Safety and efficacy of inhaled human insulin (Exubera®) during discontinuation and readministration of therapy in adults with type 1 diabetes: A 3-year randomized controlled trial

Methods Patients were randomized to receive basal insulin plus either pre-meal EXU ( n = 290) or a short-acting subcutaneous (SC) insulin ( n = 290) for 2 years (comparative phase), followed by 6 months of SC insulin (washout) and 6 months of their original therapy (readministration). Highly standardized lung function tests were performed throughout. Results Small treatment group differences favoring SC insulin in change from baseline forced expiratory volume in 1 s (FEV 1 ) and carbon monoxide diffusing capacity (DL CO ) occurred early and were non-progressive. These differences resolved during washout and recurred at the same magnitude during readministration. Both groups maintained glycemic control, and hypoglycemic event rates were similar. In the EXU group, insulin antibody (IAb) levels plateaued at 12 months, declined to near baseline levels during washout and increased during readministration to levels observed in the comparative phase. Conclusions FEV 1 and DL CO changes observed during discontinuation and readministration of EXU therapy are consistent with a reversible, non-progressive and non-pathological effect on lung function. EXU readministration is not associated with an augmented IAb response. Keywords Inhaled insulin FEV 1 DL CO A1C Hypoglycemic events Insulin antibodies 1 Introduction Clinical trials in adult patients with type 1 diabetes mellitus (T1DM) have demonstrated that inhaled human insulin (EXU; Exubera ® (insulin human [rDNA origin]) Inhalation Powder) is as effective and well tolerated as subcutaneous (SC) regular insulin when administered for up to 2 years [1–4] . Small but consistent treatment group effects in pulmonary function tests (PFTs) have been observed in patients administered EXU [1,2] . Studies designed to evaluate these effects in patients with T1DM using highly standardized PFTs [5,6] have confirmed the small treatment differences between EXU and SC insulin in lung function and have established that the differences develop early, are clinically insignificant and are non-progressive for up to 2 years of therapy [3,4,7] . The aim of the present analysis was to extend these pulmonary safety data by examining effects on lung function of (1) discontinuing EXU for a 6-month washout period following 2 years of EXU therapy, and (2) then readministering EXU treatment for 6 additional months. 2 Methods This study was originally designed as a 2-year, open-label parallel-group comparative respiratory safety study with a 6-month washout period, but the study protocol was amended to allow extended exposure to the initial randomized treatment following the washout phase. Results from the initial 2 years of comparative treatment with EXU and SC insulin have been published [4] . Results from the 6-month washout phase, when EXU therapy was discontinued and all patients received SC insulin, and a subsequent 6-month readministration phase, when patients again received their initial randomized treatment, are reported here. The study is ongoing and is being performed at 65 centers in the United States, Canada, Argentina, Mexico and Brazil. The study protocol was reviewed and approved by the institutional review boards of each participating center; all patients provided written informed consent. The study is being conducted in accordance with the principles of the Declaration of Helsinki. Patients with T1DM, aged 18–65 years, receiving a stable insulin regimen for at least 2 months, body mass index (BMI) ≤30 kg/m 2 , A1C 5.5–11% and fasting plasma C-peptide ≤0.2 pmol/ml were included in the study. Patients were excluded if they had brittle diabetes, recurrent severe hypoglycemia, poorly controlled asthma, significant chronic obstructive pulmonary disease or other respiratory disease (forced expiratory volume in 1 s [FEV 1 ] <70% of predicted; carbon monoxide diffusing capacity [DL CO ] <70% or >120% of predicted; total lung capacity [TLC] <70% or >130% of predicted), or had reported smoking in the previous 6 months. Prediction equations [8–10] were used to establish baseline predicted lung function for FEV 1 , DL CO and TLC; a 12% race adjustment in TLC and DL CO was applied for patients whose self-reported race was black. At screening, eligible patients all received SC insulin for 4 weeks and then were randomized to receive either pre-meal EXU or short-acting SC insulin (regular insulin, insulin lispro or insulin aspart). Subjects also received NPH insulin or Ultralente ® once or twice daily, or insulin glargine once daily. Randomization was performed using a computer-generated schedule. EXU was administered within 10 min before meals. The initial EXU dose was based on body weight, with subsequent adjustment to achieve blood glucose targets of 80–120 mg/dl before meals and 100–140 mg/dl at bedtime. The primary outcome measures were changes in FEV 1 and DL CO with respect to baseline during the comparative, washout and readministration phases of the study. Validated and highly standardized PFTs [5,6] were performed as per Skyler et al. [4] during the comparative phase and at washout months 1, 3 and 6, and at readministration months 1, 3 and 6. Baseline values were defined as the means of the values collected before the first dose of study drug after randomization. Safety was assessed by monitoring adverse events and clinical laboratory tests throughout the study. Serum samples for measurement of insulin antibodies (IAbs) were collected as per Skyler et al. [4] and at washout months 1, 3 and 6, and at readministration months 1, 3 and 6. During the comparative phase, efficacy assessments including A1C, fasting plasma glucose (FPG), and body weight were measured as per Skyler et al. [4] . In addition, A1C was measured at months 3 and 6 of the washout phase and months 1, 3 and 6 of the readministration phase, and FPG was measured at month 6 of the washout phase and months 1, 3 and 6 of the readministration phase. Hypoglycemia and severe hypoglycemia were defined as per Skyler et al. [4] and recorded along with insulin dose at each visit during the comparative and readministration phases only. Postprandial glucose measurements were not collected in this study. 2.1 Statistical analysis An analysis of covariance model by visit was used to estimate the mean change from baseline PFT for each treatment group for each visit, the treatment group difference in change from baseline for each visit and the corresponding 90% confidence interval (CI). The model included terms for treatment group, center, baseline PFT, age, gender and baseline height. Treatment group differences and corresponding two-sided 90% CIs for continuous secondary efficacy endpoints were estimated using a similar analysis of covariance model by visit with terms for treatment group, baseline and center. Analyses of PFT data collected during all 3 phases of the study were performed using the All Subjects Full Analysis Set (FEV 1 ), which included all patients who received at least one dose of study medication and had a baseline measurement and at least one postbaseline (FEV 1 ) measurement. Analyses of efficacy parameters (A1C, FPG, hypoglycemia and body weight) and insulin dose were performed using the Extension Full Analysis Set (A1C), which included all patients who received at least one dose of study medication and had a baseline measurement and at least one (A1C) measurement during the readministration phase. The safety population was used to report all adverse events and included all patients who received at least one dose of study drug. 3 Results A total of 582 patients were randomized, of which 222 in the EXU group and 229 in the SC insulin group completed the 2-year comparative treatment phase ( Fig. 1 ). Of these, 253 and 159 in the EXU group and 232 and 171 in the SC insulin group also completed the 6-month washout and readministration phases, respectively. The majority of patients who discontinued therapy during the 2-year comparative phase entered into the 6-month washout phase ( Fig. 1 ); however, these patients were not eligible for inclusion in the readministration phase. As previously reported [13] , demographic characteristics of the patients collected at baseline were well matched between groups. In brief, the majority of the study population was white (88.8%) and the mean A1C level at baseline was 7.37% in the EXU group and 7.41% in the SC insulin group. Observed FEV 1 and DL CO levels at baseline were 3.51 ± 0.76 l and 28.11 ± 6.21 ml/min/mmHg in the EXU group and 3.47 ± 0.78 l and 27.19 ± 6.40 ml/min/mmHg in the SC insulin group, respectively. 3.1 Pulmonary function During the 2-year comparative treatment phase, FEV 1 and DL CO declined from baseline in both treatment groups ( Fig. 2 A and B ). As described earlier [4] , treatment group differences in mean change from baseline in FEV 1 and DL CO were observed during the comparative phase that favored SC insulin. These treatment group differences were statistically significant but small, equal to 0.9% and 1.5% of baseline values for FEV 1 and DL CO , respectively. The difference between the two groups occurred during the first 3 months of treatment and did not progress beyond this time [4] . When EXU was discontinued and replaced with SC insulin during the washout phase, treatment group differences in FEV 1 and DL CO resolved completely within 1 month ( Fig. 2 ). The adjusted treatment group differences in FEV 1 during the washout were −0.013, −0.005 and −0.011 l at 1, 3 and 6 months. The equivalent treatment group differences in DL CO were 0.126, 0.191 and 0.057 ml/min/mmHg. All treatment group differences during the washout were statistically insignificant. When EXU was readministered for 6 months following the washout phase, treatment group differences in FEV 1 and DL CO recurred by 1 month ( Fig. 2 ). The magnitude of EXU-associated changes in FEV 1 and DL CO were similar to those observed in the original comparative phase. The adjusted mean change from baseline to readministration month 6 in FEV 1 was −0.169 and −0.138 l for EXU and SC insulin, respectively, with a treatment group difference of −0.031 l in favor of SC insulin (90% CI, −0.062 to −0.001). Similarly, for DL CO , the mean change from baseline to readministration month 6 was −1.468 and −0.980 ml/min/mmHg for EXU and SC insulin, respectively, with a treatment group difference of −0.488 ml/min/mmHg in favor of SC insulin (90% CI, −0.912 to −0.064). 3.2 Adverse events Both treatments were well tolerated during the three phases of the study, and the pattern of adverse events observed after readministration of EXU was similar to that reported after initiation [13] . Overall, 15 (5.2%) and four (1.4%) patients in the EXU and SC insulin groups, respectively, discontinued treatment during either the comparative or readministration phases because of adverse events. Ten of the adverse events resulting in discontinuation in the EXU group were judged to be treatment related and included cough ( n = 7), dyspnea ( n = 2) and raised insulin antibody levels ( n = 1). One treatment-related adverse event resulted in discontinuation in the SC insulin group (severe hypoglycemia). At baseline, median IAb levels were 4.50 and 4.15 μU/ml for EXU and SC insulin, respectively. Median IAb levels increased with respect to baseline in the EXU group, but not in the SC insulin group, during the 2-year comparative phase, reaching a peak at 12 months ( Fig. 2 C). In the EXU group, the median change from baseline was +128.20 μU/ml at 12 months, declining to a change from baseline of +52.95 μU/ml at 24 months. During the washout phase, median IAb levels declined further in the EXU group to levels near that of the SC insulin group ( Fig. 2 C). When EXU therapy was readministered, median IAb levels again increased to levels similar to those observed in the original comparative phase. IAb levels did not correlate with A1C, FPG, hypoglycemia or changes in FEV 1 or DL co at the end of either the comparative phase or the readministration phase. 3.3 Efficacy Glycemic control was maintained in both treatment groups throughout the three phases ( Fig. 3 A ). The adjusted mean treatment difference (EXU-SC insulin) in A1C at the end of the 6-month readministration period was 0.338 (90% CI, 0.218–0.457). It is likely that this small, clinically insignificant difference in glycemic control was driven by the notably lower mean EXU doses used at the start of the readministration phase than those used at the end of the comparative phase in the same group of subjects ( Table 1 ). Of EXU-treated patients, 67.9% (112 of 165) had taken lower total doses of EXU at readministration month 1 than at month 24 of the comparative phase, while the SC insulin-treated patients had stable doses of insulin throughout the study ( Table 1 ). As previously reported [1,2,4] , decreases in FPG were consistently greater with EXU compared with SC insulin throughout the comparative phase ( Fig. 3 B); the mean adjusted treatment difference at 2 years was −19.02 mg/dl (90% CI, −32.29 to −5.75). Mean FPG levels increased when EXU was discontinued but decreased again when EXU was readministered ( Fig. 3 B). There was significantly less weight gain with EXU versus SC insulin during the comparative phase (mean adjusted treatment difference −1.24 kg [90% CI, −2.08 to −0.40]) ( Fig. 3 C). This advantage was lost during the washout phase but returned and increased somewhat by readministration month 6 (mean adjusted treatment difference −0.58 kg [90% CI, −1.05 to −0.10]). During the comparative phase, the hypoglycemia event rates were comparable in the EXU and SC insulin groups (4.3 and 4.1 events/subject-month, respectively). Hypoglycemia was reported by 98.9% and 100.0% of the EXU and SC insulin-treated patients, respectively. The hypoglycemic event rate declined slightly during the readministration phase, with 3.7 and 3.1 events/subject-month, respectively, for EXU and SC insulin. Hypoglycemia was reported by 75.7% and 78.0% of the EXU and SC insulin-treated patients during the readministration phase, respectively. The event rate of severe hypoglycemia was lower in the EXU group compared with the SC insulin group throughout the comparative phase (1.7 and 4.5 events/100 subject-months, respectively). Severe hypoglycemia was reported by 20.9% and 31.2% of patients in the EXU and SC insulin groups, respectively. During the readministration phase, the rate was 3.2 and 3.8 events/100 subject-months, respectively, for EXU and SC insulin. Severe hypoglycemia was reported by 7.9% and 9.1% of patients in the EXU and SC insulin groups, respectively, during this time. 4 Conclusions Small treatment differences in lung function between EXU and SC insulin were observed in adult T1DM patients during the comparative phase of this ongoing study. These treatment differences developed early (within 3 months), were non-progressive, and were clinically insignificant [4] . Subsequent discontinuation and readministration of EXU therapy did not engender an augmented effect on these lung function changes. Cessation of EXU therapy for 6 months was associated with a rapid resolution of the treatment group differences in FEV 1 and DL CO . When EXU therapy was resumed, FEV 1 and DL CO treatment group differences similar to those observed in the original comparative phase re-emerged. These data are consistent with physiological, and not pathological, changes in lung function associated with EXU therapy and are strong evidence that irreversible lung injury did not occur during 2 years of continuous EXU therapy. FEV 1 and DL CO declined in both the EXU and SC treatment groups over the course of the comparative phase, consistent with the influence of age on lung function. FEV 1 and DL CO have been observed to decline by about 1–2% and 1.3% per year, respectively, in non-diabetic male and female non-smokers [8,11] . Epidemiological studies suggest that there is a modest increase in the annual rate of decline in FEV 1 in some patients with diabetes [12] , and preliminary studies have reported histopathological changes in diabetic lungs [13,14] . Although respiratory dysfunction in most patients with diabetes is subclinical and rarely the presenting complaint [15] , further studies are warranted to clarify the relationship of lung function to the development of diabetes, and to assess whether diabetes-related pulmonary dysfunction is potentially reversible following the re-establishment of normoglycemia. Increased IAb levels were detected in the EXU group within a few weeks of commencing therapy, peaking at 12 months before reaching a plateau. This finding is consistent with previous studies [16,17] . During the washout phase, IAb levels declined to near baseline levels and increased again when therapy was resumed to levels comparable to those seen after the original initiation of EXU treatment. The IAb response was not augmented by restarting EXU therapy, and there were no clinical manifestations of the increased antibodies. This conclusion is supported by previous studies that have shown that IAb formation in response to EXU did not appear to have any clinical relevance, or any correlation with glycemic control, insulin dose, hypoglycemic episodes, changes in FEV 1 or tolerability [7,16,17] . As in the comparative phase of this study [4] , glycemic control was equally maintained in the EXU and SC insulin groups throughout the washout and readministration phases, indicating that cessation and resumption of EXU therapy does not have any detrimental impact on glycemic control. Though efficacy was a secondary endpoint in this study, and this study was not designed to test for non-inferiority in A1C, a small, clinically non-meaningful difference in A1C levels was observed at the end of the readministration period. This was likely driven by the notably lower mean EXU doses used at the start of the readministration phase than those used at the end of the comparative phase. This observation is perhaps not surprising since specific instructions on EXU dosing at the start of the readministration phase were not provided in the study protocol, and the first clinic visit took place 1 month after the start of the readministration phase. Despite this, the percentage of patients reporting hypoglycemia and severe hypoglycemia in the EXU group during the readministration phase was slightly less than in the SC insulin group. Weight gain is commonly seen when patients are started on SC insulin. Ten years after the start of the United Kingdom Prospective Diabetes Study, patients who were assigned SC insulin gained more (4 kg) than those assigned the conventional therapy of chlorpropamide (2.6 kg) or glibenclamide (1.7 kg) [18] . Given that the majority of patients with type 2 diabetes are overweight or obese [19,20] , additional weight gain is often a concern for patients during insulin therapy [21] , and evidence suggests that fear of weight gain may reduce adherence to diabetes medication [22] . A recent pooled analysis revealed significantly less weight gain at 6 months with regimens including EXU therapy compared with SC insulin-only regimens in adult patients with type 1 or type 2 diabetes [23] . The present study confirms that these differences in weight gain were sustained throughout the comparative phase and subsequent readministration phase, despite producing comparable levels of glycemic control. In conclusion, the present results demonstrate that FEV 1 and DL CO changes observed during discontinuation and readministration of EXU therapy are consistent with a reversible, non-progressive and non-pathological effect on lung function in adults with type 1 diabetes. In addition, readministration of EXU results in an IAb response comparable to that seen after the original initiation of EXU treatment. Conflict of interest JSS was a recipient of research grants from Pfizer, and has been an advisor to Pfizer, Novo-Nordisk, Kos, MannKind, and Eli Lilly concerning inhaled insulins. PAH has served as a speaker and on advisory boards for Pfizer and has participated in a number of multi-center clinical trials sponsored by Pfizer. LJ received a grant for research from Pfizer to do clinical trails using inhaled insulin and has been a medical advisor for Pfizer for which she received an honorarium. SK, AK, RJR, JR, and WD are employees of Pfizer. Acknowledgments This study was supported by a research grant from Pfizer Inc. Editorial support was provided by Tom Claus, PhD, of PAREXEL and was funded by Pfizer Inc. We thank all the patients, investigators (see Supplement ) and coordinators who took part in this study. The current status of EXU, including the change in labeling regarding lung carcinoma, is summarized in Supplement . Appendix A Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.diabres.2008.08.008 . Appendix A Supplementary data References [1] T. Quattrin A. Belanger N.J.V. Bohannon S.L. Schwartz Efficacy and safety of inhaled insulin (Exubera) compared with subcutaneous insulin therapy in patients with type 1 diabetes. Results of a 6-month, randomized, comparative trial Diabetes Care 27 2004 2622 2627 [2] J.S. Skyler R.S. Weinstock P. Raskin J.-F. Yale E. Barrett J.E. Gerich Use of inhaled insulin in a basal/bolus insulin regimen in type 1 diabetic subjects: a 6-month, randomized, comparative trial Diabetes Care 28 2005 1630 1635 [3] P. Norwood R. Dumas W. Cefalu J.F. Yale R. England R.J. Riese Randomized study to characterize glycemic control and short-term pulmonary function in patients with type 1 diabetes receiving inhaled human insulin (Exubera) J. Clin. Endocrinol. Metab. 92 2007 2211 2214 [4] J.S. Skyler L. Jovanovic S. Klioze J. Reis W. Duggan Two year safety and efficacy of inhaled human insulin (Exubera ® ) in adult patients with type 1 diabetes mellitus Diabetes Care 30 2007 579 585 [5] R.L. Jensen J.G. Teeter R.D. England H.M. Howell H.J. White E.H. Pickering Sources of long-term variability in measurements of lung function: Implications for interpretation and clinical trial design Chest 132 2007 396 402 [6] R.L. Jensen J.G. Teeter R.D. England H.M. Howell H.J. White E.H. Pickering Instrument accuracy and reproducibility in measurements of pulmonary function Chest 132 2007 388 395 [7] J.G. Teeter R.J. Reise Dissociation of lung function changes with humoral immunity during inhaled human insulin therapy Am. J. Respir. Crit. Care Med. 119 2006 184 190 [8] J.L. Hankinson J.R. Odencrantz K.B. Fedan Spirometric reference values from a sample of the general U.S. population Am. J. Respir. Crit. Care Med. 159 1999 179 187 [9] R.O. Crapo A.H. Morris P.D. Clayton C.R. Nixon Lung volumes in healthy nonsmoking adults Bull. Eur. Physiopathol. Respir. 18 1982 419 425 [10] A. Miller J.C. Thornton R. Warshaw H. Anderson S.A. Teirstein I.J. Selikoff Single breath diffusing capacity in a representative sample of the population of Michigan, a large industrial state Am. Rev. Respir. Dis. 127 1983 270 277 [11] G. Viegi D.L. Sherrill L. Carrozzi F. Di Pede S. Baldacci F. Pistelli An 8-year follow-up of carbon monoxide diffusing capacity in a general population sample of northern Italy Chest 120 2001 74 80 [12] W.A. Davis M. Knuiman P. Kendall V. Grande T.M.E. Davis Glycemic exposure is associated with reduced pulmonary function in type 2 diabetes. The Fremantle Diabetes Study Diabetes Care 27 2004 752 757 [13] B. Weynand A. Jonckheere A. Frans J. Rahier Diabetes mellitus induces a thickening of the pulmonary basal lamina Respiration 66 1999 14 19 [14] J. Farina V. Furio M.J. Fernandez-Acenero M.A. Muzas Nodular fibrosis of the lung in diabetes mellitus Virchows Arch. 427 1995 61 63 [15] C.C. Hisa P. Raskin Lung function changes related to diabetes mellitus Diabetes Technol. Ther. 9 Suppl. 1 2007 S73 S82 [16] S.E. Fineberg T. Kawabata D. Finco-Kent C. Liu A. Krasner Antibody response to inhaled insulin in patients with type 1 or type 2 diabetes. An analysis of initial phase II and III inhaled insulin (Exubera) trials and a two-year extension trial J. Clin. Endocrinol. Metab. 90 2005 3287 3294 [17] T. Heise S. Bott C. Tusek J.-A. Stephan T. Kawabata D. Finco-Kent The effect of insulin antibodies on the metabolic action of inhaled and subcutaneous insulin. A prospective randomized pharmacodynamic study Diabetes Care 28 2005 2161 2169 [18] UK Prospective Diabetes Study (UKPDS) Group Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33) Lancet 352 1998 837 853 [19] C. Daousi I.F. Casson G.V. Gill I.A. MacFarlane J.P.H. Wilding J.H. Pinkney Prevalence of obesity in type 2 diabetes in secondary care: association with cardiovascular risk factors Postgrad. Med. J. 82 2006 280 284 [20] International Obesity Task Force, 2003, Diabetes Atlas second edition, Available from http://www.eatlas.idf.org/Obesity_and_type_2_diabetes . Last accessed 14th June, 2007. [21] R.R. Henry B. Gumbiner T. Ditzler P. Wallace R. Lyon H.S. Glauber Intensive conventional insulin therapy for type II diabetes Diabetes Care 16 1993 21 31 [22] A. Farmer A.-L. Kinmonth S. Sutton Measuring beliefs about taking hypoglycaemic medication among people with type 2 diabetes Diabetic Med. 23 2006 265 270 [23] P. Hollander A. Krasner S. Klioze P. Schwartz W. Duggan Body weight changes associated with insulin therapy: a retrospective pooled analysis of inhaled human insulin (Exubera ® ) versus subcutaneous insulin in five controlled phase 3 trials Diabetes Care 30 2007 2508 2510