Summary: In adults with type 1 diabetes who switched to inhaled technosphere insulin (TI) plus once-daily insulin degludec, this 13-week single-arm extension of the INHALE-3 trial showed HbA1c was sustained out to 30 weeks (7.6% at baseline, 7.6% at 17 weeks, 7.4% at 30 weeks; change from 17 to 30 weeks -0.21%, 95% CI -0.33 to -0.09, p<0.001), with time in range essentially flat at 52% to 54%. This was a durability follow-up, not a head-to-head comparison against subcutaneous insulin.
PICO Summary
| Element | Detail |
|---|---|
| Population | Adults with type 1 diabetes who had completed the 17-week randomized phase of the INHALE-3 trial; 45 of the 62 originally assigned to TI plus degludec entered this multicentre US extension. At baseline (pre-trial) 44% were using automated insulin delivery, 9% a sensor-augmented pump, and 47% multiple daily injections. |
| Intervention | Inhaled technosphere insulin (TI, Afrezza) at meals plus once-daily basal insulin degludec, continued for a further 13 weeks (to 30 weeks total) with no scheduled visits before the final visit, to approximate real-world care. Mean total daily TI dose at 30 weeks was 53 +/- 31 U/day, roughly double the pre-trial rapid-acting analog dose of 24 +/- 12 U/day. |
| Comparison | None concurrent. The extension was single-arm; the comparison was within-person across timepoints (17 weeks versus 30 weeks, and versus pre-trial baseline). The earlier 17-week randomized phase had compared TI plus degludec against usual care, but that randomized comparator did not continue into the extension. |
| Outcome | HbA1c was sustained: 7.6% +/- 1.0% at baseline, 7.6% +/- 1.0% at 17 weeks, and 7.4% +/- 1.0% at 30 weeks; mean change from 17 to 30 weeks was -0.21% (95% CI -0.33 to -0.09, p<0.001). The proportion with HbA1c <7.0% rose from 21% (baseline) to 30% (17 weeks) to 42% (30 weeks). Time in range 70-180 mg/dL was 52% +/- 18% at baseline, 53% +/- 20% at 17 weeks, and 54% +/- 20% at 30 weeks. Time <54 mg/dL was 0.4% +/- 0.6%, 0.4% +/- 0.8%, and 0.6% +/- 1.0%, respectively. No NNT or ARR applies to this single-arm durability outcome. |
Expert Commentary
The honest read is that this is a durability signal, not an efficacy verdict. The extension was designed to answer a narrow question, namely whether the glycaemic control achieved with inhaled technosphere insulin plus degludec in the randomized phase was maintained once intensive study contact was removed, and it was: HbA1c held at around 7.4% to 7.6% and the small statistically significant fall from 17 to 30 weeks (-0.21%) is clinically modest, with time in range stuck near 53%. The chief limitation is that there is no concurrent comparator in this phase, so nothing here demonstrates that inhaled insulin is better than automated insulin delivery or multiple daily injections; the apparent gain in the proportion below 7.0% is uncontrolled and vulnerable to selection, because only the participants who chose to continue entered the extension. That self-selection, plus the doubling of mealtime insulin dose needed, should temper enthusiasm. Can I use this with my patients? Cautiously yes, for a specific type: the motivated adult with type 1 diabetes who dislikes mealtime injections or pump wear and accepts inhaled dosing, rather than as a routine swap from a working AID system. Worth flagging that the manufacturer makes the inhaled product, and that lung function and cough monitoring sit outside this report. I would like to see a longer randomized trial with patient-reported and pulmonary endpoints before positioning this as anything more than an option.
References
Beck RW, Bailey RJ, Klein KR, Aleppo G, Levy CJ, Diner J, et al. Inhaled Technosphere Insulin Plus Insulin Degludec for Adults with Type 1 Diabetes: The INHALE-3 Extension Study. Diabetes Technol Ther. 2025;27(3):170-178. doi:10.1089/dia.2024.0582
