Evolution of a Pulmonary Insulin Delivery System (Exubera) for Patients With Diabetes

Abstract

Many patients with diabetes fail to meet recommended glycemic goals regardless of the recognition of optimal glycemic control as a key component for improving clinical outcomes and quality of life in patients with diabetes. Patient- and physician-related barriers to the adoption of insulin therapy include fear and anxiety about injecting insulin, concerns about side effects, and personal health beliefs in regard to the use of insulin. There is an unmet need for an alternative insulin therapy that provides optimal glycemic control, is well tolerated, and improves patient adherence. Of the several inhaled insulin devices that are in various stages of development, the Exubera (INH) formulation is the first to be approved for use in the United States and in Europe. Exubera is a novel, rapid-acting inhaled human insulin formulation that has been developed for prandial insulin use. Clinical studies have shown that INH consistently improves glycemic control, in combination with longer-acting subcutaneous (SC) insulin regimens in patients with type 1 or type 2 diabetes, or is used to supplement or replace oral antidiabetic therapy in patients with type 2 diabetes. INH has demonstrated long-term safety and tolerability, with a risk for hypoglycemia similar to that of SC insulin, and no clinically meaningful changes in pulmonary function have been noted with its use. Patients treated with INH in clinical studies reported high levels of satisfaction with treatment, and many patients with diabetes choose inhaled insulin when it is offered as a treatment option. Taken together, these findings suggest that INH represents an important new development in the treatment of diabetes that may improve glycemic control in many patients with diabetes.

Introduction

Regardless of the recognition of optimal glycemic control as a key component for improving clinical outcomes and quality of life in patients with diabetes, a large proportion of patients with diabetes fail to meet the recommended glycemic goals. In one survey conducted by the American Association of Clinical Endocrinologists (AACE), data were collected from more than 157,000 people with type 2 diabetes over a 2-year period (2003-2004), and were measured against the AACE target glycated hemoglobin (HbA1c) goal of ≤ 6.5%.[1] Two thirds of patients (67%) failed to meet that goal.[1] These findings are consistent with those of earlier cross-sectional surveys, such as the Third National Health and Nutrition Survey (NHANES III, conducted 1994–1998) and NHANES 1999–2000, which revealed that 57% to 63% of patients with diabetes failed to achieve the American Diabetes Association (ADA) recommended target HbA1c of < 7%.[2] Clearly, improved strategies are urgently needed for the optimal management of diabetes.

Both patient- and physician-related barriers contribute to the delay in initiation of insulin therapy, which has a negative impact on the achievement and maintenance of optimal glycemic control. The reluctance to start insulin treatment encompasses a wide range of patient misconceptions about injection anxiety, risk for side effects and weight gain, and feelings of being more seriously ill or having failed to control the disease.[3–5] Physician-related barriers to treatment also contribute to delayed insulin use; for example, physicians may be concerned about the potential side effects of insulin (ie, weight gain, hypoglycemia) as well as the challenges involved in educating their patients about proper insulin administration techniques.[6,7] Results from a recent cross-sectional study indicated that many healthcare providers delay insulin therapy until absolutely necessary, but that the delay is significantly less likely when insulin is viewed as more efficacious and the provider believes that the patient is more adherent to treatment.[4] Taken together, these findings suggest that elimination of the patient- and physician-related barriers to insulin therapy would have a substantial benefit in terms of improving glycemic control and patient outcomes.

Similarly, patient satisfaction with diabetes treatment may improve glycemic control, which has been associated with better quality of life.[8] Thus, diabetes treatments that eliminate treatment barriers and are well accepted by patients are expected to improve outcomes and quality of life. There is an unmet need for an alternative insulin therapy that provides optimal glycemic control, is well tolerated, and improves patient adherence.

Pulmonary Delivery as a Route for Insulin Administration

Since the discovery of insulin over 80 years ago, investigators have sought to develop a pulmonary route of insulin delivery. The pulmonary route of administration offers several advantages. First, the lung has a large surface area for drug absorption, ranging from 100 to 140 m².[9,10] In addition, the alveolar epithelium has permeability that allows for rapid absorption of solutes. Because the mucociliary clearance of the alveolar lung tissue is slower than that of the bronchiolar tissues, the alveoli provide a greater opportunity for the absorption of larger molecules (eg, insulin).[9]

Studies have shown that particle size should be between 1 and 3 micrometers (mcm) in diameter for optimal deposition in the lung, and that dry powder formulations can deliver more active drug in a single inhalation than liquid aerosol formulations.[11] Patient-controlled variables (eg, inhalation flow rate, inhaled volume, and duration of inhalation) also need to be controlled for optimal deep-lung insulin delivery.[9–11]

Although several inhaled insulin devices are in various stages of development, the Exubera (Pfizer, New York, NY) formulation and delivery system is the first inhaled insulin to be approved for use in the United States and in Europe. Additional agents in phase 2 and 3 of development include the Technosphere Insulin (TI) System (MannKind Corporation, Valencia, California), the Lilly/Alkermes inhaled insulin system, and the AERx insulin Diabetes Management System (AERx iDMS; Novo Nordisk).

The Exubera Device

Exubera (INH) is a novel inhaled insulin delivery system that has been developed for prandial insulin use in patients with type 1 or type 2 diabetes. The INH system has 3 main components: a dry powder formulation of human insulin, packaged in foil blisters for long-term stability; the insulin inhaler device with a holding chamber that facilitates appropriate and consistent insulin delivery; and an insulin or INH release unit that enables the insulin powder to be released from the foil blisters into the device-holding chamber.[12] Each inhaler activation produces minute particles (1-5 mcm in diameter, with an average of 3 mcm) of a rapid-acting, dry powder, human recombinant insulin that are inhaled as a dry cloud from the holding chamber.[12] The insulin formulation is packaged into blister packets, each containing either 1 or 3 mg of insulin, with a single 1-mg dose delivering the equivalent of approximately 3 IU of subcutaneous (SC) insulin.[12] Patients may be required to use multiple blister packs to achieve the desired dose, but should combine 1-mg and 3-mg blister packs so that the least number of blister packs per dose are taken (eg, a 4-mg dose should be administered as one 1-mg blister pack and one 3-mg blister pack).

The Pharmacokinetics and Pharmacodynamics of INH

The pharmacokinetics and pharmacodynamics of INH have been evaluated in an open-label, randomized, 3-way crossover euglycemic clamp study in 17 healthy male volunteers who received each of the following treatments in random order: inhaled INH 6 mg; insulin lispro 18 units SC; or regular human insulin 19 units SC.[13] Glucose infusion rates (GIR) and serum insulin concentrations were monitored over the following 10-hour period to allow comparison of the time-action profiles of each insulin formulation.

Pharmacokinetics

INH demonstrated a significantly more rapid mean time to maximal concentration (T_max) than that of regular SC insulin (55 minutes vs 148 minutes, respectively, P < .001).[13] However, the mean maximal serum insulin levels (C_max) were similar for both INH and regular SC insulin (66.9 mircroU/mL vs 61.0 microU/mL, respectively). Compared with regular SC insulin, the relative bioavailability of INH was 18% during the first hour after administration and 9% over the entire 10-hour study period, reflecting a rapid increase in serum insulin levels during this period.[13]

Gelfand and colleagues[14] conducted a 4-way, randomized-sequence, crossover study involving 2 inhalations of INH and 2 SC injections of regular insulin in 16 insulin-naive patients with type 2 diabetes. Analysis of postprandial glucose measurements and area under the concentration time curve (AUC) at various time points indicated that intrasubject variability was insignificant in all 16 patients, and that the pharmacokinetics of INH were equivalent to those of regular SC insulin in terms of reproducibility.[14]

A second study was conducted to ensure pharmacokinetic reproducibility in obese patients with type 2 diabetes.[15] In this randomized sequence, crossover design study, 10 men received 2 inhalations of insulin and 2 SC injections of regular insulin. As in the previous pharmacokinetic study, the pharmacokinetics and pharmacodynamics of INH showed very little intrasubject variability.[15]

The pharmacokinetics of INH have not been fully evaluated in individuals with pulmonary disease, such as mild-to-moderate asthma or chronic obstructive pulmonary disease (COPD), precluding its use in this patient population until these studies have been completed.

Pharmacodynamics

Consistent with the pharmacokinetic study results, INH demonstrated a significantly faster onset of action than both regular SC insulin and SC insulin lispro, as indicated by the shorter mean time to half-maximal effect: 32 minutes vs 48 minutes (P < .001) and 41 minutes (P < .05), respectively.[13] The mean time to maximal effect of INH was comparable to that of insulin lispro (143 minutes vs 137 minutes), but shorter than that of regular insulin (193 minutes; P < .01). Finally, the mean duration of the metabolic activity (time to late half-maximal effect) of INH was 387 minutes, significantly longer than that of insulin lispro (313 minutes, P < 0.01), and comparable to that of regular insulin (415 minutes).[13] The mean duration of metabolic activity was calculated as the time to late half-maximal effect and was assessed as the time to half of GIR_max after GIR_max (t_{GIR late 50%}).

In summary, INH has a more rapid onset of action than subcutaneously administered regular human insulin or insulin lispro, with a duration of action exceeding that of insulin lispro and comparable to regular insulin. In addition, INH has shown very little intrasubject pharmacokinetic variability, even in obese patients. Taken together, these characteristics suggest that INH is well suited for prandial insulin supplementation in patients with diabetes.

Clinical Efficacy of INH

Clinical studies have shown that INH consistently improves glycemic control in patients with type 1 or type 2 diabetes,[16–19] and consistent with our shorter-term studies, 2 recent studies have demonstrated that INH provides sustained glycemic control with better fasting plasma glucose (FPG) levels and less weight gain than SC insulin when given to type 1[20] (75.1–75.9 kg vs 73.8–75.8 kg) or type 2[21] (87.1–88.8 kg vs 88.4–91.4 kg) diabetes patients over a 2-year period. Moreover, when used in combination with longer-acting SC insulin, INH reduces the number of daily injections needed while providing similar glycemic control to that of standard basal-bolus insulin regimens.

Type 1 Diabetes

Two 24-week, open-label, randomized, multicenter studies have compared preprandial administration of INH in combination with longer-acting SC insulin with standard basal-bolus insulin regimens in patients with type 1 diabetes.[18,22] In the first of these studies by Quattrin and colleagues,[18] patients were randomly assigned to receive treatment with either preprandial INH plus bedtime ultralente insulin (n = 170) or 2–3 daily SC injections of regular and NPH insulin (n = 164). Mean decreases in HbA1c at week 24 were comparable for both treatment groups, with similar proportions of patients in each group achieving HbA1c levels < 7% (Table). Treatment with INH also resulted in a greater mean reduction in FPG (35-mg/dL reduction vs 10 mg/dL with the standard SC insulin regimen) and postprandial glucose concentrations (PPG; −30 mg/dL vs +1 mg/dL, respectively).[18]

Table 1.

Exubera (INH) Is Effective for Glycemic Control in Patients With Type 1 or Type 2 Diabetes^*[16–18, 22, 24–26]

			Outcome (%)
Study	Treatment	Study Design	Mean Reduction in HbA1c	Patients With HbA1c < 7%	Patients With HbA1c < 6.5%†
Type 1 diabetes
Skyler et al, 2005	INH (preprandial) plus twice-daily NPH insulin SC (n = 163)	24 weeks; mean baseline HbA1c was 8.0% and 7.9%, respectively	0.3	23.3	NR
	Regular insulin (preprandial) SC plus twice-daily NPH insulin SC (n = 165)		0.1	22	NR
Quattrin et al, 2004	INH plus bedtime ultralente (n = 170)	24 weeks; mean baseline HbA1c was 8.1% in both groups	0.2	15.9	NR
	2-3 daily injections of regular/NPH insulin SC (n = 164)		0.4	15.5	NR
Type 2 diabetes
Hollander et al, 2004	INH (preprandial) plus bedtime ultralente (n = 149)	24 weeks; mean baseline HbA1c was 8.1% and 8.2%, respectively; patients had received prior insulin therapy	0.7	46.9	28.7
	≥ 2 daily injections of regular/NPH insulin SC (n = 150)		0.6	31.7	17.2
Weiss et al, 2003	INH plus prestudy OAs (n = 32)	12 weeks; patients had HbA1c > 8% despite OA therapy	2.3	34	NR
	Prestudy OAs alone (n = 36)		0.1	0	NR
DeFronzo et al, 2005	INH (preprandial) (n = 76)	12 weeks; patients had HbA1c > 8% despite diet and exercise regimen	2.3	44	28
	Rosiglitazone 4 mg twice daily (n = 69)		1.4	18	7.5
Rosenstock et al, 2005	INH (n = 104)	12 weeks; patients had HbA1c 8% despite prior dual OA therapy with an insulin sensitizer and secretagogue	1.4	17	NR
	INH plus dual OAs (n = 103)		1.9	32	NR
	Dual OAs alone (n = 99)		0.2	1	NR

^*All studies were open-label, randomized, multicenter trials.

^†American College of Endocrinology/American Association of Clinical Endocrinologists recommended goal

INH = Exubera; HbA1c = glycated hemoglobin; NR = not reported; NPH = neutral protamine Hagedorn; SC = subcutaneous; OAs = oral antihyperglycemic agents

In the second study by Skyler and colleagues,[22] patients were randomly assigned to receive treatment with either preprandial INH plus morning and evening SC injections of NPH insulin (n = 163) or preprandial regular SC insulin plus morning and evening NPH insulin injections (n = 165). As in the previous study, mean decreases in HbA1c at week 24 were comparable for both treatment groups, with similar proportions of patients in each group achieving HbA1c levels < 7% (Table). Patients in the group receiving INH demonstrated greater mean reductions in FPG than those receiving the regular/NPH SC insulin regimen (−35 mg/dL vs +4 mg/dL, respectively), whereas mean reductions in PPG were comparable for both treatment groups (−21 mg/dL vs −14 mg/dL, respectively).[22]

Type 2 Diabetes

Several clinical studies have also demonstrated that INH provides effective glycemic control in patients with type 2 diabetes, in combination with longer-acting insulins or oral antihyperglycemic agents (OAs), or alone in patients who have not achieved adequate glycemic control with prior OA therapy.

INH Plus Longer-acting Insulins

In a 12-week study conducted by Cefalu and colleagues,[23] patients (N = 26) who were switched to INH in combination with bedtime ultralente SC insulin injections had significantly improved mean HbA1c concentrations compared with baseline (mean reduction, 0.7%), having received standard SC insulin regimens (ie, 2–3 insulin injections daily) prior to study enrollment. These findings were confirmed in a larger, 24-week, phase 3 study conducted by Hollander and colleagues[17] in patients who had previously received multiple daily insulin injections (≥ 2 injections daily).

Patients were randomly assigned to 6 months of treatment with preprandial INH plus bedtime ultralente insulin (n = 149) or a conventional SC insulin regimen that included regular and NPH insulin (n = 150).[17] In this study, the INH regimen provided a mean reduction in HbA1c comparable to that of regular/NPH insulin (0.7% vs 0.6%, respectively; Table). Of note, significantly more patients receiving the INH regimen were able to achieve HbA1c concentrations < 7% (46.9% vs 31.7% of patients receiving the regular/NPH SC insulin regimen; odds ratio 2.27, 95% confidence interval [CI],1.24–4.14) or < 6.5% (28.7% vs 17.2%, respectively; Table).[17,24] A significantly greater decrease in mean FPG was seen with the INH-containing regimen, with an adjusted mean difference between treatments of −16 mg/dL (95% CI, −27 mg/dL to −5 mg/dL). The adjusted mean difference in PPG concentrations was also numerically in favor of INH (−9 mg/dL, 95% CI, −17 mg/dL to 8 mg/dL).[17]

INH Plus OAs

In a randomized, controlled study conducted by Weiss and colleagues,[25] patients (N = 68) with type 2 diabetes who had OA therapy were treated for 12 weeks with either their prestudy OA (sulfonylurea and/or metformin) therapy or prestudy OA therapy plus INH. The addition of INH to the existing OA regimen improved glycemic control compared with continuing the existing OA regimen alone (Table). Mean HbA1c concentration was reduced by 2.3% in the group receiving INH, compared with only 0.1% in the group receiving OAs alone (P < .001), and 34% of patients in the INH group achieved HbA1c < 7%, compared with 0% of patients receiving OAs alone. INH treatment was also associated with a significantly greater reduction in FPG, with a difference of −61 mg/dL between adjusted mean changes from baseline for each treatment group (P < .001).[25] The postprandial increase in plasma glucose was significantly lower in patients receiving INH plus OAs than in those receiving OAs alone (P = .02).[25]

In another 12-week, randomized, comparative study, patients who were unable to achieve glycemic control (ie, HbA1c remained > 8%) with diet and exercise alone were treated for 3 months with either INH (n = 76) or rosiglitazone (n = 69), in conjunction with a regimen of diet and exercise.[16] INH provided improved glycemic control compared with rosiglitazone (Table). The absolute reduction in HbA1c was greater with INH than with rosiglitazone (−2.3% vs −1.4%), and significantly more patients achieved the HbA1c goals recommended by the ADA and AACE. Mean final HbA1c concentrations were 7.2% with INH and 8.0% with rosiglitazone, respectively.

The efficacy of INH has also been evaluated in combination with dual OA therapy. In a 12-week, randomized, comparative study, patients (N = 309) with type 2 diabetes who had baseline HbA1c levels of 8% to 11% and were receiving dual OA therapy (with both an insulin secretagogue and insulin sensitizer) were randomly assigned to treatment with either preprandial INH added to existing OA therapy (n = 103), preprandial INH alone (n = 104), or continued OA therapy alone (n = 99) for 12 weeks.[26] Baseline mean HbA1c values were 9.2%, 9.3%, and 9.3% in each of the 3 treatment groups, respectively.

Consistent with the results of the previously described studies, INH improved overall glycemic control and HbA1c levels when added to or substituted for dual OA therapy (Table).[26] Mean HbA1c decreased by 1.9 percentage points in the group treated with INH and OAs, by 1.4 percentage points in the group receiving INH alone, and by 0.2 percentage points in patients continuing dual OA therapy. A larger proportion of patients receiving INH were able to achieve HbA1c concentrations < 7%: 32.0% of those receiving INH plus OAs, and 16.7% of those receiving INH alone achieved this goal compared with only 1% of patients receiving dual OA therapy alone (Table).[26] Furthermore, significantly greater reductions in FPG and PPG levels were seen in the group receiving INH. The adjusted mean difference in FPG between INH plus OAs and OAs alone was −53 mg/dL (95% CI, from −66 mg/dL to −44 mg/dL), whereas the adjusted mean difference in PPG between these 2 groups was −76 mg/dL (95% CI, from −93 mg/dL to −58 mg/dL). The adjusted mean difference in FPG between the groups receiving INH alone and OAs alone was smaller: −24 mg/dL (95% CI, from −36 mg/dL to −11 mg/dL), whereas the adjusted mean difference in PPG between these 2 groups was −62 mg/dL (95% CI, from −79 mg/dL to −45 mg/dL).[26]

Ongoing and Future Studies

Patients with type 1 diabetes are currently being recruited for a study to determine simple, safe, and effective guidelines for the titration of the dose of inhaled insulin. Additionally, future trials comparing INH with insulin analogs or trials that evaluate the combination of INH with long-acting insulin analogs might better define the role of INH in the treatment of diabetes.

Patient-Reported Outcomes With INH

The clinical benefits of INH inhaled insulin therapy extend to improved treatment satisfaction compared with SC insulin (in both type 1 and type 2 diabetes) or OAs (in type 2 diabetes). Rosenstock and colleagues[27] performed a pooled analysis of patient satisfaction data from the two 12-week parent studies and 1-year extension studies in patients with type 1 (n = 70) or type 2 (n = 51) diabetes who received either INH or SC insulin. Of the 60 patients who received INH in the parent studies, 51 (85%) chose to continue with the same treatment in the extension studies compared with only 13 of the 61 patients (21%) who received SC insulin. Of note, 75% of patients who were receiving SC insulin elected to switch to INH during the study extension compared with only 13% who chose to switch from INH to SC insulin.

After 1 year, patients receiving INH reported greater mean improvements in overall satisfaction with treatment (37.9% vs 3.1% with SC insulin; P < .01) and in ease of use (43.2% vs −0.9% with SC insulin; P < .01).[27] It should be noted, however, that the INH treatment group had increased to 95 patients (vs only 17 patients receiving SC insulin), because most of the patients originally treated with SC insulin switched to INH when given the opportunity.

Freemantle and colleagues[28] examined how the availability of inhaled insulin affected theoretical treatment choices in 779 patients with type 2 diabetes that was poorly controlled (ie, HbA1c > 8%) despite dietary measures and treatment with up to 3 OAs. Patients received information about currently available treatment options (OAs and/or SC insulin; n = 388) or current treatments and INH inhaled insulin (n = 391), and were then asked to choose treatment.

Of the patients offered INH, 43% (169 of 391 patients) opted for an insulin-containing regimen compared with only 15% (60 of 388 patients) of those who were offered standard available therapies only (P < .0001). Furthermore, 43% of the patients offered standard therapies chose not to make any changes to their existing therapy regardless of having poor glycemic control. Of interest, 35% (138 of 391 patients) of patients who were offered INH chose it as their preferred treatment option.[28] The results of this study showed that patients would choose insulin if available in an inhaled formulation, suggesting an opportunity to improve glycemic control with insulin.

Long-term Use of INH Is Safe and Well Tolerated

INH has been shown to be safe and well tolerated with long-term use in several different clinical trials.[16–18, 22,23,25,26,29,30]

Hypoglycemia

Overall, the incidence and severity of hypoglycemia with INH are similar to that seen with SC human insulins. There are no published studies currently available comparing INH with insulin analogs (eg, insulin lispro); however, studies are under way to enable efficacy and safety comparisons between INH and insulin analogs in the future. Although there has been a trend of lower fasting blood sugar levels with INH, there have been no reports in the literature that indicate an increased incidence of nocturnal hypoglycemia with INH. In the 3 studies comparing INH with standard SC insulin regimens in patients with type 1 diabetes, the overall risk for hypoglycemia was slightly lower for patients receiving INH. In the study by Quattrin and colleagues,[18] the incidence of hypoglycemia was 8.6 events/patient-month with INH compared with 9.0 events/patient-month with the SC insulin regimen (relative risk [RR] = 0.96). Similar results were seen in the study by Skyler and colleagues[22] (9.3 events/patient-month vs 9.9 events/patient-months with the SC insulin regimen; RR = 0.89). In the study comparing INH with an SC insulin regimen in patients with type 2 diabetes, the overall incidence of hypoglycemia was 1.4 events/patient-months vs 1.6 events/patient-months with the SC insulin regimen (RR = 0.89).[17] The incidence of severe hypoglycemic events was low in all of these studies (ie, ≤ 6.5 events/100 patient-months), and was not significantly different between treatment groups.[17,18,22]

In the studies comparing INH with OA therapy in patients with type 2 diabetes, the overall incidence of hypoglycemia was generally low, with ≤ 1.7 episodes/patient-months reported.[16,23,25,26] Although the rates of hypoglycemia episodes were higher with INH than with OA therapy, rates of hypoglycemia were considered comparable to those seen with SC insulin. Given the improved HbA1c values with INH vs OAs, it is not unexpected to see a slightly higher rate of hypoglycemia in the INH group.[16,25,26]

Pulmonary Function

Numerous clinical trials have shown that INH is associated with small changes in pulmonary function (forced expiratory volume in 1 second [FEV₁] and carbon monoxide diffusing capacity [DL_CO]) that are not progressive, are reversible, and have not been associated with clinical changes.[12, 16–18, 22,23,25] In clinical studies of INH in patients with type 1 or type 2 diabetes, the average mean treatment group differences in the FEV₁ were approximately 40 mL over 2 years from a baseline of 3 L.[31] These changes occurred early (by the third month of treatment) and did not progress over the remaining 2-year follow-up. Further, in patients with type 2 diabetes, these changes reversed to that of comparator treatment group levels 6 weeks after discontinuation of INH. The reversibility of pulmonary function changes after long-term treatment with INH has not been studied in patients with type 1 diabetes. In type 1 diabetes, treatment group differences in the mean change from baseline in DL_CO after 2 years of treatment were approximately 0.5 mL/minute/mm Hg in patients, favoring SC insulin over INH, and approximately 0.1 mL/minute/mm Hg in type 2 diabetes in favor of INH. Additional, comprehensive, long-term studies are needed to verify pulmonary safety of INH use greater than 2 years and to determine any long-term risks.

Mild-to-moderate cough has been reported in approximately 21% to 31% of patients receiving INH; however, the incidence and prevalence of cough generally decreased over time, and few patients (approximately 1%) have discontinued treatment due to cough.[12, 16–18, 22,23,25]

Because smoking has been shown to increase the permeability of the alveolar-capillary barrier,[32,33] inhaled insulin dose requirements may be lower in smokers. Thus, the effects of smoking cessation on the absorption of INH have been compared in smoking (n = 38) and nonsmoking volunteers (n = 30).[34] At baseline, the smokers had 5-fold higher peak insulin concentrations, 3-fold higher total insulin exposure, and a more rapid time to peak insulin concentration than their nonsmoking counterparts (31 minutes vs 53 minutes in nonsmokers). Three weeks after quitting smoking, however, peak insulin concentrations and total insulin exposure had decreased to 49% and 59% of the levels seen while they were still smoking. After 13 weeks, the changes in inhaled insulin absorption had been partially reversed by quitting smoking.[34] Although further study is necessary to determine the clinical implications of these findings, patients should be encouraged to quit smoking prior to initiation of therapy with INH, as patients who were current smokers or who had smoked in the previous 6 months were excluded from participation in the clinical study program.

Insulin Antibodies

Another potential concern with INH includes the development of increased levels of insulin antibodies compared with SC insulin use. The insulin antibodies associated with INH are structurally identical (primarily immunoglobulin [Ig]G) to those seen in response to SC insulin administration. Long-term studies have shown that insulin antibody levels plateau by 12 months of treatment, and that insulin antibody titers rapidly decline after discontinuation of INH.[29] Of interest, no correlation has been observed between insulin antibody concentrations and HbA1c concentrations, and FPG concentrations, PPG concentrations, the incidence of hypoglycemia, insulin dose or serum concentration, the incidence of allergic and/or respiratory adverse events, or lung function.[29,30]

Patients With Lung Disease

The use of INH is not recommended for patients with underlying lung diseases (eg, asthma or chronic obstructive pulmonary disease [COPD]) because the safety and efficacy of INH in this population have not been established, and is contraindicated in patients with unstable or poorly controlled lung disease because of the potential for wide variations in lung function to affect the absorption of INH and subsequently increase the risk for hypoglycemia or hyperglycemia.[31] A study in nondiabetic subjects[31] showed that the absorption of INH was 20% lower in those with mild asthma (without bronchodilator treatment) compared with those without asthma, whereas the absorption was approximately 2-fold higher in nondiabetic subjects with COPD compared with those without COPD. The administration of a bronchodilator (ie, albuterol) 30 minutes before INH resulted in a 25% to 50% mean increase in insulin AUC and C_max in nondiabetic subjects with mild (n = 36) and moderate (n = 31) asthma.

Elderly Population

There do not appear to be any differences in the pharmacokinetics of INH in patients over the age of 65 years compared with their younger counterparts.[31] Combined safety data from controlled phase 2/3 studies[31] in which INH was administered to patients with type 1 or type 2 diabetes (n = 1975; most had type 2 diabetes) included 266 patients ≥ 65 years of age, and 30 patients were ≥ 75 years of age. There were no differences in changes in HbA1c or the rate of hypoglycemia according to age.

Other Inhaled Insulin Systems in Development

There are a few additional inhaled insulin systems that are earlier in their development process than INH. In one study,[35] either TI or SC insulin was administered to patients (N = 12) with type 2 diabetes, and the relationship between time, insulin concentration, and the glucose elimination rate (GER) was evaluated during a euglycemic clamp study. Patients achieved a greater maximal concentration of insulin at a more rapid rate with TI compared with SC insulin. Maximal GER values were achieved faster with TI, but once a maximal insulin effect occurred, the concentration-effect relationship to GER was similar for TI and SC insulin. The efficacy and safety of TI were assessed in a randomized, double-blind, placebo-controlled study[36] in patients with type 2 diabetes (N = 119) inadequately controlled on diet or oral agents. TI significantly improved HbA1c from baseline compared with placebo (mean change from baseline, −0.72% vs −0.31%, P = .0016), and no severe hypoglycemia was observed in the TI group. There were no differences between groups in pulmonary function (ie, FEV₁, DL_CO, and total alveolar volume), and no induction of insulin antibodies in the TI group during the 12-week study.

Delivery of human insulin inhalation powder (HIIP) with the Lilly/Alkermes inhaled insulin system had a similar rapid onset, a significantly prolonged duration of action (P < .001), and was tolerated as well as SC insulin lispro in an open-label, randomized, 7-period, euglycemic glucose clamp crossover study[37] in healthy subjects (N = 20). In a randomized, open-label, crossover study,[38] patients with type 1 diabetes received HIIP via the Lilly/Alkermes system or SC insulin + insulin glargine once daily for 12 weeks. Treatment with HIIP appeared to be as safe and as effective as SC insulin in the maintenance of HbA1c levels. Eighty percent of patients (94 of 119) with type 1 diabetes preferred HIIP for mealtime insulin over SC insulin in a randomized, crossover study[39] with the SF-36 Vitality Scale, subscales of the Diabetes Symptom Checklist-Revised, and the Insulin Delivery System Questionnaire at baseline, crossover, and end-of-study scores. A patient-reported (N = 119) outcome study[40] demonstrated that patient satisfaction was improved with HIIP compared with SC insulin, on the basis of significant improvements in treatment satisfaction and greater insulin delivery system satisfaction.

The onset and duration of action of inhaled insulin via the AERx iDMS was compared with SC insulin aspart and SC human regular insulin in a single-center, open-label, 3-period, crossover study[41] in patients (N = 15) with type 1 diabetes. Results from this study demonstrated that inhaled insulin via the AERx iDMS has an onset of action similar to SC insulin aspart, but faster than human regular insulin; however, the duration of action was similar to regular human insulin, but longer than insulin aspart. In patients with type 2 diabetes (N = 107) receiving evening NPH insulin, preprandial inhaled insulin administered via AERx iDMS (n = 54) resulted in similar glycemic control as preprandial SC insulin (n = 53; 7.84% ± 0.77% vs 7.76% ± 0.77%, P = .60) in a randomized, 12-week, open-label, parallel, multicenter, international study.[42] Fasting serum glucose levels were significantly lower for the AERx iDMS group than the SC insulin group (8.9 ± 3.8 mmol/L vs 10.8 ± 3.7 mmol/L, P = .01) at 12 weeks. There were no significant differences between the groups for treatment-emergent adverse events, including pulmonary function. The median total insulin antibody level increased in the AERx iDMS group (6% to 35%) but remained unchanged in the SC group (10% to 9%). There were no correlations between changes in total insulin antibody levels and metabolic control or insulin dose.

Conclusions

Reducing the barriers to insulin therapy has the potential to improve clinical outcomes and quality of life in patients with diabetes. The key barriers to the adoption of insulin therapy are patient- and physician-related barriers, which include fear and anxiety about injecting insulin, concerns about side effects, and personal health beliefs with regard to the use of insulin. New treatment approaches, such as inhaled insulin, are aimed toward eliminating some of these barriers to the currently available regimens.

Several inhaled insulin systems are currently being developed for the treatment of patients with type 1 or type 2 diabetes, of which INH is the first and only system approved for use in the United States and in Europe. Clinical studies have shown that INH consistently improves glycemic control, whether added to longer-acting SC insulin regimens in patients with type 1 or type 2 diabetes or used to supplement or replace OA therapy in patients with type 2 diabetes. INH has also demonstrated long-term safety and tolerability, with a risk for hypoglycemia similar to that of SC insulin, and no clinically important changes in pulmonary function were noted. Patients treated with INH in clinical studies reported high levels of satisfaction with treatment, and many patients with diabetes choose inhaled insulin when it is offered as a treatment option. These findings, taken together, suggest that INH provides an important new approach in the treatment of diabetes – one that offers additional choices to patients, and consequently, may improve diabetes control and the lives of many patients with diabetes.

Funding Information

Pfizer Inc.