SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

Aims:  The choice of insulin at initiation in type 2 diabetes remains controversial. The aim of this study was to assess the occurrence of self-reported severe hypoglycaemia associated with premixed insulin analogues in routine clinical care.

Methods:  A 12-month, prospective, observational, multicentre study in patients starting a commonly prescribed premixed insulin analogue (either insulin lispro 25/75 or biphasic insulin aspart 30/70, twice daily) after suboptimal glycaemic control on oral antidiabetic agents. Treatment decisions were made solely in the course of usual practice.

Results:  Study follow-up was completed by 991 (85.5%) of the 1150 patients enrolled. At baseline, mean (SD) age was 57.9 (10.1) years; mean diabetes duration was 9.2 (5.9) years; mean haemoglobin A1c (HbA1c) was 9.9 (1.8) % and the rate of severe hypoglycaemia was 0.03 episode/patient-year. At 12 months, the rate of severe hypoglycaemia was 0.04 episode/patient-year (95% CI 0.023, 0.055 episode/patient-year) and mean insulin dose was 41.5 (19.4) units. Changes from baseline to 12 months for mean fasting plasma glucose and HbA1c were −5.1 mmol/l and −2.5%, respectively.

Conclusions:  After initiation of premixed insulin analogues in patients with type 2 diabetes in real-world settings, the incidence of severe hypoglycaemia was lower than expected from previously reported studies.


What’s known

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

Hypoglycaemia is the most important limiting factor in achieving good glycaemic control of type 2 diabetes, and insulin treatment is associated with the highest risk of severe hypoglycaemia compared with other treatment regimens. As the controlled environments of clinical trials may not be fully applicable to real-world patient care, the incidence of severe hypoglycaemia with premixed insulin analogues in multicultural clinical practice remains uncertain.

What’s new

In this observational study, patient/physician decisions to initiate insulin treatment with twice daily (BID) dosing of premixed insulin analogues reflected choices in real-world settings in Africa, Asia and Europe. During the first year after initiation, glycaemic control improved considerably, and incidence of severe hypoglycaemia was lower than expected, based on previously reported studies. High mean HbA1c levels at initiation were consistent with other findings, representing opportunities for improving patient care.

Introduction

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

According to the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) expert consensus statement, the key objective of type 2 diabetes therapy is to achieve and maintain good glycaemic control with Haemoglobin A1c fraction (HbA1c) level < 7% for most patients (1). After lifestyle changes and combinations of oral antidiabetic medications (OADs), early initiation of insulin therapy is recommended to meet this goal. In numerous countries, initiating insulin therapy with premixed insulin preparations in combination with OAD therapy is a common clinical practice.

Nevertheless, insulin treatment in type 2 diabetes is associated with the highest risk of severe hypoglycaemia compared with other treatment regimens (2). Hypoglycaemia is the most important limiting factor in achieving good glycaemic control of diabetes (3,4), as it is associated with a number of adverse clinical outcomes (5) as well as the psychological fear of hypoglycaemic episodes that compromises patient adherence (6). With potentially less risk of hypoglycaemia, both rapid-acting and long-acting insulin analogues have become an attractive alternative for human insulin (7,8).

The choice of insulin initiation regimen after OAD failure in type 2 diabetes remains controversial. In a clinical trial setting, mild hypoglycaemic episodes were more frequent in patients treated with biphasic insulin in comparison with those treated with basal insulin (9,10). However, the controlled environments of clinical trial settings and cohorts may not be fully applicable to real-world patient care (11). Thus, it is unclear whether the incidence of severe hypoglycaemia is less in clinical trials than in usual clinical practice (12). For this reason, the incidence of severe hypoglycaemia with premixed insulin analogues in multicultural clinical practice remains uncertain. The objective of the present observational, multicentre study was to estimate the incidence of severe hypoglycaemia and other treatment outcomes in patients with type 2 diabetes not controlled on oral agents over the first year of insulin therapy. Patient/physician decisions to initiate insulin treatment with twice daily (BID) dosing of premixed insulin analogues reflected choices in real-world settings.

Subjects, materials and methods

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

Study design

A prospective, observational, multicentre study was conducted in outpatient settings in 10 countries in Africa, Asia and Europe. Patients with suboptimal glycaemic control on OADs who, together with the treating physician, chose to initiate insulin therapy with a premixed insulin analogue were included in the study. At baseline (insulin initiation), patients either continued or stopped OAD therapy. A commonly prescribed premixed insulin analogue was used, either as monotherapy or in combination with OAD. After insulin initiation, patients were observed over a period of 12 months. As neither randomisation nor intervention was included in this study design, no treatment groups were planned, and patient-care decisions were made solely in the course of standard clinical practice. Patients were treated in outpatient settings by general practitioners, internists or diabetes specialists.

This study was conducted in accordance with the ethical principles that have their origin in the Declaration of Helsinki and is consistent with good clinical practices and applicable laws and regulations of the countries in which the study was conducted. At least one ethical review board per country approved the study description and confirmed that the study was considered non-interventional. All participants gave informed written consent.

Patients

All patients had type 2 diabetes according to the World Health Organization classification. Inclusion criteria: 18 years or older, a body mass index lower than 40 kg/m2, treatment with one or more OADs without insulin for at least 3 months immediately before the study, and the requirement of premixed insulin to achieve metabolic control in the opinion of the investigator. Exclusion criteria: treatment with glucocorticosteroids, prescription of weight-loss medications or thiazolidinediones for over 2 weeks during the 3 months immediately before the study.

Procedure

Investigators were free to initiate premix insulin treatment either by administration of insulin lispro low mixture 25/75 (75% insulin lispro protamine suspension, 25% insulin lispro) or biphasic insulin aspart 30/70 (70% insulin aspart protamine suspension, 30% insulin aspart) BID, before morning and evening meals. The choice of therapies was made by the physician in conjunction with the patient. There were no targets set for recruitment for either of the two treatment formulations.

Patient recall information for the 3 months prior to the study was collected at baseline (0 months). Subsequently, patient information since last visit was collected at Visit 2 (4 [± 1] months), Visit 3 (8 [± 1] months) and Visit 4 (12 [± 1] months). Participating investigators recorded observations at each visit in conjunction with standard care only; no appointments specifically for the purpose of data collection were scheduled.

Assessment of severe hypoglycaemia

The occurrence of severe hypoglycaemia was self-reported during each visit. In agreement with the commonly accepted definition, severe hypoglycaemia was defined as an event that required assistance from another person who actively administered carbohydrate, glucagon or other resuscitative actions and was associated with either a blood glucose level less than 3.9 mmol/l (< 70 mg/dl) or prompt recovery after restoring normoglycaemia (13).

Study measures

The following measures routinely available in clinical practice were collected: occurrence and timing of severe hypoglycaemia since last visit (or over the 3 months prior to study for baseline), HbA1c, fasting blood glucose, body weight, waist and hip circumference and vital signs. Insulin dose and changes in OADs and concomitant therapy were collected after baseline at Visits 2, 3 and 4. Fasting triglycerides and cholesterol values (high-density lipoprotein [HDL], low-density lipoprotein [LDL] and total cholesterol) were collected at baseline and Visit 4 only in countries in which such measures are included in standard clinical practice. Reporting of compulsory blood glucose self-monitoring was not included in the study protocol.

A change in insulin treatment was defined as a dose increase, dose decrease, addition of insulin (premixed or non-premixed insulin) or cessation of insulin (premixed or non-premixed insulin). A change in OAD treatment was defined as a dose increase, dose decrease or cessation of OAD.

Statistical methods

Based on the incidence of severe hypoglycaemia in insulin-treated patients in the United Kingdom Prospective Diabetes Study (2), we expected 2.3% of patients to report at least 1 episode of severe hypoglycaemia over 1 year. We planned to enrol 1125 patients, assuming a dropout rate of up to 20%. The resulting evaluable 900 patients would permit the estimate of a two-sided 95% confidence interval of width ± 1.0%. As a supporting analysis, all eligible patients starting premixed insulin analogues were used. Episodes of self-reported, severe hypoglycaemia per patient-years were estimated from the total time on premixed insulin analogues, including a 95% bootstrap confidence interval.

For baseline characteristics and secondary objectives, summary statistics were calculated. Mean and standard deviation for continuous variables and proportions for categorical variables were derived. For means and percentages, two-sided, 95% confidence intervals (normal approximation) were calculated. For continuous variables, changes from baseline were assessed by paired t-tests. Summary statistics for 2-h postprandial blood glucose levels after morning and evening meals as well as blood-lipid values (triglycerides and total, LDL and HDL cholesterol) were assessed.

Results

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

Patient disposition and baseline characteristics

Of 1150 patients who enrolled, 11 were excluded or not treated, 1139 were treated with insulin, 148 discontinued and 991 completed the study (Figure 1). Reasons for discontinuation were: lost to follow-up (n = 63), subject decision (n = 41), physician decision (n = 32), death (n = 8), adverse event (n = 3) and lack of efficacy (n = 1).

image

Figure 1.  Patient disposition

Download figure to PowerPoint

Overall, 86.2% of patients were reported to be of Caucasian race with 52.2% men and 47.8% women (Table 1). At baseline, mean BMI was 29.2 kg/m2 and the mean HbA1c was 9.9%.

Table 1.   Demographics and patient characteristics at baseline
CharacteristicPatients, N = 1139
  1. Data are mean (SD) unless otherwise indicated. *Indian subcontinent. n = 1098. Either plasma or whole-blood glucose was measured. §n = 878. n = 256. **734 patients had observations available on pre-existing conditions, 98 patients had no pre-existing conditions and 405 patients had unknown pre-existing conditions. Each of seven patients reported only one episode of severe hypoglycaemia in the 3 months before study entry. Per cent of total number of patients (7) who experienced severe hypoglycaemia during 3 months immediately before the study. §§Ongoing or newly prescribed oral antidiabetic medications (OADs), not those discontinued at baseline. All patients were previously treated with OADs, and 943 patients either continued treatment with oral agents or received newly prescribed oral agents at baseline. ***Per cent of total number of patients (943) treated with OADs from baseline.

Countries, n (%)
 Egypt365 (32.0)
 Central and Eastern European countries (Croatia, Czech Republic, Estonia, Latvia, Poland, Romania, Slovak Republic) 489 (42.9)
 Turkey154 (13.5)
 China131 (11.5)
Gender, n (%)
 Male 595 (52.2)
 Female 544 (47.8)
Race, n (%)
 Caucasian 982 (86.2)
 East Asian 131 (11.5)
 South Asian*21 (1.8)
 African 3 (0.3)
 Hispanic 2 (0.2)
Age, years 57.9 (10.1)
Diabetes duration, years 9.3 (5.9)
Body weight, kg 81.5 (15.9)
Body mass index, kg/m2 29.2 (4.7)
Haemoglobin A1c, % 9.9 (1.8)
Fasting plasma glucose, mmol/l 12.7 (3.8)§
Fasting whole-blood glucose, mmol/l 13.1 (16.4)
Patients with ≥ 1 pre-existing condition/disorder, n (%) 636 (55.8)**
Patients with ≥ 1 severe hypoglycaemic episode in the last 3 months prior to study, n (%) 7 (0.6)
Nocturnal (after bedtime and before waking)4 (57)
Patients with ≥ 1 oral antihyperglycaemic medication§§, n (%) 943 (82.8)
 One 658 (69.8***)
 Two 263 (27.9***)
 Three 21 (2.2***)
 Four 1 (0.1***)
 Metformin 664 (70.4***)
 Sulphonylurea 400 (42.4***)
 Alpha-glucosidase inhibitors 55 (5.8***)
 Glinides 49 (5.2***)
 Thiazolidinediones 43 (4.6***)
Insulin prescribed at baseline, IU
 Morning (prebreakfast) 20.8 (8.4)
 Evening (predinner) 15.0 (6.6)
 Morning + Evening 35.8 (13.7)

Incidence of severe hypoglycaemia

In this study population in the 3 months prior to baseline, seven patients (0.6%) experienced a single episode of severe hypoglycaemia. Thus, the rate of severe hypoglycaemia during the 3 months prior to the study was 0.03 episode/patient-year (95% CI 0.01–0.04 episode/patient-year).

During the study, 28 patients (2.5%) experienced episodes of severe hypoglycaemia (Figure 2). Twenty patients (71.4% of those who experienced episodes of severe hypoglycaemia) experienced one episode each, seven patients (25.0%) experienced two episodes and one patient (3.6%) experienced three or more episodes. Thus, the rate of severe hypoglycaemia during the study was 0.04 episode/patient-year (95% CI, 0.02–0.06 episode/patient-year).

image

Figure 2.  Severe hypoglycaemia. *The timing of two episodes of severe hypoglycaemia was not reported (unknown). All patients recovered promptly

Download figure to PowerPoint

A total of 37 episodes of severe hypoglycaemia were reported. Thirty-six episodes (97.3% of total episodes) were confirmed by blood glucose readings. Seventeen episodes (45.9%) occurred between breakfast and lunch, seven (18.9%) occurred between waking and breakfast, five (13.5%) occurred between evening meal and bedtime, three (8.1%) occurred at night, two (5.4%) occurred between lunch and dinner and one (2.7%) occurred during or after exercise. Nine episodes (25.0%) resulted in changed insulin regimens, three (8.8%) resulted in hospitalisation and one (2.9%) resulted in an emergency room visit. During 25 episodes (67.6%), patients were aware of hypoglycaemic symptoms; and during 12 episodes (32.4%), patients were not aware of symptoms. All patients recovered.

For the subgroup of patients who initiated insulin as combination therapy with OAD (N = 824), 22 patients (2.7%) experienced at least one episode of severe hypoglycaemia. For the subgroup that initiated insulin as monotherapy (N = 315), six patients (1.9%) experienced at least one episode of severe hypoglycaemia.

To assess whether concomitant use of insulin with insulin secretagogues increased the risk of severe hypoglycaemia in our study, we conducted a post hoc analysis. Because of the low number of reported episodes, we pooled data for patients treated with sulphonylureas and glinides. Of the 28 patients who reported episodes of severe hypoglycaemia, five patients (17.9%) were known to have been treated with an insulin secretagogue on the day that severe hypoglycaemia was reported, and eight patients (28.6%) were known to have been treated with none of the insulin secretagogues at any time during the study. There were 12 patients (42.9%) who were treated with an insulin secretagogue during the study but had stopped treatment prior to the reported episode: 11 patients (39.3%) stopped at least 15 days prior and one patient (3.6%) stopped at least 3 days prior to the episode. Three patients (10.7%) were treated with insulin secretagogues during the study, but missing treatment dates made it unclear whether secretagogue use was concomitant on the days of the reported episodes.

Insulin therapy

The choice of premixed insulin analogue was solely that of the patient and/or physician. At insulin initiation (baseline), 964 patients (84.6%) were prescribed insulin lispro 25/75, and 175 patients (15.4%) were prescribed insulin aspart 30/70. The total (morning + evening) mean (SD) dose of premixed insulin analogue at insulin initiation was 35.8 (13.7) units (20.8 [8.4] units before the morning meal and 15.0 [6.6] units before the evening meal). The total mean dose of premixed insulin analogue at 12 months for patients who completed the study (n = 991) was 41.5 (19.4) units (intra-patient change from baseline, + 5.6 [15.5] units).

Overall, 729 patients (64.0%) experienced at least one change in insulin treatment throughout the entire study, and 410 patients (36.0%) experienced no change throughout the entire study. Four patients (0.4%) intensified from low mix BID to a basal/bolus regimen, and 42 patients (3.7%) intensified from low mix BID to low mix thrice daily (TID). Of patients who experienced change in insulin treatment, 635 (87.1%) increased insulin dose, 174 (23.9%) decreased insulin dose and 17 (2.3%) terminated at least one insulin.

Oral antidiabetic therapy

All insulin-treated patients (N = 1139) were being treated with one or more OADs prior to baseline (insulin initiation). A total of 377 patients stopped at least one OAD at baseline, the most common of which was sulphonylurea (288 patients, 79.3% of those who stopped an OAD). Ninety-three patients added at least one OAD at baseline, the most common of which was metformin (52 patients, 55.9% of those who added an OAD). At baseline, 943 patients (82.8% of insulin-treated patients) either continued their previous OAD treatment or added at least one OAD and thus initiated insulin therapy as combination therapy. The remaining 196 patients (17.2%) were not treated with OAD at baseline and thus initiated insulin as monotherapy.

Most patients (1052 or 92.4% of insulin-treated patients) had no change in OAD treatment throughout follow-up; whereas, 25 patients (2.2%) added at least one OAD, the most common of which was metformin (14 patients, 56.0% of those who added OAD). Thus, either ongoing or added during the study, 951 patients (83.5% of insulin-treated patients) received at least one OAD in combination with insulin treatment. Of these, 675 patients (71.0%) received metformin, 402 (42.3%) received a sulphonylurea, 59 (6.2%) received an alpha-glucosidase inhibitor, 51 (5.4%) received a glinide and 46 (4.8%) received a thiazolidinedione.

Metabolic control

HbA1c

At baseline, the mean (SD) HbA1c level was 9.9% (1.8%) (Figure 3A). At 12 months, the mean HbA1c level was 7.3% (1.1%), a mean intrapatient decrease from baseline of −2.5% (95% CI, 2.4–2.7%). The percentages of patients who achieved the HbA1c targets of ≤ 6.5% and < 7.0% are shown in Figure 3B.

image

Figure 3.  Glycaemic control. (A) HbA1c over 12 months. (B) Patients to HbA1c targets of ≤ 6.5% and < 7.0%. For A and B, n = 1098 at baseline, n = 973 at 4 months, n = 920 at 8 months and n = 951 at 12 months. (C) Plasma and whole-blood glucose levels over 12 months; for plasma glucose, n = 878 at baseline, n = 681 at 4 months, n = 658 at 8 months and n = 686 at 12 months; for whole-blood glucose, n = 256 at baseline, n = 396 at 4 months, n = 350 at 8 months and n = 289 at 12 months. Means (SD) are shown for 3A and 3C

Download figure to PowerPoint

Fasting blood glucose

After initiation of insulin therapy, mean fasting blood glucose decreased progressively over the study period (Figure 3C). At baseline, mean (SD) fasting whole-blood glucose was 13.1 (16.4) mmol/l, and mean fasting plasma glucose was 12.7 (3.8) mmol/l. At 12 months, mean fasting whole-blood glucose was 7.4 (2.1) mmol/l (a mean intrapatient decrease from baseline of 4.9 mmol/l [95% CI, 4.2–5.5 mmol/l]) and mean plasma glucose was 7.5 (2.4) mmol/l (a mean intrapatient decrease from baseline of 5.1 mmol/l [95% CI, 4.8–5.4 mmol/l]).

Anthropometric measures

Over 12 months, patients experienced a mean (SD) weight gain of +1.5 (4.0) kg with corresponding mean increases in body mass index (+ 0.6 [1.5] kg/m2) and waist circumference (+ 0.7 [4.0] cm) (Table 2). Waist-to-hip ratio did not change significantly throughout the study.

Table 2.   Anthropometric and blood-lipid measures
CharacteristicBaseline mean (SD) [n]After 12 months mean (SD) [n]Change from baseline mean (95% CI) [n]p-value
  1. BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Weight, kg81.5 (15.9) [1139]83.6 (14.9) [978]+ 1.5 (1.3, 1.8) [978]< 0.001
BMI, kg/m229.2 (4.7) [1138]29.9 (4.4) [978]+ 0.6 (0.5, 0.7) [978]< 0.001
Waist circumference, cm101.6 (14.9) [1078]102.9 (14.5) [903]+ 0.7 (0.5, 1.0) [897]< 0.001
Blood-lipids, mmol/l
Total cholesterol11.5 (2.6) [1057]10.6 (1.9) [806]− 1.0 (− 0.8, − 1.11) [778]< 0.001
HDL2.5 (0.7) [988]2.6 (0.9) [791]+ 0.1 (+ 0.2, + 0.05) [749]< 0.001
LDL7.0 (7.1) [982]6.4 (1.7) [791]− 0.6 (− 0.5, − 0.7) [737]< 0.001
Triglycerides11.1 (7.4) [1050]9.3 (4.6) [813]− 2.1 (− 1.6, − 2.5) [786]< 0.001
Adverse events

Forty-four patients reported a total of 59 serious adverse events or adverse events leading to discontinuation, including eight deaths. Twenty-six events resulted in hospitalisation.

Overall, 388 adverse events were reported. Forty-six events (11.9%) were considered severe, 51 (13.1%) were considered moderate and 291 (75.0%) were considered mild. One hundred eighty nine events (48.7%) were considered treatment related, and four (1.0%) were considered injection-device related.

Discussion

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

Hypoglycaemia remains the most critical factor in achieving good glycaemic control in patients with type 2 diabetes (3,4). In part, because of the risk and management of severe hypoglycaemia and the resulting changes in patient behaviour to avoid subsequent episodes, timely initiation of and choice of insulin regimen have been controversial. With potentially less hypoglycaemic risk, insulin analogues offer an alternative to human insulin (7,8).

Previous studies have reported incidences of severe hypoglycaemia from 0.12 to 0.35 episode/patient-year (14–18) and rates from 0% to 15% (16–21) in insulin-treated patients. An increased risk of severe hypoglycaemia has been observed with increased age (15,21) and longer duration of diabetes or insulin treatment (15,17,18). In our study, severe hypoglycaemia was rare, 0.04 episode/patient-year, whereas the baseline occurrence before the initiation of insulin analogues was 0.03 episode/patient-year. This finding is lower than findings previously reported in most observational or population-based analyses for patients treated with insulin (14–18). In other findings, recently published data from Sweden reported the occurrence of two major episodes of hypoglycaemia over 6 months in a cohort of 1154 patients (22). Since severe hypoglycaemia was self-reported in the present study, we cannot exclude the possibility that the much lower rate of severe hypoglycaemia in this study compared with other studies is at least partially explained by underreporting.

Wrong medication dose, missed meals and excessive dieting/weight loss have been identified by patients as reasons for severe hypoglycaemia (20). In the present study, patient-reported reasons for episodes of severe hypoglycaemia were not collected. However, since almost half (46%) of the total episodes occurred between breakfast and lunch, missed breakfast or additional, unplanned physical activity may possibly explain some of the episodes observed. Similarly, people with diabetes, like others, may take their main meal in the evening. Thus, the carbohydrate intake during breakfast may be lower than recommended, which may increase the likelihood of hypoglycaemia between breakfast and lunch. Individual insulin pharmacodynamics and prelunch insulin requirements might also influence the likelihood of hypoglycaemia (23).

Although international guidelines recommend the use of insulin therapy early in the treatment paradigm, delay in initiating insulin therapy is not uncommon. Results of analyses of patients in the UK demonstrated that insulin therapy was initiated approximately 11 years after oral agents were first prescribed (24) and 5 years after the loss of glycaemic control (25). Reasons for delay in insulin therapy are varied and may include resistance on the part of both patients and physicians (26). In the prospective, observational INSTIGATE study in five European countries (27), the mean HbA1c level at insulin initiation was 9.6% at approximately 10 years since diagnosis. Such findings correspond to the “real-life” observation in our multiethnic outpatient study. At insulin initiation in this study, the mean HbA1c value was approximately 10.0%, indicating that patients did not initiate insulin therapy until hyperglycaemia was markedly high.

In spite of low and modest increases in insulin doses (a mean intrapatient change from baseline of + 5.6 units), overall glycaemic control improved markedly (a mean intrapatient decrease from baseline in HbA1c of −2.5%), and more than one-third of patients achieved the recommended HbA1c level of less than 7% at the end of observation. In general, patients also gained weight, which is common with insulin therapy, particularly during the first year of treatment (28). An increase in mean hip-to-waist ratio was not observed.

The choice of treatment regimen at insulin initiation in type 2 diabetes is still a matter of controversy (1,29). In the INITIATE and 4-T clinical trials, premixed insulin analogues BID added to OADs appeared to be superior to basal insulin once daily for metabolic control and similar in terms of severe hypoglycaemic episodes during the first 26–52 weeks of treatment (9,10). However, treatment with premixed insulin analogues also was associated with more weight gain in these studies. These data were confirmed in a meta-analysis by Lasserson et al. (30) and corroborate the results of our study in real-life settings in which initiation of insulin treatment with premixed insulin analogues BID resulted in improved glycaemic control and a low incidence of severe hypoglycaemia. Given these findings, in patients with elevated HbA1c, insulin initiation with premixed analogues BID potentially may be more appropriate than basal insulin once daily.

A limitation of this study includes recall bias, a common limitation in observational studies. Although all treatment decisions were made by physicians and patients in the course of routine clinical care, there was an imbalance of the treatment arms in favour of the study sponsor; since, 85% of patients were treated with insulin lispro 25/75 and 15% with biphasic insulin aspart 30/70. However, available data do not indicate that the incidence of hyperglycaemic episodes was different between groups (27). In addition, patient numbers were unevenly distributed across participating countries, which also practiced varied standards of care; and no central laboratory was used. For these reasons, the results of the present study should be interpreted with caution. Strengths of this study include its prospective, observational study design and the inclusion of patients from countries with various ethnic and cultural backgrounds.

In conclusion, during the first year after initiation of insulin therapy with premixed insulin analogues in patients with type 2 diabetes in real-world settings, the observed incidence of severe hypoglycaemia was lower than expected from previously reported studies, although glycaemic control improved considerably during the study period. High mean HbA1c levels at insulin initiation are consistent findings that represent opportunities for improving patient care, as earlier and more aggressive intervention in the natural course of type 2 diabetes should prevent excessive HbA1c levels at initiation. However, long-term prospective studies are required to determine optimal insulin initiation in type 2 diabetic subjects.

Acknowledgements

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

This study was supported by Eli Lilly and Company. The following employees of Eli Lilly and Company were involved: Jacek Kiljanski, MD, for study design; Helmut Petto, MSc, for study design, analysis planning and implementation; Matthew Reaney, MSc, for data interpretation and Simon Cleall, MSc, for data interpretation. The authors thank the study investigators, who are listed in the Appendix.

Author contributions

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

VP, HED, MD, MSG, ER and EM were involved in the acquisition of data, and all authors were involved in the interpretation of data. JG drafted the manuscript, and all authors were involved in the critical revision and approval of the manuscript.

References

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

Appendix

  1. Top of page
  2. Summary
  3. What’s known
  4. Introduction
  5. Subjects, materials and methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Author contributions
  10. References
  11. Appendix

The authors thank the study investigators: Hana Romanova, Jana Psottova, Daniela Kallmunzerova, Sarka Kopecka, Petr Skala, Jana Houdova, Josef Podzimek, Yvona Pospisilova, Petra Vlckova, Josef Klein, Viera Zackova, Jirina Pavlickova, Zdenka Rihova, Adriana Ilavska, Jozef Duda, Miriam Teplanova, Tatiana Kupcova, Julia Krnacova, Karol Rummer, Maria Cizova, Gheorghe Ghise, Ella Pintilei, Rodica Avram, Anamaria Mateescu, Gelu Popescu, Cleo Simionescu, Elena Bacanu, Eva Fulop, Irena Babol, Ilze Konrade, Eve Kelk, Anu Ambos, Ali El Abbasy, Samir Assad, Ali Abd El Reheem, Mohamed Abdelaziz, Ali Sallouma, Omar El Saadany, Ashraf Atteya, Mohamed Khatab, Ashraf Okba, Ahmed Abd El Maksood, Elia Zaki, Saad El Alamy, Makram Sidhom, Magdy Hamed, Shahir Kamal, Sherif Hafez, Omar Heikal, Naglaa El Kabbany, Ebtesam Zakaria, Samir George, Inass Shaltout, Abass Orabi, Megahed Abo El Magd, Khalifa Abdallah, Yigit Sarac, Mehmet Yildiz, Yusuf Bal, Cuneyt Muderrisoglu, Osman Nuri Topal, Adem Yurumez, Nezih Tuncay, Omer Hersek, Caglayan Gerede, Halil Rakici, Esma Altunoglu, Zhong Xiaowei, Zhu Jun, Li Yanbing, Željka Crnčević-Orlić, Dubravka Jurišić-Eržen, Nenad Bićanić and Vlatka Pandžić-Jakšić.