Tissue Plasminogen Activator, Fibrin D-Dimer, and Insulin Resistance in the Relatives of Patients With Premature Coronary Artery Disease

Methods

One hundred twenty-five male patients (probands) ≤65 years old at the time of diagnostic coronary angiography and with confirmed 2- or 3-vessel CAD (World Health Organization criteria of ≥50% stenosis in a major epicardial vessel) were identified via the surgical revascularization waiting list at the Yorkshire Heart Center, Leeds. From the 125 male patients, 185 of their male, first-degree relatives ≤65 years old at the time of recruitment and free from a personal history of CAD were contacted and interviewed. An equal number of male, community control subjects ≤ 65 years old and without a personal or family history of CAD or diabetes mellitus were recruited via the Leeds Health Authority Family Health Service register. Subjects with a personal history of hypertension and/or taking cardiovascular medication (including aspirin and lipid-lowering drugs) were subsequently excluded from the study. All remaining subjects (175 in each group) were white, North European, and they gave informed consent according to a protocol approved by the United Leeds Teaching Hospitals (NHS) Trust Research Ethics Committee.

At the time of recruitment, all subjects had fasted for a minimum of 10 hours overnight. Venous blood, 50 mL, was taken from an antecubital vein with a 19-gauge needle without venous stasis with the subject in a supine position. Blood was collected in lithium heparin for lipid fraction analysis, a 10-mL tube containing 1 mL 0.9% citrate (pH 8.8) at 4°C for assay of insulin, PAI-1, tPA, and fibrin D-Dimer and a similar citrate tube at room temperature for assay of fibrinogen. The citrate samples were centrifuged at 2560g at 4°C or room temperature for 30 minutes, and 0.5-mL aliquots of plasma supernatant snap-frozen in liquid nitrogen for storage at −40°C until assay. All subjects underwent 75-g oral glucose tolerance testing with blood collected in lithium fluoride for fasting and 2-hour glucose estimation. DNA was extracted from blood collected in EDTA for subsequent genotyping at the PAI-1 4G/5G promoter polymorphism by using a previously described method.²²

Systolic and diastolic blood pressure (BP) measurements were performed manually to the nearest 2 mm Hg with subjects supine, and the measurements were calculated as a mean value of 3 consecutive readings. Relatives and control subjects with currently treated hypertension were excluded from subsequent BP analysis. Standard 12-lead electrocardiographs were recorded for all subjects, along with anthropometric data: body mass index (BMI), calculated from weight in kilograms divided by the square of the height in meters, and waist-to-hip ratio (WHR). All subjects enrolled had given negative responses to the Rose angina questionnaire and provided details concerning medical history, drug history, family history, smoking, and alcohol consumption.

A subset of first-degree relatives (n=74) and control subjects (n=90), who expressed a willingness to take part and who, in common with the remainder of the study subjects, were free from significant lung disease or other exercise limiting conditions, underwent assessment of their cardiorespiratory fitness. Subjects performed a Bruce protocol cardiopulmonary exercise test using the Medgraphics Cardio2 System for analysis of respiratory gas concentrations and flow rates. Cardiorespiratory fitness was estimated from maximal oxygen uptake (VO₂max) with anaerobic threshold as a secondary measure.

Insulin, PAI-1 antigen, tPA antigen, and fibrin D-dimer were measured with commercially available ELISA kits from BioSource (insulin) and Biopool International (PAI-1, tPA, and D-dimer). Fibrinogen was measured by using the Clauss method.²³ Inter-assay and intra-assay coefficients of variation were 3.8% and 2.5% for insulin, 9.8% and 4.4% for PAI-1, 10.4% and 7.0% for tPA, 5.3% and 3.5% for D-dimer, and 3.5% and 2.0% for fibrinogen. A glucose oxidase method was used for measurement of plasma glucose and a Hitachi 747 autoanalyzer (Boehringer Mannheim) for estimation of triglyceride and total cholesterol. HDL cholesterol was measured by a Hitachi 717 autoanalyzer after removal of LDL, chylomicrons, and VLDL by precipitation with phosphotungstic acid and magnesium chloride. LDL cholesterol was calculated by the Friedewald equation.

Values for age for the 3 study groups did not conform to a normal distribution and are presented as medians with 25th and 75th percentiles. Differences in age between groups were assessed by the Kruskal-Wallis test. To achieve a normal distribution for insulin, PAI-1, tPA, D-dimer, plasma fasting glucose, and triglyceride, their values were log-transformed and subsequent data presented as geometric mean with antilog 95% confidence intervals. Values for insulin resistance were estimated by use of the homeostasis model assessment, whereby insulin resistance=fasting glucose (mmol/L)×fasting insulin (mU/L)/22.5, and which assumes that normal-weight, healthy subjects <35 years old have 100% β-cell function and an insulin resistance of unity.²⁴ Spearman bivariate correlations were used to assess the relationship among PAI-1, tPA, and D-dimer, and features of the insulin resistance syndrome and other continuous variables. Independent Student’s t test was used to assess differences in continuous and parametric data and a χ² test for differences in category frequencies. One-way ANOVA was used to identify any relationship between 4G/5G genotype and PAI-1 levels. Multiple linear regression analysis was performed for tPA and D-dimer, and the data presented represents the best-fit model as determined by the R² value. Statistical significance was taken as P<0.05, and all analyses were performed with the Statistics Package for Social Scientists software for Windows version 9.0.

Results

Clinical, biochemical, and hemostatic assay data are shown in Table 1. D-dimer, tPA, and fibrinogen levels were significantly higher in relatives compared with controls, 55 (52 to 58) ng/mL versus 49 (45 to 53) ng/mL, P<0.01, for D-dimer, 8.0 (7.5 to 8.6) ng/mL versus 5.6 (5.2 to 6.1) ng/mL, P<0.001, for tPA, and 3.0 (2.9 to 3.1) g/L versus 2.8 (2.7 to 2.9) g/L, P<0.05, for fibrinogen. There was no difference in PAI-1 levels between the two groups. Apart from HDL cholesterol, 1.4 (1.3 to 1.4) mmol/L in relatives compared with 1.3 (1.2 to 1.3) mmol/L in controls, P<0.05, the relatives and controls were similar in terms of clinical data (including alcohol consumption and cigarette smoking) and biochemical parameter. There was a trend toward increased insulin resistance with higher BMI and systolic BP in relatives compared with controls although not statistically significant. There was no difference in the cardiorespiratory fitness of relatives and control subjects as measured by VO₂max, 34.7 (33.1 to 36.2) mL · min⁻¹ · kg⁻¹ versus 34.1 (32.2 to 36.0) mL · min⁻¹ · kg⁻¹ or anaerobic threshold, 20.8 (19.8 to 21.9) mL · min⁻¹ · kg⁻¹ versus 21.1 (20.0 to 22.2) mL · min⁻¹ · kg⁻¹.

Bivariate correlation analysis confirmed the association between PAI-1 and tPA and the insulin resistance syndrome. For tPA and age, estimated insulin resistance, BMI, WHR, systolic and diastolic BP, total cholesterol, triglyceride, HDL cholesterol, PAI-1, and fibrinogen, |r|≥0.20 in relatives and controls. For PAI-1 and the same group of correlates, |r|≥0.19 in relatives and controls. Correlation data for fibrin D-dimer is shown in Table 2. There was a strong correlation with age, r=0.32 and 0.40 (P<0.01), and fibrinogen, r=0.24 and 0.29 (P<0.01) in relatives and control subjects, respectively. The subgroup of relatives and controls who underwent the cardiopulmonary exercise test were similar in terms of age (median, 42 years) and all conventional CAD risk factors. Relatives had increased levels of tPA, D-dimer, and fibrinogen compared with controls. Significant correlation data for VO₂max is shown in Table 3.

PAI-1 4G/5G genotype was determined for all relatives and control subjects. The 4G/4G genotype was associated with elevated levels of PAI-1 in control subjects, 15.9 (12.5 to 20.2) ng/mL for 4G/4G versus 8.9 (6.6 to 12.0) ng/mL for 5G/5G control subjects, P=0.01. In the relatives group, there was a similar trend, 14.0 (10.9 to 18.1) ng/mL for 4G/4G versus 9.3 (6.4 to 13.7) ng/mL for 5G/5G. Genotype frequencies did not differ between the relatives and control groups.

Multiple linear regression models were used to calculate adjusted tPA and D-dimer levels using the significant covariates identified from bivariate correlation analysis. Only those covariates which either independentlypredicted tPA or D-dimer levels or contributed >1% variance were included in the final models (Tables 4 and 5⇓). Exclusion of the remaining correlates did not affect the R² values or the adjusted levels of tPA or D-dimer. The final models accounted for 53.7% and 21.6% of the variance in tPA and D-dimer levels, respectively. Relative/control status and PAI-1 levels were the most influential contributors to tPA variance, and relative/control status, age, and fibrinogen were the only significant factors in the D-dimer model. From the calculated effect size, relatives had a 1.4-fold increase in tPA and a 1.1-fold increase in D-dimer levels. An increase in age by 20 years was associated with a 1.3-fold rise in both tPA and D-dimer levels. Adjusted mean tPA levels were 8.1 (7.7 to 8.6) ng/mL for relatives versus 5.8 (5.5 to 6.1) ng/mL for controls, P<0.001, and adjusted D-dimer levels were 55 (52 to 59) ng/mL for relatives and 49 (46 to 52) ng/mL for controls, P=0.006.

Data concerning the probands (patients with premature, multi-vessel CAD) are included for completeness (Table 1). As expected, probands were older than the relatives and controls with higher BMI and WHR and elevated levels of tPA, PAI-1, D-dimer, fibrinogen, and estimated insulin resistance. BP and lipid profiles were affected by concurrent cardiovascular medication, which the majority of these patients were receiving at the time of recruitment.

Discussion

In this study, we have found elevated levels of tPA antigen, fibrin D-dimer, and fibrinogen in the healthy, male, first-degree relatives of patients with severe, premature CAD. tPA antigen and D-dimer remained higher in relatives compared with controls after adjustment for all significant correlates, including fibrinogen. Confounding factors cannot be ignored, but the principle strength of this study lies in the quality of subject recruitment, thereby avoiding many of the confounders that undermine studies of similar design. We have successfully enrolled more than 90% of the suitable relatives identified from our CAD patient group and simultaneously recruited a well characterized and highly comparable group of community control subjects. The lack of any significant differences between the groups in terms of history of smoking, alcohol consumption, clinical and biochemical parameters, and physical fitness would all mitigate against any significant confounding factors.