Don't Google It: Fournier's Gangrene and SGLT-2 Inhibitors

An analysis of an FDA database suggests that the novel SGLT-2 inhibitors may significant increase the risk of the life-threatening genital infection.

A new class of medication – the SGLT-2 inhibitors have been killing it lately, with study after study in the New England Journal documenting their benefits in patients with diabetes. Many of us feel that these drug will become a central part of diabetic therapy in the near future. The twitterverse is, well, all atwitter.

but this publication in the Annals of Internal Medicine may take a bit of bloom off that rose, as it links SGLT-2 Inhibitors with Fournier’s Gangrene.


For the medical folks watching this, you know what Fournier’s gangrene is.  For the non-medical people – a word of advice – do NOT google it. What we are talking about is the life-threatening necrosis and fasciitis of the perineal region – a rare condition representing around 0.02% of all inpatient hospitalizations in the US.

The primary data source for the analysis was the FDAs adverse events reporting system.

This is not a perfect data source. Reporting is voluntary, meaning we’re not getting every case of Fournier’s. Reports are also unverified. But most importantly, the database doesn’t have a denominator. For example, the authors report that from 2013 to 2019 there were 55 cases of Fournier’s gangrene among patients taking SGLT-2 inhibitors.  But we have no idea how many patients were taking the drugs total.

Nevertheless, inferences can be made. The researchers also searched for cases of Fournier’s Gangrene reported in patients taking other diabetes medications.  They found 19 cases. Over 35 years.

empagliflozin events.JPG

SGLT-2 Inhibitors – 55 cases in 6 years. All the other diabetes drugs 19 cases in 35 years. That seems real.

The FDA’s adverse event database is open access- you can do a lot of cool stuff with it.  Pick a drug, and the database can show you how much more often a given complication appears for that drug compared to the rest of the drugs in the database – a metric called the proportional reporting ratio or PRR. Here’s that list for empagliflozin, the most popular SGLT-2 inhibitor in the US, showing that DKA is reported 1200 times more often in patients on empagliflozin than the background rate.


You can go backwards too. I put “gangrene” in the dataset, and obtained this list of drugs, again ordered by how much more often gangrene appears as an adverse event compared to the background rate.

Canagliflozin has the top spot here with a gangrene rate about 100-fold above background, followed, interestingly, by insulin at around 8-fold. 

But don’t get too wrapped up in those numbers. There might be a self-fulfilling prophecy here. Since reports are voluntary, physicians, primed to worry about gangrene based on a recent focus within this drug class may be more likely to report this as an adverse drug event as opposed to attributing it to the patient’s underlying disease – and not reporting at all.

On the other hand, there is strong biologic plausibility for a causal link between SGLT-2 inhibitors and Fournier’s gangrene.

 The SGLT-2 inhibitors work by decreasing glucose reabsorption in the kidney, making patients excrete glucose in the urine. That sugary urine is no doubt an excellent culture medium.

In short, this is compelling evidence that these agents increase the risk of Fournier’s gangrene. Not airtight evidence, not perfect evidence, but compelling none-the-less. So… does it matter?

Remember that Fournier’s gangrene is vanishingly rare. I couldn’t find any cases of Fournier’s Gangrene in the landmark EMPA-Reg, CANVAS, or CREDENCE trials of these drugs, with total enrollment of 21,563 patients. Given the overall efficacy of the class, the chance is much higher that your patient will benefit from the drug than be harmed by it. But still, the risk of harm is not 0. What to do? Since early intervention is key to the successful treatment of Fournier’s gangrene, vigilance for the complication, education to the patients regarding early symptoms, and prompt referral for treatment if symptoms occur should be standard operating procedure as this drug class goes more mainstream.

This commentary first appeared on