American Heart Association
American Heart Association
January 04, 2019
5 min read

Guest Commentary: DECLARE-TIMI 58 trial validates CV benefits of SGLT2 inhibitors in type 2 diabetes

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Photo of Eldrin Lewis
Eldrin Foster Lewis

Results from the DECLARE-TIMI 58 trial, presented during a late-breaking session at the 2018 American Heart Association Scientific Sessions, add to the growing body of evidence that SGLT2 inhibitors have a favorable effect on CV outcomes in patients with type 2 diabetes. In the trial, the benefit of the SGLT2 inhibitor dapagliflozin was primarily driven by a 27% reduction in the rate of hospitalization for HF compared with placebo. In this Guest Commentary, Eldrin Foster Lewis, MD, MPH, associate professor of CV medicine at Harvard Medical School and director of the Cardiovascular Clerkship Program at Brigham and Women’s Hospital , discuss es the clinical implications of the trial and the role of SGLT2 inhibitors in diabetes and CVD.

The introduction of the glucose-lowering SGLT2 inhibitor class is exciting for patients with type 2 diabetes. The main reason is because prior to the original EMPA-REG study, there were no agents targeting glucose that were associated with improved CV outcomes.

There are three SGLT2 inhibitors with data from large clinical trials that clearly show the agents are helpful in reducing CV events, particularly HF: canagliflozin (Invokana, Janssen), which was evaluated in the CANVAS trial; dapagliflozin (Farxiga, AstraZeneca), evaluated in the DECLARE-TIMI 58 trial; and empagliflozin (Jardiance, Boehringer Ingelheim), evaluated in the EMPA-REG trial.

The similarities of these inhibitors include the dosing, most of the safety signals and drug-drug interactions. In addition, all three have been shown to reduce HF hospitalizations and/or new HF onset and have demonstrated a modest reduction in body weight. I must emphasize that there is no consistent improvement in outcomes related to atherosclerotic events, and there is no consistent reduction in stroke.

In terms of the differences between the inhibitors, empagliflozin had a mortality benefit in a fairly high-risk CV population. Neither dapagliflozin nor canagliflozin have shown an independent mortality benefit. A second difference is the safety signal. In the CANVAS trial, canagliflozin improved CV outcomes and reduced the risk for HF hospitalizations. But there was an increased risk for amputation, albeit in a relatively small group of patients with preexisting diabetic foot ulcers. This has not been observed with the other two SGLT2 inhibitors.

A third difference is driven by what population was tested with these agents. Empagliflozin was tested in the sickest population. In the DECLARE-TIMI 58 trial, dapagliflozin improved HF hospitalization rates in a lower risk population but did not improve mortality or MACE. It appears that the higher the risk of the patient, the more likely there will be a benefit.

Safety profile

In general, SGLT2 inhibitors are safe. The main adverse events patients experience are relatively minor bladder infections, which are caused by glucose that is released into the bladder and urine.

The biggest concern to date is the trend toward excess risk for amputation with canagliflozin. This could be a random effect, but it must be noted when considering safety.

Many SGLT2 trials enrolled or are enrolling patients with stage 3 chronic kidney disease, with glomerular filtration rate (GFRs) as low as 30 mL/min/1.73 m2. The FDA label will often urge caution in patients with renal disease. I would feel more comfortable prescribing SGLT2 inhibitors to patients with higher GFRs, but feel that in the right patients, we can use it in patients with Stage 3 CKD. This is something that needs to be further tested.

The last safety signal is the risk for dehydration. The true mechanism between the effects of the SGLT2 inhibitors on CV outcomes is unclear. One potential mechanism of action is the inhibitors’ ability to work as a diuretic because of the glucose that goes into the bladder and urine. Because of that, it is important to be careful with dosing. If a patient is on another diuretic, they should be on a stable dose before starting the SGLT2 inhibitor. There is a possibility the dose of the diuretic may need to be reduced.

SGLT2 inhibitors vs. other drug classes

No other drug classes for type 2 diabetes have consistently shown a CV benefit. The workhorse for type 2 diabetes has been metformin. It is relatively inexpensive and fairly safe. There was a concern for lactic acidosis, but we really do not see that often. I think the concern is greater than warranted. Still, there are no large randomized trials showing metformin can reduce CV events. The UKPDS study suggests there is some improvement, but that is clearly not as robust as the three major studies looking at SGLT2 inhibitors.


Sulfonylureas have been used consistently, with no indication of a CV signal. Similarly, there are no observed CV benefits with thiazolidinediones, DPP4 inhibitors and insulin. Early on, there was concern that thiazolidinediones increased the risk for HF. That concern prompted the 2008 FDA mandate that urged physicians to closely examine the safety profile of these agents. Non-randomized data from patients receiving insulin have shown that they actually have a higher CV risk. This could be driven, in part, by indication.

The majority of the GLP-1 receptor agonists have not shown a benefit in CV outcomes. Liraglutide is the one exception. In the LEADER study, there was an improvement among patients who were treated with liraglutide. There is no class effect, however, as several other studies evaluating GLP-1 receptor agonists have shown no CV benefit.

Future research on SGLT2 inhibitors

There is a need for more data on the mechanisms of action of SGLT2 inhibitors. In DECLARE-TIMI 58, BP was a little lower in patients randomly assigned to dapagliflozin. So, are some of the effects of SGLT2 inhibitors mediated through lower BP? I think the benefits are independent of BP, but we need more data to understand why these patients have improved outcomes.

Interestingly, in the DECLARE-TIMI 58 trial, the combination of CV death, MI and stroke was not improved with the SGLT2 inhibitor. Moving forward, we should think about the possibility of a paradigm shift from atherosclerotic risk factors to risk factors for HF. Patients with diabetes are at a very high risk for HF. Outside of death, one of the biggest CV risks for patients with diabetes is the development of HF. When looking at predictors of HF outside of hypertension and MI, diabetes is the third most important risk factor. We really should start to look at these SGLT2 inhibitors as a potential strategy for preventing HF in patients with diabetes.

Nevertheless, just because we see a HF signal with the SGLT2 inhibitors does not mean we should randomly use them in HF. I would be very cautious, because the number of patients with HF in the three big trials (EMPA-REG, CANVAS and DECLARE-TIMI 58) were limited and not well characterized. It is unknown whether these patients with a history of HF had reduced ejection fraction, preserved ejection fraction or midrange. For the most part, there was little information on HF medications and New York Heart Association class. The other reason for caution is due to previous research on the GLP-1 receptor agonist liraglutide. Although the LEADER study demonstrated an improvement in outcomes with the agent, the FIGHT study enrolled 300 recently hospitalized HF patients, and there was a trend toward an increased risk for subsequent HF hospitalizations among patients who were discharged and then randomized to liraglutide vs. placebo. So, we cannot take findings from patients who are at risk for a variety of conditions and apply it to everyone.

Finally, we need cost-effective analyses. When I was younger, I would think agnostically about cost. But in reality, we need to worry. There must be cost-effectiveness analyses examining the individual costs and costs to society for patients who are at low-risk vs. intermediate risk vs. high risk so that we can have the right drug for the right patient at the right time. If SGLT2 inhibitors prevent HF, and if patients are followed long enough, we will find cost-effectiveness. Given that HF has such a negative impact on quality of life. If HF can be prevented, it will be worth the price.


Wiviott SD, et al. LBS.02 – Late Breaking Clinical Trials: Novel Approaches to CV Prevention. Presented at: American Heart Association Scientific Sessions; Nov. 10-12, 2018; Chicago.

Disclosure: Lewis reports receiving institutional research support from Amgen, Merck, Novartis and Sanofi Pasteur, as well as consulting fees from Dal-Cor, Merck and Novartis.