Patient-Reported Outcomes in Adults with Type 2 Diabetes Using Mealtime Inhaled Technosphere Insulin and Basal Insulin Versus Premixed Insulin

Abstract

Aims:

This study assessed patient-reported outcomes in a multicenter study of adults with type 2 diabetes taking mealtime Technosphere^® inhaled insulin (MannKind Corp., Valencia, CA) and basal insulin (insulin glargine) or premixed aspart insulin 70/30.

Methods:

Subjects were 618 non-smoking adults with starting hemoglobin A1c >7.0%: 302 in the Technosphere+glargine (TI+G) arm and 316 in the biphasic rapid-acting insulin arm (premixed aspart insulin 70/30). Subjects (47% male; mean age, 56 years; mean duration of diagnosed diabetes, 13.4 years) completed a measure of health-related quality of life (the SF-36) and a measure of treatment satisfaction (the Inhaled Insulin Treatment Questionnaire [IITQ]) before starting insulin treatment and approximately 45 weeks later.

Results:

There were no significant changes in either treatment arm for SF-36 Physical or Mental Component Summary measures. IITQ Diabetes Worries declined significantly in the TI+G arm (P=0.008), and Perceptions of Insulin Therapy, Treatment Satisfaction, and Treatment Preference improved in both arms (all P<0.001); there were no significant between-arm differences in change on any of these measures.

Conclusions:

Treatment with inhaled Technosphere insulin was implemented without a negative impact on health-related quality of life and with a reduction in diabetes worries. Improvements in perceptions of insulin therapy, treatment satisfaction, and treatment preference did not differ from treatment with premixed aspart insulin.

Introduction

I nsulin is recognized as the most effective glucose-lowering agent,¹ and most patients with type 2 diabetes will eventually need insulin therapy because of progressive decline in endogenous insulin production.² Currently efforts to mimic normal endogenous insulin secretion involve subcutaneous injections of basal and bolus (mealtime) insulin,¹ but this approach remains imperfect, and glucose excursions are not adequately controlled. The relatively slow absorption and the extended duration of action for current mealtime insulins commonly result in an initial phase of postprandial hyperglycemia followed by a phase of hypoglycemia hours after they are injected.³ Weight gain is also a concern with subcutaneous mealtime insulin.⁴ Moreover, subcutaneous insulin injection at mealtimes represents a burden for patients, including the potential of pain, bruising, and risk of infection.

The ideal mealtime insulin would mimic the postprandial endogenous insulin response, reducing postprandial hyperglycemic excursions, hypoglycemia following peak insulin action, and weight gain resulting from glucose ingestion to control hypoglycemia. Inhaled Technosphere^® insulin (MannKind Corp., Valencia, CA) offers a new approach to administering bolus insulin that closely mimics endogenous insulin responses with a goal of improved glucose control.^5
–7 In the study reported here, some clinical outcomes were superior for patients taking inhaled Technosphere insulin in combination with basal insulin glargine (TI+G) compared with premixed biphasic rapid-acting insulin 70/30 (BPR 70/30).⁷ Briefly summarized, glycated hemoglobin (A1C) change in the TI+G arm (−0.68%, SE=0.077, 95% confidence interval=−0.83 to −0.53) was similar and non-inferior to that in the BPR 70/30 arm (−0.76%, SE=0.071, 95% confidence interval=–0.90 to −0.62); noninferiority was defined as between-group difference of less than 0.4%, and the difference between arms was 0.07% (SE=0.102, 95% confidence interval=–0.13 to 0.27). Yet, with similar overall glucose levels patients in the TI+G arm had significantly lower weight gain (0.9 kg, SD=0.3, 95% confidence interval=0.0 to 1.5) than those in the BPR 70/30 arm (2.5 kg, SD=0.4, 95% confidence interval=1.9 to 3.0), with a treatment difference of −1.6 kg (SD=0.4, 95% confidence interval −2.4 to −0.7). The incidence of total hypoglycemia was significantly lower in the TI+G arm (odds ratio=0.417, 95% confidence interval=0.303 to 0.573), as was the incidence of severe hypoglycemia (odds ratio=0.409, 95% confidence interval=0.215 to 0.780).

The ideal mealtime insulin would also relieve patients of the burden of injecting themselves, resulting in increased acceptance of insulin therapy. Patient acceptance is essential for the adoption of any new therapy, and acceptance is influenced by patients' perceptions of the effects of this therapy on health-related quality of life (HRQoL) and treatment satisfaction. Improving quality of life is the ultimate outcome for any therapy, and treatment satisfaction is associated with regimen adherence⁸ and with consequent clinical outcomes.^8

–11 Use of inhaled Technosphere insulin has been shown to improve perceptions of insulin therapy in insulin naïve patients.¹² But to date there are no reports comparing the effects of a regimen including inhaled Technosphere insulin and a comparator insulin regimen on patient-reported outcomes (PRO) such as HRQoL and treatment satisfaction. The current article reports PRO in a randomized, multicenter study of adults with type 2 diabetes taking either mealtime inhaled Technosphere insulin in combination with basal insulin or premixed biphasic insulin.

Subjects and Methods

This study is a 52-week randomized, open-label, parallel-group clinical trial in adult patients with type 2 diabetes previously treated with insulin, with or without oral agents, comparing safety, efficacy, and PRO of mealtime inhaled Technosphere insulin plus bedtime insulin glargine versus twice-daily biphasic aspart insulin. The Technosphere Insulin System is a mealtime insulin monomer human (recombinant DNA origin) powder that is inhaled into the deep lung using a pocket-sized, non-motorized inhaler.^13
–15

Subjects and study protocol

The recruitment disposition and treatment procedure have been described in detail elsewhere.⁷ The trial was conducted with non-smoking adults (18–80 years old) who had a A1C level of >7.0% and <11.0% on entry to the study. Prior to the study participants were taking two to three insulin injections daily and could be using oral medications other than secretagogues or α-glucosidase inhibitors. Study participants were enrolled and treated at hospitals in 10 countries. All study participants provided informed written consent before enrollment. Institutional review boards or ethics committees at clinical sites approved the study in accordance with the Declaration of Helsinki.

Study participants who were randomized to the TI+G arm received about half their dose as basal insulin. The remainder of their daily dose was given as inhaled insulin, distributed across main meals and individually adjusted in 15-U increments to a maximum of 90 U (15 U approximates 4 IU of rapid-acting analog; therefore doses were adjusted in 4-IU increments to a maximum of 24 IU), based on self-monitoring of blood glucose readings. Participants randomized to the BPR 70/30 arm received twice-daily premixed biphasic aspart insulin (a human insulin analog suspension containing 70% insulin aspart protamine suspension and 30% insulin aspart [recombinant DNA origin]) by subcutaneous injection.

The intent to treat population consisted of 302 participants in the TI+G arm and 316 participants in the BPR 70/30 arm. The characteristics of this population at baseline are described in Table 1. Respondents were primarily white, overweight, and in poor glycemic control.

Table 1.

Baseline Characteristics of Study Population

Characteristic	TI+glargine (n =302)	BPR 70/30 (n=316)
Gender [n (%)]
Male	153 (50.7)	137 (43.4)
Female	149 (49.3)	179 (56.6)
Race [n (%)]
Caucasian	202 (66.9)	215 (68.0)
Black	25 (8.3)	27 (8.5)
Asian	8 (2.6)	4 (1.3)
Hispanic	61 (20.2)	64 (20.3)
Other	6 (2.0)	6 (1.9)
Age (years)	55.9 (10.61)	55.9 (9.98)
Duration of diabetes (years)	13.0 (7.18)	13.7 (7.88)
Weight (kg)	88.28 (17.387)	85.80 (17.962)
BMI (kg/m²)	31.59 (4.849)	31.12 (4.908)
Baseline A1C (%)	8.7 (1.12)	8.7 (1.08)
Fasting plasma glucose (mg/dL)	169.3 (65.78)	176.7 (67.35)

Data are mean (SD) values unless indicated otherwise.

A1C, glycosylated hemoglobin; BMI, body mass index; BPR 70/30, premixed biphasic rapid-acting insulin 70/30; TI, Technosphere insulin.

PRO measures

In addition to measures of clinical outcomes, participants completed two PRO measures: the well-validated SF-36¹⁶ measure of HRQoL and the Inhaled Insulin Treatment Questionnaire (IITQ), a validated measure of treatment experience (the IITQ is available from the authors).¹⁷ Both PRO measures were obtained prior to treatment and approximately 14 and 45 weeks later. Participants who withdrew from the study were asked to fill out the questionnaires at the time of withdrawal. For subjects who withdrew these early termination questionnaires were used as the end of study values if available; otherwise, 14-week values were carried over. PRO analyses were performed for the intent-to-treat population.

The SF-36 instrument generates two composite scores: the Physical Component Summary (PCS) and the Mental Component Summary (MCS). The PCS measure incorporates items from four subscales (Physical Functioning, Role Physical, Bodily Pain, and General Health), and the MCS measure incorporates items from four subscales (Vitality, Social Functioning, Role Emotional, and Mental Health). Scores are derived from an empirical weighting method. These scores are norm-based with a mean of 50 and an SD of 10 in the normative population; scores range from 0 to 100, with higher scores indicating better quality of life. The QualityMetric Missing Data Estimation algorithms were used to score all SF-36 measures at all three administrations for every subject.

IITQ measures include Diabetes Worries (five questions concerning complications and blood sugar levels), Perceptions of Insulin Therapy (16 questions regarding convenience, comfort, ease of adherence, and perceived efficacy), Treatment Satisfaction (three questions assessing overall satisfaction, willingness to recommend, and desire to switch), Treatment Preference (one question comparing current treatment to past treatment), and Inhaler Performance (10 items concerning the device, the powder cartridge, and dosing). All IITQ items had a 6-point response scale (strongly disagree=1, disagree=2, mildly disagree=3, mildly agree=4, agree=5, strongly agree=6). IITQ items from each of the domains were combined into composite additive scales for analysis. Where indicated, items were reverse-scored so that all were in the direction of greater worries, more favorable attitudes toward insulin therapy and inhaler performance, or higher treatment satisfaction and preference for the medication used by the patient in the study.

Psychometric results for pretreatment IITQ scores from this study have been published elsewhere.¹⁷ Inter-item agreement (α) of baseline measures was acceptable: Diabetes Worries (0.68), Perceptions of Insulin Therapy (0.84), and Treatment Satisfaction (0.81). Test–retest reliability was acceptable for all measures: Diabetes Worries (0.77), Perceptions of Insulin Therapy (0.85), Treatment Satisfaction (0.85), and Treatment Preference (0.67). The Inhaler Performance items were administered only at follow-up and only to participants who had used inhaled insulin; inter-item agreement for the end of study administration was high (0.90).

Statistical analysis

Minimum levels of statistical significance for all analyses were set at P<0.05, two-tailed. Intent-to-treat analysis (with last observation carried forward) used mixed-effect models to estimate differences in mean group changes in PRO from baseline to Week 45 (adjusted for baseline scores); t tests assessed within-group changes from baseline to Week 45.

Results

HRQoL

In the intent-to-treat analysis there was no significant within-group change from baseline to week 45 in the MCS in either treatment group and no significant between-group difference in change (Table 2). There was no significant within-group change from baseline to week 45 in any of the MCS subscale scores in either group and no significant between-group difference in change (data not shown).

Table 2.

Change from Baseline to Week 45 in the Short-Form Health Survey SF-36 Mental Component Summary and Physical Component Summary (Intent-to-Treat Population)

	TI+glargine		BPR 70/30
Summary, time point, statistic	Reported value	Change from baseline	Reported value	Change from baseline
Mental Component Summary
Baseline
n	292		302
Mean	49.14		48.99
SD	10.653		11.074
Week 45
n	203	197	234	226
Mean	49.02	0.11	48.63	0.83
SD	10.876	10.357	11.200	10.446
Within-group comparison P value^a		0.8773		0.2344
Between-group comparison P value^{b c}				0.9854
Physical Component Summary
Baseline
n	293		301
Mean	46.62		46.80
SD	8.820		9.272
Week 45
n	201	196	234	225
Mean	46.07	−0.92	45.88	−0.43
SD	9.946	8.019	9.717	8.247
Within-group comparison P value^a		0.1105		0.4385
Between-group comparison P value^{b c}				0.3175

Within-group P values are derived from two-sided paired t tests at the significance level of 0.05.

Between-group P values are derived from an analysis of covariance model with terms for pooled site, treatment, and baseline score.

The P value was calculated with last observation carried forward.

BPR 70/30, premixed biphasic rapid-acting insulin 70/30; TI, Technosphere insulin.

There was no significant within-group change from baseline to week 45 in the PCS in either treatment group and no significant between-group difference in change (Table 2). Among PCS subscale scores the only significant within-group change from baseline to week 45 was a reduction in the Physical Functioning subscale score in the TI group (P=0.001), with no significant between-group difference in change (results not shown).

IITQ

Table 3 shows that in the intent-to-treat analysis there was a significant reduction in Diabetes Worries scores in the TI+G arm (P=0.008), whereas the change in the BPR 70/30 arm was not statistically significant. There were statistically significant improvements in Perceptions of Insulin Therapy, Treatment Satisfaction, and Treatment Preference scores (all P<0.001) in both treatment arms. There were no significant between-arm differences in change on any of these scores. Inhaler Performance scores (from inhaled insulin users at end of study only) were very positive (mean=5.01 from a maximum of 6, SD=0.69).

Table 3.

Change from Baseline to Week 45 for Inhaled Insulin Treatment Questionnaire (Intent-to-Treat Population)

	TI+glargine		BPR 70/30
Subscale, time point, statistic	Reported value	Change from baseline	Reported value	Change from baseline
Diabetes Worries
Baseline
n	300		313
Mean	4.95		5.06
SD	0.895		0.840
Week 45
n	206	204	234	232
Mean	4.80	–0.18	4.98	–0.11
SD	1.013	0.953	0.998	0.961
Within-group comparison P value^a		0.0083		0.0727
Between-group comparison P value^{b c}				0.2509
Perceptions of Insulin Therapy
Baseline
n	300		313
Mean	3.83		3.87
SD	0.751		0.793
Week 45
n	206	204	234	232
Mean	4.37	0.52	4.32	0.44
SD	0.768	0.861	0.837	0.772
Within-group comparison P value^a		<0.0001		<0.0001
Between-group comparison P value^{b c}				0.5468
Treatment Satisfaction
Baseline
n	299		312
Mean	3.68		3.75
SD	1.195		1.254
Week 45
n	206	204	234	231
Mean	4.99	1.36	4.91	1.14
SD	1.062	1.498	0.977	1.408
Within-group comparison P value^a		<0.0001		<0.0001
Between-group comparison P value^{b c}				0.1593
Treatment Preference
Baseline
n	299		309
Mean	3.92		4.02
SD	1.392		1.476
Week 45
n	204	202	231	226
Mean	4.94	1.00	4.88	0.85
SD	1.264	1.665	1.259	1.584
Within-group comparison P value^a		<0.0001		<0.0001
Between-group comparison P value^{b c}				0.5191

Within-group P values are derived from two-sided paired t tests at the significance level of 0.05.

Between-group P values are derived from an analysis of covariance model with terms for pooled site, treatment, and baseline score.

The P value was calculated for the intent-to-treat population with last observation carried forward.

BPR 70/30, premixed biphasic rapid-acting insulin 70/30; TI, Technosphere insulin.

Discussion

As reported earlier, participants who used mealtime inhaled Technosphere insulin and glargine experienced a statistically significant and non-inferior reduction in A1C, with significantly less weight gain and hypoglycemic events, compared with study participants who used biphasic aspart insulin. Participants who used inhaled Technosphere insulin also reported no negative impacts on HRQoL and significant improvements in worries about diabetes, perceptions of insulin therapy, treatment satisfaction, and treatment preference that were similar to the improvements among study participants who used biphasic aspart insulin.

Because weight gain was smaller and hypoglycemic events were less frequent in the TI+G arm, it appears that the PRO assessed in the current study were not sensitive to the statistically significant between-group differences in these clinical outcomes. The fact these outcomes (weight change and hypoglycemic events) were infrequently experienced by study participants might also have contributed to the fact that between-group differences in these outcomes were not associated with significant between-group differences in PRO. This is consistent with a report that clinical outcomes that patients experience most frequently (e.g., appetite control rather than weight change and self-monitoring of blood glucose results rather than A1C results) are the strongest predictors of PRO.¹⁸

Those PRO that did change during the trial all changed in a positive direction, with no significant difference in change between treatment groups. Diabetes worries decreased significantly among those using inhaled Technosphere insulin with glargine. In both treatment groups perceptions of insulin therapy and treatment satisfaction improved significantly, and the study treatment was significantly preferred to participants' prior treatment. Because there were no differences between treatment arms, these improvements could be attributed to a study effect; the additional attention and efforts to maximize treatment efficacy could have generated these outcomes. These outcomes could also be attributable to the reduction in A1C levels found in both treatment arms.

In an earlier report comparing active inhaled Technosphere insulin with placebo inhaled Technosphere insulin in insulin-naive subjects with type 2 diabetes, there was no significant change in any SF-36 factor or diabetes worries during the trial in either arm.¹² Perceptions of insulin therapy improved significantly during the trial in the active medication arm (effect size for composite measure=0.56, P=0.002) but not in the placebo arm, although there were no significant between-arm differences in change. The current study confirms those results, showing that initiation of treatment with inhaled Technosphere insulin occurred without deterioration in HRQoL and that treatment with Technosphere insulin was associated with significant improvement in perceptions of insulin therapy, treatment satisfaction, and treatment preference, similar to those for biphasic aspart insulin.

Strengths and limitations

The current study is the first to examine PRO associated with inhaled Technosphere insulin use in a study of long duration that included a commonly used comparator regimen, a large population of patients with type 2 diabetes, and validated measures of HRQoL and treatment experience. The study comparator regimen—bisphasic aspart insulin—was chosen because it is widely used by patients who have type 2 diabetes, but increasingly basal–bolus regimens are being used, and this regimen was not assessed in the current study. Basal–bolus therapy involves an increase in injection burden over the twice-daily injection regimen used as a comparator in this study, but it is not clear whether this would affect the treatment experience. Although participants were not blinded to treatment assignment, blinding would have made it impossible to assess impact of treatment modality on treatment perceptions, satisfaction, and preference.

Clinical implications

In this study results for PRO parallel those for A1c reported earlier;⁷ there were equivalent improvements in both treatment arms, a finding that replicates earlier research.¹⁹ The statistically significant advantages for the Technosphere Insulin System during this study in terms of weight and hypoglycemia were not reflected in patients' HRQoL or treatment satisfaction. It has been suggested that the availability of inhaled insulin may increase patients' willingness to initiate insulin therapy, but the evidence is equivocal,^19,20 and there is little compelling evidence to suggest that the availability of inhaled insulin will lead to increased insulin initiation. Although this study was not designed as a PRO non-inferiority trial, the results do suggest that it is possible to initiate basal–bolus therapy using the Technosphere Insulin System without decrement in patients' HRQoL or treatment satisfaction compared with a regimen of twice-daily injections. An earlier study of the effect of the availability of inhaled insulin found that multiple daily injection regimens were more common when inhaled insulin was available,¹⁹ so availability of the Technosphere Insulin System might enhance willingness to implement a true basal–bolus insulin regimen.

Research implications

Research suggests that many patients prefer inhaled insulin to subcutaneous insulin. During the open-label extension phase of a clinical trial most (85%) patients who received an inhaled insulin during the main phase of a clinical trial chose to continue doing so, and 75% of patients who received subcutaneous insulin during the main phase chose to switch to the inhaled insulin.²¹ In another study that used a more rigorous procedure for determining preference,¹⁹ 55% of insulin initiators chose an inhaled insulin when it was available as an alternative to insulin injections. Future studies should examine what types of patients prefer or benefit from inhaled insulin regimens.

Conclusions

Combined with the recently published report of clinical and safety outcomes, this study demonstrates that when used as a mealtime insulin in combination with bedtime insulin glargine, inhaled Technosphere insulin reduced A1C levels to a similar degree as biphasic insulin aspart, with less weight gain, fewer episodes of hypoglycemia, and similar positive effects on diabetes worries, perceptions of insulin therapy, treatment satisfaction, and treatment preference, without any deleterious effects on HRQoL.

Footnotes

Acknowledgments

This study was funded by an unrestricted grant to the authors from MannKind Corporation, which also provided access to the data. The authors are responsible for all aspects of manuscript preparation and submission.

Author Disclosure Statement

M.P. and R.R.R. received research funding and consultation fees from MannKind Corporation.

References

Nathan

, Buse

, Davidson

, Ferrannini

, Holman

, Sherwin

, Zinman

. Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the intiation, adjustment of therapy. A consensus statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care, 2009; 32:193–203.

Turner

, Cull

, Frighi

, Holman

. Glycemic control with diet, sulfonylureas, metformin, or insulin in patients with type 2 diabetes mellitus: progressive requirements for multiple therapies (UKPDS 49). UK Prospective Diabetes Study (UKPDS) Group. JAMA, 1999; 281:2005–2012.

Plank

, Siebenhofer

, Berghold

, Jeitler

, Horvath

, Mrak

, Pieber

. Systematic review and meta-analysis of short-acting insulin analogues in subjects with diabetes mellitus. Arch Intern Med, 2005; 165:1337–1344.

Holman

, Thorne

, Framer

, Davies

, Keenan

, Paul

. Levy JC; 4-T Study Group: Addition of biphasic, prandial, or basal insulin to oral therapy in type 2 diabetes. N Engl Med, 2007; 357:1716–1730.

Rave

, Heise

, Putzner

, Boss

. Coverage of postprandial blood glucose excursions with inhaled Technosphere Insulin in comparison to subcutaneously injected regular insulin in subjects with type 2 diabetes. Diabetes Care, 2007; 30:2307–2308.

Rosenstock

, Bergenstal

, DeFronzo

, Hirsch

, Klonoff

, Boss

, Kramer

, Petrucci

, Yu

, Levy

. the 0008 Study Group. Efficacy and safety of Technosphere inhaled insulin compared with Technosphere powder placebo in insulin-naïve type 2 diabetes suboptimally controlled with oral agents. Diabetes Care, 2008; 31:2177–2182.

Rosenstock

, Lorber

, Gnudi

, Howrad

, Bilheimer

, Chang

P-C

, Petrucci

, Boss

, Richardson

. Prandial inhaled insulin plus basal insulin glargine versus twice daily biaspart insulin for type 2 diabetes. Lancet, 2010; 375:2244–2253.

Charpentier

, Fleury

, Dubroca

, Vaur

, Clerson

. Electronic pill-boxes in the evaluation of oral hypoglycemic agent compliance. Diabetes Metab, 2005; 31:189–195.

Kelley

, Dempsey

. An evaluation of an insulin transfer programme delivered in a group setting. J Clin Nurs, 2007; 16:152–158.

10.

Tahrani

, Digwood

, Lee

, Moulilk

. Evaluation of glargine group-start sessions in patients with type 2 diabetes as a strategy to deliver the service. Int J Clin Pract, 2007; 61:329–335.

11.

Alvarez Guisasola

, Tofé Povedano

, Krishnaraja

, Lyu

, Mavros

, Yin

. Hyperglycaemic symptoms, treatment satisfaction, adherence and their associations with glycaemic goal in patients with type 2 diabetes mellitus: findings from the Real-Life Effectiveness and Care Patterns of Diabetes Management (RECAP-DM) Study. Diabetes Obes Metab, 2008; 10,Suppl 1:25–32.

12.

Peyrot

, Rubin

. Effect of Technosphere inhaled insulin on quality of life and treatment satisfaction. Diabetes Technol Ther, 2010; 12:49–55.

13.

Rave

, Heise

, Pfutzner

, Boss

. Coverage of postprandial glucose excursions with inhaled Technosphere insulin in comparison to subcutaneously injected regular human insulin in subjects with type 2 diabetes. Diabetes Care, 2007; 30:2307–2308.

14.

Steiner

, Pfutzner

, Wilson

, Harzer

, Heinemann

, Rave

. Technosphere Insulin—proof of concept study with a new insulin formulation for pulmonary delivery. Exp Clin Endocrinol Diabetes, 2005; 110:17–21.

15.

Pfutzner

, Forst

. Pulmonary insulin delivery by means of the Technosphere drug carrier mechanism. Expert Opin Drug Deliv, 2005; 2:1097–1106.

16.

Ware

, Sherbourne

. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care, 1992; 30:473–483.

17.

Rubin

, Peyrot

. Psychometric properties of an instrument for assessing the experience of patients treated with inhaled insulin: the Inhaled Insulin Treatment Questionnaire. Health Qual Life Outcomes, 2010; 8:32.

18.

Peyrot

, Rubin

. How does treatment satisfaction work? Modeling the determinants of treatment satisfaction and preference. Diabetes Care, 2009; 32:1411–1417.

19.

Bergenstal

, Freemantle

, Leyk

, Cutler

, Hayes

, Muchmore

. Does availability of AIR insulin increase insulin use and improve glycemic control in patients with type 2 diabetes? Diabetes Technol Ther, 11,Suppl 2:S-45–S-52.

20.

Yki-Järvinen

, Juurinen

, Alvarsson

, Bystedt

, Caldwell

, Davies

, Lahdenperä

, Nijpels

, Vähätalo

. Initiate Insulin by Aggressive Titration and Education (INITIATE): a randomized study to compare initiation of insulin combination therapy in type 2 diabetic patients individually and in groups. Diabetes Care, 2007; 30:1364–1369.

21.

Del Prato

, Blonde

, Martinez

, Goke

, Woo

, Millward

, Gomis

, Canovatchel

, Strack

, Lawrence

, Freemantle

. EXPERIENCE Trial Team. The effect of the availability of inhaled insulin on glycaemic control in patients with type 2 diabetes failing on oral therapy: the evaluation of Exubera as a therapeutic option on insulin initiation and improvement in glycaemic control in clinical practice. Diabet Med, 2008; 25:662–670.