DIABETES TECHNOLOGY & THERAPEUTICS Volume 16, Number 4, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/dia.2013.0246
ORIGINAL ARTICLE
Predictors of Response to Early Basal Insulin Treatment in Patients with Type 2 Diabetes—The EARLY Experience Markolf Hanefeld, MD, PhD,1 Holger Fleischmann, PhD,2 Guido Schiffhorst, MS,3 and Peter Bramlage, MD 4
Abstract Background: It was the aim of this study to assess baseline predictors for glycosylated hemoglobin (HbA1c) reduction, treatment-to-target, and insulin glargine dose in patients with an HbA1c level of ‡ 7.5% (58 mmol/mol) at baseline despite 3 months of maximum tolerated dose of metformin under daily conditions. Subjects and Methods: This was an open, multicenter, prospective observational study with a 6-month follow-up including 1,438 patients with type 2 diabetes. Baseline variables independently associated with HbA1c (overall reduction and achievement of target values) and insulin glargine dose used were determined using a stepwise multivariate linear regression analysis. Results: In a multivariate linear regression analysis (R2 = 0.545) baseline HbA1c (b = - 0.722; P < 0.001) and retinopathy (b = - 0.064; P = 0.007) were associated with a greater HbA1c reduction at 6 months, whereas duration of diabetes was associated with a lesser HbA1c reduction (b = 0.084; P < 0.001). In another multivariate linear regression analysis, weight (odds ratio [OR] 0.99; 95% confidence interval [CI] 0.98 to < 1.00), duration of diabetes (OR 0.96; 95% CI 0.93–0.99), and baseline HbA1c (OR 0.65; 95% CI 0.56–0.76) were associated with a reduced likelihood of achieving an HbA1c level of < 7% (53 mmol/mol); baseline HbA1c (OR 0.66; 95% CI 0.51–0.85) was the only variable associated with a reduced likelihood of achieving an HbA1c level of < 6.5% (48 mmol/mol). In a further analysis (R2 = 0.135) the insulin dose needed was increased in those with a higher body weight (b = 0.230; P < 0.001), a longer duration of diabetes (b = 0.134; P < 0.001), a higher baseline HbA1c level (b = 0.205; P < 0.001), and the presence of microalbuminuria (b = 0.096; P = 0.003). Conclusions: Identified predictors of greater HbA1c reduction, target goal achievement, and insulin dose needed may help to optimize the balance of benefits and risks with the use of insulin glargine. Background
I
nsulin use early in the course of diabetes has been shown to be an effective first-line treatment option.1–4 International and national guidelines recommend metformin as a first step in the treatment of type 2 diabetes.5–8 In the recent European Association for the Study of Diabetes/American Diabetes Association consensus statement, five second-line treatment options are considered on top of metformin.7 Under real world conditions the use of basal insulin is usually confined to those with a glycosylated hemoglobin (HbA1c) level of ‡ 7.5% or even 8.0% (‡ 58 or 64 mmol/mol), and dual or even triple combinations of other treatment options are preferred instead with insulin as the last resort.
Early insulin use, however, may protect b-cells from harmful effects of long-lasting hyperglycemia and lipotoxicity. In patients inadequately controlled by metformin a recovery of residual b-cell function was recently demonstrated by Pennartz et al.9 Chronic long-acting basal insulin improved glucose-induced insulin secretion and postprandial endogenous insulin secretion in hyperglycemic patients with type 2 diabetes.1 Fonsecca et al.,5 in a pooled analysis of randomized trials, found that insulin glargine, given on top of metformin monotherapy early in the treatment, was associated with improved efficacy and safety over a regimen including sulfonylurea. Insulin glargine treatment introduced after dual oral combination therapy with sulfonylurea was associated with less HbA1c reduction but more hypoglycemia. In the
1 Study Centre Professor Hanefeld, Company for the Transfer of Knowledge and Technology of the Dresden University of Technology, Dresden, Germany. 2 Sanofi Aventis Deutschland GmbH, Berlin, Germany. 3 IGES Institut GmbH, Berlin, Germany. 4 Institute for Pharmacology and Preventive Medicine, Mahlow, Germany.
241
242 ORIGIN trial, the early addition of insulin glargine to standard of care was more effective as the comparator standard of care treatment to improve glycemia control with no adverse effects.4 The recent EARLY study documented the use of basal insulin glargine in patients being uncontrolled on a maximal tolerated dose of metformin under real world conditions.6 The HbA1c level was reduced from 8.7% (72 mmol/mol) to 7.4% (57 mmol/mol) within 24 weeks, as was fasting plasma glucose (from 182 to 131 mg/dL [10.1 to 7.3 mmol/l]; P < 0.01). Patients had a low risk of hypoglycemia and experienced a slight decrease of weight of about 1 kg (P < 0.01). The objective of the present exploratory analysis of the EARLY cohort was to assess (1) baseline predictors for insulin glargine dose used and (2) predictors of HbA1c reduction and HbA1c on target in patients having an HbA1c level of ‡ 7.5% (58 mmol/mol) at baseline. Subjects and Methods EARLY was an open, multicenter, prospective observational study with a 6-month follow-up that was planned and conducted according to the requirements of the German Bundesinstitut fu¨r Arzneimittel und Medizinprodukte and the Kassena¨rztliche Bundesvereinigung. Ethical approval was obtained from the Saxonian State Chamber of Medicine (Dresden, Germany). Details of the study protocol have been published previously.6 Patients
HANEFELD ET AL. Table 1. Patients’ Characteristics at Baseline
Variable Age (years) Female gender Body weight (kg) Body mass index (kg/m2) Duration of diabetes (years) HbA1c [% (mmol/mol)] at baseline Fasting blood glucose [mg/dL (mmol/L)] Comorbidities at baseline Coronary heart disease Myocardial infarction Stroke or TIA Peripheral artery disease Diabetic foot syndrome Neuropathy Retinopathy Nephropathy Macroalbuminuria Microalbuminuria
EARLY insulin glargine population Eligible sample (n = 1,389) (n = 824) 65.6 – 10.7 652 (46.9) 89.3 – 17.8 30.8 – 5.5 7.1 – 5.5 8.7 – 1.3 (72 – 10.6) 181.7 – 56.6 (10.1 – 3.1) 363 84 104 125 85 359 182 185 65 428
(26.1) (6.1) (7.5) (9.0) (6.1) (25.9) (13.1) (13.3) (4.7) (30.8)
64.8 – 10.5 395 (47.9) 89.3 – 17.2 30.8 – 5.1 7.1 – 5.5 8.7 – 1.2 (72 – 9.9) 182.9 – 56.1 (10.2 – 3.1) 191 37 56 63 40 188 91 88 34 235
(23.2) (4.5) (6.8) (7.6) (4.9) (22.8) (11.0) (10.7) (4.1) (28.5)
Data are n (%) or mean – SD values. HbA1c, glycosylated hemoglobin; TIA, transitory ischemic attack.
Eligible patients had to have received at least 3 months of the maximum tolerated metformin dose, to have an HbA1c level of ‡ 7.5% (58 mmol/mol), and to be insulinnaive. They were enrolled after the attending physician had made the decision to treat the patient with either insulin glargine or (in a small subgroup) neutral protamine Hagedorn insulin.
For the identification of variables being associated with HbA1c treatment to target ( < 7.0% [ < 53 mmol/mol] and < 6.5% [48 mmol/mol]), baseline variables with a significant correlation were selected based on a univariate logistic regression analysis (Table 5). Variables with an a value of < 0.05 were eliminated and subjected to stepwise multivariate logistic regression analysis. These are expressed as odds ratios (OR) with 95% confidence intervals (CI).
Treatment
Results
Patients were titrated to achieve maximum fasting blood glucose of 100 mg/dL (5.6 mmol/L) at follow-up visits scheduled for Weeks 12 and 24. Titration was at the discretion of the attending physician, however. Blood glucose values £ 70 mg/dL ( £ 3.9 mmol/L) with symptoms were classified as mild hypoglycemia, whereas values £ 56 mg/dL ( £ 3.1 mmol/ L) or those requiring help were classified as being severe.
In total, 1,438 patients were enrolled into EARLY, of whom 1,389 (96.6%) received insulin glargine (Table 1). For the purpose of the multivariate predictor analysis the sample was further confined to those receiving insulin glargine, with available data on relevant baseline variables and the outcomes to be considered. This resulted in a total population of 824 patients available for the present analysis.
Statistical analysis
Patients
Only patients with a complete dataset for the present analysis were considered (available cases). Descriptive patients’ characteristics and changes in glycemic control/insulin dose used were displayed as percentages or mean and SD values (Tables 1 and 2). P values for Table 2 were calculated using SPSS version 20.0 software (IBM Corp., Armonk, NY) using a paired-samples t test. Baseline variables with a significant correlation to the target variable HbA1c reduction (Table 3) and insulin dose (Table 4) were selected using univariate linear regression analysis. Variables with an a value of
ORIGINAL ARTICLE
Predictors of Response to Early Basal Insulin Treatment in Patients with Type 2 Diabetes—The EARLY Experience Markolf Hanefeld, MD, PhD,1 Holger Fleischmann, PhD,2 Guido Schiffhorst, MS,3 and Peter Bramlage, MD 4
Abstract Background: It was the aim of this study to assess baseline predictors for glycosylated hemoglobin (HbA1c) reduction, treatment-to-target, and insulin glargine dose in patients with an HbA1c level of ‡ 7.5% (58 mmol/mol) at baseline despite 3 months of maximum tolerated dose of metformin under daily conditions. Subjects and Methods: This was an open, multicenter, prospective observational study with a 6-month follow-up including 1,438 patients with type 2 diabetes. Baseline variables independently associated with HbA1c (overall reduction and achievement of target values) and insulin glargine dose used were determined using a stepwise multivariate linear regression analysis. Results: In a multivariate linear regression analysis (R2 = 0.545) baseline HbA1c (b = - 0.722; P < 0.001) and retinopathy (b = - 0.064; P = 0.007) were associated with a greater HbA1c reduction at 6 months, whereas duration of diabetes was associated with a lesser HbA1c reduction (b = 0.084; P < 0.001). In another multivariate linear regression analysis, weight (odds ratio [OR] 0.99; 95% confidence interval [CI] 0.98 to < 1.00), duration of diabetes (OR 0.96; 95% CI 0.93–0.99), and baseline HbA1c (OR 0.65; 95% CI 0.56–0.76) were associated with a reduced likelihood of achieving an HbA1c level of < 7% (53 mmol/mol); baseline HbA1c (OR 0.66; 95% CI 0.51–0.85) was the only variable associated with a reduced likelihood of achieving an HbA1c level of < 6.5% (48 mmol/mol). In a further analysis (R2 = 0.135) the insulin dose needed was increased in those with a higher body weight (b = 0.230; P < 0.001), a longer duration of diabetes (b = 0.134; P < 0.001), a higher baseline HbA1c level (b = 0.205; P < 0.001), and the presence of microalbuminuria (b = 0.096; P = 0.003). Conclusions: Identified predictors of greater HbA1c reduction, target goal achievement, and insulin dose needed may help to optimize the balance of benefits and risks with the use of insulin glargine. Background
I
nsulin use early in the course of diabetes has been shown to be an effective first-line treatment option.1–4 International and national guidelines recommend metformin as a first step in the treatment of type 2 diabetes.5–8 In the recent European Association for the Study of Diabetes/American Diabetes Association consensus statement, five second-line treatment options are considered on top of metformin.7 Under real world conditions the use of basal insulin is usually confined to those with a glycosylated hemoglobin (HbA1c) level of ‡ 7.5% or even 8.0% (‡ 58 or 64 mmol/mol), and dual or even triple combinations of other treatment options are preferred instead with insulin as the last resort.
Early insulin use, however, may protect b-cells from harmful effects of long-lasting hyperglycemia and lipotoxicity. In patients inadequately controlled by metformin a recovery of residual b-cell function was recently demonstrated by Pennartz et al.9 Chronic long-acting basal insulin improved glucose-induced insulin secretion and postprandial endogenous insulin secretion in hyperglycemic patients with type 2 diabetes.1 Fonsecca et al.,5 in a pooled analysis of randomized trials, found that insulin glargine, given on top of metformin monotherapy early in the treatment, was associated with improved efficacy and safety over a regimen including sulfonylurea. Insulin glargine treatment introduced after dual oral combination therapy with sulfonylurea was associated with less HbA1c reduction but more hypoglycemia. In the
1 Study Centre Professor Hanefeld, Company for the Transfer of Knowledge and Technology of the Dresden University of Technology, Dresden, Germany. 2 Sanofi Aventis Deutschland GmbH, Berlin, Germany. 3 IGES Institut GmbH, Berlin, Germany. 4 Institute for Pharmacology and Preventive Medicine, Mahlow, Germany.
241
242 ORIGIN trial, the early addition of insulin glargine to standard of care was more effective as the comparator standard of care treatment to improve glycemia control with no adverse effects.4 The recent EARLY study documented the use of basal insulin glargine in patients being uncontrolled on a maximal tolerated dose of metformin under real world conditions.6 The HbA1c level was reduced from 8.7% (72 mmol/mol) to 7.4% (57 mmol/mol) within 24 weeks, as was fasting plasma glucose (from 182 to 131 mg/dL [10.1 to 7.3 mmol/l]; P < 0.01). Patients had a low risk of hypoglycemia and experienced a slight decrease of weight of about 1 kg (P < 0.01). The objective of the present exploratory analysis of the EARLY cohort was to assess (1) baseline predictors for insulin glargine dose used and (2) predictors of HbA1c reduction and HbA1c on target in patients having an HbA1c level of ‡ 7.5% (58 mmol/mol) at baseline. Subjects and Methods EARLY was an open, multicenter, prospective observational study with a 6-month follow-up that was planned and conducted according to the requirements of the German Bundesinstitut fu¨r Arzneimittel und Medizinprodukte and the Kassena¨rztliche Bundesvereinigung. Ethical approval was obtained from the Saxonian State Chamber of Medicine (Dresden, Germany). Details of the study protocol have been published previously.6 Patients
HANEFELD ET AL. Table 1. Patients’ Characteristics at Baseline
Variable Age (years) Female gender Body weight (kg) Body mass index (kg/m2) Duration of diabetes (years) HbA1c [% (mmol/mol)] at baseline Fasting blood glucose [mg/dL (mmol/L)] Comorbidities at baseline Coronary heart disease Myocardial infarction Stroke or TIA Peripheral artery disease Diabetic foot syndrome Neuropathy Retinopathy Nephropathy Macroalbuminuria Microalbuminuria
EARLY insulin glargine population Eligible sample (n = 1,389) (n = 824) 65.6 – 10.7 652 (46.9) 89.3 – 17.8 30.8 – 5.5 7.1 – 5.5 8.7 – 1.3 (72 – 10.6) 181.7 – 56.6 (10.1 – 3.1) 363 84 104 125 85 359 182 185 65 428
(26.1) (6.1) (7.5) (9.0) (6.1) (25.9) (13.1) (13.3) (4.7) (30.8)
64.8 – 10.5 395 (47.9) 89.3 – 17.2 30.8 – 5.1 7.1 – 5.5 8.7 – 1.2 (72 – 9.9) 182.9 – 56.1 (10.2 – 3.1) 191 37 56 63 40 188 91 88 34 235
(23.2) (4.5) (6.8) (7.6) (4.9) (22.8) (11.0) (10.7) (4.1) (28.5)
Data are n (%) or mean – SD values. HbA1c, glycosylated hemoglobin; TIA, transitory ischemic attack.
Eligible patients had to have received at least 3 months of the maximum tolerated metformin dose, to have an HbA1c level of ‡ 7.5% (58 mmol/mol), and to be insulinnaive. They were enrolled after the attending physician had made the decision to treat the patient with either insulin glargine or (in a small subgroup) neutral protamine Hagedorn insulin.
For the identification of variables being associated with HbA1c treatment to target ( < 7.0% [ < 53 mmol/mol] and < 6.5% [48 mmol/mol]), baseline variables with a significant correlation were selected based on a univariate logistic regression analysis (Table 5). Variables with an a value of < 0.05 were eliminated and subjected to stepwise multivariate logistic regression analysis. These are expressed as odds ratios (OR) with 95% confidence intervals (CI).
Treatment
Results
Patients were titrated to achieve maximum fasting blood glucose of 100 mg/dL (5.6 mmol/L) at follow-up visits scheduled for Weeks 12 and 24. Titration was at the discretion of the attending physician, however. Blood glucose values £ 70 mg/dL ( £ 3.9 mmol/L) with symptoms were classified as mild hypoglycemia, whereas values £ 56 mg/dL ( £ 3.1 mmol/ L) or those requiring help were classified as being severe.
In total, 1,438 patients were enrolled into EARLY, of whom 1,389 (96.6%) received insulin glargine (Table 1). For the purpose of the multivariate predictor analysis the sample was further confined to those receiving insulin glargine, with available data on relevant baseline variables and the outcomes to be considered. This resulted in a total population of 824 patients available for the present analysis.
Statistical analysis
Patients
Only patients with a complete dataset for the present analysis were considered (available cases). Descriptive patients’ characteristics and changes in glycemic control/insulin dose used were displayed as percentages or mean and SD values (Tables 1 and 2). P values for Table 2 were calculated using SPSS version 20.0 software (IBM Corp., Armonk, NY) using a paired-samples t test. Baseline variables with a significant correlation to the target variable HbA1c reduction (Table 3) and insulin dose (Table 4) were selected using univariate linear regression analysis. Variables with an a value of
