Increased plasma fibrinolysis and tissue-type plasminogen activator/tissue-type plasminogen activator inhibitor ratios after ethanol withdrawal in chronic alcoholics

Abstract

The effects of alcohol withdrawal on fibrinolysis were studied in 10 middle-aged male chronic alcoholics institutionalized for withdrawal therapy. All patients were sampled on admission [day 1 (D1)] and 21 days after alcohol withdrawal [day 22 (D22)]. The overall plasma fibrinolytic capacity (OFC) was assayed by measuring the ability of patient plasma to generate D-dimers from a standardized fibrin clot, and tissue-type plasminogen activator (t-PA) and t-PA inhibitor (PAI-1) levels were assayed together with serum cholesterol, triglyceride and cholesterol fractions. At D22, the OFC significantly increased in seven patients [D1 = 10 ± 0.7 μg/h (mean ± SD), D22 = 17 ± 7.4 μg/h;P< 0.01], while t-PA and PAI-1 levels decreased in all patients but two (t-PA: D1 = 16.6 ± 5 ng/ml, D22 = 10.2 ± 3.8 ng/ml;P< 0.001; and PAI-1: D1 = 46 ± 39 ng/ml, D22 = 21 ± 28 ng/ml;P< 0.01). This study clearly demonstrates an increase in overall fibrinolytic activity after alcohol withdrawal, which is mainly due to a decrease in PAI-1 levels. These changes induced by alcohol abstinence might provide clear benefit by reducing the risk of thromboembolic events and particularly of stroke associated with elevated PAI-1 levels described in heavy drinkers.

Introduction

In any individual, the effects of alcohol on coagulation and fibrinolysis may depend largely on the amount and type of alcoholic beverage ingested, and the clinical consequences associated with alcoholism are thus variable. Several epidemiological studies have shown that moderate alcohol consumption was associated with a lower risk of coronary heart disease ^[1,2]. The protective effects of alcohol against coronary heart disease is partly mediated by increased levels of high-density lipoprotein cholesterol ^[3,4], but it has also been found that alcohol may increase plasma fibrinolytic activity ^[4,5]. Tissue-type plasminogen activator (t-PA) and its inhibitor (PAI-1) have been the most frequently studied haemostasis parameters in alcoholic patients. Levels of both t-PA and PAI-1 increase after moderate alcohol consumption but the resulting overall fibrinolytic activity is usually reduced ^[6–8]. On the contrary, little information is available on the fibrinolytic status of heavy alcohol consumers. t-PA and PAI-1 antigens are reported to be raised but the consequences of such changes on fibrinolysis in such patients are not fully understood ^[9]. In addition, the effects of alcohol withdrawal on fibrinolysis have never been documented.

The aim of the present prospective study was to study fibrinolysis in 10 chronic alcoholics after alcohol withdrawal by measuring t-PA and PAI-1 levels, and to evaluate the evolution of these parameters in relation to overall fibrinolytic activity.

Patients and methods

Patients and study design

Ten male alcoholics (27–55 years) admitted to a specialized drying-out centre volunteered for this prospective study.

Each subject was still consuming alcohol on admission (100–548 g ethanol/day). The source of alcohol was only wine in two subjects, beer in two subjects, spirits in one subject, both wine and beer in four subjects, and spirits and beer in one subject. The history of alcohol consumption ranged from 10 to 20 years . All 10 patients were current smokers (10–26 cigarettes/day). None had clinical signs of liver cirrhosis and prothrombin time was consistently above 90% in all subjects. Serum aspartate aminotransferase (AST) levels were elevated in four subjects (31–49 IU/l; normal < 30 IU/l) and four had increased serum alanine aminotransferase levels (ALT) (41–94 IU/l; normal < 30 IU/l). Gamma-glutamyltransferase (GGT) levels were elevated in nine patients (44–281 IU/l; normal < 40 IU/l) and carbohydrate-deficient transferrin (CDT) levels were also high in seven patients (64–130 IU/l). Serum total cholesterol levels were > 6.20 mmol/l in nine subjects, and triglyceride levels were above 1.70 mmol/l in five subjects. Values for glucose were < 6.90 mmol/l and creatinine < 130 μmol/l in all patients . Serum orosomucoid was also measured as a marker of infectious or inflammatory disease, and was higher than 1.50 g/l in only one subject.

The purpose and design of this study were accepted by the Ethics Committee of our University Hospital, and informed written consent was obtained from each participant.

Withdrawal therapy started the day after admission (D1), and no alcohol was then allowed and abstinence was checked every day by the medical staff. Sedative drugs (meprobamate and/or benzodiazepines) and vitamins B₁, B6 and B₃ were administered to all patients. Smoking was not prohibited and no subject gave up during hospitalization . Patients were allowed to move freely within the hospital. All subjects were fed ad libitum and food was provided as three mixed meals per day. We also measured weight, height and body mass index on D1 and day 22 (D22). Nutrient and alcohol intakes were assessed by personal interview based on 7 days’ dietary recall. The dietary data were coded by food items and quantities, and then analysed by a computerized program using a French nutrient database.

Blood samples and assays

Blood sampling was carried out between 8:00 and 10:00 a.m in the laboratory of hemostasis on D1 and D22 after alcohol withdrawal. Patients had eaten or drunk nothing since the evening before and were 15 min at rest before sampling. Venous blood punctures were performed either in siliconized vacutainer tubes with 0.109 mol/l trisodium citrate for hemostasis assays or in dry vacutainer tubes with a serum separator activator for biochemical assays.

Serum ALT, AST and GGT activity was measured by kinetic photometric methods using a Hitachi 747 analyser, and the results were expressed (IU/l) at 30°C. CDT was quantified using a previously described nephelometric method ^[10] with results expressed both in milligrams per litre and as percentages of total transferrin. Measurement of lipoprotein a [Lp(a)] and orosomucoid levels was carried out by rate nephelometric assay on an Array Protein System (expressed as g/l). Triglyceride levels were measured on a SMAC II analyser (Bayer Diagnostic, Tarrytown, New York, USA) using an enzymatic ultraviolet method, and total cholesterol and glucose were assayed by colorimetric methods using a Hitachi 747 analyser, with all results expressed in millimoles per litre. High-density lipoprotein (HDL) and HDL3 cholesterol fractions were separated using a commercially available kit (`Quantolip'; Immuno-diagnostics, Vienna, Austria) based on precipitation with polyethylene glycol. The subsequent measurements of cholesterol were performed by a colorimetric method on a Hitachi 911 analyser. HDL2 cholesterol concentrations were calculated by subtracting HDL3 cholesterol from total HDL cholesterol.

Citrated samples for haemostasis evaluation were centrifuged twice at 2500 ×g at 10°C for 15 min. Platelet-poor plasma samples were then promptly separated; one fraction was immediately tested and the other was aliquoted and frozen at −80°C.

Prothrombin time (PT), activated partial thromboplastin time (aPTT) and fibrinogen were measured on the day of sampling. PT and aPTT were measured using classical procedures with Neoplastin CI and CK prest (Diagnostica Stago, Asnières, France) as reagents. Fibrinogen levels were measured according to the method of von Claus using Fibrinomat (Biomérieux, Marcy l'Étoile, France). Factor VII activity (factor VIIc) was assayed in a regular one-stage system using human thromboplastin (Dade Behring, Paris, France) and factor VII (FVII)-deficient substrate (Diagnostica Stago). FVII antigen (FVIIag) was determined with an enzyme-linked immunosorbent assay kit (Diagnostica Stago). The results were expressed in seconds for aPTT and PT, in percent of normal pooled plasma for FVIIc and FVIIag, and in grams per litre for fibrinogen.

Euglobulin clot lysis time (ECLT) was determined using the method described by von Kaulla, and results expressed in minutes. Overall fibrinolytic capacity (OFC) was evaluated by a recently developed assay ^[11] based upon the ability of plasma to degrade a standardized fibrin clot at 37°C in the presence of a limited amount of purified t-PA (5 ng) with an optimized concentration of silica (125 μg), and the results were expressed as the amount of D-dimers generated in 1 h (μg/h). D-Dimer levels were determined using a previously described immunoturbidimetric method (Liatest D-dimer; Diagnostica Stago) and expressed in micrograms per millilitre.

PAI-1 activity was measured by a two-stage amidolytic method using Staclot PAI-1 kit (Diagnostica Stago) and the results expressed in arbitrary units (AU/ml). PAI-1 antigen and t-PA antigen levels were assayed by ELISA (Asserachrom PAI-1 and Asserachrom t-PA; Diagnostica Stago), expressed in nanograms per millilitre. Clotting and chromogenic assays and D-dimer measurements were performed on a STA coagulometer. All tests, including ELISA, were performed in duplicate, and samples of D1 and D22 were tested in the same series for each patient.

Statistical analysis

The changes before and after alcohol withdrawal were compared using a paired Student t test. The Spearman correlation test was used to evaluate the relationship between variables. All statistical analyses were performed using Graphpad Prism^TM software (San Diego, California, USA).

Results

Alcohol abstinence in the 10 patients enrolled in this study was associated with a significant decrease in energy intake but without any change in fat and carbohydrate intake or in body mass index (Table 1). The protein intake was also significantly lower. Among the biochemical parameters, serum GGT and CDT were significantly reduced after 3 weeks of alcohol withdrawal but orosomucoid levels remained unchanged. Total cholesterol, total HDL cholesterol, HDL3 cholesterol and HDL2 cholesterol also decreased compared with admission levels (Table 2). Serum Lp(a) concentrations remained unchanged and triglyceride levels decreased after withdrawal, but the differences obtained were not statistically significant compared with D1.

Table 1: Characteristics of the 10 chronic alcoholic men before (D1) and after 21 days (D22) of withdrawal therapy

Table 2: Serum total cholesterol, triglycerides, high-density lipoprotein (HDL) cholesterol, HDL2 and HDL3 cholesterol, and lipoprotein a [Lp(a)] before (D1) and after 3 weeks (D22) of alcohol abstinence in the 10 subjects

On the contrary, the results of coagulation tests (i.e. PT and aPTT) were unchanged, but a significant increase in plasma fibrinogen levels was recorded on D22 (Table 3). The mean factor VII plasma levels on admission were above the normal range, reaching 128% in activity and 121% for antigenic values, particularly because four individuals exhibited levels consistently higher than 130%. In three of these patients, FVII levels consistently decreased on D22 but FVIIc/FVIIag ratios always remained close to 1.

Table 3: Haemostasis parameters before (D1) and after 3 weeks (D22) of alcohol withdrawal in the 10 subjects

Plasma fibrinolytic activity was initially evaluated by measuring ECLT, which was slightly shorter after alcohol abstinence, but the difference was not significant. Patients’ OFC was also studied by measuring the amount of D-dimers generated when the plasma degraded a standard fibrin clot in the presence of a constant concentration of t-PA. With this assay, a significant increase in plasma fibrinolysis was shown after 21 days of alcohol abstinence in seven of the 10 patients studied (Fig. 1). In one case (patient 3), the OFC remained very low and undetectable, while in another (patient 7) it only slightly decreased after ethanol abstinence. On the contrary, the OFC that was initially high on D1 was unmodified on D22 in patient 9.

Figure 1.:   Overall fibrinolytic capacity before (open square) and after (filled square) 3 weeks of alcohol withdrawal.

The t-PA antigen levels were normal on D1 in only two subjects and above the normal range in eight of the 10 alcoholic patients, varying between 14 and 26.5 ng/ml (Fig. 2). In addition, functional levels of PAI-1 were elevated (> 10 AU/ml) in every case despite normal antigen levels in six patients. In particular, two subjects exhibited very high levels of PAI-1 antigen, with values of 140 and 80 ng/ml. Finally, the calculated t-PA/PAI-1 ratio was low (< 1) in nine of the 10 enrolled patients.

Figure 2.:   Values of tissue-type plasminogen activator inhibitor (PAI-1) antigen (PAIag), PAI-1 activity, Tissue-type plasminogen activator (t-PA) antigen (t-PAag) and t-PA/PAI-1 ratios in the 10 patients before [day 1 (D1)] and after 3 weeks of alcohol withdrawal [day 22 (D22)].

The t-PA antigen levels significantly decreased after 3 weeks of alcohol withdrawal (P = 0.0002) and reached normal values, except in two patients whose plasma t-PA values remained high (13 and 19.5 ng/ml). At the same time, functional and antigen PAI-1 levels significantly decreased in most patients on D22. These changes after alcohol withdrawal were associated with increased t-PA/PAI-1 ratios, individual values being close to 1 except in two cases for whom no significant increase in OFC occurred (patients 3 and 7). These two patients had the lowest t-PA/PAI-1 ratios (0.20–0.38) and the highest PAI-1 activity levels on day 22 (Fig. 2). Finally, there was a clear significant and positive correlation between the OFC and t-PA/PAI-1 ratios (r = 0.71) for all measurements performed on days 1 and 22 in all 10 patients (Fig. 3). However, the strongest correlation was recorded when OFCs were compared with PAI-1 levels (−0.91), which also correlated with triglyceride or GGT levels. In contrast, no correlation was found with cholesterol or HDL cholesterol fractions.

Figure 3.:   Correlation (Spearman correlation test) between overall fibrinolytic capacity (OFC) and tissue-type plasminogen activator (t-PA)/tissue-type plasminogen activator inhibitor (PAI) ratios, OFC and PAI levels. Correlation between PAI values and triglyceride levels, PAI and gamma-glutamyltransferase (γ-GT) levels.

Discussion

The involvement of fibrinolysis in the pathogenesis of coronary heart disease and ischaemic stroke has been clearly recognized and the role of t-PA and its inhibitor in this regard has been emphasized in several previous reports ^[12–15]. Among the variable effects induced by alcohol consumption, those concerning fibrinolysis have been widely studied, but most reported findings have been collected in light or moderate ethanol consumers ^[4,6,7]. Several reports indicated that light to moderate alcohol consumption is associated with a reduced mortality from ischaemic heart disease, particularly if the beverage is wine, but no further risk reduction is expected if more than 30 g ethanol/day is consumed whatever the beverage ingested ^[16,17].

The 10 chronic alcoholics enrolled in this study ingested more than 100 g alcohol/day before voluntarily entering a program to stop their alcohol consumption. The effects of withdrawal on fibrinolysis were investigated.

A relatively small number of subjects was studied without a control group, and sequences of alcohol and abstinence phases could not be randomized in our study. Despite these limits, our findings clearly demonstrated that stopping alcohol consumption is associated with significant biological changes involving not only lipid profiles, but also fibrinolysis status. A recently described assay ^[11], named ‘overall fibrinolytic capacity’ (OFC), was used to evaluate plasma fibrinolytic activity in these patients. This method measured the amount of D-dimers generated in the presence of patient plasma incubated with a standardized amount of fibrin and a constant concentration of t-PA. OFC is therefore expected to be more sensitive than the classical euglobulin clot lysis time, particularly to variations in PAI-1 concentrations. Although obtained in only 10 subjects, our findings clearly support higher sensitivity to variations in t-PA and PAI-1 of OFC compared with ECLT. Indeed, ECLT values were not statistically different before and after alcohol withdrawal, despite the increased t-PA/PAI-1 ratio in most of patients. In contrast, the amount of D-dimers generated in 1 h by patient plasma in the OFC assay was strongly correlated with PAI-1 levels. The OFC method was recently developed by Amiral et al. ^[11], who showed that PAI-1 activity was the most important parameter in this assay, accounting for about 50–60% of measured values. In accordance with Amiral et al., OFCs in our study were also well correlated with t-PA antigen levels, but this protein is in fact mainly present in plasma in an inactive form and increases together with PAI-1. PAI-1 levels were also correlated with triglycerides, explaining the indirect correlation between OFC and triglycerides ^[11,18]. One of the patients studied (patient 3) had very high PAI-1 levels that remained elevated after alcohol abstinence, and this probably explains why the OFC value also did not rise on day 22. This patient had the highest GGT levels on day 1 with increased ALT levels, the latter biological parameter being related to cytolytic hepatic lesions, possibly due to severe alcohol consumption. In this particular subject, transaminase levels were still above normal values on day 22 and triglyceride levels increased, suggesting that this patient possibly did not completely stop his alcohol intake.

PAI-1 levels also correlated well with GGT and this finding, together with the significant decrease in PAI-1 after alcohol abstinence, further supports an effect of alcohol on PAI-1 levels. As shown by Marques-Vidal et al., who also studied both lipid and haemostatic parameters in a large population of drinkers ^[19], this effect is dependent on the amount of alcohol ingested. Our patients were heavy drinkers and the observed decrease in energy intake can mainly be related to alcohol withdrawal. At the same time, the fat and carbohydrate intake, and thus the body mass index, were unchanged. Furthermore, the protein intake was lower but remained in a normal range since all patients were put on a controlled and balanced diet. Serum GGT levels dramatically decreased after alcohol abstinence and, even if severe hepatic failure was absent, an improvement of hepatocytes function cannot be discarded in some patients.

The mechanisms leading to increased PAI-1 in heavy drinkers are, however, not fully understood, and both triglycerides and insulin might be involved. It was recently shown that fasting and post-glucose insulin concentrations decreased significantly with increasing amounts of regular alcohol consumption ^[20]. In addition, it was also demonstrated that alcohol withdrawal enhanced insulin secretion ^[21,22]. On the other hand, insulin stimulates the synthesis of PAI-1 by human adipose tissues ^[23] or by hepatocytes ^[24]. Therefore, as PAI-1 levels decreased in our patients between D1 and D22, variations in insulin levels are unlikely to be responsible for this effect. The alternative mechanism is thus more likely and involves triglycerides. Indeed, very low-density lipoproteins have been shown to induce the transcription of the human PAI-1 promoter in endothelial cells ^[25], and the findings in our patients are in a complete agreement with these in vitro results.

Impaired fibrinolysis due to elevated PAI-1 levels has been related clinically to a high risk of cardiovascular disease. Studies concerning heavy alcohol intake have also shown, in many cases, that heavy drinking is an independent risk factor for stroke ^[26] that decreases after alcohol withdrawal, and levels of PAI-1 could be one of the most important disease parameters involved in this regard.