March 28-31, 2014; Washington, DC Day #1-3 Highlights - Draft

Executive Highlights

We’re in our nation’s capital where, instead of seeing the spring cherry blossoms, we’re still experiencing winter snow flurries. In the first three days of the American College of Cardiology’s 2014 Scientific Sessions, diabetes received a solid amount of attention, thanks in large part to new sub-analyses from the SAVOR and EXAMINE cardiovascular outcomes trials and the first presentation of full results from the AleCardio trial of Roche/Genentech’s discontinued dual PPAR-alpha/gamma agonist aleglitazar. Additionally, there was a serious amount of conference-wide excitement regarding the first presentation of phase 3 data on PCSK9 inhibitors –Amgen’s evolocumab demonstrated striking ~53-75% LDL reductions, even (in some studies) on top of intensive statin therapy, and Sanofi’s alirocumab demonstrated 47% LDL reduction as monotherapy.

This report contains our top five highlights from the first three days of the meeting, followed by an appendix that features our complete coverage. Stay tuned for our full report on the conference – day #4 focuses largely on obesity, and will cover topics such as bariatric surgery and “healthy obesity.” We will also see corporate events on type 2 diabetes drugs (namely, Janssen on Invokana), as well as a talk on diabetes and cardiovascular disease.

Top Five Highlights

1. A poster detailing sub-analyses of heart failure outcomes in the EXAMINE trial (Takeda’s CVOT for its DPP-4 inhibitor alogliptin) seemed to indicate that Takeda’s Nesina (alogliptin) is safe even in patients at high risk for (or with a past history of) heart failure (HF). Based on the 27% increased risk for hospitalization for heart failure (HHF) seen in AZ’s SAVOR CVOT for Onglyza (saxagliptin), some have expressed concern about using Onglyza (and perhaps even all DPP-4 inhibitors, in case of a class effect) in patients at high risk for HF. Analyses detailed in the poster showed that Nesina was not associated with an increase in MACE or a HF composite endpoint (HHF + CV death) in patients with a past history of HF (HR = 0.94 for MACE, 0.90 for HF composite), or in patients in the highest quartile of baseline pro-BNP levels (HR = 0.901 for HF composite; pro-BNP is a marker for heart failure). There was no significant impact of Nesina on the HHF component on its own (HR=1.19, 95% CI: 0.90-1.58). The presenters noted that the “HF composite” of HHF + CV death is a more robust measure than HHF alone, as the latter might exclude patients that died from HF before they could be hospitalized. AZ and the SAVOR investigators have not presented an analysis of HHF + CV mortality as a composite endpoint. All the data we have seen so far on Nesina has consistently proven assuring from a safety perspective – we will be interested to see what further CVOT sub-analyses Takeda and AZ present at ADA and other coming scientific meetings.
2. Another poster on Takeda’s EXAMINE presented a sub-analysis of CV death, which was one component of the primary MACE endpoint (CV death, myocardial infarction, stroke). Rates of CV death were non-significantly lower on alogliptin (4.1%) vs. placebo (4.9%) (HR=0.85, 95% CI: 0.66, 1.10; p=0.07). Excitingly, however, subgroup analyses of CV death found some populations in which alogliptin conferred a benefit: women, patients with eGFR ≥60 ml/min/1.73 m², and those with a duration of type 2 diabetes <five years. The latter signal seems plausible, as patients with less severe diabetes might have derived a more measurable benefit from alogliptin than those with more severe diabetes at baseline (who might have already been to far down the path to complications). Dr. William White (University of Connecticut, Farmington, CT), the primary author, commented that he did not know why women might have also benefitted (that’s a pretty big sub-group). These results are quite notable given that EXAMINE enrolled a very sick patient population (type 2 diabetes post-acute coronary artery syndrome), and showing any kind of CV benefit in this population is quite difficult especially given the very short duration of the trial (mean 18 month follow up).
3. SAVOR investigators presented two new sub-analyses of the trial, including an oral presentation stratifying results by baseline renal function and a poster analyzing outcomes associated with baseline biomarker status. The renal analysis found that CV outcomes between AZ’s Onglyza and placebo were fairly consistent regardless of baseline renal function. Onglyza’s effect on improving microalbuminuria held up in people with baseline eGFR ≥30 ml/min/1.73 m², but was not significant in people with eGFR <30 ml/min/1.73 m². There was a trend towards less hypoglycemia on Onglyza in people with eGFR <30 ml/min/1.73 m², although this may have been confounded by lower use of SFUs in this population and the small number of patients in this group (only 336 of the trial’s 16,492). The baseline biomarker analysis found that saxagliptin (Onglyza) did not have an effect on CV outcomes regardless of baseline biomarker status, even in patients at highest clinical risk. We interpreted this result to mean that one cannot necessarily rely on the three biomarkers studied (high sensitivity troponin, NT-proBNP, and high sensitivity CRP) to predict how a patient will respond to saxagliptin in terms of CV outcomes.
4. Dr. A. Michael Lincoff (Cleveland Clinic, Cleveland, OH) revealed results of the AleCardio trial, the failed cardiovascular outcomes trial for Roche/Genentech’s PPAR-alpha/gamma dual agonist, in a late breaking clinical trials session. Despite promising effects on lipids observed in phase 2, aleglitazar did not prove to be cardioprotective (a futility analysis at the time 55% of planned events had accrued predicted a 1% chance of success). The primary MACE endpoint (CV death, myocardial infarction, stroke) resulted in a neutral HR of 0.96 (95% CI: 0.83, 1.11; p=0.57), and aleglitazar was associated with a host of other safety issues. These included the typical PPAR issues of edema and weight gain, heart failure, and bone fractures as well as an unexpected increase in GI hemorrhages, a dramatic (but reversible) rise in serum creatinine, and greater hypoglycemia compared to placebo. Phase 2 glycemic effects were substantiated in AleCardio, with aleglitazar providing ~0.5% A1c reduction from a baseline of 7.8% (0.6% placebo-adjusted reduction). All in all, Roche’s choice to pull the drug seemed the right one, as the benefit/risk profile was not compelling.
5. We got our first look at full phase 3 results for PCSK9 inhibitors (a novel class of injectable, non-statin, LDL lowering drugs), and wow were we impressed! For Amgen’s evolocumab (which currently leads the pack in terms of timing) we saw durable mean LDL cholesterol reductions of roughly 53-75% (see Appendix for specific details on each of the different study populations). In many studies, this reduction was from a relatively well controlled baseline (~95 mg/dl), and in addition to pre-existing statin or ezetimibe therapy. A key finding was that the magnitude of LDL lowering was not very diminished by background therapy, even when that background therapy was a high-intensity statin. The studies tested evolocumab as monotherapy in drug-naïve patients, as add-on to baseline lipid-lowering therapy, as add-on to high-intensity statin therapy, and in statin-intolerant patients – the potential for the class is likely highest in the latter two populations. We did not see any substantial signals in terms of safety and tolerability. The GAUSS-2 in statin-intolerant patients (an area of high unmet need, and a target population for PCSK9 inhibitors), led to remarkable mean LDL reductions of over 100 mg/dl (from a high baseline around 190 mg/dl). The presentation sessions on this data were absolutely packed, evincing the huge level of interest in this novel class (even as the AHA/ACC 2013 lipid guidelines de-emphasize LDL targets and emphasize statins).
   • Sanofi also presented the first phase 3 data for its PCSK9 candidate, alirocumab, in a poster detailing results of a 24-week monotherapy trial against ezetimibe. Results were not quite as impressive as Amgen’s, with the intent-to-treat population experiencing a 47% LDL reduction (compared to 16% reduction on ezetimibe).
   • While Sanofi beat Amgen to announcing topline phase 3 data in late 2013, Amgen will beat Sanofi to submission. Sanofi does not plan to submit alirocumab until 2015, and Amgen expects to submit evolocumab in 2014.

Detailed Discussion and Commentary

Posters: Acute Coronary Syndromes: Comorbid Considerations

Alogliptin in Patients with Type 2 Diabetes After Acute Coronary Syndromes: Heart Failure Outcomes and Cardiovascular Safety in Heart Failure Patients

F Zannad, CP Cannon, WC Cushman, GL Bakris, SK Heller, RM Bergenstal, V Menon, A Perez, P Fleck, R Oh, C Mehta, S Kupfer, CA Wilson, WB White

In the months since the primary results of the EXAMINE (on Takeda’s Nesina) and SAVOR (for AZ’s Onglyza) CVOTs were shared at ESC 2013, we have heard building discussion regarding a potential DPP-4 inhibitor class effect on heart failure due to the 27% statistically significant increase in heart failure seen in SAVOR and the perceived homogeneity of the DPP-4 inhibitor class. Although there is broad consensus that the current data is not sufficient to recommend a change in overall prescribing behavior for DPP-4 inhibitors, some have expressed concern about prescribing Onglyza (and perhaps all DPP-4 inhibitors, in case of a class effect) in patients with at high risk for heart failure. This poster provided evidence from EXAMINE that Takeda’s agent is safe in that patient group, which could be a notable differentiating factor in a drug class that has very few other differentiating factors, especially based on safety. 

• There was no significant impact of Nesina on hospitalization for heart failure. Hospitalization for heart failure was included as part of a pre-specified secondary expanded MACE composite endpoint (first occurrence of all-cause mortality, nonfatal MI, stroke, urgent revascularization due to unstable angina, and HHF), and heart failure was also analyzed in the classic composite endpoint of HHF plus CV death. A post-hoc analysis also broke out HHF as an individual component. When HHF was broken out from the expanded MACE composite – i.e., when it was only counted if it was the first CV event the patient experienced – the hazard ratio was 1.07 (95% CI: 0.79, 1.46) (P-value not provided). In the post-hoc analysis of HHF as an individual component, which counted all cases, even those that were preceded by another MACE event, the hazard ratio was 1.19 (95% CI: 0.90, 1.58); the latter analysis was first presented by Dr. Deepak Bhatt (Harvard Medical School, Boston, MA) at EASD 2013.
• In patients with a pre-trial history of heart failure, Nesina was not associated with any significant difference in either of two composite outcomes (the expanded MACE endpoint described above, and the composite of HHF + CV death). In those patients with a history of heart failure, the hazard ratio for the heart failure composite endpoint was 0.90 (95% CI: 0.70, 1.17; p=0.446), and the hazard ratio for expanded MACE was 0.94 (95% CI: 0.74, 1.20; p=0.622). This data would not support an assertion that prescribers should avoid using this agent in patients with a history of heart failure. SAVOR’s heart failure analysis presented at AHA 2013 also found no effect of Onglyza on HHF risk in people with a prior history of HF.
• Baseline levels of pro-BNP (a predictive biomarker for heart failure) did not predict worse outcomes in EXAMINE. A similar post-hoc analysis of SAVOR data showed that the majority of the absolute increase in heart failure hospitalization was limited to the quartile of patients with the highest baseline pro-BNP levels. However, the relative risk was not significant significantly different, which means that pro-BNP levels also cannot be used in SAVOR to predict the effect of saxagliptin, but we imagine that some prescribers are worried about avoiding use of Onglyza in patients at high baseline risk for heart failure. In EXAMINE, there was no statistically significant difference in risk for the heart failure composite endpoint (hospitalization for heart failure + CV death) in any of the baseline pro-BNP quartiles, including the uppermost quartile. Although it was not mentioned in the poster, lead investigator Dr. William White mentioned that pro-BNP levels actually decreased slightly during the trial, a promising sign with regards to CV safety.
• Questions remain about the best way to analyze heart failure. The presenters noted that the “heart failure composite” of hospitalization for heart failure + CV death is a more robust measure than hospitalization for heart failure alone, as the latter might exclude patients that died from heart failure before they could be hospitalized. AZ and the SAVOR investigators have not broken out this composite endpoint (it was not one of the pre-specified endpoints of either trial). All the data we have seen so far on Nesina has consistently proven assuring from a safety perspective – we will be interested to see what further sub-analyses Takeda and AZ present at ADA and other coming scientific meetings.

Cardiovascular Mortality in Patients with Type 2 Diabetes and Recent Acute Coronary Syndromes from the EXAMINE Trial

WB White, S Kupfer, WC Cushman, GL Bakris, SR Heller, RM Bergenstal, V Menon, AT Perez, PR Fleck, R Oh, CR Mehta, CA Wilson, F Zannad, CP Cannon

This poster presented a sub-analysis of cardiovascular (CV) death in the EXAMINE CV outcomes trial (CVOT) for Takeda’s DPP-4 inhibitor Nesina (alogliptin). As a reminder, the primary results of the EXAMINE trial (n=5,380 patients with type 2 diabetes post-acute coronary syndrome) found that alogliptin had a similar effect vs. placebo on the composite MACE (CV death, myocardial infarction, stroke) primary endpoint. This sub-analysis broke out the individual CV death component of the composite MACE endpoint. Rates of CV death were non-significantly lower on alogliptin (4.1%) vs. placebo (4.9%) (HR=0.85, 95% CI: 0.66, 1.10), and lead author Dr. William White pointed out that the p-value was very close to statistical significance (p=0.07). All-cause mortality and sudden cardiac death were also non-significantly lower on alogliptin than placebo (HR=0.88, 95% CI: 0.71-1.09 and HR=0.80, 95% CI: 0.57-1.12, respectively); p-values not specified. The analysis actually identified some subgroups in which alogliptin conferred a benefit with regard to CV death: in women (HR=0.60, 95% CI:0.40, 0.91), patients with eGFR ≥60 ml/min/1.73 m² (HR=0.67, 95% CI: 0.46, 0.98), and those with a history of type 2 diabetes <five years (HR=0.61, 95% CI:0.37, 1.00) there was a lower rate of CV death in the alogliptin group compared to placebo. This seems to make sense, as it suggests that patients with less severe diabetes achieved greater benefit from alogliptin than those with more severe diabetes at baseline. This is quite notable considering that EXAMINE enrolled a very sick patient population, sicker than the SAVOR population. These sub-group analyses for CV mortality in SAVOR have not been presented. Dr. White remarked that it is unknown why women might have derived greater benefit from saxagliptin than men.

Oral Sessions: Stable Ischemic Heart Disease: Basic Sciences and Clinical Studies

Baseline Renal Function and Cardiovascular Risk in Patients Treated with Saxagliptin: Observations from the SAVOR-TIMI 53 Trial

Jacob Udell, MD (Brigham and Women’s Hospital, Boston, MA)

Dr. Jacob Udell presented a secondary analysis of the SAVOR cardiovascular outcomes trial (CVOT) results stratified by baseline renal function. As a reminder, SAVOR’s primary results found that AZ’s DPP-4 inhibitor Onglyza (saxagliptin) had similar CV effects compared to placebo in 16,492 people with type 2 diabetes and cardiovascular disease or multiple cardiovascular risk factors (see primary results here). For this new sub-analysis, patients were stratified by baseline renal function into normal/mild (eGFR >50 ml/min/1.73 m²; n=13,916), moderate (eGFR=30-50 ml/min/1.73 m²; n=2,240), or severe (eGFR <30 ml/min/1.73 m²; n=336) renal dysfunction groups. As would be expected, CV outcomes were worse in those with worse baseline renal function. However, the effect of saxagliptin vs. placebo on CV outcomes was similar regardless of renal function. Saxagliptin’s effect on improving microalbuminuria held true in the normal/mild and moderate impairment groups, but was not statistically significant in the severe impairment group. Saxagliptin lowered A1c after one year regardless of renal function, and this effect persisted to two years except in the moderate impairment group. Interestingly, saxagliptin’s effects on hypoglycemia appeared to improve with worsening baseline renal function (although not significantly). Dr. Udell speculated in Q&A that this may have been due to the very small size of the severe impairment group or the lower use of sulfonylureas in this group. The results of this study suggest that the CV and glycemic safety and efficacy of saxagliptin are fairly consistent regardless of renal function and that saxagliptin may be used in the population of people with type 2 diabetes and CKD – a vulnerable population that has few treatment options.

• This sub-analysis of the SAVOR cardiovascular outcomes trial stratified patients by baseline renal function into normal/mild (eGFR >50 ml/min/1.73 m²; n=13,916), moderate (eGFR=30-50 ml/min/1.73 m²; n=2,240), or severe (eGFR <30 ml/min/1.73 m²; n=336) renal dysfunction groups. By design, the study capped the severe renal impairment group to roughly 300 people. People in the normal/mild group received the standard 5.0 mg/day dose of saxagliptin while people with moderate or severe renal impairment received the reduced 2.5 mg/day dose. A single dose adjustment was made to 2.5 mg/day if a patient’s renal function dropped below an eGFR of 50 ml/min/1.73 m². Otherwise, patients remained on the 2.5 mg/day dose despite eGFR declines or progression to dialysis.
• As would be expected, CV outcomes were worse in those with worse baseline renal function. After two years, the MACE (CV death, myocardial infarction, stroke) event rate in the severe renal impairment group was 17% compared to 12% and 6% in the moderate and normal/mild groups, respectively (p<0.0001). Hospitalization for heart failure followed a similar trend with even rates of 13%, 6%, and 2% in the severe, moderate, and normal/mild groups after two years (p<0.0001).
• However, CV outcomes were similar for saxagliptin vs. placebo regardless of baseline renal function. Rates of the primary MACE endpoint, the secondary MACE+ endpoint (MACE plus hospitalization for heart failure, coronary revascularization, or unstable angina), CV death alone, or MI alone were similar between saxagliptin and placebo groups regardless of baseline renal function (HRs range from 0.79 to 1.30 with all 95% CIs crossing 1.0). As a reminder, SAVOR’s primary results found that saxagliptin and placebo had nearly identical effects on CV outcomes.
  
  • The relative risk of hospitalization for heart failure (HHF) was also similar by renal function, although the absolute risk difference for saxagliptin was higher in the moderate renal impairment group. The hazard ratios for HHF for saxagliptin vs. placebo were 1.24 in the normal/mild group (95% CI: 0.99, 1.55), 1.46 in the moderate group (95% CI: 1.07, 2.00), and 0.94 in the severe group (95% CI: 0.52, 1.71); p-interaction=0.43. Dr. Udell did not dwell on this point, attributing the significant difference within the single moderate impairment category to the relatively small sizes of the moderate and severe renal cohorts. Because the p-interaction value was not significant, baseline renal function and HHF were not considered to have an interaction.
• As found in the primary analysis, saxagliptin improved microalbuminuria more than placebo; this effect held true in the normal/mild and moderate renal impairment groups, but was not significant in the severe impairment group. In the normal/mild group, microalbuminuria improved in ~10% of patients on saxagliptin compared to ~8% of those on placebo (p<0.001); in the moderate group, microalbuminuria improved in 11% of patients on saxagliptin compared to 9% on placebo (p=0.037); and in the severe group, microalbuminuria improved in 14% of those on saxagliptin compared to 11% on placebo (p=not significant).
• Saxagliptin lowered A1c after one year regardless of renal function; this effect persisted to two years except in the moderate impairment group. In the normal/mild group, saxagliptin provided a 0.3% A1c reduction after one year and after two years compared to baseline (whether this value was placebo-adjusted was not specified). In the moderate group, saxagliptin lowered A1c by 0.3% after one year and 0.2% after two years (the latter difference was not statistically significant). In the severe group, saxagliptin produced a 0.6% A1c reduction after one year and after two years.
• Interestingly, saxagliptin’s effects on hypoglycemia appeared to improve with worsening baseline renal function (although not significantly). Hospitalization for hypoglycemia tended to be more common (non-significantly) on saxagliptin in the normal/mild group compared to the moderate group and the severe group (HRs = 1.44, 1.06, and 0.77, respectively; all 95% CIs crossing 1.0). For major hypoglycemia, the HR for saxagliptin vs. placebo was 1.11 in the normal/mild group (95% CI crossing 1.0), 1.91 in the moderate group (95% CI not crossing 1.0), and 0.65 in the severe group (95% CI crossing 1.0).
  
  • Dr. Udell speculated during Q&A that this may be a product of the very small sample size in the severe renal impairment group or that it may be explained by the lower use of sulfonyulreas in the severe renal impairment group.

Questions and Answers

Q: It’s rather complicated because I have heard that DPP-4 also degrades bradykinin. And there is a difference between men and women in terms of degradation of bradykinin because women have an additional gene on the X chromosome for amino peptidase D. Did you see any difference in the % of women that developed heart failure and men who developed heart failure during the study? Secondly, was there any difference between women and men who took ACE inhibitors during the study? This could lead to a possible explanation of why saxagliptin produced the increased risk of heart failure.

A: The question is a little tangential to the talk here. Dr. Benjamin Scirica presented a heart failure analysis at AHA, and the full manuscript is still under review. To disclose full details at this point would be premature. I can say that there was no interaction between sex at baseline. The bradykinin analyses are still being performed.

Q: Any thoughts why the people with the most severe renal insufficiency would have less hypoglycemia episodes?

A: Many hypotheses have been postulated. One is to be circumspect of any outcome when the patient population was only 336 patients, especially when the interaction terms did not suggest significant heterogeneity in other outcomes. It could be play of chance. The other possibility is that these patients were on fewer SFUs, so they may be less prone to hypoglycemia.

Oral Sessions: Joint American College of Cardiology/Journal of the American Medical Association Late-Breaking Clinical Trials

Evaluation of the Dual PPAR-Alpha/Gamma Agonist Aleglitazar to Reduce Cardiovascular Events in Patients with Acute Coronary Syndrome and Type 2 Diabetes Mellitus: The AleCardio Trial

A. Michael Lincoff, MD (Cleveland Clinic, Cleveland, OH)

Dr. A. Michael Lincoff presented results of the phase 3 AleCardio trial, the terminated cardiovascular outcomes trial for Roche/Genentech’s dual PPAR-alpha/gamma agonist aleglitazar. Because of positive effects on lipids observed in phase 2, Roche decided to first pursue a CV indication for aleglitazar in patients with type 2 diabetes to differentiate the drug from other PPAR agonists (and their baggage) and to avoid the treacherous and unpredictable waters of a general diabetes indication. Notably, the futility analysis conducted when 55% of planned events had accrued found that the trial had only a 1% chance of achieving its goal of CV superiority with regards to the primary MACE endpoint (CV death, myocardial infarction, stroke). The final analysis (including 74% of planned events) showed a HR of 0.96 (95% CI: 0.83, 1.11; p=0.57). The lack of CV protection coupled with the myriad other safety concerns that emerged from the trial painted quite an unattractive picture. Aleglitazar was associated with the typical PPAR agonist issues (compared to placebo, it had a 22% trend towards elevated hospitalization for heart failure, 8 lb weight gain, and 30% increase in bone fractures) as well other safety concerns including a dramatic (but reversible) rise in serum creatinine, an unexpected 44% increase in GI hemorrhage, and 60% greater hypoglycemia compared to placebo. The decent glycemic efficacy and benefit on lipids observed in earlier phase 2 trials were reproduced, but it was very clear that the benefit/risk profile was too heavily weighted towards risk. Dr. Lincoff remarked that this trial illustrated the difficulty in developing a PPAR agonist, and he conjectured during Q&A that this may be the end for new PPAR agonist trials. These results were published simultaneously in JAMA.

• Dr. Lincoff reviewed aleglitazar’s phase 2 SYNCHRONY data showing that aleglitazar reduced A1c by 0.5% (0.8% placebo-adjusted, which was at least as good, if not better, as high-dose pioglitazone in the same trial) while also providing a roughly 30% reduction in triglycerides (44% placebo-adjusted) and a 25% increase in HDL (29% placebo-adjusted).
• AleCardio was a phase 3 cardiovascular outcomes trial for aleglitazar designed to show CV protection in people with type 2 diabetes and recent acute coronary syndrome (ACS). Roche had been investigating aleglitazar for three potential indications: (i) CV risk reduction in people with type 2 diabetes post ACS; (ii) CV risk reduction in people with type 2 diabetes and prediabetes; (iii) general type 2 diabetes. AleCardio was, thus, a superiority trial intended to inform the first indication. The trial enrolled 7,226 patients randomized 1:1 to receive once-daily 150 μg aleglitazar or placebo. It was an event driven trial planned for 950 positively adjudicated primary endpoint events (CV death, myocardial infarction, stroke), anticipated to take about 2.5 years. An interim analysis had been planned at the time when 80% of events had accrued, at which time the data safety monitoring board (DSMB) could recommend early termination for efficacy or futility.
• The trial was ultimately terminated early when the DSMB identified a higher incidence of adverse events in the aleglitazar group at a regularly scheduled safety meeting. They performed an unplanned futility analysis at the time 55% of total events had accrued (522 events). At that point, they found that the primary endpoint had a HR of 1.01 with only a 1% chance of achieving superiority by the end of the trial. As such, the DSMB recommended termination for futility, and the trial ended on July 2, 2013. The trial database was locked on December 17, 2013, at which point 74% of total events had accrued (704 events) with a median follow up of 104 weeks.
• The final analysis of the primary MACE endpoint (CV death, myocardial infarction, stroke) showed neutrality. The event rate was 9.5% on aleglitazar compared to 10% on placebo for a HR of 0.96 (95% CI: 0.83, 1.11; p=0.57). Most other CV endpoints examined (MACE plus hospitalization for unstable angina; the composite of all-cause mortality, myocardial infarction, and stroke; and the individual components of all-cause mortality, CV death, myocardial infarction, and stroke) were also neutral.
  
  • However, when analyzed as individual components, unstable angina and unplanned revascularization were each better in the aleglitazar arm (HR=0.75, 95% CI: 0.57,0.96 and HR=0.79, 95% CI: 0.69,0.90, respectively).
• Aleglitazar demonstrated decent efficacy with regards to glycemic and lipid control. From a relatively good baseline A1c of 7.8% in both groups, aleglitazar provided a 0.6% placebo-adjusted A1c reduction (~0.5% absolute reduction compared to ~0.1% rise on placebo) that was sustained through 36 months. HDL also sustained a 20% placebo-adjusted increase from a baseline of 42 mg/dl, and triglycerides a ~30% placebo-adjusted decrease from a baseline of 152-154 mg/dl.
• Not unexpectedly, aleglitazar was associated with fluid retention and, thus, a trend towards more hospitalization for heart failure. Patients on aleglitazar gained 4.6 kg (10.1 lbs) compared to 0.9 kg (2.0 lbs) weight gain on placebo. Hospitalization for heart failure occurred 22% more in the aleglitazar arm (p=0.14).
• As previously characterized in the phase 2 SYNCRHONY and AleNephro trials, aleglitazar was associated with a reversible decrease in renal function. In AleCardio, aleglitazar elevated creatinine to about 0.10-0.14 mg/dl from a baseline of 0.1 mg/dl within one week of starting the drug, but the effect disappeared within four weeks of discontinuation. The composite renal endpoint of end-stage renal disease (ESRD), doubling of serum creatinine, or 50% increase in serum creatinine leading to study drug discontinuation occurred in 7.4% of aleglitazar patients compared to 2.7%  of placebo patients (HR=2.85, 95% CIL 2.25-3.60, p<0.001). We presume this was driven by the elevations in serum creatinine; Dr. Lincoff did not comment on the incidence of ESRD, which of course is not reversible (as of yet).
• Other safety concerns included an unexpected increase in GI hemorrhage, higher rates of bone fractures, and higher hypoglycemia. GI hemorrhage occurred in roughly 1.5% of aleglitazar patients compared to roughly 0.8% of placebo patients (HR=1.44, 95% CI: 1.03, 2.00, p=0.03). Bone fractures were slightly elevated with the hazard ratio being 1.30 (95% CI: 0.94, 1.80). Hypoglycemia, defined as patients experiencing at least one event, occurred in 17% of the aleglitazar arm compared to 11% in placebo (HR=1.60, 95% CI: 1.41, 1.82, p<0.001).
• Baseline characteristics: patients on average were 61 years old, with BMI of 29 kg/m². About 10% had newly diagnosed diabetes, and the mean duration of type 2 diabetes was nine years. Ten percent had a history of heart failure, and 75% a previous myocardial infarction. About two-thirds of patients were on metformin, one-third on sulfonylureas, and 30% on insulin. The vast majority was taking heart drugs with 95-96% on aspirin and 92-93% on a statin.
• Results were published simultaneously in JAMA.

Questions and Answers

Q: It is disappointing that somehow we can’t find  a way to get the diabetics to have less CV risk. How d you explain the transient renal insufficiency?

A: Two of aleglitazar’s phase 2 studies show pretty clearly that it’s a hemodynamic effect. Like with ACE inhibitors, it is reversible.

Q: In the ACCORD study, with fenofibrate there was an increase in events with women in very dyslipidemic groups. Did you see any gender differences in AleCardio?

A: We have not conducted the subgroup analyses in all of the endpoints, but there is clearly no difference by gender in the primary endpoint.

Q: As a clinician, where do these data put the whole PPAR family for your patients who have diabetes and ACS?

A: Functionally, of course, the relevant drug is pioglitazone. The PROactive trial demonstrated CV benefit, but that is of questionable significance now in the era of high-dose statins. I think it’ll be difficult to see new development of PPARs. AleCardio shows that they are very risky drugs no matter how promising they are in phase 2. I don’t know if we can justify pioglitazone strictly on the basis of cardioprotection, but it is certainly a good diabetic agent for those not at risk for heart failure.

Q: It was interesting to see that in the pioglitazone meta-analysis, the curves began to diverge after about one-to-two years, and unfortunately perhaps we won’t have the answer for this study to see if that would in fact occur. The other comment I had was, certainly given the results of the ACCORD lipid trial, it might be interesting to see whether your subgroup with high triglycerides and low HDL may have benefitted, as well as people perhaps with lesser duration of diabetes.

A: Those are two very important points. Triglycerides, because we used an upper limit of 400 mg, were very weak. There was a borderline significant interaction of 0.06 for HDL. That is also a good point about being too late in the disease progression. If we’re trying to interfere in diabetes win people with established CVD, it may well be too late for these interventions, but to carry out trials in a low risk early diabetes population, that’s the real challenge in terms of logistics because you can’t wait a decade for return on investment.

GAUSS-2: A Phase 3 Double-blind, Randomized Study to Assess the Safety and Efficacy of Evolocumab (AMG 145) in Hypercholesterolemic Subjects Unable to Tolerate an Effective Dose of Statin

Erik Stroes, MD (Academic Medical Center, Amsterdam, Netherlands)

Statins are widely accepted as the front-line therapy for the reduction of CVD risk, but studies suggest that anywhere between 10% and 20% of patients are unable to tolerate statins in the real world. These statin-intolerant patients are a key area of unmet need that PCSK9 inhibitors can address, a theme that Dr. Eric Stroes touched upon often as he presented the results of the GAUSS-2 phase 3 trial for Amgen’s evolocumab. The 12-week, randomized, double-blind, ezetimibe-controlled study enrolled 307 patients, and tested evolocumab biweekly and monthly (which were generally clinically equivalent in this study, consistent with other phase 3 studies). All patients had demonstrated intolerance to at least two statins. Compared to other phase 3 studies, where patients were generally relatively well controlled before beginning evolocumab, the mean baseline LDL-C in GAUSS-2 was ~195 mg/dl, evincing the lack of effective therapies for statin-intolerant patients. Evolocumab led to a 56% mean reduction in LDL from baseline, a 38% reduction over ezetimibe (the primary existing option for statin-intolerant patients). Joking that Europeans such as himself “have not forgotten their numbers” and still care about LDL targets, he shared that evolocumab helped ~70-90% of statin-intolerant patients achieve a goal of under 70 mg/dl. The incidence of myalgia (muscle pain, the leading cause of statin intolerance) was high compared to other rates of myalgia in other evolocumab studies, but was lower with evolocumab (8%) than ezetimibe (18%). Dr. Stroes noted that nearly everyone’s statin intolerance is caused by statin-induced myalgia, so the fact that so many of these people were relieved of this symptom when switching to evolocumab is highly significant. Dr. Stroes concluded that an agent that can achieve LDL reductions of over 100 mg/dl in a patient population with such great unmet need is certainly exciting.

Questions and Answers:

Comment: I think it will be important to investigate to see if the effects seen in this trial last longer than 12 weeks.

A: The DESCARTES study ran for a full 52 weeks, and you can see from those results that the drug’s efficacy is sustainable for long periods of time – we are very positive on the sustainability of the effect. We have been fortunate to have most patients want to continue in the open-label continuation phase of our studies.

Q: I am intrigued by these results, but I don’t think we should discount the value of statins in the broader patient population. Do you think there is a benefit in pushing LDL as low as we can?

A: In our center, the majority of patients had tried four or more statins – we are absolutely not in competition with statins, because we have to start with statins. Is lower LDL always better? We can discuss the guidelines, but there is so much data showing that lower is indeed better. We are fond of low values, particularly in high-risk patients.

Q: Regarding myalgia, you still saw myalgia in the evolocumab group. Did you investigate that further?

A: The percentage of patients with myalgia was much higher in the ezetimibe group than the evolocumab group. This was a highly selected patient group that was prone to musculoskeletal side effects. There were around 20% of patients who were on very low doses of statins during the trial, and myalgias were likely more frequent in that patient group.

Comment: If there had been a placebo group in the study, there still probably would have been a high incidence of myalgia there. I agree that I do not think your agent has a significant musculoskeletal safety issue.

LAPLACE-2: A Phase 3, Double-blind, Randomized, Placebo and Ezetimibe Controlled, Multicenter Study to Evaluate Safety, Tolerability and Efficacy of Evolocumab (AMG 145) in Combination With Statin Therapy in Subjects With Primary Hypercholesterolemia and Mixed Dyslipidemia

Jennifer Robinson, MD, MPH (University of Iowa, Iowa City, IA)

Dr. Jennifer Robinson took the podium in the enormous “main tent” to present the full results of the 12-week phase 3 LAPLACE-2 study on Amgen’s PCSK9 inhibitor evolocumab. The trial tested the effect of monthly or biweekly evolocumab (vs. ezetimibe or placebo) when added to one of five regimens of moderate or high-intensity statin therapy. Consistent with the rest of the evolocumab phase 3 data we have seen so far, the results for monthly and biweekly dosing were clinically equivalent. The efficacy results from LAPLACE-2 were the most striking we saw out of Amgen’s five phase 3 presentations at ACC: among the various study arms, the addition of evolocumab to statin therapy led to mean LDL-C reductions of 63-75% versus placebo, taking the vast majority of patients (~85-95%, depending on particular statin therapy) to an LDL-C of under 70 mg/dl. The drug also improved other lipid parameters, and there were no major imbalances in adverse events (both consistent with the rest of the phase 3 results we have seen). A key takeaway from this trial was that evolocumab had similar efficacy regardless of whether a patient was on moderate or high-intensity statin background therapy. For patients that have trouble tolerating a high-intensity statin, we wonder if a new option will be to downshift the statin dose and add a PCSK9 inhibitor. This could be a preferred paradigm for type 2 diabetes patients, as statins can slightly increase glucose levels.

• Dr. Robinson spoke briefly on the relevance of this trial and of PCSK9 inhibitors as a whole in light of the new 2013 ACC/AHA cholesterol guidelines. She acknowledged that Amgen designed the phase 3 program for evolocumab before the release of the new guidelines, which de-emphasize specific LDL targets for most patients and are seen as not favoring PCSK9 inhibitors. However, she pointed out that non-statin therapies such as evolocumab have relevance in statin-intolerant patients, patients with familial hypercholesterolemia, and patients already on statins that need additional LDL lowering. Additionally, outside the US, most countries still recommend a specific LDL target of under 100 or 70 mg/dl, depending on patients’ risk levels.

Questions and Answers

Q: How do you think this therapy fits in with our new guidelines?

A: Trials suggest that we can use evolocumab to achieve additional LDL cholesterol lowering in patients who cannot take high-intensity statins, and in patients who are totally statin intolerant. Another potential use would be patients with familial or genetic hypercholesterolemia, including those who cannot take high-dose statins.

Q: Have you seen any neurocognitive effects with evolocumab? That is an issue that the FDA is looking into.

A: In the entire evolocumab program, there is no signal of imbalance in neurocognitive effects. That is something that we have monitored extremely carefully. When you get older people with such low LDL levels for the first time, neurocognitive function is an area in which we would want to closely examine since LDL is in every cell of the body, but we haven’t seen a signal so far. We’ve even sub-analyzed patients with LDLs below 45 and even 25 mg/dl, and there has been absolutely nothing.

Posters: The Roles of Diabetes, Obesity, and Kidney Disease in Stable Atherosclerotic Heart Disease

Prognostic Implications of Simultaneous Biomarker Assessments in Patients with Type 2 Diabetes Mellitus: Observations from the SAVOR-TIMI 53 Trial

BM Scirica, DL Bhatt, E Braunwald, I Raz, PG Steg, J Davidson, MA Cavender, B Hirschberg, A Umez-Eronini, KA Im, P Jarolim, & DA Morrow

Dr. Benjamin Scirica presented an analysis of CV outcomes in the SAVOR cardiovascular outcomes trial based on 12,182 patients’ baseline biomarker status for NT-proBNP (a marker of hemodynamic stress), high sensitivity troponin T (hsTnT; a marker of myocardial necrosis [tissue death]), and high sensitivity CRP (hsCRP; a marker of inflammation). As would be expected, elevated levels of each biomarker were independently associated with increased rates of CV death, myocardial infarction, and hospitalization for heart failure. Notably, saxagliptin did not have an effect on CV outcomes regardless of baseline biomarker status, even in patients at highest clinical risk. This means that even in people with the highest baseline levels of these biomarkers, saxagliptin did not any increased risk relative to placebo for CV events. This is consistent with Dr. Scirica’s analysis of hospitalization for heart failure (HHF) at AHA 2013 where he found no population in which the relative risk of HHF due to saxagliptin was particularly high or low but that saxagliptin’s absolute risk of heart failure was significantly elevated in people with the highest quartile of NT-proBNP levels at baseline.

Oral Sessions: Featured Clinical Research

Efficacy and Safety of Evolocumab (AMG 145) Monotherapy Compared With Ezetimibe and Placebo in Hypercholesterolemic Subjects: A Phase 3 Randomized Clinical Trial

Michael Koren, MD (Jacksonville Center for Clinical Research, Jacksonville, FL)

Dr. Michael Koren began the much-awaited series of phase 3 data presentations on Amgen’s PCSK9 inhibitor candidate evolocumab; the session was so popular that both the presentation room and the simulcast screens were packed to capacity. MENDEL-2, which tested 140 mg biweekly and 420 mg monthly doses of evolocumab compared to ezetimibe (a non-statin cholesterol lowering drug) and placebo as monotherapy in patients naïve to lipid medications, was one of the first of Amgen’s phase 3 trials to report topline data (back in December). After 12 weeks, biweekly and monthly evolocumab delivered ~55-57% reductions in LDL-C relative to placebo, and ~38-39% reductions versus ezetimibe. These effects were established by week 2, and were remarkably constant through week 12 (the DESCARTES study, which was presented next [see below], demonstrated durability through a full year). Every patient on evolocumab (n=153) saw a decrease in LDL-C from baseline (generally over 40%), something that could not be said about the placebo (n=76) or ezetimibe (n=77) arms. The drug yielded significant reductions in ApoB, and interestingly, the monthly dose of evolocumab seemed to have a slightly stronger effect on triglycerides than the biweekly dose – this was the only notable difference between the two doses, and Dr. Koren did not characterize it as clinically meaningful. The drug was well-tolerated, with no individual adverse events showing a worrying imbalance. Dr. Koren concluded that, although Amgen anticipates that evolocumab will primarily be useful in high-risk patients in conjunction with statins, it was promising to see that the drug led to marked LDL-C reductions as monotherapy.

• During Q&A, Dr. Koren noted that statin therapy can be seen as a confounding factor in studies of PCSK9 inhibitors, as they upregulate the expression of PCSK9. This counterregulatory effect, he added as an aside, is part of the reason why statins show diminishing efficacy at very high doses. Studying PCSK9 inhibitors such as evolocumab as monotherapy allows investigators to observe the true effect of the drug without this potentially confounding factor.

Questions and Answers

Q: Could you talk about the rationale of conducting a study in patients not on a statin?

A: Statins are known to upregulate PCSK9 levels. It is very possible that in patients who are not on statins, you may see different results. The point of the study was to look at safety and efficacy in patients not confounded by statins. The monotherapy study provides data on the side effect profile of the drug itself.

Q: For the non-lipidologists among us, can you speak a little more about why PCSK9 levels rise with statins?

A: Statins upregulate PCSK9 production, which is a theory explaining why statins lose their incremental effect as one increases the dosage. It is an offsetting mechanism against the primary effect of statins.

Q: Do you have any information on how patients accepted the injectable administration of the drug?

A: We have some information from previous studies and some anecdotal information from this study, and those anecdotal reports indicated that the drug administration was accepted very well. In the OSLER study, patients in a blinded phase 2 study were allowed to proceed to an open-label study phase, and most patients agreed to enroll, showing that they were very accepting of the therapy. That finding was especially notable given that the study used the older injection technology, not the more convenient auto-injector that we use now.

Q: Could you talk more about the auto-injectors?

A: The auto-injector, if I recall correctly, used a 27-gauge needle. It was very easy to use – even I could use it.

Long-Term Tolerability and Efficacy of Evolocumab (AMG 145) in Hyperlipidemic Subjects: A 52-Week Phase 3 Double-Blind, Randomized, Placebo-Controlled Study

Dirk Blom, MD, PhD (University of Cape Town, South Africa)

Dr. Dirk Blom presented the results of the phase 3 DESCARTES trial, a long-term, 52-week trial investigating the safety and efficacy of evolocumab in addition to baseline lipid-lowering therapy. The study first stabilized patients on one of four therapies, depending on lipid control: i) diet alone; ii) diet and atorvastatin 10 mg; iii) diet and atorvastatin 80 mg; and iv) diet, atorvastatin 80 mg, and ezetimibe 10 mg. As expected, there were significant differences in baseline characteristics between these four baseline treatment groups. Patients from each group were randomized to evolocumab or placebo injection in equal measure. Following stabilization of baseline therapy, patients were relatively well controlled, with mean LDL-C levels generally between 100-120 mg/dl. Patients on evolocumab saw a striking 57% reduction in LDL-C relative to placebo injection (-51.5% vs. +6.0% for evolocumab and placebo, respectively), and 82% of evolocumab patients achieved an LDL-C goal of <70 mg/dl at week 52 (relative to 6% of patients on placebo). Notably, this efficacy was fairly constant across baseline therapy subgroups, although it did seem to be slightly diminished in patients in the most intensive background therapy subgroup (diet, atorvastatin, and ezetimibe), a group that also saw slightly more modest reductions in PCSK9 levels.

Questions and Answers:

Q: When you say monthly dose, do you mean every 31 days or every 28 days?

A: We administered every 28 days.

Q: You discussed one patient with short-term memory loss, which I believe recovered spontaneously. Was that recovery while the patient was on or off of the agent?

A: I’ve only read what is in the clinical study report, that the patient started a supplement with green coffee bean extract, and then had trouble remembering things. It might have just been the caffeine making her jittery, but in any case, the patient recovered spontaneously while still on evolocumab.

Posters: Familial Hypercholesterolemia, Novel Therapies, and Cardiovascular Risk

First Data with the 75 mg Every 2 Weeks Dose of Alirocumab, a Proprotein Convertase Subtilisin/Kexin Type 9 Inhibitor: Superior Lipid-Lowering Versus Eetimibe in a Phase 3 Monotherapy Trial

EM Roth, MR Taskinen, HN Ginsberg, JJP Kastelein, HM Colhoun, JG Robinson, L Merlet, R Pordy, MT Baccara-Dinet

This poster presented the first phase 3 data for Sanofi’s PCSK9 inhibitor, alirocumab. In this 24-week monotherapy study (ODYSSEY MONO; ClinicalTrials.gov Identifier: NCT01644474), 103 patients were randomized 1:1 to 75 mg alirocumab every two weeks (Q2W) or 10 mg ezetimibe (an oral active comparator). Patients on alirocumab could be blindedly uptitrated to 150 mg if LDL remained ≥70 mg/dl at week 12 (14 patients were up-titrated). From a mean baseline LDL of ~140 mg/dl in both groups, alirocumab produced a 47% LDL reduction in the intent to treat (ITT) population and 54% reduction in the per-protocol (PP) population. In comparison, ezetimibe produced a 16% LDL reduction in the ITT population and 17% reduction in the PP population. While comparing trials is inherently challenging, this appears somewhat less impressive than the 12-week and one-year results that Amgen presented at this meeting for evolocumab (AMG 145); LDL reductions in those trials ranged from ~53-75%. In both the ITT and PP population, alirocumab also improved ApoB, non-HDL, HDL, and total cholesterol better than ezetimibe, while changes in Lp(a), triglycerides, and ApoA-1 were not significantly different (Amgen’s evolocumab did generally show differences in these parameters). While Sanofi beat Amgen to announcing topline phase 3 data in late 2013, Amgen will beat Sanofi to submission. Sanofi does not plan to submit alirocumab until 2015, and Amgen expects to submit evolocumab in 2014.

-- by Jessica Dong, Manu Venkat, and Kelly Close