Figures
Abstract
Background/Objectives
Data on the comparative effectiveness of oral antidiabetics on cardiovascular outcomes in a clinical practice setting are limited. This study sought to determine whether a differential risk of cardiovascular disease (CVD) exists for the combination of a dipeptidyl peptidase-4 (DPP-4) inhibitor plus metformin versus a sulfonylurea derivative plus metformin or pioglitazone plus metformin.
Methods
We conducted a cohort study of 349,476 patients who received treatment with a DPP-4 inhibitor, sulfonylurea, or pioglitazone plus metformin for type 2 diabetes using the Korean national health insurance claims database. The incidence of total CVD and individual outcomes of myocardial infarction (MI), heart failure (HF), and ischemic stroke (IS) were assessed using the hazard ratios (HRs) estimated from a Cox proportional-hazards model weighted for a propensity score.
Results
During follow-up, 3,881 patients developed a CVD, including 428 MIs, 212 HFs, and 1,487 ISs. The adjusted HR with 95% confidence interval (CI) for a sulfonylurea derivative plus metformin compared with a DPP-4 inhibitor plus metformin was 1.20 (1.09-1.32) for total CVD; 1.14 (1.04-1.91) for MI; 1.07 (0.71-1.62) for HF; and 1.51 (1.28-1.79) for IS. The HRs with 95% CI for total CVD, MI, HF, and IS for pioglitazone plus metformin were 0.89 (0.81-0.99), 1.05 (0.76-1.46), 4.81 (3.53-6.56), and 0.81 (0.67-0.99), respectively.
Citation: Seong J-M, Choi N-K, Shin J-Y, Chang Y, Kim Y-J, Lee J, et al. (2015) Differential Cardiovascular Outcomes after Dipeptidyl Peptidase-4 Inhibitor, Sulfonylurea, and Pioglitazone Therapy, All in Combination with Metformin, for Type 2 Diabetes: A Population-Based Cohort Study. PLoS ONE 10(5): e0124287. https://doi.org/10.1371/journal.pone.0124287
Academic Editor: Kandiah Jeyaseelan, National University of Singapore, SINGAPORE
Received: August 14, 2014; Accepted: March 12, 2015; Published: May 20, 2015
Copyright: © 2015 Seong et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The incidence of type 2 diabetes is increasing worldwide, which imposes a high burden of morbidity and mortality, mainly due to cardiovascular disease (CVD). Patients with diabetes have an increased risk of developing CVD, which is also the leading cause of mortality in patients with diabetes [1–3]. Therefore, it is important that the effects of diabetes therapies on reducing the CVD risk be characterized.
Dipeptidyl peptidase-4 (DPP-4) inhibitors are a class of agents that were recently approved for the treatment of type 2 diabetes. These agents extend the action of insulin while also suppressing the release of glucagon via increasing the bioactive form of the incretins, glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide, thereby improving glucose homeostasis [4–5]. DPP‑4 inhibitors provide comparable efficacy to other oral antidiabetics (OAs) [6] and carry lower risks of hypoglycemia and weight gain [6–7]. Meta-analyses of randomized controlled trials (RCTs) have indicated that DPP-4 inhibitors might reduce the risk of major cardiovascular events (CVEs) compared with a placebo or other OAs [8–10]. However, among the RCTs included in these meta-analyses, CVEs were reported only as adverse events and were not pre-specified as primary outcomes. Recently, post-marketing trials of DPP-4 inhibitors have shown that these drugs neither reduced nor increased the risk of major CVEs compared with placebo [11–12]. However, the results from patients enrolled in RCTs may not be generalized to the general population. Thus, cardiovascular outcome data from patients in real-world clinical settings are needed.
We performed this population-based cohort study to evaluate the differential risk of CVDs between the use of DPP-4 inhibitors and other OAs using the national health insurance claims database. Given the common and increasing use of combined OAs [13], this study focused on the administration of a DPP-4 inhibitor in combination with metformin, and this therapy was compared to treatment with a sulfonylurea derivative and metformin or pioglitazone and metformin, which have been the most commonly prescribed therapies in dual-therapy users. Study drugs were approved for the treatment of type 2 diabetes as monotherapy or combination therapy, whereas prescribing patterns of these drugs reflect the reimbursement criteria; metformin monotherapy is considered as the first-line drug for type 2 diabetes and covered by Korean health insurance program; a DPP-4 inhibitor, sulfonylurea drug or pioglitazone is generally used as combination therapy with metformin or other OAs.
Materials and Methods
Study Design, Setting, and Population
The source population for this retrospective cohort study was derived from the Health Insurance Review & Assessment Service (HIRA) database, which represents the national health insurance claims database covering approximately 50 million Koreans [14–15]. We obtained claims data that had been submitted for patients by healthcare providers between January 1, 2006 and December 31, 2010; these data contained anonymized identifiers given by HIRA to protect the patient’s privacy, according to the Act on the Protection of Personal Information Maintained by Public Agencies. The database contains longitudinal patient data including patient demographics, diagnoses (International Classification of Disease, Tenth Revision [ICD-10]), procedures, prescription drugs (brand name, generic name, prescription date, days of supply, dose, and route of administration), and type of medical utilization (outpatient, inpatient, or emergency department).
The study population consisted of type 2 diabetes patients (ICD-10 codes E11-14) who were newly treated with the study therapies between December 1, 2008 (the date that the DPP-4 inhibitors were introduced into the national insurance coverage) and December 31, 2009. These subjects ranged in age from 20 to 99 years at cohort entry (Fig 1). The study therapies were defined as follows: a DPP-4 inhibitor (sitagliptin or vildagliptin) plus metformin, a sulfonylurea drug (glibenclamide, glipizide, gliquidone, gliclazide, or glimepiride) plus metformin, or pioglitazone plus metformin. We identified all prescriptions for the study therapies. For all eligible patients, the index date was defined as the date when any of these 3 study therapies was first prescribed. To include only patients that were newly treated with the study therapies, patients who had received any of these 3 therapies in the year before the index date were excluded. To identify incident CVD, patients with a history of CVD within the 2.5 years prior to cohort entry were also excluded.
ICD-10 = the International Classification of Disease, Tenth Revision; SU = sulfonylurea; TZD = thiazolidinedione; DPP-4 = dipeptidyl peptidase-4; IHD = ischemic heart disease; HF = heart failure; TIA = transient ischemic attack; CABG = coronary artery bypass grafts; PCI = percutaneous coronary intervention.
The study patients were classified into the following three study groups: DPP-4 inhibitor (DPP-4 inhibitor+metformin), sulfonylurea (SU+metformin), and pioglitazone (pioglitazone+metformin). Patients receiving rosiglitazone plus metformin were excluded because the use of rosiglitazone has been significantly limited in Korea.
Exposure Assessment
A patient was considered continuously exposed to a study therapy if he or she filled a prescription for that therapy within the end date of the previous prescription plus 1.5 times the prescription days’ supply [16]. If a patient did not refill the prescription within this time (therapy stop) or the patient filled a prescription for a different therapy or made an addition and/or subtraction of any hypoglycemic agent to an existing therapy (therapy switch), he or she was considered to have discontinued the study therapy. The therapy discontinuation date for therapy stop was set as 1.5 times the prescription days’ supply after the end of the days’ supply from the most recent prescription of a study therapy. For a therapy switch, the therapy discontinuation date was the date a different therapy was prescribed.
Study Outcomes
The primary outcome of interest was any CVD, defined as hospitalization or a visit to an emergency department with a primary diagnosis of ischemic heart disease (ICD-10 codes I20-25), heart failure (I50; HF), ischemic stroke (I63, I65, and I66; IS), transient ischemic attack (G45), atherosclerosis (I70), or aortic aneurysm (I71) or for treatment with coronary artery bypass grafts or percutaneous coronary intervention. The secondary outcomes consisted of individual diseases including myocardial infarction (MI), HF, and IS. The secondary outcomes were defined as hospitalization or a visit to an emergency department with a primary diagnosis of each outcome (ICD-10 codes for MI: I21-23; for HF: I50; for IS: I63, I65, and I66). The outcome date was the earliest date a patient encountered a given outcome.
Covariates
The study covariates included age at index date, gender, duration of diabetes, presence of comorbidities, and use of the medications specified below. The duration of diabetes was assessed based on the date of the first medical claim with a diagnosis code for diabetes (ICD-10 codes E11-14) during the 2.5 years prior to the index date. Comorbidities were determined by ICD-10 codes for the following conditions in the 1 year prior to the index date: microvascular complications of diabetes (retinopathy, neuropathy, or nephropathy), peripheral vascular disease, hypertension, and dyslipidemia (S1 Table). Obesity was assessed based on any medical claim with diagnosis of obesity (ICD-10 code E66). The Charlson comorbidity score was calculated in the year before the index date [17]. We assessed whether a patient had a diagnosis of family history of diabetes mellitus (Z83.3) from January 1, 2006 to the end date of cohort entry. We also assessed whether a patient was hospitalized with a primary diagnosis of diabetes and the total number of hypoglycemic agent classes used in the year prior to the index date. Finally, we assessed the use of the following medications in the year prior to the index date: metformin, sulfonylureas, thiazolidinediones, α-glucosidase inhibitors, meglitinides, DPP-4 inhibitors, insulin, angiotensin converting enzyme inhibitors, angiotensin receptor blockers (ARBs), β-blockers, calcium channel blockers, thiazide diuretics, other diuretics, nitrates, digoxin, aspirin, other antiplatelet drugs, warfarin, statins, and nonsteroidal anti-inflammatory drugs.
Follow-up and Statistical Analysis
Follow-up for each patient started on the index date and ended at the earliest occurrence of a study outcome, therapy discontinuation (therapy stop or switch), death (in-hospital death), or the end of the study period (December 31, 2010).
Baseline characteristics of each study group were compared using standardized mean differences (SMD), calculated as the difference in means or proportions of a variable divided by a pooled estimate of the standard deviation of the variable. A value of 0.1 SMD or less indicates a negligible difference in means between groups [18].
Incidence rates and 95% confidence intervals (CI) of a study outcome for each of the study groups were calculated. We compared the DPP-4 inhibitor group with both the sulfonylurea and pioglitazone groups. Hazard ratios (HRs) with 95% CIs were calculated using Cox proportional hazards models. Propensity score analyses were used to balance confounders between the group receiving a DPP-4 inhibitor plus metformin and each of the 2 groups receiving a sulfonylurea drug plus metformin and pioglitazone plus metformin. The propensity scores were estimated for each patient by fitting a logistic regression model to predict a DPP-4 inhibitor plus metformin versus a sulfonylurea drug or pioglitazone plus metformin initiation, as a function of the baseline covariates. Among all covariates (Table 1), variables associated with a greater than 20% increase or reduction in each outcome rate were included in the propensity score [19–20]. We then weighted a Cox model based on propensity score values; in these analyses, we used standardized morbidity ratio (SMR) weights [21–22]. The SMR weighting method estimates the treatment effect in a population whose distribution of risk factors is equal to that found in the treated study patients. These SMR weighted analyses use as weights the value 1 for the treated (that is, users of a DPP-4 inhibitor plus metformin) and the propensity odds for the untreated (users of a sulfonylurea or pioglitazone plus metformin). To estimate the gender-specific HRs and 95% CIs, we stratified the analysis by gender (male and female). All statistical analyses were performed using SAS software Version 9.3 (SAS Institute Inc., Cary, NC, USA).
Results
Characteristics of Study Population
A total of 349,476 patients were included in the cohort, including 74,720 patients treated with a DPP-4 inhibitor plus metformin, 253,563 treated with a sulfonylurea derivative plus metformin, and 21,193 treated with pioglitazone plus metformin (Fig 1). Patients in the DPP-4 inhibitor group demonstrated a longer duration of diabetes, a higher Charlson comorbidity score, and a larger number of hypoglycemic agents used compared with both the sulfonylurea and pioglitazone groups. Patients using a DPP-4 inhibitor plus metformin were also younger and more likely to have prescriptions for ARBs, other antiplatelet drugs, and statins than the sulfonylurea plus metformin users (Table 1).
Incidence Rate of Study Outcomes
Study patients contributed between 210,878 and 211,959 person-years of follow-up, depending which the study outcome was analyzed. During follow-up, 3,881 patients developed a CVD, including 428 MIs, 212 HFs, and 1,487 ischemic strokes. Among the patients treated with a DPP-4 inhibitor plus metformin, the crude incidence per 1,000 person-years was 15.75 for total CVD, 1.45 for MI, 0.88 for HF, and 4.65 for ischemic stroke. The crude incidence per 1,000 person-years was 19.88 for total CVD, 2.30 for MI, 1.05 for HF, and 8.24 for IS among patients using a sulfonylurea drug plus metformin, and these values were 13.44 for total CVD, 1.31 for MI, 1.10 for HF, and 3.88 for IS among pioglitazone plus metformin users (Table 2).
Hazard Ratios in the Full Cohort and in the Gender-specific group
In the models weighted for a propensity score, the HRs of total CVD, MI, and IS for a sulfonylurea derivative plus metformin compared with a DPP-4 inhibitor plus metformin were 1.20 (95% CI, 1.09–1.32), 1.41 (95% CI, 1.04–1.91), and 1.51 (95% CI, 1.28–1.79), respectively. The HRs of total CVD, HF, and IS for pioglitazone plus metformin compared with a DPP-4 inhibitor plus metformin were 0.89 (95% CI, 0.81–0.99), 4.81 (95% CI, 3.53–6.56), and 0.81 (95% CI, 0.67–0.99), respectively (Table 2). Gender-specific HRs are shown in Table 3. The HR of MI for a sulfonylurea drug plus metformin compared with a DPP-4 inhibitor plus metformin was 1.10 (95% CI, 0.75–1.59) among males and 2.42 (95% CI, 1.38–4.25) among females. The HR for total CVD for pioglitazone plus metformin compared with a DPP-4 inhibitor plus metformin was 0.75 (95% CI, 0.65–0.86) among males and 1.09 (95% CI, 0.94–1.26) among females. The corresponding HR for HF was 1.08 (95% CI, 0.52–2.24) among males and 6.53 (95% CI, 4.54–9.40) among females, and the corresponding HR for IS was 0.66 (95% CI, 0.49–0.89) among males and 0.96 (95% CI, 0.74–1.25) among females.
Discussion
The results of this population-based cohort study showed that the use of a sulfonylurea derivative plus metformin was associated with increased risks of total CVD, MI, and IS compared with the use of a DPP-4 inhibitor plus metformin. In addition, we found that the use of pioglitazone plus metformin was associated with an increased risk of HF compared with a DPP-4 inhibitor plus metformin. However, lower risks for total CVD and IS were observed in patients receiving pioglitazone plus metformin.
The findings of this study were similar to those of a previous meta-analysis that examined the risk for major CVEs among patients using a DPP-4 inhibitor relative to the placebo or other OAs. In this meta-analysis, DPP-4 inhibitors were associated with a lower risk of major CVEs, although the result was not statistically significant when subgroup analysis was conducted to compare the DPP-4 inhibitor and sulfonylurea use [8]. Recently published data from RCTs including patients at a higher cardiovascular risk showed that DPP-4 inhibitor and placebo groups did not differ significantly with respect to major CVEs [11–12]; however, comparison of these data with our data were difficult due to the differences in patient characteristics and the comparison group. The present study found significantly increased risks of total CVD, MI, and IS for treatment with a sulfonylurea drug plus metformin compared to treatment with a DPP-4 inhibitor plus metformin, suggesting a reduced risk in DPP-4 inhibitor users. Indeed, some preclinical and clinical studies have demonstrated that DPP-4 inhibitors may provide beneficial effects on blood pressure, lipid metabolism, inflammatory mediators, and endothelial function, although the precise underlying mechanisms for cardiovascular benefits require further investigation [23–30].
The risk of HF was associated with pioglitazone plus metformin compared with a DPP-4 inhibitor plus metformin, and this result is consistent with the fact that pioglitazone is associated with an increased risk for HF [31]. However, our study also found that pioglitazone plus metformin was associated with lower risks of total CVD and IS than a DPP-4 inhibitor plus metformin. Although the effect of pioglitazone compared with DPP-4 inhibitors on cardiovascular outcomes has not been previously reported, a meta-analysis of 19 trials comparing pioglitazone with a placebo or other OAs found that pioglitazone was associated with a significantly lower risk of stroke and the composite of MI, stroke, or death [31]. The favorable effects of pioglitazone on stroke, death, and the composite of acute MI, stroke, HF, or death compared to rosiglitazone have also been reported in an observational study [32]. These findings suggest that the net clinical cardiovascular benefit of pioglitazone therapy could be favorable in patients without a risk of HF. However, the findings of this study must be interpreted with caution because they may be subject to biases arising from confounding variables. While differences in baseline characteristics were adjusted in our analysis, this adjustment could not completely correct for the uneven distribution of unobservable factors that may have influenced the cardiovascular risks. Moreover, channeling patients at higher risk to receive a DPP-4 inhibitor plus metformin may have occurred because this is a newer therapy with uncertain cardiovascular risk. Other authors have observed that sitagliptin users show higher proportions of comorbidities and a greater use of prescription medications and physician visits compared to users of other OAs [33–34], and these findings suggest that patients receiving a DPP-4 inhibitor may be at higher risk for clinical and health outcomes such as CVEs due to their higher baseline risk.
In the stratified analysis, the adjusted HRs for pioglitazone plus metformin compared with a DPP-4 inhibitor plus metformin were significantly reduced for total CVD and IS among males; however, the adjusted HRs were not statistically significant among females. The adjusted HR of HF for pioglitazone plus metformin was not significantly increased among males; however, the corresponding HR was statistically significantly increased among females. This result suggests a potential difference in the cardiovascular effect of pioglitazone on males and females. Indeed, pioglitazone is known to influence estrogen function and sex hormone-binding globulin expression [35–36], which may have contributed to the observed gender difference.
A study outcome occurred shortly after insulin initiation would be the effect of study therapy on cardiovascular outcomes, whereas this event was not defined as an occurrence of study outcome according to the operational definition of at-risk person time; this may bias study results. However, analysis including that event, which is an outcome of interest occurred after therapy switch, may also be subject to bias because that accounts for the different diabetes therapy conditions.
This study possesses several unique strengths. First, this study included the entire population of patients with type 2 diabetes using the national health insurance claims database; therefore, this study could show representative results in the population of approximately 50 million residing in Korea. In addition, our study population represents the patient population in real-world practice settings. Second, head‐to‐head comparisons of a DPP-4 inhibitor plus metformin with a sulfonylurea drug plus metformin and pioglitazone plus metformin were conducted. Consequently, our data provide information on the comparative effectiveness of these therapies. In addition, the combined use of DPP-4 inhibitors with metformin was covered by health insurance in Korea, whereas monotherapy with DPP-4 inhibitors was not covered. Therefore, in this study, we were able to observe most of the patients who used DPP-4 inhibitors during the study period.
Our study also had some limitations. First, the accuracy of diagnoses from claims databases is limited. However, previous studies have validated ICD-10 code-based definitions for many important diseases, including diabetes, acute MI, and ischemic stroke, which were compared with medical records reviews; these reports demonstrated positive predictive values of 72.3% to 87.2% for diabetes (E10-14), >70% for MI (I21), and 83.4% for IS (I63) [37–39]. In addition, the overall positive predictive value of the diagnosis was approximately 70% [40]. Second, clinical information, such as hemoglobin A1c levels, was not available; thus, the degree of glycemic control or the grading of diabetes severity was not well captured, which could have influenced the incidence of the outcomes. However, hospitalizations for diabetes and the total number of hypoglycemic agent classes used in the year prior to the index date were measured and adjusted for as a proxy of the degree of glycemic control. The presence of microvascular complications and peripheral vascular disease were also determined and adjusted for as a proxy of diabetes severity. Third, the maximum follow-up observation time was 2.1 years (25 months) after the index date. Even if a meta-analysis found that DPP-4 inhibitors were also associated with a significantly lower risk of major CVEs in the subgroup analysis of trials of relatively short duration (<52 weeks) [8], their long-term effects on cardiovascular outcomes remain to be established in future and ongoing studies. Fourth, although numerous potential and measurable cardiovascular risk factors were adjusted for, there is the possibility that the results were affected by unmeasured confounders, which is a universal problem with all observational studies.
In conclusion, our data suggest that, compared with a DPP-4 inhibitor plus metformin, treatment with a sulfonylurea derivative plus metformin was associated with increased risks of total CVD, MI, and ischemic stroke. Thus, adding a DPP-4 inhibitor instead of a sulfonylurea drug as second-line treatment to first-line metformin should be considered. Our results also found that the use of pioglitazone plus metformin was associated with decreased total CVD and IS risks compared with a DPP-4 inhibitor plus metformin. Although fluid retention and heart failure are more frequent with pioglitazone treatment, the beneficial effects of pioglitazone on total CVD and IS provide a well-selected treatment option for patients with type 2 diabetes and a lower risk of HF. However, this is an observational study, and there may be residual confounding variables; thus, large, ongoing long-term clinical trials to specifically address the cardiovascular risk of DPP-4 inhibitors will offer more insight.
Supporting Information
S1 Fig. Distribution of the estimated propensity of receiving dipeptidyl peptidase-4 inhibitor plus metformin versus a sulfonylurea drug plus metformin in the total cardiovascular disease outcome analysis (A), myocardial infarction (B), heart failure (C), and ischemic stroke (D).
DPP-4 = dipeptidyl peptidase-4; SU = sulfonylurea.
https://doi.org/10.1371/journal.pone.0124287.s001
(TIF)
S2 Fig. Distribution of the estimated propensity of receiving dipeptidyl peptidase-4 inhibitor plus metformin versus a pioglitazone plus metformin in the total cardiovascular disease outcome analysis (A), myocardial infarction (B), heart failure (C), and ischemic stroke (D).
DPP-4 = dipeptidyl peptidase-4.
https://doi.org/10.1371/journal.pone.0124287.s002
(TIF)
S1 Table. Baseline Covariate Definitions and Codes.
https://doi.org/10.1371/journal.pone.0124287.s003
(DOCX)
Author Contributions
Conceived and designed the experiments: JMS NKC BJP. Performed the experiments: JMS NKC BJP. Analyzed the data: JMS. Contributed reagents/materials/analysis tools: JMS NKC BJP. Wrote the paper: JMS NKC JYS YC YJK JK JL BJP.
References
- 1. Fox CS, Coady S, Sorlie PD, Levy D, Meigs JB, D'Agostino RB Sr, et al. (2004) Trends in cardiovascular complications of diabetes. JAMA 292: 2495–2499. pmid:15562129
- 2. Franco OH, Steyerberg EW, Hu FB, Mackenbach J, Nusselder W (2007) Associations of diabetes mellitus with total life expectancy and life expectancy with and without cardiovascular disease. Arch Intern Med 167: 1145–1151. pmid:17563022
- 3. Fox CS, Coady S, Sorlie PD, D'Agostino RB Sr, Pencina MJ, Vasan RS, et al. (2007) Increasing cardiovascular disease burden due to diabetes mellitus: the Framingham Heart Study. Circulation 115: 1544–1550. pmid:17353438
- 4. Deacon CF (2004) Therapeutic strategies based on glucagon-like peptide 1. Diabetes 53: 2181–2189. pmid:15331525
- 5. Drucker DJ (2006) The biology of incretin hormones. Cell Metab 3: 153–165. pmid:16517403
- 6. Phung OJ, Scholle JM, Talwar M, Coleman CI (2010) Effect of noninsulin antidiabetic drugs added to metformin therapy on glycemic control, weight gain, and hypoglycemia in type 2 diabetes. JAMA 303: 1410–1418. pmid:20388897
- 7. Pratley RE, Salsali A (2007) Inhibition of DPP-4: a new therapeutic approach for the treatment of type 2 diabetes. Curr Med Res Opin 23: 919–931. pmid:17407649
- 8. Johansen OE, Neubacher D, von Eynatten M, Patel S, Woerle HJ (2012) Cardiovascular safety with linagliptin in patients with type 2 diabetes mellitus: a pre-specified, prospective, and adjudicated meta-analysis of a phase 3 programme. Cardiovasc Diabetol 11: 3. pmid:22234149
- 9. Frederich R, Alexander JH, Fiedorek FT, Donovan M, Berglind N, Harris S, et al. (2010) A systematic assessment of cardiovascular outcomes in the saxagliptin drug development program for type 2 diabetes. Postgrad Med 122: 16–27. pmid:21084777
- 10. Monami M, Dicembrini I, Martelli D, Mannucci E (2011) Safety of dipeptidyl peptidase-4 inhibitors: a meta-analysis of randomized clinical trials. Curr Med Res Opin 27 Suppl 3: 57–64. pmid:22106978
- 11. White WB, Cannon CP, Heller SR, Nissen SE, Bergenstal RM, Bakris GL, et al.; EXAMINE Investigators (2013) Alogliptin after acute coronary syndrome in patients with type 2 diabetes. N Engl J Med 369: 1327–1335. pmid:23992602
- 12. Scirica BM, Bhatt DL, Braunwald E, Steg PG, Davidson J, Hirshberg B, et al.; SAVOR-TIMI 53 Steering Committee and Investigators (2013) Saxagliptin and cardiovascular outcomes in patients with type 2 diabetes mellitus. N Engl J Med 369: 1317–1326. pmid:23992601
- 13. United Kingdom Prospective Diabetes Study Group (1998) United Kingdom Prospective Diabetes Study 24: a 6-year, randomized, controlled trial comparing sulfonylurea, insulin, and metformin therapy in patients with newly diagnosed type 2 diabetes that could not be controlled with diet therapy. United Kingdom Prospective Diabetes Study Group. Ann Intern Med 128: 165–175. pmid:9454524
- 14. Seong JM, Choi NK, Jung SY, Chang Y, Kim YJ, Lee J, et al. (2011) Thiazolidinedione use in elderly patients with type 2 diabetes: with and without heart failure. Pharmacoepidemiol Drug Saf 20: 344–350. pmid:21328632
- 15. Choi NK, Hahn S, Park BJ (2007) Increase in mortality rate following coprescription of cisapride and contraindicated drugs. Ann Pharmacother 41: 667–673. pmid:17374629
- 16. Sikka R, Xia F, Aubert RE (2005) Estimating medication persistency using administrative claims data. Am J Manag Care 11: 449–457. pmid:16044982
- 17. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. (2005) Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 43: 1130–1139. pmid:16224307
- 18. Mamdani M, Sykora K, Li P, Normand SL, Streiner DL, Austin PC, et al. (2005) Reader's guide to critical appraisal of cohort studies: 2. Assessing potential for confounding. BMJ 330: 960–962. pmid:15845982
- 19. Patrick AR, Schneeweiss S, Brookhart MA, Glynn RJ, Rothman KJ, Avorn J, et al. (2011) The implications of propensity score variable selection strategies in pharmacoepidemiology: an empirical illustration. Pharmacoepidemiol Drug Saf 20: 551–559. pmid:21394812
- 20. Brookhart MA, Schneeweiss S, Rothman KJ, Glynn RJ, Avorn J, Stürmer T (2006) Variable selection for propensity score models. Am J Epidemiol 163: 1149–1156. pmid:16624967
- 21. Kurth T, Walker AM, Glynn RJ, Chan KA, Gaziano JM, Berger K, et al. (2006) Results of multivariable logistic regression, propensity matching, propensity adjustment, and propensity-based weighting under conditions of nonuniform effect. Am J Epidemiol 163: 262–270. pmid:16371515
- 22. Sato T, Matsuyama Y (2003) Marginal structural models as a tool for standardization. Epidemiology 14: 680–686. pmid:14569183
- 23. Fadini GP, Avogaro A (2011) Cardiovascular effects of DPP-4 inhibition: beyond GLP-1. Vascul Pharmacol 55: 10–16. pmid:21664294
- 24. van Poppel PC, Netea MG, Smits P, Tack CJ (2011) Vildagliptin improves endothelium-dependent vasodilatation in type 2 diabetes. Diabetes Care 34: 2072–2077. pmid:21788633
- 25. Fadini GP, Boscaro E, Albiero M, Menegazzo L, Frison V, de Kreutzenberg S, et al. (2010) The oral dipeptidyl peptidase-4 inhibitor sitagliptin increases circulating endothelial progenitor cells in patients with type 2 diabetes: possible role of stromal-derived factor-1alpha. Diabetes Care 33: 1607–1609. pmid:20357375
- 26. Shah Z, Kampfrath T, Deiuliis JA, Zhong J, Pineda C, Ying Z, et al. (2011) Long-term dipeptidyl-peptidase 4 inhibition reduces atherosclerosis and inflammation via effects on monocyte recruitment and chemotaxis. Circulation 124: 2338–2349. pmid:22007077
- 27. Matikainen N, Mänttäri S, Schweizer A, Ulvestad A, Mills D, Dunning BE, et al. (2006) Vildagliptin therapy reduces postprandial intestinal triglyceride-rich lipoprotein particles in patients with type 2 diabetes. Diabetologia 49: 2049–2057. pmid:16816950
- 28. Tremblay AJ, Lamarche B, Deacon CF, Weisnagel SJ, Couture P (2011) Effect of sitagliptin therapy on postprandial lipoprotein levels in patients with type 2 diabetes. Diabetes Obes Metab 13: 366–373. pmid:21226820
- 29. Ogawa S, Ishiki M, Nako K, Okamura M, Senda M, Mori T, et al. (2011) Sitagliptin, a dipeptidyl peptidase-4 inhibitor, decreases systolic blood pressure in Japanese hypertensive patients with type 2 diabetes. Tohoku J Exp Med 223: 133–135. pmid:21304217
- 30. Mistry GC, Maes AL, Lasseter KC, Davies MJ, Gottesdiener KM, Wagner JA, et al. (2008) Effect of sitagliptin, a dipeptidyl peptidase-4 inhibitor, on blood pressure in nondiabetic patients with mild to moderate hypertension. J Clin Pharmacol 48: 592–598. pmid:18353996
- 31.
Actos (pioglitazone hydrochloride) [product label]. Deerfield, IL: Takeda Pharmaceuticals America Inc; 2008.
- 32. Graham DJ, Ouellet-Hellstrom R, MaCurdy TE, Ali F, Sholley C, Worrall C, et al. (2010) Risk of acute myocardial infarction, stroke, heart failure, and death in elderly Medicare patients treated with rosiglitazone or pioglitazone. JAMA 304: 411–418. pmid:20584880
- 33. Zhang Q, Rajagopalan S, Mavros P, Engel SS, Davies MJ, Yin D, et al. (2010) Baseline characteristic differences between patients prescribed sitagliptin vs. other oral antihyperglycemic agents: analysis of a US electronic medical record database. Curr Med Res Opin 26: 1697–1703. pmid:20465367
- 34. Cai B, Katz L, Alexander CM, Williams-Herman D, Girman CJ (2010) Characteristics of patients prescribed sitagliptin and other oral antihyperglycaemic agents in a large US claims database. Int J Clin Pract 64: 1601–1608. pmid:20946268
- 35. Selva DM, Hammond GL (2009) Peroxisome-proliferator receptor gamma represses hepatic sex hormone-binding globulin expression. Endocrinology 150: 2183–2189. pmid:19179433
- 36. Starcke S, Vollmer G (2006) Is there an estrogenic component in the metabolic syndrome? Genes Nutr 1: 177–188. pmid:18850213
- 37. Park JK, Kim KS, Kim CB, Lee TY, Lee KS, Lee DH, et al. (2000) The Accuracy of ICD codes for Cerebrovascular Diseases in Medical Insurance Claims. Korean J Prev Med 33: 76–82.
- 38. Kimm H, Yun JE, Lee SH, Jang Y, Jee SH (2012) Validity of the diagnosis of acute myocardial infarction in korean national medical health insurance claims data: the korean heart study (1). Korean Circ J 42: 10–15. pmid:22363378
- 39.
Korean Diabetes Association, Health Insurance Review & Assessment Service: Report of Task Force Team for Basic Statistics Study: Diabetes in Korea 2007.
- 40.
Park BJ, Sung JH, Park KD, Seo SW, Kim SW (2003) Report of the evaluation for validity of discharged diagnoses in Korean Health Insurance database. Seoul: Seoul National University 19–52.