Hypoglycaemia is a clinically important and potentially dangerous side effect of insulin therapy in patients with Type 2 diabetes mellitus (T2DM).1,2 The basal-bolus insulin strategy has been studied in the Glulisine in Combination with Insulin Glargine in an Intensified Insulin Regimen (GINGER) study in a population with long-standing insulintreated T2DM, with or without metformin, who were not controlled adequately with their previous premixed insulins.3 The study demonstrated that an intensified basal-bolus regimen using insulin glargine plus insulin glulisine provided
Hypoglycaemia is a clinically important and potentially dangerous side effect of insulin therapy in patients with Type 2 diabetes mellitus (T2DM).1,2 The basal-bolus insulin strategy has been studied in the Glulisine in Combination with Insulin Glargine in an Intensified Insulin Regimen (GINGER) study in a population with long-standing insulintreated T2DM, with or without metformin, who were not controlled adequately with their previous premixed insulins.3 The study demonstrated that an intensified basal-bolus regimen using insulin glargine plus insulin glulisine provided significantly better glycaemic control and numerically lower overall hypoglycaemia than premixed insulin therapy.
The premixed insulin group in the GINGER study comprised two subgroups: those taking neutral protamine Hagedorn (NPH) insulin plus regular insulin and those taking biphasic insulin aspart 70/30. The aim of this post-hoc analysis was to determine the overall hypoglycaemic event rates in the insulin glargine plus insulin glulisine group with that of the premixed insulin subgroups, and to compare hypoglycaemia risk relative to the glycaemic control achieved at the end of the GINGER study.
Methods
Details of the study design and patient population have been published previously.3 In brief, this was a 52-week, open-label, active-controlled, randomised, multi-national clinical trial that compared basal insulin glargine plus mealtime insulin glulisine with an optimised conventional therapy of twice-daily premixed insulin therapy. Male and female patients aged 18 to 75 years were eligible for inclusion in the study if they had had T2DM for ≥5 years and had been treated with a stable regimen of twice-daily injections of premixed insulin (NPH plus regular insulin, or NPH plus either lispro or aspart in a ratio of 70/30), with or without metformin for ≥3 months before screening, and had HbA1cbetween 7.5 % and 11.0 % at screening and randomisation. Oral antidiabetic drugs, except metformin, were not allowed. Exclusion criteria included body mass index >38 kg/m2 and ≥2 severe hypoglycaemic episodes within the past 3 months.
Patients were randomised to an insulin regimen with glargine plus glulisine (n=153) consisting of once-daily glargine injections and three times daily glulisine injections at mealtimes or to premixed insulin combinations consisting of twice daily injections (n=157).
In the premixed insulin group, 93 patients received NPH insulin plus regular insulin, and 63 patients received biphasic insulin aspart 70/30. During the first 8 weeks of treatment, insulin doses in both groups were adjusted using a forced titration regimen. The titration targets were the same for both regimens: fasting blood glucose ≤5.5 mmol/L (≤100 mg/ dL) and postprandial blood glucose ≤7.5 mmol/L (≤135 mg/dL).
The primary endpoint of the GINGER study was the change in HbA1c from baseline to endpoint. Secondary endpoints included the rate of hypoglycaemia. A hypoglycaemic event was defined as an event with symptoms consistent with hypoglycaemia and/or a confirmed plasma glucose (PG) level ≤60 mg/dL (≤3.3 mmol/L).
For the post-hoc analysis, hypoglycaemic event rates (events per patient-year) were estimated for the total study population and for those patients who experienced ≥1 episodes of any hypoglycaemia, using a negative binomial regression model.4 In addition, hypoglycaemia event rates were plotted against the endpoint HbA1cfor each patient in each group or subgroup. All tests were adjusted for baseline HbA1c and duration of diabetes. To account for the relatively large number of patients with no event, a logistic regression analysis was performed and the p values of both methods combined to form the adjusted p values reported for each treatment group.
Results
Overall, the total study population (n=310) was 49 % female, had a mean±SD age of 61±8 years and had a body mass index of 30.1±3.7 kg/ m2. Mean duration of T2DM was 13±6 years, with mean insulin use of 5±4 years. Baseline HbA1cwas 8.56±0.85%. The demographics and baseline characteristics of the two groups were well matched.3
The glargine plus glulisine group had a greater change from baseline HbA1cat study endpoint (–1.31±1.19%) than the overall premixed insulin group (–0.80±1.01%; p=0.0001) or the biphasic insulin aspart 70/30 subgroup (–0.70±1.08%; p=0.0009).3
For overall hypoglycaemia, the adjusted event rate (episodes/patientyear), was 24.5% lower for glargine plus glulisine (14.0±24.2) than for the overall premixed insulin group (18.5±36.9), but this difference was not significant (p=0.12). However, there was a significantly lower hypoglycaemic event rate for the glargine plus glulisine group (43.3%) compared with the biphasic insulin aspart 70/30 subgroup (24.7±48.5; p=0.02).
In the analysis of patients with ≥1 episodes of hypoglycaemia during the study (n=232, 74.8% of participants), patients receiving glargine plus glulisine had a significantly lower (26.5%) event rate (18.5±26.3) than the overall premixed insulin group (25.1±41.1; p=0.044) and a 40.7% lower rate than the biphasic insulin aspart 70/30 subgroup (31.1±52.7; p=0.009).
Similar results were obtained when calculating event rates of confirmed hypoglycaemia with PG ≤3.3 mmol/L (data not shown). In addition, comparing the glargine plus glulisine group with the NPH/regular insulin premixed subgroup showed similar hypoglycaemia event rates in both groups (data not shown).
Expected and observed hypoglycaemic event rates relative to endpoint HbA1cfor the glargine plus glulisine group compared with the overall and biphasic aspart 70/30 premixed insulin subgroup are shown in Figures 1A and 1B, respectively. These graphs indicate that the glargine plus glulisine treatment maintains a numerically lower level of hypoglycaemic events over the entire range of HbA1clevels compared with the premixed insulin treatments, which showed higher hypoglycaemic event rates at lower HbA1clevels, but this declined at higher HbA1clevels. The graphs also show a more pronounced difference in the hypoglycaemic event rate between the glargine plus glulisine group and the biphasic insulin aspart 70/30 subgroup (graph B) at lower HbA1clevels than was observed between the glargine plus glulisine group and the overall premixed insulin group (graph A).
Discussion
In the GINGER study, rates of hypoglycaemic events were numerically, but not significantly, lower for patients taking once-daily insulin glargine plus mealtime insulin glulisine than for those taking twicedaily premixed insulin.3 In this post-hoc analysis of the GINGER study, actual rates of overall hypoglycaemia were determined as a function of endpoint HbA1cfor patients in each of the two groups. Using a negative binomial regression model,4 the predicted rates for the basal-bolus group were compared with those of the overall premixed insulin group. Adjusting for endpoint HbA1cdid not alter the original outcome between the glargine plus glulisine and the overall premixed insulin groups: overall hypoglycaemic events were numerically, but not significantly, lower for patients taking once-daily insulin glargine plus mealtime insulin glulisine.
A substantial number of patients in the two groups did not experience any hypoglycaemic events. When only those patients who experienced one or more events were analysed, the rate of hypoglycaemic events was significantly lower with glargine plus glulisine than with the overall premixed insulin group.
For the premixed insulin aspart subgroup, rates of overall hypoglycaemia were significantly greater than those of glargine plus glulisine whether all patients or only those who experienced a hypoglycaemic event were analysed.
Differences in dosing frequency between the glargine plus glulisine regimen (given as four injections daily) versus the premixed insulin regimen (given as two injections daily) were unlikely to have affected study endpoints.
Previous studies in inadequately controlled diabetes patients have shown that dose intensification of premixed insulins does not appear to alter endpoint HbA1clevels or the incidence of hypoglycaemic events.5–7
The study results suggest that when both endpoint HbA1cand overall hypoglycaemic events are considered, a basal−bolus regimen with insulin glargine plus insulin glulisine has a better benefit-to-risk profile than premixed insulin therapy, especially than insulin aspart 70/30, in patients with T2DM.
Summary
- A post-hoc analysis of the GINGER study was conducted to evaluate hypoglycaemia rates relative to endpoint HbA1cin patients with T2DM who were receiving insulin glargine plus insulin glulisine versus premixed insulin
- The overall adjusted hypoglycaemic event rate (events per patientyear) was numerically lower for those patients receiving insulin glargine plus insulin glulisine compared with the overall premixed insulin group (p=0.12) but significantly lower compared with the biphasic insulin aspart 70/30 subgroup (p=0.02)
- Similarly, in patients with ≥1 episode of symptomatic and/or confirmed hypoglycaemia during treatment the overall adjusted hypoglycaemic event rate was lower for those patients receiving insulin glargine plus insulin glulisine compared with both the overall premixed insulin group (p=0.044) and the biphasic insulin aspart 70/30 subgroup (p=0.009)
- Insulin glargine/glulisine treatment was associated with a numerically lower hypoglycaemia event rate that was maintained at all achieved HbA1cendpoints compared with premixed insulin treatment
Author Contributions
Andreas Fritsche, Wolfgang Landgraf and Almut Hahn researched the data, contributed to the discussion and reviewed and edited the manuscript. Hans-Ulrich Häring reviewed and edited the manuscript.
