Comparison of Basal Insulin Added to Oral Agents Versus Twice-Daily Premixed Insulin as Initial Insulin Therapy for Type 2 Diabetes

Male or female patients aged 35–75 years with a type 2 diabetes duration of at least 1 year and treated with a stable dose of sulfonylurea and metformin for at least 1 month were enrolled at 66 sites in 10 European countries. Further inclusion criteria included BMI ≤35 kg/m², HbA_1c levels between 7.5 and 10.5%, and fasting blood glucose (FBG) levels ≥120 mg/dl (≥6.7 mmol/l). Exclusion criteria included any additional use of other oral blood glucose–lowering agents, prior use of insulin exceeding 3 days, and a history of ketoacidosis. The study was conducted in accordance with the Declaration of Helsinki. Approval by institutional ethics committees was obtained for each participating site. All patients provided written informed consent before study entry.

This was a parallel group, open-label, randomized, multinational clinical trial with a 1- to 4-week screening phase and a 24-week treatment phase. A 1:1 randomization schedule stratified by center sequentially assigned treatment codes to eligible patients, using a central randomization service of the electronic case report form InForm (Phaseforward, Maidenhead, U.K.).

Previous sulfonylurea therapies were replaced with 3 or 4 mg glimepiride (Amaryl; Aventis Pharma) during the screening phase. Metformin (≥850 mg; Metformin Basics; Basics) during the study was provided and taken at the same dose as before study entry. The dosage of both agents remained unchanged throughout the study. At the baseline visit, patients were randomly assigned to either insulin glargine (Lantus; Aventis Pharma) given once daily in the morning in combination with glimepiride and metformin (glargine plus OAD) or to human premixed insulin (30% regular, 70% NPH insulin; Insulin Actraphane HM 30/70; Novo Nordisk) to be administered twice daily (before breakfast and dinner), while glimepiride and metformin were discontinued (70/30). The insulins were injected using Optipen 1E for insulin glargine and NovoPen for premixed insulin. The starting dose for insulin glargine was 10 IU in the morning and, for premixed insulin, 10 IU before breakfast and 10 IU before dinner. These starting doses could be lowered if considered clinically necessary by the investigator. Insulin doses were adjusted by a forced titration regimen calling for weekly adjustments for 8 weeks and at 2-week intervals thereafter for both groups, according to daily self-monitored capillary whole blood glucose measurements using meters (AccuChek Sensor; Roche Diagnostics). For both groups, the FBG target was 100 mg/dl (5.6 mmol/l), and the before dinner blood glucose target for the 70/30 group was 100 mg/dl (5.6 mmol/l), with a stepwise increase of insulin depending on the blood glucose values as follows: blood glucose >100–120 mg/dl, increased by 2 IU/day; blood glucose >120–140 mg/dl, increased by 4 IU/day; blood glucose >140–160 mg/dl, increased by 6 IU/day; and blood glucose >160 mg/dl, increased by 8 IU/day, unless symptoms of hypoglycemia occurred. Hypoglycemia was confirmed by blood glucose <60 mg/dl. Severe hypoglycemia was defined as an event with symptoms consistent with hypoglycemia during which the person required the assistance of another person and which was associated with a blood glucose level <36 mg/dl and/or with recovery after oral carbohydrate, intravenous glucose, or glucagon administration.

FBG values and (for patients receiving 70/30) predinner blood glucose values, as well as hypoglycemic episodes, were recorded in a standardized diary. Hematologic, clinical chemistry, and HbA_1c values at baseline and 12 and 24 weeks were measured at a central laboratory (MDS, Hamburg, Germany); HbA_1c was measured by high-performance liquid chromatography (Bio-Rad Variant; Bio-Rad, Munich, Germany) traceable to the Diabetes Control and Complications Trial reference method, with a reference range of 4.8–6.7%. An eight-point glucose profile (before and 2 h after breakfast, lunch, and dinner; at bedtime; and at 3:00 a.m.) was obtained on 2 consecutive days before a visit at baseline and 2, 4, 8, 12, and 24 weeks. The baseline eight-point profile was performed while patients were receiving only glimepiride and metformin. Adverse events were noted by the investigator at every visit or telephone contact.

The primary efficacy measure was the change in HbA_1c level from baseline to end point. Secondary efficacy measurements were HbA_1c level, mean FBG level, proportion of patients with FBG levels ≤100 mg/dl (≤5.6 mmol/l), proportion of patients with HbA_1c ≤7.0% and HbA_1c ≤7.0% with no nocturnal hypoglycemia, and mean blood glucose values from the eight-point profiles.

Safety measures were the proportion of patients with hypoglycemic events and the frequency of hypoglycemic events. Hypoglycemia was considered confirmed if documented by a blood glucose level <60 mg/dl (<3.3 mmol/l).

Statistical analyses were performed on the intent-to-treat population, defined as randomized patients who received at least one injection of insulin. Statistical testing was performed at a significance level of α = 0.05. ANCOVAs were performed to compare changes in HbA_1c and secondary continuous variables between treatment groups. Adjusted means and corresponding two-sided 95% CIs were calculated. Categoric secondary variables were analyzed using Cochran-Mantel-Haenszel tests. Statistical analyses were performed using SAS software (version 8.2; SAS Institute, Cary, NC).

With a 1:1 randomization ratio and based on the assumption of a common SD of 1.3%, an absolute difference of 0.4% for HbA_1c reductions among treatment groups can be detected with an α error of 0.05 (two sided) and a β error of 0.2 with 167 patients per treatment group.

A total of 511 patients were screened: 371 patients were eligible for randomization, and 364 patients comprised the intent-to-treat population. There were 177 patients randomly assigned to glargine plus OAD and 187 to 70/30. Baseline demographic and clinical characteristics were similar between the treatment groups (Table 1). After randomization, 7 patients on glargine plus OAD (3 lost to follow-up and 4 other reasons) and 28 patients on 70/30 (12 unwilling to continue, 5 lack of efficacy, 2 lost to follow-up, and 9 other reasons) withdrew from the study.

Over the 24-week treatment period, mean (±SD) HbA_1c levels decreased from 8.85 ± 0.98 to 7.15 ± 0.90% with glargine plus OAD and from 8.83 ± 0.87 to 7.49 ± 1.09% with 70/30 (Fig. 1A). Mean adjusted HbA_1c improvement was greater with glargine plus OAD (−1.64% [95% CI −1.51 to −1.78]) than with 70/30 (−1.31% [−1.17 to −1.44]). The adjusted mean between-treatment difference of −0.34% (−0.52 to −0.16%, P = 0.0003) significantly favored the glargine plus OAD group (Fig. 1B).

An HbA_1c level ≤7% was achieved by 49.4% of patients in the glargine plus OAD group compared with 39.0% in the 70/30 group (P = 0.0596 for the between-treatment difference). Significantly more patients on glargine plus OAD (45.5%) than on 70/30 (28.6%) reached an HbA_1c ≤7% without an episode of confirmed nocturnal hypoglycemia (P = 0.0013 for the between-treatment difference).

FBG levels decreased from 171 to 115 mg/dl (9.5 to 6.4 mmol/l) with glargine plus OAD and from 172 to 133 mg/dl (9.6 to 7.4 mmol/l) with 70/30. Improvement in FBG was significantly better with glargine plus OAD compared with 70/30 (adjusted mean between-treatment difference −17 mg/dl [−0.9 mmol/l]; 95% CI −24 to −10 mg/dl [–1.3 to −0.6 mmol/l], P < 0.0001). A greater proportion of patients reached an FBG level ≤100 mg/dl (≤5.6 mmol/l) with glargine plus OAD than with 70/30 (31.6 vs. 15.0%, P = 0.0002).

Diurnal (eight-point) glucose profiles were similar for both groups at baseline (before insulin initiation). Mean daily blood glucose level improved from 182 to 137 mg/dl (10.1 to 7.6 mmol/l) in the glargine plus OAD group compared with 184 to 151 mg/dl (10.2 to 8.4 mmol/l) in the 70/30 group (P < 0.0001 for between-treatment difference). At end point, the reduction from baseline was significantly greater with glargine plus OAD than with 70/30 for values obtained at the fasting, postlunch, dinner, postdinner, and 3:00 a.m. time points (Fig. 2).

Insulin dose increased over the study duration from a mean (±SD) daily starting dose of 9.9 ± 2.6 to 28.2 ± 15.2 IU at end point for insulin glargine. The prebreakfast dose of premixed insulin increased from the mean starting dose of 10.3 ± 2.5 to 33.5 ± 18.0 IU at end point, whereas the predinner dose increased from the mean starting dose of 10.3 ± 2.5 to 31.0 ± 16.1 IU at end point, resulting in more than twice as much daily insulin with 70/30 than with glargine plus OAD (64.5 vs. 28.2 IU). The mean daily dose was 3.4 ± 0.5 mg for glimepiride and 1,894.5 ± 475.1 mg for metformin.

One hundred nine patients (61.6%) receiving glargine plus OAD and 127 patients (67.2%) receiving 70/30 experienced at least one hypoglycemic event (P = 0.2838). During treatment, the rate of confirmed hypoglycemic events, expressed as episodes per patient-years, was ∼50% lower with glargine plus OAD than with 70/30 for the overall, symptomatic, and nocturnal categories (Table 2). Severe hypoglycemia was very uncommon in both treatment groups (Table 2).

Mean (±SD) weight gain in patients treated with glargine plus OAD and 70/30 was 1.4 ± 3.4 and 2.1 ± 4.2 kg, respectively (P = 0.0805 for between-group difference).

The incidence of adverse events was similar; 89 patients (50.3%) in the glargine plus OAD group and 92 patients (48.7%) in the 70/30 group experienced at least one adverse event. Most common were respiratory disorders (16%), nervous system disorders (10%), and gastrointestinal disorders (10%). A possible relationship to the study medication was reported for 10 adverse events in 8 glargine plus OAD patients and for 12 adverse events in 10 70/30 patients. Withdrawals due to adverse events occurred in one patient (0.6%) treated with glargine plus OAD and six patients (3.2%) treated with 70/30.

These results show that in patients with type 2 diabetes poorly controlled on oral therapy, adding a single injection of insulin glargine to glimepiride and metformin can provide more effective glycemic control than stopping OADs and starting twice-daily 70/30 insulin. The glargine plus OAD regimen enabled nearly 50% of patients to reach HbA_1c ≤7% without experiencing nocturnal hypoglycemia, whereas <30% of patients on 70/30 insulin achieved target HbA_1c ≤7% in the absence of nocturnal hypoglycemia.

The number of hypoglycemic events per patient-year and the number of events per patient were ∼50% lower in the glargine plus OAD group than in the 70/30 group. The lower rate of hypoglycemia with the basal insulin regimen is of particular interest because fear of hypoglycemia remains one of the key obstacles to both initiating and optimizing insulin therapy (11–13). The difficulty of managing multiple injections and the associated requirement for multiple daily glucose measurements is another barrier to achieving recommended glycemic control targets (14). The glargine plus OAD regimen in this study required only a single daily injection and a single before-breakfast glucose test to guide therapy and, therefore, should be easy to use in clinical practice.

Since patients randomized to the 70/30 group did not receive glimepiride or metformin, this study compared two regimens for initiating insulin rather than two specific forms of insulin. However, previous studies using NPH insulin in combination with OADs did not show better glycemic control in terms of HbA_1c reduction in comparison to insulin monotherapy with premixed insulin (15–17). In the present study, insulin treatment initiated by adding insulin glargine to OADs resulted in a significantly greater improvement in glycemic control compared with 70/30 insulin alone. In clinical practice, OADs are often discontinued once a 70/30 insulin regimen is begun, but continuing metformin might be expected to improve the effectiveness of this regimen. Clearly, many questions remain regarding the initiation of insulin therapy in patients with type 2 diabetes. The current study provides efficacy and safety data pertaining to two commonly used insulin regimens. Further studies are required to provide physicians with additional guidance. These should include addressing the benefit of 70/30 insulin plus metformin combination to ascertain the level of influence of metformin on the results obtained in the insulin glargine–treated group. In addition, it would be of interest to compare the glargine plus OAD regimen with a rapid-acting analog plus NPH insulin as use of the latter insulin regimen becomes more widespread. The relative costs of treatment with all of these regimens, including the glucose testing required by each, should also be studied. Finally, despite the improvement in control achieved by adding insulin glargine to OADs, over one-half of patients in the glargine plus OAD group did not reach HbA_1c ≤7%. The relatively low total daily insulin dose in the glargine plus OAD group and the low rate of hypoglycemia with this regimen support the feasibility of continued titration to achieve target HbA_1c in more patients. Even so, some patients will require additional prandial injections of insulin to reach the ≤7% HbA_1c target.

In conclusion, this study demonstrated that, for patients with type 2 diabetes who are inadequately controlled with metformin plus a sulfonylurea, adding a once-daily injection of insulin glargine is a simple method that is more effective in improving glycemic control than starting twice-daily injections of premixed insulin without oral agents.

This study was supported by a research grant from Aventis Pharma.

Investigators: Austria: W. Fortunat, A. Holler, R. Prager, J. Thomas, and T. Wascher; Finland: H. Yki-Järvinen; France: J.F. Blickle, J.M. Brun, M. Rodier, and B. Schmitt; Germany: P. Brommer, K. Busch, H. Dancygier, E.M. Fach, T. Feldmann, A. Fiesselmann, G. Garanin, J. Grosskopf, J. Habbig, T. Hampel, H. Herrmann, H.U. Janka, K.A. Jahnke, V. Jung, M. Kiper, C. Klein, D. Klein, V. Koch, K. Langer, E. Lohr, C. Marck, H.J. Marks, P. Mayr, S. Maxeiner, O. Mueller, F. Odemar, A. Pfuetzner, H. Pitule, H. Samer, G. Scholz, A. Sterzing, H.J. Strotmann, J. von Huebbenet, J. Wachter, and U. Wendisch; Italy: A. Civarella, G. Riccardi, G. Rosti, G. Seghieri, and G. Vespasiani; the Netherlands: R. van Doorn, L. Lieverse, and W. Venekamp; Spain: B.F. Gomis, C.F. Hawkins, R.G. Mayor, A. Novials, and J. Zurro Hernandez; Sweden: B. Lindgren and M. Palmer; Switzerland: P. Gerber; U.K.: R. Bilous, D. Gordon, C.M. Kesson, M. Sampson, and J. Vora.

Study team: M. Herbold, C. Kliebe-Frisch, S. Krull, A. Loehr, W. Messer, A. Mueller, H. Nortmeyer, J. Peiker, T. Schlink, and N. Tjia.

Data from this manuscript have been presented in abstract form [Diabetes 53 (Suppl.2):A130, 2004 and Diabetelogia 47 (Suppl. 1):A269, 2004] and presented as posters at the American Diabetes Association Scientific Sessions 2004 and the European Association for the Study of Diabetes Congress 2004.