- Email: [email protected]
The safety and efficacy of vitamin K antagonist in patients with atrial fibrillation and liver cirrhosis
International Journal of Cardiology 180 (2015) 185–191
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
The safety and efficacy of vitamin K antagonist in patients with atrial fibrillation and liver cirrhosis Seung-Jun Lee, Jae-Sun Uhm, Jong-Youn Kim, Hui-Nam Pak, Moon-Hyoung Lee, Boyoung Joung ⁎ Cardiology Division, Department of Internal Medicine, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea
a r t i c l e
i n f o
Article history: Received 6 June 2014 Received in revised form 14 October 2014 Accepted 23 November 2014 Available online 27 November 2014 Keywords: Atrial fibrillation Stroke Hemorrhage Oral anticoagulant Vitamin K antagonist Liver cirrhosis
a b s t r a c t Background: Bleeding is a major concern in treatment of atrial fibrillation (AF) with vitamin K antagonist (VKA). This concern is more emphasized in patients with high bleeding risk such as liver cirrhosis (LC). Methods and results: We retrospectively analyzed incidence of stroke and major bleeding in 321 AF patients with LC, including early [Child-Pugh (CP)-A, n = 215] and advanced [CP-B or C, n = 106] LC according to VKA prescription. The CHA2DS2-VASc, HAS-BLED and MELD scores were higher in patients with VKA. CP score was positively correlated with HAS-BLED score (rho: 0.602). The incidence of major bleeding was higher in advanced LC than in early LC (14.5%/year vs. 4.9%/year, p b 0.001). VKA reduced the risk of ischemic stroke in AF patients with LC, whereas it significantly increased the major bleeding risk. There was no difference in survival free from significant clinical events (SCEs) between the patients with or without VKA (p = 0.91). On subgroup analysis, VKA was beneficial in early LC patients, as it decreased stroke without increasing major bleeding risk. However, in advanced LC patients, VKA significantly increased the risk of major bleeding, especially variceal origin, that overwhelms stroke reduction. As a result, VKA treatment reduced the risk of SCEs in early LC patients, whereas it increased the risk of SCEs in advanced LC. Conclusions: VKA treatment might be beneficial to reduce significant clinical events in the early LC but not in the advanced LC group. However to confirm this hypothesis, a prospective randomized study is needed. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, occurring in 1–2% of the general population [1,2]. AF confers a five-fold risk of stroke, and one in five of all strokes is attributed to this arrhythmia. Multiple clinical trials have demonstrated the superior therapeutic effect of vitamin K antagonist compared with placebo in the prevention of thromboembolic events among patients with nonvalvular AF [3]. However, some AF patients have the concomitant risk of fatal bleeding, which causes clinicians to be reluctant to use vitamin K antagonist (VKA) in spite of the high stroke risk [4,5]. Chronic liver disease is mainly caused by excess alcohol intake and viral hepatitis, which has been classified as the tenth most frequent cause of death in the United States [6]. Moreover, the prevalence of liver cirrhosis (LC) is quite higher in eastern Asia or Africa, which is a major cause of morbidity and mortality in these areas [7,8]. In particular, LC patients usually suffer from conditions that cause upper
Abbreviations: AF, atrial fibrillation; CP, Child-Pugh; LC, liver cirrhosis; SCE, significant clinical event; VKA, vitamin K antagonist ⁎ Corresponding author at: Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752 Republic of Korea. E-mail address: [email protected] (B. Joung).
http://dx.doi.org/10.1016/j.ijcard.2014.11.183 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
gastrointestinal bleeding, including esophageal varices, gastric varices, and peptic ulcer disease [9,10]. In addition, the incidence of spontaneous intracerebral hemorrhage is significantly higher in LC patients with bleeding diathesis due to thrombocytopenia, splenomegaly, or prolonged international normalized ratio (INR) [11,12]. The bleeding risk schemas, such as HAS-BLED and HEMORR2HAGES, have shown that abnormal liver function or chronic liver disease is a major bleeding risk in AF patients [13]. However, sparse data exist regarding the benefit of oral anticoagulation therapy use in AF patients with LC, as most trials have excluded patients with chronic liver disease or LC [14–17]. Furthermore, anemia, thrombocytopenia and bleeding diathesis are the main clinical features of LC patients which were excluded from most of sophisticatedly designed AF trials [18,19]. Hence, little is known about the importance of VKA treatment in AF patients with LC. Meanwhile, some researches have shown that LC patients with portal vein thrombosis (PVT) are very vulnerable to life threatening bleeding complications [20,21], although VKA had clinical benefit in re-canalizing portal vein thrombosis and preventing ischemic complications [21,22]. We hypothesized that the VKA treatment in AF patients with LC would increase major bleeding and decrease thromboembolic events. Moreover, the increased bleeding risk might be outweighed by the decreased thromboembolic events in patients who were maintained in optimal INR level. This study evaluated the clinical courses, including
186
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
ischemic stroke and major bleeding events, of AF patients with early or advanced LC status according to the use of VKA. 2. Methods 2.1. Study subjects The study protocol was approved by the Institutional Review Board of Severance Cardiovascular Hospital, Seoul, Korea and complied with the Declaration of Helsinki. Between January 2001 and January 2012, using ICD-9 codes, we identified 443 consecutive patients with non-valvular AF and LC over the age of 18. We defined non-valvular AF by the discharge diagnosis of AF (ICD-9 code 427.31), no previous diagnosis of mitral stenosis or prosthetic heart valve (ICD-9 codes 394.0, 394.2, 396.0, 396.1, 396.8, V43.3, or V42.4), and no aortic or mitral valve surgery (ICD-9 codes 35.10 to 35.14 or 35.20 to 35.28). Patients with severe thrombocytopenia (platelets b50,000/mm3, n = 32) [23], insufficient follow up duration (b6 months, n = 27), evidence of renal failure (n = 26), previous intracerebral hemorrhage (n = 15), liver transplantation (n = 5), and multiple missing data (n = 17) were excluded. Finally, 321 patients were included in this study. All of VKA prescribed were warfarin (Coumadin®). LC was defined by clinical diagnosis, biochemistry, Doppler ultrasound, computed tomography scan, and biopsy, when performed [24,25]. The diagnosis was made histologically in 58 patients in whom liver biopsy was performed to confirm the status of liver cirrhosis. It was clinically diagnosed in 263 patients. The clinical diagnosis of liver cirrhosis was made based on the presence of ascites or esophageal varices without other explainable causes, along with unequivocal image on ultrasound or computed tomography. Among 263 patients diagnosed without biopsy, 176 (64%) patients had been performed transient elastography in order to evaluate the degree of hepatic fibrosis non-invasively [26]. The median liver stiffness measurement value for these patients was 18.3 (7.6– 34.3) which is highly suggestive of liver cirrhosis [27]. AF was documented by 12-lead electrocardiography or 24-hour Holter recording. Patients' medical records were reviewed for information on patient age, sex, weight and height, drug therapy, and AF duration. Patient databases were searched to identify known or putative risk factors for ischemic stroke, heart failure, hypertension, diabetes mellitus, hyperlipidemia, tobacco, or alcohol use [28]. Heart failure was defined when hospitalized patients exhibited appropriate symptoms (shortness of breath, fatigue, fluid retention, or any combination of these symptoms) and clinical signs of fluid retention (pulmonary or peripheral) with explainable abnormality of cardiac structure and function. The CHADS2 [Congestive heart failure, Hypertension, Age, Diabetes mellitus, prior Stroke or transient ischemic attack (doubled)] risk index, CHA2DS2-VASc [Congestive heart failure, hypertension, Age ≥75 (doubled), Diabetes mellitus, prior Stroke or transient ischemic attack (doubled)-Vascular disease, Age 65–74 years and Sex category (female)] risk index, and the HAS-BLED [Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly (N65), Drugs/alcohol concomitantly] scores were evaluated. We divided patients into two groups according to their liver cirrhosis status. The Child-Pugh score was used to assess the prognosis of cirrhosis. The score employs five clinical measurements of liver disease; total bilirubin (b2, 2–3, N3 mg/dl), serum albumin (N3.5, 2.8–3.5, b2.8 g/dl), PT INR (b1.7, 1.7–2.3, N2.3), ascites (none, mild, moderate to severe) and hepatic encephalopathy (none, G I–II, G III–IV). Each measure is scored from 1–3, following the order in parentheses above. In this study, the patients with cirrhosis were classified as early LC (score 5–6, n = 215) and advanced LC (score ≥7, n = 106). To quantify end-stage liver disease, the Model For End-Stage Liver Disease (MELD) score (0.957 * ln(serum Cr) +0.378 * ln(serum Bilirubin) +1.120 * ln(INR) +0.643) * 10 was calculated. 2.2. Clinical follow-up Patients visited specialized cardiologists and hepatologists with mean interval of 2.3 ± 0.7 months to check clinical symptoms, INR, biochemistry results and imaging studies. We evaluated ischemic stroke and major bleeding events during follow-up by review of the charts. Ischemic stroke was defined as a neurological deficit of sudden onset that persisted for N24 h corresponding to a vascular territory in the absence of primary hemorrhage and was not explained by other causes, including trauma, infection, or vasculitis [29] . Major bleeding was defined as any central nervous system (CNS) bleeding which includes intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), subdural hemorrhage (SDH), and epidural hemorrhage (EDH); any bleeding requiring transfusion of at least two units of red blood cells or equivalent of whole blood in 24 h; or bleeding events that caused hypotension (systolic blood pressure b90 mm Hg), multi-organ failure, or death. We defined a significant clinical event (SCE) as a first major event that occurred during follow-up period including stroke, major bleeding episode or death [30]. When a patient experienced both ischemic stroke and major bleeding events during the follow-up period, each event was counted. However, when we analyzed the Kaplan–Meier cumulative SCE free survival, we counted the first event only. 2.3. Net clinical benefit assessment We assessed the net clinical benefit (NCB) by calculating the difference of annualized incidence rate (IR) of ischemic stroke and major bleeding multiplied by weighting factors. We adopted the weighting factor derived from the large-scale clinical trial (ACTIVE trial)
by Connolly et al. [31] that measured the adjusted hazard ratio (HR) for death after the event standardized to the adjusted HR of ischemic stroke (IS, weight 1.0). The relative weights for each clinical event are 3.08 for hemorrhagic stroke (HS) which includes ICH and SAH, 0.60 for other CNS bleeding (Other_CNS) that includes SDH and EDH, and 0.67 for extracranial bleeding, respectively. As a result, we calculated the NCB according to the following equation: NCB ¼ ½IRIS
no VKA
þ 3:08 IRHS
IRExtracranial
no VKA
– ½IRIS
þ 0:67 IRExtracranial
no VKA
VKA
þ 0:60 IROther
þ 3:08 IRHS
VKA
CNS no VKA
þ 0:67
þ 0:60 IROther
CNS VKA
VKA :
The resulting values were regarded as ischemic stroke equivalents prevented by VKA per 100 patient-years. 2.4. Intensity and quality of anticoagulation In patients with concomitant indications for VKA treatment, the decision was made by the physician's clinical evaluation of the risk for thrombotic and hemorrhagic events. Among 321 patients enrolled in this study, 173 (54%) were treated with VKA. Indications for VKA therapy retrieved from the chart are as follows: atrial fibrillation, n = 144 (83%); previous portal vein thrombosis, n = 10 (6%); previous deep vein thrombosis, n = 6 (3%); and not documented, n = 13 (8%). The intensity of anticoagulation was determined by the INR values. The INR values at each outpatient clinic/emergency department visit and during hospital admission were retrieved from the medical record. Data on the first four weeks after initiation of warfarin therapy were excluded from the analysis. All INR testing was performed in the same laboratory using a fully automated hematology analyzer (ACL TOP 700 CTS, Instrumentation Laboratory, Werfen Group, Barcelona, Spain) with the same reagent (Hemosil Recombiplastin, Instrumentation Laboratory). International sensitivity index (ISI) values assigned specifically for our laboratory's reagent/instrument combination were used for calculation of INR. Local ISI verification was performed more than twice per year according to the Clinical and Laboratory Standards Institute H54-A guideline. Mean INR values and time in therapeutic range (TTR) of INR 2.0–3.0 were calculated using the linear interpolation methods proposed by Rosendaal et al. [32]. This method assumes that the INR values between two consecutive measurements vary linearly. Patients with an interval of 56 days or more between INR tests were excluded from the TTR analysis as they increase error for interpolation analysis [33]. Nine patients of early LC with VKA (8%) and 5 patients of advanced LC with VKA (8%) were excluded from TTR analysis. 2.5. Statistical analysis Continuous variables that were normally distributed were reported as mean ± standard deviation (SD) and were compared by use of a Student's t-test for parametric data and Mann–Whitney test for nonparametric data. Categorical variables were reported as count (percentage) and were compared using Chi-square or Fisher's exact test. Event rates were compared using mid-p exact test. Multivariate Cox proportional-hazard models were used to identify factors associated with ischemic stroke or major bleeding. Missing binary data, of which proportion are lesser than 2% for each covariate, were coded as “not present” and missing categorical data were coded as “missing”. Kaplan–Meier survival curves were plotted according to use of VKA and compared by means of the log-rank test. Pearson's correlation analysis was employed to assess the relationship between the CP score and the CHADS2, CHA2DS2-VASc, or HAS-BLED scoring system. The SPSS statistical package (SPSS Inc., Chicago, Illinois) version 18.0 was used to perform all statistical evaluations. A p value b 0.05 was considered statistically significant.
3. Results 3.1. Characteristics of patients Clinical characteristics of patients with or without VKA are presented in Table 1. There was no difference in the mean duration of follow-up for patients with or without VKA (3.5 ± 2.3 vs. 3.2 ± 2.4 years, p = 0.29). The mean CP scores of the patients with and without VKA were 7.5 ± 2.3 and 7.0 ± 1.9, respectively (p = 0.06). The MELD score was significantly higher in patients with VKA (14.0 ± 10.7 vs. 8.4 ± 7.3, p b 0.001). Patients without VKA had more frequent seropositivity for HBsAg (49% vs. 28%, p = 0.001). Patients on VKA had more history of ischemic stroke or transient ischemic attack (25% vs. 16%, p = 0.04). The risk of ischemic event calculated by CHADS2 and CHA2DS2-VASc was significantly higher in patients with VKA. In addition, the bleeding risk was also higher in patients with VKA (HAS-BLED score 4.1 ± 1.4 vs. 3.8 ± 1.3, p = 0.02). The characteristics of the patients in each group according to the status of LC are presented in Supplementary Table 1. Mean TTR of the VKA treated group calculated by Rosendaal's method was 59.7 ± 10.0%. Especially, early LC group showed significantly higher TTR compared to advanced LC patients (64.6 ± 10.3% vs. 51.6 ± 9.6%,
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191 Table 1 Characteristics of patients. Characteristic
VKA (n = 173)
No VKA (n = 148)
p Value
Age (years) Female, n (%) Atrial fibrillation type Permanent Persistent Paroxysmal Hypertension Diabetes mellitus Heart failure History of stroke/transient ischemic attack Peptic ulcer (active, healing, scarring) Varix (gastric or esophageal) Thrombocytopenia, moderatea Child-Pugh score MELD score CHADS2 CHA2DS2-VASc HAS-BLED Etiology B-viral C-viral Alcoholic Others Time in therapeutic range (%) Concomitant medications Anti-platelet agent Non-steroidal anti-inflammatory drug Proton pump inhibitor H2 blocker
62.1 ± 10.3 53 (31%)
62.5 ± 11.3 48 (32%)
0.77 0.73
112 (65%) 29 (17%) 31 (18%) 82 (47%) 67 (39%) 56 (32%) 44 (25%) 40 (23%) 39 (23%) 46 (27%) 7.5 ± 2.3 14.0 ± 10.7 2.1 ± 1.5 2.6 ± 1.8 4.1 ± 1.4
95 (64%) 24 (16%) 29 (20%) 62 (42%) 50 (34%) 37 (25%) 24 (16%) 29 (20%) 36 (24%) 46 (31%) 7.0 ± 1.9 8.4 ± 7.3 1.7 ± 1.3 2.2 ± 1.5 3.8 ± 1.3
0.92 0.89 0.70 0.36 0.36 0.15 0.04 0.45 0.71 0.38 0.06 b0.001 0.02 0.03 0.02
48 (28%) 29 (17%) 58 (34%) 33 (19%) 59.7 ± 10.0
72 (49%) 20 (14%) 42 (28%) 21 (14%) –
0.001 0.42 0.43 0.30 –
24 (14%) 29 (17%) 96 (55%) 92 (53%)
33 (22%) 22 (15%) 86 (58%) 93 (63%)
0.05 0.72 0.75 0.10
187
unadjusted HR (relative risk 3.01; 95% CI, 1.97–4.61, p b 0.001) and adjusted HR (relative risk 2.20; 95% CI, 1.39–3.46, p = 0.001) of major bleeding. 3.4. Screening for gastrointestinal lesion and bleeding prophylaxis There was no difference in the presence of peptic ulcer or varix between the patients with or without VKA (Table 1). Among 173 patients who were treated with VKA, 39 patients had gastric or esophageal varices. Among them, 11 patients (27%) with previous variceal bleeding, grade II esophageal varices with red signs or grade III varices were treated by endoscopic variceal ligation (EVL) before VKA prescription. All patients who underwent EVL, started VKA at least 1 month after the eradication of varices. In addition, among 173 patients on VKA, 40 (23%) individuals had peptic ulcer disease and 13 had active ulcer lesion. All patients with active ulcer (n = 13) started VKA after treatment of ulcer and 7 of them (54%) were performed follow-up esophagogastroduodenoscopy for confirmation of ulcer healing. 3.5. Reduction of ischemic stroke by VKA in LC patients
Mean values ± standard deviation or number and percentage of patients. CHADS2 = Congestive heart failure, Hypertension, Age ≥75, Diabetes mellitus, and prior Stroke or transient ischemic attack; CHA2DS2-VASc = Congestive heart failure, Hypertension, Age ≥75 (doubled), Diabetes mellitus, prior Stroke or transient ischemic attack (doubled)-Vascular disease, Age 65–74 years, and Sex category (female); HAS-BLED = Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly (N65), Drugs/alcohol concomitantly. a Platelet count: 50,000–150,000 cells/mm3.
p b 0.001). Patients without VKA showed higher prescription rate of anti-platelet agents. The indications for anti-platelet treatment in patients without VKA included stroke prevention (n = 17), ischemic heart disease (n = 12) and history of thrombosis (n = 4). However, in patients with VKA, the most common cause of anti-platelet therapy was ischemic heart disease (n = 21).
Table 2 shows clinical events during the follow-up period. The incidence of ischemic stroke was significantly lower in patients with VKA compared to those without (1.8%/year vs. 4.7%/year, p = 0.01). Fig. 1A shows the Kaplan–Meier cumulative survival free from ischemic stroke in AF patients with LC according to the use of VKA. Multivariate Cox analysis revealed that the use of VKA reduced the risk of ischemic stroke in AF patients with LC; unadjusted HR (relative risk 0.32; 95% CI, 0.15– 0.66, p = 0.002), and adjusted HR (relative risk 0.30; 95% CI, 0.14– 0.62, p = 0.001, Table 3). 3.6. VKA treatment increased the risk of major bleeding The incidence of major bleeding was significantly higher in patients with VKA (9.6%/year vs. 6.2%/year, p = 0.04). Especially, GI tract bleeding risk was significantly increased by VKA (Table 2, p = 0.03). The incidence of ulcer bleeding was not increased by VKA, whereas variceal bleeding risk was significantly higher in patients with VKA (p = 0.04). Cumulative survival free from the major bleeding is depicted in Fig. 1B (log-rank p = 0.02). VKA prescription significantly increased the risk of major bleeding; unadjusted HR (relative risk 1.68; 95% CI, 1.08– 2.61, p = 0.02), and adjusted HR (relative risk 1.87; 95% CI, 1.13–3.09, p = 0.02). 3.7. Bleeding risk canceled out the beneficial effect of VKA in LC patients
3.2. The correlation between CP score and stroke or bleeding risk estimation scoring systems We investigated the correlation between CP score and stroke or bleeding risk estimation scoring systems. The CHADS2 (Spearman's rho: 0.088; p = 0.12) and CHA2DS2-VASc (rho: 0.086; p = 0.13) scores were not significantly correlated with the CP score. However, a statistically significant correlation was found between the HAS-BLED score and the CP score (rho: 0.602; p b 0.001). 3.3. High risk of major bleeding in advanced LC patients The incidence of major bleeding was significantly higher in advanced LC patients than in early ones (14.5%/year vs. 4.9%/year, p b 0.001). Among the various types of major bleeding, gastrointestinal (p b 0.001) and CNS bleeding (p = 0.05) were more frequently observed in advanced LC group compared to early LC. Cumulative survival free from major bleeding according to hepatic functional reserve is depicted in Supplementary Figure 1. Early LC patients showed higher survival free from major bleeding compared to advanced LC patients (p b 0.001). A CP score higher than 6 (Grade B or C) increased the
We compared the composite outcome of ischemic stroke and major bleeding according to VKA treatment. There was no significant difference in the incidence of SCE between the two groups (p = 0.84; Table 2). Kaplan–Meier cumulative survival free from SCE was not changed by the VKA treatment (log-rank p = 0.91; Fig. 2). 3.8. Subgroup analysis and net clinical benefit There was no significant difference in the risk of SCE between the patients with or without VKA. We compared the clinical outcome according to hepatic functional reserve in addition to VKA prescription. Patients with CP-score lesser than 7 (CP-A) was regarded as early LC, whereas patients with signs of decompensation (CP grade B and C) was defined as advanced LC. There was significant difference in the incidence of ischemic stroke between the patients with or without VKA only in early stage LC patients (1.5%/year vs. 5.5%/year, p = 0.003, Table 2). Moreover, the incidence of major bleeding was significantly higher only in advanced LC patients (18.0%/year vs. 9.2%/year, p = 0.01, Table 2). Especially, the incidence of GI tract bleeding was significantly higher in advanced LC patients with VKA (12.3%/year vs.
188
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
Table 2 Comparison of clinical events during follow-up period in early or advanced LC patients according to VKA treatment. The numbers in parentheses represent event rates for 100 personyears.
Ischemic stroke Major bleeding GI tract Variceal bleeding (Esophageal/gastric) Ulcer bleeding (Gastric/duodenal) CNS bleeding Hemorrhagic stroke (ICH, SAH) Other (SDH, EDH) Hematoma Significant clinical event
VKA (n = 173)
No VKA (n = 148)
p Value
11 (1.82) 58 (9.61) 42 (6.96) 24 (3.98)
23 (4.74) 30 (6.18) 19 (3.91) 10 (2.06)
18 (2.98) 11 (1.82) 6 (0.99) 5 (0.83) 5 (0.83) 62 (10.27)a
Early LC (n = 215)
Advanced LC (n = 106)
VKA (n = 108)
No VKA (n = 107)
p Value
VKA (n = 65)
No VKA (n = 41)
p Value
0.01 0.04 0.03 0.04
6 (1.47) 23 (5.62) 17 (4.15) 9 (2.20)
20 (5.48) 19 (5.20) 11 (3.01) 5 (1.37)
0.003 0.81 0.41 0.41
5 (2.56) 35 (18.03) 25 (12.87) 15 (7.72)
3 (2.50) 11 (9.15) 8 (6.65) 5 (4.16)
0.98 0.01 0.04 0.12
9 (1.85)
0.12
8 (1.95)
6 (1.64)
0.76
10 (5.15)
3 (2.50)
0.14
7 (1.44) 3 (0.62) 4 (0.82) 4 (0.82) 48 (9.89)b
0.32 0.26 0.99 0.99 0.84
4 (0.97) 1 (0.24) 2 (0.48) 2 (0.48) 25 (6.11)
4 (1.10) 0 1 (0.27) 4 (1.09) 34 (9.31)
0.88 0.53 0.69 0.38 0.10
7 (3.61) 5 (2.56) 3 (1.55) 3 (1.55) 37 (19.05)
3 (2.50) 3 (2.50) 3 (2.50) 0 14 (11.65)
0.32 0.98 0.57 0.24 0.11
CNS = Central Nervous System, GI = Gastrointestinal. For definitions of other abbreviations, see Table 1. a Seven patients experienced both stroke and bleeding events. Ischemic stroke preceded in 4 patients and major bleeding in 3 patients. b Five patients experienced both stroke and bleeding events. Ischemic stroke preceded in 3 patients and major bleeding in 2 patients.
6.7%/year, p = 0.04). Multivariate Cox analyses revealed that long-term treatment of VKA in AF patients with early LC significantly reduced the risk of ischemic stroke (adjusted HR 0.23; 95% CI, 0.09–0.58, p = 0.002, Fig. 3A and Table 3) without increasing the risk of major bleeding (adjusted HR 0.89; 95% CI, 0.48–1.66, p = 0.72, Fig. 3B and Table 3). Meanwhile, in advanced LC patients, there was no significant difference in the risk of ischemic stroke between the two groups (adjusted HR 0.60; 95% CI, 0.15–2.42, p = 0.47, Fig. 3C and Table 3). Moreover, VKA prescription significantly increased the risk of major bleeding events in advanced LC patients (adjusted HR 2.98; 95% CI, 1.23–7.19, p = 0.02, Fig. 3D and Table 3). As a result, VKA significantly increased the freedom from SCEs in early stage LC patients (adjusted HR 0.57; 95% CI, 0.34–0.88, p = 0.04, Fig. 4), whereas it significantly increased the risk of SCE in advanced LC patients (adjusted HR 2.01; 95% CI, 1.06–3.83, p = 0.03, Fig. 4). We also evaluated the net clinical benefit (NCB) of VKA treatment using the model that weights ischemia or hemorrhage events by the hazard ratio for death after the event [31]. We compared the NCB according to the risk group of Child-Pugh, CHADS2, CHA2DS2-VASc and HAS-BLED scoring system. In this model, VKA treatment reduced the
3.38 ischemic stroke equivalents per 100 patient years (95% CI, 0.79 to 5.66) in patients with early LC (Fig. 5). However, in patients with advanced LC, VKA treatment significantly increased the ischemic stroke equivalents by 1.15 (95% CI, 0.32 to 1.99) in this weighted model. In addition, interestingly, VKA treatment did not show significant net clinical benefit in high-risk group of stroke (CHADS2 ≥ 2 and CHA2DS2-VASc ≥2). Also, it did not significantly worsen the outcome of patients with high risk of bleeding (HAS-BLED ≥3). 4. Discussion 4.1. Major findings The major findings of this study are as follows. Beneficial effect of VKA in reducing the risk of ischemic stroke was also observed in AF patients with LC. However, VKA significantly increased the risk of major bleeding, especially at the GI tract, which offsets the beneficial effect thus making no significant difference in net clinical outcome. On the subgroup analyses by hepatic functional reserve, however, VKA treatment was beneficial in early stage LC patients, which reduced the risk
Fig. 1. Kaplan–Meier estimates for (A) survival free from ischemic stroke or (B) major bleeding. VKA treatment significantly reduced the risk of ischemic stroke (p = 0.005), whereas increased major bleeding (p = 0.02) in AF patient with LC.
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
189
Table 3 Risk for stroke and major bleeding events associated to VKA treatment in AF patients with LC. Ischemic stroke
All
Event rate per 100 person-years
Early LC p Value
Advanced LC p Value
p Value p Value
VKA No VKA Unadjusted HR⁎ Adjusted HR (CHADS2-adjusted)⁎
1.82 (0.96–3.07) 4.74 (3.28–7.00) 0.32 (0.15–0.66) 0.30 (0.14–0.62)
– – 0.002 0.001
1.47 (0.59–3.05) 5.48 (3.44–8.31) 0.26 (0.11–0.66) 0.23 (0.09–0.58)
– – 0.004 0.002
2.56 (0.94–5.71) 2.50 (0.64–6.79) 0.59 (0.15–2.34) 0.60 (0.15–2.42)
– – 0.46 0.47
0.38 0.17 – –
Bleeding VKA No VKA Unadjusted HR⁎ Adjusted HR (HAS-BLED-adjusted)⁎
9.61 (7.36–12.33) 6.18 (4.25–8.71) 1.68 (1.08–2.61) 1.87 (1.13–3.09)
– – 0.02 0.02
5.62 (3.65–8.29) 5.20 (3.22–7.97) 1.17 (0.64–2.14) 0.89 (0.48–1.66)
– – 0.62 0.72
18.03 (12.82–24.87) 9.15 (4.81–15.91) 1.94 (0.98–3.83) 2.98 (1.23–7.19)
– – 0.05 0.02
b0.001 0.14 – –
Composite outcome of ischemic stroke and major bleeding VKA 10.27 (7.87–13.16) No VKA 9.89 (7.29–13.11) Unadjusted HR⁎ 0.98 (0.67–1.42) Adjusted HR (adjusted for CHADS2 and HAS-BLED)⁎ 0.80 (0.54–1.18)
– – 0.91 0.56
6.11 (3.95–9.01) 9.31 (6.45–13.01) 0.64 (0.38–1.07) 0.57 (0.34–0.88)
– – 0.09 0.04
19.05 (13.41–26.26) 11.65 (6.36–19.54) 1.46 (0.79–2.67) 2.01 (1.06–3.83)
– – 0.08 0.03
b0.001 0.24 – –
Numbers in parenthesis represent 95% confidence interval. CHADS2 = Congestive heart failure, Hypertension, Age ≥75, Diabetes mellitus, and prior Stroke or transient ischemic attack; HAS-BLED = Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly (N65), Drugs/alcohol concomitantly. ⁎ Hazard ratios (HR) were derived from time-dependent Cox regression analyses.
of ischemic stroke without increasing the bleeding risk. On the contrary, in advanced LC group, VKA treatment was harmful because major bleeding risk was markedly increased in these patients, whereas there was no significant difference in long-term risk of ischemic stroke. Finally, CP score was a clinically meaningful predictor for net clinical benefit by VKA treatment in AF patients with LC. 4.2. VKA and ischemic stroke in LC patients VKA effectively reduced the risk of ischemic stroke in early LC patients, however the benefit was not observed in advanced LC patients. We supposed that lower TTR in advanced LC patients with VKA offers a possible explanation for this, because mean TTR lesser than 58% is known to be of little benefit [34]. Moreover, spontaneously anticoagulation promoted by the liver cirrhosis could have affected the result. Actually, in this study, spontaneously prolonged INR was observed in 24% of advanced LC patients without VKA. It has been reported that, on rare occasions, the prevalence of ischemic stroke can be low in LC patients, probably due to increased carotid blood flow by portalsystemic shunts [35]. The overall rate of ischemic stroke in advanced LC patients was 2.5%/year (95% CI 0.63–6.79). In this study, the mean CHA2DS2-VASc score was 2.6 ± 1.8 and 2.2 ± 1.5 in early and advanced LC, respectively. Therefore, our stroke rate is comparable with the
previous report that shows annual stroke risk of 2.2% for CHA2DS2VASc score 2 and 3.2% for CHA2DS2-VASc score 3, respectively [36–38]. The casual relationship between liver cirrhosis and stroke risk in AF patients should be investigated in prospective setting. Moreover, CHADS2 score ≥ 2 was a favorable predictor for net clinical benefit in a largescale AF study cohort (ATRIA) [39], as the risk of ischemic stroke without anticoagulation surpassed the risk of bleeding increased by anticoagulation. However, interestingly, in AF patients with LC, VKA treatment in CHADS2 ≥2 or CHA2DS2-VASc ≥2 showed just the trend of net clinical benefit without statistical significance. It is supposed that increased bleeding risk by varix and coagulopathy in LC patients offsetted the beneficial effect of stroke prevention. 4.3. VKA and bleeding in LC patients Correlation analysis revealed the positive correlation between HASBLED and CP scores. Chronic liver disease is a component of the HASBLED scoring system, which accounts for 1 point out of the total 9 points. However, the LC grade was never considered in most of the bleeding risk stratification systems. Our study shows that AF patients with advanced LC are more prone to bleeding events compared to early stage LC patients. Actually, several components of bleeding risk prediction systems (e.g., previous bleeding history, predisposition, labile INR, or alcohol consumption) are well-known features of LC, and these findings are more definite in advanced stages of the disease [40–43]. Therefore, the severity of LC should be considered in the bleeding risk stratification of AF patients with LC. Actually, when we analyzed according to hepatic functional reserve, VKA did not increase the bleeding risk in early LC group, whereas significant risk increment for the GI bleeding was observed in advanced LC group probably due to higher prevalence varix in advanced LC (CP-A, 19% vs CP-B/C, 32%, p = 0.01). Furthermore, among 11 advanced LC patients who experienced major bleeding events without VKA treatment, 7 (64%) patients had spontaneously prolonged INR more than 2 with an average of 2.4 ± 0.3. This finding implies that INR prolongation with or without VKA is a strong predictor of bleeding and worse outcome in AF patients with LC. 4.4. Study limitations
Fig. 2. Kaplan–Meier estimate for cumulative survival free from significant clinical events. There was no significant difference between the patients with or without VKA (p = 0.91).
Because this study is a retrospective study and not a randomized study, it carries several limitations of such trials, including selection bias. Data abstracted from medical records is limited by the degree of
190
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
Fig. 3. Kaplan–Meier estimates for (A and C) survival free from ischemic stroke or (B and D) major bleeding in early and advanced LC patients. VKA treatment significantly reduced the risk of ischemic stroke, without increasing major bleeding risk in early LC. However, in advanced LC, there was no difference in the stroke free survival, whereas VKA increased risk of bleeding significantly.
documentation. Second, detailed reasons for using or not using VKA in each patient were not clear. Moreover, more than 10% of patients were prescribed anti-platelet agents, which incidence was higher in patients without VKA. As these agents can influence the risk of ischemic
stroke or bleeding, it should be considered in the interpretation of data. Finally, no final statement can be asserted by this retrospective study, suggesting a prospectively designed trial to clearly show the benefit of VKA treatment in early LC patients with AF.
Fig. 4. Kaplan–Meier estimates for cumulative survival free from significant clinical events (SCEs) according to hepatic functional reserve. VKA treatment showed beneficial effect in early LC patients (p = 0.04), whereas it increased SCEs in advanced LC patients (p = 0.05).
Fig. 5. Net clinical benefit of VKA treatment according to risk groups. Values represent ischemic stroke equivalents prevented per 100 patient-years by VKA treatment.
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
5. Conclusions The VKA treatment might be beneficial to reduce significant clinical events in the early LC but not in the advanced LC group. However to confirm this hypothesis, a prospective randomized study is needed.
[17] [18]
Source of funding
[19]
This study was supported in part by research grants of the Basic Science Research Program through the National Research Foundation of Korea, the Ministry of Education, Science, and Technology (NRF-20100021993, NRF-2012R1A2A2A02045367), and the Korean Healthcare Technology R&D Project, Ministry of Health & Welfare (HI12C1552).
[20]
[21]
[22]
Conflict of interest [23]
None. Acknowledgments The authors of this manuscript certify that it complies with the Principles of Ethical Publishing in the International Journal of Cardiology.
[24] [25]
[26]
Appendix A. Supplementary data [27]
Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2014.11.183.
[28]
References
[29]
[1] S. Stewart, C.L. Hart, D.J. Hole, J.J. McMurray, Population prevalence, incidence, and predictors of atrial fibrillation in the Renfrew/Paisley study, Heart 86 (2001) 516–521. [2] A.S. Go, E.M. Hylek, K.A. Phillips, et al., Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study, JAMA 285 (2001) 2370–2375. [3] K.L. Furie, S.E. Kasner, R.J. Adams, et al., Guidelines for the prevention of stroke in patients with stroke or transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association, Stroke 42 (2011) 227–276. [4] R. Pisters, D.A. Lane, R. Nieuwlaat, C.B. de Vos, H.J. Crijns, G.Y. Lip, A novel userfriendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey, Chest 138 (2010) 1093–1100. [5] C.W. Siu, H.F. Tse, Net clinical benefit of warfarin therapy in elderly Chinese patients with atrial fibrillation, Circ. Arrhythm Electrophysiol. 7 (2014) 300–306. [6] S. Vong, B.P. Bell, Chronic liver disease mortality in the United States, 1990–1998, Hepatology 39 (2004) 476–483. [7] C.W. Shepard, L. Finelli, M.J. Alter, Global epidemiology of hepatitis C virus infection, Lancet Infect. Dis. 5 (2005) 558–567. [8] H.B. El-Serag, K.L. Rudolph, Hepatocellular carcinoma: epidemiology and molecular carcinogenesis, Gastroenterology 132 (2007) 2557–2576. [9] J. Tomoda, M. Mizuno, T. Sugihara, T. Itano, T. Tsuji, Gastric mucosal lesion in liver disease: impaired gastric mucosal defence mechanism in rats with induced liver injury and in patients with liver cirrhosis, J. Gastroenterol. Hepatol. 4 (Suppl. 1) (1989) 136–139. [10] J.C. Luo, H.B. Leu, M.C. Hou, et al., Cirrhotic patients at increased risk of peptic ulcer bleeding: a nationwide population-based cohort study, Aliment. Pharmacol. Ther. 36 (2012) 542–550. [11] H. Niizuma, J. Suzuki, T. Yonemitsu, T. Otsuki, Spontaneous intracerebral hemorrhage and liver dysfunction, Stroke 19 (1988) 852–856. [12] Y. Fujii, S. Takeuchi, R. Tanaka, T. Koike, O. Sasaki, T. Minakawa, Liver dysfunction in spontaneous intracerebral hemorrhage, Neurosurgery 35 (1994) 592–596. [13] B.F. Gage, Y. Yan, P.E. Milligan, et al., Clinical classification schemes for predicting hemorrhage: results from the National Registry of Atrial Fibrillation (NRAF), Am. Heart J. 151 (2006) 713–719. [14] A. Rash, T. Downes, R. Portner, W.W. Yeo, N. Morgan, K.S. Channer, A randomised controlled trial of warfarin versus aspirin for stroke prevention in octogenarians with atrial fibrillation (WASPO), Age Ageing 36 (2007) 151–156. [15] P.T. Akins, H.A. Feldman, R.G. Zoble, et al., Secondary stroke prevention with ximelagatran versus warfarin in patients with atrial fibrillation: pooled analysis of SPORTIF III and V clinical trials, Stroke 38 (2007) 874–880. [16] A.S. Go, E.M. Hylek, L.H. Borowsky, K.A. Phillips, J.V. Selby, D.E. Singer, Warfarin use among ambulatory patients with nonvalvular atrial fibrillation: the anticoagulation
[30] [31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40] [41]
[42] [43]
191
and risk factors in atrial fibrillation (ATRIA) study, Ann. Intern. Med. 131 (1999) 927–934. S.J. Connolly, M.D. Ezekowitz, S. Yusuf, et al., Dabigatran versus warfarin in patients with atrial fibrillation, N. Engl. J. Med. 361 (2009) 1139–1151. P. Petersen, G. Boysen, J. Godtfredsen, E.D. Andersen, B. Andersen, Placebo-controlled, randomised trial of warfarin and aspirin for prevention of thromboembolic complications in chronic atrial fibrillation. The Copenhagen AFASAK study, Lancet 1 (1989) 175–179. H.G. Site, Clopidogrel plus aspirin versus oral anticoagulation for atrial fibrillation in the Atrial fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events (ACTIVE W): a randomised controlled trial, Lancet 367 (2006) 1903–1912. H. Decousus, V.F. Tapson, J.F. Bergmann, et al., Factors at admission associated with bleeding risk in medical patients: findings from the IMPROVE investigators, Chest 139 (2011) 69–79. M.G. Delgado, S. Seijo, I. Yepes, et al., Efficacy and safety of anticoagulation on patients with cirrhosis and portal vein thrombosis, Clin. Gastroenterol. Hepatol. 10 (2012) 776–783. K.I. Rodriguez-Castro, P. Simioni, P. Burra, M. Senzolo, Anticoagulation for the treatment of thrombotic complications in patients with cirrhosis, Liver Int. 32 (2012) 1465–1476. S.D. Berkowitz, D.C. Sane, K.N. Sigmon, et al., Occurrence and clinical significance of thrombocytopenia in a population undergoing high-risk percutaneous coronary revascularization. Evaluation of c7E3 for the Prevention of Ischemic Complications (EPIC) Study Group, J. Am. Coll. Cardiol. 32 (1998) 311–319. M.H. Tsai, Y.S. Peng, Y.C. Chen, et al., Adrenal insufficiency in patients with cirrhosis, severe sepsis and septic shock, Hepatology 43 (2006) 673–681. U. Heemann, U. Treichel, J. Loock, et al., Albumin dialysis in cirrhosis with superimposed acute liver injury: a prospective, controlled study, Hepatology 36 (2002) 949–958. L. Sandrin, B. Fourquet, J.M. Hasquenoph, et al., Transient elastography: a new noninvasive method for assessment of hepatic fibrosis, Ultrasound Med. Biol. 29 (2003) 1705–1713. J. Foucher, E. Chanteloup, J. Vergniol, et al., Diagnosis of cirrhosis by transient elastography (FibroScan): a prospective study, Gut 55 (2006) 403–408. G.Y. Lip, S.J. Edwards, Stroke prevention with aspirin, warfarin and ximelagatran in patients with non-valvular atrial fibrillation: a systematic review and meta-analysis, Thromb. Res. 118 (2006) 321–333. H.D. White, M. Gruber, J. Feyzi, et al., Comparison of outcomes among patients randomized to warfarin therapy according to anticoagulant control: results from SPORTIF III and V, Arch. Intern. Med. 167 (2007) 239–245. D.B. Matchar, A. Jacobson, R. Dolor, et al., Effect of home testing of international normalized ratio on clinical events, N. Engl. J. Med. 363 (2010) 1608–1620. S.J. Connolly, J.W. Eikelboom, J. Ng, et al., Net clinical benefit of adding clopidogrel to aspirin therapy in patients with atrial fibrillation for whom vitamin K antagonists are unsuitable, Ann. Intern. Med. 155 (2011) 579–586. F.R. Rosendaal, S.C. Cannegieter, F.J. van der Meer, E. Briet, A method to determine the optimal intensity of oral anticoagulant therapy, Thromb. Haemost. 69 (1993) 236–239. E.M. Hylek, A.S. Go, Y. Chang, et al., Effect of intensity of oral anticoagulation on stroke severity and mortality in atrial fibrillation, N. Engl. J. Med. 349 (2003) 1019–1026. S.J. Connolly, J. Pogue, J. Eikelboom, et al., Benefit of oral anticoagulant over antiplatelet therapy in atrial fibrillation depends on the quality of international normalized ratio control achieved by centers and countries as measured by time in therapeutic range, Circulation 118 (2008) 2029–2037. A. Berzigotti, A. Bonfiglioli, A. Muscari, et al., Reduced prevalence of ischemic events and abnormal supraortic flow patterns in patients with liver cirrhosis, Liver Int. 25 (2005) 331–336. J.B. Olesen, G.Y. Lip, M.L. Hansen, et al., Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study, BMJ 342 (2011) d124. G.Y. Lip, R. Nieuwlaat, R. Pisters, D.A. Lane, H.J. Crijns, Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation, Chest 137 (2010) 263–272. European Heart Rhythm A, European Association for Cardio-Thoracic S, A.J. Camm, et al., Guidelines for the management of atrial fibrillation: the task force for the management of atrial fibrillation of the European Society of Cardiology (ESC), Eur. Heart J. 31 (2010) 2369–2429. L. Friberg, M. Rosenqvist, G.Y. Lip, Net clinical benefit of warfarin in patients with atrial fibrillation: a report from the Swedish atrial fibrillation cohort study, Circulation 125 (2012) 2298–2307. P. Gentilini, G. Laffi, G. La Villa, et al., Long course and prognostic factors of virusinduced cirrhosis of the liver, Am. J. Gastroenterol. 92 (1997) 66–72. P. Jepsen, P. Ott, P.K. Andersen, H.T. Sorensen, H. Vilstrup, Clinical course of alcoholic liver cirrhosis: a Danish population-based cohort study, Hepatology 51 (2010) 1675–1682. C.S. Lieber, Medical disorders of alcoholism, N. Engl. J. Med. 333 (1995) 1058–1065. W. Blonski, T. Siropaides, K.R. Reddy, Coagulopathy in liver disease, Curr. Treat. Options Gastroenterol. 10 (2007) 464–473.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
The safety and efficacy of vitamin K antagonist in patients with atrial fibrillation and liver cirrhosis Seung-Jun Lee, Jae-Sun Uhm, Jong-Youn Kim, Hui-Nam Pak, Moon-Hyoung Lee, Boyoung Joung ⁎ Cardiology Division, Department of Internal Medicine, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea
a r t i c l e
i n f o
Article history: Received 6 June 2014 Received in revised form 14 October 2014 Accepted 23 November 2014 Available online 27 November 2014 Keywords: Atrial fibrillation Stroke Hemorrhage Oral anticoagulant Vitamin K antagonist Liver cirrhosis
a b s t r a c t Background: Bleeding is a major concern in treatment of atrial fibrillation (AF) with vitamin K antagonist (VKA). This concern is more emphasized in patients with high bleeding risk such as liver cirrhosis (LC). Methods and results: We retrospectively analyzed incidence of stroke and major bleeding in 321 AF patients with LC, including early [Child-Pugh (CP)-A, n = 215] and advanced [CP-B or C, n = 106] LC according to VKA prescription. The CHA2DS2-VASc, HAS-BLED and MELD scores were higher in patients with VKA. CP score was positively correlated with HAS-BLED score (rho: 0.602). The incidence of major bleeding was higher in advanced LC than in early LC (14.5%/year vs. 4.9%/year, p b 0.001). VKA reduced the risk of ischemic stroke in AF patients with LC, whereas it significantly increased the major bleeding risk. There was no difference in survival free from significant clinical events (SCEs) between the patients with or without VKA (p = 0.91). On subgroup analysis, VKA was beneficial in early LC patients, as it decreased stroke without increasing major bleeding risk. However, in advanced LC patients, VKA significantly increased the risk of major bleeding, especially variceal origin, that overwhelms stroke reduction. As a result, VKA treatment reduced the risk of SCEs in early LC patients, whereas it increased the risk of SCEs in advanced LC. Conclusions: VKA treatment might be beneficial to reduce significant clinical events in the early LC but not in the advanced LC group. However to confirm this hypothesis, a prospective randomized study is needed. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, occurring in 1–2% of the general population [1,2]. AF confers a five-fold risk of stroke, and one in five of all strokes is attributed to this arrhythmia. Multiple clinical trials have demonstrated the superior therapeutic effect of vitamin K antagonist compared with placebo in the prevention of thromboembolic events among patients with nonvalvular AF [3]. However, some AF patients have the concomitant risk of fatal bleeding, which causes clinicians to be reluctant to use vitamin K antagonist (VKA) in spite of the high stroke risk [4,5]. Chronic liver disease is mainly caused by excess alcohol intake and viral hepatitis, which has been classified as the tenth most frequent cause of death in the United States [6]. Moreover, the prevalence of liver cirrhosis (LC) is quite higher in eastern Asia or Africa, which is a major cause of morbidity and mortality in these areas [7,8]. In particular, LC patients usually suffer from conditions that cause upper
Abbreviations: AF, atrial fibrillation; CP, Child-Pugh; LC, liver cirrhosis; SCE, significant clinical event; VKA, vitamin K antagonist ⁎ Corresponding author at: Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752 Republic of Korea. E-mail address: [email protected] (B. Joung).
http://dx.doi.org/10.1016/j.ijcard.2014.11.183 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
gastrointestinal bleeding, including esophageal varices, gastric varices, and peptic ulcer disease [9,10]. In addition, the incidence of spontaneous intracerebral hemorrhage is significantly higher in LC patients with bleeding diathesis due to thrombocytopenia, splenomegaly, or prolonged international normalized ratio (INR) [11,12]. The bleeding risk schemas, such as HAS-BLED and HEMORR2HAGES, have shown that abnormal liver function or chronic liver disease is a major bleeding risk in AF patients [13]. However, sparse data exist regarding the benefit of oral anticoagulation therapy use in AF patients with LC, as most trials have excluded patients with chronic liver disease or LC [14–17]. Furthermore, anemia, thrombocytopenia and bleeding diathesis are the main clinical features of LC patients which were excluded from most of sophisticatedly designed AF trials [18,19]. Hence, little is known about the importance of VKA treatment in AF patients with LC. Meanwhile, some researches have shown that LC patients with portal vein thrombosis (PVT) are very vulnerable to life threatening bleeding complications [20,21], although VKA had clinical benefit in re-canalizing portal vein thrombosis and preventing ischemic complications [21,22]. We hypothesized that the VKA treatment in AF patients with LC would increase major bleeding and decrease thromboembolic events. Moreover, the increased bleeding risk might be outweighed by the decreased thromboembolic events in patients who were maintained in optimal INR level. This study evaluated the clinical courses, including
186
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
ischemic stroke and major bleeding events, of AF patients with early or advanced LC status according to the use of VKA. 2. Methods 2.1. Study subjects The study protocol was approved by the Institutional Review Board of Severance Cardiovascular Hospital, Seoul, Korea and complied with the Declaration of Helsinki. Between January 2001 and January 2012, using ICD-9 codes, we identified 443 consecutive patients with non-valvular AF and LC over the age of 18. We defined non-valvular AF by the discharge diagnosis of AF (ICD-9 code 427.31), no previous diagnosis of mitral stenosis or prosthetic heart valve (ICD-9 codes 394.0, 394.2, 396.0, 396.1, 396.8, V43.3, or V42.4), and no aortic or mitral valve surgery (ICD-9 codes 35.10 to 35.14 or 35.20 to 35.28). Patients with severe thrombocytopenia (platelets b50,000/mm3, n = 32) [23], insufficient follow up duration (b6 months, n = 27), evidence of renal failure (n = 26), previous intracerebral hemorrhage (n = 15), liver transplantation (n = 5), and multiple missing data (n = 17) were excluded. Finally, 321 patients were included in this study. All of VKA prescribed were warfarin (Coumadin®). LC was defined by clinical diagnosis, biochemistry, Doppler ultrasound, computed tomography scan, and biopsy, when performed [24,25]. The diagnosis was made histologically in 58 patients in whom liver biopsy was performed to confirm the status of liver cirrhosis. It was clinically diagnosed in 263 patients. The clinical diagnosis of liver cirrhosis was made based on the presence of ascites or esophageal varices without other explainable causes, along with unequivocal image on ultrasound or computed tomography. Among 263 patients diagnosed without biopsy, 176 (64%) patients had been performed transient elastography in order to evaluate the degree of hepatic fibrosis non-invasively [26]. The median liver stiffness measurement value for these patients was 18.3 (7.6– 34.3) which is highly suggestive of liver cirrhosis [27]. AF was documented by 12-lead electrocardiography or 24-hour Holter recording. Patients' medical records were reviewed for information on patient age, sex, weight and height, drug therapy, and AF duration. Patient databases were searched to identify known or putative risk factors for ischemic stroke, heart failure, hypertension, diabetes mellitus, hyperlipidemia, tobacco, or alcohol use [28]. Heart failure was defined when hospitalized patients exhibited appropriate symptoms (shortness of breath, fatigue, fluid retention, or any combination of these symptoms) and clinical signs of fluid retention (pulmonary or peripheral) with explainable abnormality of cardiac structure and function. The CHADS2 [Congestive heart failure, Hypertension, Age, Diabetes mellitus, prior Stroke or transient ischemic attack (doubled)] risk index, CHA2DS2-VASc [Congestive heart failure, hypertension, Age ≥75 (doubled), Diabetes mellitus, prior Stroke or transient ischemic attack (doubled)-Vascular disease, Age 65–74 years and Sex category (female)] risk index, and the HAS-BLED [Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly (N65), Drugs/alcohol concomitantly] scores were evaluated. We divided patients into two groups according to their liver cirrhosis status. The Child-Pugh score was used to assess the prognosis of cirrhosis. The score employs five clinical measurements of liver disease; total bilirubin (b2, 2–3, N3 mg/dl), serum albumin (N3.5, 2.8–3.5, b2.8 g/dl), PT INR (b1.7, 1.7–2.3, N2.3), ascites (none, mild, moderate to severe) and hepatic encephalopathy (none, G I–II, G III–IV). Each measure is scored from 1–3, following the order in parentheses above. In this study, the patients with cirrhosis were classified as early LC (score 5–6, n = 215) and advanced LC (score ≥7, n = 106). To quantify end-stage liver disease, the Model For End-Stage Liver Disease (MELD) score (0.957 * ln(serum Cr) +0.378 * ln(serum Bilirubin) +1.120 * ln(INR) +0.643) * 10 was calculated. 2.2. Clinical follow-up Patients visited specialized cardiologists and hepatologists with mean interval of 2.3 ± 0.7 months to check clinical symptoms, INR, biochemistry results and imaging studies. We evaluated ischemic stroke and major bleeding events during follow-up by review of the charts. Ischemic stroke was defined as a neurological deficit of sudden onset that persisted for N24 h corresponding to a vascular territory in the absence of primary hemorrhage and was not explained by other causes, including trauma, infection, or vasculitis [29] . Major bleeding was defined as any central nervous system (CNS) bleeding which includes intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), subdural hemorrhage (SDH), and epidural hemorrhage (EDH); any bleeding requiring transfusion of at least two units of red blood cells or equivalent of whole blood in 24 h; or bleeding events that caused hypotension (systolic blood pressure b90 mm Hg), multi-organ failure, or death. We defined a significant clinical event (SCE) as a first major event that occurred during follow-up period including stroke, major bleeding episode or death [30]. When a patient experienced both ischemic stroke and major bleeding events during the follow-up period, each event was counted. However, when we analyzed the Kaplan–Meier cumulative SCE free survival, we counted the first event only. 2.3. Net clinical benefit assessment We assessed the net clinical benefit (NCB) by calculating the difference of annualized incidence rate (IR) of ischemic stroke and major bleeding multiplied by weighting factors. We adopted the weighting factor derived from the large-scale clinical trial (ACTIVE trial)
by Connolly et al. [31] that measured the adjusted hazard ratio (HR) for death after the event standardized to the adjusted HR of ischemic stroke (IS, weight 1.0). The relative weights for each clinical event are 3.08 for hemorrhagic stroke (HS) which includes ICH and SAH, 0.60 for other CNS bleeding (Other_CNS) that includes SDH and EDH, and 0.67 for extracranial bleeding, respectively. As a result, we calculated the NCB according to the following equation: NCB ¼ ½IRIS
no VKA
þ 3:08 IRHS
IRExtracranial
no VKA
– ½IRIS
þ 0:67 IRExtracranial
no VKA
VKA
þ 0:60 IROther
þ 3:08 IRHS
VKA
CNS no VKA
þ 0:67
þ 0:60 IROther
CNS VKA
VKA :
The resulting values were regarded as ischemic stroke equivalents prevented by VKA per 100 patient-years. 2.4. Intensity and quality of anticoagulation In patients with concomitant indications for VKA treatment, the decision was made by the physician's clinical evaluation of the risk for thrombotic and hemorrhagic events. Among 321 patients enrolled in this study, 173 (54%) were treated with VKA. Indications for VKA therapy retrieved from the chart are as follows: atrial fibrillation, n = 144 (83%); previous portal vein thrombosis, n = 10 (6%); previous deep vein thrombosis, n = 6 (3%); and not documented, n = 13 (8%). The intensity of anticoagulation was determined by the INR values. The INR values at each outpatient clinic/emergency department visit and during hospital admission were retrieved from the medical record. Data on the first four weeks after initiation of warfarin therapy were excluded from the analysis. All INR testing was performed in the same laboratory using a fully automated hematology analyzer (ACL TOP 700 CTS, Instrumentation Laboratory, Werfen Group, Barcelona, Spain) with the same reagent (Hemosil Recombiplastin, Instrumentation Laboratory). International sensitivity index (ISI) values assigned specifically for our laboratory's reagent/instrument combination were used for calculation of INR. Local ISI verification was performed more than twice per year according to the Clinical and Laboratory Standards Institute H54-A guideline. Mean INR values and time in therapeutic range (TTR) of INR 2.0–3.0 were calculated using the linear interpolation methods proposed by Rosendaal et al. [32]. This method assumes that the INR values between two consecutive measurements vary linearly. Patients with an interval of 56 days or more between INR tests were excluded from the TTR analysis as they increase error for interpolation analysis [33]. Nine patients of early LC with VKA (8%) and 5 patients of advanced LC with VKA (8%) were excluded from TTR analysis. 2.5. Statistical analysis Continuous variables that were normally distributed were reported as mean ± standard deviation (SD) and were compared by use of a Student's t-test for parametric data and Mann–Whitney test for nonparametric data. Categorical variables were reported as count (percentage) and were compared using Chi-square or Fisher's exact test. Event rates were compared using mid-p exact test. Multivariate Cox proportional-hazard models were used to identify factors associated with ischemic stroke or major bleeding. Missing binary data, of which proportion are lesser than 2% for each covariate, were coded as “not present” and missing categorical data were coded as “missing”. Kaplan–Meier survival curves were plotted according to use of VKA and compared by means of the log-rank test. Pearson's correlation analysis was employed to assess the relationship between the CP score and the CHADS2, CHA2DS2-VASc, or HAS-BLED scoring system. The SPSS statistical package (SPSS Inc., Chicago, Illinois) version 18.0 was used to perform all statistical evaluations. A p value b 0.05 was considered statistically significant.
3. Results 3.1. Characteristics of patients Clinical characteristics of patients with or without VKA are presented in Table 1. There was no difference in the mean duration of follow-up for patients with or without VKA (3.5 ± 2.3 vs. 3.2 ± 2.4 years, p = 0.29). The mean CP scores of the patients with and without VKA were 7.5 ± 2.3 and 7.0 ± 1.9, respectively (p = 0.06). The MELD score was significantly higher in patients with VKA (14.0 ± 10.7 vs. 8.4 ± 7.3, p b 0.001). Patients without VKA had more frequent seropositivity for HBsAg (49% vs. 28%, p = 0.001). Patients on VKA had more history of ischemic stroke or transient ischemic attack (25% vs. 16%, p = 0.04). The risk of ischemic event calculated by CHADS2 and CHA2DS2-VASc was significantly higher in patients with VKA. In addition, the bleeding risk was also higher in patients with VKA (HAS-BLED score 4.1 ± 1.4 vs. 3.8 ± 1.3, p = 0.02). The characteristics of the patients in each group according to the status of LC are presented in Supplementary Table 1. Mean TTR of the VKA treated group calculated by Rosendaal's method was 59.7 ± 10.0%. Especially, early LC group showed significantly higher TTR compared to advanced LC patients (64.6 ± 10.3% vs. 51.6 ± 9.6%,
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191 Table 1 Characteristics of patients. Characteristic
VKA (n = 173)
No VKA (n = 148)
p Value
Age (years) Female, n (%) Atrial fibrillation type Permanent Persistent Paroxysmal Hypertension Diabetes mellitus Heart failure History of stroke/transient ischemic attack Peptic ulcer (active, healing, scarring) Varix (gastric or esophageal) Thrombocytopenia, moderatea Child-Pugh score MELD score CHADS2 CHA2DS2-VASc HAS-BLED Etiology B-viral C-viral Alcoholic Others Time in therapeutic range (%) Concomitant medications Anti-platelet agent Non-steroidal anti-inflammatory drug Proton pump inhibitor H2 blocker
62.1 ± 10.3 53 (31%)
62.5 ± 11.3 48 (32%)
0.77 0.73
112 (65%) 29 (17%) 31 (18%) 82 (47%) 67 (39%) 56 (32%) 44 (25%) 40 (23%) 39 (23%) 46 (27%) 7.5 ± 2.3 14.0 ± 10.7 2.1 ± 1.5 2.6 ± 1.8 4.1 ± 1.4
95 (64%) 24 (16%) 29 (20%) 62 (42%) 50 (34%) 37 (25%) 24 (16%) 29 (20%) 36 (24%) 46 (31%) 7.0 ± 1.9 8.4 ± 7.3 1.7 ± 1.3 2.2 ± 1.5 3.8 ± 1.3
0.92 0.89 0.70 0.36 0.36 0.15 0.04 0.45 0.71 0.38 0.06 b0.001 0.02 0.03 0.02
48 (28%) 29 (17%) 58 (34%) 33 (19%) 59.7 ± 10.0
72 (49%) 20 (14%) 42 (28%) 21 (14%) –
0.001 0.42 0.43 0.30 –
24 (14%) 29 (17%) 96 (55%) 92 (53%)
33 (22%) 22 (15%) 86 (58%) 93 (63%)
0.05 0.72 0.75 0.10
187
unadjusted HR (relative risk 3.01; 95% CI, 1.97–4.61, p b 0.001) and adjusted HR (relative risk 2.20; 95% CI, 1.39–3.46, p = 0.001) of major bleeding. 3.4. Screening for gastrointestinal lesion and bleeding prophylaxis There was no difference in the presence of peptic ulcer or varix between the patients with or without VKA (Table 1). Among 173 patients who were treated with VKA, 39 patients had gastric or esophageal varices. Among them, 11 patients (27%) with previous variceal bleeding, grade II esophageal varices with red signs or grade III varices were treated by endoscopic variceal ligation (EVL) before VKA prescription. All patients who underwent EVL, started VKA at least 1 month after the eradication of varices. In addition, among 173 patients on VKA, 40 (23%) individuals had peptic ulcer disease and 13 had active ulcer lesion. All patients with active ulcer (n = 13) started VKA after treatment of ulcer and 7 of them (54%) were performed follow-up esophagogastroduodenoscopy for confirmation of ulcer healing. 3.5. Reduction of ischemic stroke by VKA in LC patients
Mean values ± standard deviation or number and percentage of patients. CHADS2 = Congestive heart failure, Hypertension, Age ≥75, Diabetes mellitus, and prior Stroke or transient ischemic attack; CHA2DS2-VASc = Congestive heart failure, Hypertension, Age ≥75 (doubled), Diabetes mellitus, prior Stroke or transient ischemic attack (doubled)-Vascular disease, Age 65–74 years, and Sex category (female); HAS-BLED = Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly (N65), Drugs/alcohol concomitantly. a Platelet count: 50,000–150,000 cells/mm3.
p b 0.001). Patients without VKA showed higher prescription rate of anti-platelet agents. The indications for anti-platelet treatment in patients without VKA included stroke prevention (n = 17), ischemic heart disease (n = 12) and history of thrombosis (n = 4). However, in patients with VKA, the most common cause of anti-platelet therapy was ischemic heart disease (n = 21).
Table 2 shows clinical events during the follow-up period. The incidence of ischemic stroke was significantly lower in patients with VKA compared to those without (1.8%/year vs. 4.7%/year, p = 0.01). Fig. 1A shows the Kaplan–Meier cumulative survival free from ischemic stroke in AF patients with LC according to the use of VKA. Multivariate Cox analysis revealed that the use of VKA reduced the risk of ischemic stroke in AF patients with LC; unadjusted HR (relative risk 0.32; 95% CI, 0.15– 0.66, p = 0.002), and adjusted HR (relative risk 0.30; 95% CI, 0.14– 0.62, p = 0.001, Table 3). 3.6. VKA treatment increased the risk of major bleeding The incidence of major bleeding was significantly higher in patients with VKA (9.6%/year vs. 6.2%/year, p = 0.04). Especially, GI tract bleeding risk was significantly increased by VKA (Table 2, p = 0.03). The incidence of ulcer bleeding was not increased by VKA, whereas variceal bleeding risk was significantly higher in patients with VKA (p = 0.04). Cumulative survival free from the major bleeding is depicted in Fig. 1B (log-rank p = 0.02). VKA prescription significantly increased the risk of major bleeding; unadjusted HR (relative risk 1.68; 95% CI, 1.08– 2.61, p = 0.02), and adjusted HR (relative risk 1.87; 95% CI, 1.13–3.09, p = 0.02). 3.7. Bleeding risk canceled out the beneficial effect of VKA in LC patients
3.2. The correlation between CP score and stroke or bleeding risk estimation scoring systems We investigated the correlation between CP score and stroke or bleeding risk estimation scoring systems. The CHADS2 (Spearman's rho: 0.088; p = 0.12) and CHA2DS2-VASc (rho: 0.086; p = 0.13) scores were not significantly correlated with the CP score. However, a statistically significant correlation was found between the HAS-BLED score and the CP score (rho: 0.602; p b 0.001). 3.3. High risk of major bleeding in advanced LC patients The incidence of major bleeding was significantly higher in advanced LC patients than in early ones (14.5%/year vs. 4.9%/year, p b 0.001). Among the various types of major bleeding, gastrointestinal (p b 0.001) and CNS bleeding (p = 0.05) were more frequently observed in advanced LC group compared to early LC. Cumulative survival free from major bleeding according to hepatic functional reserve is depicted in Supplementary Figure 1. Early LC patients showed higher survival free from major bleeding compared to advanced LC patients (p b 0.001). A CP score higher than 6 (Grade B or C) increased the
We compared the composite outcome of ischemic stroke and major bleeding according to VKA treatment. There was no significant difference in the incidence of SCE between the two groups (p = 0.84; Table 2). Kaplan–Meier cumulative survival free from SCE was not changed by the VKA treatment (log-rank p = 0.91; Fig. 2). 3.8. Subgroup analysis and net clinical benefit There was no significant difference in the risk of SCE between the patients with or without VKA. We compared the clinical outcome according to hepatic functional reserve in addition to VKA prescription. Patients with CP-score lesser than 7 (CP-A) was regarded as early LC, whereas patients with signs of decompensation (CP grade B and C) was defined as advanced LC. There was significant difference in the incidence of ischemic stroke between the patients with or without VKA only in early stage LC patients (1.5%/year vs. 5.5%/year, p = 0.003, Table 2). Moreover, the incidence of major bleeding was significantly higher only in advanced LC patients (18.0%/year vs. 9.2%/year, p = 0.01, Table 2). Especially, the incidence of GI tract bleeding was significantly higher in advanced LC patients with VKA (12.3%/year vs.
188
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
Table 2 Comparison of clinical events during follow-up period in early or advanced LC patients according to VKA treatment. The numbers in parentheses represent event rates for 100 personyears.
Ischemic stroke Major bleeding GI tract Variceal bleeding (Esophageal/gastric) Ulcer bleeding (Gastric/duodenal) CNS bleeding Hemorrhagic stroke (ICH, SAH) Other (SDH, EDH) Hematoma Significant clinical event
VKA (n = 173)
No VKA (n = 148)
p Value
11 (1.82) 58 (9.61) 42 (6.96) 24 (3.98)
23 (4.74) 30 (6.18) 19 (3.91) 10 (2.06)
18 (2.98) 11 (1.82) 6 (0.99) 5 (0.83) 5 (0.83) 62 (10.27)a
Early LC (n = 215)
Advanced LC (n = 106)
VKA (n = 108)
No VKA (n = 107)
p Value
VKA (n = 65)
No VKA (n = 41)
p Value
0.01 0.04 0.03 0.04
6 (1.47) 23 (5.62) 17 (4.15) 9 (2.20)
20 (5.48) 19 (5.20) 11 (3.01) 5 (1.37)
0.003 0.81 0.41 0.41
5 (2.56) 35 (18.03) 25 (12.87) 15 (7.72)
3 (2.50) 11 (9.15) 8 (6.65) 5 (4.16)
0.98 0.01 0.04 0.12
9 (1.85)
0.12
8 (1.95)
6 (1.64)
0.76
10 (5.15)
3 (2.50)
0.14
7 (1.44) 3 (0.62) 4 (0.82) 4 (0.82) 48 (9.89)b
0.32 0.26 0.99 0.99 0.84
4 (0.97) 1 (0.24) 2 (0.48) 2 (0.48) 25 (6.11)
4 (1.10) 0 1 (0.27) 4 (1.09) 34 (9.31)
0.88 0.53 0.69 0.38 0.10
7 (3.61) 5 (2.56) 3 (1.55) 3 (1.55) 37 (19.05)
3 (2.50) 3 (2.50) 3 (2.50) 0 14 (11.65)
0.32 0.98 0.57 0.24 0.11
CNS = Central Nervous System, GI = Gastrointestinal. For definitions of other abbreviations, see Table 1. a Seven patients experienced both stroke and bleeding events. Ischemic stroke preceded in 4 patients and major bleeding in 3 patients. b Five patients experienced both stroke and bleeding events. Ischemic stroke preceded in 3 patients and major bleeding in 2 patients.
6.7%/year, p = 0.04). Multivariate Cox analyses revealed that long-term treatment of VKA in AF patients with early LC significantly reduced the risk of ischemic stroke (adjusted HR 0.23; 95% CI, 0.09–0.58, p = 0.002, Fig. 3A and Table 3) without increasing the risk of major bleeding (adjusted HR 0.89; 95% CI, 0.48–1.66, p = 0.72, Fig. 3B and Table 3). Meanwhile, in advanced LC patients, there was no significant difference in the risk of ischemic stroke between the two groups (adjusted HR 0.60; 95% CI, 0.15–2.42, p = 0.47, Fig. 3C and Table 3). Moreover, VKA prescription significantly increased the risk of major bleeding events in advanced LC patients (adjusted HR 2.98; 95% CI, 1.23–7.19, p = 0.02, Fig. 3D and Table 3). As a result, VKA significantly increased the freedom from SCEs in early stage LC patients (adjusted HR 0.57; 95% CI, 0.34–0.88, p = 0.04, Fig. 4), whereas it significantly increased the risk of SCE in advanced LC patients (adjusted HR 2.01; 95% CI, 1.06–3.83, p = 0.03, Fig. 4). We also evaluated the net clinical benefit (NCB) of VKA treatment using the model that weights ischemia or hemorrhage events by the hazard ratio for death after the event [31]. We compared the NCB according to the risk group of Child-Pugh, CHADS2, CHA2DS2-VASc and HAS-BLED scoring system. In this model, VKA treatment reduced the
3.38 ischemic stroke equivalents per 100 patient years (95% CI, 0.79 to 5.66) in patients with early LC (Fig. 5). However, in patients with advanced LC, VKA treatment significantly increased the ischemic stroke equivalents by 1.15 (95% CI, 0.32 to 1.99) in this weighted model. In addition, interestingly, VKA treatment did not show significant net clinical benefit in high-risk group of stroke (CHADS2 ≥ 2 and CHA2DS2-VASc ≥2). Also, it did not significantly worsen the outcome of patients with high risk of bleeding (HAS-BLED ≥3). 4. Discussion 4.1. Major findings The major findings of this study are as follows. Beneficial effect of VKA in reducing the risk of ischemic stroke was also observed in AF patients with LC. However, VKA significantly increased the risk of major bleeding, especially at the GI tract, which offsets the beneficial effect thus making no significant difference in net clinical outcome. On the subgroup analyses by hepatic functional reserve, however, VKA treatment was beneficial in early stage LC patients, which reduced the risk
Fig. 1. Kaplan–Meier estimates for (A) survival free from ischemic stroke or (B) major bleeding. VKA treatment significantly reduced the risk of ischemic stroke (p = 0.005), whereas increased major bleeding (p = 0.02) in AF patient with LC.
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
189
Table 3 Risk for stroke and major bleeding events associated to VKA treatment in AF patients with LC. Ischemic stroke
All
Event rate per 100 person-years
Early LC p Value
Advanced LC p Value
p Value p Value
VKA No VKA Unadjusted HR⁎ Adjusted HR (CHADS2-adjusted)⁎
1.82 (0.96–3.07) 4.74 (3.28–7.00) 0.32 (0.15–0.66) 0.30 (0.14–0.62)
– – 0.002 0.001
1.47 (0.59–3.05) 5.48 (3.44–8.31) 0.26 (0.11–0.66) 0.23 (0.09–0.58)
– – 0.004 0.002
2.56 (0.94–5.71) 2.50 (0.64–6.79) 0.59 (0.15–2.34) 0.60 (0.15–2.42)
– – 0.46 0.47
0.38 0.17 – –
Bleeding VKA No VKA Unadjusted HR⁎ Adjusted HR (HAS-BLED-adjusted)⁎
9.61 (7.36–12.33) 6.18 (4.25–8.71) 1.68 (1.08–2.61) 1.87 (1.13–3.09)
– – 0.02 0.02
5.62 (3.65–8.29) 5.20 (3.22–7.97) 1.17 (0.64–2.14) 0.89 (0.48–1.66)
– – 0.62 0.72
18.03 (12.82–24.87) 9.15 (4.81–15.91) 1.94 (0.98–3.83) 2.98 (1.23–7.19)
– – 0.05 0.02
b0.001 0.14 – –
Composite outcome of ischemic stroke and major bleeding VKA 10.27 (7.87–13.16) No VKA 9.89 (7.29–13.11) Unadjusted HR⁎ 0.98 (0.67–1.42) Adjusted HR (adjusted for CHADS2 and HAS-BLED)⁎ 0.80 (0.54–1.18)
– – 0.91 0.56
6.11 (3.95–9.01) 9.31 (6.45–13.01) 0.64 (0.38–1.07) 0.57 (0.34–0.88)
– – 0.09 0.04
19.05 (13.41–26.26) 11.65 (6.36–19.54) 1.46 (0.79–2.67) 2.01 (1.06–3.83)
– – 0.08 0.03
b0.001 0.24 – –
Numbers in parenthesis represent 95% confidence interval. CHADS2 = Congestive heart failure, Hypertension, Age ≥75, Diabetes mellitus, and prior Stroke or transient ischemic attack; HAS-BLED = Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly (N65), Drugs/alcohol concomitantly. ⁎ Hazard ratios (HR) were derived from time-dependent Cox regression analyses.
of ischemic stroke without increasing the bleeding risk. On the contrary, in advanced LC group, VKA treatment was harmful because major bleeding risk was markedly increased in these patients, whereas there was no significant difference in long-term risk of ischemic stroke. Finally, CP score was a clinically meaningful predictor for net clinical benefit by VKA treatment in AF patients with LC. 4.2. VKA and ischemic stroke in LC patients VKA effectively reduced the risk of ischemic stroke in early LC patients, however the benefit was not observed in advanced LC patients. We supposed that lower TTR in advanced LC patients with VKA offers a possible explanation for this, because mean TTR lesser than 58% is known to be of little benefit [34]. Moreover, spontaneously anticoagulation promoted by the liver cirrhosis could have affected the result. Actually, in this study, spontaneously prolonged INR was observed in 24% of advanced LC patients without VKA. It has been reported that, on rare occasions, the prevalence of ischemic stroke can be low in LC patients, probably due to increased carotid blood flow by portalsystemic shunts [35]. The overall rate of ischemic stroke in advanced LC patients was 2.5%/year (95% CI 0.63–6.79). In this study, the mean CHA2DS2-VASc score was 2.6 ± 1.8 and 2.2 ± 1.5 in early and advanced LC, respectively. Therefore, our stroke rate is comparable with the
previous report that shows annual stroke risk of 2.2% for CHA2DS2VASc score 2 and 3.2% for CHA2DS2-VASc score 3, respectively [36–38]. The casual relationship between liver cirrhosis and stroke risk in AF patients should be investigated in prospective setting. Moreover, CHADS2 score ≥ 2 was a favorable predictor for net clinical benefit in a largescale AF study cohort (ATRIA) [39], as the risk of ischemic stroke without anticoagulation surpassed the risk of bleeding increased by anticoagulation. However, interestingly, in AF patients with LC, VKA treatment in CHADS2 ≥2 or CHA2DS2-VASc ≥2 showed just the trend of net clinical benefit without statistical significance. It is supposed that increased bleeding risk by varix and coagulopathy in LC patients offsetted the beneficial effect of stroke prevention. 4.3. VKA and bleeding in LC patients Correlation analysis revealed the positive correlation between HASBLED and CP scores. Chronic liver disease is a component of the HASBLED scoring system, which accounts for 1 point out of the total 9 points. However, the LC grade was never considered in most of the bleeding risk stratification systems. Our study shows that AF patients with advanced LC are more prone to bleeding events compared to early stage LC patients. Actually, several components of bleeding risk prediction systems (e.g., previous bleeding history, predisposition, labile INR, or alcohol consumption) are well-known features of LC, and these findings are more definite in advanced stages of the disease [40–43]. Therefore, the severity of LC should be considered in the bleeding risk stratification of AF patients with LC. Actually, when we analyzed according to hepatic functional reserve, VKA did not increase the bleeding risk in early LC group, whereas significant risk increment for the GI bleeding was observed in advanced LC group probably due to higher prevalence varix in advanced LC (CP-A, 19% vs CP-B/C, 32%, p = 0.01). Furthermore, among 11 advanced LC patients who experienced major bleeding events without VKA treatment, 7 (64%) patients had spontaneously prolonged INR more than 2 with an average of 2.4 ± 0.3. This finding implies that INR prolongation with or without VKA is a strong predictor of bleeding and worse outcome in AF patients with LC. 4.4. Study limitations
Fig. 2. Kaplan–Meier estimate for cumulative survival free from significant clinical events. There was no significant difference between the patients with or without VKA (p = 0.91).
Because this study is a retrospective study and not a randomized study, it carries several limitations of such trials, including selection bias. Data abstracted from medical records is limited by the degree of
190
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
Fig. 3. Kaplan–Meier estimates for (A and C) survival free from ischemic stroke or (B and D) major bleeding in early and advanced LC patients. VKA treatment significantly reduced the risk of ischemic stroke, without increasing major bleeding risk in early LC. However, in advanced LC, there was no difference in the stroke free survival, whereas VKA increased risk of bleeding significantly.
documentation. Second, detailed reasons for using or not using VKA in each patient were not clear. Moreover, more than 10% of patients were prescribed anti-platelet agents, which incidence was higher in patients without VKA. As these agents can influence the risk of ischemic
stroke or bleeding, it should be considered in the interpretation of data. Finally, no final statement can be asserted by this retrospective study, suggesting a prospectively designed trial to clearly show the benefit of VKA treatment in early LC patients with AF.
Fig. 4. Kaplan–Meier estimates for cumulative survival free from significant clinical events (SCEs) according to hepatic functional reserve. VKA treatment showed beneficial effect in early LC patients (p = 0.04), whereas it increased SCEs in advanced LC patients (p = 0.05).
Fig. 5. Net clinical benefit of VKA treatment according to risk groups. Values represent ischemic stroke equivalents prevented per 100 patient-years by VKA treatment.
S.-J. Lee et al. / International Journal of Cardiology 180 (2015) 185–191
5. Conclusions The VKA treatment might be beneficial to reduce significant clinical events in the early LC but not in the advanced LC group. However to confirm this hypothesis, a prospective randomized study is needed.
[17] [18]
Source of funding
[19]
This study was supported in part by research grants of the Basic Science Research Program through the National Research Foundation of Korea, the Ministry of Education, Science, and Technology (NRF-20100021993, NRF-2012R1A2A2A02045367), and the Korean Healthcare Technology R&D Project, Ministry of Health & Welfare (HI12C1552).
[20]
[21]
[22]
Conflict of interest [23]
None. Acknowledgments The authors of this manuscript certify that it complies with the Principles of Ethical Publishing in the International Journal of Cardiology.
[24] [25]
[26]
Appendix A. Supplementary data [27]
Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2014.11.183.
[28]
References
[29]
[1] S. Stewart, C.L. Hart, D.J. Hole, J.J. McMurray, Population prevalence, incidence, and predictors of atrial fibrillation in the Renfrew/Paisley study, Heart 86 (2001) 516–521. [2] A.S. Go, E.M. Hylek, K.A. Phillips, et al., Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study, JAMA 285 (2001) 2370–2375. [3] K.L. Furie, S.E. Kasner, R.J. Adams, et al., Guidelines for the prevention of stroke in patients with stroke or transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association, Stroke 42 (2011) 227–276. [4] R. Pisters, D.A. Lane, R. Nieuwlaat, C.B. de Vos, H.J. Crijns, G.Y. Lip, A novel userfriendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey, Chest 138 (2010) 1093–1100. [5] C.W. Siu, H.F. Tse, Net clinical benefit of warfarin therapy in elderly Chinese patients with atrial fibrillation, Circ. Arrhythm Electrophysiol. 7 (2014) 300–306. [6] S. Vong, B.P. Bell, Chronic liver disease mortality in the United States, 1990–1998, Hepatology 39 (2004) 476–483. [7] C.W. Shepard, L. Finelli, M.J. Alter, Global epidemiology of hepatitis C virus infection, Lancet Infect. Dis. 5 (2005) 558–567. [8] H.B. El-Serag, K.L. Rudolph, Hepatocellular carcinoma: epidemiology and molecular carcinogenesis, Gastroenterology 132 (2007) 2557–2576. [9] J. Tomoda, M. Mizuno, T. Sugihara, T. Itano, T. Tsuji, Gastric mucosal lesion in liver disease: impaired gastric mucosal defence mechanism in rats with induced liver injury and in patients with liver cirrhosis, J. Gastroenterol. Hepatol. 4 (Suppl. 1) (1989) 136–139. [10] J.C. Luo, H.B. Leu, M.C. Hou, et al., Cirrhotic patients at increased risk of peptic ulcer bleeding: a nationwide population-based cohort study, Aliment. Pharmacol. Ther. 36 (2012) 542–550. [11] H. Niizuma, J. Suzuki, T. Yonemitsu, T. Otsuki, Spontaneous intracerebral hemorrhage and liver dysfunction, Stroke 19 (1988) 852–856. [12] Y. Fujii, S. Takeuchi, R. Tanaka, T. Koike, O. Sasaki, T. Minakawa, Liver dysfunction in spontaneous intracerebral hemorrhage, Neurosurgery 35 (1994) 592–596. [13] B.F. Gage, Y. Yan, P.E. Milligan, et al., Clinical classification schemes for predicting hemorrhage: results from the National Registry of Atrial Fibrillation (NRAF), Am. Heart J. 151 (2006) 713–719. [14] A. Rash, T. Downes, R. Portner, W.W. Yeo, N. Morgan, K.S. Channer, A randomised controlled trial of warfarin versus aspirin for stroke prevention in octogenarians with atrial fibrillation (WASPO), Age Ageing 36 (2007) 151–156. [15] P.T. Akins, H.A. Feldman, R.G. Zoble, et al., Secondary stroke prevention with ximelagatran versus warfarin in patients with atrial fibrillation: pooled analysis of SPORTIF III and V clinical trials, Stroke 38 (2007) 874–880. [16] A.S. Go, E.M. Hylek, L.H. Borowsky, K.A. Phillips, J.V. Selby, D.E. Singer, Warfarin use among ambulatory patients with nonvalvular atrial fibrillation: the anticoagulation
[30] [31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40] [41]
[42] [43]
191
and risk factors in atrial fibrillation (ATRIA) study, Ann. Intern. Med. 131 (1999) 927–934. S.J. Connolly, M.D. Ezekowitz, S. Yusuf, et al., Dabigatran versus warfarin in patients with atrial fibrillation, N. Engl. J. Med. 361 (2009) 1139–1151. P. Petersen, G. Boysen, J. Godtfredsen, E.D. Andersen, B. Andersen, Placebo-controlled, randomised trial of warfarin and aspirin for prevention of thromboembolic complications in chronic atrial fibrillation. The Copenhagen AFASAK study, Lancet 1 (1989) 175–179. H.G. Site, Clopidogrel plus aspirin versus oral anticoagulation for atrial fibrillation in the Atrial fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events (ACTIVE W): a randomised controlled trial, Lancet 367 (2006) 1903–1912. H. Decousus, V.F. Tapson, J.F. Bergmann, et al., Factors at admission associated with bleeding risk in medical patients: findings from the IMPROVE investigators, Chest 139 (2011) 69–79. M.G. Delgado, S. Seijo, I. Yepes, et al., Efficacy and safety of anticoagulation on patients with cirrhosis and portal vein thrombosis, Clin. Gastroenterol. Hepatol. 10 (2012) 776–783. K.I. Rodriguez-Castro, P. Simioni, P. Burra, M. Senzolo, Anticoagulation for the treatment of thrombotic complications in patients with cirrhosis, Liver Int. 32 (2012) 1465–1476. S.D. Berkowitz, D.C. Sane, K.N. Sigmon, et al., Occurrence and clinical significance of thrombocytopenia in a population undergoing high-risk percutaneous coronary revascularization. Evaluation of c7E3 for the Prevention of Ischemic Complications (EPIC) Study Group, J. Am. Coll. Cardiol. 32 (1998) 311–319. M.H. Tsai, Y.S. Peng, Y.C. Chen, et al., Adrenal insufficiency in patients with cirrhosis, severe sepsis and septic shock, Hepatology 43 (2006) 673–681. U. Heemann, U. Treichel, J. Loock, et al., Albumin dialysis in cirrhosis with superimposed acute liver injury: a prospective, controlled study, Hepatology 36 (2002) 949–958. L. Sandrin, B. Fourquet, J.M. Hasquenoph, et al., Transient elastography: a new noninvasive method for assessment of hepatic fibrosis, Ultrasound Med. Biol. 29 (2003) 1705–1713. J. Foucher, E. Chanteloup, J. Vergniol, et al., Diagnosis of cirrhosis by transient elastography (FibroScan): a prospective study, Gut 55 (2006) 403–408. G.Y. Lip, S.J. Edwards, Stroke prevention with aspirin, warfarin and ximelagatran in patients with non-valvular atrial fibrillation: a systematic review and meta-analysis, Thromb. Res. 118 (2006) 321–333. H.D. White, M. Gruber, J. Feyzi, et al., Comparison of outcomes among patients randomized to warfarin therapy according to anticoagulant control: results from SPORTIF III and V, Arch. Intern. Med. 167 (2007) 239–245. D.B. Matchar, A. Jacobson, R. Dolor, et al., Effect of home testing of international normalized ratio on clinical events, N. Engl. J. Med. 363 (2010) 1608–1620. S.J. Connolly, J.W. Eikelboom, J. Ng, et al., Net clinical benefit of adding clopidogrel to aspirin therapy in patients with atrial fibrillation for whom vitamin K antagonists are unsuitable, Ann. Intern. Med. 155 (2011) 579–586. F.R. Rosendaal, S.C. Cannegieter, F.J. van der Meer, E. Briet, A method to determine the optimal intensity of oral anticoagulant therapy, Thromb. Haemost. 69 (1993) 236–239. E.M. Hylek, A.S. Go, Y. Chang, et al., Effect of intensity of oral anticoagulation on stroke severity and mortality in atrial fibrillation, N. Engl. J. Med. 349 (2003) 1019–1026. S.J. Connolly, J. Pogue, J. Eikelboom, et al., Benefit of oral anticoagulant over antiplatelet therapy in atrial fibrillation depends on the quality of international normalized ratio control achieved by centers and countries as measured by time in therapeutic range, Circulation 118 (2008) 2029–2037. A. Berzigotti, A. Bonfiglioli, A. Muscari, et al., Reduced prevalence of ischemic events and abnormal supraortic flow patterns in patients with liver cirrhosis, Liver Int. 25 (2005) 331–336. J.B. Olesen, G.Y. Lip, M.L. Hansen, et al., Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study, BMJ 342 (2011) d124. G.Y. Lip, R. Nieuwlaat, R. Pisters, D.A. Lane, H.J. Crijns, Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation, Chest 137 (2010) 263–272. European Heart Rhythm A, European Association for Cardio-Thoracic S, A.J. Camm, et al., Guidelines for the management of atrial fibrillation: the task force for the management of atrial fibrillation of the European Society of Cardiology (ESC), Eur. Heart J. 31 (2010) 2369–2429. L. Friberg, M. Rosenqvist, G.Y. Lip, Net clinical benefit of warfarin in patients with atrial fibrillation: a report from the Swedish atrial fibrillation cohort study, Circulation 125 (2012) 2298–2307. P. Gentilini, G. Laffi, G. La Villa, et al., Long course and prognostic factors of virusinduced cirrhosis of the liver, Am. J. Gastroenterol. 92 (1997) 66–72. P. Jepsen, P. Ott, P.K. Andersen, H.T. Sorensen, H. Vilstrup, Clinical course of alcoholic liver cirrhosis: a Danish population-based cohort study, Hepatology 51 (2010) 1675–1682. C.S. Lieber, Medical disorders of alcoholism, N. Engl. J. Med. 333 (1995) 1058–1065. W. Blonski, T. Siropaides, K.R. Reddy, Coagulopathy in liver disease, Curr. Treat. Options Gastroenterol. 10 (2007) 464–473.