participants on insulin (HR=1.00, 95% CI: 0.61-1.24) and (ii) the European cohort (HR=1.13, 95% CI: 0.85-1.51); interaction p-values were significant for insulin vs. non-insulin (p=0.04) and geographic (p=0.04) subgroups, though interpretation of this results should take into account the testing of 33 pre-specified subgroups for interaction without correction for multiplicity. The insulin interaction in particular calls to mind a talk we heard at ESC 2018, in which Dr. Deborah Cosmi argued for further investigation of an observed link between insulin use and risk of heart failure in patients with type 2 diabetes, based on pooled and observational data – of course, a HR or 1.00 is still completely neutral. This being said, given the complex history of DPP-4 inhibitors and heart failure, these results certainly bode well for Tradjenta and the DPP-4 class. As a reminder, initial concerns about DPP-4s and HHF risk were sparked by the SAVOR-TIMI trial for AZ’s Onglyza (saxagliptin), which found a small but concerning signal for increased risk on this endpoint with saxagliptin (HR=1.27, 95% CI: 1.07-1.51, p=0.007). However, in the group that exhibited excess heart failure risk in SAVOR, there was no increased risk of all-cause mortality, occurrence of the primary endpoint (MACE), or occurrence of the secondary endpoint (MACE plus hospitalization for heart failure, unstable angina, or coronary revascularization).

  • Dr. McGuire highlighted that linagliptin is now the 2nd DPP-4 with no signal for increased HHF risk, along with sitagliptin, with each of the 2 analyzed by pre-specified exploratory analyses. As stated above, SAVOR-TIMI found a concerning increased risk for HHF with Onglyza, while EXAMINE for Takeda’s Nesina (alogliptin) trended strongly in favor of placebo (HR=1.19, 95% CI: 0.89-1.59). TECOS for Merck’s Januvia (sitagliptin) was resoundingly neutral (HR=1.00, 95% CI: 0.83-1.20). When asked about this within-class heterogeneity on HHF, Dr. McGuire posited that saxagliptin’s relatively short half-life may have contributed to its concerning signal, based on emerging evidence that DPP4i exposure induces increases circulating DPP4. Indeed, saxagliptin’s half-life is 2.5 hours, compared to 12.4 for sitagliptin, ~12 hours for linagliptin, and 21.4 hours for alogliptin. Considering that all are taken once daily and that the DPP-4 enzyme degrades a number of substrates in the body in addition to GLP-1, Dr. McGuire explained that any number of factors could be at play.

2. EMPA-HEART Shows Significant Reduction in Left Ventricular Mass with Jardiance, Offering Evidence of SGLT-2 Mechanism for Heart Failure Benefit in Humans

Dr. Subodh Verma (University of Toronto, Canada) presented primary results from the EMPA-HEART CardioLink-6 Trial, demonstrating that Jardiance (empagliflozin) significantly reduced left ventricular mass (LVM) in patients (n=97) with type 2 diabetes and established coronary artery disease. The study just completed in September, and these results represent an important mechanistic finding that helps unshroud some of the uncertainty surrounding SGLT-2 inhibitors’ mechanism of action for cardiovascular benefit, and more specifically for their effects on heart failure. Dr. Verma explained that results from EMPA-HEART indicate that empagliflozin treatment may promote reverse remodeling of the heart, contributing to observed CV benefits. In the trial, 97 patients between the ages of 40 and 80 years old with a history of type 2 diabetes and established, well-treated CAD were randomized to either once-daily 10 mg empagliflozin treatment or placebo and tracked for six months. For the primary outcome of LVM, patients were assessed by cardiac MRI, which is considered the “gold standard” for this endpoint; further, LVM is thought to be a strong predictor of CV outcomes. On this primary endpoint, empagliflozin treatment was associated with a significant reduction in LVM from baseline, reducing LVM by 2.6 g/m2 vs. 0.01 g/m2 with placebo. The adjusted mean difference between groups was -3.35 g/m2 (95% CI: -5.9, -0.81; p=0.01). This result remained statistically significant even after conducting sensitivity regression analyses over height and weight, reinforcing the robustness of the result.

  • As the discussant for these results, Dr. Elliot Antman (Brigham and Women’s Hospital, Boston, MA) called EMPA-HEART a “very important mechanistic study” and posited that the reduction in systolic blood pressure along with the increase in hematocrit seen with empagliflozin treatment may be potential drivers of the LVM reduction. He also noted the robustness of the primary endpoint result and called for more similar studies to be conducted (e.g. with other SGLT-2 inhibitors) in order to evaluate physiologic effects of SGLT inhibition. Importantly, he particularly called for studies that would investigate direct myocardial effects from these agents – for example, animal model results have suggested an increase in cardiac energy production through glucose and fatty acid oxidation, and Dr. Verma himself has also hypothesized an improvement in myocardial energetics through reduction of intracellular sodium and calcium levels via inhibition of the sodium-hydrogen exchanger in myocytes, which drives a downstream sodium-calcium antiporter. Other model studies have supported a shift to ketone metabolism that lowers left ventricular remodeling. All in all, our understanding of the effects SGLT-2s have on human physiology continues to evolve and is far from complete, from EMPA-HEART represents an important step in examining these agents in humans specifically.

3. Computational EMPA-REG OUTCOME Analysis Demonstrates Cost Effectiveness of Jardiance in Patients with Type 2 Diabetes and Heart Failure; ICER of ~$22,000 per QALY; ~1.2 Year Average Increase in Lifespan

An interesting computational analysis based on results from the EMPA-REG OUTCOME CVOT suggested that Lilly/BI’s Jardiance (empagliflozin) is a cost-effective treatment option for patients with type 2 diabetes and heart failure, from the perspective of commercial payers in the US – see the full poster here. The model projected outcomes for 10,000 patients over a 3-year mean trial duration (EMPA-REG OUTCOME enrolled 7,020 for a median 3.1 years) and utilized the hazard ratios and 95% confidence intervals for 11 endpoints (first HHF, subsequent HHF, CV death, non-fatal MI, non-fatal stroke, unstable angina, transient ischemic attack, revascularization, macroalbuminuria, renal injury, and renal failure) in the CVOT’s population of patients with heart failure at baseline (n=706). For each event that occurred, an associated quality of life reduction and average cost was applied, informed by the US Department of HHS, the IBM Micromedex RED BOOK, Medicare and Medicaid data, and other publicly available sources; the model included an estimation which accounted for overlapping QoL decrements as patients accumulated multiple clinical effects. Cost to the health plan assumed a 50% rebate for empagliflozin (we’re not sure how realistic this rebate is, and our understanding is that this data is not public), yielding a $401 monthly cost to payers per patient on empagliflozin. No other pharmacy costs were considered. In the base case analysis, clinical event costs were reduced by $5,957/patient, which partially off-set the cost of empagliflozin ($21,144/patient) and gave an incremental cost of $15,187/patient; longer survival of patients and reduced rate of clinical events translated to an incremental 0.67 Quality Adjusted Life Years (QALYs). Dividing the incremental cost by the incremental QALYs produced an incremental cost-effectiveness ratio (ICER) with empagliflozin plus standard of care vs. standard of care only of $22,644/QALY – which is well below the $100,000/QALY US cost-effectiveness threshold (this benchmark  is flexible and commonly ranges from $50,000/QALY to $150,000/QALY). The QALY threshold indicates the highest price the system is willing to pay for improvements in health – the lower the threshold, the more frugal. Investigators also performed sensitivity analyses examining the effects of varying discount rates (0-5%), drug costs (±20%), and other factors, which gave ICERs ranging from $12,465/QALY to $32,248/QALY, robustly supporting the conclusion that empagliflozin is cost-effective for patients with type 2 diabetes and heart failure across a variety of assumptions and at a threshold of $100,000/QALY.

  • Modeled patients treated with empagliflozin survived 1.22 years longer compared to standard of care alone. This aligns with a recent actuarial analysis of EMPA-REG OUTCOME which found an ~1 to 4.5 year increase in life expectancy with empagliflozin vs. placebo. Our understanding is that the 1.22 year figure refers to patients with existing heart failure included in the model, and it’s not surprising that these patients would exist on the low end of the alternative actuarial spectrum.

  • We would love to see this type of analysis at varying rebate levels and with different SGLT-2 inhibitors/GLP-1 agonists. Heart failure has been thrust even more strongly into the diabetes care spotlight following positive results from the DECLARE CVOT (for AZ’s SGLT-2 inhibitor dapagliflozin) at AHA 2018. Moreover, AZ’s CEO recently proclaimed that heart failure is “as costly to the healthcare system as cancer.” Perhaps more accurate models can be constructed based on results from the Dapa-HF (dapagliflozin in heart failure with reduced ejection fraction; n=4,500) and DELIVER (dapagliflozin in heart failure with preserved ejection fraction, n=4,700) trials for Farxiga, or the analogous EMPEROR-Reduced and EMPEROR-Preserved studies for Jardiance.

  • As with all computational studies, assumptions and limitations abound. Foremost, EMPA-REG OUTCOME was not powered or designed to assess treatment benefit in its HF subgroup, and even baseline HF status isn’t great – data collection on this comorbidity was, to our understanding, a check box for the investigator to fill out. Moreover, the analysis assumed a constant treatment effect for each event type regardless of changes in event or treatment history, and that changes in risk for clinical events due to changes in treatment were implicitly captured in event rate trajectories.

4. Diabetes Drugs and Heart Failure: Drs. Verma and McMurray Detail What We’ve Yet to Learn About Prevention and Treatment of HF

In a session on the metabolic aspects of heart failure, a true dynamic duo of Drs. Subodh Verma and John McMurray discussed the knowns and unknowns of diabetes drugs, with a particular eye toward heart failure. Dr. Verma was crystal clear in enumerating nine key unknowns and pointing to anticipated answers on four points:

  • Will DECLARE change clinical practice guidelines for a broader use of SGLT-2 inhibitors for HHF and renal benefit in people without ASCVD?

  • Are GLP-1 agonists efficacious in patients without diabetes but with ASCVD?

    • The SELECT CVOT for a high dose (2.4 mg) of Novo Nordisk’s Ozempic in obesity (with or without type 2 diabetes) will offer an answer. SELECT recently began recruiting 17,500 participants and should complete September 2023.

  • Are GLP-1 agonists benefit in type 2 diabetes without ASCVD?

    • The REWIND CVOT for Lilly’s Trulicity will offer an answer. While last week’s topline release indicates the trial was positive overall, we’ll have to see subgroup analyses for primary vs. secondary prevention to have a definitive answer on this question.

  • Do SGLT-2 inhibitors offer renal protection in type 2 diabetes?

    • The CREDENCE renal and CV outcomes trial for J&J’s Invokana will offer an answer. CREDENCE was stopped ~one year early this summer for meeting its primary endpoint ahead of schedule; results are anticipated “around Christmastime” this year.

  • Do SGLT-2 inhibitors offer renal protection outside of diabetes?

    • Both Dapa-CKD and EMPA-KIDNEY will offer answers. Dapa-CKD for AZ’s Farxiga is still enrolling 4,000 participants with CKD and should complete November 2020, while EMPA-KIDNEY is just about to start enrolling 5,000 participants with CKD and should complete June 2020.

  • How do we prioritize and individualize secondary prevention therapies based on ischemic risk vs. heart failure risk? (i.e., how do you pick and choose for a specific patient?)

  • Will GLP-1 agonist and SGLT-2 inhibitor combination therapy offer the greatest CV risk benefit?

  • Will metformin ever be displaced as first-line therapy in diabetes?

  • Will SGLT-2 inhibitors be used in the treatment of HF with preserved and/or reduced ejection fraction?

Dr. McMurray delved deeply into the final question, first drawing the key distinction between prevention of heart failure (“where we are today”) and treatment of heart failure (“the next step”). Going forward, this distinction will only become more important: As he explained, we’ve learned much about the prevention of incident heart failure with SGLT-2 inhibitors, but we know very little about those who have HF at baseline. In EMPA-REG OUTCOME and CANVAS, no information was collected on ejection fraction or biomarkers at baseline; investigators simply checked a box. Very positively, however, in the three completed SGLT-2 inhibitor CVOTs, participants benefitted on the composite of CV death and HHF regardless of recorded baseline heart failure status, as per this very-recent meta-analysis from Dr. Marc Sabatine et al. While more data is expected to come out of DECLARE, little can be concluded about the specificity of these “tantalizing” signals. Forthcoming data will reveal (i) whether HHF risk reductions apply to only patients with diabetes or to everyone with heart failure, (ii) whether the reduction applies to HF with preserved ejection fraction and/or with reduced ejection fraction, and (iii) whether the reduction impacts existing HF, or only prevents new HF episodes. Dr. McMurray found it encouraging that, so far, it doesn’t seem that the mechanism of CV benefit with SGLT-2s is mediated by any glucose or “metabolic” effects, particularly because the benefit appears so quickly. That said, even if benefit is dependent on dysglycemia, the majority of at-risk patients with HF will fall under that umbrella. Sooner rather than later, though, we’ll have a more definite answer (or a few of them), from five dedicated trials of SGLT inhibitors in heart failure:

  • Dapa-HF is investigating dapagliflozin in 4,744 participants with HFrEF and should complete December 2019. Dr. McMurray noted that the average eGFR in this study will be much lower than in DECLARE (85 ml/min/1.73 m2), which could be important.

  • DELIVER is investigating dapagliflozin in 4,700 participants with HFpEF and should complete June 2021.

  • EMPEROR-Reduced is investigating empagliflozin in 2,850 participants with HFrEF and should complete June 2020.

  • EMPEROR-Preserved is also investigating empagliflozin in 4,126 participants with HFpEF and should also complete June 2020.

  • SOLOIST-WHF is investigating sotagliflozin in 4,000 participants with worsening HF and type 2 diabetes and should complete January 2021.

5. “Are We Ready to Bell the Cat?” – Dr. Mikhail Kosiborod Advocates for Cardiologists to Take a More Active Role in Type 2 Diabetes Management and Better Utilize GLP-1 Agonists + SGLT-2 Inhibitors

In two separate presentations, Dr. Mikhail Kosiborod lamented that diabetes drugs proven to lower CV risk are still woefully underutilized by cardiologists and called for his field to take a greater role in managing CVD in patients with diabetes. Referencing a recent editorial, which he co-authored in Circulation, Dr. Kosiborod compared cardiologists’ current relationship to managing ASCVD in people with diabetes to the old fable “Belling the Cat”: A concerned group of mice are debating the best way to minimize the threat from a cat that has tormented them; one proposes putting a bell around the cat to signal when the threat will arrive. This solution is met with support until a wise old mouse asks, “Who is to bell the cat?” Each of the mice offer reasons why they cannot be the one to do so, so the cat continues to hunt the mice. In the words of Dr. Kosiborod, “It is one thing to say that something should be done but quite a different matter to actually do it.” He explained that the current state of cardiologists treating patients with diabetes displays a similar dynamic: The benefit of certain glucose-lowering drugs in reducing CV risk is readily apparent, yet their actual use by cardiologists in patients with diabetes is still stunningly low. Both empagliflozin (Jardiance) and liraglutide (Victoza) have robustly demonstrated MACE reductions and received indications from FDA and EMA for reducing CV risk in patients with diabetes; nevertheless, an overwhelming number of patients who stand to benefit from these drugs are not receiving them. In his editorial, Dr. Kosiborod cited previous analyses that have shown (i) only 5% of patients who met the inclusion criteria for EMPA-REG OUTCOME were on an SGLT-2 inhibitor, and (ii) only 6% of those who met LEADER inclusion criteria were on a GLP-1. As early as the year 2000, thought leaders were calling for cardiologists to be more involved in diabetes – see this JACC editorial from Drs. James Gavin and Gottlieb Friesinger. Why then, he asks, are cardiologists still sitting on the sidelines, when now more than ever they have powerful tools at their disposal?

Dr. Kosiborod placed significant emphasis for this underuse on cardiologists being hesitant to prescribe what are, in name, primarily glucose-lowering therapies. He acknowledged several other oft-cited justifications for this phenomenon: Cardiologists may not see it as “their place” to get involved with glucose management and may feel ill-equipped to prescribe diabetes therapies or follow up on diabetes issues. Moreover, they may be hesitant to “step on the toes” of referring clinicians, be they endocrinologists or PCPs. Cardiologists may fear hypoglycemia and other side effects, or be unfamiliar with diabetes reimbursement and patient education specifics. In response, he argued for a decoupling of glycemic control and cardiovascular benefits: “Cardiologists should not shy away from the initiation of these agents simply because they also happen to lower blood glucose, just like we would not shy away from initiating high-intensity statins in [patients with diabetes] with established ASCVD.” Dr. Kosiborod has previously shared the view that it’s completely within a cardiologist’s scope of practice to prescribe a GLP-1 or SGLT-2 for CV risk lowering. Moreover, he sees a huge opportunity for cardiologists to be coordinators of care for patients with diabetes by improving outcomes in both volume- and value-based models of care delivery. He floated the idea of “Cardiometabolic Centers of Excellence” that could coordinate the care of patients with type 2 diabetes and established CVD within preventative cardiology practices – ideally involving close collaboration with endos and PCPs and with a focus on event reduction. We’re always impressed by Dr. Kosiborod’s relentless advocacy for greater involvement of cardiologists in lowering CV risk and improving outcomes for people with diabetes.

  • Dr. Kosiborod questioned the notion that SGLT-2 inhibitors exert their effects on heart failure simply by acting as a diuretic. Session chair Dr. Naveed Sattar referenced a well-published recent comment that SGLT-2s are little more than a more expensive diuretic, questioning whether diabetologists might get closer to outcomes that matter with better use of diuretics. Dr. Kosiborod explained that while evidence does clearly suggest a volume-related effect with SGLT-2 inhibitors, thinking of these agents simply as diuretics alone is not sufficient. While it can be helpful to conceptualize them as such, he says, it seems that there’s much more going on physiologically, given their now well-established cardio- and nephron-protective properties. Whether there are within-class differences in how well these emerging effects are exerted, we think, remains to be seen. Further, this comment reflects the opinion we’ve heard from a variety of thought leaders, notably Dr. Subodh Verma at ESC 2018. Additionally, during Dr. Kosiborod’s Sunday presentation, he highlighted the increasingly well-demonstrated benefits of SGLT-2s on CKD (specifically, for slowing the progression of kidney disease), adding another layer to the impressive and unique benefits of the class.

6. CGM-Measured MAGE and Hyperglycemia Predict Recurrent Expanded MACE Events in Post-Acute Coronary Syndrome Patients; Cause vs. Correlation Remains Unclear

Researchers at Yokohama City University Medical Center found that mean amplitude of glycemic excursion (MAGE; measured and calculated via CGM) is a significant predictor of MACCE events (CV death, stroke, recurrent ACS, angina requiring revascularization, acute decompensated heart failure, stroke) in patients with prior acute coronary syndrome (ACS). The study (n=417), with a median 39-month follow-up, fitted patients with Medtronic’s iPro2 for at least 24 consecutive hours during a stable phase after ACS. For reference, MAGE is the mean difference between consecutive peaks and valleys if the difference was >1 standard deviation of the overall mean glucose level; it serves as a well-accepted marker of glycemic variability. It was found that 34% of the patient population (n=140) had type 2 diabetes, and average A1c was 5.9%. During follow-up, 66 patients experienced MACCE events: There were 5 CV deaths, 14 recurrences of ACS, 27 anginas requiring revascularization, 8 acute decompensated heart failure events, and 16 strokes. A univariate logistic regression of 25 demographic and metabolic factors revealed that high MAGE (>52 mg/dl) was the strongest predictor of a MACCE event (HR=2.35, 95% CI: 1.38-4.00, p=0.001). This was closely followed by high-sensitivity c-reactive protein >0.1355 mg/dL in stable phase post-ACS (HR=2.28, 95% CI: 1.33-3.92, p=0.002) and a glucose >180 mg/dl on admission (HR=2.24, 95% CI: 1.29-3.89, p=0.004). The Kaplan-Meier survival for patients grouped by MAGE (low vs. high, with 52 mg/dl as the threshold) separated at ~one year (below).

  • The most important question, in our assessment, is whether glycemic variability has an associative or causative relationship with CV risk in post-ACS patients (a particularly high risk group from a cardiovascular perspective) – a relationship that Dr. Lawrence Leiter discussed in detail at Diabetes Canada 2018. Given that the association between hyperglycemia and macrovascular risk seems to be grounded more in a common causative pathophysiology (i.e. insulin resistance) than a directional relationship, we imagine high glycemic variability serves more to “mark” high risk patients than to drive that risk. On the other hand, the specific question of hypoglycemia’s relationship to subsequent CV events remains open, and there is certainly data to suggest that episodes of severe hypoglycemia promote a high-risk state. Still, the question is far from answered, and it’s hard to imagine a definitive resolution coming anytime soon given the inherent impossibility of conducting an RCT to test this hypothesis. Taken together, we do feel these results reinforce the importance of CGM for patients with diabetes; at this point, that data is hardly actionable from a cardiovascular risk perspective, but could certainly play a role in helping to minimize hypoglycemia in those at risk.

7. Amarin Product Theater Offers Granularity on REDUCE-IT Results: A Dramatic 25% CV Benefit with Vascepa is Independent of Baseline Triglycerides (TG) and TG Lowering; Dr. Eliot Brinton Downplays Possible Confounding by Use of Mineral Oil Placebo

Following presentation of full results for REDUCE-IT on Saturday, an Amarin-sponsored Vascepa product theater discussed the science behind the drug, REDUCE-IT results, and implications for future use of Vascepa. We certainly sensed a buzz as conference attendees shuffled into this session, eager to learn more about a previously below-the-radar drug (pure prescription EPA) that’s caused a splash in the media. For our part, we remain very excited about Vascepa’s potential to improve outcomes for people with diabetes – who comprised 58% of REDUCE-IT’s enrolled population. We’ve picked out some of the most salient points from the presentation given by Dr. Eliot Brinton (Utah Lipid Center, Salt Lake City, UT and REDUCE-IT Steering Committee) below:

  • Dr. Brinton emphasized the surprising findings that Vascepa’s CV benefit did not appear to be dependent on baseline TG or even on TG lowering. He noted that this was “an astounding result” of REDUCE-IT: comparable CV benefit was seen across the full range of baseline TG levels studied (even in the 10% of subjects with TG < 135 mg/dL). As a lipidologist, Dr. Brinton joked that he felt “threatened” by these surprising findings. Further, citing comparable CV benefit above and below an on-treatment TG of 150 mg/dL, equally unexpected, he also said there is no need to “think Vascepa fails if it doesn’t get your patient’s TG below 150.” We find these points quite intriguing, as they suggest Vascepa might benefit a very broad patient population; of course, any official indication would have to reflect the study population, with elevated TG levels, that was enrolled in REDUCE-IT.  
  • Vascepa treatment had no effect on A1c; information is still being collected on diabetes agents added on during the trial, but no imbalance is expected. As a reminder, a major fraction (~58%) of the study population had diabetes at baseline, and the entire primary prevention cohort (~30%) in REDUCE-IT was comprised of patients with diabetes. Dr. Brinton remarked that he’s “sure that [some] diabetes agents were added [during] the trial” as patients tried to manage their A1c levels, but Amarin’s (very reasonable) belief is that there was no imbalance between these agents between the two groups, given that Vascepa did not impact A1c. Still, the impact on drop-in medication use must be assessed, particularly given that two classes of diabetes drugs are now known to be cardioprotective. Dr. Brinton mentioned that more details on diabetes-specific care in REDUCE-IT will be released at a later date – we’ll be sure to look out for this.

  • In response to chatter that the use of a mineral oil as the placebo may have contributed to the apparent magnitude of CV benefit with Vascepa by increasing placebo group risk, Dr. Brinton pointed toward the JELIS trial to reinforce Vascepa’s CV benefit. In that earlier randomized clinical CV outcomes trial (also testing pure EPA) there was no placebo agent used at all, and yet a 19% CV event reduction was still seen, roughly comparable to that in REDUCE-IT. Dr. Brinton leveraged these data to suggest that concerns over the effect of a mineral oil in placebo pills have been overblown: “If mineral oil is the entire explanation for what happened in REDUCE-IT, then tell me what happened in JELIS?” That being said, our understanding is not that critics have alleged the effect was entirely due to placebo, but perhaps exaggerated relative to a slight increased risk due to mineral oil in the placebo group. More reassuring was that, in a conversation with our team, Amarin CMO Dr. Craig Granowitz explained that increases in LDL-C and CRP values in the placebo group were relatively small, and that these lab values were not collected in an extremely rigorous fashion – therefore, he doesn’t put much stock into the significance of these results. He emphasized that the net difference on LDL-C at trial end was 5 mg/dl (+7 mg/dl in placebo, +2 mg/dl in Vascepa), which has arguably low clinical relevance; on CRP, he drove home the points that inflammation is highly variable and EPA has anti-inflammatory properties, which likely dampened any increase in inflammation in the Vascepa group. Dr. Granowitz also mentioned that separate analyses were performed for placebo group participants whose LDL-C levels increased vs. those whose level did not, and no difference in outcomes was observed.

8. Belviq Superior to Placebo on CAMELLIA’s Composite Renal Endpoint, Driven by Improvements in Albuminuria, CKD; Small but Significant Improvement in eGFR Over Three Years

On the heels of full metabolic results at EASD, Dr. Erin Bohula presented renal outcomes from the CAMELLIA CVOT for Arena/Eisai’s obesity therapy Belviq (lorcaserin) – see results presented at ESC 2018. As a reminder, Lorcaserin achieved superiority vs. placebo in the composite microvascular endpoint of persistent microalbuminuria, diabetic retinopathy, and diabetic neuropathy in patients with diabetes at baseline (HR=0.79, 95% CI: 0.69-0.92, p=0.001). This was driven by a significant (p=0.001) reduction in persistent microalbuminuria vs. placebo (HR=0.77, 95% CI 0.66-0.90, p=0.53). Only non-inferiority was established on diabetic retinopathy (HR=0.84, 95% CI: 0.50-1.43) and diabetic neuropathy (HR=0.94, 95% CI: 0.67-1.32), making us eager to see full renal outcomes – Belviq did not disappoint. On the prespecified renal composite outcome of new or worsening albuminuria, new or worsening CKD, doubling of serum creatinine, end-stage renal disease, renal transplant, or renal death, Belviq demonstrated a 13% relative risk reduction vs. placebo (HR=0.87, 95% CI: 0.79-0.96, p=0.006) with event rates of 4.2% and 4.9% per year, respectively. Superiority on the composite was driven by (i) persistent new or worsening albuminuria (HR=0.86, 95% CI: 0.76-0.97, p=0.017) and (ii) persistent new or worsening CKD (HR=0.81, 95% CI: 0.72-0.93, p=0.0018). Persistent doubling of creatinine trended toward Belviq (HR=0.75, 95% CI: 0.26-2.15, p=0.59) while ESRD trended towards placebo (HR=1.40, 95% CI: 0.72-2.71), both with very wide confidence intervals. Renal transplant/death rates were too low to establish a confidence interval. Lorcaserin also significantly improved eGFR vs. placebo with sustained mean treatment differences between 0.8 and 1.2 ml/min/m2 at each 6, 12, 24, and 36 months (p<0.001 for all) – the clinical significant of this difference, however, seems quite mild. UACR was not different between the two groups, though both experienced an increase of ~25 mg/g at the end of the three-year trial. An abbreviated Kaplan-Meier curve for the endpoint, consisting of annual data points for the three-year trial, demonstrated that the two groups began to diverge before the one-year mark (HR=0.86, 95% CI: 0.74, 1.01, p=0.061). The overall composite endpoint trended in favor of Belviq across all subgroups of primary vs. secondary CV prevention and baseline eGFR, UACR, BMI, and glycemic (diabetes, prediabetes, normoglycemia) levels.

  • We’re fairly surprised to see significant benefit with lorcaserin, given very mild effects on weight and A1c in CAMELLIA; Dr. Bohula said as much during her presentation. Over 3.3 years of follow-up, Belviq gave a mean treatment difference vs. placebo of ~four pounds and a 0.2%-0.3% drop in A1c. We’re not certain about the magnitude and clinical relevance of the improvements seen with Belviq, but it seems possible that lorcaserin is exerting an effect outside of A1c and weight lowering.

  • We’ll be interested to see if these results at all convince Eisai to focus more strongly on Belviq, after deprioritizing it for an Alzheimer’s candidate in 2Q18. Our guess is no, and we acknowledge the desperate need for better obesity pharmacotherapy in the ever-challenging obesity landscape. As Dr. Naveed Sattar put it at EASD 2018, “We still need safe weight loss drugs…We have excellent cholesterol, blood pressure, and diabetes drugs. Things are really, really good there. Yet, obesity and associated complications are rising, in part due to the obesogenic environment…Obesity is now, as I see it, the biggest challenge to those of us in the metabolic profession.”

Exhibit Hall


Capitalizing on the buzz around Vascepa and positive REDUCE-IT results (see our full coverage of these results here), Amarin made its first appearance in an exhibit hall we’ve covered. Amarin’s booth was near the entrance of the exhibit hall and featured huge banners touting impressive REDUCE-IT results, describing Vascepa as “A breakthrough in CV risk reduction” and cleverly proclaiming “What does Vascepa do to CV Risk? REDUCE-IT!” Amarin has undertaken a sizeable expansion of its sales force following the release of positive topline results from REDUCE-IT in September, growing from 150 to ~400 sales reps in anticipation of (and to help drive) a boom in sales; from our conversations in the booth, the novelty of the sales reps was fairly evident. The vibe we got from cardiologists visiting the booth was one of curiosity – it seemed as many had (understandably) not heard of Vascepa before. Therein lies the challenge for Amarin, as Vascepa will need to overcome a lack of prescriber familiarity – and build its class alone – while also straightening out confusion among patients over how Vascepa is different from dietary fish oil and EPA supplements. However, there’s no denying the enormous potential that Vascepa now has to become a widely used drug for CV risk reduction. Amarin has repeatedly indicated its desire to utilize a mass marketing approach with Vascepa and aims to minimize cost while attempting to reach as many patients as possible. On this note, we fully expect to see Amarin bring a full-throated presence to future conferences as the company spreads the word on Vascepa.


Amgen had a sizable booth near the center of the exhibit hall. PCSK9 inhibitor Repatha (alirocumab) was the obvious focus of Amgen’s exhibit, as the booth touted FOURIER results demonstrating Repatha’s significant CV risk reduction and LDL-C lowering vs. placebo. Representatives in the booth emphasized a new slogan that Amgen has rolled out – “Lower Together” – referencing Repatha’s reductions in LDL-C, CV risk, hassle, and cost. The final is a recent addition, as Amgen just last month announced a significant 60% list price reduction, taken in an attempt to increase access to Repatha, which carried a notoriously high list price for all of its previous time on the market. Access remains a tremendous issue for the PCSK9 inhibitor class: Reps underscored that only 0.2% of patients who could benefit from one of these agents are actually using either Repatha or Praluent currently. A lower list price, if it does improve access, would be a big step toward increasing the number of patients who will actually get Repatha. We’re definitely excited and curious to see what impact this decision will have on uptake; for more granularity on Repatha’s recent commercial performance, see our report on Amgen’s 3Q18 financial update. As an additional side note, we were excited and intrigued to see that Amgen’s booth featured an “Escape Room” that visitors could complete together – Amgen hoped that the activity could increase collaboration between attendees. We loved to see it!


A sprawling booth in the middle of the hall promoted AZ’s SGLT-2 Farxiga and antiplatelet therapy Brilinta in equal turn – we were quite struck by the level of material dedicated to Farxiga. Indeed, signing called out to attendees, “Welcome to the Farxiga booth,” and it was clear that AZ expected significant interest in the product following readout of full DECLARE results as a Saturday late breaker. Our sense is that AZ and Farxiga benefitted from a very strong focus on heart failure at AHA 2018 (we also think the DECLARE readout drove some of the robust discussion on HF). AZ reps echoed our excitement over a potential first-ever heart failure focused indication, likely including primary prevention, for a diabetes therapy.

We were, however, disappointed not to see any real estate for GLP-1 agonist Bydureon. To be sure, we understand that AZ is not able to promote its GLP-1 for cardioprotection given the narrow statistical miss on superiority for three-point MACE in the EXSCEL CVOT. However, given growing consensus around the idea of cardioprotection as a class effect of GLP-1 agonists, we would love to see more robust promotion of these agents to cardiologists. GLP-1s need to be prescribed more often and become more easily accessible to people with type 2 diabetes, and cardiologists doubtless have a key role to play in promoting their use.


The Lilly/BI brought a colossal, two-level (!) booth to AHA, almost exclusively devoted to SGLT-2 inhibitor Jardiance (empagliflozin). In line with AHA’s focus, reps were naturally concentrated on the products’ CV death indication in the US, which was granted based on an impressive 38% relative risk reduction vs. placebo in EMPA-REG OUTCOME. The “crow’s nest” of the booth featured several VR headsets fashioned as binoculars overlooking the exhibit hall floor. Fittingly, these headsets provided users with skyline views of several major US cities, using famous landmarks to illustrate the magnitude of Jardiance’s associated CV death reduction. For example, in Chicago, 25,000 people visit the Willis Tower (formerly the Sears Tower) every day. If every one of those people had type 2 diabetes, established CV disease, and was taking Jardiance, then 550 CV deaths could be prevented over the course of their lifetimes. By the same logic, 2,567 CV deaths could be prevented amongst the 116,667 yearly visitors to Alcatraz, in our home city of San Francisco. While we found this to be a somewhat roundabout way of conveying Lilly/BI’s point, we acknowledge that these numbers are certainly striking and the visualizations stuck with us – we wonder if a similar visualization could help convey Jardiance’s CV death reduction to patients. The lower level of the exhibit was devoted to a coffee bar and social milieu, where we asked a rep for his thoughts on Invokana’s recent US CV indication for reducing risk of full three-point MACE, despite achieving very comparable results in CANVAS as Jardiance achieved in EMPA-REG OUTCOME. We were thrilled to hear the rep express excitement at the indication, stating that it could only help the SGLT-2 class and, most importantly, patients– nice! We are always beyond inspired when companies and reps prioritize the larger picture of diabetes care over in-class competition – hats off to Lilly/BI and this rep.

There was no mention of DPP-4 Tradjenta at the booth, despite the recent readout of the CARMELINA CVOT at EASD. Given that the trial demonstrated only CV safety, and not CV superiority, vs. placebo (as was expected), we find it unsurprising that Lilly/BI prioritized their cardioprotective agent in this display.


Janssen was split between two booths at AHA; the larger was devoted exclusively to blood-thinner Xarelto while the smaller promoted SGLT-2 inhibitor Invokana’s very recently FDA approved CV indication for reducing three-point MACE in patients with type 2 diabetes and established CV disease in the US. This was certainly a massive win for J&J, and it was clear the company is trying to capitalize commercially. Four large screens repeated flashy infographics promoting Invokana as the first SGLT-2 inhibitor to be recognized for the reduction of all MACE components (Lilly/BI’s Jardiance – the only other SGLT-2 with a CV indication in the US – is only recognized for its impressive 38% reduction in CV death, despite a significant effect on MACE of the same magnitude as Invokana). Particularly impressive to us was how quickly this booth must have been pulled together – Invokana’s CV indication was announced just 10 days ago (we imagine production started before then ; >)! Contrast this to the company’s more defensive booth from AADE 2018, where J&J was mainly promoting its wealth (6.5 years) of safety data on Invokana: This was almost certainly attempt to staunch falling sales attributable to the amputation signal associated with canagliflozin in CANVAS (a signal that we continue to feel is not reflective of the risk faced by most patients). To this end, Invokana’s CV indication is a welcome tailwind for the product’s steadily falling sales; as it stands, the SGLT-2 inhibitor 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 declines in 3Q18. Needless to say, we are beyond thrilled that J&J seems to re-investing in this highly efficacious therapy once again.

Notably, there was no mention of upcoming data from CREDENCE (canagliflozin in CKD), which was stopped a year early this summer due to overwhelming efficacy of Invokana; full results are expected “around Christmastime” or early January. When we asked two company reps about CREDENCE, we were surprised that to hear that neither was aware of the trial’s discontinuation.


Merck occupied an unassuming and indistinct, single-person booth at AHA, with no promotional material – there were a few logos and charging ports available for attendees (we’ve seen an equivalent presence at other cardiology meetings). This was slightly surprising to us, given that the company heavily advertised DPP-4 inhibitor Januvia (sitagliptin) last year at AHA 2017. While some said that given increasing recognition of cardioprotection as a class effect of SGLT-2 inhibitors, they would have loved to see promotion of Pfizer-partnered Steglatro (ertugliflozin) at the meeting, since the associated CVOT hasn’t reported yet, we imagine that they have to be very careful about this. Results from the VERTIS CV CVOT will be critical to the product’s commercial success, given that Jardiance already has a CV indication, Invokana’s US indication was just approved, and full results from Farxiga’s DECLARE CVOT could support a potential CV indication for heart failure in both primary and secondary populations.

Novo Nordisk

Novo Nordisk brought a smaller booth than we’re used to (at diabetes meetings) to AHA 2018, and it was also located toward the back of the exhibit hall. Still, we were glad to see prominent promotion of both Ozempic and Victoza, the company’s duo of GLP-1 agonists. Booth materials leaned in favor of Ozempic – the product name was emblazoned across the display’s main panel – and we noted that the once-weekly was being promoted as superior to Trulicity on both A1c lowering and weight loss, with panels displaying very positive results from SUSTAIN 7. Reps indicated that there was substantial interest from conference attendees in these results and in learning more about Novo Nordisk’s GLP-1 agonists.

To be sure, however, Victoza was still very prominently featured, and understandably so: The injectable remains the only GLP-1 agonist to hold a CV indication from FDA, for reducing risk of MACE (CV death, non-fatal MI, non-fatal stroke) in people with type 2 diabetes and established CVD (based on LEADER). Of course, that could all change in the next year or so, as Lilly’s Trulicity has also demonstrated CV benefit in the REWIND CVOT, for which topline results were only just announced. We note that Ozempic has also demonstrated CV benefit in the SUSTAIN 6 trial, but that study alone is too small (n=3,297) and short (median FU 2.1 years) to support a CV indication, at least to FDA – in Europe, Novo Nordisk is allowed to promote Ozempic as superior to placebo on three-point MACE. For more on Novo Nordisk’s CVOT plans for Ozempic, see here.

Our understanding is very much that cardiology has been more receptive to the SGLT-2 than the GLP-1 class – and even the former remains underutilized for CV risk reduction (see Dr. Kosiborod, above). We’ve observed that cardiologists find SGLT-2s easier to prescribe than injectable GLP-1s (thought leaders consistently emphasize at cardiology meetings how simple the GLP-1 injection process is), and we also sense that the field has found consistent CVOT results for SGLT-2s somewhat more compelling. We’re glad to see Novo Nordisk continue to target the cardiology community on its mission to get GLP-1s to more patients with CVD who could benefit from them.


Sanofi/Regeneron had a well-sized booth at AHA 2018, with much of the focus on PCSK9 inhibitor Praluent (evolocumab). Praluent’s dosing flexibility was featured prominently throughout the booth, as the companies highlighted that Praluent is the only agent in the (two-member) class to offer two levels of efficacy at varying doses. Regarding patient access to Praluent, representatives in the booth explained that patients can qualify for a $0 copay card for Praluent – however, we do note that this only applies to patients on commercial insurance plans, and not those who are insured through Medicare or Medicaid or are uninsured. Access for PCSK9s continues to be a major point of discussion for both Sanofi/Regeneron and Amgen, especially in light of Amgen’s recent list price cut for Repatha. On this front, however, we imagine that the just-presented cost-effectiveness analysis for Praluent, based on ODYSSEY Outcomes data, will help with identifying patients who can benefit most from this therapy while also increasing justification for broader uptake of PCSK9s in this group.

Next-gen basal insulin Toujeo (insulin glargine) and combination therapy Soliqua (insulin glargine/lixisenatide) were also displayed on hovering screens above the booth, which was great to see; however, it was clear that these products weren’t the focus of the booth’s messaging efforts.


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