Reduction in Postprandial Peak Glucose With Increased Technosphere Insulin Dosage

Abstract

Background:

Technosphere Insulin (TI) is an ultra-rapid-acting inhaled insulin. This study assessed the mean peak two-hour postprandial glucose concentration with the initial TI dose (dose 1) calculated per the current label (United State Prescribing Information) compared with a ~2× higher dose (dose 2). Secondary objectives were to evaluate hypoglycemia within the two-hour postprandial period, evaluate change in forced expiratory volume in one second (FEV₁) before and after the two-hour postprandial period, and monitor for other adverse events.

Methods:

Twenty patients with diabetes, on basal-bolus insulin therapy, received an initial dose 1 of TI followed by the higher dose 2, one to three days later. Subjects received an identical meal for both visits, and TI doses were administered immediately prior to the meal.

Results:

The higher dose 2 provided significant reductions in mean postprandial glucose excursion (PPGE) in the two-hour postprandial period starting from 45 minutes (P = .008) to 120 minutes (P < .0001). Mean peak glucose was reduced from 228.6 to 179.3 mg/dL (P < .001) at two hours. Two hypoglycemic events (one level 1, one level 2) were observed in a single subject during the two-hour postprandial period with dose 2. There were no significant changes in FEV₁ after either dose of TI.

Conclusions:

The higher dose 2 reduced PPGE versus the current label recommended dose 1 within the two-hour postprandial timeframe without any new safety concerns. When confirmed with a larger study, this higher TI dosing recommendation may help patients and clinicians minimize immediate postprandial hyperglycemia when titrating TI for prandial glucose control.

Introduction

Individuals with insulin-treated diabetes are often unable to achieve and maintain their glycemic targets without excessive hypoglycemia.^1-5 Although the causes of hypoglycemia are multifactorial, the type of mealtime insulin used can often contribute to hypoglycemia risk. For example, high rates of hypoglycemia were observed in the Diabetes Control and Complications Trial (DCCT) in which regular insulin was used for mealtime and correction doses. ⁶ With subcutaneously injected rapid-acting insulin analogs (RAA), these rates were significantly lower in subsequent clinical studies. ⁷ The contribution of postprandial hyperglycemia is often overlooked, and it continues to be a major factor contributing to glycated hemoglobin (HbA1c) as patients get closer to the target HbA1c of 6.5% or 7%.^8-11

Subcutaneously injected RAA administered as prandial insulin can take up to 15 to 30 minutes to start having a measurable effect, requires 52 to 60 minutes to reach maximum concentration, and stays in the system for up to five hours, sometimes long after the meal glucose is absorbed.^12-14 Current RAA formulations do allow for administration at the start^12,13 or immediately after starting a meal, ¹⁴ but achieving optimal postprandial control commonly requires patients to administer their prandial insulin 15 to 20 minutes before the meal.^15,16 The extended duration of insulin activity following subcutaneous injection of RAA formulations (up to five hours)^17-19 can also increase hypoglycemia risk.

The rapid time to peak effect and clearance of Technosphere Insulin (TI), brand name Afrezza^® (MannKind Corporation, Westlake Village, CA, USA), a dry powder formulation of recombinant human insulin for oral inhalation, ²⁰ addresses some of the limitations of subcutaneously injected prandial RAA. Administered at the time of the meal using a breath-powered inhaler, TI particles dissolve immediately in the lungs and insulin is rapidly absorbed systematically. After administration of TI in adults, the maximum serum insulin concentration occurs in approximately 10 to 20 minutes^21,22 and returns to near-baseline concentrations in approximately 180 minutes. ²³

The use of prandial TI has been studied when administered at the dosage calculation recommended in Figure 1 of the current labeling for initial dose conversion. ²⁴ This dosage often requires administration of a second dose after meals for correction of postprandial hyperglycemia. ²⁵ The current study was undertaken to evaluate a modified conversion from subcutaneously injected RAA to a higher TI dose and compare it with the current conversion recommendation in the United States Prescribing Information (USPI) by assessing the impact on postprandial glucose excursion (PPGE) and safety.

Figure 1.

Mean PPGE (±SE) in the two-hour postprandial period (dose 1 vs dose 2). Abbreviation: PPGE, postprandial glucose excursion.

Methods

Study Design

This was an open-label, single-arm, proof-of-concept dosing study of TI in adult subjects with type 1 diabetes (T1D) or type 2 diabetes (T2D) treated with a basal-bolus insulin therapy. The study was conducted at two investigational sites between April 7, 2021, and May 28, 2021. The primary objective was to evaluate the effects of TI administered at two dosages on PPGE for two hours. The initial dose 1 was determined using the conversion table specified in the current label (USPI) based on each subject’s calculated prandial dosage of RAA. ²⁴ The higher dose 2 was twice each subject’s calculated prandial RAA dosage rounded down to the nearest TI dose, which is a multiple of four units. The doses were inhaled immediately prior to identical standardized meals under medical supervision. The study was registered with ClinicalTrials.gov (NCT04849845) and was approved by an independent institutional review board. All participants provided written informed consent.

Subjects aged ≥18 years with a clinical diagnosis of T1D or T2D on a stable, subcutaneously administered basal-bolus insulin regimen were eligible to participate in the study. Subjects using continuous subcutaneous insulin infusion (CSII) were eligible, but automated insulin delivery was disabled at least three hours prior to dosing visits. These subjects continued to receive their basal insulin delivery via the pump, based on their programmed manual settings which were not changed. Exclusion criteria were the following: history of asthma, chronic obstructive pulmonary disease, or any other clinically important pulmonary disease (eg, cystic fibrosis, bronchopulmonary dysplasia); use of any medications to treat such conditions within the last year; significant congenital or acquired cardiopulmonary disease; history of smoking; respiratory tract infection within 14 days before screening; use of antiadrenergic drugs; women who were pregnant, breastfeeding, intend to become pregnant or of child-bearing potential, and not using adequate contraceptive methods; had a severe hypoglycemic event or episode of diabetic ketoacidosis within 90 days prior to screening.

Study Procedures

After informed consent, the study comprised four clinic visits, including a screening visit, two treatment visits, and a follow-up visit. Two individual doses of TI were administered during visits 2 (initial USPI dose 1) and 3 (higher dose 2) with a standardized meal challenge. Subjects completed the study within two weeks.

At visit 1, a forced expiratory volume in one second (FEV₁) measurement was taken for each subject. The dosing visits were done in a fasting state, with no drink or food for ≥6 hours before the clinic visit. Subjects were also instructed not to inject correction bolus insulin during the six hours immediately preceding each dosing visit but to continue their basal insulin regimen.

Fingerstick glucose was measured with CONTOUR® meters (Ascensia Diabetes Care, Parsippany, NJ, USA). At the dosing visits (visits 2 and 3), fasting glucose concentration was required to be between 100 and 200 mg/dL for starting the meal challenge. If the subject’s fasting glucose value was <100 mg/dL, the Investigator could provide 15 g of fast-acting carbohydrates and re-measure fingerstick glucose after 30 minutes to determine whether the subject should proceed to the meal challenge. Initial FEV₁ measurements were taken before dose 1 and dose 2 of TI.

Subjects were trained on proper inhalation technique using the BluHale^® Pro device and application along with the inhaler device. Fingerstick glucose values were measured 15 minutes before the expected TI dose. The meal consisted of one to two bottles of nutritional shake (Boost^®; Nestlé, Leeds, England) to be fully consumed within 15 minutes. Each nutritional shake bottle contained approximately 240 calories, 41 g carbohydrate, 10 g protein, and 4 g fat. Subjects were instructed to consume an identical amount of nutritional shake for both meal challenges. Fingerstick glucose measurements were obtained at 15, 30, 45, 60, 90, and 120 minutes after each TI dose. Only correction glucose could be administered for hypoglycemia within the two-hour period.

Dose 1 of TI was based on each subject’s current calculated prandial RAA dose for the carbohydrates to be consumed, which was converted according to Figure 1 of the current label. Dose 2 was based on each subject’s calculated prandial RAA dose for the carbohydrates to be consumed and then multiplied by 2 and rounded down to the nearest TI dose subject to Investigator’s discretion. TI is available in three color-coded cartridge dosages: 4U (blue), 8U (green), and 12U (yellow).

The FEV₁ measurement for each subject was taken at 120 minutes after each TI dose. All subjects returned to the clinic for a follow-up visit (visit 4), within 24 to 72 hours after dose 2, for safety assessments. A final FEV₁ measurement was taken for each subject, and adverse events (AEs), glycemic status, and general health were assessed.

The primary efficacy endpoint was change in fingerstick glucose from immediately before administering the dose of TI (T = 0; baseline) to 15, 30, 45, 60, 90 and 120 minutes after TI dosing.

Safety endpoints were event rates and incidence of level 1 (<70 mg/dL) and level 2 (<54 mg/dL) hypoglycemia; the incidence of severe hypoglycemia and other AEs; and change in FEV₁ from T = 0 to 120 minutes after TI dosing.

Statistical Analysis

Statistical analyses were performed using Minitab 20.2 or Excel 2016 at the time of the analysis. Summary tables (descriptive statistics and/or frequency tables) are provided for screening and/or baseline variables, efficacy variables, and safety variables. Continuous variables are described by descriptive statistics (n, mean, standard deviation). Frequency counts and percentage of subjects within each category are provided for categorical data. No imputation method was applied for any missing post-baseline values. The number and incidence of hypoglycemia episodes are also summarized. When not specifically associated with a prespecified study hypothesis, P values should be considered hypothesis-generating only; test methods are annotated within the relevant tables.

A paired t test was used to compare the means of the actual finger stick glucose measurements (postprandial glucose [PPG]) between the two doses of TI, as well as the means of the glucose excursion (ie, the change in glucose concentration from baseline; PPGE) between the two doses of TI. A P value <.05 was considered significant.

Results

A total of 20 subjects were enrolled (10 at each site). Subjects were mostly T1D and white with an average age of 51.2 years and well-controlled diabetes (with a mean A1c of 6.66% [Table 1]).

Table 1.

Subject Characteristics.

Subject characteristics	Total (n = 20)
Demography
Age, y
Mean ± SD	51.2 ± 13.90
Gender
Male	50% (10/20)
Female	50% (10/20)
Race and Ethnicity
White	75% (15/20)
Asian	5% (1/20)
Hispanic or Latino	20% (4/20)
Medical history
Glycated hemoglobin (%)
Mean ± SD	6.66 ± 0.84
Diabetes mellitus
Type 1	90% (18/20)
Type 2	10% (2/20)
Weight, kg
Mean ± SD (n)	77.5 ± 27.03
Body mass index
Mean ± SD (n)	26.9 ± 6.97
Insulin: Carb ratio—breakfast
Mean ± SD (n)	9.93 ± 3.48
Insulin: Carb ratio—lunch
Mean ± SD (n)	9.98 ± 3.32

Significantly smaller PPGE (ie, the change in glucose value from baseline; PPGE) were observed with dose 2 at 45, 60, 90, and 120 minutes compared with dose 1 (Figure 1). The average doses of TI were 6.4 and 11.4 units for dose 1 and dose 2, respectively.

Mean peak PPG (actual glucose value measured by finger stick; PPG) in the two-hour postprandial period was reduced by 49.3 mg/dL (from 228.6 mg/dL to 179.3 mg/dL) at 120 minutes (Table 2) with dose 2 of TI, compared with dose 1 of TI.

Table 2.

Postprandial Glucose (PPG) in the Two-Hour Postprandial Period for Dose 1 and Dose 2.

Time (min)	Dose 1 (n = 20)		Dose 2 (n = 20)		P value a
	Mean (mg/dL)	SD	Mean (mg/dL)	SD
0	140.0	19.8	142.4	26.5	.714
15	146.8	29.2	148.0	33.0	.882
30	165.9	39.5	164.3	43.9	.868
45	188.9	43.5	169.3	48.6	.075
60	204.2	44.7	176.6	52.8	.017
90	222.8	64.3	178.3	53.3	.002
120	228.6	77.3	179.3	61.8	<.001

^aP value calculated using paired t test.

Two hypoglycemia events were observed following dose 2 in a single subject, one level 1 (0.8%) and one level 2 (0.8%). Two incidents of cough in one participant were classified as mild treatment-emergent adverse events (TEAEs). No serious AEs or adverse events of special interest (AESIs) were observed. Changes in percent predicted FEV₁ after TI dosing showed no statistically significant difference in percent predicted FEV₁ for dose 1 (P = .737), dose 2 (P = .063), and combined doses.

Discussion

This study enrolled 20 adult subjects with T1D and T2D treated with basal-bolus therapy to assess the two-hour postprandial mean peak glucose concentration following a standardized meal when TI was administered using a higher dosing conversion algorithm. As expected, analyses demonstrated significant reductions in glucose concentrations at 45, 60, 90, and 120 minutes of postprandial periods, with one level 1 (<70 mg/dL) and one level 2 (<54 mg/dL) hypoglycemic event (in a single subject). Importantly, no serious AEs or AESIs were reported, with no significant difference in percent predicted FEV₁, after TI administration.

The new dose conversion algorithm is based, in part, on findings from two recent open-label, randomized, multicenter clinical studies.^21,25 Akturk et al ²⁵ assessed the efficacy of TI versus injected insulin aspart at meals in 60 adult T1D patients. Subjects in the TI group used the current recommended dose conversion at each meal but were advised to take supplemental TI at one and two hours after meals if indicated based on PPG values. Although use of TI demonstrated significant reductions in PPGE and hypoglycemia compared with insulin aspart at mealtime, subjects in the TI group administered the supplemental doses 27% of the time, suggesting that an upward adjustment in the dose conversion algorithm was needed.

Similar findings were recently reported by Grant et al ²¹ in a randomized, controlled, cross-over study in which 29 subjects received three different doses of TI and insulin lispro over a series of six euglycemic clamps. Investigators reported that treatment with TI produced a lesser total effect per unit insulin than insulin lispro due to its faster absorption and shorter duration of exposure. The difference required significantly higher TI doses to achieve the same total effect as insulin lispro, demonstrating that a unit-for-unit dose conversion is not appropriate.

A key strength of this study was the use of standardized meals with a known carbohydrate content, allowing for a direct comparison of the effects of TI at dose 1 versus dose 2. Although less postprandial hyperglycemia was demonstrated with the higher dosage, the broader applicability of these findings is limited by the small number of subjects and the short duration of postprandial evaluation that were included in this analysis.

Conclusions

This study demonstrated that administration of a dosing regimen for TI that doubled the calculated subcutaneously injected prandial dose of RAA immediately before a meal challenge resulted in significant reductions in glucose concentrations during the two-hour postprandial period, with no new safety concerns in subjects treated with basal-bolus therapy. While these findings suggest the modified conversion ratio might allow reduction in post-meal hyperglycemia, computer modeling or studies with larger cohorts will lead to further understanding as to the optimal conversion from RAA to TI.