Dipeptidyl peptidase-4 (DPP-4) inhibitors may affect the serum levels of plasminogen activator inhibitor-1 (PAI-1) associated with triglyceride (TG) metabolism, which is a prognostic factor for cardiovascular disease, in diabetic patients. We conducted an 8-week, prospective, randomized study in which we assigned type 2 diabetic patients who were inadequately controlled with antidiabetic therapy to the vildagliptin group (50 mg bid, n = 49) or the control group (n = 49). The primary efficacy parameter was the change in the serum level of PAI-1, and the secondary end point was the change in the serum levels of TG-rich lipoproteins. In the vildagliptin group, significant decrease of the serum PAI-1 level by 16.3% (p <0.0001) and significant decreases of the serum TG, remnant-like particle cholesterol, and apolipoprotein B levels by 12.1% (p = 0.002), 13.9% (p = 0.003), and 9.5% (p <0.0001), respectively, were observed. No such changes were observed in the control group. Multivariate regression analyses identified the absolute change from the baseline (Δ) of the PAI-1, but not that of the fasting blood glucose or hemoglobin A1c, as independent predictors of the ΔTG, Δ remnant-like particle cholesterol, and Δ apolipoprotein B. In conclusion, treatment of type 2 diabetes with vildagliptin might prevent the progression of atherosclerotic cardiovascular disease in diabetic patients by decreasing the serum PAI-1 levels and improving TG metabolism.
Progression of atherothrombotic disease in diabetes has been linked with elevated levels of various coagulation factors including plasminogen activator inhibitor-1 (PAI-1). Moreover, increase in the serum level of a cardiovascular risk marker, PAI-1, which is produced by adipocytes, the cells that are responsible for the physiological activities of adipose tissue, has been reported to be related to increase in the serum levels of triglyceride (TG)-rich lipoproteins (TRLs), which have a strong effect of inducing the progression of arteriosclerosis. Basic studies have reported that glucagon-like peptide 1 (GLP-1) receptor and GLP-1 activated by dipeptidyl peptidase-4 (DPP-4) inhibitors are involved in the inhibition of PAI-1 production through direct or indirect effects on the adipose tissue. However, there is no consensus as yet. We hypothesized that DPP-4 inhibitors might directly or indirectly inhibit the physiological activities of the adipocytes, thereby reducing PAI-1 production. As a result, TG metabolism might be improved, leading to a reduction in the serum levels of TRLs. The purpose of this study was to investigate the effect of DPP-4 inhibitors on the serum levels of PAI-1, to elucidate the relation between TG metabolism and PAI-1 and to discuss, at least in part, the mechanism underlying the prevention of cardiovascular events by DPP-4 inhibitors.
Methods
We conducted an 8-week, prospective, open-label study in which we randomized type 2 diabetic patients who were inadequately controlled with antidiabetic therapy to either a vildagliptin group (50 mg bid, n = 50) or a control group (n = 50) by a simple sealed envelope method. The primary efficacy parameter was the change in the serum level of PAI-1, and the secondary end point was the change in the serum levels of the TRLs after vildagliptin therapy.
From among the 123 outpatients with fasting blood glucose (FBG) levels of ≥126 mg/dl and hemoglobin A1c (HbA1c) level of ≥6.5% treated at Cardiovascular Center, Surugadai Nihon University Hospital, who were recommended a 6-month lifestyle improvement by exercising, diet therapy, and medical treatment, the 110 patients who failed to meet the above diabetic control criteria despite making the recommended lifestyle improvements were encouraged to participate in this study. One hundred patients who gave consent for participation in this trial were randomized with a simple sealed envelope method to vildagliptin treatment (the vildagliptin group, n = 50) or non-vildagliptin treatment (the control group, n = 50) in this trial between April 12, 2013, and April 25, 2013; vildagliptin, 50 mg, orally, twice daily, total daily dose 100 mg, was prescribed. There were no changes or additions to any of the antidiabetic drugs or lipid-modifying drugs during the period of this trial. Blood and urine tests were performed at the start of vildagliptin therapy and 8 weeks after the start. Patients were not enrolled if they met any of the following exclusion criteria: hepatic dysfunction (alanine aminotransferase and aspartate aminotransferase ≥2 times the upper limit of the normal values) or known malignant disease.
A diagnosis of hypertension was made when the systolic pressure was ≥140 mm Hg and/or the diastolic pressure was ≥90 mm Hg or the subjects gave a history of taking antihypertensive medication. Smoking was defined as current smoking or smoking cessation within 1 year before the start of the study. A diagnosis of lipid disorder was made when the low-density lipoprotein cholesterol (LDL-C) level was ≥140 mg/dl, the TG level was ≥150 mg/dl, or the high-density lipoprotein cholesterol (HDL-C) level was <40 mg/dl or if the patient was already on lipid-modified treatment. A diagnosis of hyperuricemia was made when the serum uric acid level was ≥7.0 mg/dl or taking medications. A history of coronary artery disease was the following: documented myocardial infarction, previous coronary revascularization intervention (coronary artery bypass graft surgery or percutaneous coronary intervention), ≥50% stenosis in 1 or more coronary arteries identified during cardiac catheterization, and coronary spastic angina. The estimated glomerular filtration rate was calculated using the abbreviated Modification of Diet in Renal Disease Study equation modified by a Japanese coefficient. The homeostasis model assessment-insulin resistance (HOMA-IR) was calculated as the fasting insulin level (μU/ml) × FBG level (mg/dl)/405. This study was conducted in accordance with the ethical principles of the Declaration of Helsinki. The study protocol was approved by the institutional review boards at our institute, and written informed consent was obtained from all participants.
Fasting blood samples were collected early in the morning after a 12-hour fast. The serum total cholesterol, HDL-C, and TG levels were measured by the standard methods. The serum LDL-C level was estimated using the Friedewald formula. The remnant-like particle cholesterol (RLP-C) level was measured by an immunoadsorption assay (SRL Co., Ltd, Tokyo, Japan). The serum apolipoprotein (apo) level was determined by turbidimetric latex agglutination assays (Daiichi Pure Chemicals Co., Ltd, Tokyo, Japan). The serum immunoreactive insulin (IRI) level was measured by the chemiluminescence enzyme immunoassay (SRL).
We used the LipoPhor system (Joko, Tokyo, Japan) for the evaluation of the LDL particle size in this study. Relative LDL migration (LDL-Rm value), an indicator of LDL particle size, was measured relative to the mobility value of LDL by performing polyacrylamide gel electrophoresis with the LipoPhor system. LDL-Rm value was calculated as the distance between the very LDL peak and the LDL peak divided by the distance between the very LDL peak and the HDL peak. LDL-Rm values obtained by this method have been demonstrated to show strong negative correlation with the LDL particle size. A decrease in the LDL-Rm value indicates an increase of the LDL particle size.
We performed all the statistical analyses using the SPSS Window, version 12.0 software program (Statistical Package for the Social Sciences; SPSS Inc., Chicago, Illinois). Data were expressed as the mean ± SD for continuous variables and as percentages for discrete variables. Differences between the groups in data for continuous variables were analyzed by Student’s t test, and the data for categorical variables were analyzed by the chi-square test. Data that did not have a normal distribution were expressed as medians (interquartile range). Mann-Whitney’s U test was used to evaluate differences in data between the groups, and Wilcoxon’s signed-rank test was used to analyze differences in data within the same group. Univariate and multivariate regression analyses with the serum absolute change in (Δ) TRLs (TG, RLP-C, and apo B) as the dependent variable were performed to identify patient characteristics and laboratory markers as independent predictors of the ΔTRL levels. Furthermore, we performed multivariate regression analyses with ΔPAI-1 as the dependent variable after adjustments for the conventional cardiovascular risk factors, ΔLDL-C, ΔHDL-C, and lipid-modifying drug use to identify whether ΔFBG and ΔHbA1c might affect the ΔPAI-1. Regression analysis was performed to calculate linear regression coefficients and Pearson correlation coefficients. p Values <0.05 were considered to indicate statistical significance.
Results
Liver dysfunction was noted as adverse reactions in 1 patient in the vildagliptin group. In the control group, pertinent baseline data were not available for 1 patient. After exclusion of these patients, the data of 98 patients each in the 2 groups were subjected to this analysis, and Table 1 lists their baseline characteristics. There were no significant differences between the 2 study groups. Thus, the 2 groups were well matched in terms of the baseline characteristics ( Table 1 ). There were no changes in the diet or other medication use during the study period. Neither group exhibited any significant change of the body weight from the baseline to the 8-weeks follow-up examination vildagliptin group: 70.8 ± 14.2 kg at baseline versus 70.6 ± 14.3 kg at 8 weeks, p = 0.269; and control group: 70.6 ± 16.7 kg at baseline versus 70.8 ± 16.8 kg at 8 weeks, p = 0.192.
| Variable | All Cases (n = 98) | Control (n = 49) | Vildagliptin (n = 49) | p value |
|---|---|---|---|---|
| Male, female | 76 (79%), 20 (21%) | 10 (20%), 39 (80%) | 10 (20%), 39 (80%) | > 0.999 |
| Age (years) | 65 ± 11 | 65 ± 11 | 64 ± 10 | 0.202 |
| Body mass index (kg/m 2 ) | 25.9 ± 4.5 | 25.7 ± 4.3 | 26.1 ± 4.8 | 0.625 |
| Hypertension | 76 (76%) | 41 (84%) | 37 (76%) | 0.316 |
| Lipid disorders | 87 (91%) | 43 (88%) | 44 (90%) | 0.749 |
| Cigarette smoking | 12 (13%) | 6 (12%) | 6 (12%) | > 0.999 |
| Hyperuricemia | 29 (30%) | 18 (37%) | 13 (27%) | 0.277 |
| Coronary artery disease | 49 (51%) | 28 (57%) | 23 (47%) | 0.312 |
| eGFR (ml/min/1.73 m 2 ) | 67.0 (54.4, 77.0) | 63.2 (50.2, 79.0) | 69.2 (62.2, 76.6) | 0.206 |
| Concomitant drugs | ||||
| Antiplatelets | 54 (56%) | 30 (61%) | 26 (53%) | 0.414 |
| ACEIs or ARBs | 52 (54%) | 27 (55%) | 27 (55%) | > 0.999 |
| β blockers | 28 (29%) | 16 (33%) | 14 (29%) | 0.661 |
| Calcium channel blockers | 53 (55%) | 29 (59%) | 26 (53%) | 0.541 |
| Statins | 65 (68%) | 33 (67%) | 32 (65%) | 0.943 |
| Ezetimibe | 2 (2.1%) | 1 (2%) | 1 (2%) | > 0.999 |
| Fibrates | 3 (3.1%) | 2 (4%) | 1 (2%) | 0.558 |
| Sulfonylureas | 33 (34%) | 15 (31%) | 18 (37%) | 0.521 |
| Metformin | 27 (28%) | 13 (27%) | 14 (29%) | 0.821 |
| α glucosidase inhibitors | 25 (26%) | 13 (27%) | 12 (24%) | 0.817 |
| Thiazolidine | 17 (18%) | 8 (16%) | 9 (18%) | 0.790 |
In the vildagliptin group, the serum PAI-1 level was significantly decreased by 16.3 ± 31.3% (p <0.0001; Figure 1 ). The serum levels of TG, RLP-C, and apo B were also significantly decreased by 12.1 ± 31.3% (p = 0.002), 14.2 ± 30.0% (p = 0.001), and 9.5 ± 14.6% (p <0.0001), respectively. In the vildagliptin group, no significant changes of the serum total cholesterol, LDL-C, HDL-C, non–HDL-C, or MDA-LDL levels were noted. The LDL-Rm value was significantly decreased, and LDL-C/apo B, a rough marker of LDL particle size, was significantly increased, indicating an increase in LDL particle size. The FBG, HbA1c, and HOMA-IR were significantly decreased, but the IRI was not changed. No significant changes in any of the aforementioned variables were observed in the control group ( Table 2 ).
| Measurement (unit) | Baseline | 6- month | Change from baseline | p compared baseline | |
|---|---|---|---|---|---|
| PAI-1 (ng/mL) | Control | 26.0 ± 11.0 | 25.3 ± 11.1 | -0.6 ± 9.2 | 0.632 |
| Vildagliptin | 27.4 ± 11.4 | 20.9 ± 7.8 | -6.5 ± 9.1 | < 0.0001 | |
| p = 0.511 | p = 0.025 | p = 0.019 | |||
| TG (mg/dL) | Control | 142 ± 86 | 153 ± 106 | 10.8 ± 64.1 | 0.243 |
| Vildagliptin | 153 ± 89 | 124 ± 59 | -29.0± 58.7 | 0.0012 | |
| p = 0.807 | p = 0.092 | p = 0.0018 | |||
| RLP-C (mg/dL) | Control | 4.5 (3.4, 5.8) | 4.1 (3.1, 7.8) | 0.95 ± 2.53 | 0.252 |
| Vildagliptin | 5.3 (3.8, 6.5) | 4.1 (2.9, 5.6) | -1.21 ± 2.47 | 0.001 | |
| p = 0.297 | p = 0.572 | p = 0.0002 | |||
| apo B (mg/dL) | Control | 87 ± 19 | 83 ± 20 | -3.1 ± 10.1 | 0.268 |
| Vildagliptin | 91 ± 16 | 82 ± 14 | -9.6 ± 12.0 | < 0.0001 | |
| p = 0.259 | p = 0.200 | p = 0.005 | |||
| TC (mg/dL) | Control | 173 ± 32 | 168 ± 31 | -4.6 ± 17.8 | 0.089 |
| Vildagliptin | 174 ± 25 | 167± 24 | -7.2 ± 24.9 | 0.053 | |
| p = 0.745 | p = 0.900 | p = 0.555 | |||
| LDL-C (mg/dL) | Control | 94 ± 24 | 94 ± 25 | -0.3 ± 10.5 | 0.846 |
| Vildagliptin | 94 ± 20 | 91 ± 17 | -3.2 ± 18.8 | 0.236 | |
| p = 0.949 | p = 0.468 | p = 0.350 | |||
| HDL-C (mg/dL) | Control | 49 ± 12 | 47 ± 12 | -0.10 ± 7.9 | 0.298 |
| Vildagliptin | 51 ± 14 | 49 ± 14 | -2.2 ± 7.7 | 0.052 | |
| p = 0.356 | p = 0.633 | p = 0.387 | |||
| non HDL-C (mg/dL) | Control | 123 ± 29 | 120 ± 31 | -3.4 ± 15.3 | 0.094 |
| Vildagliptin | 122 ± 24 | 120 ± 21 | -4.9 ± 22.4 | 0.138 | |
| p = 0.979 | p = 0.878 | p = 0.785 | |||
| MDA-LDL (U/L) | Control | 124 ± 41 | 123 ± 42 | -1.6 ± 33.2 | 0.750 |
| Vildagliptin | 120 ± 37 | 120 ± 38 | -0.8 ± 26.4 | 0.849 | |
| p = 0.824 | p = 0.712 | p = 0.908 | |||
| LDL-C/apoB | Control | 1.081 ± 0.130 | 1.123 ± 0.170 | 0.042 ± 0.130 | 0.058 |
| Vildagliptin | 1.035 ± 0.177 | 1.115 ± 0.121 | 0.079 ± 0.135 | < 0.001 | |
| p = 0.151 | p = 0.784 | p = 0.169 | |||
| LDL-Rm value | Control | 0.366 ± 0.035 | 0.356 ± 0.035 | -0.010 ± 0.035 | 0.056 |
| Vildagliptin | 0.363± 0.024 | 0.348 ± 0.027 | -0.016 ± 0.025 | < 0.0001 | |
| p = 0.715 | p = 0.234 | p = 0.334 | |||
| FBG (mg/dL) | Control | 159 ± 67 | 136 (117, 176) | 1.0 ± 48 | 0.864 |
| Vildagliptin | 152 ± 49 | 119 (106, 140) | -25 ± 43 | < 0.0001 | |
| p = 0.573 | p = 0.002 | p = 0.005 | |||
| HbA 1c (%) | Control | 7.39 ± 1.41 | 7.44 ± 1.53 | 0.05 ± 0.44 | 0.441 |
| Vildagliptin | 7.46 ± 0.80 | 6.84 ± 0.77 | -0.68 ± 0.49 | < 0.0001 | |
| p = 0.741 | p = 0.016 | p < 0.0001 | |||
| IRI (mIU/mL) | Control | 10.9 (6.0, 21.5) | 9.1 (6.1, 17,3) | -0.3 (-7.0, 3.0) | 0.499 |
| Vildagliptin | 8.6 (5.6, 12.3) | 7.7 (5.5, 15.3) | -0.1 (-1.4, 1.7) | 0.866 | |
| p = 0.229 | p = 0.212 | p = 0.689 | |||
| HOMA-IR | Control | 3.29 (1.93, 6.46) | 3.09 (1.98, 6.94) | -0.005 (-2.95, 1.43) | 0.554 |
| Vildagliptin | 2.83 (1.64, 5.91) | 2.26 (1.49, 4.17) | -0.31 (-1.39, 0.25) | 0.049 | |
| p = 0.330 | p = 0.057 | p = 0.368 |
To identify independent predictors of the ΔTRLs levels, we conducted a univariate linear regression analysis to determine the significance of associations with patient characteristics and laboratory markers, and the variables that were found to be related to the ΔTRL levels at a p value <0.1 in the univariate regression analysis were entered into the multivariate model. We thus created a multivariate regression analysis model with the changes in 3 TRLs (ΔTG, ΔRLP-C, and Δapo B) serving as dependent variables. Both ΔPAI-1 and ΔHDL-C were independent predictors of the ΔTG. The results showed that the presence of hypertension was significantly associated with the increase in the serum TG level. Smoking and hypertension were significantly associated with the increase in the serum RLP-C level. ΔPAI-1 was an independent predictor of the ΔRLP-C. ΔPAI-1 and ΔLDL-C were independent predictors of the Δapo B ( Table 3 ). The associations of ΔTG, ΔRLP-C, and ΔapoB with the ΔPAI are shown in Figure 2 .
| Independent variables | ΔTG | ΔRLP-C | ΔapoB | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| r | p | β | p | r | p | β | p | r | p | β | p | |
| Age | 0.109 | 0.292 | 0.014 | 0.894 | 0.007 | 0.946 | ||||||
| Male | -0.138 | 0.183 | -0.121 | 0.245 | -0.058 | 0.577 | ||||||
| BMI | 0.139 | 0.176 | 0.172 | 0.096 | 0.117 | 0.214 | 0.007 | 0.950 | ||||
| Smoking | 0.084 | 0.417 | 0.219 | 0.033 | 0.283 | 0.002 | 0.060 | 0.561 | ||||
| Hypertension | 0.371 | 0.0002 | 0.309 | 0.001 | 0.294 | 0.004 | 0.225 | 0.019 | 0.128 | 0.216 | ||
| Dyslipidemia | 0.044 | 0.668 | 0.081 | 0.434 | 0.064 | 0.536 | ||||||
| ΔFBG | 0.053 | 0.606 | 0.150 | 0.147 | 0.04 | 0.703 | ||||||
| ΔHb A1c | 0.172 | 0.095 | 0.101 | 0.274 | 0.233 | 0.023 | 0.124 | 0.177 | 0.06 | 0.562 | ||
| ΔLDL-C | 0.044 | 0.672 | 0.138 | 0.198 | 0.666 | < 0.0001 | 0.637 | < 0.0001 | ||||
| ΔHDL-C | -0.179 | 0.080 | -0.211 | 0.022 | -0.018 | 0.864 | 0.043 | 0.679 | ||||
| ΔPAI-1 | 0.306 | 0.0024 | 0.270 | 0.005 | 0.358 | 0.0004 | 0.336 | 0.0004 | 0.292 | 0.004 | 0.200 | 0.009 |
| Lipid-modifying drug use | 0.004 | 0.966 | 0.103 | 0.321 | 0.049 | 0.637 | ||||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
