Insulin Analogs or Premixed Insulin Analogs in Combination With Oral Agents for Treatment of Type 2 Diabetes

Philip Levy, MD, FACE


April 16, 2007

Approaches to Insulin Initiation in Type 2 Diabetes: Lessons From Clinical Trials

Three approaches for initiation of insulin therapy have been widely used: (1) addition of basal insulin to OADs,[32] (2) addition of a premixed insulin to OADs,[43,44] or (3) addition of mealtime insulin to ongoing treatment with 1 or more oral agents. Each of these approaches has been investigated using insulin analogs as compared to human insulin formulations. In clinical practice, treatment should be individualized to meet patients' needs. In some cases, a complete transition to insulin without continuing oral medication provides effective control and is appropriate.

Studies show that addition of a long-acting insulin analog to oral therapy (secretagogues and sensitizers) is effective for improving glycemic control in patients with type 2 diabetes. The Treat-to-Target Study[23] compared adding glargine or NPH insulin once daily to ongoing treatment with 1 or 2 oral drugs in 756 overweight individuals and an A1C greater than 7.5%. Insulin was titrated using a simple algorithm aimed at achieving a target FPG ≤ 100 mg/dL. After 24 weeks of treatment, FPG declined from 198 to 117 mg/dL for patients who received insulin glargine and from 194 to 120 mg/dL for those who received NPH insulin. The respective declines in A1C were from 8.61% to 6.96% and from 8.56% to 6.97%. The major difference between regimens was that nearly 25% more patients using insulin glargine achieved these improvements without documented nocturnal hypoglycemia (33.2% vs 26.7%). The rate of symptomatic hypoglycemia was also lower with insulin glargine.[23]

In the 9-month Lantus plus Metformin (LANMET) study[45] of 110 patients, an A1C of approximately 7.15% was reached with either insulin glargine or NPH. The incidence of symptomatic hypoglycemia was lower with glargine (4.1 vs 9.0 episodes per patient-year; P < .05) during the first 12 weeks but not significantly different during the remaining 6 months of therapy.[45] In the Glycemic Optimization with Algorithms and Labs at Point of Care (GOAL A1C) study[46] with 7893 adults, glargine dosed with an active titration schedule (monitored each week by physician) showed significantly greater A1C reduction (1.5% vs 1.3%, P < .0001) than glargine dosed with a usual titration schedule (monitored every 6 weeks by physician with no unsolicited contact between visits).[46] A meta-analysis of results from several 24- to 28-week studies that included 1142 patients also indicated that once-daily insulin glargine was as effective as NPH insulin in achievement of A1C ≤ 7%, but with significantly decreased risk for hypoglycemia (54.2% vs 61.2%; P = .0006).[47]

Insulin detemir has also been used effectively as basal therapy in type 2 diabetes. In a 26-week study, addition of insulin detemir to oral therapy decreased A1C similarly by 1.8% to 6.8% vs 1.9% to 6.6% for NPH insulin, while insulin detemir was associated with significantly less weight gain (1.2 vs 2.8 kg; P < .001).[48] A 47% reduction in the risk of all hypoglycemia was also seen with insulin detemir compared with NPH (P < .001), while the risk for nocturnal hypoglycemia was reduced by 55% with insulin detemir compared with NPH.[48] Treatment with insulin detemir has consistently been shown to result in less weight gain compared with NPH insulin.[48,49,50] In 1 study, patients treated with insulin glargine also gained less weight (0.4 kg) than those treated with NPH insulin (1.4 kg; P = .0007), although trials with insulin glargine have not reported weight benefits as consistently as those with insulin detemir.[51,52] Preliminary results from another trial also suggest that insulin detemir in combination with OADs may be associated with less weight gain than insulin glargine and OADs.[53]

Premixed insulin analogs permit delivery of both basal and prandial insulin in a single injection at mealtimes. Three premixed insulin analog formulations are currently available in the United States: insulin lispro 75/25, which contains 75% insulin lispro protamine suspension and 25% insulin lispro; insulin lispro 50/50, which contains equal proportions of the same; and biphasic insulin aspart 70/30 (BIAsp 70/30), which contains 70% insulin aspart protamine suspension and 30% insulin aspart. These formulations are more convenient compared with basal-bolus therapy (ie, fewer injections and less monitoring) or with older premixed human insulin formulations that contained NPH and RHI as the basal and prandial components, respectively. The rapid-acting component of premixed insulin analogs is absorbed and cleared more quickly, thereby allowing mealtime administration and providing better postprandial glycemic control, as well as potentially lower risk for hypoglycemia. The protaminated portion of the analogs impart a longer duration of action.[43,54]

The ease of transitioning patients to insulin therapy with premixed insulin analogs is illustrated by the 1-2-3 Study,[44] an observational trial that included 100 patients not achieving control on OAD therapy with or without basal insulin. The trial evaluated whether addition and self-titration of biphasic insulin aspart 70/30 (BIAsp 30) could achieve AACE or ADA targets for A1C (≤ 6.5% and < 7.0%, respectively). Patients continued oral agents, but stopped basal insulin and added 1 injection of BIAsp 30 (12 units or 70%-100% of the prior basal insulin dose) within 15 minutes of supper initiation. Patients self-titrated their BIAsp 30 dose with investigator guidance every 3 or 4 days to achieve a pre-breakfast FPG of 80-110 mg/dL. At 16 weeks, a pre-breakfast injection of 6 units of BIAsp 30 was added if week 15 A1C was > 6.5% and secretagogues were discontinued. After an additional 16 weeks, 3 units of pre-lunch BIAsp 30 was added if A1C was still above this goal. Addition of once-daily BIAsp 30 before supper enabled 21% of the patients to achieve A1C ≤ 6.5% and 41% to achieve A1C < 7.0%. The respective values with 2 daily injections of BIAsp 30 were 52% and 70%, and those with 3 injections were 60% and 77% (Figure 1).[44]

Percent of patients reaching A1C ≤ 6.5% and < 7.0% in the 1-2-3 Study.[44]

Designed for delivery at mealtimes to control PPG excursions, rapid-acting insulin analogs (aspart, glulisine, or lispro) can be added to a basal or premixed insulin regimen or, in some patients, used separately without basal coverage. For patients using basal insulin, one strategy is to add a rapid-acting analog to the largest meal and then cover additional meals as needed. For patients using a premixed insulin regimen at breakfast and supper, a rapid-acting insulin analog can be added to provide lunch coverage when needed. Rapid-acting analogs can be dosed within 15 minutes of starting a meal and provide better postprandial control compared to RHI.[39] The more rapid pharmacodynamic effects of rapid-acting analogs also decrease the risk of post-absorptive (between-meal) hypoglycemia compared to RHI.[40]

For patients who may prefer a simpler insulin regimen, premixed insulin analogs offer a solution without compromising glycemic control. The need for fewer daily injections may be easier to accept, especially when first transitioning to injections; acceptance of therapy and adherence should improve, also shown to improve A1C levels.[55] Results from several large-scale clinical trials have indicated that the use of a premixed insulin analog twice daily (BID) is superior to insulin glargine daily (QD) for lowering A1C in patients with type 2 diabetes. There may be several reasons for this. In a randomized, crossover study in 12 subjects in which the glucose level was held constant using a euglycemic clamp technique, the glucose-lowering effect of BIAsp 70/30 given BID was 34% greater than with QD insulin glargine. Additionally, the insulin area under the curve (AUC) was 28% higher, and endogenous insulin secretion was suppressed to a greater degree with BIAsp 70/30 (potentially sparing beta-cells) compared with insulin glargine QD.[56]

The Initiation of Insulin to Reach A1C Target (INITIATE) study was a 28-week parallel-group trial of 233 insulin-naive patients with an A1C ≥ 8.0% on > 1000 mg/day metformin alone or in combination with other OADs.[57] Metformin was adjusted up to 2550 mg/day and patients using thiazolidinediones continued on or switched to pioglitazone. Secretagogues were discontinued in both groups. Insulin therapy was started with either 5-6 units BIAsp 30 BID or 10-12 units insulin glargine at bedtime. Treatment was targeted to achieve an FPG of 80-110 mg/dL. At the end of 28 weeks, A1C was 6.91% for the BIAsp 30 group vs 7.41% for those treated with insulin glargine. In addition, more BIAsp 30-treated patients reached the A1C targets of ≤ 6.5% (42% vs 28%) and < 7.0% (66% vs 40%) than those who received insulin glargine. Patients who received BIAsp 30 experienced 3.4 episodes of minor hypoglycemia per patient-year vs 0.7 for insulin glargine. Weight gain with BIAsp 30 (5.4 kg) was also higher than that observed for patients who received insulin glargine (3.5 kg).[57] In a similar study in which secretagogues were continued in patients remaining on insulin glargine, the decrease in A1C was 0.5% lower with BIAsp 30 plus metformin compared with glargine plus glimepiride.[58] In a recent study, more than 75% of patients who were inadequately controlled on optimal doses of metformin and pioglitazone reached an A1C < 7% after BIAsp 30 BID was added.[59] Finally, a regimen using BIASP 30 BID in 715 insulin-naive type 2 diabetes patients was as effective in lowering A1C as a basal-bolus regimen of insulin detemir-insulin aspart (1.42% vs 1.69%, P = .106) while resulting in fewer major hypoglycemic events.[60]

Malone and colleagues compared 75% insulin lispro protamine suspension and 25% insulin lispro (lispro 75/25) BID plus metformin vs insulin glargine QD plus metformin in 105 patients who were starting insulin treatment. The premixed insulin analog resulted in a significantly lower A1C than insulin glargine (7.4% vs 7.8%) and a higher percentage of patients who achieved A1C ≤ 7.0% (42% vs 18%). However, weight gain was significantly higher (2.3 vs 1.6 kg), and hypoglycemia occurred significantly more often (0.68 vs 0.39 episodes/patient per 30 days) with lispro 75/25.[61] In a second study of 97 patients inadequately controlled with insulin alone or insulin plus oral drugs, reductions from baseline in A1C were significantly greater with metformin plus lispro 75/25 vs metformin plus glargine (1.0% and 0.42%, respectively; P < .0001).[62]

The greater A1C lowering obtained with premixed insulin analogs compared with insulin glargine in clinical trials has also been documented in clinical practice. Based on a retrospective analysis of electronic medical records from 2.4 million patients in the United States, including 8166 insulin-naive individuals, patients using a premixed insulin analog BID had a 0.49% to 0.65% greater decrease in A1C compared to those using insulin glargine QD or human insulin 70/30 BID.[63]

These results obtained with premixed insulin analogs vs basal long-acting insulin analogs contrast sharply with those from a study of insulin glargine plus OADs vs premixed human insulin 70/30 alone.[64] In this 24-week study, 371 insulin-naive patients with A1C 7.5% to 10% while on OADs (sulfonylurea plus metformin) were randomly assigned to QD morning insulin glargine plus glimepiride and metformin or premixed human insulin 70/30 BID with discontinuation of the OADs. Insulin dosages were titrated to achieve FPG ≤ 100 mg/dL. Patients who received insulin glargine and OADs had a significantly greater decrease from baseline in A1C (-1.64%) vs those treated with premixed human insulin alone (-1.31%). In addition, the rate of nocturnal hypoglycemia was significantly lower with insulin glargine and metformin vs the conventional premixed insulins alone (28.6% vs 45.5%).[64]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.