The impact of DPP-4 inhibitors on long-term survival among diabetic patients after first acute myocardial infarction

Since 1995, the National Health Insurance Program has provided healthcare to approximately 99% of the residents in Taiwan. The data for this study were collected from National Health Insurance Research Database (NHIRD) from January 2000 through December 2012.

The NHIRD includes detailed information from medical inpatient records including age, gender, diagnosis, interventional procedures, medical orders, and relevant survival data. The database provides researchers with de-identified data via encryption of the identification codes to preserve patient anonymity and has been extensively used in epidemiologic studies in Taiwan. This study was approved by the Human Research Committee of Kaohsiung Veterans General Hospital.

A total of 186,326 patients admitted to hospitals in Taiwan between January 2000 and December 2012 with a primary diagnosis of AMI (ICD: 410–410.92) were retrieved from Taiwan’s NHIRD which consisted of data collected from more than 23,000,000 patients. From this group of 186,326 patients, those with a previous admission for AMI, those who were ≤18 or ≥120 years old, and those patients whose gender was undetermined were excluded. The remaining 186,112 patients were included in the analysis (Fig. 1).

Among the 186,112 patients who were hospitalized for first AMI, 72,924 cases with DM were identified. The remaining 113,188 patients without DM were excluded. A propensity score matching technique was applied to minimize baseline differences between the control group and the DPP-4i group. One-to-one matching was performed using the following variables: gender, age, hypertension, dyslipidemia, diabetes, HF, peripheral vascular disease (PVD), cerebrovascular accident (CVA), end-stage renal disease (ESRD), chronic obstructive pulmonary disease (COPD), and percutaneous coronary intervention (PCI) (Table 1). The data from 2672 AMI patients receiving DPP-4i and 2672 matched controls were included in the final analysis (Fig. 1).

Survival was defined based on the difference between the date of hospitalization and the end date of National Health Insurance (NHI) coverage. Since the NHI premium is paid monthly, coverage can easily be canceled at the time of death. Measurement of mortality was valid via the record of the end date of NHI coverage, within a maximum error of 1 month [14, 15].

The SAS version 9.4 software (SAS Institute, Inc., Cary, NC) was used to analyze the data in this study. All variables were calculated using descriptive statistics. Categorical data were expressed as percentile values and continuous variables were expressed as a mean and standard deviation (SD). Paired t test for continuous variables and Chi squared test for categorical variables were applied to evaluate between-group differences. A P < 0.05 was considered statistically significant.

Cox proportional hazard regression analysis was used to calculate the hazard ratio (HR) and the associated 95% confidence intervals (95% CIs) for significant variables. Kaplan–Meier cumulative survival curves were used to compare survival between patients who received DPP-4i compared with those who did not receive DPP-4i in order to compare survival between the two groups as a whole, based on gender, and also based on age. Log-rank tests which used a P < 0.05 were considered statistically significant.

Most of the prior literature has focused on the relationship between DPP-4i and cardiovascular safety. Several recent studies (including pre-clinical data, small mechanistic studies, and post hoc analyses of randomized clinical trials) support the benefit of DPP-4i in patients with cardiovascular disease [16‐18]. In a nationwide longitudinal study, DPP4i-treated T2DM patients were shown to have a lower risk for cardiovascular disease as compared with non-DPP4i users [19]. Furthermore, in DM patients with pre-existing heart failure, the use of DPP-4i has resulted in a lower risk of mortality in patients with the combination of myocardial infarction and ischemic stroke [16]. Some studies have discussed the effect of DPP-4i on cardiovascular outcomes, but were limited regarding their ability to evaluate the impact of DPP-4i on the long-term outcomes after first AMI.

In a murine experimental diabetes model, DPP-4 inhibition was shown to attenuate cardiac dysfunction and adverse remodeling in the post-MI setting [20]. Furthermore, chronic treatment with DPP-4i in an animal model reduced infarct size and improved LV function, via the GLP-1 receptor-protein kinase A pathway, in a glucose-dependent manner [21, 22]. Peak plasma troponin I did not differ between patients with myocardial infarction on DPP4i and those who did not receive such therapy [23]. Rather than a neutral or negative effect, our study provides evidence that DPP-4i therapy can increase the survival of patients following a first AMI regardless of gender. In this study, DPP-4i therapy after first AMI improved long-term survival rather than resulting in just a “do no harm” effect. Major prospective clinical trials are investigating the various uses and cardiovascular outcomes of DPP-4i in diabetic patients. These trials include the Sitagliptin Cardiovascular Outcome Study (TECOS), the saxagliptin assessment of vascular outcomes recorded in patients with diabetes mellitus-thrombolysis in myocardial infarction (SAVOR-TIMI), the cardiovascular outcomes study of alogliptin in subjects with type 2 diabetes and acute coronary syndrome (EXAMINE), and the cardiovascular outcome study of linagliptin versus glimepiride in patients with type 2 diabetes (CAROLINA) trial [24‐26]. It would be an important milestone and significant influence on the management of diabetes if these trials confirm that DPP-4i can contribute to a reduction in the cardiovascular complications of DM.

There are several possible explanations for the benefit of DPP-4i in patients with AMI. First, DPP4i may reduce reperfusion injury via protection of mitochondrial function [27]. DPP-4i has been shown to elevate the activity of reperfusion injury salvage kinase and reduce reperfusion injury caused by reperfusion-related cardiac tissue damage and instigated arrhythmias [2, 22, 27, 28]. Furthermore, mitochondria are both a major energy and oxidative stress production site. DPP-4i can rescue cardiac mitochondrial dysfunction and decrease reactive oxygen species production in those patients with obesity-related insulin resistance and DM with ischemia/reperfusion injury. In addition, DPP4i can reduce oxidative stress [27, 29]. Diabetes is one of the comorbidities that can induce a systemic inflammatory state, while DPP-4i can reduce the systemic proinflammatory state and decrease oxidative stress, which may explain why DPP-4i decrease coronary microvascular endothelial inflammation and make a contribution to post-AMI remodeling during advanced heart failure [20]. Improvements in survival after AMI in T2 DM might also arise from modification of autophagy in the non-infarcted region of the heart [30]. Diabetes carries a two-fold increased risk of heart failure following myocardial infarction, due to an excessive loss of cardiac microvasculature. Stromal cell-derived factor-1alpha (SDF-1alpha) is a chemokine that is elaborated by ischemic tissue but is rapidly degraded by DPP-4i, thus, DPP-4i attenuates cardiac dysfunction and adverse remodeling in the post-MI setting [20].

Finally, DPP4i might inhibit atherosclerosis and proliferation of vascular smooth muscle cell. The possible impact of DPP-4i on post-AMI survival may derive from their cardioprotective effects and their beneficial effect on cardiovascular risk factors such as atherosclerosis and hypertension [4‐8, 31‐36]. In Japanese patients with diabetes and multiple CV risk factors, DPP4i was shown to decrease blood pressure associated with an improvement in albuminuria in addition to glycemic control [17]. In a sub-analysis of the PROLOGUE study, DPP-4i was shown not to alter endothelial function at the 2-year follow-up [37], a result which was further confirmed by another study [38].

Both DPP-4i and metformin have been shown to improve insulin resistance and attenuate myocardial injury caused by ischemia/reperfusion injury, and several studies have indicated that the combined therapy provided better outcomes than monotherapy with a reduction in arrhythmia scores and reduced all-cause mortality rates [39‐43]. Previous studies revealed that both DPP-4i and metformin users had a significantly lower risk for composite cardiovascular diseases [18]. The combination therapy of DPP4i with conventional OHA led to an improvement in passive left ventricular compliance [20]. Interestingly, there were pleiotropic effects on cardiovascular protection using several OHAs, and their use was also associated with a lower risk of aortic aneurysm growth in Metformin-, sulfonylurea-, and TZD-treated patients but not in patients treated with DPP-4i or alpha-glucosidase inhibitors [19]. In our study, 1168 (43.71%) patients received both metformin and DPP-4i therapy, and the overall survival rates were better in those patients receiving DPP-4i. This benefit may be attributed to metformin (HR = 0.77; 95% CI 0.68–0.86) or DPP-4i (HR = 0.86; 95% CI 0.78–0.95). Furthermore, combined therapy with DPP-4i contributed to the improvement in survival rates independent of gender.

Our study had several limitations. We did not use an objective indicator, such as glycated hemoglobin (HbA1c), as a standard for assessing diabetic compensation and treatment, due to bias in the frequency and interval of HbA1c check-ups. In addition, the DPP-4i group had a higher proportion of patients prescribed β-blockers, ARBs, spironolactone, and OHAs, which implied intractable hypertension, poorly controlled DM, or an unstable status. Regardless of the severity of diabetes or underlying conditions in the DPP-4i group, DPP-4i still made an independent contribution to the long-term survival of patients, which was also confirmed by Cox proportional hazard regression analysis.

Previous studies analyzed relatively small samples, while our study enrolled 186,326 patients with first AMI, based on an entire population comprising 23,000,000 patients. Our large sample size reduced the variability in sampling statistics and also utilized a propensity score, one-to-one matching technique to minimize confounding factors between the DPP-4i and control groups. However, future prospective randomized studies are required to confirm our findings.