Exploring the characteristics of suboptimally controlled patients after 24weeks of basal insulin treatment: An individualized approach to intensification

Open AccessPublished:December 09, 2016DOI:https://doi.org/10.1016/j.diabres.2016.11.028

      Highlights

      • Post hoc analysis of patients with type 2 diabetes who started insulin detemir.
      • HbA1c ⩾ 7.0% is associated with baseline parameters related to disease severity.
      • Supports intensive titration/treatment intensification in subjects with HbA1c ⩾ 7.0%.

      Abstract

      Aim

      To identify characteristics of suboptimally controlled patients with type 2 diabetes (T2DM) on basal insulin treatment who may benefit from intensive titration or further intensification of treatment.

      Methods

      A post hoc analysis of SOLVE: a 24-week, international, observational study conducted in 17,374 patients with T2DM inadequately controlled on oral antidiabetic drugs (OADs) started on once-daily insulin detemir. Patients were divided into two groups based on whether they achieved HbA1c < 7.0% (<53.0 mmol/mol) or not at final visit.

      Results

      Suboptimal glycemic control (HbA1c ⩾ 7.0 [⩾53.0 mmol/mol]) was independently associated with several baseline characteristics including higher baseline HbA1c (odds ratio [95% confidence interval]: 1.56 [1.50;1.62]; p < 0.0001) and body mass index (BMI) (1.03 [1.02;1.04]; p < 0.0001), longer duration of diabetes (5–10 years: 1.44 [1.25;1.66]; >10 years: 1.44 [1.17;1.77]; p < 0.0001), and greater number of OADs (two OADs: 1.27 [1.12;1.44]; >2 OADs: 1.38 [1.14;1.66]; p = 0.0003). Overall reporting of hypoglycemia was low; fewer patients with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) reported hypoglycemic events compared with patients with HbA1c < 7.0% (9.8% vs. 12.5%, respectively; p < 0.001).

      Conclusions

      Baseline characteristics related to severity of disease were strongly associated with suboptimal glycemic control in patients with T2DM receiving basal insulin. These factors may help clinicians in identifying patients who may require an individualized approach to titration or intensification of treatment.

      Trial registration

      NCT00740519.

      Keywords

      1. Introduction

      Treatment guidelines for people with type 2 diabetes (T2DM) recommend the use of a variety of antihyperglycemic drugs after the failure of diet and lifestyle changes to help optimize glycemic control [
      AACE/ACE comprehensive diabetes management algorithm.
      ,
      • American Diabetes Association
      Standard of Medical Care in Diabetes–2016.
      ,

      National Institute for Health and Care Excellence. Type 2 diabetes in adults: management. NG28. December 2015.

      ]. The sequence in which these drugs should be used is contentious, so treatment decisions should be made based on individual patient characteristics. Nevertheless, as T2DM progresses, prompt treatment intensification of these drugs, which includes the use of insulin, will be required. It is of concern that studies have identified a considerable potential delay (up to 7 years) when intensifying treatment [
      • Khunti K.
      • Wolden M.L.
      • Thorsted B.L.
      • et al.
      Clinical inertia in people with type 2 diabetes: a retrospective cohort study of more than 80,000 people.
      ], as this, in turn, will impact upon the achievement of treatment goals and put poorly controlled patients at a high risk of diabetes-related complications [
      • Gerstein H.C.
      • Miller M.E.
      • Genuth S.
      • et al.
      Long-term effects of intensive glucose lowering on cardiovascular outcomes.
      ].
      Although most patients with T2DM will eventually require treatment with insulin [
      • Holman R.R.
      • Farmer A.J.
      • Melanie M.J.
      • et al.
      Three-year efficacy of complex insulin regimens in type 2 diabetes.
      ], patients and physicians alike may be discouraged from starting this line of therapy due to concerns regarding the increased risk of hypoglycemia [
      • Peyrot M.
      • Rubin R.R.
      • Lauritzen T.
      • et al.
      Resistance to insulin therapy among patients and providers: results of the cross-national Diabetes Attitudes, Wishes, and Needs (DAWN) study.
      ,
      • Stargardt T.
      • Gonder-Frederick L.
      • Krobot K.J.
      • et al.
      Fear of hypoglycaemia: defining a minimum clinically important difference in patients with type 2 diabetes.
      ] and potential weight gain [
      • Peyrot M.
      • Rubin R.R.
      • Khunti K.
      Addressing barriers to initiation of insulin in patients with type 2 diabetes.
      ]. Studies, however, have shown that starting basal insulin therapy in patients suboptimally controlled with oral antidiabetic drugs (OADs) may improve glycemic control with a low risk for hypoglycemia and weight gain [
      • Blonde L.
      • Merilainen M.
      • Karwe V.
      • et al.
      Patient-directed titration for achieving glycaemic goals using a once-daily basal insulin analogue: an assessment of two different fasting plasma glucose targets – the TITRATETM study.
      ,
      • Khunti K.
      • Caputo S.
      • Damci T.
      • et al.
      The safety and efficacy of adding once-daily insulin detemir to oral hypoglycaemic agents in patients with type 2 diabetes in a clinical practice setting in 10 countries.
      ]. Notwithstanding the reported improvement in clinical outcomes, observational findings suggest that basal insulin dose in patients with T2DM could still be titrated further, thus indicating the potential for additional improvement in outcomes with dose escalation [
      • Khunti K.
      • Caputo S.
      • Damci T.
      • et al.
      The safety and efficacy of adding once-daily insulin detemir to oral hypoglycaemic agents in patients with type 2 diabetes in a clinical practice setting in 10 countries.
      ]. Once a patient is receiving basal insulin, there is a need to ensure that the treatment is monitored, titrated, and intensified in a timely and appropriate manner. Indeed, a recent UK study demonstrated that only 30.9% of patients who were receiving basal insulin and were eligible for treatment intensification (defined by HbA1c ⩾ 7.5% [⩾58.5 mmol/mol]) actually had their treatment intensified after a median 3.7-year delay; more concerning, 32.1% of patients stopped the use of basal insulin altogether [
      • Khunti K.
      • Nikolajsen A.
      • Thorsted B.L.
      • et al.
      Clinical inertia with regard to intensifying therapy in people with type 2 diabetes treated with basal insulin.
      ]. Readily identifying patients who would benefit from more intensive titration or intensification (with postprandial glucose-controlling drugs) of their basal insulin regimen to prevent treatment inertia and/or withdrawal is an important challenge facing clinicians.
      SOLVE was an international observational study conducted in 17,374 patients with T2DM inadequately controlled on OADs who were started on once-daily basal insulin with insulin detemir. After 24 weeks of treatment, 32.8% of patients achieved HbA1c < 7.0% (<53.0 mmol/mol) with a change in HbA1c from baseline (8.9%; 74 mmol/mol) of −1.3% (−14 mmol/mol; p < 0.001). Moreover, few adverse events, minor weight reduction, and low hypoglycemia rates were reported [
      • Khunti K.
      • Caputo S.
      • Damci T.
      • et al.
      The safety and efficacy of adding once-daily insulin detemir to oral hypoglycaemic agents in patients with type 2 diabetes in a clinical practice setting in 10 countries.
      ]. Based on SOLVE data, this post hoc analysis sets out to explore the characteristics of suboptimally controlled patients (defined as HbA1c ⩾ 7.0% at final visit) after basal insulin treatment, and therefore identify the characteristics of patients who are likely to benefit from intensive titration of basal insulin or further intensification.

      2. Materials and methods

      2.1 SOLVE design

      SOLVE methodology has been reported previously [
      • Khunti K.
      • Caputo S.
      • Damci T.
      • et al.
      The safety and efficacy of adding once-daily insulin detemir to oral hypoglycaemic agents in patients with type 2 diabetes in a clinical practice setting in 10 countries.
      ]. Briefly, SOLVE was a 24-week, observational, open-label study in patients with T2DM who were not achieving treatment targets with diet, exercise, and ⩾1 OAD. Once-daily insulin detemir was added to each patient’s existing therapy. The study was conducted at 2817 sites across 10 countries: Canada, China, Germany, Israel, Italy, Poland, Portugal, Spain, Turkey, and the UK. Baseline demographic data were collected for all patients who entered the study; this included any previous history of macro- or microvascular disease (defined elsewhere [
      • Khunti K.
      • Caputo S.
      • Damci T.
      • et al.
      The safety and efficacy of adding once-daily insulin detemir to oral hypoglycaemic agents in patients with type 2 diabetes in a clinical practice setting in 10 countries.
      ]). In addition to baseline, data were collected at 12 and 24 weeks of treatment.

      2.2 Post hoc analysis population and outcomes

      All patients enrolled in SOLVE who were prescribed insulin detemir and included in the full analysis set (FAS) were considered for the analysis. Patients had to have information on HbA1c at the 12-week interim visit and at the 24-week final visit. Suboptimal glycemic control was defined as an HbA1c level ⩾7.0% at final visit, as this is the threshold recommended for most patients by the American Diabetes Association and European Association for the Study of Diabetes [
      • Inzucchi S.E.
      • Bergenstal R.M.
      • Buse J.B.
      • et al.
      Management of hyperglycaemia in type 2 diabetes, 2015: a patient-centred approach. Update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes.
      ]. Patients were divided into two groups based on whether they achieved HbA1c < 7.0% (<53 mmol/mol) or not at final visit. Supplementary analyses were conducted using an HbA1c cut-off of 7.5% (58.5 mmol/mol) and outcomes of both HbA1c cut-offs were investigated after 12 weeks of treatment.
      The analysis explored the relationship between baseline patient characteristics and the attainment of HbA1c level ⩾7.0% at 24 weeks. Baseline patient characteristics included demographics and indicators of disease severity. Clinical study outcomes recorded both at baseline and at the 24-week visit included glycemic control (HbA1c and fasting blood glucose [FBG]), hypoglycemia, weight, and insulin dose. For the purpose of this analysis, the parameter ‘any hypoglycemia’ includes minor (daytime or nocturnal) and major episodes; these were combined, as there were only a few major episodes reported. An episode was considered minor if blood glucose was <3.1 mmol/l with or without symptoms (recalled within the last 4 weeks prior to visit). A major episode required third-party assistance with the administration of food, glucagon, or intravenous glucose (recalled within the last 12 weeks in all countries except the UK; the recall period in the UK was 4 weeks). Data on antihypertensive and/or lipid-lowering drugs were not collected in the study.

      2.3 Statistical analysis

      Patient characteristics were examined by simple frequency tables for categorical variables, and means and standard deviations for continuous variables. Non-parametric tests for differences in characteristics between patients optimally and sub-optimally controlled were performed using chi-square test/Fisher’s exact test (when cells had expected counts less than 5) and Wilcoxon/Kruskal Wallis test. The resultant p-value of a chi-square test/Fisher’s exact test represents the association between two variables of a contingency table, each with at least two categorical levels (e.g. gender [male/female] vs. glycaemic control [< or ⩾HbA1c 7.0%]). The analysis was conducted for individual participating countries and all countries combined; a pooled analysis was feasible given that all data were collected in a standardized way across countries. Two-sided testing at α = 0.05 level of significance was used. The analysis was performed only on available data (complete case analysis), as it was assumed that any missing data would have been due to random chance; in addition, differences in baseline characteristics of included and excluded individuals were examined.
      To account for mutual confounding, the association between selected exposure variables and HbA1c at 24 weeks was examined by multiple logistic regression (SAS enterprise guide 7.1, Cary, NC, USA). Wald estimates, 95% confidence interval (CI), and statistical significance were calculated by iterative maximum likelihood methods. Included variables were selected a priori based on prior knowledge of their association with HbA1c. Crude and adjusted results are presented. The main model was adjusted for country, HbA1c before insulin initiation, insulin dose at interim/final visit, sex, age, duration of diabetes, body mass index (BMI) before insulin initiation, microvascular complications, macrovascular complications, number of OADs before insulin initiation, duration of OAD treatment, change in number of OADs during study, any hypoglycemia during study, and any hypoglycemia before insulin initiation. FBG was not included in the main model, as there were too many missing values for this variable; instead, an identical model to the one described above, but with FBG included, was conducted separately to assess the association of FBG and HbA1c.

      3. Results

      Of the SOLVE population (n = 17,374), 10,763 patients had HbA1c information at the interim and final visits and therefore were included in the post hoc analysis. The excluded population were marginally younger (61.4 ± 11.7 years vs. 61.7 ± 11.4 years; p = 0.046), and had greater baseline HbA1c (9.0 ± 1.7% [74.9 ± 18.6 mmol/mol] vs. 8.8 ± 1.6% [72.7 ± 17.5 mmol/mol]; p< 0.0001), FBG (10.8 ± 3.52 mmol/l [194 ± 63 mg/dl] vs. 10.2 ± 3.0 mmol/l [184 ± 54 mg/dl]; p< 0.0001) and BMI (29.7 ± 5.5 kg/m2 vs. 29.0 ± 5.2 kg/m2; p< 0.0001) compared with the analysis population. In addition, the excluded population presented a greater use of sulphonylureas (61.9% vs. 57.8%, p < 0.0001) and reduced use of glinides (14.7% vs. 16.9%, p < 0.0001) and α-glucosidase inhibitors (10.2% vs. 13.4%, p < 0.0001) compared with the analysis population.
      After 24 weeks of basal insulin treatment, 66.3% (n = 7138) of the analysis population had an HbA1c level ⩾ 7.0% [⩾53.0 mmol/mol]).

      3.1 Baseline characteristics

      Based on a non-adjusted analysis of the pooled population, significantly different baseline characteristics in patients who had HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) after 24 weeks included older age (p < 0.0001), longer duration of diabetes (p< 0.0001), higher weight (p< 0.0001), greater use of OADs (p< 0.0001; specifically biguanides, sulphonylureas, and α-glucosidase inhibitors), and more frequent micro- and macrovascular complications (p< 0.0001 and p < 0.05, respectively) compared with patients who achieved HbA1c < 7.0% (Table 1). Prevalence of HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) at final visit was also related to country, with the highest prevalence of HbA1c ⩾ 7.0% in the UK (86.9%) and the lowest in China (49.7%). Similar trends in baseline characteristics were observed for individual countries (data not shown).
      Table 1Baseline characteristics of patients with type 2 diabetes mellitus optimally (HbA1c < 7.0% [<53.0 mmol/mol]) and suboptimally controlled (HbA1c ⩾ 7.0% [⩾53.0 mmol/mol]) on basal insulin after 24 weeks.
      HbA1c < 7.0% (<53.0 mmol/mol) n = 3625HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) n = 7138P-value
      Test for differences using chi-square test and Kruskal Wallis test. BMI, body mass index; OAD, oral antidiabetic drug.
      Gender, % (n)
       Men54.87 (1988)52.15 (3719)0.007
       Women45.13 (1635)47.85 (3413)
      Mean age, years (SD)60.4 (11.8)62.36 (11.1)<0.0001
      Age group frequency, % (n)
       <55 years62.73 (2260)57.61 (4090)<0.0001
       55–75 years14.40 (519)16.68 (1184)
       >75 years22.87 (824)25.72 (1826)
      Mean age at diagnosis, years (SD)52.1 (10.9)51.70 (10.7)0.112
      Mean duration of diabetes, years (SD)8.3 (6.8)10.6 (7.2)<0.0001
      Duration of diabetes group frequency, % (n)
       <5 years33.01 (1193)18.71 (1331)<0.0001
       5–10 years38.30 (1384)40.17 (2857)
       >10 years28.69 (1037)41.11 (2924)
      Insulin dose at baseline, U/kg (SD)0.15 (0.08)0.16 (0.08)0.242
      Mean weight at baseline, kg (SD)77.9 (16.5)81.2 (17.5)<0.0001
      Mean BMI at baseline, kg/m2 (SD)27.9 (4.9)29.5 (5.3)<0.0001
      Microvascular complications, % (n)
       No70.72 (2534)65.29 (4584)<0.0001
       Yes29.28 (1049)34.71 (2437)
      Macrovascular complications, % (n)
       No75.34 (2688)73.16 (5120)0.017
       Yes24.66 (880)26.84 (1878)
      Number of OADs at insulin initiation, % (n)
       137.56 (1335)26.26 (1851)<0.0001
       248.87 (1737)56.31 (3969)
       >213.56 (482)17.44 (1229)
      Mean duration of OADs, years (SD)7.2 (6.2)9.2 (6.8)<0.0001
      Type of OADs, % (n)
       Biguanides77.63 (2759)82.68 (5828)<0.0001
       Sulphonylureas48.26 (1715)62.62 (4414)<0.0001
       Glinides17.64 (627)16.58 (1169)0.170
       α-glucosidase inhibitors9.99 (355)13.14 (926)<0.0001
       Thiazolidinediones17.30 (615)11.41 (804)<0.0001
       DPP-4 inhibitors6.50 (231)6.13 (432)0.470
       Other0.06 (2)0.17 (12)0.162
      Country, % (n)
       Canada20.60 (89)79.40 (343)<0.0001
       Turkey34.07 (445)65.93 (861)
       Italy21.66 (664)78.34 (2401)
       Spain37.98 (196)62.02 (320)
       Germany40.08 (620)59.92 (927)
       Poland34.93 (160)65.07 (298)
       Israel18.87 (57)81.13 (245)
       UK13.11 (54)86.89 (358)
       China50.29 (1288)49.71(1273)
       Portugal31.71 (52)68.29 (112)
      a Test for differences using chi-square test and Kruskal Wallis test. BMI, body mass index; OAD, oral antidiabetic drug.

      3.2 Association between baseline characteristics and suboptimal glycemic control

      Crude and adjusted multiple logistic regression results are described in Table 2. In the adjusted model, the baseline characteristics that remained significantly and independently associated with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) at final visit were duration of diabetes, number of OADs, and duration of OAD treatment. Gender, age, micro- and macrovascular complications (although borderline significant) were not independently associated with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol).
      Table 2Multiple logistic regression of patient characteristics on the probability of HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) after 24 weeks.
      Unadjusted odds ratio (95% CI)P-valueAdjusted odds ratio (95% CI)
      Model adjusted for country, HbA1c before insulin initiation, insulin dose at final visit, sex, age, duration of diabetes, BMI before insulin initiation, microvascular complications, macrovascular complications, number of OADs before insulin initiation, duration of OAD treatment, change in OD treatment during study, any hypoglycemia during study, and any hypoglycemia before insulin initiation.
      P-value
      HbA1c before insulin initiation in %1.65 (1.59, 1.71)<0.00011.56 (1.50, 1.62)<0.0001
      FBG before insulin initiation in mmol/l
      FBG was calculated using a separate model as there were many missing FBG values.
      1.16 (1.14, 1.19)<0.00011.01 (0.99, 1.05)0.290
      BMI before insulin initiation in kg/m21.06 (1.05, 1.07)<0.00011.03 (1.02, 1.04)<0.0001
      Duration of OAD treatment in years1.05 (1.05, 1.06)<0.00011.03 (1.01, 1.04)<0.0001
      Gender
       Men1.0 (ref)0.0071.0 (ref)0.367
       Women1.11 (1.03, 1.21)1.05 (0.94, 1.17)
      Age in years
       <55 years1.0 (ref)<0.00011.0 (ref)0.533
       55–75 years1.26 (1.12, 1.41)0.95 (0.81, 1.11)
       >75 years1.22 (1.11, 1.35)0.92 (0.80, 1.07)
      Duration of diabetes in years
       <5 years1.0 (ref)<0.00011.0 (ref)<0.0001
       5–10 years1.85 (1.67, 2.04)1.44 (1.25, 1.66)
       >10 years2.52 (2.27, 2.80)1.44 (1.17, 1.77)
      Microvascular complications
       No1.0 (ref)<0.00011.0 (ref)0.116
       Yes1.28 (1.18, 1.40)1.10 (0.98, 1.23)
      Macrovascular complications
       No1.0 (ref)0.0161.0 (ref)0.082
       Yes1.12 (1.02, 1.23)0.89 (0.79, 1.01)
      Number of OADs before insulin initiation
       11.0 (ref)<0.00011.0 (ref)0.0003
       21.65 (1.50, 1.80)1.27 (1.12, 1.44)
       >21.84 (1.62, 2.09)1.38 (1.14, 1.66)
      Country
       Canada1.06 (0.83, 1.37)<0.00011.03 (0.76, 1.38)<0.0001
       Turkey0.53 (0.46, 0.62)0.41 (0.33, 0.52)
       Italy1.0 (ref)1.0 (ref)
       Spain0.45 (0.37, 0.55)0.44 (0.33, 0.59)
       Germany0.41 (0.36, 0.47)0.52 (0.43, 0.62)
       Poland0.51 (0.41, 0.63)0.76 (0.58, 0.99)
       Israel1.19 (0.88, 1.60)1.01 (0.71, 1.45)
       UK1.83 (1.36, 2.47)1.28 (0.89, 1.83)
       China0.27 (0.24, 0.30)0.47 (0.40, 0.56)
       Portugal0.59 (0.44, 0.84)0.58 (0.38, 0.90)
      Any hypoglycemia before study
       No1.0 (ref)0.8021.0 (ref)0.120
       Yes0.98 (0.81, 1.17)1.20 (0.95, 1.53)
      Any hypoglycemia during study
       No1.0 (ref)<0.00011.0 (ref)<0.0001
       Yes0.76 (0.67, 0.86)0.71 (0.60, 0.84)
      Insulin dose at final visit in 0.1 units/kg1.23 (1.19, 1.27)<0.00011.13 (1.09, 1.18)<0.0001
      Change in number of OADs during study
       Decreased1.05 (0.95, 1.16)0.6140.99 (0.86, 1.14)0.588
       Unchanged1.0 (ref)1.0 (ref)
       Increased1.00 (0.85, 1.12)0.89 (0.72, 1.10)
      FBG, fasting blood glucose; OAD, oral antidiabetic drug; OD, once daily.
      a Model adjusted for country, HbA1c before insulin initiation, insulin dose at final visit, sex, age, duration of diabetes, BMI before insulin initiation, microvascular complications, macrovascular complications, number of OADs before insulin initiation, duration of OAD treatment, change in OD treatment during study, any hypoglycemia during study, and any hypoglycemia before insulin initiation.
      b FBG was calculated using a separate model as there were many missing FBG values.

      3.3 Glycemic parameters

      Patients with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) at final visit had significantly greater HbA1c and FBG levels at baseline and 24 weeks compared with patients achieving HbA1c target (Fig. 1). After 24 weeks of treatment, patients with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) experienced a statistically significant reduction in HbA1c of −1.10 ± 1.5% (−12.0 ± 16.4 mmol/mol; p < 0.0001) from baseline, whereas patients with HbA1c < 7.0% had a reduction of −1.73 ± 1.5% (−18.9 ± 16.4 mmol/mol; p < 0.0001) from baseline.
      Figure thumbnail gr1
      Fig. 1Glycemic control after 24 weeks of treatment in patients with type 2 diabetes mellitus optimally (HbA1c < 7.0% [<53.0 mmol/mol]) and suboptimally controlled (HbA1c ⩾ 7.0% [⩾53.0 mmol/mol]) on basal insulin: (A) HbA1c and (B) fasting blood glucose. *p < 0.0001.
      After 24 weeks of treatment, FBG was significantly reduced from baseline by −3.1 ± 3.1 mmol/l (−55.8 ± 55.8 mg/dl) in patients with HbA1c ⩾ 7.0% (53.0 mmol/mol) at final visit, and −3.2 ± 2.8 mmol/l (−57.6 ± 50.4 mg/dl) in patients with HbA1c < 7.0% (both p < 0.0001) at final visit. Final values were 7.7 ± 2.0 mmol/l (139 ± 36 mg/dl) and 6.3 ± 1.0 mmol/l (113 ± 18 mg/dl), respectively. In the adjusted regression model, HbA1c ⩾ 7.0% at final visit) was positively associated with the HbA1c level at baseline, but not with the FBG level at baseline (Table 2).

      3.4 Insulin dose

      Insulin dose significantly increased by 0.12 ± 0.17 U/kg from baseline in patients with final HbA1c ⩾ 7.0 (53.0 mmol/mol) (p < 0.0001), and by 0.08 ± 0.13 U/kg in patients with final HbA1c < 7.0% (p< 0.0001). Final insulin dose values were significantly different from each other: 0.28 ± 0.17 U/kg (HbA1c ⩾ 7.0%) and 0.23 ± 0.14 U/kg (HbA1c < 7.0%) (p < 0.0001). In the adjusted regression model, HbA1c ⩾ 7.0% at final visit was positively associated with the insulin dose at 24 weeks (Table 2).

      3.5 Safety

      After 24 weeks of treatment, the reporting of any hypoglycemia during the study was significantly less in patients with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) compared with patients with HbA1c < 7.0% at final visit (Fig. 2). In terms of major hypoglycemia, 0.04% (n = 3) of patients with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) experienced an episode after 24 weeks compared with 0.19% (n = 7) of patients with HbA1c < 7.0% at final visit. The adjusted regression model supports the negative relationship between hypoglycemia during treatment and achieving HbA1c ⩾ 7.0% at 24 weeks (Table 2).
      Figure thumbnail gr2
      Fig. 2Hypoglycemia reported after 24 weeks of treatment in patients with type 2 diabetes mellitus optimally (HbA1c < 7.0% [<53.0 mmol/mol]) and suboptimally controlled (HbA1c ⩾ 7.0% [⩾53.0 mmol/mol]) on basal insulin. Any hypoglycemia includes minor (daytime or nocturnal) and major episodes.
      Mean weight and BMI were significantly greater at baseline in patients with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) at final visit compared with patients achieving HbA1c target (Table 1). Body weight was significantly reduced by 0.4 ± 5.5 kg from baseline in patients with HbA1c ⩾ 7.0%, and by 0.9 ± 5.6 kg in patients with HbA1c < 7.0% (both p < 0.0001). In the adjusted model, HbA1c ⩾ 7.0% at final visit was positively associated with BMI at baseline (Table 2).

      4. Discussion

      Patients with suboptimal glycemic control, defined as HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) at final visit, were positively associated with baseline characteristics related to severity of disease; this included baseline HbA1c (the odds of not achieving glycemic target were 1.56 times greater for every 1.0% higher baseline HbA1c), duration of diabetes, BMI, number of OADs, and duration of OAD treatment. In contrast, suboptimal glycemic control was negatively associated with the reporting of hypoglycemia, i.e. more hypoglycemia was reported in patients who achieved glycemic target compared with those who did not. Suboptimally controlled patients also had a significantly higher insulin dose requirement after 24 weeks of treatment compared with patients with good glycemic control. These results support the findings of similar studies that have investigated predictors of glycemic control [
      • Balkau B.
      • Francois C.G.
      • Freemantle N.
      • et al.
      Predictors of HbA1c over 4 years in people with type 2 diabetes starting insulin therapies: the CREDIT study.
      ,
      • Home P.D.
      • Shen C.
      • Hasan M.I.
      • et al.
      Predictive and explanatory factors of change in HbA1c in a 24-week observational study of 66,726 people with type 2 diabetes starting insulin analogs.
      ,
      • Riddle M.C.
      • Vlajnic A.
      • Zhou R.
      • et al.
      Baseline HbA1c predicts attainment of 7.0% HbA1c target with structured titration of insulinglargine in type 2 diabetes: a patient-level analysis of 12 studies.
      ]. They all unequivocally demonstrate that baseline HbA1c is the strongest predictor associated with glycemic control. In a long-term 4-year study in patients with T2DM starting any insulin, suboptimal glycemic control was also associated with other baseline factors including BMI, age, diabetes duration, and use of OADs [
      • Balkau B.
      • Francois C.G.
      • Freemantle N.
      • et al.
      Predictors of HbA1c over 4 years in people with type 2 diabetes starting insulin therapies: the CREDIT study.
      ]. Balkau and colleagues also reported a strong association between high insulin dose and high HbA1c level. As in our study, this finding is rather unexpected, given that a higher insulin dose should reduce HbA1c and, indeed, HbA1c was reduced over the 24 weeks of the SOLVE study [
      • Khunti K.
      • Caputo S.
      • Damci T.
      • et al.
      The safety and efficacy of adding once-daily insulin detemir to oral hypoglycaemic agents in patients with type 2 diabetes in a clinical practice setting in 10 countries.
      ]. This finding is likely to be driven by reverse causation, i.e. high HbA1c caused the high insulin requirement, and not the other way around.
      Glycemic control varied across the participating countries, with the UK presenting the largest proportion of patients with HbA1c ⩾ 7.0% (86.9%) and China the lowest (49.7%). This difference is likely to be due to heterogeneity between the two populations prior to initiating the study. HbA1c level, insulin dose, population age, weight, and the incidence of macrovascular disease at baseline were all lower in the Chinese SOLVE population (data not shown); thus, with fewer comorbidities, the Chinese population may indeed be less insulin-resistant than the UK population. Furthermore, differences in OAD use may also contribute to the resulting glycemic profile; for example, the use of α-glucosidase inhibitors and glinides was more pronounced in the Chinese population compared with the UK population (data not shown).
      Although glycemic control was significantly improved in both groups of patients, there remains a need to further optimize the treatment of those patients who did not achieve glycemic target. At final visit, patients with HbA1c ⩾ 7.0% (⩾53.0 mmol/mol) had an FBG level of 7.7 mmol/l (138.6 mg/dl), which stands closely above the recommended FBG targets of 4.4–7.2 mmol/l (80–130 mg/dl) [
      • American Diabetes Association
      Standard of Medical Care in Diabetes–2016.
      ] or <6.1 mmol/l (110 mg/dl) [
      AACE/ACE comprehensive diabetes management algorithm.
      ]. In a randomized, treat-to-target, 20-week study in insulin-naïve patients with T2DM starting basal insulin therapy, Blonde and colleagues demonstrated that the majority of patients could achieve HbA1c < 7.0% (<53.0 mmol/mol) while titrating to an FBG target of 3.9–5.0 mmol/l (70.2–90 mg/dl) or 4.4–6.1 mmol/l (79.2–110 mg/dl) [
      • Blonde L.
      • Merilainen M.
      • Karwe V.
      • et al.
      Patient-directed titration for achieving glycaemic goals using a once-daily basal insulin analogue: an assessment of two different fasting plasma glucose targets – the TITRATETM study.
      ]. This was achieved while accompanied by low hypoglycemia rates. In our analysis, patients with suboptimal glycemic control also experienced significantly less hypoglycemia during the study compared with well-controlled patients, which suggests that their treatment could indeed be further titrated towards glycemic targets without adversely increasing the risk of hypoglycemia.
      The above is also supported by the fact that, although insulin doses increased in both groups, the final dose in the suboptimally controlled group (0.28 U/kg) was relatively low compared with the insulin doses (0.4–0.6 U/kg) reported in randomized clinical trials that assessed the efficacy and safety of insulin detemir in insulin-naïve patients with T2DM [
      • Blonde L.
      • Merilainen M.
      • Karwe V.
      • et al.
      Patient-directed titration for achieving glycaemic goals using a once-daily basal insulin analogue: an assessment of two different fasting plasma glucose targets – the TITRATETM study.
      ,
      • Rosenstock J.
      • Davies M.
      • Home P.D.
      • et al.
      A randomised, 52-week, treat-to-target trial comparing insulin detemir with insulin glargine when administered as add-on to glucose-lowering drugs in insulin-naïve people with type 2 diabetes.
      ,
      • Philis-Tsimikas A.
      • Charpentier G.
      • Clauson P.
      • et al.
      Comparison of once-daily insulin detemir with NPH insulin added to a regimen of oral antidiabetic drugs in poorly controlled type 2 diabetes.
      ]. This may not be that unexpected given that titration procedures in the SOLVE study were conducted at the investigators’ discretion in routine clinical care. Furthermore, physicians may be more careful when titrating insulin in patients who present the features identified in the suboptimally controlled population (i.e. long duration of diabetes, high BMI, etc.) in order to reduce the risk of severe hypoglycemia and/or cardiovascular complications. Nevertheless, the findings of the analysis appear to support that, in some patients, there is a need for more aggressive titration towards recommended glycemic targets while being mindful of the risk of severe hypoglycemia. Treatment intensification with prandial insulin or other glucose-lowering drugs (e.g. glucagon-like peptide-1 receptor agonists [GLP-1RAs]) that target postprandial glucose could also be considered, particularly as final FBG levels were not far from the recommended target [
      • Inzucchi S.E.
      • Bergenstal R.M.
      • Buse J.B.
      • et al.
      Management of hyperglycaemia in type 2 diabetes, 2015: a patient-centred approach. Update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes.
      ]; however, American Diabetes Association guidelines only recommend this approach once the basal insulin dose is greater than 0.5 U/kg or if HbA1c remains uncontrolled despite achieving FBG target [
      • Inzucchi S.E.
      • Bergenstal R.M.
      • Buse J.B.
      • et al.
      Management of hyperglycaemia in type 2 diabetes, 2015: a patient-centred approach. Update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes.
      ]. Given that neither of the above requirements strictly apply to our population, more intensive titration would appear to be the best initial option to help attain treatment goals.
      Intensive treatment, particularly with insulin, can have an adverse effect on weight gain [
      • DCCT
      Influence of intensive diabetes treatment on body weight and composition of adults with type 1 diabetes in the Diabetes Control and Complications Trial.
      ,
      • UKPDS
      Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group.
      ], and this may result in hesitation when further titration of basal insulin treatment is required. However, it should be noted that, in the SOLVE study, regardless of glycemic control, minor weight reduction was seen in both groups. Although extrinsic factors may play a role, the weight-sparing effect of insulin detemir could be attributed to a CNS-mediated reduction in energy intake [
      • Russell-Jones D.
      • Danne T.
      • Hermansen K.
      • et al.
      Weight-sparing effect of insulin detemir: a consequence of central nervous system-mediated reduced energy intake?.
      ]. Furthermore, randomized clinical trials have demonstrated a less adverse effect on weight gain with insulin detemir compared with other basal insulin, including NPH insulin [
      • Philis-Tsimikas A.
      • Charpentier G.
      • Clauson P.
      • et al.
      Comparison of once-daily insulin detemir with NPH insulin added to a regimen of oral antidiabetic drugs in poorly controlled type 2 diabetes.
      ] and insulin glargine [
      • Rosenstock J.
      • Davies M.
      • Home P.D.
      • et al.
      A randomised, 52-week, treat-to-target trial comparing insulin detemir with insulin glargine when administered as add-on to glucose-lowering drugs in insulin-naïve people with type 2 diabetes.
      ].
      Other factors may hinder the intensification of basal insulin treatment, as demonstrated in a recent survey that assessed the perceptions on control and insulin intensification among physicians and patients with suboptimally controlled T2DM on basal insulin [
      • Brod M.
      • Pfeiffer K.M.
      • Barnett A.H.
      • et al.
      Perceptions of diabetes control among physicians and people with type 2 diabetes uncontrolled on basal insulin in Sweden, Switzerland, and the United Kingdom.
      ]. There was considerable reluctance by physicians to intensify treatment when an agreement could not be reached with the patient, while patients themselves were more concerned about weight gain and the perception of a worsening disease. A UK qualitative study also highlighted healthcare provider perceptions on clinical inertia, which identified patient factors (comorbidities and fallibility) and system-level factors (time constraints) as key contributors; however, the participants themselves seemed reluctant to accept a degree of responsibility for the inertia [
      • Zafar A.
      • Stone M.A.
      • Davies M.J.
      • et al.
      Acknowledging and allocating responsibility for clinical inertia in the management of Type 2 diabetes in primary care: a qualitative study.
      ]. These factors are likely to be among several that contribute to the clinical inertia reported in patients on basal insulin. Remarkably, a UK retrospective analysis in T2DM showed that it took approximately 4 years to intensify patients’ treatment following basal insulin, and this was conducted in only one-third of the patients who were actually eligible for intensification [
      • Khunti K.
      • Nikolajsen A.
      • Thorsted B.L.
      • et al.
      Clinical inertia with regard to intensifying therapy in people with type 2 diabetes treated with basal insulin.
      ]. Duration of diabetes, older age, and OAD usage were significantly associated with the reported inertia in intensifying treatment [
      • Khunti K.
      • Nikolajsen A.
      • Thorsted B.L.
      • et al.
      Clinical inertia with regard to intensifying therapy in people with type 2 diabetes treated with basal insulin.
      ], factors that, based on our analysis, also appear to be associated with suboptimal glycemic control. Early identification of these factors in clinical practice is important to help physicians identify in a timely manner those patients likely to require modifications to their treatment.
      In terms of strengths, our analysis had a large sample size and is fairly representative of primary and specialist care across 10 countries, and thus has relatively high external validity. Importantly, countries/regions may use different glycemic targets and, to account for this, we also conducted the analysis with an HbA1c cut-off of 7.5% (58.5 mmol/mol); this did not impact upon the findings (Supplementary Material). In addition, no differences were observed when conducting the analysis after 12 weeks of treatment (Supplementary Material); this was studied to help account for any potential differences in outcomes that may arise due to more active insulin titration typical of the first 3 months of treatment. Potential confounding was accounted for by multiple logistic regression, although this would not have been able to account for socioeconomic status and lifestyle factors. Limitations of the analysis include the possibility of reverse causation (as reported with insulin dose and HbA1c) and false-positive findings driven by multiple testing, although the data were consistently reported across countries. Of the overall SOLVE population, 38% of subjects were not eligible for this analysis due to missing data; it cannot be ruled out that inclusion of these data could influence the findings. There were many missing values for FBG, thus it was not included in the main adjusted model. Although some factors were statistically significant in the adjusted model (e.g. BMI, number of OADs), the odds ratios were very close to 1 and therefore the magnitude of change in these outcomes should be weighed against their clinical relevance. It should be highlighted that SOLVE was an observational study and, as such, any observed associations may not be causal; however, the observational nature of the study also helps reflect attitudes of real-world clinical practice.
      In conclusion, we identified several baseline characteristics that were strongly associated with suboptimal glycemic control in patients with T2DM who were receiving once-daily basal insulin. These factors may help clinicians in identifying patients who may require individualized titration or intensification of treatment. Indeed, in the suboptimally controlled group of the SOLVE population, insulin dose and the risk of hypoglycemia were relatively low, which, in combination with suboptimal FPG levels, suggests that a more aggressive titration regimen could be implemented to improve glycemic control while still maintaining a low risk of hypoglycemia. Alternatively, intensification of basal insulin treatment by adding postprandial glucose-lowering drugs like bolus insulin or GLP-1RAs can be considered in this patient group.

      Author contributions

      KK, AL, TD, and VB were involved with analysis design, conceptualization of the manuscript, and contributed to the discussion and writing of the manuscript. LLNH performed the statistical analyses, and contributed to the discussion and writing of the manuscript. All authors have approved the final article.

      Role of the funding source

      The analysis was sponsored by Novo Nordisk, who were involved in the analysis and interpretation of data, and the discussion and writing of the manuscript.

      Declaration of interests

      LLNH and VB are employees at Novo Nordisk. KK, AL and TD received support from Novo Nordisk to attend meetings to discuss the design, analysis, and interpretation of the original SOLVE study. KK was a principle investigator for the study in his respective country. KK, AL, and TD have received funding for membership on Novo Nordisk Advisory Boards and/or consulting services.

      Acknowledgements

      The authors would like to thank Steven Barberini and Helen Marshall of Watermeadow Medical, an Ashfield company, for editorial assistance. The analysis was sponsored by Novo Nordisk . KK also acknowledges support for this article from the UK National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care – East Midlands (NIHR CLAHRC – EM) and the NIHR Diet, Lifestyle & Physical Activity Biomedical Research Unit based at University Hospitals of Leicester and Loughborough University.

      Appendix A. Supplementary data

      References

      1. AACE/ACE comprehensive diabetes management algorithm.
        Endocr Pract. 2015; 21
        • American Diabetes Association
        Standard of Medical Care in Diabetes–2016.
        Diabetes Care. 2016; 39: S1
      2. National Institute for Health and Care Excellence. Type 2 diabetes in adults: management. NG28. December 2015.

        • Khunti K.
        • Wolden M.L.
        • Thorsted B.L.
        • et al.
        Clinical inertia in people with type 2 diabetes: a retrospective cohort study of more than 80,000 people.
        Diabetes Care. 2013; 36: 3411-3417
        • Gerstein H.C.
        • Miller M.E.
        • Genuth S.
        • et al.
        Long-term effects of intensive glucose lowering on cardiovascular outcomes.
        N Engl J Med. 2011; 364: 818-828
        • Holman R.R.
        • Farmer A.J.
        • Melanie M.J.
        • et al.
        Three-year efficacy of complex insulin regimens in type 2 diabetes.
        N Engl J Med. 2009; 361: 1736-1747
        • Peyrot M.
        • Rubin R.R.
        • Lauritzen T.
        • et al.
        Resistance to insulin therapy among patients and providers: results of the cross-national Diabetes Attitudes, Wishes, and Needs (DAWN) study.
        Diabetes Care. 2005; 28: 2673-2679
        • Stargardt T.
        • Gonder-Frederick L.
        • Krobot K.J.
        • et al.
        Fear of hypoglycaemia: defining a minimum clinically important difference in patients with type 2 diabetes.
        Health Qual Life Outcomes. 2009; 7: 91
        • Peyrot M.
        • Rubin R.R.
        • Khunti K.
        Addressing barriers to initiation of insulin in patients with type 2 diabetes.
        Prim Care Diabetes. 2010; 4: S11-S18
        • Blonde L.
        • Merilainen M.
        • Karwe V.
        • et al.
        Patient-directed titration for achieving glycaemic goals using a once-daily basal insulin analogue: an assessment of two different fasting plasma glucose targets – the TITRATETM study.
        Diabetes Obes Metab. 2009; 11: 623-631
        • Khunti K.
        • Caputo S.
        • Damci T.
        • et al.
        The safety and efficacy of adding once-daily insulin detemir to oral hypoglycaemic agents in patients with type 2 diabetes in a clinical practice setting in 10 countries.
        Diabetes Obes Metab. 2012; 14: 1129-1136
        • Khunti K.
        • Nikolajsen A.
        • Thorsted B.L.
        • et al.
        Clinical inertia with regard to intensifying therapy in people with type 2 diabetes treated with basal insulin.
        Diabetes Obes Metab. 2016; 18: 401-409
        • Inzucchi S.E.
        • Bergenstal R.M.
        • Buse J.B.
        • et al.
        Management of hyperglycaemia in type 2 diabetes, 2015: a patient-centred approach. Update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes.
        Diabetologia. 2015; 58: 429-442
        • Balkau B.
        • Francois C.G.
        • Freemantle N.
        • et al.
        Predictors of HbA1c over 4 years in people with type 2 diabetes starting insulin therapies: the CREDIT study.
        Diabetes Res Clin Prac. 2015; 108: 432-440
        • Home P.D.
        • Shen C.
        • Hasan M.I.
        • et al.
        Predictive and explanatory factors of change in HbA1c in a 24-week observational study of 66,726 people with type 2 diabetes starting insulin analogs.
        Diabetes Care. 2014; 37: 1237-1245
        • Riddle M.C.
        • Vlajnic A.
        • Zhou R.
        • et al.
        Baseline HbA1c predicts attainment of 7.0% HbA1c target with structured titration of insulinglargine in type 2 diabetes: a patient-level analysis of 12 studies.
        Diabetes Obes Metab. 2013; 15: 819-825
        • Rosenstock J.
        • Davies M.
        • Home P.D.
        • et al.
        A randomised, 52-week, treat-to-target trial comparing insulin detemir with insulin glargine when administered as add-on to glucose-lowering drugs in insulin-naïve people with type 2 diabetes.
        Diabetologia. 2008; 51: 408-416
        • Philis-Tsimikas A.
        • Charpentier G.
        • Clauson P.
        • et al.
        Comparison of once-daily insulin detemir with NPH insulin added to a regimen of oral antidiabetic drugs in poorly controlled type 2 diabetes.
        Clin Ther. 2006; 28: 1569-1581
        • DCCT
        Influence of intensive diabetes treatment on body weight and composition of adults with type 1 diabetes in the Diabetes Control and Complications Trial.
        Diabetes Care. 2001; 24: 1711-1721
        • UKPDS
        Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group.
        Lancet. 1998; 352: 837-853
        • Russell-Jones D.
        • Danne T.
        • Hermansen K.
        • et al.
        Weight-sparing effect of insulin detemir: a consequence of central nervous system-mediated reduced energy intake?.
        Diabetes Obes Metab. 2015; 17: 919-927
        • Brod M.
        • Pfeiffer K.M.
        • Barnett A.H.
        • et al.
        Perceptions of diabetes control among physicians and people with type 2 diabetes uncontrolled on basal insulin in Sweden, Switzerland, and the United Kingdom.
        Curr Med Res Opin. 2016; 7: 1-9
        • Zafar A.
        • Stone M.A.
        • Davies M.J.
        • et al.
        Acknowledging and allocating responsibility for clinical inertia in the management of Type 2 diabetes in primary care: a qualitative study.
        Diabet Med. 2015; 32: 407-413