Lilly releases topline REWIND results: Trulicity achieves superiority to placebo on three-point MACE in majority primary prevention population – November 5, 2018

Executive Highlights

  • Lilly just announced that GLP-1 agonist Trulicity, in the REWIND CVOT, significantly reduced risk of three-point MACE vs. placebo. No numbers were given – and apparently won’t be until a presentation at ADA 2019. Of high note, only 31% of participants in REWIND had baseline CVD, compared to 73% to 100% for all other GLP-1 CVOTs. With an average A1c of 7.3%, the REWIND population is substantially lower-risk than traditional CVOTs, making the results much more generalizable. Median follow-up was 5.4 years, also much longer than average.

  • Will REWIND support a primary prevention indication for Trulicity? It’s hard to say with so few details, including effects in the primary vs. secondary prevention cohorts. But REWIND does seem to support the use of Trulicity – and likely other GLP-1s – to prevent CV disease, not just recurrent CV events. Lilly will submit data to regulatory agencies in 2019, and we strongly hope that these results can bring more attention to the comorbidity of diabetes and CV disease. We look very forward to seeing more on other secondary endpoints, particularly weight.

  • REWIND offers further support to an already-strong case for a class cardioprotective effect with GLP-1 agonists. To date, LEADER for Novo Nordisk’s Victoza, SUSTAIN 6 for Novo Nordisk’s Ozempic, and HARMONY for GSK’s discontinued Tanzeum have demonstrated CV benefit, and EXSCEL for AZ’s Bydureon is widely considered to belong in that group despite a narrow miss on statistical superiority. And while some have asserted that CV protection with GLP-1s should reasonably extend to lower-risk patients (it simply takes a longer time to demonstrate this in clinical trials), REWIND is the strongest evidence to date for primary prevention cardioprotection with the class.

    • We’re also very eager to see data on REWIND’s secondary microvascular endpoint. Evidence from LEADER especially points to a possible microvascular (specifically renal) benefit with GLP-1s, and we would tentatively hypothesize that REWIND’s longer duration of follow-up might make any microvascular effects easier to detect.

Lilly just announced positive topline results for the REWIND CVOT of GLP-1 agonist Trulicity (dulaglutide). Trulicity demonstrated superiority over placebo on a primary endpoint of three-point MACE (CV death, non-fatal MI, non-fatal stroke); however, no numbers were released.

Importantly, 69% of REWIND participants did not have baseline CVD, which represents a sizable increase over the primary prevention populations enrolled in other completed GLP-1 CVOTs. Aside from REWIND, the EXSCEL CVOT for AZ’s Bydureon enrolled the largest primary prevention cohort at 27% (see table below). And while the age, BMI, and diabetes duration of REWIND’s population are not very different from other GLP-1 CVOTs, the baseline A1c in REWIND was very low at 7.3% – compared to 7.7% for ELIXA, 8.0% for EXSCEL, and 8.7% for HARMONY, SUSTAIN 6, and LEADER.

Overall, this overall lower-risk population makes REWIND much more generalizable than prior CVOTs. The “real-world” applicability of these results, we agree, is exciting: It seems to us that REWIND supports the use of Trulicity, and possibly other GLP-1s, to prevent cardiovascular disease in the first place, not just to prevent CV events in those with baseline CVD. We’ll be very interested to see any analyses (such as this, for SGLT-2 inhibitors – a presentation called  “Are Diabetes CVOT Results Generalizable? Analysis of Eligibility Requirements of SGLT-2 Inhibitor CVOTs (EMPA-REG, CANVAS, DECLARE) Suggests Not,” by Dr. Eric Wittbrodt at ACC earlier this year) of what proportion of diabetes patients overall would have qualified for REWIND.

The generalizability of CVOT results has emerged as one of the biggest scientific and clinical questions in diabetes. Historically, these trials have enrolled very high-risk patients to achieve a high event rate. REWIND balanced its lower-risk population against a median follow up of 5.4 years, which is almost two years longer than the next-longest GLP-1 CVOT (LEADER; 3.8 years). As such, the only GLP-1 agonist to garner a CV indication to date – Novo Nordisk’s Victoza – was limited to use for CV risk reduction in those with established cardiovascular disease.

At this point in time, however, REWIND results do come with a big caveat: There is no data available on effects seen in the primary vs. secondary prevention cohorts, nor is there an interaction statistic available. Dr. Philip Home aptly pointed out to us that if the 31% of participants with prior CVD were followed for 5.4 years and saw a hazard ratio advantage similar to that of SUSTAIN-6 and HARMONY, then the number could be sufficient to give statistical superiority for the whole study even if the non-CVD group saw no benefit whatsoever. Of yet, generalizability is more of a possibility than a definite. Combining high and low risk populations, he says, comes with this inherent challenge. 

We’re very hopeful that REWIND results can, more than previous studies, bring attention to the issue of diabetes and cardiovascular disease. Of course, diabetes in and of itself is a risk factor for CVD, but studies have shown that awareness of this connection is less than ideal. Moreover, both GLP-1s and SGLT-2 inhibitors – both have which have demonstrated cardioprotection – should be used more broadly to lower CV risk.  

Given that Novo Nordisk has cited Victoza’s CV indication as a key factor in sustaining – and even boosting – Victoza sales in the face of Ozempic’s launch, we would expect a primary prevention indication to be a sizable commercial tailwind for Trulicity. Of course, Trulicity has done well regardless: Lilly’s GLP-1 recently became the US market leader by volume and sold $780 million on a steep upward trajectory in 2Q18.

We do wish that more information could be given in “topline” releases: we’ll have to wait nearly eight months to learn more about REWIND results, which Lilly hopes to present in full at ADA 2019 in San Francisco. There’s no doubt that these findings should impact clinical practice, and we would be glad to see that happen sooner rather than later – if we were practicing medicine, we do think that this release would make it more likely that we’d prescribe Trulicity. We note that topline releases for EMPA-REG OUTCOME, DECLARE, LEADER, EXSCEL, and SUSTAIN 6 also did not contain even topline numbers. As far as we know, there was no topline release for CANVAS or HARMONY ahead of result presentations.

Lilly plans to submit REWIND results to regulatory authorities sometime in 2019. For more details on REWIND’s design and baseline characteristics, see here.

GLP-1 Agonist CVOT Results

Trial and Status

Enrollment, Percent w/ Baseline CV Disease

Mean Age (years), BMI (kg/m2), A1c, and Diabetes Duration at Baseline

Study Design:

Median Follow Up, Concomitant Medications Allowed?

Primary Endpoint Result (components)

Mortality Results:

CV Death, All-Cause Mortality

Other Notable Findings

REWIND for dulaglutide (Lilly’s Trulicity)

Topline results announced; full results at ADA 2019



Age: 66

BMI: 32

A1c: 7.3%

Diabetes Duration: 10 years

5.4 years

No other GLP-1 agonists or DPP-4 inhibitors

Demonstrated superiority on three-point MACE



LEADER for liraglutide (Novo Nordisk’s Victoza) (outcomes paper)

Completed – reported ADA 2016



Age: 64

BMI: 33

A1c: 8.7%

Diabetes Duration: 13 years

3.8 years

No other GLP-1 agonists or DPP-4 inhibitors

13% RRR

(HR=0.87, 95% CI: 0.78-0.97, p=0.01 for superiority, p<0.001 for non-inferiority)

3-point MACE (CV death, non-fatal MI, non-fatal stroke)

CV: 22% RRR

(HR=0.78, 95% CI: 0.66-0.93)

All-Cause: 15% RRR

(HR=0.85, 95% CI: 0.74-0.97)

Indication for reducing MACE approved by FDA August 2017

Heart Failure: NS

(HR=0.87, 95% CI; 0.73-1.05)

EXSCEL for exenatide once-weekly (AZ’s Bydureon) (outcomes paper)

Completed – reported EASD 2017



Age: 63

BMI: 32

A1c: 8.0%

Diabetes Duration: 12 years

3.2 years

No other GLP-1 agonists


(HR=0.91, 95% CI: 0.83-1.00, p<0.001 for non-inferiority, p=0.06 for superiority)

3-point MACE (CV death, non-fatal MI, non-fatal stroke)


(HR=0.88, 95% CI: 0.76-1.02)

All-Cause: exploratory

(HR=0.86, 95% CI: 0.77-0.97)

Heart Failure: NS

(HR=0.94, 95% CI: 0.78-1.13)

SUSTAIN-6 for semaglutide (Novo Nordisk’s Ozempic) (outcomes paper)

Completed – reported EASD 2016



Age: 65

BMI: 33

A1c: 8.7%

Diabetes Duration: 14 years

2.1 years

No other GLP-1 agonists or DPP-4 inhibitors

26% RRR

(HR=0.74, 95% CI: 0.58-0.95, p=0.02 for superiority, p<0.001 for non-inferiority)

3-point MACE (CV death, non-fatal MI, non-fatal stroke)


(HR=0.98, 95% CI: 0.65-1.48)

All-Cause: NS

(HR=1.05, 95% CI: 0.74-1.50)


76% increased risk for retinopathy w/ semaglutide (HR=1.76, 95% CI: 1.11-2.78); Premarket CVOT not designed for superiority; does not support CV indication

Heart Failure: NS

(HR=1.11, 95% CI: 0.77-1.61)

ELIXA for lixisenatide (Sanofi’s Adlyxin) (outcomes paper)

Completed – reported ADA 2015



Age: 60

BMI: 30

A1c: 7.7%

Diabetes Duration: 9 years

2.1 years

No other GLP-1 agonists or DPP-4 inhibitors


(HR=1.02, 95% CI: 0.89-1.17, p=0.81)

4-point MACE (CV death, non-fatal MI, non-fatal stroke, hospitalization for unstable angina)


(HR=0.98, 95% CI: 0.78-1.22)

All-Cause: NS

(HR=0.94, 95% CI: 0.78-1.13)

Heart Failure: NS

(HR=0.96, 95% CI: 0.75-1.23)

HARMONY for albiglutide (GSK’s Tanzeum)

(outcomes paper)

Completed – reported EASD 2018



Age: 64

BMI: 32

A1c: 8.7%

Diabetes Duration: 14 years

1.6 years

No other GLP-1 agonists

22% RRR

(HR=0.78, 95% CI: 0.68-0.90, p=0.0006 for superiority, p<0.0001 for non-inferiority)


(HR=0.93, 95% CI: 0.73-1.19)

All-Cause: NS

(HR=0.95, 95% CI: 0.79-1.16)

GSK has withdrawn commercial support for Tanzeum; current commercial status unclear

GLP-1 Agonist CVOT Design Comparisons

From the accompanying infographic released with the REWIND announcement:

Class Implications

  • REWIND bolsters already-strong evidence for a cardioprotective class effect with GLP-1 agonists and also offers the strongest-ever evidence that this benefit extends to people with type 2 without established CVD. REWIND joins three other GLP-1 agonist CVOTs in demonstrating superiority on three-point MACE:

    • LEADER for Novo Nordisk’s Victoza (liraglutide)

    • SUSTAIN 6 for Novo Nordisk’s Ozempic (semaglutide), and

    • HARMONY for GSK’s discontinued Tanzeum (albiglutide).

    • Additionally, EXSCEL for AZ’s Bydureon (exenatide) is widely considered a positive trial for both Bydureon and the GLP-1 class despite a narrow miss on statistical superiority (attributed to aspects of study design).

    • The ELIXA trial for Sanofi’s Adlyxin (lixisenatide) is somewhat of an outlier due to its tremendously high-risk population (recent ACS patients), and lixisenatide is also a very short-acting molecule (half-life of ~3 hours).

  • In our observation, the release of HARMONY data at EASD 2018 brought the field together in favor of a class effect for GLP-1 agonists – read comments from seven thought leaders here. The fact that Tanzeum conferred very nominal improvements on A1c and weight over the duration of HARMONY very well supports that cardioprotection through GLP-1 agonism occurs independently of metabolic effects. In the words of Dr. John Buse, “The main thing is that [HARMONY] demonstrates that the GLP-1 cardiovascular effect is probably mediated through the GLP-1 receptor.”

  • The larger debate now is whether cardioprotective benefits extend to lower-risk patients. From a very strict evidence-based perspective, there’s a gap on primary prevention with GLP-1s – though REWIND is doing much to fill that gap! That said, many feel that the CV benefits of GLP-1s should extend to lower-risk patients; unfortunately, it simply takes longer to show that in a clinical trial, due to a lower baseline event rate. At AACE 2018, for example, Dr. Mikhail Kosiborod argued that it doesn’t physiologically make sense for GLP-1 benefits to be limited to secondary CV prevention. Particularly with a hypothesized anti-atherosclerotic mechanism, which confers risk reduction gradually over time, one can imagine that GLP-1s given for a long duration could give very substantial risk reductions. REWIND is the most powerful evidence to date in favor of primary prevention with GLP-1s, and we’re incredibly eager to see the full data.  

  • Microvascular outcomes from REWIND could have a tremendous impact on our understanding of the benefit GLP-1s may offer on eye and kidney outcomes. REWIND included a secondary microvascular endpoint comprised of (i) diabetic retinopathy requiring laser therapy, vitrectomy, or anti-VEGF, (ii) clinical proteinuria, (ii) a 30% decline in eGFR, and (iv) need for chronic renal replacement therapy.

    • From where we stand, the renal benefits associated with the SGLT-2 inhibitor class have prompted significant buzz in the field – and dedicated outcomes trial data, from CREDENCE, now supports this benefit directly – but relatively less attention has been given to GLP-1 agonists.

    • Our sense is that so far GLP-1s haven’t demonstrated as robust of an effect on microvascular outcomes, but the data is nothing to scoff at: In LEADER, liraglutide gave a 16% risk reduction on a composite microvascular endpoint encompassing renal and ophthalmic adverse outcomes (HR=0.84, 95% CI: 0.73-0.97, p=0.02), driven by a 22% improvement in renal outcomes (95% CI: 0.67-0.92, p=0.003). Further, there was very notable attention dedicated to renal outcomes and GLP-1s at EASD, where we saw post hocs from SUSTAIN 6, ELIXA, and others.

    • We would tentatively speculate that REWIND is poised to demonstrate an even greater magnitude of benefit on microvascular outcomes. Given that microvascular disease is a longer-term complication, it’s possible that the significantly longer duration of follow up in REWIND allowed for demonstration of a greater microvascular benefit. At the same time, the mechanism of action by which GLP-1s may give improvements on renal or other outcomes is not known (i.e. through glucose lowering or another pathway). We very much hope that microvascular outcomes will be presented with CV outcomes at ADA 2019.

Close Concerns Questions

  1. Will FDA grant a cardiovascular indication for primary prevention to Trulicity?

  2. Was there any imbalance in cardioprotective or other effects between the primary and secondary prevention cohorts? That is, did Trulicity do more to prevent initial CV events or reduce recurrent CV events?

  3. How can the diabetes treatment paradigm shift more toward prevention of CV disease and microvascular complications, particularly by taking GLP-1 agonists?

  4. What will the microvascular/renal endpoint show? Does REWIND’s longer follow-up (median 5.4 years) give it a better chance of showing an effect on microvascular outcomes?

  5. How was the benefit on three-point MACE balanced across individual components?

  6. What are the barriers to longer-term follow-up for CVOTs, beyond cost? What could be learned from yearly check-ins out to 10 or 15 years? Could more cognitive or patient-reported outcomes be incorporated?

  7. What percent of the general diabetes population would have qualified for REWIND, compared to other GLP-1 agonist and SGLT-2 inhibitor CVOTs?

  8. How does the design and conduct of REWIND compare to others, for example in terms of concomitant medication use and study visit setup?

  9. How can REWIND be leveraged to expand patient and HCP awareness of CVD risk and burden among people with diabetes?


--by Ann Carracher and Kelly Close