original article
original article
Diabetes, Obesity and Metabolism 16: 426–432, 2014. © 2013 John Wiley & Sons Ltd
Intensification of medication and glycaemic control among patients with type 2 diabetes – the ADVANCE trial S. van Dieren1,2 , A. P. Kengne1,2,3 , J. Chalmers1 , J. W. J. Beulens2 , T. M. E. Davis4 , G. Fulcher5 , S. R. Heller6 , A. Patel1 , S. Colagiuri7 , P. Hamet8 , G. Mancia9 , M. Marre10 , B. Neal1 , B. Williams11 , L. M. Peelen2 , Y. T. van der Schouw2 , M. Woodward1 & S. Zoungas1,12 1 The George Institute for Global Health, University of Sydney, Sydney, Australia 2 Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands 3 Department of medicine, South African Medical Research Council & University of Cape Town, Cape Town, South Africa 4 School of Medicine and Pharmacology, Fremantle Hospital, University of Western Australia, Fremantle, Australia 5 Department of Diabetes, Endocrinology and Metabolism, Royal North Shore Hospital, Sydney, Australia 6 Academic Unit of Diabetes, Endocrinology and Metabolism, School of Medicine and Biomedical Sciences, Sheffield, UK 7 Boden Institute for Obesity, Nutrition and Exercise, Sydney University, Sydney, Australia 8 Research Centre, Centre hospitalier del’Universite´ de Montreal, ´ Montreal, Canada 9 IRCCS Istituto Auxologico Italiano, Milan, Italy 10 Department of Endocrinology, Diabetology and Nutrition, Hopital ˆ Bichat-Claude Bernard and Universite´ Paris 7, Paris, France 11 School of Medicine, University of Leicester, Leicester, UK 12 School of Public Health and Preventive Medicine, Monash University, Melbourne, Australia
Aims: The aim of this study was to assess associations between patient characteristics, intensification of blood glucose-lowering treatment through oral glucose-lowering therapy and/or insulin and effective glycaemic control in type 2 diabetes.
Methods: 11 140 patients from the Action in Diabetes and Vascular disease: preterAx and diamicroN-MR Controlled Evaluation (ADVANCE) trial who were randomized to intensive glucose control or standard glucose control and followed up for a median of 5 years were categorized into two groups: effective glycaemic control [haemoglobin A1c (HbA1c) ≤ 7.0% or a proportionate reduction in HbA1c over 10%] or ineffective glycaemic control (HbA1c > 7.0% and a proportionate reduction in HbA1c less than or equal to 10%). Therapeutic intensification was defined as addition of an oral glucose-lowering agent or commencement of insulin. Pooled logistic regression models examined the associations between patient factors, intensification and effective glycaemic control. Results: A total of 7768 patients (69.7%), including 3198 in the standard treatment group achieved effective glycaemic control. Compared to patients with ineffective control, patients with effective glycaemic control had shorter duration of diabetes and lower HbA1c at baseline and at the time of treatment intensification. Treatment intensification with addition of an oral agent or commencement of insulin was associated with a 107% [odds ratio, OR: 2.07 (95% confidence interval, CI: 1.95–2.20)] and 152% [OR: 2.52 (95% CI: 2.30–2.77)] greater chance of achieving effective glycaemic control, respectively. These associations were robust after adjustment for several baseline characteristics and not modified by the number of oral medications taken at the time of treatment intensification. Conclusions: Effective glycaemic control was associated with treatment intensification at lower HbA1c levels at all stages of the disease course and in both arms of the ADVANCE trial. Keywords: glycaemic control, HbA1c, therapy, treatment intensification, type 2 diabetes Date submitted 29 May 2013; date of first decision 2 July 2013; date of final acceptance 2 November 2013
Introduction Although achieving and maintaining normal or near normal blood glucose levels remains a key strategy in the treatment of patients with type 2 diabetes it can be difficult to implement [1–3]. Indeed most recent guidelines have recommended individualization of glycaemic targets based on patient characteristics [4,5]. In large trials of blood glucose-lowering Correspondence to: Sophia Zoungas MD, PhD, The George Institute for Global Health, University of Sydney, PO Box M201, Missenden Road, NSW 2050 Sydney, Australia. E-mail: [email protected]
monotherapy in new or recently diagnosed patients there was strong evidence of progressive β cell failure and glycaemic escape [6,7]. In the UK Prospective Diabetes Study (UKPDS), there was a gradual rise in HbA1c over time in both conventional and intensive treatment arms, and a need to intensify pharmacotherapy to meet protocol-driven glycaemic targets as a result [6]. In the A Diabetes Outcome Progression Trial (ADOPT) there was a progressive rate of increase in HbA1c that was dependent on the allocated monotherapy [7]. More recent studies involving patients with longer duration diabetes have employed more pragmatic poly-pharmaceutical approaches to achieving pre-specified glycaemic targets for, and
original article
DIABETES, OBESITY AND METABOLISM
separation between, intensive versus conventional treatment arms. In Action in Diabetes and Vascular disease: preterAx and diamicroN-MR Controlled Evaluation (ADVANCE) trial, for example, the mean HbA1c values at the end of the study were 6.5% in the intensive group and 7.3% in the standard group, levels that were maintained for the 5 years of follow-up [2]. This intensified treatment resulted in a significant reduction in the combined primary endpoint of major macrovascular and microvascular events, and separately for microvascular events but not macrovascular events [2]. Stable mean HbA1c levels during intensive and conventional management were also a feature of Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial and Veterans Affairs Diabetes Trial (VADT) [8,9]. The reasons for this maintained therapeutic efficacy remain unknown. The ADVANCE trial data provide an opportunity to determine which patient-, drug- and treatment algorithmspecific factors were responsible for the stable glycaemia that was not seen in the monotherapy studies. Such data could be used to tailor glycaemic management to the clinical situation in routine care and provide further insight into mechanisms underlying the durability of different therapeutic approaches in type 2 diabetes.
Methods Study Population and Design ADVANCE was a 2 × 2 factorial randomized controlled trial evaluating the effects of lowering blood pressure and intensive blood glucose control on vascular outcomes in participants with type 2 diabetes. Detailed descriptions of the design have been published previously [2,10]. In brief, 11 140 participants with type 2 diabetes, aged 55 years or older, with a history of major macrovascular or microvascular disease, or at least one other risk factor for vascular disease, were recruited from 215 centres in 20 countries. Approval for the trial was obtained from the institutional ethics committee of each centre and all participants signed an informed consent form. All potentially eligible participants entered a 6-week active run-in period during which they continued their usual methods of glucose control and received a fixed combination of perindopril–indapamide (2 mg/0.625 mg). Participants who tolerated and were compliant with the run-in treatment were subsequently randomized to continued treatment with perindopril–indapamide or matching placebo and randomly assigned to intensive glucose control strategy aiming for a HbA1c ≤ 6.5% or standard glucose-lowering therapy. Participants randomized to the intensive glucose control regimen commenced gliclazide modified release (gliclazide MR; 30–120 mg daily), and were required to discontinue any other sulphonylurea. The timing, selection and dose of all other glucose-lowering therapies were at the discretion of the treating physician, however, a conventional approach of incremental progression from oral monotherapy to combination therapy to basal insulin was advocated [10]. Participants randomized to the standard glucose control regimen were managed according to the relevant national or regional glucose-lowering guidelines. Participants were seen at two pre-randomization visits, at 3,
Volume 16 No. 5 May 2014
4 and 6 months after randomization and subsequently every 6 months. The median follow-up time for the glucose control arm of the study was 5 years [2]. HbA1c was measured at baseline and every 6 months for the intensive group and at baseline, 6 months and every 12 months in the standard group. Clinical and laboratory examinations were performed at each study centre. Weight, height and systolic blood pressure were recorded at baseline.
Definition of Effective and Ineffective Glycaemic Control Definition of effective and ineffective glycaemic control are based upon the target level used in the standard group (HbA1c < 7.0%). Therefore, effective glycaemic control was defined as a decline in absolute HbA1c level to ≤7.0% or a proportionate reduction in HbA1c of over 10% (e.g. if HbA1c was 10% an absolute reduction of 1%). Participants with a starting HbA1c > 7.0% who did not achieve these reductions or those with a starting HbA1c ≤ 7% who were unable to maintain this HbA1c level, were considered to have ineffective glycaemic control. To compare the baseline characteristics and depict the HbA1c trajectory over time, participants were categorized according to the reduction in HbA1c from baseline to the end of follow-up. To examine the effects of intensification by commencement of insulin or additional oral glucose-lowering therapy during follow-up, HbA1c reductions were calculated between consecutive annual visits until the end of follow-up. The HbA1c measurements used for these purposes were limited to those taken at visits that were common to both the intensive and the standard glucose control arms.
Statistical Analyses All participants with at least one follow-up measurement were included in these analyses (n = 10 780). Baseline characteristics were compared separately by randomized glucose control strategy in participants with effective and ineffective glycaemic control, between baseline and the end of follow-up. Interactions between randomized treatment, glycaemic control and baseline variables were examined by including the interaction term in generalized linear regression models. Pooled logistic regression models were used to assess the associations between effective glycaemic control and treatment intensification by addition of an oral glucoselowering medication or commencement of insulin during each year of follow-up. Participants contributed a new observation at the start of each consecutive year of follow-up so that any treatment intensification in the previous year of follow-up was captured. Intensification by oral glucose-lowering therapy was defined by the commencement of a new oral glucoselowering medication between two consecutive annual visits. Intensification by insulin was defined as commencement of insulin between two visits. As such a given participant could contribute more than one treatment intensification during follow-up. Analyses were first adjusted for age and sex and then further adjusted for age at diabetes diagnosis, diabetes
doi:10.1111/dom.12238 427
original article duration, ethnicity, weight, systolic blood pressure, history of macrovascular disease, history of microvascular disease, randomized treatment arms, the number of oral medications at intensification and HbA1c concentration at the start of each year of follow-up. The effects of commencement of an added oral glucose treatment or insulin on glycaemic control was stratified by the number of oral medication at time of intensification to examine the effects of intensification when patients were already on mono-, dual- or triple-oral therapy. In sensitivity analyses the same associations were assessed (i) defining effective glycaemic control using a cut-off of 6.5% and using different cut-offs for the two treatment arms (6.5% for the intensive arm and 7.0% for the standard arm) and (ii) separately analysing effective glycaemic control on the basis of an HbA1c < 7.0 at end of follow-up or a proportionate reduction in HbA1c over 10% alone. sas version 9.2 for Windows (SAS Institute, Cary, NC, USA) was used for all statistical analyses.
Results All patients with an HbA1c measurement at baseline and at least one follow-up (n = 10 780) were included in the analyses. In the standard glucose control arm 3198 participants (59.5%) were classified as achieving effective glycaemic control between baseline and the end of follow-up while in the intensive arm 4570 participants (84.5%) were classified in this manner. In the standard arm from baseline to 60 months of follow-up HbA1c decreased by a mean of 0.88% in the effective group and increased by a mean of 0.78% in the ineffective group (Figure 1a). In the intensive arm from baseline to 60 months of follow-up HbA1c decreased by a mean of 1.20% in the effective group and increased by a mean of 0.52% in the ineffective group (all p values 10% (2.14% in standard arm and 2.49% in intensive arm; Figure S1). The total number (minimum–maximum) of intensifications in the effective and ineffective groups was 8196 (0–5) and 3012 (0–4), respectively.
Patient Characteristics Associated With Effective Glycaemic Control In both the standard and intensive arm of the study, duration of diabetes was shorter and baseline HbA1c levels higher in the effective glycaemic control group compared with the ineffective glycaemic control group (Table 1). In addition, in the intensive arm of the study only, weight and body mass index (BMI) were lower and non-Caucasian ethnicity more frequent in the effective glycaemic control group compared with the ineffective glycaemic control group. A significant interaction by treatment arm was only observed for ethnicity (p < 0.001). In multivariable analyses, duration of diabetes and HbA1c at the time of intensification were both independently associated with effective glycaemic control. For every additional 5 years of diagnosed diabetes and every 1% higher HbA1c level at the time of intensification, the relative likelihood of achieving effective glycaemic control was reduced by 3% [95% confidence interval (CI): 1–5%] and 81% (95% CI: 80–81%) respectively (Table S1). Of note, prevalent microvascular and macrovascular diseases and randomized treatment assignment to perindopril–indapamide or placebo were not independently associated with effectiveness of glycaemic control.
Time to Treatment Intensification and HbA1c at Intensification Time to oral treatment intensification in the intensive arm of the study was 14.4 months for the effective glycaemic control group and 17.6 months for the ineffective glycaemic control group, mean difference 3.2 months (95% CI 1.7–4.7). Equivalents in the standard arm were 14.8 versus 15.7 months, mean difference 0.9 months (95% CI −0.2–1.9, Table 2). Time to
Figure 1. HbA1c during follow-up stratified by ineffective (red) and effective (green) glycaemic control for (a) standard arm and (b) intensive arm.
428 van Dieren et al.
Volume 16 No. 5 May 2014
original article
DIABETES, OBESITY AND METABOLISM
Table 1. Baseline characteristics by effectiveness of glycaemic control treatment of all patients included in the analysis (n = 10 780). Effective glycaemic control is defined as having an HbA1c below 7.0% at end of follow-up or an HbA1c reduction of more than 10%. Standard arm
Age (years), mean (s.d.) Sex (female), n (%) Weight (kg), mean (s.d.) BMI (kg/m2 ), mean (s.d.) Diabetes duration (years), mean (s.d.) Systolic blood pressure (mmHg), mean (s.d.) Glucose (mmol/l), mean (s.d.) Baseline HbA1c (%), mean (s.d.) Ethnicity (Caucasian), n (%) History macrovascular disease, n (%) History microvascular disease, n (%)
Intensive arm
Effective (n = 3198)
Ineffective (n = 2174)
65.9 (6.34) 1380 (43.2) 77.5 (16.4) 28.2 (5.25) 7.65 (6.26) 145 (21.4) 8.48 (2.95) 7.56 (1.77) 1889 (59.1) 1014 (31.7) 318 (9.9)
65.6 (6.40) 892 (41.0) 78.3 (16.5) 28.4 (5.08) 8.37 (6.49) 145 (21.2) 8.45 (2.46) 7.44 (1.14) 1301 (59.8) 705 (32.4) 236 (10.9)
p Value
Effective (n = 4570)
Ineffective (n = 838)
p Value
p Value interaction
0.205 0.122 0.073 0.081
original article
Diabetes, Obesity and Metabolism 16: 426–432, 2014. © 2013 John Wiley & Sons Ltd
Intensification of medication and glycaemic control among patients with type 2 diabetes – the ADVANCE trial S. van Dieren1,2 , A. P. Kengne1,2,3 , J. Chalmers1 , J. W. J. Beulens2 , T. M. E. Davis4 , G. Fulcher5 , S. R. Heller6 , A. Patel1 , S. Colagiuri7 , P. Hamet8 , G. Mancia9 , M. Marre10 , B. Neal1 , B. Williams11 , L. M. Peelen2 , Y. T. van der Schouw2 , M. Woodward1 & S. Zoungas1,12 1 The George Institute for Global Health, University of Sydney, Sydney, Australia 2 Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands 3 Department of medicine, South African Medical Research Council & University of Cape Town, Cape Town, South Africa 4 School of Medicine and Pharmacology, Fremantle Hospital, University of Western Australia, Fremantle, Australia 5 Department of Diabetes, Endocrinology and Metabolism, Royal North Shore Hospital, Sydney, Australia 6 Academic Unit of Diabetes, Endocrinology and Metabolism, School of Medicine and Biomedical Sciences, Sheffield, UK 7 Boden Institute for Obesity, Nutrition and Exercise, Sydney University, Sydney, Australia 8 Research Centre, Centre hospitalier del’Universite´ de Montreal, ´ Montreal, Canada 9 IRCCS Istituto Auxologico Italiano, Milan, Italy 10 Department of Endocrinology, Diabetology and Nutrition, Hopital ˆ Bichat-Claude Bernard and Universite´ Paris 7, Paris, France 11 School of Medicine, University of Leicester, Leicester, UK 12 School of Public Health and Preventive Medicine, Monash University, Melbourne, Australia
Aims: The aim of this study was to assess associations between patient characteristics, intensification of blood glucose-lowering treatment through oral glucose-lowering therapy and/or insulin and effective glycaemic control in type 2 diabetes.
Methods: 11 140 patients from the Action in Diabetes and Vascular disease: preterAx and diamicroN-MR Controlled Evaluation (ADVANCE) trial who were randomized to intensive glucose control or standard glucose control and followed up for a median of 5 years were categorized into two groups: effective glycaemic control [haemoglobin A1c (HbA1c) ≤ 7.0% or a proportionate reduction in HbA1c over 10%] or ineffective glycaemic control (HbA1c > 7.0% and a proportionate reduction in HbA1c less than or equal to 10%). Therapeutic intensification was defined as addition of an oral glucose-lowering agent or commencement of insulin. Pooled logistic regression models examined the associations between patient factors, intensification and effective glycaemic control. Results: A total of 7768 patients (69.7%), including 3198 in the standard treatment group achieved effective glycaemic control. Compared to patients with ineffective control, patients with effective glycaemic control had shorter duration of diabetes and lower HbA1c at baseline and at the time of treatment intensification. Treatment intensification with addition of an oral agent or commencement of insulin was associated with a 107% [odds ratio, OR: 2.07 (95% confidence interval, CI: 1.95–2.20)] and 152% [OR: 2.52 (95% CI: 2.30–2.77)] greater chance of achieving effective glycaemic control, respectively. These associations were robust after adjustment for several baseline characteristics and not modified by the number of oral medications taken at the time of treatment intensification. Conclusions: Effective glycaemic control was associated with treatment intensification at lower HbA1c levels at all stages of the disease course and in both arms of the ADVANCE trial. Keywords: glycaemic control, HbA1c, therapy, treatment intensification, type 2 diabetes Date submitted 29 May 2013; date of first decision 2 July 2013; date of final acceptance 2 November 2013
Introduction Although achieving and maintaining normal or near normal blood glucose levels remains a key strategy in the treatment of patients with type 2 diabetes it can be difficult to implement [1–3]. Indeed most recent guidelines have recommended individualization of glycaemic targets based on patient characteristics [4,5]. In large trials of blood glucose-lowering Correspondence to: Sophia Zoungas MD, PhD, The George Institute for Global Health, University of Sydney, PO Box M201, Missenden Road, NSW 2050 Sydney, Australia. E-mail: [email protected]
monotherapy in new or recently diagnosed patients there was strong evidence of progressive β cell failure and glycaemic escape [6,7]. In the UK Prospective Diabetes Study (UKPDS), there was a gradual rise in HbA1c over time in both conventional and intensive treatment arms, and a need to intensify pharmacotherapy to meet protocol-driven glycaemic targets as a result [6]. In the A Diabetes Outcome Progression Trial (ADOPT) there was a progressive rate of increase in HbA1c that was dependent on the allocated monotherapy [7]. More recent studies involving patients with longer duration diabetes have employed more pragmatic poly-pharmaceutical approaches to achieving pre-specified glycaemic targets for, and
original article
DIABETES, OBESITY AND METABOLISM
separation between, intensive versus conventional treatment arms. In Action in Diabetes and Vascular disease: preterAx and diamicroN-MR Controlled Evaluation (ADVANCE) trial, for example, the mean HbA1c values at the end of the study were 6.5% in the intensive group and 7.3% in the standard group, levels that were maintained for the 5 years of follow-up [2]. This intensified treatment resulted in a significant reduction in the combined primary endpoint of major macrovascular and microvascular events, and separately for microvascular events but not macrovascular events [2]. Stable mean HbA1c levels during intensive and conventional management were also a feature of Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial and Veterans Affairs Diabetes Trial (VADT) [8,9]. The reasons for this maintained therapeutic efficacy remain unknown. The ADVANCE trial data provide an opportunity to determine which patient-, drug- and treatment algorithmspecific factors were responsible for the stable glycaemia that was not seen in the monotherapy studies. Such data could be used to tailor glycaemic management to the clinical situation in routine care and provide further insight into mechanisms underlying the durability of different therapeutic approaches in type 2 diabetes.
Methods Study Population and Design ADVANCE was a 2 × 2 factorial randomized controlled trial evaluating the effects of lowering blood pressure and intensive blood glucose control on vascular outcomes in participants with type 2 diabetes. Detailed descriptions of the design have been published previously [2,10]. In brief, 11 140 participants with type 2 diabetes, aged 55 years or older, with a history of major macrovascular or microvascular disease, or at least one other risk factor for vascular disease, were recruited from 215 centres in 20 countries. Approval for the trial was obtained from the institutional ethics committee of each centre and all participants signed an informed consent form. All potentially eligible participants entered a 6-week active run-in period during which they continued their usual methods of glucose control and received a fixed combination of perindopril–indapamide (2 mg/0.625 mg). Participants who tolerated and were compliant with the run-in treatment were subsequently randomized to continued treatment with perindopril–indapamide or matching placebo and randomly assigned to intensive glucose control strategy aiming for a HbA1c ≤ 6.5% or standard glucose-lowering therapy. Participants randomized to the intensive glucose control regimen commenced gliclazide modified release (gliclazide MR; 30–120 mg daily), and were required to discontinue any other sulphonylurea. The timing, selection and dose of all other glucose-lowering therapies were at the discretion of the treating physician, however, a conventional approach of incremental progression from oral monotherapy to combination therapy to basal insulin was advocated [10]. Participants randomized to the standard glucose control regimen were managed according to the relevant national or regional glucose-lowering guidelines. Participants were seen at two pre-randomization visits, at 3,
Volume 16 No. 5 May 2014
4 and 6 months after randomization and subsequently every 6 months. The median follow-up time for the glucose control arm of the study was 5 years [2]. HbA1c was measured at baseline and every 6 months for the intensive group and at baseline, 6 months and every 12 months in the standard group. Clinical and laboratory examinations were performed at each study centre. Weight, height and systolic blood pressure were recorded at baseline.
Definition of Effective and Ineffective Glycaemic Control Definition of effective and ineffective glycaemic control are based upon the target level used in the standard group (HbA1c < 7.0%). Therefore, effective glycaemic control was defined as a decline in absolute HbA1c level to ≤7.0% or a proportionate reduction in HbA1c of over 10% (e.g. if HbA1c was 10% an absolute reduction of 1%). Participants with a starting HbA1c > 7.0% who did not achieve these reductions or those with a starting HbA1c ≤ 7% who were unable to maintain this HbA1c level, were considered to have ineffective glycaemic control. To compare the baseline characteristics and depict the HbA1c trajectory over time, participants were categorized according to the reduction in HbA1c from baseline to the end of follow-up. To examine the effects of intensification by commencement of insulin or additional oral glucose-lowering therapy during follow-up, HbA1c reductions were calculated between consecutive annual visits until the end of follow-up. The HbA1c measurements used for these purposes were limited to those taken at visits that were common to both the intensive and the standard glucose control arms.
Statistical Analyses All participants with at least one follow-up measurement were included in these analyses (n = 10 780). Baseline characteristics were compared separately by randomized glucose control strategy in participants with effective and ineffective glycaemic control, between baseline and the end of follow-up. Interactions between randomized treatment, glycaemic control and baseline variables were examined by including the interaction term in generalized linear regression models. Pooled logistic regression models were used to assess the associations between effective glycaemic control and treatment intensification by addition of an oral glucoselowering medication or commencement of insulin during each year of follow-up. Participants contributed a new observation at the start of each consecutive year of follow-up so that any treatment intensification in the previous year of follow-up was captured. Intensification by oral glucose-lowering therapy was defined by the commencement of a new oral glucoselowering medication between two consecutive annual visits. Intensification by insulin was defined as commencement of insulin between two visits. As such a given participant could contribute more than one treatment intensification during follow-up. Analyses were first adjusted for age and sex and then further adjusted for age at diabetes diagnosis, diabetes
doi:10.1111/dom.12238 427
original article duration, ethnicity, weight, systolic blood pressure, history of macrovascular disease, history of microvascular disease, randomized treatment arms, the number of oral medications at intensification and HbA1c concentration at the start of each year of follow-up. The effects of commencement of an added oral glucose treatment or insulin on glycaemic control was stratified by the number of oral medication at time of intensification to examine the effects of intensification when patients were already on mono-, dual- or triple-oral therapy. In sensitivity analyses the same associations were assessed (i) defining effective glycaemic control using a cut-off of 6.5% and using different cut-offs for the two treatment arms (6.5% for the intensive arm and 7.0% for the standard arm) and (ii) separately analysing effective glycaemic control on the basis of an HbA1c < 7.0 at end of follow-up or a proportionate reduction in HbA1c over 10% alone. sas version 9.2 for Windows (SAS Institute, Cary, NC, USA) was used for all statistical analyses.
Results All patients with an HbA1c measurement at baseline and at least one follow-up (n = 10 780) were included in the analyses. In the standard glucose control arm 3198 participants (59.5%) were classified as achieving effective glycaemic control between baseline and the end of follow-up while in the intensive arm 4570 participants (84.5%) were classified in this manner. In the standard arm from baseline to 60 months of follow-up HbA1c decreased by a mean of 0.88% in the effective group and increased by a mean of 0.78% in the ineffective group (Figure 1a). In the intensive arm from baseline to 60 months of follow-up HbA1c decreased by a mean of 1.20% in the effective group and increased by a mean of 0.52% in the ineffective group (all p values 10% (2.14% in standard arm and 2.49% in intensive arm; Figure S1). The total number (minimum–maximum) of intensifications in the effective and ineffective groups was 8196 (0–5) and 3012 (0–4), respectively.
Patient Characteristics Associated With Effective Glycaemic Control In both the standard and intensive arm of the study, duration of diabetes was shorter and baseline HbA1c levels higher in the effective glycaemic control group compared with the ineffective glycaemic control group (Table 1). In addition, in the intensive arm of the study only, weight and body mass index (BMI) were lower and non-Caucasian ethnicity more frequent in the effective glycaemic control group compared with the ineffective glycaemic control group. A significant interaction by treatment arm was only observed for ethnicity (p < 0.001). In multivariable analyses, duration of diabetes and HbA1c at the time of intensification were both independently associated with effective glycaemic control. For every additional 5 years of diagnosed diabetes and every 1% higher HbA1c level at the time of intensification, the relative likelihood of achieving effective glycaemic control was reduced by 3% [95% confidence interval (CI): 1–5%] and 81% (95% CI: 80–81%) respectively (Table S1). Of note, prevalent microvascular and macrovascular diseases and randomized treatment assignment to perindopril–indapamide or placebo were not independently associated with effectiveness of glycaemic control.
Time to Treatment Intensification and HbA1c at Intensification Time to oral treatment intensification in the intensive arm of the study was 14.4 months for the effective glycaemic control group and 17.6 months for the ineffective glycaemic control group, mean difference 3.2 months (95% CI 1.7–4.7). Equivalents in the standard arm were 14.8 versus 15.7 months, mean difference 0.9 months (95% CI −0.2–1.9, Table 2). Time to
Figure 1. HbA1c during follow-up stratified by ineffective (red) and effective (green) glycaemic control for (a) standard arm and (b) intensive arm.
428 van Dieren et al.
Volume 16 No. 5 May 2014
original article
DIABETES, OBESITY AND METABOLISM
Table 1. Baseline characteristics by effectiveness of glycaemic control treatment of all patients included in the analysis (n = 10 780). Effective glycaemic control is defined as having an HbA1c below 7.0% at end of follow-up or an HbA1c reduction of more than 10%. Standard arm
Age (years), mean (s.d.) Sex (female), n (%) Weight (kg), mean (s.d.) BMI (kg/m2 ), mean (s.d.) Diabetes duration (years), mean (s.d.) Systolic blood pressure (mmHg), mean (s.d.) Glucose (mmol/l), mean (s.d.) Baseline HbA1c (%), mean (s.d.) Ethnicity (Caucasian), n (%) History macrovascular disease, n (%) History microvascular disease, n (%)
Intensive arm
Effective (n = 3198)
Ineffective (n = 2174)
65.9 (6.34) 1380 (43.2) 77.5 (16.4) 28.2 (5.25) 7.65 (6.26) 145 (21.4) 8.48 (2.95) 7.56 (1.77) 1889 (59.1) 1014 (31.7) 318 (9.9)
65.6 (6.40) 892 (41.0) 78.3 (16.5) 28.4 (5.08) 8.37 (6.49) 145 (21.2) 8.45 (2.46) 7.44 (1.14) 1301 (59.8) 705 (32.4) 236 (10.9)
p Value
Effective (n = 4570)
Ineffective (n = 838)
p Value
p Value interaction
0.205 0.122 0.073 0.081
