Healthcare Costs of the Combination of Metformin/Dipeptidyl Peptidase-4 Inhibitors Compared with Metformin/Other Oral Antidiabetes Agents in Patients with Type 2 Diabetes and Metabolic Syndrome

Abstract

Objectives:

This study assessed the health costs resulting from the combination of metformin/dipeptidyl peptidase-4 (DPP-4) inhibitors compared with metformin/oral antidiabetes drugs in patients with type 2 diabetes and metabolic syndrome (MS).

Patients and Methods:

An observational retrospective study was performed. Patients ≥30 years of age who were receiving treatment with metformin who started a second oral antidiabetes therapy in 2008 and 2009 were included. Patients were divided into two groups: (a) metformin plus DPP-4 inhibitors and (b) metformin plus other oral antidiabetes drugs. The main measures were compliance, persistence, metabolic control (glycosylated hemoglobin level of <7%), and complications (hypoglycemia and cardiovascular events). Healthcare and non-healthcare costs were calculated. Patients were followed up for 2 years. An analysis of covariance was carried out (P<0.05 was considered significant).

Results:

Of the 1,435 patients (mean age, 67.3 years; 53.1% male) who were enrolled, 442 (30.8%) were receiving metformin plus DPP-4 inhibitors, and 993 (69.2%) were receiving metformin plus other oral antidiabetes drugs. The prevalence of MS was 72.2% (95% confidence interval, 71.1–73.3%). Patients treated with DPP-4 inhibitors had better compliance (69.1% vs. 63.8%), persistence (63.8% vs. 53.1%), and metabolic control (69.9% vs. 64.3%) (P<0.01) compared with those receiving other antidiabetes drugs, lower rates of hypoglycemia (14.3% vs. 41.1%) and cardiovascular events (2.9% vs. 5.7%) (P<0.01), and a lower mean adjusted unit cost (€2,278 vs. €2,631; P=0.003).

Conclusions:

Despite the limitations of this observational study, diabetes patients with MS who were treated with metformin plus DPP-4 inhibitors had better compliance, greater metabolic control, and lower rates of hypoglycemia, causing lower costs for the Spanish national health system than patients receiving metformin plus other antidiabetes drugs.

Introduction

Cardiovascular disease and the complications due to atherosclerosis are the major cause of morbidity and mortality in patients with type 2 diabetes mellitus (T2DM).^1,2 Diabetes should not be regarded as a single disease entity but, like other forms of glucose intolerance, should be considered as part of a wider, underlying health problem: metabolic syndrome (MS).³

MS, or insulin resistance syndrome, consists of a combination of impaired glucose metabolism, abdominal obesity, dyslipidemia, and hypertension.⁴ The prevalence of MS in patients with T2DM depends on the definition of the constituent components but, according to the National Cholesterol Education Program Adult Treatment Panel III diagnostic criteria, generally ranges from 65% to 80%.^5,6 The importance of MS lies in the fact that its detection can help identify individuals at increased cardiovascular risk, as MS triples the risk of cardiovascular disease.^3,6

Metformin is recommended as the first therapeutic choice in diabetes, together with dietary measures, and when metabolic control is not achieved, the addition of a second drug in combination with metformin is recommended.⁷ Dipeptidyl peptidase-4 (DPP-4) inhibitors, which are used in combination therapy, have the advantage over traditional secretagogues in that they significantly reduce hypoglycemia because their insulin secretion-stimulating mechanism is glucose-dependent.^8,9 Studies have demonstrated the efficacy and safety of DPP-4 inhibitors in high-risk patients with MS.^10,11 However, the available evidence on the economic effects of these treatments in the context of routine clinical practice and a population-based context is limited, and therefore the results of this Spanish study may be of interest. The aim of this study was to describe the health costs resulting from the combination of metformin plus DPP-4 inhibitors compared with metformin plus other oral antidiabetes drugs (includes sulfonylureas and glitazones) in patients with T2DM and MS over a 2-year period.

Patients and Methods

Design and study population

We performed a retrospective observational, multicenter study through review of computerized medical records of outpatients and inpatients treated with metformin. The study population consisted of patients assigned to six primary care centers and two reference hospitals (Hospital Municipal de Badalona and Hospital Universitari Germans Trias i Pujol of Badalona, Barcelona, Spain). The population assigned to these centers is mostly urban, with middle–low socioeconomic status and predominantly industrial occupations.

Inclusion and exclusion criteria

The study included all patients who started a second antidiabetes treatment in 2008 and 2009 and fulfilled the following conditions: (a) age ≥30 years; (b) diagnosis of T2DM at least 12 months before the study date; (c) patients who regularly followed the cardiovascular risk protocol/guidelines of the participating centers; (d) patients currently treated with metformin as the first therapeutic option (monotherapy); and (e) patients in whom the follow-up could be guaranteed. Patients transferring out to other health areas were excluded. There were two study groups: (a) patients treated with metformin plus DPP-4 inhibitors and (b) patients treated with metformin plus other oral antidiabetes drugs. Patients were followed up for 24 months, a sufficient period to observe complications and costs.

Diagnosis of T2DM and MS

The diagnosis of T2DM was obtained from the International Classification of Primary Care (ICPC-2; code T90)¹² and the International Classification of Diseases, 9^th Revision, Clinical Modification (ICD-9-CM, code 250). Baseline data were obtained on the following microvascular complications: (a) retinopathy, (b) neuropathy, (c) nephropathy, and (d) microalbuminuria.

MS was defined as fulfillment of three of the following five criteria, modified from the Adult Treatment Panel III report¹³: (a) triglyceride level of ≥150 mg/dL; (b) high-density lipoprotein cholesterol level of <40 mg/dL in males or <50 mg/dL in females; (c) systolic/diastolic blood pressure of ≥130/85 mm Hg or antihypertensive treatment; (d) baseline fasting plasma glucose level of ≥110 mg/dL, hypoglycemic treatment, or previously diagnosed diabetes; and (e) body mass index. We replaced one of the usual MS components, the measurement of waist circumference, by the determination of body mass index, which is one of the modified National Cholesterol Education Program Adult Treatment Panel III criteria. A body mass index of ≥28.8 kg/m² was considered equivalent to the criterion of abdominal adiposity (waist circumference of >102 cm in males and >88 cm in females), the original criterion followed by other authors.¹⁴

Sociodemographic and comorbidity variables

The variables collected were age, sex, time since diagnosis of T2DM and MS, and the personal history (Table 1). The general comorbidity summary variables used for each patient treated were as follows: (a) the Charlson Comorbidity Index,¹⁵ which is used as a proxy of the severity of the health status; and (b) the individual causality index, obtained from the Adjusted Clinical Groups, which is a patient classification system based on iso-resource use.¹⁶ The Adjusted Clinical Groups application provides resource utilization bands with each patient, according to general morbidity, placed in one of five mutually exclusive categories.

Table 1.

Baseline Characteristics of the Study Patients

	Study group
	Metformin+OOAD (n=993)	Metformin+DPP-4 inhibitors (n=442)	Total (n=1,435)	P
Sociodemographic characteristics
Mean age (years)	67.5 (10.4)	67.0 (10.2)	67.3 (10.4)	0.339
Sex (male)	53.2%	52.9%	53.1%	0.935
General comorbidity
Mean diagnoses	5.6 (2.5)	5.3 (2.3)	5.5 (2.4)	0.082
Mean Charlson Comorbidity Index	1 (0.4)	1 (0.5)	1 (0.5)	0.490
Mean RUB	3 (0.6)	2.9 (0.5)	3 (0.6)	0.131
RUB-1	1.1%	1.4%	1.2%
RUB-2	13.4%	13.6%	13.5%
RUB-3	73.2%	78.0%	74.6%
RUB-4	10.7%	6.5%	9.4%
RUB-5	1.6%	0.5%	1.2%	0.177
Associated comorbidities
Hypertension	77.9%	74.7%	76.9%	0.173
Dyslipidemia	72.3%	71.5%	72.1%	0.751
Obesity	29.3%	30.1%	29.5%	0.763
Active smokers	21.1%	24.2%	22.1%	0.197
Alcoholism	4.2%	2.7%	3.8%	0.164
Ischemic heart disease	14.2%	13.3%	13.9%	0.667
Cerebrovascular accident	15.8%	13.2%	14.4%	0.526
Organ failure	18.7%	16.1%	17.9%	0.224
Bronchial asthma	5.4%	4.5%	5.2%	0.470
COPD	6.1%	6.8%	6.3%	0.644
Neuropathies	2.1%	1.5%	1.9%	0.226
Dementias (all types)	4.9%	4.0%	4.6%	0.335
Depressive syndrome	22.3%	25.1%	23.5%	0.313
Malignancies	10.0%	8.6%	9.5%	0.414
Relationship with diabetes
Time since diagnosis (years)	11.4 (11.1)	12.1 (4.5)	11.6 (9.6)	0.294
Diabetic retinopathy	25.1%	28.2%	27.1%	0.197
Diabetic neuropathy	8.5%	8.6%	8.5%	0.772
Diabetic nephropathy	7.0%	7.7%	7.2%	0.751
Microalbuminuria	22.6%	21.4%	22.3%	0.584

Data are expressed as percentages or mean (SD) values as indicated.

COPD, chronic obstructive pulmonary disease; DPP-4, dipeptidyl peptidase-4; OOAD, other oral antidiabetes agents (including sulfonylureas and glitazones); RUB, resource utilization bands.

Treatment compliance and persistence and metabolic control

Information was collected on the following oral antidiabetes drugs according to the Anatomical Therapeutic Chemical Classification System¹⁷: (a) metformin (A10BA*), (b) insulin release stimulators [sulfonylureas (A10BB*) and glinides (A10BX*)], (c) glitazones (A10BG*), and (d) DPP-4 inhibitors (A10BH*) in monotherapy or combination. We did not include patients receiving α-glucosidase inhibitors because of the insufficient sample size. Compliance during the study period was calculated by dividing the total number of tablets dispensed by those recommended or prescribed. Treatment persistence was defined as the time, measured in months, without abandoning the initial treatment or with no change to another medication for at least 30 days after the initial prescription. Metabolic control was established as levels of glycosylated hemoglobin that were <7%.⁶

Macrovascular complications and cardiovascular events

The variables collected were as follows: (a) heart disease, including cardiac ischemia, acute myocardial infarction, and heart failure, as defined by the diagnostic criteria of the World Health Organization; (b) cerebrovascular disease, including stroke (ischemic or hemorrhagic; according to the American Heart Association) and transient ischemic attack; and (c) peripheral arterial disease (all types). The cumulative incidence rate was defined as the proportion of healthy individuals who developed the complication (number of new cases during the study period). All cases of symptomatic hypoglycemia during the follow-up were also identified by medical judgment during the regular follow-up visits.

Resource use and costs of health model

Direct healthcare costs were defined as costs related to medical care (medical visits, days of hospitalization, emergency department visits, diagnostic tests and therapies, etc.). Indirect costs were defined as the days of work productivity lost due to disability. Costs were expressed as the mean cost per patient during the 2-year follow-up (unit cost). The unit costs used in the study are expressed in euros (2011 values). The tariffs were obtained from the cost accounting of each center¹⁸ except for medications (retail price).

Confidentiality of information and statistical analysis

The study was classified by the Spanish Agency for Medicines and Health Products (Non-Post Licensing Study) and was subsequently approved by the Ethics Committee for Clinical Research of the Hospital Universitari Germans Trias i Pujol, Badalona. A descriptive analysis was performed, and the 95% confidence intervals (CI) were calculated. The normality of the distribution of quantitative variables was verified using the Kolmogorov–Smirnov test. In the bivariate analysis, analysis of variance, the χ² test, Pearson's linear correlation, and the Mann–Whitney–Wilcoxon nonparametric test were used. A logistic regression analysis was performed to determine the variables associated with DPP-4 inhibitors (presence/absence) using an enter procedure (Wald statistic). A multiple linear regression model was constructed to assess the variables associated with health costs. The comparison of costs was made according to the recommendations of Thompson and Barber¹⁹ using analysis of covariance, with sex, age, resource utilization bands, the Charlson Comorbidity Index, and the time of evolution of the diagnosis as covariates.

Results

Of an initial population of 62,370 patients ≥30 years of age assigned to the centers, 48,295 sought care. Of these, 6,620 were diagnosed with T2DM (prevalence, 10.6%; 95% CI, 10.4–10.8%). The prevalence of MS (n=6,620) in patients with T2DM was 72.2% (95% CI, 71.1–73.3%).

In total, 5,185 patients were excluded from the study: 1,840 had no criteria for MS, and 3,345 were removed for other reasons (988 were not receiving drug treatment, 564 were receiving other drug therapies [insulin, n=372], 233 discontinued treatment, 887 changed therapy during the follow-up, 385 were considered lost to the study, and 288 for reasons not known). The percentage distribution of patients excluded according to the two study groups was similar. The components of MS (n=6,620) were as follows: body mass index of >28.8 kg/m², 63.4%; blood pressure of >130/85 mm Hg (or treatment), 75.9%; triglyceride level of >150 mg/dL, 55.6%; fasting blood glucose level of >110 mg/dL, 78.9%; and high-density lipoprotein level of <40 mg/dL (males) or <50 mg/dL (females), 66.8%.

In total, 1,435 patients (T2DM and MS) treated with oral antidiabetes agents in combination who had MS were finally included in the study: 30.8% (n=442) were receiving DPP-4 inhibitors (vildagliptin, 65.5%), and 69.2% (n=993) were receiving other oral antidiabetes drugs (85.5% with sulfonylureas and 14.5% with glitazones). The baseline characteristics of the patients included according to study group are shown in Table 1. The mean (SD) age was 67.3 (10.4) years, and 53.1% were male. Patients receiving DPP-4 inhibitors had similar overall comorbidity (5.3 vs. 5.6 diagnoses; P=0.082) as those receiving other oral antidiabetes drugs. There were no significant between-group differences in baseline characteristics.

Patients receiving DPP-4 inhibitors had better compliance (69.1% vs. 63.8%; P=0.038) and treatment persistence (63.8% vs. 53.1%, P<0.001). There was an acceptable correlation between the degree of compliance and the time of treatment persistence (r=0.417, P<0.001). Metabolic control (glycosylated hemoglobin level of <7%) of T2DM was 62.8% versus 60.0%, respectively (P=0.360), at baseline and 69.9% versus 64.3%, respectively (P=0.032), at 24 months. In the logistic model, treatment with DPP-4 inhibitors was associated with compliance (odds ratio=1.4; 95% CI, 1.2–1.7; P=0.011), treatment persistence (odds ratio=1.1; 95% CI, 1.0–1.3; P=0.045), and metabolic control (odds ratio=1.2; 95%, CI, 1.0–1.5; P=0.033).

Patients receiving DPP-4 inhibitors had fewer primary care (21.5 vs. 29.7; P<0.001) and specialist care (1.8 vs. 2.2; P=0.026) medical visits compared with those receiving other oral antidiabetes drugs. The pattern of gross and adjusted costs (covariates) during the 2-year follow-up according to the study groups is shown in Table 2. The total cost of care of the patients included was €3.6 million. The mean unit health costs were lower in patients receiving DPP-4 inhibitors (€2,252 vs. €2,621; P<0.001). In the analysis of covariance model adjusted for covariates, these figures were €2,278 (95% CI, €2,074–€2,472) versus €2,631 (95% CI, €2,499–€2,762) (P=0.003). These differences were apparent for all components of the health costs analyzed except for lost productivity.

Table 2.

Model of Gross Costs and Adjusted Model According to the Study Groups During the 2-Year Follow-Up

	Study group
	Metformin+OOAD	Metformin+DPP-4 inhibitors	Total	P
Unadjusted cost model
Healthcare costs	2,495 (1,411)	2,242 (1,055)	2,417 (1,316)	0.001
Primary care costs	2,210 (1,217)	2,020 (972)	2,151 (1,150)	0.004
Medical visits	689 (349)	498 (325)	630 (353)	<0.001
Laboratory tests	70 (40)	50 (33)	64 (39)	<0.001
Conventional radiology	20 (23)	20 (23)	20 (23)	0.862
Complementary tests	19 (29)	17 (25)	18 (28)	0.092
Drugs	1,412 (1030)	1,435 (823)	1,419 (970)	0.675
Metformin	598 (332)	430 (425)	522 (477)	0.031
OOAD vs. DPP-4	665 (439)	952 (558)	787 (611)	0.012
Other drugs	149 (112)	53 (45)	110 (101)	0.257
Specialized care costs	285 (464)	223 (270)	266 (415)	0.009
Days of hospitalization	21 (264)	4 (43)	16 (221)	0.180
Medical visits	233 (359)	191 (253)	220 (331)	0.026
Emergency	30 (62)	27 (63)	29 (62)	0.451
Non-health costs	127 (1,911)	10 (0)	88 (1,591)	0.164
Total costs	2,621 (2,383)	2,252 (1,055)	2,504 (2,074)	<0.001
Adjusted cost model^a
Healthcare costs	2,503	2,276	−228	0.002
95% CI	2,423–2,583	2,154–2,396
Non-health costs	128	2	−126	0.177
95% CI	24–230	0–154
Total costs	2,631	2,278	−353	0.003
95% CI	2,499–2,762	2,074–2,472

Data are expressed as mean (SD) values in euros.

In the analysis of covariance model, the contrasts are based on linearly independent pairwise comparisons of estimated marginal means.

CI, confidence interval; DPP-4, dipeptidyl peptidase-4; OOAD, other oral antidiabetes agents (includes sulfonylureas and glitazones).

During follow-up, the rate of new cardiovascular events was 4.9% (95% CI, 3.8–6.0%) and was slightly lower in patients receiving DPP-4 inhibitors (2.9% vs. 5.7%; P=0.011). The percentage of patients with hypoglycemia was 32.8%, which was lower in patients receiving DPP-4 inhibitors (14.3% vs. 41.1%; P<0.001). Healthcare costs (adjusted for covariates) were associated with age (β=0.124), poor metabolic control (β=0.103), and hypoglycemia (β=0.091).

Discussion

Our results show that diabetes patients with MS treated with metformin and DPP-4 inhibitors had higher levels of compliance and disease control than patients receiving other oral antidiabetes agents, a difference that was related to lower rates of hypoglycemia and healthcare costs during the disease course.

The prevalence of MS in patients with T2DM was 72.2%. Our results are consistent with the literature reviewed.^5,6 Consensus guidelines recommend that these patients should be considered as at very high cardiovascular risk. In diabetes, the intensive treatment of dyslipidemia reduces cardiovascular deaths by 17–50%, overall mortality by 12–40%, coronary events by 24–40%, and stroke by 27–40%, and therefore strategies aimed at improving prevention, control, and individualized treatment should be prioritized.²⁰

Our results show that patients receiving DPP-4 inhibitors had better treatment compliance and persistence and metabolic control and a lower rate of hypoglycemia than those receiving other oral antidiabetes drugs. There are few studies on treatment compliance and persistence with oral agents in general, and the differing methodologies used to measure these factors mean that existing studies are difficult to compare. Studies show a rate of compliance ranging from 40% to 80%.²¹ A recent study by Rathmann et al.²² of primary care patients found that patients receiving DPP-4 inhibitors had less interruption of therapy and a lower rate of hypoglycemia and incidence of macrovascular events compared with patients receiving sulfonylureas. The results of various studies suggest that DPP-4 inhibitors seem to have a better tolerance and safety profile and a possible cardioprotective effect, which would favor lower rates of hypoglycemia and tend to reduce cardiovascular events.^9,23,24 These results are consistent with ours, although long-term studies comparing the use of combination antidiabetes therapy are required to strengthen the consistency of our results. It seems clear that the role of DPP-4 inhibitors in the T2DM therapeutic arsenal is evolving rapidly.²⁵

The cost analysis was a strength of our study. Higher costs were associated with poor metabolic control and hypoglycemia. The few published studies on this subject show that the greater the degree of compliance and metabolic control in these patients, the lower the risk of hospitalization.^26,27 However, there was a wide range of variability in these studies, making comparison of the results difficult. Overall, our results are consistent with these studies and highlight the association between hypoglycemia and health costs.²⁷ Giorda et al.²⁸ underlined the fact that effective therapy with good metabolic control can reduce the risk of complications and is a valid economic strategy. An observational study by Genovese and Tedeschi²⁹ suggested that therapy with vildagliptin/metformin reduces indirect costs by increasing work productivity and also improves the quality of life. Our results are in line with these studies.^29,30

Possible study limitations are concerned with the accurate diagnosis of T2DM and the possible bias of patient classification and of the operational measurement of costs attributable to the information system developed. This type of study design is subject to bias (factors not taken into account such as the socioeconomic, cultural, or education level, compliance, pharmacological or therapeutic doses consumed, the correctness of treatment, etc.), which should be minimized. The main limitation of the study is the undoubted selection bias due to the attending physician deciding which drugs to administer. Thus, the results should be interpreted with caution. Another limitation was the measurement of hypoglycemia because only episodes in which the patient required medical care and this was documented were identified, leading to a possible under-diagnosis of cases.

Despite the possible limitations of our study, our results show a good approximation to the real behavior of the disease. Further research is required on the long-term cost-effectiveness of these therapies, diagnostic and treatment delays, and the replication of our results in other health organizations. In conclusion, our results show that patients with diabetes and MS treated with DPP-4 inhibitors in combination with metformin had better compliance, better metabolic control, and lower rates of hypoglycemia than patients treated with metformin and other oral antidiabetes drugs. This resulted in lower healthcare costs for the Spanish national health system.

Footnotes

Acknowledgments

This study was funded by Novartis Pharmaceuticals SA.

Author Disclosure Statement

No competing financial interests exist.

A.S.M. was responsible for the conception and design of the manuscript and statistical analysis, A.S.M. and R.N.A. were responsible for data collection, and both authors were involved in the interpretation of the data, drafting, revision, and approval of the submitted manuscript.

References

Pearson

, Palaniappan

, Artinian

, et al.: American Heart Association Guide for Improving Cardiovascular Health at the Community Level, 2013 update: a scientific statement for public health practitioners, healthcare providers, and health policy makers. Circulation, 2013; 127:1730–1753.

Garber

, Abrahamson

, Barzilay

, et al.: AACE comprehensive diabetes management algorithm 2013. Endocr Pract, 2013; 19:327–336.

Cascio

, Schiera

, Di Liegro

: Dietary fatty acids in metabolic syndrome, diabetes and cardiovascular diseases. Curr Diabetes Rev, 2012; 8:2–17.

Linton

, Fazio

; National Cholesterol Education Program (NCEP)–the Third Adult Treatment Panel (ATP III): A practical approach to risk assessment to prevent coronary artery disease and its complications. Am J Cardiol, 2003; 92:19i–26i.

Tziomalos

, Athyros

, Karagiannis

, et al.: Endothelial dysfunction in metabolic syndrome: prevalence, pathogenesis and management. Nutr Metab Cardiovasc Dis, 2010; 20:140–146.

Alberti

, Eckel

, Grundy

, et al.: Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation, 2009; 120:1640–1645.

Inzucchi

, Bergenstal

, Buse

, et al.: Management of hyperglycemia 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). Diabetes Care, 2012; 35:1364–1379.

Grunberger

: Clinical utility of the dipeptidyl peptidase-4 inhibitor linagliptin. Postgrad Med, 2013; 125:79–90.

Scheen

: Cardiovascular effects of dipeptidyl peptidase-4 inhibitors: from risk factors to clinical outcomes. Postgrad Med, 2013; 125:7–20.

10.

Fadini

, de Kreutzenberg

, Gjini

, et al.: The metabolic syndrome influences the response to incretin-based therapies. Acta Diabetol, 2011; 48:219–225.

11.

Svacina

: Incretin therapy and the metabolic syndrome [in Czech]. Vnitr Lek, 2011; 57:417–421.

12.

Lamberts

, Wood

, Hofmans-Okkes

, eds.: The International Classification of Primary Care in the European Community. With a Multi-Language Layer. Oxford, United Kingdom: Oxford University Press, 1993.

13.

National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment on High Blood Cholesterol in Adults (Adult Treatment Panel III): Third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation, 2002; 106:3143–3421.

14.

Sattar

, Gaw

, Scherbakova

, et al.: Metabolic syndrome with and without C-reactive protein as a predictor of coronary heart disease and diabetes in the West of Scotland Coronary Prevention Study. Circulation, 2003; 108:414–419.

15.

Charlson

, Pompei

, Ales

, et al.: A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis, 1987; 40:373–383.

16.

Weiner

, Starfield

, Steinwachs

, et al.: Development and application of a population-oriented measure of ambulatory care case-mix. Med Care, 1991:29:452–472.

17.

World Health Organization: The Anatomical Therapeutic Chemical Classification System. Geneva: World Health Organization, 1991.

18.

Sicras Mainar

, Roldán Suárez

, Font Ramos

, et al.: Clinical and economical consequences of the combination of metformin with dipeptidyl peptidase inhibitors in type 2 diabetes patients [in Spanish, English]. Rev Clin Esp, 2013; 213:377–384.

19.

Thompson

, Barber

: How should cost data in pragmatic randomised trials be analysed?. BMJ, 2000; 320:1197–1200.

20.

Varughese

, Tomson

, Lip

GYH

: Type 2 diabetes mellitus: a cardiovascular perspective. Int J Clin Pract, 2005; 59:798–816.

21.

Asche

, LaFleur

, Conner

: A review of diabetes treatment adherence and the association with clinical and economic outcomes. Clin Ther, 2011; 33:74–109.

22.

Rathmann

, Kostev

, Gruenberger

, et al.: Treatment persistence, hypoglycaemia and clinical outcomes in type 2 diabetes patients with dipeptidyl peptidase-4 inhibitors and sulphonylureas: a primary care database analysis. Diabetes Obes Metab, 2013; 15:55–61.

23.

Erlich

, Slawson

, Shaughnessy

: Diabetes update: new drugs to manage type 2 diabetes. FP Essent, 2013; 408:20–24.

24.

Baetta

, Corsini

: Pharmacology of dipeptidyl peptidase-4 inhibitors: similarities and differences. Drugs, 2011; 71:1441–1467.

25.

Scheen

: DPP-4 inhibitors in the management of type 2 diabetes: a critical review of head-to-head trials. Diabetes Metab, 2012; 38:89–101.

26.

Breitscheidel

, Stamenitis

, Dippel

, et al.: Economic impact of compliance to treatment with antidiabetes medication in type 2 diabetes mellitus: a review paper. J Med Econ, 2010; 13:8–15.

27.

Almutairi

, Alkharfy

: Direct medical cost and glycemic control in type 2 diabetic Saudi patients. Appl Health Econ Health Policy, 2013; 11:671–675.

28.

Giorda

, Manicardi

, Diago Cabezudo

: The impact of diabetes mellitus on healthcare costs in Italy. Expert Rev Pharmacoecon Outcomes Res, 2011; 11:709–719.

29.

Genovese

, Tedeschi

: Effects of vildagliptin/metformin therapy on patient-reported outcomes: work productivity, patient satisfaction, and resource utilization. Adv Ther, 2013; 30:152–164.

30.

Viereck

, Boudes

: An analysis of the impact of FDA's guidelines for addressing cardiovascular risk of drugs for type 2 diabetes on clinical development. Contemp Clin Trials, 2011; 32:324–332.