Addition of Biphasic, Prandial, or Basal Insulin to Oral Therapy in Type 2 Diabetes

Holman RR, Thorne KI, Farmer AJ, et al; 4-T Study Group. N Engl J Med. 2007;357:1716-1730.

Patients with type 2 diabetes mellitus (T2DM) are afflicted with progressive deterioration of glycemic control that occurs in conjunction with decline in pancreatic β-cell function, causing a majority of patients to eventually require insulin. The addition of insulin to treatment regimens usually occurs when glycemic control becomes suboptimal despite high doses of oral antidiabetics (OADs), such as metformin, sulfonylureas (SUs), and thiazolidinediones (TZDs). Although the addition of insulin can improve A1C levels, many patients with T2DM still cannot reach A1C goals when receiving conventional insulin regimens, and there have been few studies designed to provide direct comparisons of insulin analog regimens in combination with OADs. Treating to Target in Type 2 Diabetes (4-T) is a 3-year, multicenter, open-label, randomized, controlled clinical trial being conducted by the Diabetes Trials Unit. This report contains results obtained during the first year, including efficacy and safety of adding a biphasic, prandial, or basal insulin analog to the treatment regimen for patients with T2DM who had suboptimal glycemic control while receiving the maximum tolerated doses of metformin and SU.

Men and women ≥ 18 years of age who had T2DM for ≥ 12 months, A1C of 7% to 10%, received maximum tolerated doses of metformin and SU (or only 1 if they did not tolerate the other) for ≥ 4 months, and body mass index ≤ 40 but had not been treated with insulin were recruited from 58 clinical centers in Ireland and the UK. Patients who had received TZDs or triple OAD treatment within the previous 6 months were excluded, as were patients with hypoglycemia unawareness or recurrent major hypoglycemia, anticipated changes in concomitant medication that could affect their glucose regulation, or uncontrolled hypertension as defined by systolic blood pressure (BP) ≥ 180 mm Hg or diastolic BP ≥ 105 mm Hg. A total of 936 patients were screened, 708 of whom were eligible to participate. Participants were randomized to receive biphasic insulin aspart 30 (n = 235) twice a day, prandial insulin aspart (n = 239) before meals 3 times a day, or basal insulin detemir (n = 234) once daily. They had a mean (± SD) age of 61.7 ± 9.8 years and a median duration of disease of 9 years.

Follow-up visits for each patient were scheduled at Weeks 2, 6, 12, 24, 38, and 52, along with interim telephone contact, during which they provided information obtained from capillary glucose profiles performed daily by those in the biphasic and basal groups before breakfast and evening meals, and by those in the prandial group before meals, 2 hours after, and again at bedtime. Glucose profiles and self-reported hypoglycemia were used to calculate changes in insulin doses with the trial-management system recommendations of 72 to 99 mg/dL before meals and 90 to 126 mg/dL 2 hours after meals. Hypoglycemia among patients was categorized as grade 1 when they had symptoms along with capillary glucose levels ≥ 56 mg/dL, grade 2 if they had symptoms and glucose levels < 56 mg/dL, and grade 3 if they required third-party assistance. If there was unacceptable hyperglycemia, defined as A1C > 10% or 2 consecutive values ≥ 8% after 24 weeks of therapy, a second insulin was added; if the patient was taking an SU, it was discontinued. Insulin aspart was added to the midday meal for patients receiving biphasic insulin, insulin detemir at bedtime for those treated with prandial insulin, and aspart 3 times daily with meals for those in the basal insulin group. The primary outcome of this study was A1C after 1 year; secondary outcomes included the proportion of patients in each group that had A1C levels ≤ 6.5% and those with A1C ≤ 6.5% who had no grade 2 or 3 hypoglycemia during Weeks 48 to 52. Additional secondary outcomes included rates of hypoglycemia and weight gain, the self-measured capillary glucose profile, the proportion of patients requiring 2 injections of insulin detemir, the proportion of those who suffered unacceptable hyperglycemia, the ratio of albumin to creatinine, and quality of life.

Within the first 2 weeks of receiving insulin, the mean rates of grade 2 hypoglycemia were 0.045 events per patient each week among the biphasic group, 0.031 for the prandial group, and 0.024 for the basal group; no patients experienced grade 3 episodes. The percentages of patients whose insulin doses were within ± 10% of the trial-management system recommendation averaged 89.7% in the biphasic group, 80.4% in the prandial group, and 90.2% in the basal group. Among patients in the basal insulin group, 33.8% required an additional morning injection of insulin detemir. From Weeks 24 to 52, a second type of insulin was required to address unacceptable hyperglycemia among 8.9% of patients in the biphasic group, 4.2% in the prandial group, and 17.9% in the basal group (P < .001 for all comparisons). Reductions of A1C levels from baseline to Week 52 were 1.3% in the biphasic group, 1.4% in the prandial group, and 0.8% in the basal group. Although there were no significant between-group differences in A1C level at Week 52 for the biphasic (7.3%) and prandial (7.2%) groups, A1C was significantly higher in the basal group (7.6%) than the others (P < .001 vs basal group for both).

At Week 52, all groups had greater proportions of patients with A1C levels ≤ 6.5% than they had at baseline (P < .001 for all comparisons). Although there was no significant difference in the proportions of patients in the biphasic (17%) and prandial (23.9%) groups achieving that goal, the proportion achieving this goal in the basal group (8.1%) was lower than both (P = .001 vs biphasic, P < .001 vs prandial). In a similar manner, the proportion of patients with A1C level ≤ 7% was also significantly lower in the basal (27.8%) group than the biphasic (41.7%) or prandial (48.7%) groups (P < .001 for both). A1C levels ≤ 6.5% without grade 2 or 3 hypoglycemia were reached during Weeks 48 to 52 by 52.5%, 43.9%, and 78.9% of the biphasic, prandial, and basal groups, respectively (P = .001).

Assessment of patients with baseline A1C level ≤ 8.5% revealed no significant difference in the likelihood of achieving values ≤ 6.5% between the prandial and biphasic groups (odds ratio [OR] for prandial = 1.76; 95% confidence interval [CI], 0.96-3.26; P = .07) or between the basal and biphasic groups (OR for basal = 0.50; 95% CI, 0.24-1.03; P = .06). Patients with baseline A1C > 8.5% were less likely to achieve values of ≤ 6.5% in the basal group than were patients in the biphasic group (OR for basal = 0.21; 95% CI, 0.07-0.65; P = .007), but patients in the prandial group did not differ significantly from those in the biphasic group (OR for prandial = 1.24; CI = 0.62-2.51; P = .54).

Median insulin doses increased during the study, and at Week 52 doses were similar in the biphasic and basal groups but higher in the prandial group. Although patients treated with each regimen gained weight, such increases were greatest among those in the prandial group, less among the biphasic group, and least among the basal group, although changes were similar among all patients who reached A1C levels ≤ 6.5% at Week 52. Reductions of mean FPG values were greater in the basal than biphasic group, and greater in the biphasic than prandial group. In contrast, reductions in mean postprandial glucose levels were greatest among the prandial group, followed by the biphasic group, and were least in the basal group. Median rates of grade 2 or higher hypoglycemia were were highest in the prandial group, lower in the biphasic group, and lowest in the basal group. Mean numbers of hypoglycemic events per patient per year were 5.7,12, and 2.3 in the biphasic, prandial, and basal groups, respectively.

There were no significant between-group differences in the proportions of patients with serious adverse events (AEs) or the total number of such events. Levels of fasting plasma insulin that were more than 3 times the upper limit of normal occurred more frequently in the biphasic (9.6%, P = .004) than the prandial (2.8%) or basal (1.8%) groups. Only 2 patients discontinued metformin due to 2 successive measures of plasma creatinine values > 1.7 mg/dL.

This report provides results obtained during the first 12 months of the 3-year 4-T trial, which is anticipated to be completed in December 2008. Patients with T2DM who had suboptimal glycemic control despite treatment with metformin and SU were randomized to 1 of 3 different analog insulin regimens, each of which provided clinically relevant, sustainable reductions of A1C levels. Unfortunately, only 16% of the participants achieved the goal of A1C ≤ 6.5% and 39% achieved A1C ≤ 7%; however, the data obtained from this trial thus far provide information that could help clinicians tailor insulin regimens for each of their patients. For example, few studies completed to date have provided direct comparisons between basal insulin detemir and other insulins among patients with T2DM. Although patients treated with biphasic or prandial insulin regimens had comparable reductions in A1C levels, these decreases were greater than those provided by basal insulin among individuals with A1C > 8.5%, possibly reflecting the increased impact of postprandial glycemia on A1C levels as glycemic control worsens. All 3 groups experienced weight gain and increased risk of hypoglycemia, both of which were greatest among the group receiving prandial insulin. Such AEs, along with the relative simplicity of fewer injections each day, indicate that basal insulin may be preferable to other regimens for initiating insulin among patients already treated with OADs. However, the majority of patients with T2DM will require more than 1 type of insulin to achieve goals of glycemic control, and the final 2 years of this study will provide a great deal of information regarding the use of such insulin regimens among these patients.

Note: this study was supported by Novo Nordisk and Diabetes UK