Sage Journals: Discover world-class research

Abstract

We investigated the effects of 12-month exercise training on hypercoagulability in patients with combined type 2 diabetes mellitus and coronary artery disease. Associations with severity of disease were further explored. Patients (n = 131) were randomized to exercise training or a control group. Blood was collected at inclusion and after 12 months. Tissue factor, free and total tissue factor pathway inhibitor, prothrombin fragment 1 + 2 (F1 + 2) and D-dimer were determined by enzyme-linked immunosorbent assay and ex vivo thrombin generation by the calibrated automated thrombogram assay. Tissue factor and ex vivo thrombin generation increased from baseline to 12 months (p < 0.01, all), with no significant differences in changes between groups. At baseline, free and total tissue factor pathway inhibitor significantly correlated to fasting glucose (p < 0.01, both) and HbA1c (p < 0.05, both). In patients with albuminuria (n = 34), these correlations were strengthened, and elevated levels of D-dimer, free and total tissue factor pathway inhibitor (p < 0.01, all) and decreased ex vivo thrombin generation (p < 0.05, all) were observed. These results show no effects of exercise training on markers of hypercoagulability in our population with combined type 2 diabetes mellitus and coronary artery disease. The association between poor glycaemic control and tissue factor pathway inhibitor might indicate increased endothelial activation. More pronounced hypercoagulability and increased tissue factor pathway inhibitor were demonstrated in patients with albuminuria.

Keywords

Type 2 diabetes mellitus coronary artery disease exercise training hypercoagulability tissue factor pathway inhibitor albuminuria

Introduction

Patients with type 2 diabetes mellitus (T2DM) are at significant risk of developing cardiovascular disease (CVD) and having diabetes has been proposed as an equivalent to coronary artery disease (CAD) regarding risk for future cardiovascular events.¹ Endothelial cell activation, with change of phenotype to be proinflammatory and procoagulant, is thought to be the initial step in the pathogenesis of atherosclerosis and plaque development, the most important underlying condition of CAD.^2,3 By plaque rupture or erosion, collagen and sub endothelial tissue factor (TF) are exposed to circulating blood, thus triggering the coagulation cascade with subsequent thrombin formation, which may lead to acute manifestations like acute coronary syndrome.²

Patients with combined T2DM and CAD have more extensive atherosclerosis and thrombotic burden,⁴ and hyperglycaemia is supposed to be one of the driving forces in this process. Also insulin resistance, typically present in T2DM, triggers atherothrombosis by inducing endothelial dysfunction, and hyperinsulinaemia has further been shown to increase TF procoagulant activity in T2DM patients, with subsequent accelerated thrombin generation.^5–8 Albuminuria characterizes the development of kidney disease in diabetic patients and these patients are known to be hypercoagulable and often die of CVD complications.^9,10

Regular physical exercise is well established in the management of both T2DM and CAD.^11,12 In addition to improvement of conventional risk factors and glycaemic control, exercise training may have a favourable effect on the delicate balance of coagulation and fibrinolysis.¹³

The calibrated automated thrombogram (CAT) is an overall function test of the haemostatic system that is sensitive to prothrombotic changes in plasma.^14,15 The most used parameter, the endogenous thrombin potential (ETP), represents the total ex vivo thrombin activity and reflects the interplay between all relevant proteases and inhibitors in plasma. Furthermore, soluble tissue factor (sTF), free and total tissue factor pathway inhibitor (TFPI) and the coagulation activation markers prothrombin fragment 1 + 2 (F1 + 2) and D-dimer are useful tools to estimate the degree of in vivo hypercoagulability.

The aim of this study was to investigate the effects of 1 year of organized physical exercise on the degree of hypercoagulability, assessed by the above-mentioned parameters representative for both in vivo and ex vivo status, in patients with combined T2DM and CAD with the hypothesis that exercise has a favourable effect on the prothrombotic state in these high-risk patients. In addition, any associations between the haemostatic and glucometabolic variables and severity of disease were explored.

Methods

Study population

The study population (n = 137) consisted of patients with angiographically verified stable CAD and known T2DM. Subjects were enrolled at Department of Cardiology, Oslo University Hospital, Ullevål, Norway, between August 2010 and March 2012. The main study was a randomized clinical trial investigating the effects of 12 months physical exercise on progression of atherosclerosis and glucometabolic control.^16,17 Exclusion criteria included severe diabetic complications, cancer, stroke or acute coronary syndrome within the last 3 months prior to inclusion, unstable angina, uncompensated heart failure, serious arrhythmia, serious valvular disease, serious rheumatological diseases, chronic obstructive pulmonary disease stadium GOLD IV, ongoing infections, thromboembolic disease, severe musculoskeletal disorders and other illnesses seriously limiting the ability for participating in physical activity. At inclusion, the subjects were randomized 1:1 to a combination of endurance and strength training or a control group with conventional follow-up by their general practitioner.

The regional Ethics Committee has approved the study, and all patients have given written informed consent to participate. The study was conducted in accordance with the Declaration of Helsinki and is registered at http://www.clinicaltrials.gov, NCT01232608.

Physical exercise intervention

The physical exercise programme was developed and conducted in collaboration with the Norwegian school of Sport Sciences. Details on the exercise programme have been previously described.¹⁶ Briefly, the exercise group had two group-based sessions per week with qualified instructors, and a third home-based session, for a total volume of at least 150 min/week. About two-thirds were aerobic, and one-third resistance training and each of the supervised sessions included parts with high-intensity interval training.

Laboratory methods

Venous blood samples were drawn in fasting condition by standard venipuncture (21-gauge needle) between 08:00 and 10:00 AM, before intake of morning medication, at inclusion and after 1 year. Blood collected in Vacutainer tubes, containing 0.129 M trisodium citrate in dilution 1:10, was kept on ice until centrifugation at 3000g for 20 min at 4°C within 1 h. Plasma was stored frozen at −80°C until analysed. Routine blood samples, including HbA_1c and insulin, were determined by conventional methods. To estimate insulin resistance, the updated homeostatic model assessment 2 of insulin resistance (HOMA2-IR) was applied.

Levels of the coagulation markers sTF, TFPI, F1 + 2 and D-dimer were determined by conventional enzyme immunoassays; Imubind TF kit recognizing TF-apolipoprotein, sTF and TF-VII complexes (American Diagnostic Inc., Greenwich, CT, USA), Asserachrom^® free and total TFPI antigen, recognizing the full-length TFPI molecules, and the full-length and truncated TFPI molecules including TFPI bound to lipoproteins, respectively (Stago Diagnostica, Asniere, France), Enzygnost^® F1 + 2 (monoclonal) (Siemens, Marburg, Germany) and Asserachrom^® D-dimer (Stago Diagnostica), respectively. Inter-assay coefficients of variations (CVs) were 10.5%, 9.0%, 6.5%, 10.0% and 7.2%, respectively.

Ex vivo potential to generate thrombin was analysed by the CAT assay, conducted according to the manufacturer’s instruction (Thrombinoscope BV, Maastricht, The Netherlands). The method is described in detail elsewhere.¹⁴ Briefly, citrated platelet poor plasma (PPP) was mixed with recombinant relipidated TF and phospholipids, with the final concentrations of 5 pM and 4 µM, respectively. Every experiment was performed with two separate readings, one for thrombin generation and a second, to which an inert thrombin calibrator with known concentration is added, to enable calculation of the results. A freshly made starting reagent containing CaCl₂ (100 mM) and a thrombin-specific fluorogenic substrate (Z-Gly-Gly-Arg-AMC) (2.5 mM) in Hepes buffer were automatically added. The fluorescence intensity was recorded by the Fluoroskan Ascent^® microplate fluorimeter (Thermo Fisher Scientific Oy, Vantaa, Finland). Reagents were provided by Thrombinoscope BV. The software (Thrombinoscope BV, version 3.0.0.29) enabled calculation of the thrombogram parameters, which include the lag time (LT, min), time to peak (ttPeak, min), peak height (PeakH, nM), the ETP (nM/min) and the slope of thrombin formation, named velocity index (VI). The respective inter assay CVs were 14.6%, 10.2%, 12.4%, 8.5% and 23.2%.

Statistical methods

Continuous data were mostly skewed and the data are presented as median values with 25th and 75th percentiles. Categorical data are presented as number and proportions. Spearman rank test was used for correlation analyses. Adjusted correlations were calculated by partial rank correlations. Mann–Whitney U test, independent samples T-test, Kruskal–Wallis test and chi-square test were applied for group comparisons, as appropriate. Wilcoxon’s test was used to explore within group changes, and for differences in changes between the randomized groups, Mann–Whitney U test was applied. p-values ≤0.05 were considered statistically significant. The statistical analyses were performed by IBM^© SPSS^© statistics for windows, v 21.0 (IBM Corp., New York, NY, USA).

Results

Of the total 137 patients included in the main study, 6 patients were on warfarin and have been excluded. Baseline characteristics of the population (n = 131) are presented in Table 1. Median age was 64 years, 83% were men, 57 patients (44%) presented with previous acute myocardial infarction (AMI) and 32 (25%) with albuminuria. A total of 14 patients did not complete the study and 9 participants were excluded due to low adherence to exercise intervention. These patients had an average exercise adherence of <40%. For the 108 patients who were analysed for the intervention effects, the average adherence was 67% (55%, 83%) and 21 patients had an average of >70%.¹⁶ Clinical data did not differ between the randomized groups (Table 1).

Table 1.

Baseline characteristics of the total population (n = 131) and of the cohort completing the study according to the randomized groups (n = 108).

	All (n = 131)	Control (n = 59)	Exercise (n = 49)	p
Age (years)	64 (58, 69)	63 (60, 69)	65 (58, 70)	0.32
Men/women, n (%)	108/22 (83/17)	47/12 (80/20)	42/7 (86/14)	0.56
BMI (kg/m²)^a	29.1 ± 4.9	29.1 ± 5.6	28.9 ± 4.0	0.82
Waist circumference (cm)^a	106 ± 12	106 ± 14	105 ± 11	0.79
SBP (mmHg)	138 (127, 150)	139 (124, 150)	136 (129, 151)	0.99
DBP (mmHg)	78 (71, 86)	80 (71, 86)	76 (71, 82)	0.17
Previous AMI, n (%)	57 (44)	28 (48)	19 (39)	0.30
Hypertension, n (%)	94 (72)	44 (75)	36 (74)	1.00
Current smokers, n (%)	22 (17)	8 (14)	9 (18)	0.88
Albuminuria, n (%)	32 (25)	12 (20)	13 (27)	0.48
Years of diabetes	9 (6, 15)	8 (5, 9)	11 (5, 15)	0.29
Years since AMI	3 (1, 11)	3 (1, 11)	3 (1, 11)	0.92
Total cholesterol (mmol/L)	3.9 (3.4, 4.5)	4.0 (3.3, 4.8)	3.8 (3.4, 4.4)	0.38
HDL cholesterol (mmol/L)	1.12 (0.94, 1.39)	1.13 (0.93, 1.14)	1.13 (0.99, 1.40)	0.86
Triglycerides (mmol/L)	1.42 (1.06, 1.92)	1.36 (0.94, 1.88)	1.44 (1.09, 1.91)	0.62
Fasting glucose (mmol/L)	8.1 (6.9, 9.9)	8.0 (6.7, 9.7)	7.6 (6.8, 9.9)	0.64
HbA1c (%)	7.4 (6.8, 8.3)	7.3 (6.7, 8.0)	7.4 (6.8, 8.5)	0.62
Insulin (pmol/L)	58 (33, 102)	65 (32, 109)	50 (30, 86)	0.17
HOMA2-IR	1.3 (0.7, 2.1)	1.3 (0.7, 2.3)	1.1 (0.7, 1.8)	0.30
Creatinine (µmol/L)	78 (68, 91)	81 (66, 89)	75 (69, 92)	0.90
Medication %
Statins	94	93	96	1.00
Aspirin	92	97	88	0.08
Clopidogrel	23	22	22	0.96
Insulin/insulin analogue	19	20	18	1.00
Metformin	73	73	78	0.91
Gliptins	12	10	18	0.31
Sulphonylurea	35	41	27	0.16

BMI: body mass index; SBP/DBP: systolic/diastolic blood pressure; AMI: acute myocardial infarction; HDL: high-density lipoprotein; LDL: low-density lipoprotein; HOMA2-IR: homeostatic model assessment 2 of insulin resistance; SD: standard deviation.

p-values refer to differences between the randomized groups.

Mean ± SD.

Effects of exercise training

After 12-month intervention, an increase in TF (p = 0.004) and ex vivo thrombin generation, the latter expressed by increased PeakH (p = 0.002) and VI (p = 0.001), and a reduction in ttPeak (p = 0.006) were observed in the exercise group. Increased ex vivo thrombin generation was also observed in the control group, shown by PeakH (p = 0.002), VI (p < 0.001) and a reduced ttPeak (p = 0.001). There were, however, no significant differences in changes between the groups (Table 2), also when baseline levels were taken into account for free and total TFPI which differed between the groups at baseline (p = 0.003 and p = 0.036, respectively).

Table 2.

Markers of procoagulant activity at baseline and after 12 months in the randomized groups.

	Exercise (n = 49)		p1	Controls (n = 59)		p2	Δp
	Baseline	12 months	p1	Baseline	12 months	p2	Δp
TF (pg/mL)	277 (224, 319)	309 (246, 386)	0.004	244 (203, 356)	288 (210, 391)	0.057	0.279
TFPI (free) (ng/mL)	18.9* (13.8, 21.7)	18.4 (14.1, 20.9)	0.943	13.6 (9.7, 17.7)	14.5 (11.4, 19.0)	0.098	0.234^a
TFPI (total) (ng/mL)	50.1* (42.0, 58.9)	50.5 (44.2, 59.9)	0.372	43.8 (37.7, 53.8)	45.4 (34.9, 51.2)	0.380	0.169^a
F1 + 2 (pmol/L)	251 (204, 320)	250 (205, 334)	0.870	255 (210, 312)	261 (214, 328)	0.296	0.856
D-dimer (ng/mL)	393 (266, 607)	433 (308, 622)	0.325	364 (271, 493)	375 (269, 528)	0.815	0.268
Lag time (min)	2.50 (2.08, 2.79)	2.34 (2.00, 2.88)	0.908	2.50 (2.16, 2.83)	2.41 (2.13, 2.80)	0.474	0.337
ETP (nM/min)	1185 (1090, 1297)	1217 (1126, 1311)	0.521	1246 (1158, 1355)	1279 (1206, 1382)	0.472	0.695
Peak height (nM)	200.6 (168.8, 254.9)	224.5 (191.2, 264.5)	0.002	223.6 (188.6, 247.6)	239.9 (210.4, 269.8)	0.002	0.746
ttPeak (min)	5.76 (5.11, 6.67)	5.42 (4.55, 6.46)	0.006	5.83 (4.83, 6.50)	5.33 (4.67, 6.33)	0.001	0.617
Velocity index (nM/min)	63 (42, 89)	77 (55, 103)	0.001	73 (48, 88)	81 (62, 111)	<0.001	0.865

TF: tissue factor; TFPI: tissue factor pathway inhibitor; ETP: endogenous thrombin potential.

p1 and p2 refer to differences from baseline to 12 months in the exercise and control groups, respectively. Δp refers to difference in changes between groups from baseline to 12 months (Mann–Whitney U).

Relative changes between groups.

Refers to differences between groups at baseline (p = 0.003 and p = 0.036, respectively).

Bold entries indicate exact p-values are given.

Excluding patients with albuminuria (n = 25) did not have any impact on the results (data not shown). There were also no differences in changes between groups when analysed separately for patients with previous myocardial infarction (MI) (n = 47) or not (data not shown).

There was no difference in changes during the intervention period between the groups in HbA1c levels, although a numerical reduction was observed in the exercise group [7.3 (6.8, 8.0)% to 7.4 (6.5, 8.2)% in controls vs 7.4 (6.8, 8.4)% to 7.2 (6.6, 7.8)% in the exercise group] (p = 0.24, for difference in changes between groups). There were also no significant differences in changes between the randomized groups in weight, waist circumference, energy intake of main nutrients or diabetes medication during the study period (data not shown).

Baseline associations

In the total population (n = 131), free and total TFPI were significantly correlated to fasting glucose (p < 0.01, both) and to HbA1c (p < 0.05, both) (Table 3).

Table 3.

Baseline correlations (Spearman’s rho) between markers of procoagulant activity and glucometabolic variables in the total population (n = 131).

		HbA1c (%)	Glucose (mmol/L)	HOMA2-IR	Insulin (pmol/L)	Years of diabetes
sTF (pg/mL)	r	−0.081	−0.050	−0.087	−0.098	0.147
	p	0.362	0.576	0.354	0.275	0.104
TFPI (free) (ng/mL)	r	0.195	0.239	0.070	0.057	0.068
	p	0.027	0.007	0.455	0.526	0.449
TFPI (total) (ng/mL)	r	0.178	0.261	0.040	−0.015	0.102
	p	0.044	0.003	0.666	0.863	0.256
F1 + 2 (pmol/L)	r	−0.126	−0.021	−0.157	−0.156	0.024
	p	0.158	0.817	0.090	0.078	0.787
D-dimer (ng/mL)	r	−0.111	−0.074	−0.028	−0.038	0.062
	p	0.211	0.413	0.764	0.668	0.494
LT (min)	r	−0.026	0.106	0.179	0.138	−0.123
	p	0.775	0.236	0.054	0.122	0.171
ETP (nM/min)	r	0.001	0.001	0.012	0.090	−0.192
	p	0.995	0.989	0.898	0.312	0.032 ^a
ttPeak (min)	r	−0.016	0.139	0.120	0.111	−0.195
	p	0.859	0.121	0.197	0.213	0.029 ^a
PeakH (nM)	r	0.020	−0.058	−0.003	0.040	0.023
	p	0.821	0.521	0.976	0.656	0.798
VI (nM/min)	r	0.021	−0.083	−0.014	0.003	0.148
	p	0.815	0.355	0.879	0.976	0.099

HOMA2-IR: homeostatic model assessment 2 of insulin resistance; sTF: soluble tissue factor; TFPI: tissue factor pathway inhibitor; LT: lag time; ETP: endogenous thrombin potential; VI: velocity index.

Adjusted for age and HbA1c: p = 0.60 and p = 0.26, respectively.

Bold entries indicate exact p-values are given.

Years of diabetes was inversely associated with ETP and ttPeak (p < 0.05, both) (Table 3); however, no longer significant after correction for age and HbA1c (p = 0.60 and p = 0.26, respectively). No other significant correlations between glucometabolic variables and the coagulation markers were observed. Furthermore, the levels of free and total TFPI were significantly inter-correlated to TF (r = 0.207 and r = 0.296, respectively, p < 0.05, both).

In the subgroup of patients having albuminuria (n = 32), the correlations between free and total TFPI and fasting glucose and HbA1c were strengthened (r = 0.476 and r = 0.503, respectively, p < 0.01, both) and (r = 0.352 and r = 0.444, respectively, p < 0.05, both).

These patients, as compared to those without albuminuria, had elevated levels of D-dimer, free and total TFPI (p < 0.05, all) (Figure 1, upper panel) in addition to delayed, slower and lower ex vivo thrombin generation, expressed by LT, ttPeak, VI and PeakH (p < 0.01, all) (Figure 1, lower panel). In patients with previous MI, no differences in any of the markers were observed (data not shown).

Figure 1.

Levels of in vivo haemostatic variables (upper panel) and ex vivo thrombin generation (lower panel) according to patients with albuminuria (n = 32) (white column) and without (black column). Values refer to medians and 25th and 75th percentiles.

Discussion

In this randomized study, 1 year of organized physical exercise in patients with combined T2DM and CAD did not result in significant changes different from controls in the selected markers of procoagulant activity, also not when excluding patients with more severe disease states, that is, albuminuria or previous MI. In both groups, increased levels of TF and ex vivo thrombin generation were recorded at the end of the study.

At baseline, poor glycaemic control, assessed by HbA1c and fasting glucose, was significantly associated with levels of TFPI (free and total). These associations were strengthened when analysed in the subgroup of patients presenting with albuminuria, in which also the levels of TFPI and D-dimer were elevated and the ex vivo thrombin generation was decreased.

To the best of our knowledge, long-term effect of exercise training on markers of hypercoagulability in patients with combined CAD and T2DM has previously not been reported on. We observed, however, no differences in changes between the randomized groups in any of the measured variables, although our hypothesis was to achieve a shift towards a less prothrombotic state in the exercise group. This was also our hypothesis with regard to TFPI, which might be discussed to reflect endothelial activation.¹⁸

Compliance to the study intervention for a period of 1 year is a challenge in our patient population. The mean adherence to the exercise intervention was 67%¹⁶ which might be discussed not to be sufficient to achieve an effect on procoagulant activity. However, 21 of these patients obtained an adherence of >70%, assuming in total the compliance to be satisfactory.

In a study on patients with T2DM alone, randomized to only aerobic exercise or not, significantly reduced concentration of tissue plasminogen activator (tPA) antigen was shown, which could reflect reduced thrombotic potential as well as improved endothelial activation.¹⁹ In a group of overweight men supposed to have prothrombotic traits similar to T2DM, a 3-month period with daily endurance exercise training resulted in significantly reduced levels of ETP in the exercise group compared to controls, also independent of exercise dose.²⁰ This is somewhat in contrast to our results; however, CAD patients were not included in that particular study. Also in post-MI patients, it has been shown that exercise training results in reduced procoagulant activity, assessed by activated prothrombin time (aPTT).²¹

The combination of endurance and resistance exercise in our study was originally chosen in order to have good impact on glycaemic control. We did, however, obtain limited changes in HbA1c and also no weight change during the intervention, and it might be discussed if this combination is the best modality for dampening hypercoagulability. However, Ponjee et al.²² who investigated prolonged effects of endurance exercise alone in sedentary subjects could also not demonstrate any changes in levels of von Willebrand factor, an endothelial marker, or in thrombin–antithrombin complex concentration.

The increase in levels of TF, the main activator of the extrinsic pathway of coagulation, shown in our study, might reflect a raise in prothrombotic potential, although in vivo coagulation per se might not be activated. These findings are strengthened by the increased and accelerated ex vivo thrombin generation. Why these changes occur also in controls is not easily explainable. Elevated levels of circulating TF have previously been demonstrated in both diabetes mellitus (DM) and CAD.^23,24

Duration of diabetes was shown to be inversely related to ex vivo prothrombotic activity in univariate analyses; however, the correlation was no longer significant after adjusting for age and glycaemic control.

Diabetes patients who have developed albuminuria are supposed to have more severe and complicated disease state. We could, however, not find any effect on the measured markers also when those with albuminuria were excluded. We could, nevertheless, show that in this subgroup of patients, the coagulation system was activated, demonstrated by increased D-dimer levels. Furthermore, both free and total TFPI were elevated and the baseline correlations between the prothrombotic markers, especially TFPI, were strengthened, indicating that these patients have increased endothelial activation. Wakabayashi and Masuda²⁵ also showed D-dimer to be related to microalbuminuria, which is in accordance with our findings, suggesting a link between endothelial dysfunction, activation of the coagulation system and diabetic nephropathy. The reduced ex vivo potential for thrombin generation found in our study might be discussed to be due to ongoing in vivo coagulation, resulting in lower potential for ex vivo generation.

In contrast to our finding, enhanced ex vivo thrombin generation in T2DM patients with macrovascular disease, of which 46% of the total population (n = 160) had CVD, has been reported, and also significantly elevated ETP in patients with albuminuria, as compared to those without.²⁶ In this study, patients with previous MI did not present with higher procoagulant activity. When analysed separately, no benefit of exercise training was obtained in this subpopulation. Thus, it seems that coronary heart disease severity did not influence the results.

The significant correlations shown at baseline between levels of TFPI and HbA1c and fasting glucose, but not other glucometabolic variables, indicate that hyperglycaemia rather than insulin resistance causes the raise in TFPI. Our results are in line with previous reports showing a positive correlation between HbA1c and plasma activity levels of TFPI,^27,28 and a decrease in TFPI levels after optimization of glycaemic control has also been demonstrated.²⁹ Several studies have demonstrated increased levels of TFPI in diabetic patients compared to healthy controls.^30,31 The significant interrelation between TF and TFPI we observed might indicate a compensating mechanism in order to attenuate the TF-dependent pathway of coagulation. This might also be discussed along with the lack of associations between glucometabolic control and variables of thrombin generation. This is supported by the report on insulin-dependent DM patients, in which TFPI was found to dampen the hypercoagulable state, demonstrated by inhibition of factor Xa activity, however not sufficient to inhibit in vivo thrombin generation, shown by increased levels of F1 + 2.³⁰ In another study, accelerated thrombin generation was demonstrated in patients with diabetes as compared to healthy controls, with the most pronounced increase in the poorly controlled group.³¹

Study limitations

The mean adherence to the exercise intervention in our study was 67%¹⁶ which might have been too liberal in order to achieve an effect on procoagulant activity. In addition, study participants being engaged in physical exercise at the time of inclusion were not excluded, but their training programme should not exceed the one in the study. The main study was powered for between-group difference in HbA1c¹⁴ and not for change in markers of hypercoagulability. The study participants were well controlled for their diabetes.

Conclusion

We could not demonstrate any effects of exercise training on markers of procoagulant activity in our population of patients with combined T2DM and CAD. However, both groups experienced increased coagulation potential after 12 months, which might be discussed along with progression of diseases. The significant association between poor glycaemic control and levels of TFPI observed might be indicative of increased endothelial activation. A more pronounced in vivo procoagulant state, as well as increased TFPI, was demonstrated in a subgroup of patients presenting with albuminuria.

Footnotes

Acknowledgements

The authors thank Sissel Åkra and Beate Vestad for valuable laboratory assistance to the study.

Authors’ contribution

H.A., S.S., R.B., I.S. and V.B. contributed to study conception and design; V.B., R.B, I.U.N. and I.S. contributed to execution of experiments and data analyses; V.B., H.A., S.S. and I.S. contributed to result interpretation; V.B. and I.S. contributed to drafting the manuscript; V.B., R.B., I.U.N., S.S., H.A. and I.S. contributed to manuscript revision and editing; and V.B. and I.S. contributed to manuscript approval.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

This work was supported by Stein Erik Hagen Foundation for Clinical Heart Research and ‘Ada og Hagbart Waages Humanitære og veldedige stiftelse’, Oslo, Norway.

References

Juutilainen

Lehto

Ronnemaa

. Type 2 diabetes as a ‘coronary heart disease equivalent’: an 18-year prospective population-based study in Finnish subjects. Diabetes Care 2005; 28: 2901–2907.

Ross

Atherosclerosis – an inflammatory disease. N Engl J Med 1999; 340: 115–126.

Libby

Ridker

Maseri

Inflammation and atherosclerosis. Circulation 2002; 105: 1135–1143.

Otel

Ledru

Danchin

Ischemic heart disease in type 2 diabetes. Metabolis 2003; 52(Suppl. 1): 6–12.

Paneni

Beckman

Creager

. Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy: part I. Eur Heart J 2013; 34: 2436–2443.

Beijers

Ferreira

Spronk

. Impaired glucose metabolism and type 2 diabetes are associated with hypercoagulability: potential role of central adiposity and low-grade inflammation – the Hoorn Study. Thromb Res 2012; 129: 557–562.

Grant

PJ.

Diabetes mellitus as a prothrombotic condition. J Intern Med 2007; 262: 157–172.

Carr

ME.

Diabetes mellitus: a hypercoagulable state. J Diabetes Complications 2001; 15: 44–54.

Domingueti

Dusse

Carvalho

. Hypercoagulability and cardiovascular disease in diabetic nephropathy. Clin Chim Acta 2013; 415: 279–285.

10.

Domingueti

Foscolo

Reis

. Association of haemostatic and inflammatory biomarkers with nephropathy in type 1 diabetes mellitus. J Diabetes Res 2016; 2016: 2315260.

11.

Colberg

Albright

Blissmer

. Exercise and type 2 diabetes: American College of Sports Medicine and the American Diabetes Association: joint position statement. Exercise and type 2 diabetes. Med Sci Sports Exerc 2010; 42: 2282–2303.

12.

Marwick

Hordern

Miller

. Exercise training for type 2 diabetes mellitus: impact on cardiovascular risk: a scientific statement from the American Heart Association. Circulation 2009; 25: 3244–3262.

13.

Womack

Nagelkirk

Coughlin

AM.

Exercise-induced changes in coagulation and fibrinolysis in healthy populations and patients with cardiovascular disease. Sports Med 2003; 33: 795–807.

14.

Hemker

Al Dieri

De Smedt

. Thrombin generation, a function test of the haemostatic-thrombotic system. Thromb Haemost 2006; 96: 553–561.

15.

Orbe

Zudaire

Serrano

. Increased thrombin generation after acute versus chronic coronary disease as assessed by the thrombin generation test. Thromb Haemost 2008; 99: 382–387.

16.

Byrkjeland

Njerve

Anderssen

. Effects of exercise training on HbA1c and VO2peak in patients with type 2 diabetes and coronary artery disease: a randomised clinical trial. Diab Vasc Dis Res 2015; 12: 325–333.

17.

Byrkjeland

Stensaeth

Anderssen

. Effects of exercise training on carotid intima-media thickness in patients with type 2 diabetes and coronary artery disease. Influence of carotid plaques. Cardiovasc Diabetol 2016; 15: 13.

18.

Kato

Regulation of functions of vascular wall cells by tissue factor pathway inhibitor: basic and clinical aspects. Arterioscler Thromb Vasc Biol 2002; 22: 539–548.

19.

Dunstan

Mori

Puddey

. A randomised, controlled study of the effects of aerobic exercise and dietary fish on coagulation and fibrinolytic factors in type 2 diabetics. Thromb Haemost 1999; 81: 367–372.

20.

Gram

Bladbjerg

Skov

. Three months of strictly controlled daily endurance exercise reduces thrombin generation and fibrinolytic risk markers in younger moderately overweight men. Eur J Appl Physiol 2015; 115: 1331–1338.

21.

Suzuki

Yamauchi

Yamada

. Blood coagulability and fibrinolytic activity before and after physical training during the recovery phase of acute myocardial infarction. Clin Cardiol 1992; 15: 358–364.

22.

Ponjee

Janssen

van Wersch

JW.

Prolonged endurance exercise and blood coagulation: a 9 month prospective study. Blood Coagul Fibrinolysis 1993; 4: 21–25.

23.

Boden

Rao

AK.

Effects of hyperglycemia and hyperinsulinemia on the tissue factor pathway of blood coagulation. Curr Diab Rep 2007; 7: 223–227.

24.

Singh

Boden

Rao

AK.

Tissue factor and Toll-like receptor (TLR)4 in hyperglycaemia-hyperinsulinaemia. Effects in healthy subjects, and type 1 and type 2 diabetes mellitus. Thromb Haemost 2015; 113: 750–758.

25.

Wakabayashi

Masuda

Association of D-dimer with microalbuminuria in patients with type 2 diabetes mellitus. J Thromb Thrombolysis 2009; 27: 29–35.

26.

Hoellerl

. Thrombin generation in type 2 diabetes with albuminuria and macrovascular disease. Eur J Clin Invest 2012; 42: 470–477.

27.

El-Ghoroury

El-Din

Abdel-Kader

. Study of factor VII, tissue factor pathway inhibitor and monocyte tissue factor in noninsulin-dependent diabetes mellitus. Blood Coagul Fibrinolysis 2008; 19: 7–13.

28.

Rigla

Mateo

Fontcuberta

. Normalisation of tissue factor pathway inhibitor activity after glycaemic control optimisation in type 1 diabetic patients. Thromb Haemost 2000; 84: 223–227.

29.

Leurs

van Oerle

Wolffenbuttel

. Increased tissue factor pathway inhibitor (TFPI) and coagulation in patients with insulin-dependent diabetes mellitus. Thromb Haemost 1997; 77: 472–476.

30.

Yokoyama

Myrup

Rossing

. Increased tissue factor pathway inhibitor activity in IDDM patients with nephropathy. Diabetes Care 1996; 19: 441–445.

31.

Aoki

Shimoyama