The stakes are high because the link between hyperinsulinemia and laminitis is now firmly established. In research settings, experimentally induced high insulin concentrations produce structural and metabolic changes in the lamellae of the hoof that worsen with time and can progress to clinical laminitis in as few as three days. Horses that require corticosteroids for conditions such as purpura hemorrhagica or immune-mediated myopathy are particularly concerning if they also carry underlying metabolic risk, such as equine metabolic syndrome. Until now, veterinarians have had little in the way of adjunctive therapy to protect these animals; the new findings suggest a pharmacological strategy that could fill that gap.

The study, conducted at the University of California, Davis, enrolled seven healthy research horses free of endocrine disease, with a mean age of 12.9 years. The investigators used a randomized crossover design spanning 11 weeks in the summer. In each of two 15-day blocks, horses received either dexamethasone sodium phosphate at 0.06 mg/kg intravenously every 48 hours, or the same dexamethasone regimen plus oral ertugliflozin at 0.05 mg/kg once daily. After a seven-week washout, each horse crossed over to the opposite treatment, allowing every animal to serve as its own control.

During the seven treatment days, blood was drawn at three time points daily, at 0600 hours in a fasted state, at 1000 hours post-prandially, and at 1800 hours in a fed state, to capture insulin and glucose concentrations across different prandial states and circadian influences. The primary analysis used mixed-effects linear regression models. Mean daily insulin concentrations were markedly lower in the combination group: 35.9 microIU/mL versus 50.6 in block 1, and 18.9 versus 34.5 in block 2, a highly significant difference (P < .001). Glucose concentrations were also lower in the ertugliflozin group, significantly so on days 3 and 5, the days dexamethasone was administered, with means of 101.0 mg/dL versus roughly 117 mg/dL in the steroid-only group.

An intriguing pattern emerged in the glucose data: the difference between treatment groups was significant precisely on the days dexamethasone was given, and nearly so on the following day. This suggests the SGLT-2 inhibitor acted to dampen the exaggerated glucose response that each steroid dose provoked. The authors note that ertugliflozin’s effect on glucose was of lower magnitude and less consistent than its effect on insulin, which they interpret as clinically advantageous, since a robust glucose-lowering effect could risk hypoglycemia, while the dominant insulin-lowering effect addresses the actual therapeutic goal.

To confirm that the dexamethasone regimen truly induced insulin dysregulation, the team performed oral sugar tests and insulin tolerance tests before and after each treatment period. The oral sugar test, which involves administering corn syrup by mouth and measuring the insulin response 60 minutes later, told a sobering story: five of the seven horses in the dexamethasone-only group and four in the combination group crossed the diagnostic threshold of 65 microIU/mL after treatment, despite having normal responses beforehand. In other words, even this relatively conservative steroid dose converted most healthy horses into insulin-dysregulated animals on dynamic testing, reinforcing the authors’ recommendation that serum insulin be monitored serially in any horse undergoing corticosteroid therapy.

However, the dynamic tests also delivered the study’s key caveat. After drug discontinuation, there was no difference between the two groups in oral sugar test results or in the glucose area under the curve, indicating that ertugliflozin co-administration did not measurably improve the underlying insulin dysregulation once the drugs were stopped. The authors caution that this may reflect a design limitation rather than a true failure: dynamic testing began 24 hours after the last ertugliflozin dose, and the drug’s renal glucose-excreting effect wanes within 24 hours in humans, whereas dexamethasone’s endocrine effects can persist for two to three days. Studies continuing both drugs through the post-treatment testing window are needed to resolve the question.

More surprisingly, the insulin tolerance test revealed that horses in the ertugliflozin group had a higher residual percentage of glucose 30 minutes after insulin administration after treatment, a pattern consistent with transient peripheral insulin resistance. The authors propose several non-mutually exclusive explanations rooted in human pharmacology. SGLT-2 inhibitors in people increase plasma glucagon and endogenous glucose production through hepatic gluconeogenesis, and recent work has located SGLT-2 receptors in equine pancreatic islets where glucagon-secreting alpha-cells reside. Renal nerve signaling that stimulates hepatic glucose production independently of glucagon is another suspected pathway. The relatively small insulin dose used in the test, and the 72-hour gap between ertugliflozin discontinuation and testing, may have further limited the assay’s ability to suppress glucose production in these animals.

The trial also documented metabolic side effects that veterinarians should weigh. Horses receiving ertugliflozin lost significantly more weight than those on dexamethasone alone, a mean of 40.1 kg versus 21.5 kg, consistent with the fat-mass reduction seen in human patients, and all horses regained weight during the washout. Triglyceride concentrations rose, though the increase remained subclinical and all horses maintained good appetite, and modest increases in the liver enzymes sorbitol dehydrogenase and gamma glutamyl transferase were observed, of unclear clinical significance. No adverse clinical signs were attributed directly to either drug, and all horses completed the study, though one horse developed transient seizure-like activity during a pre-treatment insulin tolerance test and recovered promptly with oral syrup and intravenous dextrose.

One unplanned lesson concerns diet. Insulin concentrations were higher in the first study block than the second, and hay analysis offers a plausible explanation: the hay fed during block 1 contained 22.4 to 26.7 percent nonstructural carbohydrates as sampled, versus only 10.5 percent in block 2. Even in non-endocrinopathic horses, dietary nonstructural carbohydrates clearly modulated serum insulin during corticosteroid treatment, a finding with practical implications for feeding horses on steroid therapy. The authors acknowledge the small sample size and the use of metabolically healthy horses as limitations, and they call for trials in animals with naturally occurring metabolic disease or systemic illness, where the true clinical benefit of pairing SGLT-2 inhibitors with corticosteroids will ultimately be decided.

Subject of Research: The effect of the SGLT-2 inhibitor ertugliflozin on corticosteroid-induced insulin dysregulation in healthy horses

Article Title: The effect of ertugliflozin on serum insulin and glucose in healthy horses receiving dexamethasone

Article References: Dupuis-Dowd, F., Kass, P. H., & Berryhill, E. H. (2026). The effect of ertugliflozin on serum insulin and glucose in healthy horses receiving dexamethasone. Journal of Veterinary Internal Medicine, 40(5), Article aalag204. https://doi.org/10.1093/jvimsj/aalag204

Image Credits: AI Generated

DOI: 10.1093/jvimsj/aalag204

Keywords: ertugliflozin, SGLT-2 inhibitors, dexamethasone, horses, insulin dysregulation, hyperinsulinemia, laminitis, glucose, equine metabolic syndrome, veterinary medicine, insulin tolerance test, oral sugar test