Jason Baker, MD

Expert Perspectives

SGLT2 Inhibitors in the Treatment of Type 2 Diabetes

Practice Community
New York, NY
Hospital and Institutional Affiliations
Attending endocrinologist and assistant professor of clinical medicine at Weill Cornell Medicine in New York City
Practice Niche


How do you approach risk-benefit discussions regarding sodium-glucose cotransporter2 (SGLT2) inhibitor use in patients who have type 2 diabetes?


A lot of interesting data have come out about a class of medications — the SGLT2 inhibitors — that have made them much more attractive for use in multiple patient populations, specifically patients who have type 2 diabetes and cardiovascular disease (CVD) or who we consider to be at extraordinarily high risk for CVD. Recent well-designed trials, which include EMPA-REG and CANVAS, demonstrate that there seems to be a CV benefit from using SGLT2 inhibitors in patients with type 2 diabetes.

So when I talk about treating patients with diabetes who need glucose-lowering medications, I think it’s best if we continue to use the traditional medications, which also promote weight loss. When you look at the guidelines from the American Diabetes Association (ADA) and the American Association of Clinical Endocrinologists (AACE), lifestyle and exercise rightfully remain the biggest factors to be discussed first in patients in order to deal with their diabetes and lower their CV risk. I think we would all agree on that.

But when it comes to the need to add medications to these patients’ regimens, the addition of something like metformin first, which is obviously the cheapest and one of the older medications — “tried and true,” if you will — is very helpful for patients. But once you’ve maximized the metformin dose and you’re wondering, “What can we do next?” the conversation I have with patients is: “We can add an SGLT2 inhibitor, which we know will lower your glucose levels. It will also probably help you lose some weight and result in some lowering of your blood pressure.” And now we have evidence that supports also saying, “…and probably it will lower your risk for CVD, mortality related to CVD, and possibly heart failure leading to hospitalization.” These are exciting claims.

With every medication we also have to discuss the potential adverse effects. So, when we talk to patients who have diabetes and are considering medications such as SGLT2 inhibitors, we need to emphasize these benefits but also delineate all the potential risks that may be associated with them.

I think it is always important to tell people that they must stay well hydrated while they are on SGLT2 inhibitors as it is fairly easy for a patient to become dehydrated because of the mechanism of action of these agents. If a patient has symptoms of dehydration, we need to remember that if they are on antihypertensive agents, particularly diuretics, we may need to decrease the SGLT2 inhibitor dose or perhaps adjust their other medications accordingly.


What SGLT2 inhibitor treatment initiation strategies have you employed in the past, particularly with regard to titrating baseline diabetes medication?


Another thing we need to discuss with patients is the risk for hypoglycemia. If you are using an SGLT2 inhibitor without concurrently giving a patient a sulfonylurea or insulin, the risk is much lower. However, when we start an SGLT2 inhibitor, we must keep in mind that it is a powerful medication that starts to work within just a few hours of administration. Therefore, we often have to decrease the insulin and/or sulfonylurea dose as we titrate an SGLT2 inhibitor. We have to have a discussion with patients about how to safely start an SGLT2 inhibitor.

We have a lot of data on diabetic ketoacidosis and the risk for diabetic ketoacidosis in some patients. Although this is a rare phenomenon, particularly in patients with type 2 diabetes, when you are dealing with a patient with type 2 diabetes who may already be rather insulinopenic — not producing enough insulin — who requires higher doses of insulin to maintain blood sugar levels, there is some as yet unknown mechanism that causes some people to go more easily into diabetic ketoacidosis, even people with lower-than-expected glucose levels. So, we have to have the same discussion with our patients with type 2 diabetes that we did previously with patients with type 1 diabetes: What are the symptoms of diabetic ketoacidosis, and what are the situations in which this class of medications should be held back to ensure that we can use them safely?

I like to tell patients that if you get sick, if you get dehydrated for any reason (eg, gastroenteritis; you are vomiting from food poisoning or something of that nature and are just not feeling right), it’s prudent to hold SGLT2 inhibitors because they can ultimately worsen dehydration and flip them into diabetic ketoacidosis. This can obviously be dangerous and sometimes leads to hospitalization and in some settings, death. So again the conversation I have with patients when it comes to starting an SGLT2 inhibitor includes emphasizing the potential benefits but also all the potential adverse events.


What are the long-term benefits of SGLT2 inhibitor treatment you discuss with your patients when considering SGLT2 inhibitor therapy?


We know that patients with type 2 diabetes — just because of the nature of type 2 diabetes and the comorbidities that go along with it such as high blood pressure, high cholesterol, or a family history of CVD — are already at fairly high risk. So for the first time ever, a study 18 months ago — the EMPA-REG study — showed that empagliflozin, one of the SGLT2 inhibitors, helped prevent mortality related to CVD and hospitalization related to heart failure. That was again, as I mentioned, in a group of people who already had established CVD. So when that study was completed it still left us with many unanswered questions, including “What about people who we think are probably at high risk but don’t yet have true established CVD? They’ve never had a stroke. They’ve never had a heart attack. They have no diagnosis of peripheral vascular disease. Do those people garner protection from those events, from mortality related to those events, and also hospitalization-related heart failure from SGLT2 inhibitors?” The CANVAS trial helped us to better understand that. CANVAS looked at a cohort of people with and without established CVD who had type 2 diabetes, and what it showed was a similar type of outcome as the EMPA-REG trial: decreased mortality related to CVD and decreased incidence of hospitalization related to heart failure.

So we have 2 well-designed trials that really gave us a wealth of information that says that these medications are, it seems, protective from a CV standpoint. Again, we are dealing with a very high-risk patient population. So, if they can safely be started on a medication like an SGLT2 inhibitor, and there is an indication for this because their glycemic control is not good enough, or if you’re interested in transitioning them away from some of their insulin or some or all of their sulfonylurea, then SGLT2 inhibitors are an ideal choice for both a glycemic and a CV benefit.

The CANVAS and EMPA-REG trials did demonstrate decreased hospitalization related to heart failure in patients with type 2 diabetes both with and without known CVD. When it comes to primary prevention of heart failure, however, because of the risk-to-benefit profile of SGLT2 inhibitors, I’m not sure we have enough data yet to say that we should be prescribing them in patients who do not have an indication for improved glycemic control simply to prevent or treat heart failure or CVD. I think we need to go back to the basics. Can we improve lifestyle and diet? If we need help with glycemic control and we have a patient in whom the potential risks associated with SGL2 inhibitors are counterweighed by the glycemic benefits and we are concerned about CVD and the risk of heart failure, then this could be a good class of agents to add to their regimen. But, when you look at any medication and the relevant clinical trials, we have an increasing amount of data that say that we think these medications could be helpful in this population. However, I don’t think we are quite ready to prescribe SGLT2 inhibitors for primary prevention unless there is another indication such as worsening glycemic control that requires the addition of another agent to the medication regimen.

The EMPA-REG and CANVAS studies were designed to look at CV outcomes related to the use of the SGLT2 inhibitors empagliflozin and canagliflozin. When you look at the data, there were some ethnic differences in the studies regarding who seemed to benefit more vs who seemed to benefit less in terms of CV protection and decreased hospitalization related to heart failure, but the studies were not designed to look at specific ethnic groups. So all we can infer from the results is that there seems to be a trend — and not always a statistically significant trend — when it comes to the EMPA-REG trial, for an increased benefit in white and Asian subjects with regard to CVD than black subjects. On the other hand, in the CANVAS trial Asian subjects appeared to benefit a little less than black and white subjects. It was interesting that there did seem to be some differences between ethnic groups in their response to these medications.

But again, I emphasize that the studies were not powered to truly show an ethnic difference in the response to SGLT2 inhibitors, and the numbers of black and Asian participants were smaller than the number of white participants. In fact, the percentage of black subjects in many of these trials is less than 5%. So when looking at trends in study data on SGLT2 inhibitors, there needs to be wider recognition of the study results and we need studies that are appropriately designed.

We do know that there are many medications that elicit different responses in different ethnic groups. There may be SGLT2 receptor and transporter density and location that may vary from one ethnic group to another, which might potentially account for some of these trends and ethnic differences. But the jury is still out as to what this really means and how it applies on a broader scale outside of these trials.


Canagliflozin treatment has demonstrated potential renal benefits in certain patients with type 2 diabetes, as shown by the lower incidence of albuminuria progression in the CANVAS trial. Can you review the role of medication as part of a renoprotective strategy in type 2 diabetes?


We know that SGLT2 inhibitors have the potential for improving renal function. We are always afraid of using these medications; there are contraindications depending upon the SGLT2 inhibitor to using them in people with a glomerular filtration rate (GFR) that is decreased. We have to be mindful that the GFR and renal function can decline in patients who use SGLT2 inhibitors, and the mechanisms are multi-fold. This may be a result of pre-renal dehydration, and it is something we have to discuss with our patients. But the data out there have also shown that, although again trials were not necessarily designed specifically to look at renal outcomes, there appears to be a decline in the amount of albuminuria seen in patients treated with SGLT2 inhibitors.

Clinically, we have been happily surprised by that result. The mechanism for this improvement in albuminuria in patients who have type 2 diabetes who are treated with SGLT2 inhibitors is thought to be related to a feedback loop that decreases hyperfiltration at the level of the glomerulus as it relates to the renin-angiotensin system. So, it seems that using SGLT2 inhibitors in people with hyperfiltration leading to albuminuria decreases the feedback loop that harnesses the renin-angiotensin system, leading to decreased filtration pressures in the glomerular units and a resulting decrease in albuminuria. We need more studies to better analyze these mechanisms, but I think it is exciting to think that we have yet another potential benefit from the SGLT2 inhibitors. However, I wouldn’t recommend starting them in a primary prevention scenario unless there is another indication present, such as the need for improving glycemic control.


How do you approach conversations with your patients regarding the risk for diabetic ketoacidosis when being treated with SGLT2 inhibitors?


There have been multiple case reports, both in studies and since the SGLT2 inhibitors have been more widely used clinically, of diabetic ketoacidosis in patients with much lower glucose levels than expected. This has been most commonly seen in patients with type 1 diabetes who were prescribed SGLT2 inhibitors, either in a study or off-label in the community. However, we also see it in patients with type 2 diabetes who were treated with SGLT2 inhibitors, albeit less commonly than what we see in patients with type 1 diabetes. So in which patients are we more likely to see this? A patient with a decreased amount of insulin reserve — and frankly all patients who present with diabetes have this to a greater or lesser degree — who has progressed to needing insulin (perhaps they are receiving basal bolus insulin for type 2 diabetes or have an increased decline in beta-cell function and therefore require more insulin). We are also likely to see this in patients who have decreased beta-cell function reserve and thus decreased insulin secretory capacity and are faced with another comorbidity such as infection or dehydration from another cause.