The Effect of Liraglutide Added to U-500 Insulin in Patients with Type 2 Diabetes and High Insulin Requirements

Introduction

O bese patients with type 2 diabetes and severe insulin resistance often require extremely high insulin doses (>100 units daily). However, treatment with high-dose insulin therapy often leads to weight gain, which in turn worsens insulin resistance. This then leads to further increases in insulin requirement and persistent hyperglycemia, with the end result of a vicious cycle of poor metabolic control, ever-increasing insulin doses, and continuing weight gain. Thus, the development of a type 2 diabetes treatment regimen that promotes weight loss, or ameliorates insulin-induced weight gain, would be extremely beneficial.

Obese, insulin-resistant patients are ideal candidates for treatment with glucagon-like peptide 1 (GLP-1) receptor agonists, which improve hyperglycemia even in insulinopenic patients by reducing glucagon secretion, slowing gastric emptying, delaying carbohydrate absorption, and increasing satiety, leading to reduced caloric intake. ¹ Combined GLP-1 receptor agonist treatment with exenatide and insulin has been shown to have an insulin-sparing effect in insulin-treated patients with type 2 diabetes as well as a weight loss benefit in several small uncontrolled studies. Furthermore, these studies report a low incidence of severe hypoglycemia attributed to premeal insulin dose reduction prior to initiation of exenatide therapy in these studies. ² However, combination treatment with GLP-1 receptor agonists and insulin is not an approved use of these compounds owing to a lack of prospective randomized trials of this combination therapy. Furthermore, no studies of combined treatment specifically with the long-acting human GLP-1 analog liraglutide (Novo Nordisk, Bagsvaerd, Denmark) and insulin have been published to date.

We examined the effects of adding liraglutide to intensive high-dose insulin therapy in 15 patients with type 2 diabetes treated with concentrated U-500 insulin by either multiple daily injections or continuous subcutaneous insulin infusion. Specifically, we examined the effect of adding liraglutide to U-500 insulin on hemoglobin A1c (HbA1c), total daily insulin dose (TDID), weight, and incidence of hypoglycemia.

Subjects and Methods

In this retrospective, observational study, 15 obese patients with insulin-treated type 2 diabetes using U-500 insulin (average insulin dose, 192 ± 77 units/day; range, 80–325 units/day) in combination with liraglutide for at least 12 weeks were identified at routine office visits at Mountain Diabetes and Endocrine Center in Asheville, NC. The data were obtained by chart review of these patients from office visits that took place between March 1 and October 15, 2010. The cohort consisted of eight men and seven women (average age, 56 ± 7.2 years) who had a mean baseline HbA1c of 8.48 ± 0.84%, a mean body mass index of 45.5 ± 6.4 kg/m², and a mean initial weight of 300.9 ± 55.7 lbs (136.8 ± 25.3 kg). All patients were liraglutide-naive and had been on a stable regimen of intensive therapy (multiple daily injections or continuous subcutaneous insulin infusion) with U-500 insulin, with or without oral hypoglycemic agents or exenatide, for at least 6 months, to be included in the study cohort. Liraglutide was then added to long-term background U-500 insulin therapy. The only concomitant glucose-lowering therapies these patients used were metformin and/or exenatide, as all subjects had advanced type 2 diabetes (mean duration, 17.2 years) and had failed sulfonylureas many years prior. Also, no patients were treated with thiazolidinediones in the previous 6 months as most subjects had not tolerated these agents in combination with high-dose insulin therapy because of edema, and thus our physicians do not routinely combine U-500 insulin with these agents. Upon addition of liraglutide, no new oral agents were added or dose changes of oral agents were made; exenatide was discontinued upon initiation of liraglutide in the two patients treated with exenatide at baseline. The baseline characteristics of the study cohort are shown in Table 1.

Liraglutide was initiated at a dose of 0.6 mg injected subcutaneously once per day and increased to 1.2 mg/day after 1 week. Nine patients (60%) underwent a subsequent increase in liraglutide to 1.8 mg/day to optimize glycemic control. Seven patients (40%) were determined to have an optimal glycemic response with achievement of target HbA1c on 1.2 mg of liraglutide daily and were maintained on this dose in combination with high-dose insulin therapy. Insulin doses were reduced by 0–30% upon initiation of liraglutide based upon the prescribing physician's judgment. The insulin reduction was customized subsequently based on each patient's self-monitored glucose levels. Patients continued their usual diet and exercise regimens as well as any concomitant glucose-lowering medications. Liraglutide and insulin dosing, whereas generally consistent between patients, reflected real-world practice where treatment was customized based on the clinician's judgment, and not performed according to a specific protocol.

Patients were seen in follow-up in 4–12 weeks after the initiation of liraglutide based on the physician's discretion. Seven patients were seen in follow-up at 4–6 weeks; all 15 patients were seen in follow-up at 12 weeks after initiation of liraglutide. TDID and weight were obtained at 4–6 weeks for the seven patients who had follow-up visits in that interval and at 12 weeks after initiation of liraglutide therapy for all patients; insulin doses were titrated to maintain euglycemia while avoiding hypoglycemia.

Paired two-tailed t tests were performed using XLSTAT 2009 (Windows Edition) (Microsoft, Redmond, WA) to assess change in HbA1c, TDID, and weight from baseline to 12 weeks.

Results

The results of adding liraglutide to U-500 insulin at 12 weeks on HbA1c, TDID, and weight are shown in Table 2. At 4–6 weeks, adding liraglutide to U-500 insulin resulted in an average reduction in TDID of 25% (range of absolute reduction, 0–100 units/day; mean absolute reduction, 44 ± 30 units) and an average weight loss of 11.2 lbs (5.1 kg) (n = 7). At 12 weeks, adding liraglutide to U-500 insulin resulted in an average reduction in HbA1c from baseline of 1.4 ± 0.71% (P = 0.0001), an average reduction in TDID of 28% (range of change of insulin dose, −100 to +30 units/day; mean change in daily insulin dose, −53 ± 35 units/day) (P = 0.0001). Fourteen of the 15 patients (93%) underwent insulin reduction by 12 weeks. The patients (n = 15) experienced an average weight loss of 11.2 ± 8.5 lbs (5.1 ± 3.9 kg) (P = 0.0001) (n = 15) after 12 weeks of treatment with liraglutide in combination with U-500 insulin. The range of weight change was from −26.8 lbs (−12.2 kg) to +0.8 lbs (+0.36 kg).

Hypoglycemia, defined as a blood glucose level of <70 mg/dL, was seen soon after the addition of liraglutide to U-500 insulin in eight of the 15 (53%) patients. No severe hypoglycemic episodes (requiring assistance for treatment) occurred in any patient, and all patients experienced minimal or no nausea.

Two patients had been treated with U-500 insulin in combination with exenatide, 10 μg twice daily, at baseline. These subjects were switched from exenatide directly to 1.2 mg of liraglutide once daily. Switching from exenatide to liraglutide in combination with U-500 insulin resulted in further HbA1c reduction of 1.2%, a 36% reduction in TDID, and a weight loss of 14.6 lbs (6.6 kg) at 3 months in one patient and a reduction of 0.9% in HbA1c, a 34% reduction in TDID, and a weight gain of 0.8 lbs (0.36 kg) in the other. (The latter subject was the only subject in the cohort who gained weight.)

Discussion

This retrospective observational study demonstrates the beneficial effect of adding the long-acting GLP-1 analog liraglutide to high doses of insulin in U-500-treated, insulin-resistant obese patients. Addition of liraglutide at either 1.2 or 1.8 mg daily to high-dose U500 insulin therapy in this small case series resulted in a significant 1.4% decrease in HbA1c in the first 12 weeks of combination therapy with additional benefits of weight loss and insulin reduction. The reduction in HbA1c seen in this study is greater than that achieved by the addition of liraglutide to any combination of oral agents. ^{4 –7} Furthermore, this improvement in glycemic control occurred without any incidence of severe (i.e., requiring assistance for treatment) hypoglycemia.

GLP-1 analog therapy is not associated with hypoglycemia unless combined with insulin secretagogues. To avoid hypoglycemia in patients treated with oral insulin secretagogues, a dose reduction of the oral secretagogue at the time of initiation of GLP-1 analog therapy is sometimes recommended. ¹ Similarly, in most published studies combining exenatide and insulin, mealtime insulin was reduced upon initiation of exenatide, resulting in a low incidence of severe hypoglycemia. ² In our study, the U-500 insulin dose was reduced from 0% to 30% of baseline upon initiation of liraglutide based upon the baseline HbA1c and self-monitored glucose readings of each patient according to the managing physician's judgment; subjects with higher baseline HbA1c values had less insulin reduction on beginning liraglutide. Although no assisted episodes of hypoglycemia occurred in any patient during the 12-week observation period, patients who achieved the lowest 12-week HbA1c values had the most frequent episodes of hypoglycemia (blood glucose <70 mg/dL). Four patients (27%) achieved HbA1c below 6.0% at 12 weeks. Of these patients who achieved normal HbA1c values, two reported no blood glucose levels under 70 mg/dL (and none was seen on any glucose download), one had a single blood glucose reading below 50 mg/dL and 10 blood glucose readings below 70 mg/dL captured on meter download over the 12 weeks, and one had nearly daily episodes of blood glucose values below 70 mg/dL, with the lowest being 36 mg/dL, requiring a further reduction of U-500 insulin at 6 weeks. None of these episodes was considered severe by the patients, and none required assistance for treatment.

A mean weight loss of 11.2 lbs (5.1 kg) occurred by 4–6 weeks in the subset of patients who presented for 4–6-week follow-up visits (n = 7) and was sustained at 12 weeks; it is higher than that observed in any of the preclinical trials of liraglutide as monotherapy or as add-on therapy to oral agents. ^{3 –7} The weight loss observed in our cohort probably resulted from a combination of reduced caloric intake caused by the appetite-suppressing effect of liraglutide coupled to reduced lipogenesis achieved by insulin reduction.

The weight loss and HbA1c reduction achieved by using liraglutide as an adjunct to U-500 insulin in our cohort also exceed those observed in patients with type 2 diabetes treated with pramlintide and insulin, which resulted in a drop in HbA1c of 0.66% and a weight reduction of only 2.3 kg (5.1 lbs) at 3 months in one published study. ⁸ Pramlintide is a pancreatic hormone co-secreted with insulin from the β-cell in response to nutrient intake that acts in concert with insulin to maintain postprandial glucose homeostasis by slowing gastric emptying, suppressing postprandial glucagon secretion, and suppressing food intake, ⁸ all actions shared by GLP-1 and its analogs. ¹ Currently, pramlintide, which is not a GLP-1 receptor agonist, is the only Food and Drug Administration–approved injectable agent for use in combination with insulin for the treatment of diabetes.

All agents for the treatment of type 2 diabetes are associated with either weight gain or weight neutrality, except for pramlintide and the GLP-1 receptor agonist class. Only the GLP-1 analogs and pramlintide are associated with weight loss resulting from central appetite suppression, leading to reduced energy intake. Insulin, although the most effective therapeutic agent for lowering the blood glucose, is particularly associated with weight gain, and one of the limitations of using U-500 insulin is its tendency to cause undesirable weight gain in an already obese population. A meta-analysis of studies of U-500 insulin trials showed an average weight gain of 4.2 kg in patients who switch from other insulin regimens to U-500 insulin. ⁹ These patients might benefit from use of GLP-1 receptor agonists for their salubrious effect on weight; however, the GLP-1 agents are not Food and Drug Administration–approved for use in combination with insulin. Although our study suggests that combining high-dose insulin therapy with liraglutide may yield benefits of improved glycemic control and attenuation of insulin-induced weight gain, the uncontrolled, observational nature of the study and its small sample size are limitations that require that these results, while encouraging, be interpreted with caution. Larger randomized, prospective studies are required to define the best practices for combination therapy with insulin and liraglutide.

Conclusions

The addition of liraglutide, 1.2 or 1.8 mg injected once daily, to high-dose insulin therapy with U-500 insulin resulted in a significant improvement in glycemic control, reduction in insulin requirement, and reduction of weight in obese, insulin-resistant patients with type 2 diabetes. Adding liraglutide to high-dose insulin therapy appears to attenuate or reverse the weight gain associated with insulin therapy with an insulin-sparing effect and without significant hypoglycemia. These results indicate that there is an important potential role for the GLP-1 receptor agonist class in combination with high-dose insulin therapy in the most insulin-resistant population with type 2 diabetes and that prospective controlled trials of GLP-1 analogs in combination with insulin should be undertaken.