The Threat That Never Fades: Ketoacidosis on SGLT2s

Key Takeaway: A landmark Scandinavian study of more than 282,000 patients with type 2 diabetes reveals that the risk of ketoacidosis from popular SGLT2 inhibitors like Jardiance, Farxiga, and Invokana does not diminish over time. Patients with poor blood sugar control, low body weight, or a history of ketoacidosis face a dramatically elevated danger, and infections are the single most common trigger.

A Risk That Refuses to Retreat

Imagine a patient who has been faithfully taking their diabetes medication for two years. They come down with the flu, continue their medications as usual, and within 48 hours, find themselves in the emergency room—their blood dangerously acidic, their consciousness clouded, their breathing rapid. Most clinicians had assumed this type of crisis, known as ketoacidosis, was primarily a concern in the initial weeks after starting an SGLT2 inhibitor. A massive new study from Scandinavia, covering more than 282,000 patients, reveals a much different and more alarming picture: the threat doesn’t disappear. It persists for as long as you take the drug, and for certain patients, the risk remains alarmingly high.

What the Researchers Found

This population-based cohort study leveraged the national health registries of three Scandinavian countries to analyze real-world data from 282,282 adults with type 2 diabetes who were prescribed SGLT2 inhibitors. The overall incidence of ketoacidosis was found to be 2.43 per 1,000 person-years, meaning that for every thousand patients taking the drug for a year, roughly two to three experienced this potentially life-threatening complication^[1]. While that rate may seem small as a percentage, when millions of people worldwide are using these medications, the absolute numbers become alarming.

The risk was highest shortly after patients started the drug, but—and this is the most critical finding—it did not decrease over time. It continued throughout the follow-up period, challenging the common clinical assumption that early vigilance is sufficient.

Perhaps most striking were the risk factors. A high HbA1c, the blood test reflecting average blood sugar over the last three months, had an odds ratio of 15.37, meaning patients with poor glucose control were more than fifteen times as likely to develop ketoacidosis. Malnutrition (odds ratio 10.54), a prior episode of ketoacidosis (odds ratio 10.40), and low body mass index (odds ratio 9.98) completed this quartet of powerful predictors. Infection was the most common precipitating event, identified in 31.8% of all ketoacidosis cases. And in a telling detail about the aftermath, insulin use among affected patients, which stood at 31.9% before the event, skyrocketed to 73.0% one year later. This dramatic shift suggests that many of these individuals harbored a previously unrecognized insulin deficiency.

Why Do SGLT2 Inhibitors Create This Vulnerability?

To understand why these drugs carry this particular risk, one must know what they do at the level of the kidney. SGLT2 inhibitors block a protein called the sodium-glucose cotransporter 2, located in the proximal tubule of the kidney, which normally reabsorbs glucose from the urine back into the bloodstream^[2]. By blocking this transporter, the drugs cause excess glucose to be excreted in the urine, lowering blood sugar without needing insulin.

But this glucose-lowering effect has a metabolic consequence. When the body senses reduced glucose availability, it shifts to burning fat for fuel. The liver accelerates the production of ketone bodies, which are acidic molecules that serve as an alternative energy source. In most patients, this shift is modest and well-tolerated. But in someone whose insulin reserves are already limited—whether from long-standing type 2 diabetes, undiagnosed latent autoimmune diabetes of adults (LADA), or the metabolic stress of an acute illness—the brake on ketone production is released^[3]. Ketones accumulate, the blood becomes dangerously acidic, and the result is diabetic ketoacidosis.

What makes SGLT2 inhibitor-associated ketoacidosis particularly insidious is that blood sugar levels may be only mildly elevated or even normal—a phenomenon called “euglycemic ketoacidosis”^[4]. This can delay diagnosis, as both patients and clinicians are conditioned to associate ketoacidosis with very high blood sugar readings. The patient may feel nauseous and breathe rapidly, attributing it to a stomach bug and losing precious hours before seeking medical attention.

The finding from the Scandinavian data that infection was the most common trigger aligns perfectly with this mechanism. During an acute infection, stress hormones like cortisol, catecholamines, and glucagon surge, pushing the liver to produce more ketones while simultaneously increasing insulin resistance^[5]. When an SGLT2 inhibitor is added to this metabolic storm, the stage is set for ketoacidosis in a susceptible patient.

What This Means for Patients and Clinicians

The practical implications of this study are immediate and concrete. First, the data paint a clear portrait of the high-risk patient: someone with a high HbA1c, low body weight, poor nutritional status, or a prior history of ketoacidosis. The decision to prescribe an SGLT2 inhibitor for these individuals requires careful consideration and, if the drug is used, intensified monitoring.

Second, the persistence of the risk over time means that patient education cannot be a one-time conversation at the pharmacy. Every patient on an SGLT2 inhibitor needs to understand “sick day rules”—the practice of temporarily stopping the medication during any illness that causes fever, vomiting, diarrhea, or reduced food intake, and seeking urgent medical attention. The patient’s relatives should also keep these rules in mind. These actions may be the most effective way to prevent a catastrophic event.

Third, the dramatic spike in post-event insulin use—from roughly one-third to nearly three-quarters of affected patients—raises a significant diagnostic question. It is possible that a subset of these patients were misclassified as having type 2 diabetes when they actually had LADA or another insulin-deficient form of diabetes^[3]. Checking for autoantibodies like GAD65 before prescribing an SGLT2 inhibitor, particularly in leaner patients, may identify those at highest risk before a crisis occurs.

Notable Limitations

No single study, no matter how large, settles a question completely. This was an observational study based on registry data, meaning some cases of ketoacidosis may have been miscoded or missed, and unmeasured confounding factors like alcohol use or dietary habits could influence the results. The Scandinavian population is also relatively homogeneous, and the findings may not translate one-to-one to more diverse populations. Still, the sheer size of the cohort and the consistency of the risk factors lend significant weight to this study.

SGLT2 inhibitors remain truly transformative drugs; they protect the heart, preserve kidney function, and lower blood sugar with a low risk of hypoglycemia. But this study is a sharp reminder that their safety profile commands respect. The risk of ketoacidosis is real, it is persistent, and it is preventable—but only if patients and their physicians remain vigilant long after the first prescription is written.

Scientific Sources

1. Kadesjö E, et al. Ketoacidosis with SGLT2 inhibitors in routine clinical practice of type 2 diabetes: Scandinavian cohort and nested case-control study. Lancet Diabetes Endocrinol. 2026. PubMed: https://pubmed.ncbi.nlm.nih.gov/42679839/
2. Vallon V, et al. SGLT2 inhibitor empagliflozin reduces renal growth and albuminuria in proportion to hyperglycemia and prevents glomerular hyperfiltration in diabetic Akita mice. Am J Physiol Renal Physiol. 2014. DOI: 10.1152/ajprenal.00520.2013
3. Buzzetti R, et al. Management of Latent Autoimmune Diabetes in Adults: A Consensus Statement From an International Expert Panel. Diabetes. 2020. DOI: 10.2337/dbi20-0017
4. Rosenstock J, et al. Euglycemic diabetic ketoacidosis: a predictable, detectable, and preventable safety concern with SGLT2 inhibitors. Diabetes Care. 2015. DOI: 10.2337/dc15-1380
5. Marik PE, et al. Stress hyperglycemia: an essential survival response. Crit Care Med. 2013. DOI: 10.1097/CCM.0b013e318283d124