- Email: [email protected]
Markers of fibrinolysis as predictors for maintenance of sinus rhythm after electrical cardioversion
Thrombosis Research 127 (2011) 189–192
Contents lists available at ScienceDirect
Thrombosis Research j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t h r o m r e s
Regular Article
Markers of fibrinolysis as predictors for maintenance of sinus rhythm after electrical cardioversion Jonas Andersson a,⁎, Henrik Almroth b, Niklas Höglund c, Steen Jensen c, Per Tornvall d, Anders Englund b, Mårten Rosenqvist e, Kurt Boman a a
Department of Medicine and Geriatrics, Skellefteå County Hospital, S-931 86 Skellefteå, Sweden Department of Cardiology, Örebro University Hospital, S-701 85 Örebro, Sweden Department of Cardiology, Heart Centre, University Hospital, S-901 85 Umeå, Sweden d Department of Cardiology, Karolinska University Hospital, S-171 76 Stockholm, Sweden e Department of Cardiology, Karolinska Institutet at Stockholm South Hospital, S-118 83 Stockholm, Sweden b c
a r t i c l e
i n f o
Article history: Received 31 March 2010 Received in revised form 26 May 2010 Accepted 2 June 2010 Available online 14 January 2011
a b s t r a c t Introduction: Inflammation, endothelial dysfunction and metabolic pathways provide possible links between the inflammatory and hypofibrinolytic states in atrial fibrillation. Our aim was to explore the role of mass concentrations of PAI-1 and tPA, activities of PAI-1 and tPA as predictors of recurrence of atrial fibrillation adjusted for CRP. Materials and methods: The study included 129 patients with persistent atrial fibrillation. Laboratory analyses were performed including PAI-1 activity, PAI-1 mass, tPA activity, tPA mass and CRP in baseline. Patients were then randomized to atorvastatin (40 mg, two tablets once daily) or placebo, initiated at least 14 days before the elective cardioversion. Further samples and follow-up were made at day 2 and 30 days after cardioversion. Results: In univariate logistic regression no fibrinolytic variable was significantly correlated with rhythm in day 30. In multivariate analysis lower PAI-1 mass was significantly associated with sinus rhythm in all models including fibrinolytic variables, CRP, metabolic components, age, hypertension and smoking. After adding treatment allocation to the fully adjusted model, PAI-1 mass remained significantly associated with sinus rhythm both at day 2 and 30 (OR 0.98; 95% CI 0.95–1.00). Conclusions: No fibrinolytic component alone was found to be a predictor of recurrence of atrial fibrillation. In multivariate models lower PAI-1 mass was associated with sinus rhythm even after adjusting for CRP, markers of the metabolic syndrome and treatment with atorvastatin. Our findings suggest a patophysiological link between AF and PAI-1 mass but the relation to inflammation remains unclear. © 2010 Elsevier Ltd. All rights reserved.
Introduction The first evidence of atrial fibrillation (AF) as an inflammatory disease emerged in the late 1990s [1,2]. Subsequent studies demonstrated inflammatory infiltrates in atrial biopsies from AF patients and found that CRP was a predictor not only of the presence of and the risk of developing future AF but also the recurrence of AF after electrical cardioversion [3–7]. A hypofibrinolytic state with increased plasminogen activator inhibitor type 1 mass concentration (PAI-1 mass) and its complex with tissue plasminogen activator (t-PA/PAI-1 complex) levels was also described in AF [8–10] and PAI-1 has been found to predict the recurrence of AF [11,12]. Inflammation, endothelial dysfunction and metabolic pathways provided possible links between the inflamma⁎ Corresponding author. Department of Medicine and Geriatrics Skellefteå County Hospital SE–931 86 Skellefteå (Sweden). Tel.: +46 910 771 294; fax: +46 910 771 125. E-mail address: [email protected] (J. Andersson). 0049-3848/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2010.06.002
tory and hypofibrinolytic states in AF [9,13,14]. There is a well known interaction between inflammation and fibrinolysis as components of the fibrinolytic system act as acute phase reactants. The described association between PAI-1 and recurrence of AF may thus reflect an inflammatory state rather than a disturbed fibrinolytic system. Moreover, treatment with statins has been found to reduce both PAI-1 and C-reactive protein (CRP) and suggested as a new treatment option to prevent recurrence of AF [15,16]. We here present new data on components of the fibrinolytic system taking into account the inflammatory state with and without treatment with atorvastatin. Our null hypothesis was the components of the fibrinolytic system do not predict maintenance of SR when adjusted for CRP. Aim Our primary aim was to explore the role of mass concentrations of PAI-1 and tPA, activities of PAI-1 and tPA as predictors of maintenance of SR in a non-postoperative situation with adjustment for CRP.
190
J. Andersson et al. / Thrombosis Research 127 (2011) 189–192
Method
Ethics
We conducted a pre-specified hypothesis generating study within a double-blinded, placebo-controlled, randomized, prospective, investigator-initiated multicentre study of the effects of atorvastatin on recurrence of AF after cardioversion which has previously been described in detail [17]. Between August 2004 and January 2007, a total of 234 patients with persistent AF and an indication for cardioversion were included. Five of the participating centres collected blood samples from their 129 patients for the present study. Persistent AF was defined as AF with duration more than 7 days. Patients below 18 years and over 80 years of age were excluded as were patients with paroxysmal AF, atrial flutter, contraindications against atorvastatin, ongoing treatment with lipidlowering drugs, ongoing treatment with class I or class III antiarrhythmic treatment, known liver disease or a myopathy, as well as patients with a previous electrical cardioversion the last 12 month. Treatment with betareceptor-blocking agents, calcium antagonists, diuretics, digoxin, ACE-inhibitors, ARBs, and acetylsalicylic acid were used when clinically indicated. Warfarin was given to all patients according to national guidelines, i.e. international ratio (INR) of 2–3, Nthree weeks before cardioversion, checked weekly. Before randomization, i.e. baseline, a detailed medical history was obtained and transthoracic echocardiographic examination, 12-lead electrocardiography, and basic laboratory analysis were performed including PAI1 activity, PAI-1 mass, tPA activity, tPA mass, von Willebrand factor (vWF), CRP and blood lipids. Patients were then randomized to atorvastatin (40 mg, two tablets once daily) or placebo, initiated at least 14 days before the elective cardioversion. Further samples and follow-up were made at day 2 and 30 days after cardioversion. Compliance was monitored at cardioversion and 30 days after cardioversion (end of treatment) by pill counting. For analysis of fibrinolytic markers blood were collected into Stabilyte tubes and analysis was carried out using an enzyme-linked immunosorbent assay (ELISA). Chromolize PAI-1 and Chromolize t-PA from Biopool AB, Sweden, was used to determine activities of PAI-1 and t-PA respectively. Tintelize PAI-1 and Tintelize t-PA was used to determine mass concentrations of PAI-1 and tPA respectively. Von Willebrand factor was determined with enzyme-linked immunosorbent assays (ELISA) from DAKO in Denmark. Plasma CRP was determined with an automated high sensitive C-reactive protein method, (IMMULITE, Diagnostic Products Corporation, USA). Cardioversion was performed according to local standard clinical practice on an elective outpatient basis. Patients were instructed to contact the study team if they experienced symptoms suggesting AF between scheduled visits. No patient was lost during follow-up.
The Ethics Committee at Karolinska Institute and the Swedish Medical Product Agency approved this study that also complied with the Declaration of Helsinki. Written informed consent was obtained from all patients.
Statistical analysis The Mann-Whitney U test was used for calculations between groups and the Wilcoxon signed ranks test for calculations within the groups. Chi-square was used for grouped variables when comparing the subgroup with the entire population. Correlation between variables was tested with Spearman´s test. Logistic regression was used to calculate the odds ratio (OR) and 95% confidence interval for components of the fibrinolytic system to predict rhythm at 30 days adjusting for potential confounding factors. In the multivariate analysis for fibrinolytic variables, PAI-1 activity was excluded due to a high correlation (r = 0.80) with tPA activity. In the initial model fibrinolytic variables were analysed, thereafter was the contribution of inflammation (CRP) analysed. The components of the metabolic syndrome (hypertension, triglycerides, BMI and cholesterol) were then added and at age and smoking. For all analyses p b 0.05 (two-tailed) was considered statistically significant. All analysis was performed using Statistical Package for the Social Sciences (SPSS) version 17.0 (Chicago, IL, USA).
Results In total 129 patients (97 males, 75%, 32 females, 25%, and mean age 65) were eligible for the substudy. Baseline characteristics according to rhythm at the 30-day follow up are presented in Table 1. The substudy was well-balanced with the main study with regard to baseline characteristics and treatment (placebo = 64, atorvastatin = 65). Levels of biomarkers at baseline, two days after cardioversion and at the 30 day are presented in Table 2. Only HDL (p = 0.048) was significantly different in baseline between the two treatment groups in the substudy. Effect of cardioversion and rhythm At day 30 recurrent AF was observed in 74 patients and atrial flutter in two. Two days after cardioversion PAI-1 activity, PAI-1 mass and tPA mass had decreased significantly and CRP had increased significantly compared to baseline (Table 2). Among the 53 individuals who remained in sinus rhythm after 30 days PAI-1 activity, tPA activity and tPA mass was significantly lower at day two after cardioversion compared to baseline and in the AF group only PAI-1 mass was significantly lower. In within group analysis there were no significant change between baseline and day 30 in any fibrinolytic variable or vWF among neither those who remained in sinus rhythm nor among those with recurrent AF at day 30. However CRP was significantly lower at day 30 compared to baseline among those who remained in sinus rhythm (p = 0.04). In between group analysis, there were no significant differences for any fibrinolytic variable, vWF or CRP neither at baseline nor at day 30. Among those with recurrent AF at day 30 there were no significant changes in fibrinolytic variables from baseline to day 30. Predictors of rhythm In univariate logistic regression no fibrinolytic variable or vWF was significantly correlated with rhythm in day 30. However, in the multivariate analysis lower baseline PAI-1 mass was significantly associated with sinus rhythm in all models, Table 3. Lower baseline PAI-1 mass was also predictive of sinus rhythm at day 2 (OR 0.98; 95% CI 0.97–1.00, p = 0.02) and remained significant after adjusting for all Table 1 Baseline characteristics according to rhythm at 30-days post cardioversion.
Number (n) Age Sex M/F Smokers Weight Body mass index Total cholesterol HDL LDL Triglycerides Hypertension Stroke ICH Left atrial size Normal ejection fraction
(years) n (% men) n (%) (kg) (kg/m2) (mmol/L) (mmol/L) (mmol/L) (mmol/L) n (%) n (%) n (%) (mm) n (%)
Sinus rhythm
Atrial fibrillation
53
74
67 (59; 75) 42/11 (79) 2 (3.8) 84 (75; 100) 27 (25; 31) 5.20 (4.65; 6.00) 1.26 (1.04; 1.38) 3.27 (2.79; 3.69) 1.59 (1.12; 2.01) 25 (47.2) 2 (3.8) 4 (7.5) 45 (40; 49) 35 (66)
66 (59; 72) 53/21 (72) 9 (12.2) 87 (77; 95) 28 (25; 31) 5.30 (4.90; 5.90) 1.24 (1.06; 1.53) 3.31 (2.79; 3.92) 1.58 (1.00; 1.85) 28 (37.8) 2 (2.7) 4 (5.4) 44 (40; 48) 60 (81)
Median and 25th and 75th percentiles are given.
J. Andersson et al. / Thrombosis Research 127 (2011) 189–192
191
Table 2 Median values for the markers of inflammation and fibrinolysis and VWF.
PAI-1 activity (IU/mL) PAI-1 mass (ng/mL) tPA activity (IU/mL) tPA mass (ng/mL) CRP (mg/L) VWF (% of normal)
Baseline
Day 2 after cardioversion
p
Day 30
p
11.9 27.1 0.79 10.7 2.30 87
7.42 22.8 0.89 0.7 2.30 184
0.001 0.003 0.084 0.000 0.025 0.662
12.6 28.1 0.70 11.1 1.75 179
0.15 0.28 0.53 0.73 0.15 0.88
(5.52; 21.5) (19.4; 40.2) (0.46; 1.14) (8.63; 138.8) (1.20; 3.60) (141; 245)
(3.36; 18.0) (15.2; 35.0) (0.55; 1.26) (8.63; 13.8) (1.20; 5.40) (143; 243
(4.6; 24.7) (18.9; 43.8) (0.43; 1.05) (8.70; 13.7) (0.92; 3.45) (140; 230)
Median and 25th and 75th percentiles are given. P is calculated between baseline and day 2 and baseline and day 30, respectively (entire group: both atorvastain and placebo).
variables that correlated with a p-value of b0.05 (OR 0.98; 95% CI 0.97–1.00, p = 0.04). After adding treatment allocation to the fully adjusted model, PAI-1 mass remained significantly associated with sinus rhythm both at day 2 and 30 (OR 0.98; 95% CI 0.95–1.00). Increased tPA mass was significantly associated with sinus rhythm at day 30 (OR 1.13; 95% CI 1.01–1.27) with the fibrinolytic variables PAI1 mass and tPA activity and CRP, but not when it was further adjusted, Table 3. Discussion Our main finding was that in univariate analysis no component of the fibrinolytic system or vWF was significantly associated with maintenance of sinus rhythm one month after cardioversion. However, in all the other models and after further adjustment for treatment, lower PAI-1 mass was still predictive of sinus rhythm both at 2 days and 30 days after cardioversion. Our findings differ somewhat from those of the CAPRAF study, using a design similar to ours, where Tveit [12] found that baseline PAI-1 activity, rather than PAI mass, was significantly lower in patients in sinus rhythm six months after cardioversion for AF. Their results might depend on different patient selection as our patients had almost twice as high frequency of hypertension, which in CAPRAF was correlated to PAI-1 activity and the fact that they used candesartan instead of atorvastatin in the active treatment arm. Different times for primary endpoint and exclusion of patients with previous cardioversion (one month vs. one year) might also explain the different results. Pretorius [11] has also shown that PAI-1 level was an independent predictor of development of AF after cardiac surgery. The fibrinolytic function in AF has been proposed to depend on endothelial dysfunction, inflammation and insulin resistance [9]. A septic chock with disseminated intravascular coagulation is an apparent example of the link between inflammation and coagulation. The fibrinolytic factors and CRP are known to interact, CRP induces PAI-1 expression and activity in endothelial cells [18] and PAI-1 in turn acts as an acute-phase reactant [13,19]. PAI-1 as a predictor of maintenance of SR seems however independent of inflammatory
effects as adding CRP to our model did not change the result. On the other hand was increased tPA mass significantly associated to maintenance of sinus rhythm in a model with the fibrinolytic variables only when adjusting for CRP. The prognostic value of PAI-1 levels often disappears after adjustment for metabolic factors as PAI-1 is elevated in the insulin resistance syndrome [20]. Contrary we found that PAI-1 mass remained significantly associated with sinus rhythm when adding BMI, cholesterol and hypertension to our model. These results question the role of metabolic factors and inflammation as explanatory factors for PAI-1 as a predictor for rhythm. Future studies on the interaction between inflammatory, metabolic and fibrinolytic factors in AF may enlighten the pathophysiology behind AF. Fibrinolysis has been suggested to affect arterial wall remodelling thereby causing AF [21,22]. PAI-1 inhibits t-PA which means that less plasmin will be formed from plasminogen [23]. Plasmin is important for degradation of extracellular matrix in two ways, it is an activator of matrix metalloproteinase's which regulates collagen matrix homeostasis and is itself a proteolytic enzyme [23,24]. Together metalloproteinase's and plasmin can degrade most extracellular matrix components [23,24]. These interrelated proteolytic systems are thus involved in matrix degradation and arterial wall remodelling and both are inhibited by PAI-1 providing a possible mechanism behind the link between PAI-1 mass and maintenance of SR. Two days after cardioversion there were lower levels of PAI-1 mass, PAI-1 activity and tPA mass suggesting a rapid improvement of fibrinolysis, possible an effect of sinus rhythm, but no difference was seen after 30 days, maybe because of recurrence of AF in some individuals during this time. The finding may thus suggest an improved fibrinolysis in sinus rhythm compared to AF but further research is needed to confirm this finding. When studying those 53 individuals still in sinus rhythm after one month only CRP was significantly lower compared to baseline suggesting that rhythm may have an effect on CRP levels. The increase of CRP two days after cardioversion may be a result of the electrical damage to the underlying tissue and this change did not remain at day 30. The clinical implications of our findings are so far uncertain. There is a relationship between PAI-1 mass and recurrence of AF one month after
Table 3 Univariate and multivariate logistic regression analysis for predicting sinus rhythm 30 days after cardioversion. Odds ratio per unit with 95% confidence interval are presented.
PAI-1 mass (ng/mL) PAI-1 activity (IU/mL) tPA activity (IU/mL) tPA mass (ng/mL) VWF (% of normal) CRP (mg/L) Triglycerides (mmol/L) BMI (kg/m2) Hypertension (yes/no) Cholesterol (mmol/L) Age (years) Smoker (yes/no)
Univariate
Fibrinolytic components
+ hsCRP
+ Metabolic components
All variables, p b 0.05
0.99; 0.97–1.00 0.99; 0.97–1.01 0.91; 0.45–1.82 1.05; 0.96–1.15 1.00; 1,00–1.01 0.95; 0.87–1.05 1.44; 0.96–2.17 0.98; 0.92–1.05 0.68; 0.33–1.39 1.01; 0.70–1.46 1.00; 0.97–1.04 0.28; 0,06–1.37
0.98; 0.96–1.00* Excluded 0,75; 0,34–1,64 1.09; 0.98–1.20 NiM 0.95; 0.85–1.06 NiM NiM NiM NiM NiM NiM
0.98; 0.96–1.00* Excluded 0.75; 0.34–1.68 1.11; 1.00–1.24* NiM NiM NiM NiM NiM NiM NiM
0.98;0.96–1.00* Excluded 0.62; 0.26–1.49 1.13; 1.00–1.28 NiM 0.95; 0.84–1.06 1.47; 0.87–2.48 0.94; 0.86–1.03 0.54; 0,24–1.25 0.73; 0.44–1.20 NiM NiM
0.98;0.96–1.00* Excluded 0.62; 0.25–1.50 1.13; 0.99–1.29 NiM 0.94; 0.84–1.06 1.44; 0.85–2.45 0.94; 0.85–1.04 0.60; 0.25–1.43 0.72; 0.43–1.18 1.00; 0.95–1.04 0.46; 0.08–2.55
Excluded = In multivariate analysis PAI-1 activity was excluded due to high correlation with tPA activity (r = 0,8). NiM = Variable not included in the present model. * denotes p b 0,05.
192
J. Andersson et al. / Thrombosis Research 127 (2011) 189–192
electrical cardioversion but the study does not give us an answer on the clinical usefulness of fibrinolytic markers as predictors for recurrence in AF. It should rather be regarded as an insight to patophysiological mechanisms behind AF. Limitations This is an explorative substudy why the results should only be regarded as hypothesis generating. The number of patients in our explorative study is a limitation; it is underpowered to detect weaker associations. The studied predictors may be dependent on the studied population and there was a relatively large proportion of screened patients who already were on statin treatment (39% in the main study), and therefore not eligible for the study. Our patients may thus represent a cardiovascular low risk group. Another limitation is that periods of asymptomatic AF, which is known to be common, was not detected unless they coincided with the follow-up after cardioversion. Conclusion No component alone of the fibrinolytic system or vWF was found to be a predictor for maintenance of sinus rhythm, while in multivariate models lower PAI-1 mass was associated with sinus rhythm. This finding was maintained after adjusting for CRP and some markers of the metabolic syndrome as well as treatment with atorvastatin. Our findings suggest a patophysiological link between AF and PAI-1 mass but the relation to inflammation remains unclear. Conflict of interest We declare that none of the authors have any financial or other relationships that might lead to a conflict of interest regarding this paper. The manufacturer of atorvastatin, Pfizer, provided study medication and an unrestricted research grant but had no other involvement in the study. Acknowledgement The manufacturer of atorvastatin, Pfizer, provided study medication and an unrestricted research grant but had no other involvement in the study. References [1] Bruins P, Velthuis H, Yazdanbakhsh AP, Jansen PG, Hardevelt FW, Beaumont EM, et al. Activation of the complement system during and after cardiopulmonary bypass surgery: postsurgery activation involves C-reactive protein and is associated with postoperative arrhythmia. Circulation 1997;96:3542–8.
[2] Frustaci A, Chimenti C, Bellocci F, Morgante E, Russo MA, Maseri A. Histological substrate of atrial biopsies in patients with lone atrial fibrillation. Circulation 1997;96:1180–4. [3] Boos CJ, Anderson RA, Lip GY. Is atrial fibrillation an inflammatory disorder? Eur Heart J 2006;29:136–49. [4] Engelmann MD, Svendsen JH. Inflammation in the genesis and perpetuation of atrial fibrillation. Eur Heart J 2005;26:2083–92. [5] Lombardi F, Tundo F, Belletti S, Mantero A, MelziD'eril GV. C-reactive protein but not atrial dysfunction predicts recurrences of atrial fibrillation after cardioversion in patients with preserved left ventricular function. J Cardiovasc Med 2008;9: 581–8. [6] Loricchio ML, Cianfrocca C, Pasceri V, Bianconi L, Auriti A, Calo L, et al. Relation of Creactive protein to long-term risk of recurrence of atrial fibrillation after electrical cardioversion. Am J Cardiol 2007;99:1421–4. [7] Watanabe E, Arakawa T, Uchiyama T, Kodama I, Hishida H. High-sensitivity Creactive protein is predictive of successful cardioversion for atrial fibrillation and maintenance of sinus rhythm after conversion. Int J Cardiol 2006;108:346–53. [8] Freestone B, Lip GY. The endothelium and atrial fibrillation. The prothrombotic state revisited. Hamostaseologie 2008;28:207–12. [9] Marín F, Roldán V, Lip GY. Fibrinolytic function and atrial fibrillation. Thromb Res 2003;109:233–40. [10] Roldán V, Marín F, Blann AD, García A, Marco P, Sogorb F, et al. Interleukin-6, endothelial activation and thrombogenesis in chronic atrial fibrillation. Eur Heart J 2003;24:1373–80. [11] Pretorius M, Donahue BS, Yu C, Greelish JP, Roden DM, Brown NJ. Plasminogen activator inhibitor-1 as a predictor of postoperative atrial fibrillation after cardiopulmonary bypass. Circulation 2007;116:I1–7. [12] Tveit A, Seljeflot I, Grundvold I, Abdelnoor M, Smith P, Arnesen H. Levels of PAI-1 and outcome after electrical cardioversion for atrial fibrillation. Thromb Res 2008;121:447–53. [13] Faber DR, Groot PG, Visseren FL. Role of adipose tissue in haemostasis, coagulation and fibrinolysis. Obes Rev 2009;10:554–63. [14] Shantsila E, Watson T, Lip GY. Cardioversion and remodelling in atrial fibrillation: Insights beyond the prothrombotic state. Thromb Res 2008;121:443–5. [15] Dangas G, Badimon JJ, Smith DA, Unger AH, Levine D, Shao JH, et al. Pravastatin therapy in hyperlipidemia: effects on thrombus formation and the systemic hemostatic profile. J Am Coll Cardiol 1999;33:1294–304. [16] Ridker PM, Cannon CP, Morrow D, Rifai N, Rose LM, et al. C-reactive protein levels and outcomes after statin therapy. N Engl J Med 2005;352:20–8. [17] Almroth H, Höglund N, Boman K, Englund A, Jensen S, Kjellman B, et al. Atorvastatin and persistent atrial fibrillation following cardioversion: a randomized placebo-controlled multicentre study. Eur Heart J 2009;30:827–33. [18] Devaraj S, Xu DY, Jialal I. C-reactive protein increases plasminogen activator inhibitor-1 expression and activity in human aortic endothelial cells: implications for the metabolic syndrome and atherothrombosis. Ciculation 2003;107:398–404. [19] Sawdey MS, Loskutoff DJ. Regulation of murine type 1 plasminogen activator inhibitor gene expression in vivo. Tissue specificity and induction by lipopolysaccharide, tumor necrosis factor-alpha, and transforming growth factor-beta. J Clin Invest 1991;88:1346–53. [20] Juhan-Vague I, Alessi MC, Mavri A, Morange PE. Plasminogen activator inhibitor-1, inflammation, obesity, insulin resistance and vascular risk. J Thromb Haemost 2003;1:1575–9. [21] Moe GW, Laurent G, Doumanovskaia L, Konig A, Hu X, Dorian P. Matrix metalloproteinase inhibition attenuates atrial remodeling and vulnerability to atrial fibrillation in a canine model of heart failure. J Card Fail 2004;14:768–76. [22] Xu J, Cui G, Esmailian F, Plunkett M, Marelli D, Ardehali A, et al. Atrial extracellular matrix remodeling and the maintenance of atrial fibrillation. Ciculation 2004;109: 363–8. [23] Lijnen HR. Plasmin and matrix metalloproteinases in vascular remodeling. Thromb Haemost 2001;86:324–33. [24] Rerolle JP, Hertig A, Nguyen G, Sraer JD, Rondeau EP. Plasminogen activator inhibitor type 1 is a potential target in renal fibrogenesis. Kidney Int 2000;58: 1841–50.
Contents lists available at ScienceDirect
Thrombosis Research j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t h r o m r e s
Regular Article
Markers of fibrinolysis as predictors for maintenance of sinus rhythm after electrical cardioversion Jonas Andersson a,⁎, Henrik Almroth b, Niklas Höglund c, Steen Jensen c, Per Tornvall d, Anders Englund b, Mårten Rosenqvist e, Kurt Boman a a
Department of Medicine and Geriatrics, Skellefteå County Hospital, S-931 86 Skellefteå, Sweden Department of Cardiology, Örebro University Hospital, S-701 85 Örebro, Sweden Department of Cardiology, Heart Centre, University Hospital, S-901 85 Umeå, Sweden d Department of Cardiology, Karolinska University Hospital, S-171 76 Stockholm, Sweden e Department of Cardiology, Karolinska Institutet at Stockholm South Hospital, S-118 83 Stockholm, Sweden b c
a r t i c l e
i n f o
Article history: Received 31 March 2010 Received in revised form 26 May 2010 Accepted 2 June 2010 Available online 14 January 2011
a b s t r a c t Introduction: Inflammation, endothelial dysfunction and metabolic pathways provide possible links between the inflammatory and hypofibrinolytic states in atrial fibrillation. Our aim was to explore the role of mass concentrations of PAI-1 and tPA, activities of PAI-1 and tPA as predictors of recurrence of atrial fibrillation adjusted for CRP. Materials and methods: The study included 129 patients with persistent atrial fibrillation. Laboratory analyses were performed including PAI-1 activity, PAI-1 mass, tPA activity, tPA mass and CRP in baseline. Patients were then randomized to atorvastatin (40 mg, two tablets once daily) or placebo, initiated at least 14 days before the elective cardioversion. Further samples and follow-up were made at day 2 and 30 days after cardioversion. Results: In univariate logistic regression no fibrinolytic variable was significantly correlated with rhythm in day 30. In multivariate analysis lower PAI-1 mass was significantly associated with sinus rhythm in all models including fibrinolytic variables, CRP, metabolic components, age, hypertension and smoking. After adding treatment allocation to the fully adjusted model, PAI-1 mass remained significantly associated with sinus rhythm both at day 2 and 30 (OR 0.98; 95% CI 0.95–1.00). Conclusions: No fibrinolytic component alone was found to be a predictor of recurrence of atrial fibrillation. In multivariate models lower PAI-1 mass was associated with sinus rhythm even after adjusting for CRP, markers of the metabolic syndrome and treatment with atorvastatin. Our findings suggest a patophysiological link between AF and PAI-1 mass but the relation to inflammation remains unclear. © 2010 Elsevier Ltd. All rights reserved.
Introduction The first evidence of atrial fibrillation (AF) as an inflammatory disease emerged in the late 1990s [1,2]. Subsequent studies demonstrated inflammatory infiltrates in atrial biopsies from AF patients and found that CRP was a predictor not only of the presence of and the risk of developing future AF but also the recurrence of AF after electrical cardioversion [3–7]. A hypofibrinolytic state with increased plasminogen activator inhibitor type 1 mass concentration (PAI-1 mass) and its complex with tissue plasminogen activator (t-PA/PAI-1 complex) levels was also described in AF [8–10] and PAI-1 has been found to predict the recurrence of AF [11,12]. Inflammation, endothelial dysfunction and metabolic pathways provided possible links between the inflamma⁎ Corresponding author. Department of Medicine and Geriatrics Skellefteå County Hospital SE–931 86 Skellefteå (Sweden). Tel.: +46 910 771 294; fax: +46 910 771 125. E-mail address: [email protected] (J. Andersson). 0049-3848/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2010.06.002
tory and hypofibrinolytic states in AF [9,13,14]. There is a well known interaction between inflammation and fibrinolysis as components of the fibrinolytic system act as acute phase reactants. The described association between PAI-1 and recurrence of AF may thus reflect an inflammatory state rather than a disturbed fibrinolytic system. Moreover, treatment with statins has been found to reduce both PAI-1 and C-reactive protein (CRP) and suggested as a new treatment option to prevent recurrence of AF [15,16]. We here present new data on components of the fibrinolytic system taking into account the inflammatory state with and without treatment with atorvastatin. Our null hypothesis was the components of the fibrinolytic system do not predict maintenance of SR when adjusted for CRP. Aim Our primary aim was to explore the role of mass concentrations of PAI-1 and tPA, activities of PAI-1 and tPA as predictors of maintenance of SR in a non-postoperative situation with adjustment for CRP.
190
J. Andersson et al. / Thrombosis Research 127 (2011) 189–192
Method
Ethics
We conducted a pre-specified hypothesis generating study within a double-blinded, placebo-controlled, randomized, prospective, investigator-initiated multicentre study of the effects of atorvastatin on recurrence of AF after cardioversion which has previously been described in detail [17]. Between August 2004 and January 2007, a total of 234 patients with persistent AF and an indication for cardioversion were included. Five of the participating centres collected blood samples from their 129 patients for the present study. Persistent AF was defined as AF with duration more than 7 days. Patients below 18 years and over 80 years of age were excluded as were patients with paroxysmal AF, atrial flutter, contraindications against atorvastatin, ongoing treatment with lipidlowering drugs, ongoing treatment with class I or class III antiarrhythmic treatment, known liver disease or a myopathy, as well as patients with a previous electrical cardioversion the last 12 month. Treatment with betareceptor-blocking agents, calcium antagonists, diuretics, digoxin, ACE-inhibitors, ARBs, and acetylsalicylic acid were used when clinically indicated. Warfarin was given to all patients according to national guidelines, i.e. international ratio (INR) of 2–3, Nthree weeks before cardioversion, checked weekly. Before randomization, i.e. baseline, a detailed medical history was obtained and transthoracic echocardiographic examination, 12-lead electrocardiography, and basic laboratory analysis were performed including PAI1 activity, PAI-1 mass, tPA activity, tPA mass, von Willebrand factor (vWF), CRP and blood lipids. Patients were then randomized to atorvastatin (40 mg, two tablets once daily) or placebo, initiated at least 14 days before the elective cardioversion. Further samples and follow-up were made at day 2 and 30 days after cardioversion. Compliance was monitored at cardioversion and 30 days after cardioversion (end of treatment) by pill counting. For analysis of fibrinolytic markers blood were collected into Stabilyte tubes and analysis was carried out using an enzyme-linked immunosorbent assay (ELISA). Chromolize PAI-1 and Chromolize t-PA from Biopool AB, Sweden, was used to determine activities of PAI-1 and t-PA respectively. Tintelize PAI-1 and Tintelize t-PA was used to determine mass concentrations of PAI-1 and tPA respectively. Von Willebrand factor was determined with enzyme-linked immunosorbent assays (ELISA) from DAKO in Denmark. Plasma CRP was determined with an automated high sensitive C-reactive protein method, (IMMULITE, Diagnostic Products Corporation, USA). Cardioversion was performed according to local standard clinical practice on an elective outpatient basis. Patients were instructed to contact the study team if they experienced symptoms suggesting AF between scheduled visits. No patient was lost during follow-up.
The Ethics Committee at Karolinska Institute and the Swedish Medical Product Agency approved this study that also complied with the Declaration of Helsinki. Written informed consent was obtained from all patients.
Statistical analysis The Mann-Whitney U test was used for calculations between groups and the Wilcoxon signed ranks test for calculations within the groups. Chi-square was used for grouped variables when comparing the subgroup with the entire population. Correlation between variables was tested with Spearman´s test. Logistic regression was used to calculate the odds ratio (OR) and 95% confidence interval for components of the fibrinolytic system to predict rhythm at 30 days adjusting for potential confounding factors. In the multivariate analysis for fibrinolytic variables, PAI-1 activity was excluded due to a high correlation (r = 0.80) with tPA activity. In the initial model fibrinolytic variables were analysed, thereafter was the contribution of inflammation (CRP) analysed. The components of the metabolic syndrome (hypertension, triglycerides, BMI and cholesterol) were then added and at age and smoking. For all analyses p b 0.05 (two-tailed) was considered statistically significant. All analysis was performed using Statistical Package for the Social Sciences (SPSS) version 17.0 (Chicago, IL, USA).
Results In total 129 patients (97 males, 75%, 32 females, 25%, and mean age 65) were eligible for the substudy. Baseline characteristics according to rhythm at the 30-day follow up are presented in Table 1. The substudy was well-balanced with the main study with regard to baseline characteristics and treatment (placebo = 64, atorvastatin = 65). Levels of biomarkers at baseline, two days after cardioversion and at the 30 day are presented in Table 2. Only HDL (p = 0.048) was significantly different in baseline between the two treatment groups in the substudy. Effect of cardioversion and rhythm At day 30 recurrent AF was observed in 74 patients and atrial flutter in two. Two days after cardioversion PAI-1 activity, PAI-1 mass and tPA mass had decreased significantly and CRP had increased significantly compared to baseline (Table 2). Among the 53 individuals who remained in sinus rhythm after 30 days PAI-1 activity, tPA activity and tPA mass was significantly lower at day two after cardioversion compared to baseline and in the AF group only PAI-1 mass was significantly lower. In within group analysis there were no significant change between baseline and day 30 in any fibrinolytic variable or vWF among neither those who remained in sinus rhythm nor among those with recurrent AF at day 30. However CRP was significantly lower at day 30 compared to baseline among those who remained in sinus rhythm (p = 0.04). In between group analysis, there were no significant differences for any fibrinolytic variable, vWF or CRP neither at baseline nor at day 30. Among those with recurrent AF at day 30 there were no significant changes in fibrinolytic variables from baseline to day 30. Predictors of rhythm In univariate logistic regression no fibrinolytic variable or vWF was significantly correlated with rhythm in day 30. However, in the multivariate analysis lower baseline PAI-1 mass was significantly associated with sinus rhythm in all models, Table 3. Lower baseline PAI-1 mass was also predictive of sinus rhythm at day 2 (OR 0.98; 95% CI 0.97–1.00, p = 0.02) and remained significant after adjusting for all Table 1 Baseline characteristics according to rhythm at 30-days post cardioversion.
Number (n) Age Sex M/F Smokers Weight Body mass index Total cholesterol HDL LDL Triglycerides Hypertension Stroke ICH Left atrial size Normal ejection fraction
(years) n (% men) n (%) (kg) (kg/m2) (mmol/L) (mmol/L) (mmol/L) (mmol/L) n (%) n (%) n (%) (mm) n (%)
Sinus rhythm
Atrial fibrillation
53
74
67 (59; 75) 42/11 (79) 2 (3.8) 84 (75; 100) 27 (25; 31) 5.20 (4.65; 6.00) 1.26 (1.04; 1.38) 3.27 (2.79; 3.69) 1.59 (1.12; 2.01) 25 (47.2) 2 (3.8) 4 (7.5) 45 (40; 49) 35 (66)
66 (59; 72) 53/21 (72) 9 (12.2) 87 (77; 95) 28 (25; 31) 5.30 (4.90; 5.90) 1.24 (1.06; 1.53) 3.31 (2.79; 3.92) 1.58 (1.00; 1.85) 28 (37.8) 2 (2.7) 4 (5.4) 44 (40; 48) 60 (81)
Median and 25th and 75th percentiles are given.
J. Andersson et al. / Thrombosis Research 127 (2011) 189–192
191
Table 2 Median values for the markers of inflammation and fibrinolysis and VWF.
PAI-1 activity (IU/mL) PAI-1 mass (ng/mL) tPA activity (IU/mL) tPA mass (ng/mL) CRP (mg/L) VWF (% of normal)
Baseline
Day 2 after cardioversion
p
Day 30
p
11.9 27.1 0.79 10.7 2.30 87
7.42 22.8 0.89 0.7 2.30 184
0.001 0.003 0.084 0.000 0.025 0.662
12.6 28.1 0.70 11.1 1.75 179
0.15 0.28 0.53 0.73 0.15 0.88
(5.52; 21.5) (19.4; 40.2) (0.46; 1.14) (8.63; 138.8) (1.20; 3.60) (141; 245)
(3.36; 18.0) (15.2; 35.0) (0.55; 1.26) (8.63; 13.8) (1.20; 5.40) (143; 243
(4.6; 24.7) (18.9; 43.8) (0.43; 1.05) (8.70; 13.7) (0.92; 3.45) (140; 230)
Median and 25th and 75th percentiles are given. P is calculated between baseline and day 2 and baseline and day 30, respectively (entire group: both atorvastain and placebo).
variables that correlated with a p-value of b0.05 (OR 0.98; 95% CI 0.97–1.00, p = 0.04). After adding treatment allocation to the fully adjusted model, PAI-1 mass remained significantly associated with sinus rhythm both at day 2 and 30 (OR 0.98; 95% CI 0.95–1.00). Increased tPA mass was significantly associated with sinus rhythm at day 30 (OR 1.13; 95% CI 1.01–1.27) with the fibrinolytic variables PAI1 mass and tPA activity and CRP, but not when it was further adjusted, Table 3. Discussion Our main finding was that in univariate analysis no component of the fibrinolytic system or vWF was significantly associated with maintenance of sinus rhythm one month after cardioversion. However, in all the other models and after further adjustment for treatment, lower PAI-1 mass was still predictive of sinus rhythm both at 2 days and 30 days after cardioversion. Our findings differ somewhat from those of the CAPRAF study, using a design similar to ours, where Tveit [12] found that baseline PAI-1 activity, rather than PAI mass, was significantly lower in patients in sinus rhythm six months after cardioversion for AF. Their results might depend on different patient selection as our patients had almost twice as high frequency of hypertension, which in CAPRAF was correlated to PAI-1 activity and the fact that they used candesartan instead of atorvastatin in the active treatment arm. Different times for primary endpoint and exclusion of patients with previous cardioversion (one month vs. one year) might also explain the different results. Pretorius [11] has also shown that PAI-1 level was an independent predictor of development of AF after cardiac surgery. The fibrinolytic function in AF has been proposed to depend on endothelial dysfunction, inflammation and insulin resistance [9]. A septic chock with disseminated intravascular coagulation is an apparent example of the link between inflammation and coagulation. The fibrinolytic factors and CRP are known to interact, CRP induces PAI-1 expression and activity in endothelial cells [18] and PAI-1 in turn acts as an acute-phase reactant [13,19]. PAI-1 as a predictor of maintenance of SR seems however independent of inflammatory
effects as adding CRP to our model did not change the result. On the other hand was increased tPA mass significantly associated to maintenance of sinus rhythm in a model with the fibrinolytic variables only when adjusting for CRP. The prognostic value of PAI-1 levels often disappears after adjustment for metabolic factors as PAI-1 is elevated in the insulin resistance syndrome [20]. Contrary we found that PAI-1 mass remained significantly associated with sinus rhythm when adding BMI, cholesterol and hypertension to our model. These results question the role of metabolic factors and inflammation as explanatory factors for PAI-1 as a predictor for rhythm. Future studies on the interaction between inflammatory, metabolic and fibrinolytic factors in AF may enlighten the pathophysiology behind AF. Fibrinolysis has been suggested to affect arterial wall remodelling thereby causing AF [21,22]. PAI-1 inhibits t-PA which means that less plasmin will be formed from plasminogen [23]. Plasmin is important for degradation of extracellular matrix in two ways, it is an activator of matrix metalloproteinase's which regulates collagen matrix homeostasis and is itself a proteolytic enzyme [23,24]. Together metalloproteinase's and plasmin can degrade most extracellular matrix components [23,24]. These interrelated proteolytic systems are thus involved in matrix degradation and arterial wall remodelling and both are inhibited by PAI-1 providing a possible mechanism behind the link between PAI-1 mass and maintenance of SR. Two days after cardioversion there were lower levels of PAI-1 mass, PAI-1 activity and tPA mass suggesting a rapid improvement of fibrinolysis, possible an effect of sinus rhythm, but no difference was seen after 30 days, maybe because of recurrence of AF in some individuals during this time. The finding may thus suggest an improved fibrinolysis in sinus rhythm compared to AF but further research is needed to confirm this finding. When studying those 53 individuals still in sinus rhythm after one month only CRP was significantly lower compared to baseline suggesting that rhythm may have an effect on CRP levels. The increase of CRP two days after cardioversion may be a result of the electrical damage to the underlying tissue and this change did not remain at day 30. The clinical implications of our findings are so far uncertain. There is a relationship between PAI-1 mass and recurrence of AF one month after
Table 3 Univariate and multivariate logistic regression analysis for predicting sinus rhythm 30 days after cardioversion. Odds ratio per unit with 95% confidence interval are presented.
PAI-1 mass (ng/mL) PAI-1 activity (IU/mL) tPA activity (IU/mL) tPA mass (ng/mL) VWF (% of normal) CRP (mg/L) Triglycerides (mmol/L) BMI (kg/m2) Hypertension (yes/no) Cholesterol (mmol/L) Age (years) Smoker (yes/no)
Univariate
Fibrinolytic components
+ hsCRP
+ Metabolic components
All variables, p b 0.05
0.99; 0.97–1.00 0.99; 0.97–1.01 0.91; 0.45–1.82 1.05; 0.96–1.15 1.00; 1,00–1.01 0.95; 0.87–1.05 1.44; 0.96–2.17 0.98; 0.92–1.05 0.68; 0.33–1.39 1.01; 0.70–1.46 1.00; 0.97–1.04 0.28; 0,06–1.37
0.98; 0.96–1.00* Excluded 0,75; 0,34–1,64 1.09; 0.98–1.20 NiM 0.95; 0.85–1.06 NiM NiM NiM NiM NiM NiM
0.98; 0.96–1.00* Excluded 0.75; 0.34–1.68 1.11; 1.00–1.24* NiM NiM NiM NiM NiM NiM NiM
0.98;0.96–1.00* Excluded 0.62; 0.26–1.49 1.13; 1.00–1.28 NiM 0.95; 0.84–1.06 1.47; 0.87–2.48 0.94; 0.86–1.03 0.54; 0,24–1.25 0.73; 0.44–1.20 NiM NiM
0.98;0.96–1.00* Excluded 0.62; 0.25–1.50 1.13; 0.99–1.29 NiM 0.94; 0.84–1.06 1.44; 0.85–2.45 0.94; 0.85–1.04 0.60; 0.25–1.43 0.72; 0.43–1.18 1.00; 0.95–1.04 0.46; 0.08–2.55
Excluded = In multivariate analysis PAI-1 activity was excluded due to high correlation with tPA activity (r = 0,8). NiM = Variable not included in the present model. * denotes p b 0,05.
192
J. Andersson et al. / Thrombosis Research 127 (2011) 189–192
electrical cardioversion but the study does not give us an answer on the clinical usefulness of fibrinolytic markers as predictors for recurrence in AF. It should rather be regarded as an insight to patophysiological mechanisms behind AF. Limitations This is an explorative substudy why the results should only be regarded as hypothesis generating. The number of patients in our explorative study is a limitation; it is underpowered to detect weaker associations. The studied predictors may be dependent on the studied population and there was a relatively large proportion of screened patients who already were on statin treatment (39% in the main study), and therefore not eligible for the study. Our patients may thus represent a cardiovascular low risk group. Another limitation is that periods of asymptomatic AF, which is known to be common, was not detected unless they coincided with the follow-up after cardioversion. Conclusion No component alone of the fibrinolytic system or vWF was found to be a predictor for maintenance of sinus rhythm, while in multivariate models lower PAI-1 mass was associated with sinus rhythm. This finding was maintained after adjusting for CRP and some markers of the metabolic syndrome as well as treatment with atorvastatin. Our findings suggest a patophysiological link between AF and PAI-1 mass but the relation to inflammation remains unclear. Conflict of interest We declare that none of the authors have any financial or other relationships that might lead to a conflict of interest regarding this paper. The manufacturer of atorvastatin, Pfizer, provided study medication and an unrestricted research grant but had no other involvement in the study. Acknowledgement The manufacturer of atorvastatin, Pfizer, provided study medication and an unrestricted research grant but had no other involvement in the study. References [1] Bruins P, Velthuis H, Yazdanbakhsh AP, Jansen PG, Hardevelt FW, Beaumont EM, et al. Activation of the complement system during and after cardiopulmonary bypass surgery: postsurgery activation involves C-reactive protein and is associated with postoperative arrhythmia. Circulation 1997;96:3542–8.
[2] Frustaci A, Chimenti C, Bellocci F, Morgante E, Russo MA, Maseri A. Histological substrate of atrial biopsies in patients with lone atrial fibrillation. Circulation 1997;96:1180–4. [3] Boos CJ, Anderson RA, Lip GY. Is atrial fibrillation an inflammatory disorder? Eur Heart J 2006;29:136–49. [4] Engelmann MD, Svendsen JH. Inflammation in the genesis and perpetuation of atrial fibrillation. Eur Heart J 2005;26:2083–92. [5] Lombardi F, Tundo F, Belletti S, Mantero A, MelziD'eril GV. C-reactive protein but not atrial dysfunction predicts recurrences of atrial fibrillation after cardioversion in patients with preserved left ventricular function. J Cardiovasc Med 2008;9: 581–8. [6] Loricchio ML, Cianfrocca C, Pasceri V, Bianconi L, Auriti A, Calo L, et al. Relation of Creactive protein to long-term risk of recurrence of atrial fibrillation after electrical cardioversion. Am J Cardiol 2007;99:1421–4. [7] Watanabe E, Arakawa T, Uchiyama T, Kodama I, Hishida H. High-sensitivity Creactive protein is predictive of successful cardioversion for atrial fibrillation and maintenance of sinus rhythm after conversion. Int J Cardiol 2006;108:346–53. [8] Freestone B, Lip GY. The endothelium and atrial fibrillation. The prothrombotic state revisited. Hamostaseologie 2008;28:207–12. [9] Marín F, Roldán V, Lip GY. Fibrinolytic function and atrial fibrillation. Thromb Res 2003;109:233–40. [10] Roldán V, Marín F, Blann AD, García A, Marco P, Sogorb F, et al. Interleukin-6, endothelial activation and thrombogenesis in chronic atrial fibrillation. Eur Heart J 2003;24:1373–80. [11] Pretorius M, Donahue BS, Yu C, Greelish JP, Roden DM, Brown NJ. Plasminogen activator inhibitor-1 as a predictor of postoperative atrial fibrillation after cardiopulmonary bypass. Circulation 2007;116:I1–7. [12] Tveit A, Seljeflot I, Grundvold I, Abdelnoor M, Smith P, Arnesen H. Levels of PAI-1 and outcome after electrical cardioversion for atrial fibrillation. Thromb Res 2008;121:447–53. [13] Faber DR, Groot PG, Visseren FL. Role of adipose tissue in haemostasis, coagulation and fibrinolysis. Obes Rev 2009;10:554–63. [14] Shantsila E, Watson T, Lip GY. Cardioversion and remodelling in atrial fibrillation: Insights beyond the prothrombotic state. Thromb Res 2008;121:443–5. [15] Dangas G, Badimon JJ, Smith DA, Unger AH, Levine D, Shao JH, et al. Pravastatin therapy in hyperlipidemia: effects on thrombus formation and the systemic hemostatic profile. J Am Coll Cardiol 1999;33:1294–304. [16] Ridker PM, Cannon CP, Morrow D, Rifai N, Rose LM, et al. C-reactive protein levels and outcomes after statin therapy. N Engl J Med 2005;352:20–8. [17] Almroth H, Höglund N, Boman K, Englund A, Jensen S, Kjellman B, et al. Atorvastatin and persistent atrial fibrillation following cardioversion: a randomized placebo-controlled multicentre study. Eur Heart J 2009;30:827–33. [18] Devaraj S, Xu DY, Jialal I. C-reactive protein increases plasminogen activator inhibitor-1 expression and activity in human aortic endothelial cells: implications for the metabolic syndrome and atherothrombosis. Ciculation 2003;107:398–404. [19] Sawdey MS, Loskutoff DJ. Regulation of murine type 1 plasminogen activator inhibitor gene expression in vivo. Tissue specificity and induction by lipopolysaccharide, tumor necrosis factor-alpha, and transforming growth factor-beta. J Clin Invest 1991;88:1346–53. [20] Juhan-Vague I, Alessi MC, Mavri A, Morange PE. Plasminogen activator inhibitor-1, inflammation, obesity, insulin resistance and vascular risk. J Thromb Haemost 2003;1:1575–9. [21] Moe GW, Laurent G, Doumanovskaia L, Konig A, Hu X, Dorian P. Matrix metalloproteinase inhibition attenuates atrial remodeling and vulnerability to atrial fibrillation in a canine model of heart failure. J Card Fail 2004;14:768–76. [22] Xu J, Cui G, Esmailian F, Plunkett M, Marelli D, Ardehali A, et al. Atrial extracellular matrix remodeling and the maintenance of atrial fibrillation. Ciculation 2004;109: 363–8. [23] Lijnen HR. Plasmin and matrix metalloproteinases in vascular remodeling. Thromb Haemost 2001;86:324–33. [24] Rerolle JP, Hertig A, Nguyen G, Sraer JD, Rondeau EP. Plasminogen activator inhibitor type 1 is a potential target in renal fibrogenesis. Kidney Int 2000;58: 1841–50.