Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis

ARTICLE IN PRESS

PCD-448; No. of Pages 8

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

Contents lists available at ScienceDirect

Primary Care Diabetes journal homepage: http://www.elsevier.com/locate/pcd

Original research

Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis Stefan Pscherer a , Engels Chou b , Franz-Werner Dippel c , Wolfgang Rathmann d,1 , Karel Kostev e,∗,1 a

Klinisches Diabeteszentrum Süd-Ostbayern, Traunstein, Germany Sanofi-Aventis, NJ, USA c Sanofi-Aventis Deutschland GmbH, Berlin, Germany d Institute of Biometrics and Epidemiology, German Diabetes Center, Duesseldorf, Germany e IMS HEALTH, Frankfurt, Germany b

a r t i c l e

i n f o

a b s t r a c t

Article history:

Aims: To compare persistence and its predictors in type 2 diabetes patients in primary care,

Received 23 July 2014

initiating either basal supported oral therapy (BOT) or intensified conventional therapy (ICT)

Received in revised form

with glargine, detemir, or NPH insulin.

8 January 2015

Methods: In the BOT cohort, 1398 glargine (mean age: 68 years), 292 detemir (66 years), and

Accepted 25 January 2015

874 NPH (65 years) users from 918 practices were retrospectively analyzed (Disease Analyzer,

Available online xxx

Germany: 2008–2012). The ICT group incorporated 866 glargine (64 years), 512 detemir (60 years), and 1794 NPH (64 years) new users. Persistence was defined as proportion of patients

Keywords:

remaining on the initial basal insulin (glargine, detemir and NPH insulin) over 2 years. Per-

Insulin initiation

sistence was evaluated by Kaplan–Meier curves (log-rank tests) and Cox regression adjusting

Type 2 diabetes

for age, sex, diabetes duration, antidiabetic co-therapy, comorbidities, specialist care, and

Persistence

private health insurance.

Basal insulin

Results: In BOT, two-year persistence was 65%, 53%, and 59% in glargine, detemir, and NPH

Primary care

users, respectively (p < 0.001). In ICT, persistence was higher without differences between groups: 84%, 85%, 86% in glargine, detemir, and NPH, respectively (p = 0.536). In BOT, detemir and NPH users were more likely to discontinue basal insulin compared with glargine (detemir vs. glargine: adjusted Hazard Ratio; 95% CI: 1.56; 1.31–1.87; NPH vs. glargine: 1.22; 1.07–1.38). Heart failure (1.39; 1.16–1.67) was another predictor of non-persistence, whereas higher age (per year: 0.99; 0.98–0.99), metformin (0.61; 0.54–0.69), and sulfonylurea co-medication (0.86; 0.77–0.97) were associated with lower discontinuation. Conclusions: In BOT, treatment persistence among type 2 diabetes patients initiating basal insulin is influenced by type of insulin, antidiabetic co-medication, and patient characteristics. © 2015 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

∗ Corresponding author at: CES LifeLink-Epidemiology & Pharmacovigilance, IMS HEALTH GmbH & Co OHG, Darmstädter Landstrasse 108, Frankfurt am Main 60598, Germany. E-mail address: [email protected] (K. Kostev). 1 These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.pcd.2015.01.011 1751-9918/© 2015 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011

PCD-448; No. of Pages 8

ARTICLE IN PRESS

2

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

1.

Introduction

In the long run insulin will be required in most type 2 diabetic patients to maintain glycemic control due to the progressive beta cell dysfunction [1]. Current guidelines from the American Diabetes Association and the European Association for the Study of Diabetes note that the majority of patients with type 2 diabetes requiring insulin therapy can be successfully treated with basal insulins [2]. Either intermediate-acting (neutral protamine Hagedorn [NPH]) or long-acting analogs such as insulin glargine or insulin detemir may be used [2]. However, because of progressive loss of endogenous insulin secretion, some patients will require additional prandial insulin therapy with short-acting insulin [2]. A Cochrane Review of randomized clinical trials comparing insulin glargine and detemir concluded that there were no clinically relevant differences in efficacy or safety, however, to achieve the same glycemic control, insulin detemir was often injected twice-daily and higher doses were needed [3]. Compared to NPH and detemir insulin glargine in addition showed fewer hypoglycemic events [4]. Several studies suggested that a large proportion of type 2 diabetes patients have difficulties managing their antidiabetic medications including insulin [5]. This often results in low treatment persistence, which has been defined as the proportion of patients who remained on treatment for a specific time or the duration of time from initiation to discontinuation of therapy [5,6]. Few studies have examined the persistence of type 2 diabetes patients on basal insulins in a real-world setting and previous results have been somewhat controversial [7,8]. In Germany, there is a lack of real-world evidence studies on the basal insulin treatment persistence and related factors in type 2 diabetes patients in primary care. Thus, the objectives of this study were to describe persistence of basal insulin use (glargine, detemir, NPH) in type 2 diabetes patients in primary care and to determine risk factors for poor persistence. Two patient cohorts with either basal supported oral therapy (BOT) or intensified conventional therapy (ICT) were analyzed. The study applied a retrospective approach using a nationwide primary care database in Germany.

2.

Patients and methods

The Disease Analyzer database (IMS HEALTH) assembles drug prescriptions, diagnoses, and basic medical and demographic data directly obtained from the practice computer system of general practitioners and diabetologists [9]. Diagnoses (ICD-10), prescriptions (Anatomical Therapeutic Chemical (ATC) Classification System) and the validity of reported data were monitored by IMS based on a number of quality assurance criteria (e.g. completeness of documentation, linkage of diagnoses and prescriptions). Because of the retrospective analysis of anonymized data from primary care practices all over Germany no specific ethical consent was obtained. The analyzed database period was January 2008 to December 2012, and included 918 general or specialist practices throughout Germany. Patients ≥18 years old with type

2 diabetes, who had basal insulin (glargine, GLA; detemir, DET; human insulin NPH) initiated, whichever came first (defined as index), were identified. The practice visit records were used to determine 12-month prior and 24-month post index continuous follow-up, respectively. Then, type of insulin therapy (basal supported oral therapy, BOT; intensified conventional therapy, ICT) was assessed. BOT was defined if patients received basal insulin in combination with ≥1 oral antidiabetic prescription within 183 days prior and post to index and no short acting insulin (ATC: A10C1) during these time periods. ICT was defined if patients received a basal-bolus treatment with ≥1 short acting insulin within 183 days post to index. The selection process of the patients from the database is shown in detail in Fig. 1. Basal insulin treatment persistence was defined as the proportion of patients who remained on the basal insulin treatment over 2 years after index prescription. In addition, the duration from initiation to change of insulin therapy was determined. Thus, discontinuation of basal insulin in BOT was indicated by ≥1 of the following events: prescription of another type of basal insulin, additional short acting insulin, or premixed insulin. In ICT, discontinuation was defined as use of another basal insulin or switch to premixed insulin formulations. Potential predictors of persistence considered in the present analysis were age, sex, diabetes duration, baseline comorbidity, co-medication with oral antidiabetic drugs, diabetologist care, private health insurance and geographical region. Macrovascular complications were determined based on primary care diagnoses (ICD-10 codes) for coronary heart disease (I20, I24, I25), myocardial infarction (I21, I22, I23, I25.2), stroke (I63, I64, G45), peripheral vascular disease (E11.5, E14.5, I73.9) and heart failure (I50). Microvascular complications included retinopathy (E11.3, E14.2), neuropathy (E11.4, E14.4), and nephropathy (E11.2, E14.2, N18, N19). If available, the last HbA1c and the last recorded body mass index (BMI) before index date was also considered in the analysis. Furthermore, lipid disorders, hypertension and related drug treatment were assessed as potential confounders. In addition, the Charlson co-morbidity index was used as general marker of co-morbidity. The Charlson index is a weighted index that accounts for the number and severity of co-morbidities in administrative database studies [10]. The conditions included in the Charlson index cover a wide range of co-morbidities (macrovascular diseases, dementia, pulmonary diseases, gastrointestinal, liver and renal diseases, diabetes, tumors and AIDS). Descriptive statistics are given for the above-mentioned variables. The analyses of persistence were carried out using Kaplan–Meier curves and log-rank tests separately for BOT and ICT. The median time (days) to discontinuation was assessed. Univariate und multivariate Cox regression models were fitted with persistence as dependent variable and the potential predictors. Stepwise regression models were fitted with type of insulin, age, sex, diabetes duration, diabetologist care, and private health insurance keeping as fixed variables (forced entry). Two sided tests were used and a p-value of <0.05 was considered as statistically significant. All analyses were carried out using SAS 9.3. (SAS Institute, Cary, USA). The analysis was carried out following established national [11] and international

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011

PCD-448; No. of Pages 8

ARTICLE IN PRESS p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

3

Fig. 1 – Flowchart of the selection process of type 2 diabetes patients with newly prescribed basal insulins stratified for BOT and ICT (Disease Analyzer Germany).

good practice recommendations of secondary data analysis [12].

3.

Results

3.1.

Characteristics of BOT and ICT users

The clinical characteristics stratified by basal insulin and type of insulin treatment regimen (BOT, ICT) are shown in Tables 1 and 2. In the BOT cohort, patients treated with glargine were slightly older than the other groups (p < 0.001) (Table 1). The proportions with long-standing diabetes (>5 years) were higher in glargine and detemir users compared to NPH (p < 0.001). Patients with NPH insulin were substantially more often treated by diabetologists. There were also

differences in use of antidiabetic agents between the three groups: sulfonylureas were slightly more often prescribed among glargine and detemir users than in NPH, whereas metformin was less frequently used in glargine patients than the other two groups (p < 0.05). Overall, macro- and microvascular comorbidity was comparable among the three insulin groups. Significant differences were only found for prevalence of heart failure (more often in glargine) and neuropathy (most frequent in NPH users). Finally, antihypertensive drugs were slightly more often prescribed in glargine and detemir users compared to NPH patients. In subgroup analyses for patients with available data, HbA1c values before insulin initiation were higher in glargine and detemir than in NPH patients (p < 0.001). In the ICT cohort, only small age differences were found (somewhat lower in detemir patients) (p < 0.001). Similar to BOT, diabetologist care was most often found in NPH users

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011

PCD-448; No. of Pages 8

ARTICLE IN PRESS

4

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

Table 1 – Baseline characteristics of primary care patients with glargine, detemir or NPH insulin therapy (BOT): Disease Analyzer database, Germany. Variables N Age (years) Diabetes duration (>5 years) Males (%) Private health insurance (%) Diabetologist care (%) Region (West Germany) (%) Antidiabetic treatmenta Biguanides (%) Sulfonylureas (%) Other (GLP-1, DPP-4, TZD, AGI)b Last HbA1c (%) before ID Last BMI (kg/m2 ) before ID Hypoglycaemia (≥1 per year) (%) Diagnosed co-morbidityc (%) Coronary heart disease Myocardial infarction Stroke Peripheral vascular disease Heart failure Hypertension Hyperlipidemia Retinopathy Nephropathy Neuropathy Other drug treatmentd Antihypertensive drugs Lipid-lowering drugs Charlson comorbidity score

Glargine

Detemir

NPH

p value

1398 67.7 (11.3) 44.0 54.2 6.1 13.0 73.1

292 66.4 (11.4) 48.6 54.8 6.9 10.6 67.8

874 65.0 (11.1) 35.6 54.9 4.2 26.5 77.8

<0.0001 <0.0001 0.9316 0.1008 <0.0001 0.0015

77.2 67.4 47.8 8.7 (1.6) 30.4 (5.8) 0.8

82.5 65.4 53.8 8.7 (1.7) 32.6 (7.3) 1.4

81.4 58.8 44.3 8.2 (1.4) 30.8 (5.9) 1.4

0.0194 0.0002 0.0163 <0.0001 0.1374 0.3516

26.8 4.7 8.8 12.5 18.8 75.3 51.8 5.4 13.9 14.5

25.7 4.5 7.2 17.8 14.7 80.8 52.7 7.9 11.0 16.8

24.9 4.5 6.8 13.4 14.0 72.4 53.4 5.6 12.0 19.7

0.6032 0.9517 0.1891 0.0545 0.0066 0.0151 0.7432 0.2577 0.2496 0.0056

80.3 44.6 1.9 (1.4)

81.5 48.6 1.9 (1.2)

73.8 45.5 1.9 (1.3)

0.0004 0.4571 0.1469

Data are means (SD) or proportions (%). At least one prescription prior to index date. b GLP-1 agonists, DPP-4 inhibitors, thiazolidinediones (glitazones), alpha-glucose inhibitors. c Primary care diagnoses prior to index date. d At least one prescription within 365 days prior to index date. ID: index date of incident basal insulin prescription. HbA1c: n = 1447. BMI: n = 472. a

(p < 0.001). In addition, biguanides were more frequently prescribed in the NPH group (p < 0.001). Glargine users had the highest baseline HbA1c values. In line with the BOT results, only small differences were found for comorbidity: hypertension was more frequently diagnosed in NPH users (p = 0.002).

3.2.

patients who switched within the basal insulin class, 59% of detemir patients and 77% of NPH patients switched to glargine. Finally, 60% of glargine patients switched to detemir (40% to NPH).

Persistence of patients under basal insulins

Overall, the median time to discontinuation was about one year after basal insulin initiation. In general, patients stayed longer on their index treatment in the ICT cohort (median days, IQR): 421 (252–574) days, 361 (185–560) days, 483 (288–683) days in glargine, detemir, and NPH users (GLA vs. NPH: p = 0.971, detemir vs. NPH: p = 0.149), respectively, than in the BOT cohort (glargine: 371 (203–524) days; detemir: 323 (196–447) days; NPH: 334 (195–542) days; GLA vs. NPH: p = 0.215, detemir vs. NPH: p = 0.426). In the BOT cohort, 65%, 53%, and 59% of glargine, detemir, and NPH patients were still staying on their index basal insulin after 2 years, respectively (log rank test p < 0.001; Fig. 2). Overall, 63% of the non-persistent BOT patients were switched to ICT, and 12–15% changed to premixed insulin formulations. Of

Fig. 2 – Kaplan–Meier curves for the two-year basal insulin treatment persistence in the BOT cohort.

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011

PCD-448; No. of Pages 8

ARTICLE IN PRESS 5

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

Table 2 – Baseline characteristics of primary care patients with glargine, detemir or NPH insulin therapy (ICT): Disease Analyzer database, Germany. Variables

Glargine

Detemir

NPH

N Age (years) Diabetes duration (>5 years) Males (%) Private health insurance (%) Diabetologist care (%) Region (West Germany) (%) Antidiabetic treatmenta Biguanides (%) Sulfonylureas (%) Other (GLP-1, DPP-4, TZD, AGI)b Last HbA1c (%) before ID Last BMI (kg/m2 ) before ID Hypoglycemia (≥1 per year) (%) Diagnosed co-morbidityc (%) Coronary heart disease Myocardial infarction Stroke Peripheral vascular disease Heart failure Hypertension Hyperlipidemia Retinopathy Nephropathy Neuropathy Other drug treatmentd Antihypertensive drugs Lipid-lowering drugs Charlson comorbidity score

866 63.8 (12.8) 32.3 57.4 5.9 31.5 68.1

512 60.4 (12.9) 26.6 53.7 6.5 34.8 69.0

1794 63.9 (11.9) 30.2 53.8 3.6 39.2 69.6

<0.0001 0.0787 0.1895 0.0044 0.0004 0.7526

42.3 25.1 20.9 8.5 (1.9) 31.2 (6.3) 2.5

44.0 21.3 22.5 8.3 (1.6) 31.5 (6.7) 2.2

51.2 26.6 21.1 8.1 (1.5) 32.1 (6.9) 1.8

<0.0001 0.0508 0.7609 0.0009 0.4560 0.4911

25.3 4.3 7.5 14.8 14.0 65.8 45.4 10.5 17.1 20.9

24.4 4.5 4.9 15.6 11.3 61.7 44.3 12.9 16.2 23.2

27.4 3.9 7.3 15.1 13.4 69.5 45.7 10.1 17.5 23.9

0.2935 0.7576 0.1268 0.9137 0.3562 0.0024 0.8700 0.2033 0.7889 0.2212

65.6 37.6 2.0 (1.2)

57.8 40.4 2.0 (1.3)

64.9 38.9 2.1 (1.3)

0.0066 0.5862 0.3549

p value

Data are means (SD) or proportions (%). At least one prescription prior to index date. b GLP-1 agonists, DPP-4 inhibitors, thiazolidinediones (glitazones), alpha-glucose inhibitors. c Primary care diagnoses prior to index date. d At least one prescription within 365 days prior to index date. ID: index date of incident basal insulin prescription. HbA1c: n = 1489. BMI: n = 488. a

In the ICT cohort, there were no differences in the 2-year persistence between the three basal insulins (84–86%) (Fig. 3). Forty-four (44) to 49% switched within basal insulin class, 9–13% switched to premix, and 35–39% stayed on ICT. Most

Fig. 3 – Kaplan–Meier curves for the two-year basal insulin treatment persistence in the ICT cohort.

of the patients kept their prandial insulin, although they had their basal insulin treatment changed.

3.3.

Predictors of discontinuation

The results of the multivariate regression analyses for the BOT cohort are shown in Table 3. Glargine patients were more likely to stay on their basal insulin treatment compared to detemir or NPH insulin users (p < 0.001). Patients were also more likely to be persistent to their index basal insulin if they were treated with metformin or sulfonylureas (Table 3). Higher age was another independent predictor of persistence. Among comorbidity, patients were less likely to be persistent to their index basal insulin if they had heart failure diagnosed at baseline. In a subgroup analysis including patients with documented HbA1c values, glycemic control (per 1 HbA1c % increase) was significantly related to a higher likelihood for discontinuation (Hazard Ratio; 95% CI: 1.07; 1.02–1.12; p = 0.004). The results of the Cox regression analysis for the ICT cohort are shown in Table 4. There were no significant differences between detemir and glargine, or NPH and glargine users for the risk of discontinuation. Patients with diabetologist care were less likely to discontinue basal insulin

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011

PCD-448; No. of Pages 8

ARTICLE IN PRESS

6

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

Table 3 – Persistence of type 2 diabetes patients on BOT using different long-acting insulins: Cox regression analyses. Outcome variables Detemir vs. glargine NPH vs. glargine Age (years) Males Diabetes duration >5 years Diabetologist care Private health insurance Metformin Sulfonylurea Heart failure

Hazard Ratioa

95% CI

1.56 1.22 0.99 0.94 0.97 1.09 0.93 0.61 0.86 1.39

1.31–1.87 1.07–1.38 0.98–0.99 0.84–1.06 0.86–1.09 0.93–1.27 0.72–1.20 0.54–0.69 0.77–0.97 1.16–1.67

p value <0.0001 0.0029 <0.0001 0.3089 0.6099 0.2895 0.5747 <0.0001 0.0106 0.0004

Table 4 – Persistence of type 2 diabetes patients on ICT using different long-acting insulins: Cox regression analyses. Outcome variables Detemir vs. glargine NPH vs. glargine Age (years) Males Diabetes duration Diabetologist care Private health insurance Other OADa a

Hazard Ratioa 1.02 1.11 0.995 1.00 0.99 0.86 0.93 1.28

0.81–1.28 0.94–1.31 0.989–1.001 0.87–1.15 0.84–1.16 0.74–1.00 0.66–1.30 1.06–1.55

p value 0.8725 0.2303 0.0867 0.9803 0.8596 0.0526 0.6581 0.0111

GLP-1 agonists, DPP-4 inhibitors, thiazolidinediones (glitazones), alpha-glucose inhibitors.

treatment (p = 0.053). ICT patients using GLP-1 agonists or DPP-4 inhibitors or other antidiabetics (thiazolidinediones, alpha-glucose inhibitors) were less persistent than nonusers. In subgroup analyses, no significant associations with HbA1c or BMI were found (data not shown).

4.

95% CI

Discussion

This real-world study shows that treatment persistence among type 2 diabetes patients initiating basal insulin is influenced by type of insulin, antidiabetic co-medication, and some disease-specific patient characteristics. In line with previous observational studies, prescription of insulin glargine was associated with a higher two-year persistence than use of insulin detemir or NPH. Wei et al. pooled data from three observational retrospective studies to evaluate real-world treatment persistence among patients with type 2 diabetes mellitus initiating insulin glargine or insulin detemir [13]. The average insulin persistence over 1-year follow-up was 65%, and insulin glargine was associated with significantly higher persistence [13]. Davis et al. retrospectively analyzed data from the US General Electric Centricity electronic medical records database from patients with type 2 diabetes initiating insulin glargine or detemir [14]. Patients initiating insulin glargine were more persistent compared with detemir (80% vs. 68%, p < 0.0001) during ≥12 months follow-up [14]. A retrospective cohort study from an analysis of claims data from prescriptions for ambulatory patients within the German Statutory Health Insurance scheme also showed that type 2 diabetes patients under BOT with insulin glargine stayed significantly longer on the initial therapy before switching to ICT than patients on BOT using NPH insulin [15]. A Markov model was developed simulating the transition from BOT to ICT during a treatment course of 10 years based on German data

(Disease Analyzer, Statutory Health Insurance) [16]. After 2 years, 53% of glargine-treated patients and 31% of NPH-treated patients continued the BOT. After 6.5 years, all NPH-treated patients had switched to ICT, whereas complete transition to ICT of glargine-treated patients occurred 1.75 years later [15]. Unfortunately, data on reasons for switching were not available in these studies. Recently, a cohort study based on data from the Swedish National Diabetes Register, including 5077 patients indicated that insulin glargine, insulin detemir, NPH and premixed insulins were equally effective in lowering HbA1c in insulin-naïve patients with type 2 diabetes in routine clinical care [17]. Thus, further studies are necessary to investigate the underlying reasons for the observed differences in persistence. It is conceivable that basal insulin treatment with comparable effectiveness but less hypoglycemia [4] and only a once daily injection such as glargine [18] can improve treatment persistence. A novel finding in the present study was that diagnosed heart failure was independently associated with a 40% increased risk of discontinuing basal insulin therapy in BOT. Congestive heart failure may cause myocardial insulin resistance [19]. Recently, it has been shown that restoration of normal cardiac output after left ventricular assist devices implantation improves diabetic control in patients with advanced heart failure, which could not be ascribed to weight loss due to increased physical activity [20]. The underlying reasons are unclear. Heart failure is associated with unfavorable impact on neurohormonal activity and with an elevation of pro-oxidant and pro-inflammatory circulating cytokines, which may decrease insulin sensitivity [20]. The observed relation of metformin with higher basal insulin persistence in the present study is most likely due to its well-known effects on reducing hepatic glucose production [2]. In addition, sulfonylurea prescriptions were also associated with higher basal insulin persistence, which is controversial to

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011

PCD-448; No. of Pages 8

ARTICLE IN PRESS p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

recent findings [21,22]. Finally, older age was associated with better basal insulin persistence in the present study, which is confirming previous studies [13]. However, the other explanation for this finding could be the known high correlation between the older patient’s age and older physician’s age, whereby older physicians are not so keen to change a therapy. In patients with ICT, two-year persistence was somewhat higher than in BOT. There were no significant differences in the proportions of patients remaining on the initial basal insulin after 2 years of ICT. Most likely this reflects that ICT is a final stage of complex insulin treatment in most type 2 diabetes patients, with only few possibilities to improve metabolic control such as dose optimization and/or lifestyle modifications. Several limitations of the present study should be mentioned. First, no valid information on diabetes type, prescribed daily doses, and important outcome measures (e.g. hypoglycaemia) were available in the database. Furthermore, no valid information on diabetes duration was provided. Unfortunately, socioeconomic factors, educational level or smoking habits are not covered in the database. Also assessment of comorbidity relied on ICD codes by primary care physicians only and no information about disease severity (for example, NYHA classification for heart failure) was available. Second, HbA1c and body mass index values were only available for a subgroup at baseline but not during the course of insulin treatment. Finally, the low prevalence for microvascular complications observed in the present study needs to be further evaluated. The information of the insulin used is based on receipt to pharmacies and bias is possible because patient intake cannot be validated. Finally, patients in this study are localized in Germany and the results may not be generalizable to all countries and regions. In conclusion, treatment persistence among type 2 diabetes patients initiating basal insulin is influenced by a variety of factors including type of insulin, antidiabetic comedication, and patient characteristics (age, comorbidity: heart failure). Further studies should be performed to explore the underlying reasons for these associations in order to improve insulin initiation in type 2 diabetes patients.

Transparency The information contained in this manuscript was presented at the 74th Scientific Sessions of the American Diabetes Association, San Francisco, California, June 13–17, 2014 (poster 1546) and at the 49th Annual Conference of the German Diabetes Association, Berlin, Germany, May 28–31, 2014 (poster 282).

Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: the study was supported by an unrestricted grant from Sanofi-Aventis, Berlin, Germany.

7

Conflict of interest W.R. received an honorary from IMS HEALTH for consulting on this article.

Authors’ contributions All authors were involved in the conception and design of the research. FWD, KK, and WR developed the analysis plan; KK analyzed and all authors interpreted the results. WR wrote the first draft of the manuscript; all authors drafted the final version of the manuscript. All authors had full access to all of the data in the study.

Acknowledgment The authors thank Hsing-Wen Chung for statistical and data validation support.

references

[1] UK Prospective Diabetes Study (UKPDS) Group, Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33), Lancet 352 (1998) 837–853. [2] S.E. Inzucchi, R.M. Bergenstal, J.B. Buse, M. Diamant, E. Ferrannini, M. Nauck, A.L. Peters, A. Tsapas, R. Wender, D.R. Matthews, Management of hyperglycaemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD), Diabetologia 55 (2012) 1577–1596. [3] S.G. Swinnen, A.C. Simon, F. Holleman, J.B. Hoekstra, J.H. Devries, Insulin detemir versus insulin glargine for type 2 diabetes mellitus, Cochrane Database Syst. Rev. 7 (2011) CD006383. [4] Institute for Quality and Efficiency in Health Care, Long-acting insulin analogues in the treatment of type 2 diabetes. Executive summary of final report A05-03. Version 1.1, 2009 www.iqwig.de (accessed 11.06.14). [5] J.A. Cramer, A systematic review of adherence with medications for diabetes, Diabetes Care 27 (2004) 1218–1224. [6] J.A. Cramer, A. Roy, A. Burrell, et al., Medication compliance and persistence: terminology and definitions, Value Health 11 (2008) 44–47. [7] W. Wei, S. Zhou, R. Miao, C. Pan, L. Xie, O. Baser, J. Gill, Much ado about nothing? A real-world study of patients with type 2 diabetes switching basal insulin analogs, Adv. Ther. 31 (2014) 539–560. [8] B.J. Borah, T. Darkow, J. Bouchard, M. Aagren, F. Forma, B. Alemayehu, A comparison of insulin use, glycemic control, and health care costs with insulin detemir and insulin glargine in insulin-naive patients with type 2 diabetes, Clin. Ther. 31 (2009) 623–631. [9] H. Becher, K. Kostev, D. Schröder-Bernhardi, Validity and representativeness of the “Disease Analyzer” patient database for use in pharmacoepidemiological and pharmacoeconomic studies, Int. J. Clin. Pharmacol. Ther. 47 (2009) 617–626.

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011

PCD-448; No. of Pages 8

ARTICLE IN PRESS

8

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 5 ) xxx–xxx

[10] H. Quan, V. Sundararajan, P. Halfon, et al., Coding algorithms for defining co-morbidities in ICD-9-CM and ICD-10 administrative data, Med. Care 43 (2005) 1130–1139. [11] E. Swart, Good practice of secondary data analysis, first revision, Gesundheitswesen 70 (2008) 54–60. [12] B. Motheral, J. Brooks, M.A. Clark, et al., A checklist for retrospective database studies – report from the ISPOR task force on retrospective databases, Value Health 6 (2003) 90–97. [13] W. Wei, C. Pan, L. Xie, O. Baser, Real-world insulin treatment persistence among patients with type 2 diabetes, Endocr. Pract. 20 (2014) 52–61. [14] K.L. Davis, M. Tangirala, J.L. Meyers, W. Wei, Real-world comparative outcomes of US type 2 diabetes patients initiating analog basal insulin therapy, Curr. Med. Res. Opin. 29 (2013) 1083–1091. [15] R. Quinzler, M. Ude, A. Franzmann, S. Feldt, K. Schüssel, K. Leuner, W.E. Müller, F.W. Dippel, M. Schulz, Treatment duration (persistence) of basal insulin supported oral therapy (BOT) in type-2 diabetic patients: comparison of insulin glargine with NPH insulin, Int. J. Clin. Pharmacol. Ther. 50 (2012) 24–32. [16] M. Pfohl, F.W. Dippel, K. Kostev, S. Fuchs, W. Kotowa, Different persistence on initial basal supported oral therapy in type 2 diabetics is associated with unequal distributions of insulin treatment regimens under real-life conditions in Germany, Int. J. Clin. Pharmacol. Ther. 48 (2010) 761–766.

[17] B. Eliasson, N. Ekström, S. Bruce Wirta, A. Odén, M.P. Fard, A.M. Svensson, Metabolic effects of basal or premixed insulin treatment in 5077 insulin-naïve type 2 diabetes patients: registry-based observational study in clinical practice, Diabetes Ther. (2014), May 15 (Epub ahead of print). [18] W. Rathmann, F.W. Dippel, K. Kostev, Different injection frequencies of basal insulins in type 2 diabetes patients under real-life conditions: a retrospective database analysis, J. Diabetes Sci. Technol. 7 (2013) 1354–1358. [19] L.A. Nikolaidis, A. Sturzu, C. Stolarski, D. Elahi, Y.T. Shen, R.P. Shannon, The development of myocardial insulin resistance in conscious dogs with advanced dilated cardiomyopathy, Cardiovasc. Res. 61 (2004) 297–306. [20] N. Uriel, Y. Naka, P.C. Colombo, M. Farr, S.W. Pak, V. Cotarlan, J.B. Albu, D. Gallagher, D. Mancini, H.N. Ginsberg, U.P. Jorde, Improved diabetic control in advanced heart failure patients treated with left ventricular assist devices, Eur. J. Heart Fail. 13 (2011) 195–199. [21] S.E. Holden, C.J. Currie, Mortality risk with sulphonylureas compared to metformin, Diabetes Obes. Metab. (2014), http://dx.doi.org/10.1111/dom.12280. [22] O.J. Phung, E. Schwartzman, R.W. Allen, S.S. Engel, S.N. Rajpathak, Sulphonylureas and risk of cardiovascular disease: systematic review and meta-analysis, Diabet. Med. 30 (2013) 1160–1171.

Please cite this article in press as: S. Pscherer, et al., Treatment persistence after initiating basal insulin in type 2 diabetes patients: A primary care database analysis, Prim. Care Diab. (2015), http://dx.doi.org/10.1016/j.pcd.2015.01.011