Summary: In 13 adults with longstanding type 1 diabetes, seven days of dapagliflozin did not significantly alter the primary GLP-1 response compared with placebo. Ketone concentrations increased in both experimental clamp settings. No diabetic ketoacidosis occurred, but the small, short study cannot establish safety for routine adjunctive use.
PICO Summary
| Element | Detail |
|---|---|
| Population | 13 adults with type 1 diabetes for more than five years, C-peptide below 0.1 nmol/L and BMI 20-29 kg/m squared, recruited at a Swiss centre. All completed both treatment periods and were analysed per protocol. |
| Intervention | Dapagliflozin 10 mg daily added to insulin for seven days, followed by metabolic testing. This was an experimental study regimen, not a recommendation for type 1 diabetes treatment. |
| Comparator | Placebo for seven days in the alternate crossover period, separated by a 14-day washout. Participants and investigators were not blinded. |
| Outcomes | Primary endpoint: GLP-1 during the oral glucose tolerance test clamp. Median concentrations were 192.8 versus 176.3 pmol/L (P=0.7). GLP-1, glucagon and somatostatin did not differ significantly in the tested settings. Secondary total ketone concentrations were higher with dapagliflozin: median 0.15 versus 0.03 mmol/L during the hyperinsulinaemic-euglycaemic clamp, and 0.10 versus 0.03 mmol/L during the oral glucose tolerance test clamp; both P<0.001. Ketone-related adverse events were mild and asymptomatic; no severe adverse events or diabetic ketoacidosis were observed. The trial was not powered to quantify ketoacidosis risk. |
| Study Design | Randomised, open-label, placebo-controlled mechanistic crossover trial with two seven-day periods. NCT04035031. Funded by the Swiss National Science Foundation; the authors reported no relevant conflicts. Biochemical outcomes are not evidence of long-term clinical benefit. |
Expert Commentary
I regard this study as a mechanistic warning rather than evidence supporting routine dapagliflozin use in type 1 diabetes. The prespecified GLP-1 hypothesis was not supported, while the rise in ketone concentrations was consistent across both experimental settings. Those findings should not be combined into a claim that the drug improves hormonal regulation. Can I use this with my patients? The results may help explain why a lower glucose concentration does not necessarily imply metabolic safety during SGLT2 inhibition. They do not establish that adding this treatment is appropriate, nor do they provide an individual insulin-adjustment protocol. The principal limitation is the combination of thirteen selected participants and very short exposure. Absence of ketoacidosis in that setting cannot exclude an uncommon but serious clinical hazard. Clamp experiments also impose conditions that differ substantially from illness, fasting or disrupted insulin delivery in everyday life. I would therefore preserve a clear distinction between the observed biochemical increase and the unmeasured absolute risk of ketoacidosis. The open-label design and per-protocol analysis further constrain interpretation, although laboratory measurements are less subjective than symptom reporting. Future studies should test clinically meaningful safety questions in appropriately monitored populations; the present experiment should not be used as reassurance that routine adjunctive treatment is safe.
References
Gübeli A, Steiner N, Limacher A, et al. Dapagliflozin's impact on hormonal regulation and ketogenesis in type 1 diabetes: a randomised controlled crossover trial. Diabetologia. 2025;68(10):2116-2125. doi:10.1007/s00125-025-06481-9. PMID: 40629004.
