14 CLINICAL STUDIES

The efficacy and safety of LEVEMIR given once-daily at bedtime or twice-daily (before breakfast and at bedtime, before breakfast and with the evening meal, or at 12-hour intervals) was compared to that of once-daily or twice-daily NPH insulin in open-label, randomized, parallel trials of 1155 adults with type 1 diabetes mellitus, 347 pediatric patients with type 1 diabetes mellitus, and 869 adults with type 2 diabetes mellitus. The efficacy and safety of LEVEMIR given twice-daily was compared to once-daily insulin glargine in an open-label, randomized, parallel trial of 320 patients with type 1 diabetes. The evening LEVEMIR dose was titrated in all trials according to pre-defined targets for fasting blood glucose. The pre-dinner blood glucose was used to titrate the morning LEVEMIR dose in those trials that also administered LEVEMIR in the morning. In general, the reduction in HbA1c with LEVEMIR was similar to that with NPH insulin or insulin glargine.

14.1 Clinical Studies in Adult Patients with Type 1 Diabetes

In a 16-week open-label clinical trial (Study A, n=409), adults with type 1 diabetes were randomized to treatment with either LEVEMIR at 12-hour intervals, LEVEMIR administered in the morning and bedtime or NPH insulin administered in the morning and bedtime. Insulin aspart was also administered before each meal. At 16 weeks of treatment, the combined LEVEMIR-treated patients had similar HbA1c and fasting plasma glucose (FPG) reductions compared to the NPH-treated patients (Table 7). Differences in timing of LEVEMIR administration had no effect on HbA1c, fasting plasma glucose (FPG), or body weight.

In a 26-week, open-label clinical trial (Study B, n=320), adults with type 1 diabetes were randomized to twice-daily LEVEMIR (administered in the morning and bedtime) or once-daily insulin glargine (administered at bedtime). Insulin aspart was administered before each meal. LEVEMIR-treated patients had a decrease in HbA1c similar to that of insulin glargine-treated patients.

In a 24-week, open-label clinical trial (Study C, n=749), adults with type 1 diabetes were randomized to once-daily LEVEMIR or once-daily NPH insulin, both administered at bedtime and in combination with regular human insulin before each meal. LEVEMIR and NPH insulin had a similar effect on HbA1c.

Table 7: Type 1 Diabetes Mellitus – Adult
	Study A		Study B		Study C
Treatment duration	16 weeks		26 weeks		24 weeks
Treatment in combination with	NovoLog (insulin aspart)		NovoLog (insulin aspart)		Human Soluble Insulin (regular insulin)
	Twice-daily LEVEMIR	Twice-daily NPH	Twice-daily LEVEMIR	Once-daily insulin glargine	Once-daily LEVEMIR	Once-daily NPH
Number of patients treated	276	133	161	159	492	257
HbA1c (%)
Baseline HbA1c	8.6	8.5	8.9	8.8	8.4	8.3
Adj. mean change from baseline	-0.8*	-0.7*	-0.6**	-0.5**	-0.1*	0.0*
LEVEMIR – NPH 95% CI for Treatment difference	-0.2 (-0.3, -0.0)		-0.0 (-0.2, 0.2)		-0.1 (-0.3, 0.0)
Fasting blood glucose (mg/dL)
Baseline mean	209	220	153	150	213	206
Adj. mean change from baseline	-44*	-9*	-38**	-41**	-30*	-9*

*From an ANCOVA model adjusted for baseline value and country.

**From an ANCOVA model adjusted for baseline value and study site.

14.2 Clinical Studies in Pediatric Patients with Type 1 Diabetes

Two open-label, randomized, controlled clinical trials have been conducted in pediatric patients with type 1 diabetes. One trial (Study D) was 26 weeks in duration and enrolled patients 6 to 17 years of age. The other trial (Study I) was 52 weeks in duration and enrolled patients 2 to 16 years of age. In both trials, LEVEMIR and NPH insulin were administered once- or twice-daily. Bolus insulin aspart was administered before each meal. In the 26-week trial, LEVEMIR-treated patients had a mean decrease in HbA1c similar to that of NPH insulin (Table 8). In the 52-week trial, the randomization was stratified by age (2-5 years, n=82, and 6-16 years, n=265) and the mean HbA1c increased in both treatment arms, with similar findings in the 2-5 year-old age group (n=80) and the 6-16 year-old age group (n=258) (Table 8).

Table 8: Type 1 Diabetes Mellitus – Pediatric
	Study D		Study I
Treatment duration	26 weeks		52 weeks
Treatment in combination with	NovoLog (insulin aspart)		NovoLog (insulin aspart)
	Once- or Twice Daily LEVEMIR	Once- or Twice Daily NPH	Once- or Twice Daily LEVEMIR	Once- or Twice Daily NPH
Number of subjects treated	232	115	177	170
HbA1c (%)
Baseline HbA1c	8.8	8.8	8.4	8.4
Adj. mean change from baseline	-0.7*	-0.8*	0.3**	0.2**
LEVEMIR – NPH 95% CI for Treatment difference	0.1 -0.1; 0.3		0.1 -0.1; 0.4
Fasting blood glucose (mg/dL)
Baseline mean	181	181	135	141
Adj. mean change from baseline	-39	-21	-10**	0**

*From an ANCOVA model adjusted for baseline value, geographical region, gender and age (covariate).

**From an ANCOVA model adjusted for baseline value, country, pubertal status at baseline and age (stratification factor).

14.3 Clinical Studies in Adult Patients with Type 2 Diabetes

In a 24-week, open-label, randomized, clinical trial (Study E, n=476), LEVEMIR administered twice-daily (before breakfast and evening) was compared to NPH insulin administered twice-daily (before breakfast and evening) as part of a regimen of stable combination therapy with one or two of the following oral antidiabetic medications: metformin, an insulin secretagogue, or an alpha–glucosidase inhibitor. All patients were insulin-naïve at the time of randomization. LEVEMIR and NPH insulin similarly lowered HbA1c from baseline (Table 9).

In a 22-week, open-label, randomized, clinical trial (Study F, n=395) in adults with type 2 diabetes, LEVEMIR and NPH insulin were given once- or twice-daily as part of a basal-bolus regimen with insulin aspart. As measured by HbA1c or FPG, LEVEMIR had efficacy similar to that of NPH insulin.

Table 9: Type 2 Diabetes Mellitus – Adult

	Study E		Study F
Treatment duration	24 weeks		22 weeks
Treatment in combination with	oral agents		insulin aspart
	Twice-daily LEVEMIR	Twice-daily NPH	Once- or Twice Daily LEVEMIR	Once- or Twice Daily NPH
Number of subjects treated	237	239	195	200
HbA1c (%)
Baseline HbA1c	8.6	8.5	8.2	8.1
Adj. mean change from baseline	-2.0*	-2.1*	-0.6**	-0.6**
LEVEMIR – NPH 95% CI for Treatment difference	0.1 (-0.0, 0.3)		-0.1 (-0.2, 0.1)
Fasting blood glucose1 (mg/dL)
Baseline mean	179	173	-	-
Adj. mean change from baseline	-69*	-74*	-	-

1Study F - Fasting blood glucose data not collected

*From an ANCOVA model adjusted for baseline value, country and oral antidiabetic treatment category.

**From an ANCOVA model adjusted for baseline value and country.

Combination Therapy with Metformin and Liraglutide

This 26-week open-label trial enrolled 988 patients with inadequate glycemic control (HbA1c 7-10%) on metformin (≥1500 mg/day) alone or inadequate glycemic control (HbA1c 7-8.5%) on metformin (≥1500 mg/day) and a sulfonylurea. Patients who were on metformin and a sulfonylurea discontinued the sulfonylurea then all patients entered a 12-week run-in period during which they received add-on therapy with liraglutide titrated to 1.8 mg once-daily. At the end of the run-in period, 498 patients (50%) achieved HbA1c <7% with liraglutide 1.8 mg and metformin and continued treatment in a non-randomized, observational arm. Another 167 patients (17%) withdrew from the trial during the run-in period with approximately one-half of these patients doing so because of gastrointestinal adverse reactions. The remaining 323 patients with HbA1c ≥7% (33% of those who entered the run-in period) were randomized to 26 weeks of once-daily LEVEMIR administered in the evening as add-on therapy (n=162) or to continued, unchanged treatment with liraglutide 1.8 mg and metformin (n=161). The starting dose of LEVEMIR was 10 units/day and the mean dose at the end of the 26-week randomized period was 39 units/day. During the 26-week randomized treatment period, the percentage of patients who discontinued due to ineffective therapy was 11.2% in the group randomized to continued treatment with liraglutide 1.8 mg and metformin and 1.2% in the group randomized to add-on therapy with LEVEMIR.

Treatment with LEVEMIR as add-on to liraglutide 1.8 mg + metformin resulted in statistically significant reductions in HbA1c and FPG compared to continued, unchanged treatment with liraglutide 1.8 mg + metformin alone (Table 10). From a mean baseline body weight of 96 kg after randomization, there was a mean reduction of 0.3 kg in the patients who received LEVEMIR add-on therapy compared to a mean reduction of 1.1 kg in the patients who continued on unchanged treatment with liraglutide 1.8 mg + metformin alone.

Table 10: Results of a 26-week Open-label Trial of LEVEMIR as Add-on to Liraglutide + Metformin Compared to Continued Treatment with Liraglutide + Metformin Alone in Patients Not Achieving HbA1c <7% After 12 Weeks of Metformin and Liraglutide

	Study H
	LEVEMIR + Liraglutide +Metformin	Liraglutide+ Metformin
Intent-to-Treat Population (N)a	162	157
HbA1c (%) (Mean)
Baseline (week 0)	7.6	7.6
Adjusted mean change from baseline	-0.5*	0*
Difference from liraglutide + metformin arm (LS mean)b 95% Confidence Interval	-0.5*** (-0.7, -0.4)
Percentage of patients achieving A1c <7%	43**	17**
Fasting Plasma Glucose (mg/dL) (Mean)
Baseline (week 0)	166	159
Adjusted mean change from baseline	-38*	-7*
Difference from liraglutide + metformin arm (LS mean)b 95% Confidence Interval	-31*** (-39, -23)

aIntent-to-treat population using last observation on study

bLeast squares mean adjusted for baseline value

*From an ANCOVA model adjusted for baseline value, country and previous oral antidiabetic treatment category.

**From a logistic regression model adjusted for baseline HbA1c.

***p-value <0.0001