Introduction
Insulin administration in patients with type 2 diabetes mellitus (T2DM) has proved to be the most effective treatment modality, which should be initiated early in the course of the disease in case lifestyle modification and/or metformin administration fail to achieve recommended glycaemic standards[
1,
2]. Insulin can be administered in addition to oral hypoglycaemic therapy and in different therapeutic schemes which can be prospectively modified based on close glycosylated haemoglobin A1c (HbA1c) and blood glucose (BG) monitoring[
3].
The primary objective of the present retrospective study (LANTus utilisation in real life versus PREmix Insulin, PRELANTI) was to assess the efficacy of intensification of treatment, by measuring the change in HbA1c, of two main insulin treatment schemes: (a) glargine once daily in the evening and (b) biphasic human insulin twice daily (bd) in patients with T2DM on concomitant therapy with oral antidiabetic drugs (OAD). Secondary objectives were to examine for changes in fasting BG, body weight, treatment with OAD or insulin and safety, by assessing the frequency and severity of hypoglycaemic episodes. In both treatment groups, fast acting insulin was added if needed.
Research design and methods
A retrospective parallel two-arm study was designed to include 332 patients with T2DM, equally distributed in two treatment groups. Taking into consideration that difference in HbA1c change is 0.34% in favour of insulin glargine versus premixed insulins in similar comparative studies between insulin treatment strategies[
4], a sample size of 332 participants were required to be enrolled in the study with an alpha risk (2-sided test) of 5% and 80% power, as well as with a standard deviation of 1.1%. Finally, 301 patients, already on treatment with biphasic human insulin bd in combination with OADs, were recruited in 10 centers (including both National Health Services and university hospitals and private practice doctors) throughout Greece. The study was approved by hospitals research ethics committees and by National Organization for Medicines.
Patients’ selection by investigators and inclusion in the study was taking place consecutively, starting from most recent patient’s data (from each patient’s last visit) and going backwards in time. In each centre, data were collected retrospectively from 142 consecutive patients with T2DM, who had been switched from human biphasic insulin to insulin glargine once daily in a period of 6 – 12 months prior to their inclusion in the study (active group), and compared to data collected retrospectively from 159 consecutive patients who continued treatment with biphasic human insulin bd for the same time period of 6 – 12 months prior to their inclusion in the study (control group). Fast acting insulin was added in both groups if considered necessary, depending on the values of HbA1c and BG at the attending physician’s discretion. OADs were also added or withdrawn on the basis of the attending physician’s global assessment of the efficacy of treatment schemes and the changes were reported on the patients’ case records. Baseline characteristics of study population are given in Table
1.
Table 1
Baseline characteristics of the study population
Sex, n (%)
| | | |
Male | 73 (45.9) | 60 (42.3) | 0.523 |
Female | 86 (54.1) | 82 (57.8) | |
Age (years)
| 66.7 ± 9.6 | 63.2 ± 10.1 | 0.002 |
BMI (kg/m
2
)
| 31.1 ± 8.6 | 29.3 ± 5.4 | 0.029 |
Height (cm)
| 162.8 ± 12.5 | 164.6 ± 8.4 | 0.146 |
Weight (kg)
| 80.8 ± 12.7 | 79.3 ± 14.2 | 0.331 |
Years diagnosed
| 16.0 ± 8.6 | 13.5 ± 7.8 | 0.011 |
Duration of monitoring period (months)
| 9.3 ± 1.9 | 9.1 ± 2.0 | 0.291 |
HbA1c (%)
| 8.18 ± 1.33 | 8.53 ± 1.29 | 0.019 |
FBG (mg/dl)
| 168 ± 47.2 | 182.6 ± 57.1 | 0.025 |
Data were collected from the medical records for all studied variables and for total insulin dose administered, number of injections and changes in body weight during the monitoring period. Regarding hypoglycaemia, we reported the number and the severity of hypoglycaemic episodes as well nocturnal hypoglycaemia. Severe hypoglycaemia was defined as any hypoglycaemic episode that the patient was unable to self-treat requiring the assistance of another person to deal with it[
5]. All the other cases of hypoglycaemia (documented symptomatic hypoglycaemia, asymptomatic hypoglycaemia, probable symptomatic hypoglycaemia and relative hypoglycaemia) were considered as mild/moderate hypoglycaemia[
5].
Statistical analysis
Data were analyzed separately in each treatment group (control/active group). Continuous variables were presented by measures of central tendency and dispersion, whereas categorical variables by frequency distribution tables. The comparability of the two treatment groups regarding baseline data was assessed by means of t-student distribution tests for continuous variables and via chi-square Pearson’s tests of independence or Fisher’s exact tests for categorical variables, respectively. The least squares (LS) mean differences and two-sided 95% confidence intervals (CI) were estimated for the comparisons of the changes in HbA1c, FBG and body weight between the active and the control group.The comparability of treatment groups regarding the number of hypoglycaemic incidences was assessed by means of Wilcoxon rank-sum test. The evaluation of the change in the proportion of patients receiving orally administered antidiabetic treatment from the start until the end of the monitoring period was performed by means of McNemar test. The statistically significant variables were identified according to the analysis of covariance (ANCOVA) methodology. The statistical significance of body weight changes in each treatment group was evaluated by one-way analysis of variance (ANOVA) tests. The hypothesis of equal hypoglycaemic incidence rates between the two treatment groups was assessed by the estimation of relative risk (RR) and 95% confidence intervals (CI), while the hypothesis of equal rapid acting insulin intake was assessed by the estimation of odds ratio (OR, 95% CI). All statistical tests were two-tailed and the significance level was set to 5%.
Discussion
The importance of good glycaemic control constitutes the cornerstone of every therapeutic modality in patients with diabetes. Prospective randomised studies have clearly shown that effective diabetes control, as judged by HbA1c, is associated with lower incidence of chronic complications, especially microangiopathy, both in patients with type 1 and type 2 diabetes[
6,
7]. The same applies to satisfactory targeting of postprandial hyperglycaemia that constitutes an independent risk factor for the development of diabetic complications[
8].
The therapeutic goals in terms of HbA1c level s and of fasting and post prandial blood glucose values have been based on large epidemiological studies associating glucose control with the incidence of diabetic complications, although the therapeutic target and timing of intervention for the prevention of macroangiopathic complications have not been well documented compared to relevant action for the reduction of microangiopathy complications[
9‐
11].
The need for effective glycaemic control in the early stages of the disease has modified treatment of T2DM in patients not adequately controlled with life style changes and oral hypoglycaemic agents, bringing insulin into prominence as an effective therapeutic intervention even in the early stages of the disease and before secretory failure of β-cells occurs. The combination of oral hypoglycaemic agents and insulin administration has become a favourable type of treatment achieving effective control with lower insulin dosage in T2DM. The benefits of insulin administration on preservation of β-cell secretory capacity, the antinflammatory action and the dose dependent hypoglycaemic activity are combined with the favourable effects of some OAD like metformin on insulin resistance and glucagon-like-peptide-1 (GLP-1) agonists and dipeptyl-peptidase-4 (DPP-4) inhibitors on β-cell mass preservation and appetite as well as on weight control[
12,
13].
The advent of insulin analogues and especially the basal insulins have facilitated initiation of insulin treatment and reduced the incidence of hypoglycaemic episodes which was a certain barrier for insulin therapy in the early stages of T2DM[
14]. The treatment guidelines suggested by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD)[
1,
2,
15] and the introduction of simple instructions for insulin titration[
14‐
18] resulted in an increase of the use of insulin in patients with T2DM.
Insulin may be given in different therapeutic schemes either alone or in combination with OAD depending on the severity of the disease and appointed glycaemic targets. It can be administered once daily as basal insulin usually in combination with 1 or 2 OAD, basal insulin in combination with fast acting insulin, premeal bolus short acting insulin or in the form of a premixed intermediate acting/fast acting combinations (biphasic) insulin[
15].
In the present study we assessed the efficacy and safety of two insulin regimens in subjects with T2DM who had been already on treatment with biphasic human insulin in combination with OAD, without achieving optimal glycaemic control.
A total of 142 patients who had switched treatment from biphasic human insulin to insulin glargine daily were included in the active group and compared to 159 patients who continued treatment with biphasic human insulin bd and were included in the control group. We have shown that the reduction in HbA1c and fasting BG values achieved were significantly greater in the active than in the control group. In particular, HbA1c declined by 1.13% in the active and by 0.59% in the control group and fasting BG declined by 47 mg/dl in the active and 20 mg/dl in the control group.
It must also be emphasized that despite more intensive management in the active group, the number of hypoglycaemic episodes did not differ and body weight did not increase in comparison with the control group. Moreover, only one third of the patients in the active treatment group required three injections of fast acting insulin in addition to basal insulin glargine, the dosage of which was considerably reduced at the end of the monitoring period as a consequence of the addition of fast acting insulin.
Well controlled studies on the comparative evaluation of the effectiveness and safety of the different insulin schemes are limited. Raskin et al.[
19] in a 28-week parallel group randomized study involving 233 insulin-naïve patients with T2DM on treatment with metformin, found that biphasic insulin twice daily was more effective in lowering HbA1c compared to treatment with once daily insulin glargine. However, weight gain was greater and minor hypoglycaemic episodes were more often in the group treated with biphasic insulin. Similar results were presented by Holman et al.[
17] in an open label one year controlled study involving 708 patients on metformin and sulfonylurea treatment randomized in three insulin treatment groups, biphasic insulin twice daily, fast acting insulin three times daily or basal insulin once daily (or twice if required). However, their results were modified three years after randomization and patients who added a basal or fast acting insulin-based regimen had better glycaemic control than patients who added biphasic insulin[
20].
Another study examined the efficacy and safety of two treatment regimens (intensified basal-bolus glargine/glulisine regimen and biphasic insulin twice daily) for 52 weeks in 312 patients with long-standing T2DM initially treated with biphasic insulin. The study showed that patients allocated in the basal-bolus regimen had superior glycaemic control vs. those allocated in the premix therapy with no increase in the rates of hypoglycaemia[
21]. In addition more patients reached HbA1c ≤ 7% in the basal-bolus than in the biphasic treatment group (46.6% vs. 27.9%).
In a meta-analysis, Lasserson et al.[
22] reviewed 22 trials that randomized 4,379 insulin-naïve patients. They found greater HbA1c reductions with biphasic and fast acting insulin compared with basal insulin, but at the end of the study larger doses were needed in the biphasic and fast acting arms with minor hypoglycaemic events being inconsistently higher in biphasic and fast acting groups and weight gain greater in the fast acting group compared to the group on basal insulin.
In our study the addition of fast acting insulin when necessary to basal insulin glargine resulted in greater HbA1c and fasting BG reduction, which almost reached the suggested by the ADA/EASD guidelines levels[
1,
2,
15]. This was achieved with no increase in weight or in the rate of hypoglycaemic episodes. The recent guidelines for the management of T2DM suggest that not all patients need or benefit from aggressive glucose management and that it is important to individualize treatment targets[
15]. Because most of the T2DM patients maintain some endogenous insulin secretion even in late stages of disease, addition of basal insulin [neutral protamine Hagedorn (NPH) or long-acting insulin glargine or insulin detemir)] to the OAD if glucose control is inadequate is an effective, safe and simple approach, unless the patient is markedly hyperglycemic[
15]. Our results showed that almost 30% of the participants managed to achieve glucose targets only with insulin glargine once daily. However, almost 40% of the participants in the control and 54% of the participants in the active group needed 3or more insulin injections per day. Moreover, we found that with basal insulin regimens, the addition of medications-beyond metformin-that increase insulin secretion like sulfonylureas or meglitinides is necessary in some patients for the maintenance of glycaemic control; thus, more patients in the active group received such medications during follow-up. Therefore, our data suggest that addition of basal insulin to OAD is an effective approach for the management of hyperglycaemia in T2DM and intensification of insulin therapy with the addition of fast-acting pre-meal insulin may be necessary to achieve glucose targets.
The combination of basal insulin with fast acting insulin when needed offers greater flexibility and better glycaemic control mimicking the physiologic insulin secretion. Due to its pharmacodynamic profile, insulin glargine substitutes basal insulin secretion effectively controlling morning glycaemia, whereas fast acting insulin reduces postprandial hyperglycaemia.
The strength of this study is that it examined in everyday clinical practice with no intervention at all and with adequate power the effectiveness of intensification of treatment with insulin based on two insulin regimens (basal insulin glargine vs. biphasic human insulin) on glycemic control. In addition, the study has adequate power to support the findings. However, this study being observational and retrospective by design has several limitations. First, the two groups differ significantly at baseline in terms of age, BMI, known duration of diabetes and diabetes control. Moreover, at the end of the monitoring period there were significant differences in concomitant treatments for diabetes; thus the results should be interpreted taking into consideration the changes in medications used for the treatment of diabetes during the study. Second, the number of mild/moderate hypoglycaemic episodes might have been under-reported in medical records and therefore, underestimated. Third, detailed data for documented symptomatic hypoglycaemia, asymptomatic hypoglycaemia, probable symptomatic hypoglycaemia and relative hypoglycaemia were not available and all cases of hypoglycaemia other than severe hypoglycaemia were classified as mild or moderate hypoglycaemia. Fourth, causality between the treatments and the outcomes cannot be established and these findings must be confirmed by a randomized clinical trial.
Study investigators
N. Tentolouris and I. Mpalla, 1st Department of Propaedeutic and Internal Medicine, Athens University Medical School, Laiko General Hospital, Athens; V. Kyriazopoulou and M. Michalaki, Department of Internal Medicine, Division of Endocrinology and Diabetes, University Hospital of Patras, Patras; I. Ioannidis, 2nd Department of Internal Medicine, Konstantopoulio Hospital, Nea Ionia; E. Pagkalos, 1st Department of Internal Medicine, Papageorgiou Hospital, Thessaloniki; C. Vasilopoulos Department of Endocrinology, General Hospital of Athens “Evangelismos”, Athens; V. Klisiaris, Department of Internal Medicine, General Hospital of Larisa “Koutllibanio-Triantafilio”, Larisa; D. Makrigiannis, Department of Internal Medicine, General Hospital of Ioannina “Hatzikosta”, Ioannina; I. Mallias, Second Department of Internal Medicine General Hospital of Serres, Serres; K. Kazakos, Internist, Diabetologist, Thessaloniki; A. Likoudi Internist - Diabetologist, Marathonas, Athens.
Competing interests
Nicholas Tentolouris has received honoraria from NovoNordisk, Sanofi, Elli-Lilly, Bristol-Myers-Squibb and Novartis for memberships on advisory boards and for speaking on subjects related to treatment with antidiabetic drugs and insulin. Venetsana Kyriazopoulou declares no conflict of interest. Dimitrios Makrigiannis has received honoraria from NovoNordisk, Sanofi, Elli-Lilly and Novartis for memberships on the advisory boards and for speaking on subjects related to treatment with antidiabetic drugs and insulin. Barbara Baroutsou is an employee of Sanofi working as Medical Director for Greece and Cyprus.
Authors’ contributions
NT was principal investigator, study co-ordinator and investigator, participated in all stages of recruitment of the patients and in analysis of the data, drafted and reviewed critically the manuscript. VK was study co-ordinator and investigator, participated in all stages of recruitment of the patients and in analysis of the data, drafted and reviewed critically the manuscript; DM was study investigator, participated in all stages of recruitment of patients and reviewed critically the manuscript; BB was study sponsor co-ordinator, wrote the protocol, drafted and reviewed critically the manuscript. All other study investigators conducted the study and collected the data. All authors read and approved the final manuscript.