Abstract
OBJECTIVE
To evaluate inhaled technosphere insulin (TI) in children with diabetes.
RESEARCH DESIGN AND METHODS
A total of 230 youth 4–17 years old with type 1 (98%) or type 2 (2%) diabetes treated with multiple daily injections of insulin were randomly assigned 1:1 to TI or rapid-acting analog (RAA) insulin plus continuation of long-acting basal insulin and continuous glucose monitoring (CGM) for 26 weeks. The primary outcome was change in HbA1c, tested for noninferiority with margin of 0.4%.
RESULTS
In intent-to-treat analysis, mean HbA1c (% ± SD) was 8.22 ± 0.87 at baseline and 8.41 ± 1.38 at 26 weeks with TI and 8.21 ± 0.96 and 8.21 ± 1.10, respectively, with RAA (adjusted difference = 0.18; 95% CI −0.07, 0.43; noninferiority P = 0.091). CGM-measured time in range 70–180 mg/dL was not significantly different between groups (adjusted difference −2.2%; 95% CI −7.0, 2.7; P = 0.38). Two severe hypoglycemic events occurred in the TI group and one in the RAA group. Change in forced expiration volume in 1 s from baseline to 26 weeks did not differ comparing TI and RAA (P = 0.53). The TI group reported greater treatment satisfaction (P = 0.004) and had less gain in weight and BMI percentile (P = 0.009) than did the RAA group.
CONCLUSIONS
The primary analysis did not meet the prespecified criteria for HbA1c noninferiority. However, TI use was safe over 26 weeks without affecting pulmonary function and was associated with greater treatment satisfaction and less weight gain compared with RAA, supporting TI as a treatment option for some pediatric patients with type 1 diabetes.
Graphical Abstract
Introduction
Achieving optimal glycemic control in youth with diabetes remains a significant challenge (1). Pediatric patients often miss or delay meal and snack boluses, which negatively affects glycemic control (2). Even with automated insulin delivery (AID) now becoming a mainstay of pediatric type 1 diabetes care in the U.S., most children with type 1 diabetes, using subcutaneously injected or infused insulin, cannot achieve optimal glycemic control (1). The prolonged time required to achieve maximal effects of subcutaneously delivered insulin may play a role. Insulins with a more physiologic, faster onset of action are needed.
One such insulin is inhaled Technosphere insulin (TI) (brand name Afrezza; MannKind Corp., Danbury, CT), which is a dry-powder formulation of recombinant human insulin. Pharmacokinetic and pharmacodynamic studies have shown that inhaled TI has both a faster time to peak concentration (10–15 min) and more rapid dissipation (nearly complete return to predose concentration by 120–180 min) than subcutaneous RAA insulin, which reaches peak concentration 45–60 min after injection and takes 4–5 h to fully dissipate (3–6). TI was approved in 2014 by the Food and Drug Administration (FDA) for use in adults with diabetes (7). Recently, a multicenter trial (INHALE-3: Afrezza Combined with Insulin Degludec versus Usual Care in Adults with Type 1 Diabetes) was conducted with 123 adults with type 1 diabetes using a higher TI dose than given on the FDA label. The trial demonstrated promising efficacy results, even when compared with AID, without safety concerns over 30 weeks of TI use (8–10). Additionally, an in-clinic meal challenge showed that postmeal glucose excursions were reduced with TI compared with RAA (11).
In view of TI’s pharmacokinetics compared with those of injectable RAA and the benefits demonstrated in adults along with a favorable safety profile, inhaled insulin may offer advantages over traditional injectable RAA for youth with diabetes. The rapid onset of action may provide more flexible dosing for meals and the noninjectable nature of inhaled insulin may improve engagement in self-management by avoiding frequent injections. A pharmacokinetics study of TI in children demonstrated no safety concerns and a similar insulin action profile to what has been reported for adults (3). However, there has not been a randomized trial to evaluate TI in a pediatric diabetes population, and TI is not currently approved for pediatric use.
We evaluated the efficacy and safety of TI for pediatric diabetes in a randomized, open-label trial, comparing TI versus RAA in conjunction with basal insulin. Herein, we describe the results of the study.
Research Design and Methods
The randomized controlled trial (RCT) was conducted at 38 endocrinology practices in the U.S. (n = 29 academic and 9 community-based centers) for 26 weeks. The protocol was approved by a central institutional review board. Informed consent was obtained from the parent or guardian of each participant and assent from participants ≥7 years old. A Data and Safety Monitoring Board provided oversight. Cumulative reports of safety events were reviewed semiannually and cases of bronchospasm, hypersensitivity reactions, severe hypoglycemia, diabetic ketoacidosis (DKA), and any serious adverse events (AEs) possibly related to TI were reviewed at the time of occurrence.
Study Participants and Eligibility
Major eligibility criteria included age 4 to <18 years old, clinical diagnosis of type 1 diabetes for ≥6 months or type 2 diabetes for at least 3 months treated with multiple daily injections (MDIs) of insulin for at least 2 weeks, and HbA1c of 7.0–11.0%. Key exclusion criteria included asthma treatment in the prior 12 months, smoking in the prior 6 months or positive cotinine test, pregnancy or breastfeeding, and DKA or severe hypoglycemia in the prior 90 days (see Supplementary Table 1 for a complete listing of eligibility criteria and additional methods).
After initial screening, baseline continuous glucose monitoring (CGM) data were collected. For the baseline CGM data collection, participants who were CGM-naïve or had not used CGM in the past 3 months wore a blinded Dexcom G6 Pro sensor for 2 weeks. Participants who had used CGM in the past 3 months were provided an unblinded Dexcom G6 sensor to use for 2 weeks. To be eligible for the trial, CGM utilization needed to be at least 70% of the 2-week period. The CGM run-in period could be repeated if necessary. During baseline and follow-up, CGM data for analysis were obtained using both Dexcom Clarity and the Dexcom eCRO app.
Study Treatments
Eligible participants were randomly assigned (1:1) to either the TI or RAA group. Randomization was stratified by type of diabetes (type 1 or type 2), baseline HbA1c (<8.5% or ≥8.5%), and age (4 to <13 years, 13 to <18 years) and used a permuted blocks design with blocks of size 2 and 4. Both groups continued to use a basal insulin, generally once daily in the evening with either insulin glargine, insulin degludec, or insulin detemir. For titrating the dose of basal insulin, the protocol recommended for both groups adjusting basal insulin doses with the goal of achieving fasting glucose values of 70–120 mg/dL without hypoglycemia. Both groups were provided with unblinded Dexcom CGM G6 sensors to use continuously and a blood glucose meter (Ascencia Contour Next One).
TI Dosing
TI-group participants received TI through a small inhaler for meal/snack and correction dosing (Supplementary Fig. 1). They were instructed to administer their meal insulin dose at the start of a meal or no later than 20 min after the start of the meal. The first TI dose, given in the clinic with ingestion of a standardized meal, was based on doubling the calculated RAA dose for the size of the meal and then rounding down to the nearest multiple of 4 units. Based on the glucose response to the standardized meal, the initial TI dosing regimen was prescribed.
For meals, participants were advised to add 4 units if the premeal glucose level was >160 mg/dL or for a larger carbohydrate meal than usual and to reduce by 4 units for a smaller-than-usual meal. They also were instructed to give a 4-unit correction dose as soon as 60–90 min after the last dose if the glucose level was >160 mg/dL and to give a correction, if indicated, at bedtime. Participants could take a correction dose >4 units at investigator discretion. At follow-up, the meal dose could be increased as indicated, titrated to target a 60- to 90-min postprandial glucose level of 70–160 mg/dL. Treatment adherence with the prescribed TI and basal insulin regimen was assessed at each visit by having the clinical site personnel evaluate whether the participant was following the prescribed TI dosing regimen and answering the following question on the case report form: “Is the participant taking study drug as prescribed/instructed at prior visit? Yes/No. If No, describe how the participant is dosing including amount and frequency.”
A titration monitoring committee reviewed CGM data, blinded to treatment group, and notified site investigators when they assessed that a dosing change should be considered, based on the following guidelines: time at <70 mg/dL >4%, time at <54 mg/dL >1%, time at >180 mg/dL >25%, or time at >250 mg/dL >5%.
RAA Dosing
The RAA group continued to use their prestudy insulin aspart, insulin lispro, or insulin glulisine for meals/snacks and corrections, and were advised to dose before eating. Dose adjustments for breakfast were based on the lowest prelunch glucose level over a 3-day period, for lunch based on the lowest predinner glucose level, and for dinner based on the lowest prebedtime glucose level. Adherence to the prescribed treatment regimen was not formally assessed in the RAA group.
Study Follow-up and Procedures
The follow-up period for the RCT was 26 weeks. Scheduled visits or contacts occurred after 1, 2, 4, 13, 18, and 26 weeks. The occurrence of AEs was solicited at each visit or contact. At baseline, 13 weeks, and 26 weeks, HbA1c and fasting plasma glucose level were measured at a central laboratory, and spirometry was performed to assess pulmonary function for measurement of forced expiratory volume in 1 s (FEV1). Participants ≥13 years old completed the Diabetes Treatment Satisfaction Questionnaire original status version at baseline and 26 weeks, and the TI group also completed the DTSQ change version (DTSQc) at 26 weeks. A parent or guardian of participants <13 years old completed the parent version of the surveys.
Study Outcomes
The primary outcome was change in HbA1c from baseline to 26 weeks, tested for noninferiority between groups (noninferiority margin 0.4%). Additional outcomes included binary HbA1c metrics, CGM metrics, and DTSQc scores. Safety outcomes included all reported AEs, severe hypoglycemia (defined as severe neuroglycopenia usually resulting in loss of consciousness, coma, or a seizure requiring external assistance by another person to actively administer carbohydrates or glucagon or take other corrective actions requiring third-party assistance for treatment), DKA, other serious AEs, and change in percent predicted FEV1 on pulmonary function testing.
Statistical Methods
Sample size was computed for the primary end point of noninferiority for HbA1c. Assuming a noninferiority margin of 0.4%, true difference between groups of 0, one-sided α of 0.025, power of 80%, SD of 26-week HbA1c of 1.0%, 1:1 randomization, and a dropout rate of 24%, the total sample size to be randomized was calculated to be 264 in order to have the minimum number of 26-week completers be at least 200.
The primary analysis followed the intent-to-treat principle. HbA1c was compared between groups using an ANCOVA model adjusted for type of diabetes, age, and baseline HbA1c. A prespecified per-protocol analysis included participants who met the following criteria: 1) had nonmissing HbA1c data at baseline and 26 weeks, 2) followed the randomized treatment for at least 80% of the 26-week treatment period, and 3) had no major protocol deviations that would affect the primary end point assessment. A post hoc sensitivity analysis excluded TI participants for whom the site indicated that at >80% of visits, the participant was not adhering to the prescribed TI regimen.
Binary outcomes were analyzed using logistic regression models adjusted for the randomization stratification variables and the baseline continuous version of the outcome as a covariate. Multiple imputation was used to estimate missing outcome HbA1c data. The false discovery rate method was used to adjust for multiplicity (noted in table legends when used). Statistical methods for other outcomes are described in table legends. SAS, version 9.4, was used for the analyses.
Results
Participants and Follow-up
Between 29 September 2021 and 28 February 2024, 313 participants were screened. Of these, 230 were randomly assigned to the TI group (n = 117) or the RAA group (n = 113). The mean (± SD) age was 12.6 ± 3.0 years (range 4–17 years), 38% were female, 77% were of White race, and 20% were of Hispanic ethnicity. Type 1 diabetes was present in 225 (98%) and type 2 diabetes in 5 (2%) participants. Mean (± SD) HbA1c at randomization was 8.2% ± 0.9%. Participant characteristics appeared balanced between groups (Table 1).
Table 1.
Participant characteristics by treatment group
| Characteristic | Overall (N = 230) | TI (n = 117) | RAA (n = 113) |
|---|---|---|---|
| Age (years) | |||
| Mean ± SD | 12.6 ± 3.0 | 12.7 ± 2.9 | 12.4 ± 3.2 |
| Range | 4.0–17.0 | 6.0–17.0 | 4.0–17.0 |
| 4 to <13 | 105 (46) | 53 (45) | 52 (46) |
| 13 to <18 | 125 (54) | 64 (55) | 61 (54) |
| Sex | |||
| Female | 88 (38) | 45 (38) | 43 (38) |
| Male | 142 (62) | 72 (62) | 70 (62) |
| Race | |||
| White | 176 (77) | 91 (78) | 85 (75) |
| Black/African American | 24 (10) | 10 (9) | 14 (12) |
| Asian | 7 (3) | 5 (4) | 2 (2) |
| American Indian/Alaska Native | 1 (<1) | 0 | 1 (<1) |
| More than one race | 14 (6) | 7 (6) | 7 (6) |
| Unknown or not reported | 8 (3) | 4 (3) | 4 (4) |
| Ethnicity | |||
| Hispanic or Latino | 45 (20) | 24 (21) | 21 (19) |
| Not Hispanic or Latino | 183 (80) | 92 (79) | 91 (81) |
| Unknown or not reported | 2 (<1) | 1 (<1) | 1 (<1) |
| Type of diabetes | |||
| Type 1 | 225 (98) | 113 (97) | 112 (>99) |
| Type 2 | 5 (2) | 4 (3) | 1 (<1) |
| Diabetes duration (years) | |||
| Mean ± SD | 4.4 ± 3.2 | 4.5 ± 3.3 | 4.4 ± 3.2 |
| Range | 0.5–15.0 | 0.5–14.0 | 0.6–15.0 |
| BMI | |||
| Mean ± SD | 74.4 ± 24.7 | 73.3 ± 24.7 | 75.5 ± 24.9 |
| Percentile | |||
| <85th | 124 (54) | 67 (57) | 57 (50) |
| 85th to <95th | 52 (23) | 21 (18) | 31 (27) |
| ≥95th percentile | 54 (23) | 29 (25) | 25 (22) |
| Prior asthma diagnosis | 5 (2) | 3 (3) | 2 (2) |
| Parent’s highest completed education level | |||
| Less than high school | 3 (1) | 3 (3) | 0 |
| High school graduate/diploma/GED | 45 (20) | 18 (15) | 27 (24) |
| Associate degree/technical/vocational | 56 (24) | 35 (30) | 21 (19) |
| Bachelor’s degree | 75 (33) | 34 (29) | 41 (36) |
| Advanced degree | 45 (20) | 24 (21) | 21 (19) |
| Unknown or not reported | 6 (3) | 3 (3) | 3 (3) |
| Annual household income (USD) | |||
| <50,000 | 37 (16) | 18 (15) | 19 (17) |
| 50,000 to <100,000 | 65 (28) | 36 (31) | 29 (26) |
| 100,000 to <200,000 | 63 (27) | 30 (26) | 33 (29) |
| ≥200,000 | 27 (12) | 13 (11) | 14 (12) |
| Unknown or not reported | 38 (17) | 20 (17) | 18 (16) |
| Health insurance | |||
| Private only | 154 (67) | 81 (69) | 73 (65) |
| Medicaid only | 38 (17) | 19 (16) | 19 (17) |
| Private and Medicaid | 13 (6) | 7 (6) | 6 (5) |
| Medicare | 3 (1) | 1 (<1) | 2 (2) |
| Other government insurance | 14 (6) | 4 (3) | 10 (9) |
| None | 5 (2) | 3 (3) | 2 (2) |
| Unknown or not reported | 3 (1) | 2 (2) | 1 (<1) |
| Baseline HbA1c* (%) | |||
| Mean ± SD | 8.2 ± 0.9 | 8.2 ± 0.9 | 8.2 ± 1.0 |
| Range | 6.3–11.0 | 6.7–10.7 | 6.3–11.0 |
| <7.0 (<53 mmol/mol) | 8 (3) | 3 (3) | 5 (4) |
| 7.0–8.4 (53–68 mmol/mol) | 138 (60) | 72 (62) | 66 (58) |
| 8.5–9.9 (69–85 mmol/mol) | 73 (32) | 37 (32) | 36 (32) |
| ≥10.0 (≥86 mmol/mol) | 11 (5) | 5 (4) | 6 (5) |
| DKA event in the past 12 months | 20 (9) | 10 (9) | 10 (9) |
| Severe hypoglycemia event in the past 12 months | 9 (4) | 1 (1) | 8 (7) |
| Daily insulin use, units/kg/day | |||
| Total daily insulin | 1.02 ± 0.39 | 1.06 ± 0.42 | 0.98 ± 0.36 |
| Total bolus insulin | 0.58 ± 0.29 | 0.60 ± 0.30 | 0.56 ± 0.28 |
| Total basal insulin | 0.44 ± 0.19 | 0.46 ± 0.21 | 0.41 ± 0.17 |
| RAA insulin used | |||
| Lispro | 136 (59) | 72 (62) | 64 (57) |
| Aspart | 87 (38) | 39 (33) | 48 (42) |
| Fast-acting insulin aspart | 7 (3) | 6 (5) | 1 (<1) |
| Basal insulin used | |||
| Glargine | 155 (67) | 78 (67) | 77 (68) |
| Degludec | 64 (28) | 30 (26) | 34 (30) |
| Detemir | 11 (5) | 9 (8) | 2 (2) |
| Current CGM user | 213 (93) | 110 (94) | 103 (91) |
Data are reported as n (%) or mean ± SD, unless otherwise indicated.
*HbA1c results from central laboratory at time of randomization (screening HbA1c value from the central laboratory was used to determine eligibility).
The 26-week primary outcome visit was completed by 105 of the 117 participants (90%) in the TI group and by 108 of the 113 (96%) in the RAA group (Supplementary Fig. 2). One participant was dropped from the study due to using a medication that was an eligibility exclusion. In addition to the other 11 participants in the TI group who dropped from the trial prior to the 26-week primary outcome visit, 15 participants discontinued TI but remained in the trial through 26 weeks. Ten of the 26 early treatment discontinuations were related to inadequate glucose control (hyperglycemia), 4 to poor adherence to the treatment regimen, 5 to coughing with inhalation, 5 to an AE not related to cough, and 2 to difficulty with inhalation (Supplementary Table 2). Overall, TI treatment was discontinued early, with or without trial completion, in 26 of the 117 participants (22%) in the TI group (excluding one participant who was dropped from the study after being determined to be ineligible).
HbA1c Outcomes
In the primary intent-to-treat analysis, mean HbA1c (%) was 8.2 ± 0.9 at baseline and 8.4 ± 1.4 at 26 weeks in the TI group and 8.2 ± 1.0 and 8.2 ± 1.1, respectively, in the RAA group (adjusted difference at 26 weeks = 0.18%; 95% CI −0.07, 0.43; P value for noninferiority = 0.091 for 0.4% margin) (Table 2 and Supplementary Fig. 3). In the prespecified per-protocol analysis, the mean difference was 0.10% (95% CI −0.15, 0.34; P value for noninferiority = 0.016) and in a post hoc sensitivity analysis, which excluded one extreme outlier with respect to treatment adherence (nonadherence recorded at five of six follow-up visits), the mean difference was 0.14% (95% CI −0.10, 0.37; P = 0.026 for noninferiority).
Table 2.
HbA1c and CGM outcomes at 26 weeks
| Baseline | 26 Weeks | Adjusted difference: TI minus RAA (95% CI) [P value]*† | |||
|---|---|---|---|---|---|
| TI | RAA | TI | RAA | ||
| HbA1c (%) | |||||
| Primary intent-to-treat analysis | n = 117 | n = 113 | n = 105 | n = 105 | |
| 8.22 ± 0.87 | 8.21 ± 0.96 | 8.41 ± 1.38 | 8.21 ± 1.10 | 0.18 (−0.07, 0.43) [0.091‡] | |
| Per-protocol analysis‖ | n = 92 | n = 102 | n = 92 | n = 102 | |
| 8.18 ± 0.91 | 8.17 ± 0.95 | 8.32 ± 1.26 | 8.21 ± 1.12 | 0.10 (−0.15, 0.34) [0.016‡] | |
| Sensitivity analysis§ | n = 116 | n = 113 | n = 104 | n = 105 | |
| 8.20 ± 0.86 | 8.21 ± 0.96 | 8.35 ± 1.23 | 8.21 ± 1.10 | 0.14 (−0.10, 0.37) [0.026‡] | |
| CGM metrics (%) | n = 117 | n = 113 | n = 102 | n = 107 | |
| Time at 70–180 mg/dL | 41.9 ± 13.7 | 42.7 ± 15.5 | 39.3 ± 15.7 | 41.4 ± 13.5 | −2.2 (−7.0, 2.7) [0.38] |
| Time at 70–140 mg/dL | 24.5 ± 10.3 | 25.4 ± 11.9 | 22.5 ± 11.2 | 24.3 ± 9.1 | −1.6 (−5.2, 2.1) [0.38] |
| Mean glucose (mg/dL) | 206 ± 34 | 204 ± 36 | 213 ± 44 | 208 ± 36 | 4.4 (−8.3, 17.1) [0.49] |
| Time at <70 mg/dL¶ | 1.87 ± 1.67 | 2.09 ± 1.67 | 1.97 ± 1.77 | 2.24 ± 1.56 | −0.19 (−0.92, 0.91) [0.70] |
| Time at <54 mg/dL¶ | 0.37 ± 0.40 | 0.50 ± 0.57 | 0.43 ± 0.48 | 0.55 ± 0.52 | −0.06 (−0.34, 0.37) [0.70] |
| Time at >180 mg/dL | 56.0 ± 14.8 | 54.9 ± 16.6 | 58.4 ± 17.0 | 56.2 ± 14.5 | 2.1 (−3.2, 7.4) [0.42] |
| Time at >250 mg/dL | 29.2 ± 15.7 | 28.7 ± 16.0 | 31.8 ± 19.1 | 30.3 ± 16.0 | 1.4 (−4.2, 7.0) [0.64] |
| Coefficient of variation | 38.9 ± 6.6 | 38.9 ± 6.5 | 37.7 ± 7.0 | 39.7 ± 5.3 | −1.5 (−3.7, 0.6) [0.13] |
Data are reported as mean ± SD unless otherwise indicated.
*HbA1c was compared between treatment groups using a linear regression model adjusted for randomization strata (type of diabetes [type 1 or type 2], baseline HbA1c stratum [<8.5% or ≥8.5%], and age stratum [≥4 and <13 years, or ≥13 and <18 years]) as factors and baseline HbA1c value as a covariate. For the primary analysis and sensitivity analysis, missing data at 26 weeks were imputed using the return-to-baseline imputation method.
†Each CGM metric was compared between treatment groups using a repeated measures linear regression model with the end point at baseline and 26 weeks as the dependent variable. The model was adjusted for randomization strata (type of diabetes [type 1 or type 2], baseline HbA1c stratum [<8.5% or ≥8.5%], and age stratum [≥4 and <13 years, or ≥13 and <18 years]) as factors. The model is adjusted for the baseline value of the end point by forcing the treatment groups to have the same mean value at baseline. Missing data were handled using direct likelihood, which assumes missing at random. P values and CI assess superiority and were adjusted to control the false discovery rate.
‡P value for noninferiority based on a margin of 0.4%.
§The sensitivity analysis excluded one participant randomized to TI who was an outlier with respect to being extremely noncompliant.
‖The per-protocol population included participants who met the following criteria: 1) had nonmissing HbA1c data at baseline and 26 weeks, 2) followed randomized treatment for at least 80% of the 26-week treatment period, and 3) had no major protocol deviations that would affect the primary end point assessment.
¶Because the hypoglycemia end points had skewed distributions, values were winsorized at the 10th and 90th percentiles. P values and CIs for the treatment effects for these outcomes were calculated using bootstrap.
HbA1c improved from baseline to 26 weeks by >0.5% (5.5 mmol/mol) in 16 participants (15%) in the TI group versus 24 in the RAA group (23%) and worsened by >0.5% (5.5 mmol/mol) in 30 participants (29%) in the TI group versus 21 in the RAA group (20%) (Supplementary Tables 3–5). At 26 weeks, 11 in the TI group (10%) versus 9 in the RAA group (9%) had HbA1c <7.0% (53 mmol/mol). Results at 13 weeks were similar to those at 26 weeks (Supplementary Table 6).
An exploratory analysis suggested that participants 4–12 years old and participants with baseline HbA1c ≥8.5% had worse outcomes with TI compared with RAA than did older participants or participants with lower HbA1c levels (Supplementary Tables 7–9). Analyses in other subgroups did not suggest a differing treatment effect according to baseline characteristics (Supplementary Fig. 4).
CGM Outcomes
CGM-measured time in range (TIR) 70–180 mg/dL was 41.9% ± 13.7% at baseline and 39.3% ± 15.7% at 26 weeks in the TI group and 42.7% ± 15.5% and 41.4% ± 13.5%, respectively, in the RAA group (adjusted difference −2.2%, 95% CI −7.0, 2.7; P = 0.38) (Table 2). TIR was >70.0% at 26 weeks in only 3 participants (3%) in each group (Supplementary Table 5). Time at <54 mg/dL was low at baseline (0.37% ± 0.40% in the TI group and 0.50% ± 0.57% in the RAA group) and had changed little at 26 weeks (0.43% ± 0.48% and 0.55% ± 0.52%, respectively), with a treatment-group difference of −0.06% (95% CI −0.34, 0.37; P = 0.70). Other CGM metrics also did not show meaningful differences between groups overall (Table 2) or when analyzed separately for daytime and nighttime (Supplementary Table 10). Assessing glucose levels over the 24 h of the day indicated little difference between groups overnight, with a suggestion of lower mean glucose with TI during the morning hours and lower mean glucose with RAA during the afternoon hours (Fig. 1 and Supplementary Fig. 5).
Figure 1.
CGM-measured mean glucose by hour at 26 weeks. The circles denote the median values and the lower and upper boundaries of the shaded region represent the 25th and 75th percentiles, respectively.
Insulin Dosing
In the TI group, the mean total daily insulin dose was 1.06 ± 0.42 units/kg at baseline prior to initiating TI, with 0.46 ± 0.21 units/kg basal insulin and 0.60 ± 0.30 units/kg RAA insulin. At 26 weeks, mean daily basal insulin was 0.50 ± 0.24 units/kg and mean daily TI insulin was 1.45 ± 0.84 “TI” units/kg (a TI unit is bioequivalent to 0.33–0.50 RAA units). The mean ratio of daily TI insulin at 26 weeks to daily RAA insulin at baseline was 2.96 ± 1.83 (Supplementary Table 11).
Among the 91 participants in the TI group who were using TI at 26 weeks (or within 1 week of the 26-week visit), 61 (67%) were considered by the clinical site to be following the prescribed regimen at all completed follow-up visits and 72 (79%) at ≥80% of completed visits; 6 (7%) were considered nonadherent at least 50% of visits (Supplementary Table 12).
In the RAA group, the mean total daily insulin dose was 0.98 ± 0.36 units/kg at baseline and 1.06 ± 0.37 units/kg at 26 weeks. Changes in total daily basal and RAA insulin doses were small (Supplementary Table 11).
Weight and BMI
Weight change from baseline was +1.6 ± 3.7 kg in the TI group and +3.0 ± 3.1 kg in the RAA group (difference at 26 weeks −1.4 kg; 95% CI −2.6, −0.3; P = 0.009) (Supplementary Table 13). Change in mean BMI percentile paralleled the change in weight (treatment group difference at 26 weeks −4.2; 95% CI −7.4, −0.9; P = 0.009).
Diabetes Treatment Satisfaction Questionnaire
The treatment satisfaction subscale scores (potential range 0–6) suggested higher treatment satisfaction in the TI group than RAA group (P = 0.004) when combining the surveys completed by parents and by teenagers (Supplementary Table 14). Among the parents of participants <13 years old, the mean treatment satisfaction subscale score increased from 4.23 ± 0.91 at baseline to 4.99 ± 0.76 at 26 weeks in the TI group and from 4.28 ± 0.98 to 4.55 ± 1.12, respectively, in the RAA group (change from baseline 0.72 ± 1.06 vs. 0.28 ± 0.92; P = 0.08). Among participants 13–17 years old, the mean treatment satisfaction subscale score increased from 4.63 ± 0.76 at baseline to 4.99 ± 0.74 at 26 weeks in the TI group and from 4.62 ± 1.03 to 4.60 ± 0.95, respectively, in the RAA group (0.38 ± 0.89 vs. −0.04 ± 0.92; P = 0.07). In addition, on the DTSQc survey (potential score range −3 to +3) completed by the TI group at 26 weeks, treatment satisfaction was high for the participants aged 13–17 years and the parents of the younger participants, when comparing their current regimen of TI plus basal insulin to their prestudy RAA–basal insulin regimen (Supplementary Table 14). No significant or meaningful differences between treatment groups were found for the perceived diabetes control and the perceived frequency of hypoglycemia subscales.
Safety Outcomes
Overall, 240 AEs were reported by 88 participants (75%) in the TI group and 195 events by 74 participants (65%) in the RAA group (Table 3 and Supplementary Table 15). A severe hypoglycemia event occurred in two participants in the TI group and in one participant in the RAA group; DKA occurred in none and one participant, respectively, in the two groups.
Table 3.
Safety outcomes
| All reported AEs | TI (n = 117 participants) | RAA (n = 113 participants) |
|---|---|---|
| No. of events | 240 | 195 |
| Events per participant | ||
| 0 | 29 (25) | 38 (34) |
| 1 | 36 (31) | 31 (27) |
| 2 | 17 (15) | 18 (16) |
| 3 | 13 (11) | 10 (9) |
| 4 | 6 (5) | 7 (6) |
| ≥5 | 16 (14) | 9 (8) |
| Severe hypoglycemia events per participant* | ||
| 0 | 115 (98) | 112 (>99) |
| 1 | 2 (2) | 1 (<1) |
| DKA events per participant | ||
| 0 | 117 (100) | 112 (>99) |
| 1 | 0 | 1 (<1) |
| Other serious AEs per participant† | ||
| 0 | 116 (>99) | 112 (>99) |
| 1 | 1 (<1) | 0 |
| 2 | 0 | 1 (<1) |
| Other AEs of special interest‡ | 3 (3) | – |
| % Predicted FEV1 | ||
| At baseline, n | 117 | 113 |
| Mean ± SD | 99.6 ± 11.5 | 102.3 ± 11.2 |
| At 13 weeks, n | 105 | 110 |
| Mean ± SD | 96.7 ± 13.4 | 98.3 ± 13.3 |
| At 26 weeks, n | 99 | 103 |
| Mean ± SD | 96.5 ± 12.0 | 98.1 ± 11.7 |
| Change from baseline to 26 weeks, n | 99 | 103 |
| Mean ± SD | −2.9 ± 7.7 | −4.2 ± 8.0 |
| 26-Week adjusted group difference (95% CI) [P value]§ | 0.7 (−1.5, 2.8) [0.53] | |
| % Predicted FEV1 decrease ≥15% | 6 | 10 |
Data are reported as n (%) unless otherwise indicated.
*Of the two severe hypoglycemia events in the TI group, one was considered unlikely related to TI because there had not been a TI dose for several hours prior to the event; it was assessed by the site investigator and medical monitor that an increase in the glargine dose may have precipitated the event. In the other case, the participant inadvertently inhaled a correction dose that was twice the number of intended TI units and developed cognitive impairment requiring the parent to provide juice, which resolved the event.
†The other serious AEs were spinal operation in the TI group and osteomyelitis and suicide attempt in the RAA group.
‡Certain AEs were specified in the protocol as AEs of special interest: defined as acute bronchospasm, clinically relevant decline in pulmonary function (>15% decline from baseline in percent predicted FEV1 accompanied by respiratory symptoms), hypersensitivity reactions (including anaphylaxis), severe hypoglycemia, and DKA. In addition to the two severe hypoglycemia events, there were three AEs of special interest in the TI group. One was shortness of breath for several minutes after TI inhalation, although not consistently, with some exercise intolerance; FEV1 unchanged from baseline. Another involved chest tightness, shortness of breath, and wheezing lasting about 30 min and resolving without treatment, possibly consistent with bronchospasm. Episodes only occurred after a dose of TI ≥36 units. Physical exam findings were normal and FEV1 was unchanged from baseline. The third was cough and wheezing after an upper respiratory infection. Wheezing present on exam was indicative of possible bronchospasm. FEV1 was unchanged from baseline, but the medical monitor considered it possible that TI was exacerbating persistent airway hyperactivity after the respiratory infection.
§Difference is TI minus RAA. Percent predicted FEV1 was compared between treatment groups using a linear regression model adjusted for randomization strata (type of diabetes [type 1 or type 2], baseline HbA1c stratum [<8.5% or ≥8.5%], and age stratum [≥4 and <13 years, or ≥13 and <18 years]) as factors and baseline percent predicted FEV1 value as a covariate. The analysis was based on observed data only.
In the TI group, there were 49 AEs in 35 participants (30%) considered possibly, probably, or definitely related to TI (Supplementary Table 16). TI-associated cough, which was usually transient and mild, was reported by 20 of the participants (17%), 5 of whom dropped from the trial or discontinued TI but remained in the trial. Three nonserious respiratory AEs were considered AEs of special interest per the protocol (see Table 3 for description of events).
The mean percent predicted FEV1 was 99.6 ± 11.5 at baseline and 96.5 ± 12.0 at 26 weeks in the TI group and 102.3 ± 11.2 and 98.1 ± 11.7, respectively, in the RAA group (difference = 0.7; 95% CI −1.5, 2.8; P = 0.53). At 26 weeks, FEV1 had decreased by ≥15% from baseline in 6 participants (6%) in the TI group and 10 in the RAA group (10%).
Conclusions
In this multicenter RCT, inhaled TI was shown to be safe, without affecting pulmonary function, in youth with diabetes (98% with type 1 diabetes) compared with RAA insulin for meal and correction dosing, with greater treatment satisfaction and less weight gain. Although TI did not meet the prespecified criteria for HbA1c noninferiority to RAA in the intent-to-treat analysis, it was noninferior in the prespecified per-protocol analysis and in a post hoc sensitivity analysis, which excluded one extreme outlier with respect to treatment adherence. Participants ≥13 years old, particularly those with baseline HbA1c <8.5%, appeared to do better with TI than participants <13 years old, particularly those with HbA1c >8.5%.
The study results indicate use of TI in a pediatric population has a favorable safety profile, comparable to that of an adult population (8,12). Notably, the frequency of hypoglycemia was low and similar with TI versus RAA. There were no TI-related serious AEs other than one case of severe hypoglycemia that was caused by inadvertent doubling of the intended correction dose. Although TI-associated cough was reported by 17% of participants, it was generally mild and transient, typically lasting just a few seconds, and its frequency was similar or less than the frequency reported in adult studies (8,12). Pulmonary effects were infrequent and none was serious. Importantly, pulmonary function, as measured by change in FEV1, changed similarly between the TI and RAA groups, reinforcing the safety of TI in children.
Adolescent participants and parents of younger participants reported higher satisfaction with TI compared with RAA. The finding of less weight and BMI increase with TI compared with RAA also is important and is consistent with what has been reported for adults using TI (8). The mechanisms to explain less weight gain in youth and in adults using TI versus RAA are not precisely known. Hypothetically, lower weight gain may be related to better matching of TI’s onset and peak of action in comparison with RAA, potentially with less need to treat hypoglycemia or impending hypoglycemia in the postprandial state with TI. Similarly, the reduced time spent with relative hyperinsulinism, due to the rapid clearance of TI, may result in a less anabolic postmeal state in TI versus RAA users.
As noted, the pharmacokinetic and pharmacodynamic profiles of TI reflect a more rapid effect or more rapid dissipation of effect than that of RAA (3,6). In addition to the potential benefits associated with less weight gain, the pharmacokinetics and pharmacodynamics of TI provide for greater meal dosing flexibility in that TI can be inhaled at the time of meal rather than 10–15 min before eating, as recommended for RAA insulin. Furthermore, the rapid dissipation of TI’s effects may reduce the risk of postmeal and exercise-induced hypoglycemia. Given TI’s ability to blunt glycemic excursions related to meals (11,13), it may be a superior meal insulin but may require repeated dosing for optimal effect.
The current U.S. labeling of TI is to base initial dosing on an approximate 1:1 conversion of the usual RAA dose for a patient. However, this is based on the amount of powdered insulin in a cartridge, not on the amount actually absorbed in the lungs. In this study, the starting TI dose was approximately twice the RAA dose, and, by 26 weeks, this was titrated to be almost three times the baseline RAA dose. Our results, which have also been supported by adult studies using a two- to three-times ratio of TI to RAA (8,14), indicate this dosing approach is safe.
Despite the theoretical advantages of TI, relatively few participants in the TI group achieved targets for TIR (>70%) or HbA1c (<7%). These results contrast with those of the INHALE-3 adult trial, in which 30% of adult MDI users achieved an HbA1c level <7% with TI compared with 4% continuing RAA, and 24% versus 0%, respectively achieved TIR >70% (8). We speculate that number of participants who discontinued TI due to dissatisfaction with glucose levels and the lack of improvement in glucose metrics with TI in many participants were due to several factors. First, to achieve an optimal effect of using TI, it is necessary to not only estimate the amount of TI needed for each meal bolus but also to inhale a correction dose if hyperglycemia is present 1–2 h after the prior bolus. Frequent correction dosing and dosing for snacks might have been challenging for many participants, particularly on school days. Second, there may have been a tendency to underdose TI for meals, snacks, corrections, and at bedtime due to concerns about hypoglycemia. The fact that the number of TI units is 2–3 times the equivalent RAA dose might have factored into participant and parental hesitancy to titrate the dose of TI needed to optimize glycemic control. Future studies as well as clinical use of TI must emphasize the difference in TI units versus RAA units and ensure users are able and willing to give frequent corrections throughout the day.
The study has limitations that should be acknowledged. The small number of participants with type 2 diabetes limits our ability to generalize these findings to the broader pediatric population affected by type 2 diabetes. Assessment of adherence to insulin dosing in the TI group relied on self-report, which may not have provided an accurate representation. The relatively short duration of 26 weeks limits the ability to evaluate long-term efficacy and safety. However, a 26-week extension of the present study will provide data on TI use for 52 weeks, which will be reported separately. The study included only MDI users, the percentage of whom is being reduced by the increasing adoption of AID systems. There is a need to evaluate the use of TI as an adjunct to use of AID to limit postmeal glucose excursions.
In conclusion, the INHALE-1 study results demonstrate that inhaled TI use by youth with diabetes was safe and did not affect pulmonary function over a 26-week period. Although the primary intent-to-treat analysis did not meet the prespecified criteria for HbA1c noninferiority, the safety results plus the findings of greater treatment satisfaction and less weight gain support TI being a treatment option for some pediatric patients with type 1 diabetes, particularly for youth who choose not to or are unable to use an AID system.
This article contains supplementary material online at https://doi.org/10.2337/figshare.30456095.
Article Information
Acknowledgments. C.Z. and F.B.H. are editors of Diabetes Care but were not involved in any of the decisions regarding review of the manuscript or its acceptance.
Representatives of MannKind participated in protocol development with the investigators and were involved with the Jaeb Center for Health Research in study oversight. The manuscript was written by the authors without a medical writer and the content is solely the responsibility of the authors. MannKind reviewed the manuscript and provided comments but had no rights of approval of the content of the manuscript or the decision on journal submission.
Duality of Interest. M.J.H. reports receiving consultancy fees and serving as a scientific advisory board member for MannKind and SAB BIO and consultancy fees from Sanofi. G.P. reports receiving consultancy fees and reports serving as a principal investigator in MannKind-funded research. L.A.D. reports research support to her institution from Lilly, Sanofi, and Zealand and consulting fees from Tandem Diabetes. L.M.L. reports consultant and advisory Board fees from Boehringer Ingelheim, Janssen, Dexcom, Medtronic, Sequel, Tandem Diabetes, Sanofi, MannKind, Arbor Biotech, Vertex. S.M.W. reports receiving advisory board fees from Bohringer-Ingelheim and Mannkind. B.A.N. reports receiving advisory board fees from Sanofi. J.R.W. reports receiving research funding from Insulet and Medtronic. K.L. reports receiving advisory board fees from Mannkind. M.K. reports receiving research grants from 89bio, Medtronic, Abbott, Akero Therapeutics, Biomea Fusion, Carmot, Dexcom, AbbVie, Amgen, Biolinq, Diamyd, Endogenex, Insulet, IONIS Pharmaceuticals, Kowa, Novo Nordisk, Tandem, Zucara Therapeutics, Zydus, Sinocare, Boehringer Ingelheim, Corcept, and Lilly; and advisory board fees from AP stem and Corcept. H.R. reports receiving advisory board fees from Mannkind. A.B. reports receiving research support from Verdiva Bio, Amgen Inc, Endogenix, Gasherbrum Bio, VtV Therapeutics, Boehringer Ingelheim Pharmaceuticals, Abbott Diabetes Care, AbbVie, Covance, Dexcom, Lilly, Madrigal Pharmaceuticals, Medtronic, Novo Nordisk, MannKind, Carmot, Quintiles, Akero, 89bio, Viking Therapeutics, PPD, and Zydus. L.E. receives salary support from the National Institute of Diabetes and Digestive and Kidney Diseases, served on the advisory boards of Abbott, Diabetes Center Berne, Sequel, MannKind, and Medtronic; reports receiving consulting fees from Jaeb and Tandem Diabetes Care; and has received honorarium fees from Med Learning Group (Sanofi-sponsored grant), Tandem Diabetes Care, Medtronic, and Insulet. L.E.’s institution has received research support from Breakthrough T1D, Medtronic, Mannkind, and Abbott. D.M.M. reports research support from the National Institutes of Health, National Science Foundation, Breakthrough T1D, and the Helmsley Charitable Trust; and has consulted for Abbott, Sanofi, Lilly, Medtronic, Biospex, Kriya, and Enable Biosciences. D.M.M. reports that his institution has received research support from Dexcom. M.A.W. reports receiving speaker fees from Alexion, Sanofi, and Mannkind. M.V.N. reports receiving consultancy fees from Soleno therapeutics and Novo Nordisk, and that her institution has received research funding or support on her behalf from Provention Bio and Novo Nordisk. S.M.B. reports receiving grants paid to their institution by the Cystic Fibrosis Foundation, Jaeb Center for Health Research, and MannKind, and reports study support to their institution from Dexcom. M.A.C. reports receiving research support from Dexcom and Abbot and consultancy fees from Glooko. R.W.B. reports his institution has received funding on his behalf as follows: grant funding, study supplies, and consulting fees from Insulet, Tandem Diabetes Care, and Beta Bionics; grant funding and study supplies from Dexcom and Abbott; grant funding from Bigfoot Biomedical, embecta, Sequel Med Tech, and MannKind; study supplies from Medtronic; consulting fees and study supplies from Novo Nordisk; consulting fees from Vertex, Hagar, DreaMed, Ypsomed, Abata Therapeutics, Lilly and Zucara. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. M.J.H., R.W.B., and R.M. wrote the first draft of the manuscript and vouch for the completeness and accuracy of the data and for the fidelity of the trial to the protocol. All other authors had a key role in the study and provided review and critical input into the writing of the manuscript. R.W.B. is the guarantor of this work and, as such, had full access to all study data and takes responsibility for the integrity of the data and the accuracy of the analysis.
Prior Presentation. The results were presented at the 85th meeting of the American Diabetes Association, Chicago, IL, 20–23 June 2024.
Handling Editors. The journal editor responsible for overseeing the review of the manuscript was Stephen S. Rich.
Funding Statement
Funding for the study was provided by MannKind.
Footnotes
Clinical trial reg. no. NCT04974528, clinicaltrials.gov
*A complete list of the INHALE-1 Study Group can be found in the supplementary material online.
Contributor Information
Michael J. Haller, Email: hallemj@peds.ufl.edu.
INHALE-1 Study Group*:
Lori Laffel, Elvira Isganaitis, Hannah Desrochers, Louise Ambler-Osborn, Jade Doolan, Kerry Milaszewski, David Maahs, Lisa Norlander, Priya Prahalad, Laya Ekhlaspour, Sejal Shah, Ilenia Balistreri, Noor Alramahi, Eliana Frank, Nora Arrizon-Ruiz, Michelle Van Name, Kate Weyman, Amy Steffen, Melinda Zgorski, Himala Kashmiri, Nikta Forghani, Heather Speer, Ana Cruz, Diana Martinez, Marissa Erickson, Hunter Vallejo, Perrin White, Jimmy Penn, Abha Choudhary, Colten Youngblood, Yazmin Molina, Michelle Murphy, Lisa Staples-Wherry, Chantal Nwosu, Michael Gottschalk, Anna Cymbaluk, Carla Demeterco-Berggren, Ron Newfield, Marcela Vargas Trujillo, Jane Kim, Marla Hashiguchi, Maja Marinkovic, Mary Patterson, Michelle Rivera-Vega, Susan Phillips, Angela Lam, Lisa Paglia, Shannan Davis, Kristina Cossen, Andrew Muir, Eric Felner, Lynette Gonzalez, Linton Cuff, Xiaomiao LanPidhainy, Amber Antich, Michael Haller, Laura Jacobsen, Brittany Bruggeman, Timothy Foster, Sarah Peeling, Joshua Modeste, Miriam Cintron, Angela Chen, Janey Adams, Sarah Doll, Diane Biernacki, Jessica Thomas, Jerylen Caraecle, Roberta Brunson, Janet King, Charlie Church, Jamie Thomas, Kimberly Garrett, Anuj Bhargava, Christine Langel, Kathy Fitzgerald, Christine Burke, Kendra, Kara Turner, Allie Bohr, Konda Reddy, Christine Truong, Daniella Orpi, Maria McFadden, Pamela Hedrick, Calla Boatright, Joshua Smith, Jonathan Williams, Brittani Holden, Myleigh Brink, Kelsie Wetzel, Roshanak Monzavi, Nancy Chang, Juliana Austin, David Geller, Debra Miller, Pamela Parcon, Daniel Brimberry, Henry Landers, Lara Keskin, Steven Willi, Pantea Minnock, Sarah, Meghan Mack, Jean “Yvener” Deverlis, Diana Olivos-Stewart, Joshua Moyer, Sakai Young, Lisa Kalnik, Catherine Chang, Emma Bickfard, Aiyana Mate, Michelle Wu, Lisa Sher, Chanpreet Toor, Mark Clements, Cintya Schweisberger, Heather Harding, Casey McClain, Kelsye Howell, Richard Dugan-Starr, Kupper Wintergerst, Sara Watson, Emily Montgomery, Heather Rush, Lauren Bauer, Paul Hiers, Gwendolyn Pierce, Shannon Hall, Shylaja Srinivasan, Eda Cengiz, Jasmine Hosseinipour, Avani Narayan, Rebecca Wesch, Nicole Sears, Jamie Wood, Sarah MacLeish, Terri Casey, Paul McGuigan, Wendy Campbell, Ramon Adams, Faisal Malik, Beth Loots, Britney Ellisor, Samantha Garcia Perez, Ciara DeGraff, Andria Bonney, Kathleen Bethin, Lucy Mastrandrea, Teresa Quattrin, Amanda House, Hannah Elsinghorst, Emily Tabaczynski, Alora Rumfola, William Hibbard, Meaghan Gleason, Nicholas Torre, Danielle Handley, Krista Willard, Linda DiMeglio, Tamara Hannon, Amy Hatton, Stephanie Woerner, Nate De Jong, Hannah Lease, Dana Chatila, Sarah Holloway, Muna Sunni, Antoinette Moran, Shannon Beasley, Janice Leschyshyn, Kali Johnson, Beth Pappenfus, Veronica Jones-Carr, Melissa Kelly, Henry Rodriguez, Dorothy Shulman, Janet Rodriguez, Stephenie Yapchanyk, Shannon Carper, Ponja Hemphill, Elizabeth Doble, Megha Patel, Gregory Spires, Brenden Teaman, Hillary Henry, Michael Tansey, Julie Coffey, Eva Tsalikian, Alexandra LaCarte, Paige Aossey, Holly Hemann, Brooke Burnett, Carter Johnson, Michael Wood, Mary McFeters, Mark Kipnes, Lauren Pankratz, Sheila Calumpiano, Terri Ryan, Stephanie Beltran, Kashif Latif, Jennifer Jurado, Katherine Dean, Marcy Kaufman, Yovana Avalos, Lesley Draffin, Brittany Lewis, Kanika Shanker, Kristin Sabanosh, Carrie Gregory, Marie Fox, Asheesh K. Dewan, Carlos Torres, Uenah Yun, Travis Boman, Tayler Romprey, Chase Larson, Lissette Velasquez, Lesley Perez, Gnanagurudasan Prakasam, Ulhas Nadgir, Natalie Marlen, Mila Melnik, Robin Nemery, Lital Reitblat, Martha Taboada, Sara Hart-Unger, Lisa Kenigsberg Fetcher, Paulette Smith, Jean Barton, Barry J. Reiner, Lee Bromberger, Rosanna Fiallo-Scharer, Samuel Engle, Joanna Kramer, LaTonda Tyler, Cristen Berry, Brittany Powroznik, Sara Agamaite, Hannah Hubert, Joe Zack, Dawn Dziubek, Mary Pat Gallagher, Kayla Wagner, Jeniece Ilkowitz, Juana Gonzalez, David Sparling, Kruti Shah, Joni Beck, Deidre Graham, Linda Weber, Bryce Nelson, Erica Memoli, Scott Blackman, Risa Wolf, Lee Bromberger, Dhruva Patel, Surya Narayan Mulukutla, Jose David Gamez-Godoy, Adrienne Casciato, Luis Cantu, Victoria Mancillas, Nicole Sheanon, Larry Dolan, Ryan Brady, Cierra Farrell, Bliss Magella, Kaitlyn Witt, Patrick Hanley, Chijioke Ikomi, Taylor Sullivan, Dhwani Shah, Jennifer Abuzzahab, Melinda Pierce, Jennifer Kyllo, Michelle Smith, Brittany Machus, Briana Escobar, Abby Schmitt, Kevin Kaiserman, Johanna Ulloa, Jennifer Pleitez, Joseph Sylvan, Lynn Badgett, Robin Gal, Thomas J. Mouse, Kamille Janess, Melanie Christian, Nicole Cagnina, Emma Smith, Nicole Reese, Heidi Strayer, Lauren Kanapka, Craig Kollman, Roy W. Beck, Deanna Gabrielson, and Van Le
Supporting information
References
- 1. Ebekozien O, Mungmode A, Sanchez J, et al. Longitudinal trends in glycemic outcomes and technology use for over 48,000 people with type 1 diabetes (2016-2022) from the T1D Exchange Quality Improvement Collaborative. Diabetes Technol Ther 2023;25:765–773 [DOI] [PubMed] [Google Scholar]
- 2. Christie J, Clements MA, Williams DD, et al. Mealtime insulin BOLUS score more strongly predicts HbA1c than the self-care inventory in youth with type 1 diabetes. J Diabetes Sci Technol 2025;19:340–344 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Haller MJ, Jones MC, Bhavsar S, et al. Time-action profile of Technosphere insulin in children with type 1 diabetes. Diabetes Ther 2023;14:611–617 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Lyumjev (Insulin Lispro-AABC). Prescribing Information. Eli Lilly and Company, 2020. Accessed 2 October 2025. Available from https://www.accessdata.fda.gov/drugsatfda_docs/label/2020/761109s000lbl.pdf. [Google Scholar]
- 5. Humalog (Insulin Lispro Injection, USP [rDNA Origin]). Prescribing Information. Eli Lilly and Company, 1996. Accessed 2 October 2025. Available from https://www.accessdata.fda.gov/drugsatfda_docs/label/2013/020563s115lbl.pdf. [Google Scholar]
- 6. Heinemann L, Baughman R, Boss A, et al. Pharmacokinetic and pharmacodynamic properties of a novel inhaled insulin. J Diabetes Sci Technol 2017;11:148–156 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Bode BW, McGill JB, Lorber DL, et al.; Affinity 1 Study Group . Inhaled Technosphere insulin compared with injected prandial insulin in type 1 diabetes: a randomized 24-week trial. Diabetes Care 2015;38:2266–2273 [DOI] [PubMed] [Google Scholar]
- 8. Hirsch IB, Beck RW, Marak MC, et al.; INHALE-3 Study Group . A randomized trial comparing inhaled insulin plus basal insulin versus usual care in adults with type 1 diabetes. Diabetes Care 2025;48:353–360 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Beck RW, Bailey RJ, Akturk HK, et al.; INHALE-3 Study Group . A 13-week single-arm evaluation of inhaled Technosphere insulin plus insulin degludec for adults with type 1 diabetes. Diabetes Technol Ther 2025;27:161–169 [DOI] [PubMed] [Google Scholar]
- 10. Beck RW, Bailey RJ, Klein KR, et al.; INHALE-3 Study Group . Inhaled Technosphere insulin plus insulin degludec for adults with type 1 diabetes: the INHALE-3 Extension Study. Diabetes Technol Ther 2025;27:170–178 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Hirsch IB, Beck RW, Marak MC, et al.; INHALE-3 Study Group . A randomized comparison of postprandial glucose excursion using inhaled insulin versus rapid-acting analog insulin in adults with type 1 diabetes using multiple daily injections of insulin or automated insulin delivery. Diabetes Care 2024;47:1682–1687 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. McGill JB, Peters A, Buse JB, et al. Comprehensive pulmonary safety review of inhaled Technosphere® insulin in patients with diabetes mellitus. Clin Drug Investig 2020;40:973–983 [Google Scholar]
- 13. Heinemann L, Parkin CG. Rethinking the viability and utility of inhaled insulin in clinical practice. J Diabetes Res 2018;2018:4568903. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Kaiserman KB, Christiansen M, Bhavsar S, et al. Reduction in postprandial peak glucose with increased Technosphere insulin dosage. J Diabetes Sci Technol 2024;18:397–401 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.