Expert Blog

Who is the Right Patient for SGLT-2 Inhibitors?

Irl B. Hirsch, MD
Dr. Leahy’s comments are right-on target. Some further observations by another endocrinologist who can’t determine who the best patient for this new class of drugs will be:
1.     The advisory committee had problems with the bump in LDL-cholesterol. True, this is a surrogate, not a true “outcome” that will be determined in the CANVAS study. But do you recall the last time we had this debate with a diabetes drug? If you don’t recall, please Google “rosiglitazone”.
2.     How is the early reduction in renal function going to impact a more elderly, frail population who may be receiving non-steroidal anti-inflammatory agents or other nephrotoxic agents?
3.     While the vulvovaginitis has been “played down” in study presentations for all in the SGLT-2 class, how many of these would a patient have to have before deciding the side effect was not worth the modest benefit in glucose control? Similarly, what is the overall cost for a population using this drug (compared to other generic or branded agents) for treating all of the infectious adverse events? I will be interested to see, if available, how many hospitalizations for pyelonephritis occur once the drug is released.
4.     I’m thinking about how this drug will complicate our treating of  patients that have other conditions. Some of this may be addressed in the label of the drug when it is released. But how will use of this agent impact the patient with mild heart failure with heavy glycosuria? When the heart failure is exacerbated to the point more diuretics are required, what is the impact of the SGLT-2 class of agents? Or the patient truly dehydrated after running a marathon on a hot summer day, ending up in the emergency room with a serum sodium of 150 mEq/L? I don’t think that patient is a good match for this class of agents. As an endocrinologist, I also think of the hyperfiltration from glycosuria that can cause false positives in 24 hour urine collections for cortisol, or for that matter the 1,5 ahydroglucitol (1,5 AG) test, a new biomarker we are using for the assessment of glycemic variability. 1,5 AG is lost with glucose in the urine, and thus serum levels reflect the amount of time above the renal threshold of glucose, therefore the exposure to a hyperglycemic spike. This test can’t be used with someone receiving an SGLT-2 inhibitor.
5.     The ADA/EASD algorithm has moved to an individualized, or personalized prescription, since outside of general agreement for the use of metformin for first line therapy, we don’t have good data about what should follow. So for now, I will assume (perhaps incorrectly) the next version of this algorithm will include canagliflozin. But as we move further along into the Accountable Care Organization era where all of us will be graded on how efficiently we lower A1C (in case you can’t read between the lines, the only part of “efficiently” to be concerned about is cost), should we routinely use these more costly drugs without more data? While there may be excellent patients who would be good candidates for any agent for type 2 diabetes, I’m having difficulty understanding how this would be good for the majority of patients without more safety data.
6.     While Dr. Leahy speaks about how both primary care and endocrinologists are becoming confused with all of the choices, think about our students, residents, and fellows! At my institution, due to many factors (including cost), I can’t recall even one prescription of one of the thiazolidinediones coming from our primary care providers. Wasn’t even on their radar. Perhaps that will now change with pioglitazone losing its patent protection. But I also don’t see GLP-1 agonists used from my academic colleagues, and rarely do I see a DPP-4 inhibitor used. The reason is simple. Most of the patients seen in training programs (county hospitals, VA hospitals, and in general, a more disadvantaged population) can’t afford the newer, more expensive agents. The students eventually end up in practice, and they have no experience with the new drugs.
As I reflect on the FDA Advisory Committee’s vote this week, I wonder what the final label will say, how the drug will be marketed, and what type of uptake we will see in the US. I may be surprised, but with all of our choices, confusion, and disagreement, I would be surprised if we will use a lot of this drug 5 years from now. I would like to be proven wrong as my greatest hope is we have a new drug that is truly uniquely beneficial showing improvements in the cardiovascular complications of diabetes. Five years from now, let’s look at this blog and review what happened.


Add new comment