[
Original Research Cardiovascular Disease
]
Effect of Active Smoking on Comparative Efficacy of Antithrombotic Therapy in Patients With Atrial Fibrillation The Loire Valley Atrial Fibrillation Project Denis Angoulvant, MD; Olivier Villejoubert, MD; Theodora Bejan-Angoulvant, MD; Fabrice Ivanes, MD; Christophe Saint Etienne, MD; Gregory Y. H. Lip, MD; and Laurent Fauchier, MD
Active smoking is associated with elevated thrombotic risk. Smoking status has recently been incorporated into the SAMe-TT2R2 (sex female, age , 60 years, medical history [more than two comorbidities], treatment [interacting drugs, eg, amiodarone for rhythm control], tobacco use [doubled], race [doubled]) score that can help predict poor international normalized ratio control in patients with atrial fibrillation (AF) treated with vitamin K antagonists (VKAs). The clinical benefit of antiplatelet therapy (APT) has been seen primarily in smokers. We hypothesized that active smoking may differently influence the risks of stroke and bleeding in patients with AF treated with VKAs or with APT. BACKGROUND:
METHODS: We examined the clinical course of 7,809 consecutive patients with AF seen between 2000 and 2010. Outcomes in patients who were active smokers were compared with those in other patients.
Among 7,809 patients with AF, 1,034 (13%) were active smokers. APT was prescribed on an individual basis for 2,761 patients (35%) and VKAs for 4,534 (57%). After a follow-up of 929 1,082 days (median 5 463 days, interquartile range 5 1,564 days), smoking was not independently associated with a higher risk of stroke/thromboembolic event in patients with AF (hazard ratio [HR], 0.95; 95% CI, 0.78-1.22; P 5 .66). On multivariate analysis, smoking was independently associated with a worse prognosis for the risk of severe bleeding (HR, 1.23; 95% CI, 1.01-1.49; P 5 .04) and for the risk of major Bleeding Academic Research Consortium bleeding (HR, 1.40; 95% CI, 1.02-1.90; P 5 .03). Smoking was independently associated with a higher risk of bleeding in patients treated with VKAs (HR, 1.32; 95% CI, 1.04-1.67; P 5 .02), whereas the risk was nonsignificant in patients treated with APT (HR, 1.28; 95% CI, 0.94-1.74; P 5 .11).
RESULTS:
In AF, there was a higher risk of severe bleeding in smokers, mainly in those CHEST 2015; 148(2):491-498 treated with VKAs.
CONCLUSIONS:
Manuscript received December 2, 2014; revision accepted February 2, 2015; originally published Online First March 26, 2015. ABBREVIATIONS: AF 5 atrial fibrillation; APT 5 antiplatelet therapy; BARC 5 Bleeding Academic Research Consortium; CHA2DS2-VASc 5 congestive heart failure, hypertension, age 75 years (doubled), diabetes, stroke (doubled), vascular disease, age 65 to 74 years, sex category (female); HAS-BLED 5 hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly (. 65 years), drugs (antiplatelet drugs or nonsteroidal antiinflammatory drugs)/alcohol excess concomitantly; HR 5 hazard ratio; INR 5 international normalized ratio; SAMe-TT2R2 5 sex female, age , 60 years, medical history (more than two comorbidities), treatment
journal.publications.chestnet.org
(interacting drugs, eg, amiodarone for rhythm control), tobacco use (doubled), race (doubled); TTR 5 time in therapeutic range; VKA 5 vitamin K antagonist AFFILIATIONS: From the Service de Cardiologie (Drs Angoulvant, Villejoubert, Bejan-Angoulvant, Ivanes, Saint Etienne, and Fauchier), Centre Hospitalier Universitaire Trousseau et Faculté de Médecine, Université François Rabelais, Tours, France; and the University of Birmingham Centre for Cardiovascular Sciences (Dr Lip), City Hospital, Birmingham, England. Drs Lip and Fauchier were the joint senior authors of this manuscript. FUNDING/SUPPORT: The authors have reported to CHEST that no funding was received for this study. 491
Atrial fibrillation (AF) confers a fivefold increased risk of stroke, and 15% to 20% of all strokes are caused by AF. Importantly, mortality, morbidity, and economic burden from stroke complicating AF are particularly high. The European Society of Cardiology guidelines recommend the use of the CHA2DS2-VASc (congestive heart failure, hypertension, age 75 years [doubled], diabetes, stroke [doubled], vascular disease, age 65 to 74 years, and sex category [female]) score to evaluate the individual thromboembolic risk associated with AF and to determine the risk to benefit ratio of antithrombotic therapy.1,2 Until recently, the latter essentially referred to oral anticoagulants (mainly vitamin K antagonists [VKAs]), or antiplatelet therapy (APT), usually aspirin.1,3 The main adverse effect of antithrombotic therapy is bleeding and, as recommended by the European Society of Cardiology guidelines, this risk may be assessed using the HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly [. 65 years], drugs [antiplatelet drugs or nonsteroidal antiinflammatory drugs]/alcohol excess concomitantly) score.1,4
Active smoking is a frequent cardiovascular risk factor usually associated with a higher risk of thrombotic events. Smoking independently influences poor international normalized ratio (INR) control and has been incorporated into a simple score (the SAMe-TT2R2 [sex female, age , 60 years, medical history (more than two comorbidities), treatment (interacting drugs, eg, amiodarone for rhythm control), tobacco use (doubled), race (doubled)] score) in patients with AF to predict the likelihood of good anticoagulation control in a patient initiated on VKAs.5 Moreover, the clinical benefit of clopidogrel in reducing myocardial infarction and stroke in randomized clinical trials has been seen primarily in smokers, with little benefit seen among nonsmokers.6 Smoking status could, therefore, expose patients treated for AF to an excess risk of thromboembolic events linked to “low efficacy of their antithrombotic treatment” and/or an excess risk of bleeding complications associated with a “larger efficacy of their antithrombotic treatment.” In the current study, we investigated whether active smoking would differently influence the risks of stroke and bleeding in patients with AF treated with VKAs or with APT.
Materials and Methods
diabetes or vascular disease, and sex category [female]; two points for a prior stroke or transient ischemic attack or age 75 years)2 and the HAS-BLED bleeding risk score (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile INR, elderly [. 65 years], drugs/alcohol concomitantly) to assess bleeding risk, where a score of 3 indicates high risk.4
We included all patients with a diagnosis of AF, atrial flutter, or both seen in the cardiology department of the Centre Hospitalier Universitaire Trousseau, Tours, France between January 2000 and December 2010. Patient characteristics were obtained from the records of the institution’s computerized codification system for each patient. Extensive information on date of admission, discharge, diagnosis, clinical presentation, comorbidities, medication, and subsequent hospitalization were collected. For each patient, the thromboembolic risk was estimated using the CHA2DS2-VASc score and the hemorrhagic risk using the HAS-BLED bleeding risk score. The SAMe-TT2R2 score was used to predict poor INR control in patients with AF treated with VKAs. The SAMe-TT2R2 score was calculated as the sum of points after the addition of one point each for female sex, age , 60 years, medical history of more than two comorbidities (from among hypertension, diabetes, coronary artery disease/myocardial infarction, peripheral arterial disease, congestive heart failure, previous stroke, pulmonary disease, and hepatic or renal disease), treatment (interacting drugs [eg, amiodarone for rhythm control]), and two points each for tobacco use and nonwhite race. Because recording a patient’s race is not allowed in any electronic file in France, the last item was actually not included in the addition, but the local population was essentially white. Patients with a SAMe-TT2R2 score of 0 to 2 were deemed to have low-moderate risk and those with a score . 2 to have a high risk of poor INR control with a VKA.5 We also calculated the CHA2DS2-VASc score (one point each for a history of heart failure, a history of hypertension, age 65 to 75 years, the presence of CORRESPONDENCE TO: Laurent Fauchier, MD, Service de Cardiologie et Laboratoire d’Electrophysiologie Cardiaque, Centre Hospitalier Universitaire Trousseau, 37044 Tours, France; e-mail: [email protected] © 2015 AMERICAN COLLEGE OF CHEST PHYSICIANS. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details. DOI: 10.1378/chest.14-3006
492 Original Research
Information on stroke, thromboembolism, and bleeding events was recorded whenever documented within the institution, which includes a total of four hospitals covering all specialties, and was obtained by searching in the medical computerized database. Severe bleeding was defined as a decrease of . 5.0 g/dL in the blood hemoglobin level (including the period around the coronary interventional procedure), the need for transfusion of one or more units of blood, the need for corrective surgery, the occurrence of an intracranial or retroperitoneal hemorrhage, or any combination of these events. We also considered major bleedings, using the Bleeding Academic Research Consortium (BARC) bleeding definitions: intracranial hemorrhage, intraocular compromising vision, overt bleeding plus hemoglobin drop of . 5 g/dL, tamponade, bleeding requiring surgical or percutaneous intervention for control (excluding dental/nose/skin/hemorrhoids) or inotropes (BARC type 3A), any transfusion with overt bleeding, overt bleeding plus hemoglobin drop of 3 to 5 g/dL (BARC type 3B), or fatal bleeding.7,8 Finally, we considered separately the GI location of bleeding because smoking is a known risk factor for GI bleeding. The Centre Hospitalier Regional et Universitaire de Tours serves a population of approximately 400,000 and is the only public institution in an area of around 4,000 km2. Patients were followed up for stroke and thromboembolic and bleeding events. In addition, deaths were identified using an online search tool dedicated to local news, covering an area of 35,000 km2 (http://nrco.lanouvellerepublique.fr/dossiers/necro/ index.php).The outcomes in patients with active smoking were compared with those in other patients.9 Statistical Analysis Patients’ characteristics were reported as percentages or means SD. Comparisons between groups were made using x2 tests to compare
[
148#2 CHEST AUGUST 2015
]
categorical variables and the Student t test or nonparametric KruskalWallis test where appropriate for continuous variables. Event rates of stroke/thromboembolism and bleeding were calculated, and multivariate analysis with a proportional hazard model was used to investigate the association between active smoking and each of the outcomes. The results were expressed as relative risk and 95% CI. A P value , .05 was considered statistically significant. Statistical analysis was carried out with Statview 5.0 software (Abacus Concepts).
Ethics Approval The study was approved by the institutional review board of the Pole Coeur Thorax Vaisseaux from the Trousseau University Hospital (Tours, France) on December 7, 2010, and was registered as a clinical audit. Ethical review was, therefore, not required. Patient consent was not sought. Patient data were used only to facilitate the cross-referencing of data sources, and records were otherwise anonymous. The study was conducted retrospectively, patients were not involved in its conduct, and there was no impact on their care.
Results
analysis after adjustment for possible confounders as detailed in Table 1, there was no significant difference in AF between nonsmoker and active smoker patients.
A total of 7,809 patients with AF (62% men; mean age, 71 15 years) were included in this study, which covered the period 2000 to 2010. VKA and APT distribution between smokers and nonsmokers are depicted in Figure 1. Smokers were younger than nonsmokers, and there were more men and comorbidities (hypertension, diabetes, renal insufficiency, coronary artery disease, heart failure, and lower left ventricular ejection fraction) among smokers. The CHA2DS2-VASc score was similar in the two groups, whereas smokers had higher HAS-BLED and SAMe-TT2R2 scores. There was a similar number of patients given VKAs in the smoking group and in the nonsmoking group, whereas there were more patients treated with APT in the smoking group (Table 1). Follow-up: Stroke/Thromboembolism
We recorded 631 strokes/thromboembolic events during a 929 1,082-day follow-up (median, 63 days; interquartile range, 1,564 days). Figure 2 shows stroke or systemic thromboembolism events in patients with AF with or without smoking during follow-up. Smoking was not independently associated with a higher risk of stroke in those patients with AF (relative risk, 0.95; 95% CI, 0.78-1.22; P 5 .66). On multivariate
Figure 1 – Antithrombotic therapy of the patients with AF with or without active smoking. AF 5 atrial fibrillation; APT 5 antiplatelet therapy; VKA 5 vitamin K antagonist.
journal.publications.chestnet.org
Follow-up: Bleeding Events
During follow-up, 707 severe bleedings and 248 major BARC bleedings were observed. Figure 3 shows bleeding events in patients with AF with or without smoking during follow-up. After adjustment for age, smoking status, CHA2DS2-VASc score, HAS-BLED bleeding risk score, use of VKAs, and use of APT, there was an excess risk of severe bleeding in smokers (hazard ratio [HR], 1.23; 95% CI, 1.01-1.49) and patients with VKA use (HR, 1.26; 95% CI, 1.06-1.50). Older patients and those with a higher HAS-BLED bleeding risk score also had an excess risk of severe bleeding. This excess risk of severe bleeding in smokers persisted significantly in patients receiving VKAs and was nonsignificant in patients receiving APT (Table 2). We also performed a multivariate analysis for bleeding events with COPD added to the model because this condition was more prevalent in smokers. COPD was an independent predictor of bleeding events, but the overall findings were similar: Smoking was an independent predictor of bleeding events in patients receiving VKAs (HR, 1.37; 95% CI, 1.08-1.75; P 5 .01 for COPD; HR, 1.26; 95% CI, 1.01-1.60; P 5 .05 for active smoking) and was nonsignificant in patients receiving APT (HR, 1.48; 95% CI, 1.07-2.04; P 5 .02 for COPD; HR, 1.24; 95% CI, 0.92-1.68; P 5 .15 for active smoking). For major BARC bleeding, active smoking and use of VKAs were also associated with a significant excess risk in the multivariate analysis using a model including age, smoking status, HAS-BLED bleeding risk score, CHA2DS2-VASc score, use of VKAs, and use of APT (Table 3). Smoking was not significantly associated with a higher risk of GI bleeding in those patients with AF (relative risk, 1.28; 95% CI, 0.84-1.94; P 5 .25). Finally, smoking was found to be an independent predictor of all-cause death and cardiovascular death in patients with AF (Table 3). 493
TABLE 1
] Characteristics of Patients With AF With or Without Active Smoking
Variable
AF and No Smoking (n 5 6775)
AF and Smoking (n 5 1034)
P Value
Age, y
72 15
68 13
, .0001
Women
43
11
, .0001
Heart failure
54
60
, .0001
Hypertension
40
52
, .0001
Diabetes mellitus
, .0001
14
25
Previous ischemic stroke
8
8
Coronary artery disease
27
51
, .0001
Valve disease
23
23
.84
.42
CHA2DS2-VASc score
3.1 1.8
3.0 1.7
.22
HAS-BLED score
1.5 1.1
1.9 1.2
, .0001
SAMe-TT2R2 score
1.2 0.9
3.2 0.8
, .0001
9
11
.004
9
20
, .0001
39
38
.38
Renal insufficiency COPD Permanent AF Pacemaker or ICD
17
15
.06
47 16
45 15
.02
VKA
57
55
.21
Aspirin
29
40
, .0001
LVEF (n 5 1934) Medication during follow-up
Clopidogrel
9
18
, .0001
VKA 1 single APT
7
11
, .0001
2
4
.0002
ACE inhibitor or ARB
VKA 1 dual APT
39
49
, .0001
b-Blocker
44
46
.37
Diuretic
47
51
.07
Class 3 antiarrhythmic agent
45
44
.54
Data are presented as mean SD or %. ACE 5 angiotensin-converting enzyme; AF 5 atrial fibrillation; APT 5 antiplatelet therapy; ARB 5 angiotensin II receptor blocker; CHA2DS2-VASc 5 congestive heart failure, hypertension, age 75 y (doubled), diabetes, stroke (doubled), vascular disease, age 65 to 74 y, and sex category (female); HAS-BLED 5 hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly (. 65 y), drugs (antiplatelet drugs or nonsteroidal antiinflammatory drugs)/alcohol excess concomitantly; ICD 5 implantable cardioverter defibrillator; LVEF 5 left ventricular ejection fraction; SAMe-TT2R2 5 sex female, age , 60 y, medical history (more than two comorbidities), treatment (interacting drugs, eg, amiodarone for rhythm control), tobacco use (doubled), race (doubled); VKA 5 vitamin K antagonist.
Discussion Our principal findings were that actively smoking patients with AF did not have a higher risk of thromboembolic events but had a significantly higher risk of severe bleeding, particularly so in patients treated with VKAs. To our knowledge, this is one of the first reports from a large series to highlight an excess bleeding risk in smoking patients with AF. Smoking could simply reflect the additional bleeding risk factors but may also more directly predispose to bleeding. Our observations also indirectly confirm the relevance of the HAS-BLED bleeding score, which was 494 Original Research
higher in the smoking group.1,4 When investigating whether this excess bleeding risk among smokers with AF in our cohort was drug related, the increase in bleeding risk was significant in patients taking VKAs, whereas it was not significant in patients treated with APT. Actively smoking patients with AF were significantly younger, with a significantly higher male predominance, and had significantly more comorbidities, with a higher HAS-BLED score. Nonetheless, there was still an independent and significant association between active smoking and bleeding risk on multivariate analysis (Table 2). The benefit of adding smoking to the HAS-BLED scoring formula is debatable because this additional parameter may offer
[
148#2 CHEST AUGUST 2015
]
TABLE 2
] Cox Regression Analysis for Prediction of Severe Bleeding Events
Predictors of Severe Bleeding (707 Events)
Hazard Ratio (95% CI)
P Value
1.02 (1.01-1.02)
, .0001
All patients Age Smoking
1.23 (1.01-1.49)
.04
HAS-BLED score
1.24 (1.13-1.37)
, .0001
CHA2DS2-VASc score
0.99 (0.93-1.07)
.88
VKA at discharge
1.26 (1.06-1.50)
.009
APT at discharge
1.04 (0.87-1.25)
.66
Age
1.01 (1.00-1.02)
.005
Smoking
1.32 (1.04-1.67)
.02
Patients treated with VKA
Figure 2 – Stroke or systemic thromboembolism in patients with AF with or without active smoking (mean 884 1,054 d of follow-up; n 5 631 events). RR 5 relative risk. See Figure 1 legend for expansion of other abbreviation.
marginal improvement in predictive ability at the cost of losing simplicity, ease of use, and everyday practicality. Numerous studies have shown that the rates of stroke or thromboembolic events and major bleeding are lowest in patients with a high average individual time in therapeutic range (TTR). In addition, a recent European consensus recommended an optimal average TTR of . 70% in patients treated with VKAs.10,11 Apostolakis et al5 proposed the SAMe-TT2R2 score, which was useful in predicting those patients with a high TTR . 70% (SAMe-TT2R2 score 0-1) who would do well taking VKAs and those who would do less well (SAMe-TT2R2 score 2). In this score, active smoking is given two points, predicting in itself a smaller TTR. Moreover, we recently showed that the SAMe-TT2R2 score was predictive for an increased risk of stroke or thromboembolic events, severe bleeding, major BARC bleeding, and death, reflecting poor anticoagulation control in patients with
Figure 3 – Bleeding events in patients with AF with or without active smoking (mean 877 1,052 d of follow-up; n 5 707 events). See Figure 1 and 2 legends for expansion of abbreviations.
journal.publications.chestnet.org
HAS-BLED score
1.21 (1.07-1.32)
.002
CHA2DS2-VASc score
1.04 (0.96-1.13)
.36
APT at discharge
1.12 (0.88-1.42)
.34
Age
1.03 (1.02-1.05)
, .0001
Smoking
1.28 (0.94-1.74)
.11
HAS-BLED score
1.28 (1.10-1.49)
.001
CHA2DS2-VASc score
0.88 (0.79-0.99)
.03
VKA at discharge
1.36 (1.05-1.76)
.02
Aspirin
0.88 (0.60-1.30)
.53
Clopidogrel
1.19 (0.88-1.60)
.25
Patients treated with APT
See Table 1 legend for expansion of abbreviations.
AF treated with VKAs.8 The current study may explain some of the results of this previous observation because we found an independent excess bleeding risk in actively smoking patients with AF treated with VKAs.8 Warfarin is eliminated through hepatic metabolism, and some experiments in tissue, animal, and human models have indicated that active smoking is associated with enhanced drug clearance through hepatic microsomal enzyme induction.12,13 A meta-analysis by Nathisuwan et al14 suggested that smoking may cause significant interactions with warfarin by increasing warfarin clearance, which leads to reduced warfarin effects. They found a significant increase in the warfarin dosage requirement in patients who were active smokers when compared with nonsmokers.14,15 Moreover, previous studies demonstrated that chronic smoking cessation was associated with increased INR values.16 This increased INR value after smoking cessation could explain the higher number of hemorrhagic events in active smokers in our cohort. Indeed, one of our hypotheses to explain this excess bleeding risk 495
TABLE 3
] Cox Regression Analysis for Prediction of Major BARC Bleeding Events
Predictors of Outcomes
OR (95% CI)
P Value
1.00 (0.99-1.01)
.96
Major BARC bleeding (248 events) Age Smoking
1.40 (1.02-1.90)
.03
HAS-BLED score
1.40 (1.20-1.64)
, .0001
CHA2DS2-VASc score
0.98 (0.88-1.10)
.77
VKA at discharge
1.72 (1.27-2.31)
.0004
APT at discharge
1.27 (0.95-1.74)
.11
1.05 (1.04-1.06)
, .0001
Smoking
1.39 (1.17-1.64)
.0002
HAS-BLED score
1.10 (1.01-1.20)
.02
CHA2DS2-VASc score
1.12 (1.06-1.19)
.0001
VKA at discharge
0.58 (0.50-0.66)
, .0001
APT at discharge
0.70 (0.60-0.82)
, .0001
1.05 (1.04-1.06)
, .0001
Smoking
1.41 (1.14-1.75)
.0002
HAS-BLED score
1.13 (1.01-1.26)
.03
CHA2DS2-VASc score
1.16 (1.07-1.24)
.0001
VKA at discharge
0.61 (0.51-0.73)
, .0001
APT at discharge
0.74 (0.61-0.90)
.003
All-cause death (989 events) Age
Cardiovascular death (595 events) Age
BARC 5 Bleeding Academic Research Consortium. See Table 1 legend for expansion of other abbreviations.
among smokers in our population of patients with AF treated with VKAs is that smoking induces INR instability, which is known to increase bleeding risk. (Of note, INR instability while taking VKAs was a component of the HAS-BLED risk score.4) Indeed, Nathisuwan et al14 recommended close INR monitoring between 1 week and 3 months after smoking cessation. Unfortunately, smoking cessation was not mentioned in our cohort, so we cannot measure its impact on bleeding risk. On the other hand, previous studies found no significant relationship between smoking and nontherapeutic INR (sub- and supratherapeutic INR).17,18 The pharmacologic interaction between smoking and antithrombotic effect is not linked to the importance of tobacco intoxication, but to active smoking. Other plausible hypotheses may also explain our findings. Active smokers may tend to pay less attention to alcohol, diet, vegetables, herbal products, or other environmental factors, which may interfere 496 Original Research
with warfarin response. Thus, the excess risk may be explained partly by INR instability in active smokers. The Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial analyzed the risk of major bleeding in patients with acute coronary syndrome according to smoking status and found that smokers had a higher risk of bleeding complications.19,20 Tobacco smoking results in increased platelet aggregability, and some authors have suggested that low-dose aspirin cannot sufficiently inhibit the platelet aggregability in smokers with coronary artery disease.21,22 In our work, we found no excess risk of thromboembolic events in smokers treated with aspirin. Of note, Zhao et al23 suggested in a recent metaanalysis that current smokers have significantly lower on-clopidogrel platelet reactivity. This could explain this higher risk of bleeding complications in active smokers treated with clopidogrel. Because of the small number of patients taking clopidogrel in our study population, the influence of this drug on bleeding risk in patients with AF cannot be evaluated. Importantly, whereas smoking may be a risk factor for GI bleeding, the excess bleeding risk in active smokers compared with nonsmokers in our cohort did not result from an increase in GI bleeding. The “smoker’s paradox” refers to the observation of a favorable prognosis in current smokers following an acute myocardial infarction.24 Several hypotheses have been proposed to explain it. Some involve a modification of the response to APT in smokers, especially for clopidogrel.6,23 In the current study, we did not show such a paradox or a higher risk of bleeding complications in patients who were active smokers treated with clopidogrel, because they did not have fewer thromboembolisms or more bleeding events. Indeed, we did not find an increased thromboembolic risk in active smokers in our cohort, consistent with the relevance of the CHA2DS2-VASc score, which was similar in both groups. In a recent study with a smaller number of patients, Nakagawa et al25 found more fatal strokes, higher all-cause mortality, and more intracranial bleeding in patients who smoked, even when controlling for risk factors. However, persistent smoking did not independently predict the risk of stroke, whereas it predicted intracranial bleeding independent of age, and these results are consistent with our findings. Study Limitations
Our work was an observational study, not a randomized trial. Data were obtained retrospectively from a search
[
148#2 CHEST AUGUST 2015
]
in the hospital discharge records, with the limitations of diagnostic coding and case ascertainment. It was also a monocentric study, and our results should be interpreted with caution when applied to the general population. Although we had information on major episodes of labile INR for some patients, data regarding a precise TTR were are not available, which is a limitation for the analysis of bleeding risk. Finally, we did not have data regarding the intensity and regularity of tobacco use or the possibility of smoking cessation during follow-up.
Acknowledgments Author contributions: L. F. is the guarantor of the study. D. A., T. B.-A., and L. F. contributed to the study conception and design; D. A., O. V., and L. F. contributed to the analyses and production of the initial manuscript; L. F. contributed to the data collection; and D. A., O. V., T. B.-A., F. I., C. S. E., G. Y. H. L., and L. F. contributed to the interpretation of the results, critical revision of the manuscript for important intellectual content, and approval of the final manuscript. Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Angoulvant has received funding for conference travel and educational symposia from AstraZeneca plc, Eli Lilly and Company, Novartis, Bayer AG, MSD, Amgen, and Pfizer Inc. Dr Lip has served as a consultant for Bayer AG, Astellas Pharma Inc, Merck & Co Inc, AstraZeneca plc, Sanofi SA, Biotronik SE & CoKG, Bristol-Myers Squibb Co/Pfizer Inc, and Boehringher Ingelheim GmbH and has been on the speakers bureau for Bayer AG, Bristol-Myers Squibb Co/Pfizer Inc, Boehringher Ingelheim GmbH, and Sanofi Aventis LLC. Dr Fauchier has served as a consultant for Bayer AG, Bristol-Myers Squibb Co/Pfizer Inc, Boehringher Ingelheim GmbH, Medtronic Inc, Novartis, and Sanofi Aventis LLC and has received funding for conference travel and educational symposia from Bayer AG, Bristol-Myers Squibb Co/Pfizer Inc, Boehringher Ingelheim GmbH, Boston Scientific, Medtronic Inc, Novartis, Sorin, and Sanofi Aventis LLC. Drs Villejoubert, Bejan-Angoulvant, Ivanes, and Saint Etienne have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
References 1. Camm AJ, Lip GY, De Caterina R, et al; ESC Committee for Practice Guidelines (CPG). 2012 focused update of the ESC guidelines for the management of atrial fibrillation: an update of the 2010 ESC guidelines for the management of atrial fibrillation. Developed with the special contribution of the European Heart Rhythm Association. Eur Heart J. 2012;33(21):2719-2747.
journal.publications.chestnet.org
Conclusions In our large cohort of consecutive patients with AF, smoking was independently associated with a higher risk of severe bleeding, particularly in patients treated with VKAs. However, smoking was not associated with an excess of thromboembolic risk. Whether our observation is driven mainly by pharmacologic interaction(s) between tobacco and VKA and/or by the direct effect of tobacco on bleeding risk remains to be investigated.
2. Olesen JB, Torp-Pedersen C, Hansen ML, Lip GYH. The value of the CHA2DS2VASc score for refining stroke risk stratification in patients with atrial fibrillation with a CHADS2 score 0-1: a nationwide cohort study. Thromb Haemost. 2012; 107(6):1172-1179. 3. Lip GYH, Tse HF, Lane DA. Atrial fibrillation. Lancet. 2012;379(9816):648-661. 4. Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJGM, Lip GYH. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138(5):1093-1100. 5. Apostolakis S, Sullivan RM, Olshansky B, Lip GYH. Factors affecting quality of anticoagulation control among patients with atrial fibrillation on warfarin: the SAMe-TT2R2 score. Chest. 2013;144(5): 1555-1563. 6. Gagne JJ, Bykov K, Choudhry NK, Toomey TJ, Connolly JG, Avorn J. Effect of smoking on comparative efficacy of antiplatelet agents: systematic review, meta-analysis, and indirect comparison. BMJ. 2013;347:f5307. 7. Mehran R, Rao SV, Bhatt DL, et al. Standardized bleeding definitions for cardiovascular clinical trials: a consensus report from the Bleeding Academic Research Consortium. Circulation. 2011; 123(23):2736-2747. 8. Lip GYH, Haguenoer K, Saint-Etienne C, Fauchier L. Relationship of the SAMe-TT2R2 score to poor-quality anticoagulation, stroke, clinically relevant bleeding and mortality in patients with atrial fibrillation. Chest. 2014;146(3): 719-726. 9. Gorin L, Fauchier L, Nonin E, Charbonnier B, Babuty D, Lip GYH. Prognosis and guideline-adherent antithrombotic treatment in patients with atrial fibrillation and atrial flutter: implications of undertreatment and overtreatment in real-life clinical practice; the Loire Valley Atrial Fibrillation Project. Chest. 2011;140(4):911-917. 10. De Caterina R, Husted S, Wallentin L, et al; European Society of Cardiology Working Group on Thrombosis Task Force on Anticoagulants in Heart Disease. General mechanisms of coagulation and targets of anticoagulants (section I). Position paper of the ESC Working Group on Thrombosis—Task Force on
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Anticoagulants in Heart Disease. Thromb Haemost. 2013;109(4):569-579. De Caterina R, Husted S, Wallentin L, et al. Vitamin K antagonists in heart disease: current status and perspectives (section III). Position paper of the ESC Working Group on Thrombosis—Task Force on Anticoagulants in Heart Disease. Thromb Haemost. 2013;110(6): 1087-1107. Holbrook AM, Pereira JA, Labiris R, et al. Systematic overview of warfarin and its drug and food interactions. Arch Intern Med. 2005;165(10):1095-1106. Zevin S, Benowitz NL. Drug interactions with tobacco smoking. An update. Clin Pharmacokinet. 1999;36(6):425-438. Nathisuwan S, Dilokthornsakul P, Chaiyakunapruk N, Morarai T, Yodting T, Piriyachananusorn N. Assessing evidence of interaction between smoking and warfarin: a systematic review and meta-analysis. Chest. 2011;139(5): 1130-1139. Bachmann K, Shapiro R, Fulton R, Carroll FT, Sullivan TJ. Smoking and warfarin disposition. Clin Pharmacol Ther. 1979;25(3):309-315. Colucci VJ, Knapp JF. Increase in international normalized ratio associated with smoking cessation. Ann Pharmacother. 2001;35(3):385-386. McGriff-Lee NJ, Csako G, Chen JT, et al; Anticoagulation Working Group. Search for predictors of nontherapeutic INR results with warfarin therapy. Ann Pharmacother. 2005;39(12): 1996-2002. Pamboukian SV, Nisar I, Patel S, et al. Factors associated with non-adherence to therapy with warfarin in a population of chronic heart failure patients. Clin Cardiol. 2008;31(1):30-34. Berger JS, Bhatt DL, Steinhubl SR, et al; CHARISMA Investigators. Smoking, clopidogrel, and mortality in patients with established cardiovascular disease. Circulation. 2009;120(23):2337-2344. Hirschl MM. Smoking status and the effects of antiplatelet drugs. BMJ. 2013;347:f5909. Li W-J, Zhang H-Y, Miao C-L, et al. Cigarette smoking inhibits the antiplatelet activity of aspirin in patients with coronary heart disease. Chin Med J (Engl). 2011;124(10):1569-1572.
497
22. Ikonomidis I, Lekakis J, Vamvakou G, Andreotti F, Nihoyannopoulos P. Cigarette smoking is associated with increased circulating proinflammatory and procoagulant markers in patients with chronic coronary artery disease: effects of aspirin treatment. Am Heart J. 2005;149(5):832-839.
498 Original Research
23. Zhao Z-G, Chen M, Peng Y, et al. The impact of smoking on clinical efficacy and pharmacodynamic effects of clopidogrel: a systematic review and meta-analysis. Heart. 2014;100(3): 192-199. 24. Aune E, Røislien J, Mathisen M, Thelle DS, Otterstad JE. The “smoker’s paradox”
in patients with acute coronary syndrome: a systematic review. BMC Med. 2011; 9:97. 25. Nakagawa K, Hirai T, Ohara K, et al. Impact of persistent smoking on longterm outcomes in patients with nonvalvular atrial fibrillation. J Cardiol. 2015; 65(5):429-433.
[
148#2 CHEST AUGUST 2015
]
Original Research Cardiovascular Disease
]
Effect of Active Smoking on Comparative Efficacy of Antithrombotic Therapy in Patients With Atrial Fibrillation The Loire Valley Atrial Fibrillation Project Denis Angoulvant, MD; Olivier Villejoubert, MD; Theodora Bejan-Angoulvant, MD; Fabrice Ivanes, MD; Christophe Saint Etienne, MD; Gregory Y. H. Lip, MD; and Laurent Fauchier, MD
Active smoking is associated with elevated thrombotic risk. Smoking status has recently been incorporated into the SAMe-TT2R2 (sex female, age , 60 years, medical history [more than two comorbidities], treatment [interacting drugs, eg, amiodarone for rhythm control], tobacco use [doubled], race [doubled]) score that can help predict poor international normalized ratio control in patients with atrial fibrillation (AF) treated with vitamin K antagonists (VKAs). The clinical benefit of antiplatelet therapy (APT) has been seen primarily in smokers. We hypothesized that active smoking may differently influence the risks of stroke and bleeding in patients with AF treated with VKAs or with APT. BACKGROUND:
METHODS: We examined the clinical course of 7,809 consecutive patients with AF seen between 2000 and 2010. Outcomes in patients who were active smokers were compared with those in other patients.
Among 7,809 patients with AF, 1,034 (13%) were active smokers. APT was prescribed on an individual basis for 2,761 patients (35%) and VKAs for 4,534 (57%). After a follow-up of 929 1,082 days (median 5 463 days, interquartile range 5 1,564 days), smoking was not independently associated with a higher risk of stroke/thromboembolic event in patients with AF (hazard ratio [HR], 0.95; 95% CI, 0.78-1.22; P 5 .66). On multivariate analysis, smoking was independently associated with a worse prognosis for the risk of severe bleeding (HR, 1.23; 95% CI, 1.01-1.49; P 5 .04) and for the risk of major Bleeding Academic Research Consortium bleeding (HR, 1.40; 95% CI, 1.02-1.90; P 5 .03). Smoking was independently associated with a higher risk of bleeding in patients treated with VKAs (HR, 1.32; 95% CI, 1.04-1.67; P 5 .02), whereas the risk was nonsignificant in patients treated with APT (HR, 1.28; 95% CI, 0.94-1.74; P 5 .11).
RESULTS:
In AF, there was a higher risk of severe bleeding in smokers, mainly in those CHEST 2015; 148(2):491-498 treated with VKAs.
CONCLUSIONS:
Manuscript received December 2, 2014; revision accepted February 2, 2015; originally published Online First March 26, 2015. ABBREVIATIONS: AF 5 atrial fibrillation; APT 5 antiplatelet therapy; BARC 5 Bleeding Academic Research Consortium; CHA2DS2-VASc 5 congestive heart failure, hypertension, age 75 years (doubled), diabetes, stroke (doubled), vascular disease, age 65 to 74 years, sex category (female); HAS-BLED 5 hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly (. 65 years), drugs (antiplatelet drugs or nonsteroidal antiinflammatory drugs)/alcohol excess concomitantly; HR 5 hazard ratio; INR 5 international normalized ratio; SAMe-TT2R2 5 sex female, age , 60 years, medical history (more than two comorbidities), treatment
journal.publications.chestnet.org
(interacting drugs, eg, amiodarone for rhythm control), tobacco use (doubled), race (doubled); TTR 5 time in therapeutic range; VKA 5 vitamin K antagonist AFFILIATIONS: From the Service de Cardiologie (Drs Angoulvant, Villejoubert, Bejan-Angoulvant, Ivanes, Saint Etienne, and Fauchier), Centre Hospitalier Universitaire Trousseau et Faculté de Médecine, Université François Rabelais, Tours, France; and the University of Birmingham Centre for Cardiovascular Sciences (Dr Lip), City Hospital, Birmingham, England. Drs Lip and Fauchier were the joint senior authors of this manuscript. FUNDING/SUPPORT: The authors have reported to CHEST that no funding was received for this study. 491
Atrial fibrillation (AF) confers a fivefold increased risk of stroke, and 15% to 20% of all strokes are caused by AF. Importantly, mortality, morbidity, and economic burden from stroke complicating AF are particularly high. The European Society of Cardiology guidelines recommend the use of the CHA2DS2-VASc (congestive heart failure, hypertension, age 75 years [doubled], diabetes, stroke [doubled], vascular disease, age 65 to 74 years, and sex category [female]) score to evaluate the individual thromboembolic risk associated with AF and to determine the risk to benefit ratio of antithrombotic therapy.1,2 Until recently, the latter essentially referred to oral anticoagulants (mainly vitamin K antagonists [VKAs]), or antiplatelet therapy (APT), usually aspirin.1,3 The main adverse effect of antithrombotic therapy is bleeding and, as recommended by the European Society of Cardiology guidelines, this risk may be assessed using the HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly [. 65 years], drugs [antiplatelet drugs or nonsteroidal antiinflammatory drugs]/alcohol excess concomitantly) score.1,4
Active smoking is a frequent cardiovascular risk factor usually associated with a higher risk of thrombotic events. Smoking independently influences poor international normalized ratio (INR) control and has been incorporated into a simple score (the SAMe-TT2R2 [sex female, age , 60 years, medical history (more than two comorbidities), treatment (interacting drugs, eg, amiodarone for rhythm control), tobacco use (doubled), race (doubled)] score) in patients with AF to predict the likelihood of good anticoagulation control in a patient initiated on VKAs.5 Moreover, the clinical benefit of clopidogrel in reducing myocardial infarction and stroke in randomized clinical trials has been seen primarily in smokers, with little benefit seen among nonsmokers.6 Smoking status could, therefore, expose patients treated for AF to an excess risk of thromboembolic events linked to “low efficacy of their antithrombotic treatment” and/or an excess risk of bleeding complications associated with a “larger efficacy of their antithrombotic treatment.” In the current study, we investigated whether active smoking would differently influence the risks of stroke and bleeding in patients with AF treated with VKAs or with APT.
Materials and Methods
diabetes or vascular disease, and sex category [female]; two points for a prior stroke or transient ischemic attack or age 75 years)2 and the HAS-BLED bleeding risk score (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile INR, elderly [. 65 years], drugs/alcohol concomitantly) to assess bleeding risk, where a score of 3 indicates high risk.4
We included all patients with a diagnosis of AF, atrial flutter, or both seen in the cardiology department of the Centre Hospitalier Universitaire Trousseau, Tours, France between January 2000 and December 2010. Patient characteristics were obtained from the records of the institution’s computerized codification system for each patient. Extensive information on date of admission, discharge, diagnosis, clinical presentation, comorbidities, medication, and subsequent hospitalization were collected. For each patient, the thromboembolic risk was estimated using the CHA2DS2-VASc score and the hemorrhagic risk using the HAS-BLED bleeding risk score. The SAMe-TT2R2 score was used to predict poor INR control in patients with AF treated with VKAs. The SAMe-TT2R2 score was calculated as the sum of points after the addition of one point each for female sex, age , 60 years, medical history of more than two comorbidities (from among hypertension, diabetes, coronary artery disease/myocardial infarction, peripheral arterial disease, congestive heart failure, previous stroke, pulmonary disease, and hepatic or renal disease), treatment (interacting drugs [eg, amiodarone for rhythm control]), and two points each for tobacco use and nonwhite race. Because recording a patient’s race is not allowed in any electronic file in France, the last item was actually not included in the addition, but the local population was essentially white. Patients with a SAMe-TT2R2 score of 0 to 2 were deemed to have low-moderate risk and those with a score . 2 to have a high risk of poor INR control with a VKA.5 We also calculated the CHA2DS2-VASc score (one point each for a history of heart failure, a history of hypertension, age 65 to 75 years, the presence of CORRESPONDENCE TO: Laurent Fauchier, MD, Service de Cardiologie et Laboratoire d’Electrophysiologie Cardiaque, Centre Hospitalier Universitaire Trousseau, 37044 Tours, France; e-mail: [email protected] © 2015 AMERICAN COLLEGE OF CHEST PHYSICIANS. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details. DOI: 10.1378/chest.14-3006
492 Original Research
Information on stroke, thromboembolism, and bleeding events was recorded whenever documented within the institution, which includes a total of four hospitals covering all specialties, and was obtained by searching in the medical computerized database. Severe bleeding was defined as a decrease of . 5.0 g/dL in the blood hemoglobin level (including the period around the coronary interventional procedure), the need for transfusion of one or more units of blood, the need for corrective surgery, the occurrence of an intracranial or retroperitoneal hemorrhage, or any combination of these events. We also considered major bleedings, using the Bleeding Academic Research Consortium (BARC) bleeding definitions: intracranial hemorrhage, intraocular compromising vision, overt bleeding plus hemoglobin drop of . 5 g/dL, tamponade, bleeding requiring surgical or percutaneous intervention for control (excluding dental/nose/skin/hemorrhoids) or inotropes (BARC type 3A), any transfusion with overt bleeding, overt bleeding plus hemoglobin drop of 3 to 5 g/dL (BARC type 3B), or fatal bleeding.7,8 Finally, we considered separately the GI location of bleeding because smoking is a known risk factor for GI bleeding. The Centre Hospitalier Regional et Universitaire de Tours serves a population of approximately 400,000 and is the only public institution in an area of around 4,000 km2. Patients were followed up for stroke and thromboembolic and bleeding events. In addition, deaths were identified using an online search tool dedicated to local news, covering an area of 35,000 km2 (http://nrco.lanouvellerepublique.fr/dossiers/necro/ index.php).The outcomes in patients with active smoking were compared with those in other patients.9 Statistical Analysis Patients’ characteristics were reported as percentages or means SD. Comparisons between groups were made using x2 tests to compare
[
148#2 CHEST AUGUST 2015
]
categorical variables and the Student t test or nonparametric KruskalWallis test where appropriate for continuous variables. Event rates of stroke/thromboembolism and bleeding were calculated, and multivariate analysis with a proportional hazard model was used to investigate the association between active smoking and each of the outcomes. The results were expressed as relative risk and 95% CI. A P value , .05 was considered statistically significant. Statistical analysis was carried out with Statview 5.0 software (Abacus Concepts).
Ethics Approval The study was approved by the institutional review board of the Pole Coeur Thorax Vaisseaux from the Trousseau University Hospital (Tours, France) on December 7, 2010, and was registered as a clinical audit. Ethical review was, therefore, not required. Patient consent was not sought. Patient data were used only to facilitate the cross-referencing of data sources, and records were otherwise anonymous. The study was conducted retrospectively, patients were not involved in its conduct, and there was no impact on their care.
Results
analysis after adjustment for possible confounders as detailed in Table 1, there was no significant difference in AF between nonsmoker and active smoker patients.
A total of 7,809 patients with AF (62% men; mean age, 71 15 years) were included in this study, which covered the period 2000 to 2010. VKA and APT distribution between smokers and nonsmokers are depicted in Figure 1. Smokers were younger than nonsmokers, and there were more men and comorbidities (hypertension, diabetes, renal insufficiency, coronary artery disease, heart failure, and lower left ventricular ejection fraction) among smokers. The CHA2DS2-VASc score was similar in the two groups, whereas smokers had higher HAS-BLED and SAMe-TT2R2 scores. There was a similar number of patients given VKAs in the smoking group and in the nonsmoking group, whereas there were more patients treated with APT in the smoking group (Table 1). Follow-up: Stroke/Thromboembolism
We recorded 631 strokes/thromboembolic events during a 929 1,082-day follow-up (median, 63 days; interquartile range, 1,564 days). Figure 2 shows stroke or systemic thromboembolism events in patients with AF with or without smoking during follow-up. Smoking was not independently associated with a higher risk of stroke in those patients with AF (relative risk, 0.95; 95% CI, 0.78-1.22; P 5 .66). On multivariate
Figure 1 – Antithrombotic therapy of the patients with AF with or without active smoking. AF 5 atrial fibrillation; APT 5 antiplatelet therapy; VKA 5 vitamin K antagonist.
journal.publications.chestnet.org
Follow-up: Bleeding Events
During follow-up, 707 severe bleedings and 248 major BARC bleedings were observed. Figure 3 shows bleeding events in patients with AF with or without smoking during follow-up. After adjustment for age, smoking status, CHA2DS2-VASc score, HAS-BLED bleeding risk score, use of VKAs, and use of APT, there was an excess risk of severe bleeding in smokers (hazard ratio [HR], 1.23; 95% CI, 1.01-1.49) and patients with VKA use (HR, 1.26; 95% CI, 1.06-1.50). Older patients and those with a higher HAS-BLED bleeding risk score also had an excess risk of severe bleeding. This excess risk of severe bleeding in smokers persisted significantly in patients receiving VKAs and was nonsignificant in patients receiving APT (Table 2). We also performed a multivariate analysis for bleeding events with COPD added to the model because this condition was more prevalent in smokers. COPD was an independent predictor of bleeding events, but the overall findings were similar: Smoking was an independent predictor of bleeding events in patients receiving VKAs (HR, 1.37; 95% CI, 1.08-1.75; P 5 .01 for COPD; HR, 1.26; 95% CI, 1.01-1.60; P 5 .05 for active smoking) and was nonsignificant in patients receiving APT (HR, 1.48; 95% CI, 1.07-2.04; P 5 .02 for COPD; HR, 1.24; 95% CI, 0.92-1.68; P 5 .15 for active smoking). For major BARC bleeding, active smoking and use of VKAs were also associated with a significant excess risk in the multivariate analysis using a model including age, smoking status, HAS-BLED bleeding risk score, CHA2DS2-VASc score, use of VKAs, and use of APT (Table 3). Smoking was not significantly associated with a higher risk of GI bleeding in those patients with AF (relative risk, 1.28; 95% CI, 0.84-1.94; P 5 .25). Finally, smoking was found to be an independent predictor of all-cause death and cardiovascular death in patients with AF (Table 3). 493
TABLE 1
] Characteristics of Patients With AF With or Without Active Smoking
Variable
AF and No Smoking (n 5 6775)
AF and Smoking (n 5 1034)
P Value
Age, y
72 15
68 13
, .0001
Women
43
11
, .0001
Heart failure
54
60
, .0001
Hypertension
40
52
, .0001
Diabetes mellitus
, .0001
14
25
Previous ischemic stroke
8
8
Coronary artery disease
27
51
, .0001
Valve disease
23
23
.84
.42
CHA2DS2-VASc score
3.1 1.8
3.0 1.7
.22
HAS-BLED score
1.5 1.1
1.9 1.2
, .0001
SAMe-TT2R2 score
1.2 0.9
3.2 0.8
, .0001
9
11
.004
9
20
, .0001
39
38
.38
Renal insufficiency COPD Permanent AF Pacemaker or ICD
17
15
.06
47 16
45 15
.02
VKA
57
55
.21
Aspirin
29
40
, .0001
LVEF (n 5 1934) Medication during follow-up
Clopidogrel
9
18
, .0001
VKA 1 single APT
7
11
, .0001
2
4
.0002
ACE inhibitor or ARB
VKA 1 dual APT
39
49
, .0001
b-Blocker
44
46
.37
Diuretic
47
51
.07
Class 3 antiarrhythmic agent
45
44
.54
Data are presented as mean SD or %. ACE 5 angiotensin-converting enzyme; AF 5 atrial fibrillation; APT 5 antiplatelet therapy; ARB 5 angiotensin II receptor blocker; CHA2DS2-VASc 5 congestive heart failure, hypertension, age 75 y (doubled), diabetes, stroke (doubled), vascular disease, age 65 to 74 y, and sex category (female); HAS-BLED 5 hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly (. 65 y), drugs (antiplatelet drugs or nonsteroidal antiinflammatory drugs)/alcohol excess concomitantly; ICD 5 implantable cardioverter defibrillator; LVEF 5 left ventricular ejection fraction; SAMe-TT2R2 5 sex female, age , 60 y, medical history (more than two comorbidities), treatment (interacting drugs, eg, amiodarone for rhythm control), tobacco use (doubled), race (doubled); VKA 5 vitamin K antagonist.
Discussion Our principal findings were that actively smoking patients with AF did not have a higher risk of thromboembolic events but had a significantly higher risk of severe bleeding, particularly so in patients treated with VKAs. To our knowledge, this is one of the first reports from a large series to highlight an excess bleeding risk in smoking patients with AF. Smoking could simply reflect the additional bleeding risk factors but may also more directly predispose to bleeding. Our observations also indirectly confirm the relevance of the HAS-BLED bleeding score, which was 494 Original Research
higher in the smoking group.1,4 When investigating whether this excess bleeding risk among smokers with AF in our cohort was drug related, the increase in bleeding risk was significant in patients taking VKAs, whereas it was not significant in patients treated with APT. Actively smoking patients with AF were significantly younger, with a significantly higher male predominance, and had significantly more comorbidities, with a higher HAS-BLED score. Nonetheless, there was still an independent and significant association between active smoking and bleeding risk on multivariate analysis (Table 2). The benefit of adding smoking to the HAS-BLED scoring formula is debatable because this additional parameter may offer
[
148#2 CHEST AUGUST 2015
]
TABLE 2
] Cox Regression Analysis for Prediction of Severe Bleeding Events
Predictors of Severe Bleeding (707 Events)
Hazard Ratio (95% CI)
P Value
1.02 (1.01-1.02)
, .0001
All patients Age Smoking
1.23 (1.01-1.49)
.04
HAS-BLED score
1.24 (1.13-1.37)
, .0001
CHA2DS2-VASc score
0.99 (0.93-1.07)
.88
VKA at discharge
1.26 (1.06-1.50)
.009
APT at discharge
1.04 (0.87-1.25)
.66
Age
1.01 (1.00-1.02)
.005
Smoking
1.32 (1.04-1.67)
.02
Patients treated with VKA
Figure 2 – Stroke or systemic thromboembolism in patients with AF with or without active smoking (mean 884 1,054 d of follow-up; n 5 631 events). RR 5 relative risk. See Figure 1 legend for expansion of other abbreviation.
marginal improvement in predictive ability at the cost of losing simplicity, ease of use, and everyday practicality. Numerous studies have shown that the rates of stroke or thromboembolic events and major bleeding are lowest in patients with a high average individual time in therapeutic range (TTR). In addition, a recent European consensus recommended an optimal average TTR of . 70% in patients treated with VKAs.10,11 Apostolakis et al5 proposed the SAMe-TT2R2 score, which was useful in predicting those patients with a high TTR . 70% (SAMe-TT2R2 score 0-1) who would do well taking VKAs and those who would do less well (SAMe-TT2R2 score 2). In this score, active smoking is given two points, predicting in itself a smaller TTR. Moreover, we recently showed that the SAMe-TT2R2 score was predictive for an increased risk of stroke or thromboembolic events, severe bleeding, major BARC bleeding, and death, reflecting poor anticoagulation control in patients with
Figure 3 – Bleeding events in patients with AF with or without active smoking (mean 877 1,052 d of follow-up; n 5 707 events). See Figure 1 and 2 legends for expansion of abbreviations.
journal.publications.chestnet.org
HAS-BLED score
1.21 (1.07-1.32)
.002
CHA2DS2-VASc score
1.04 (0.96-1.13)
.36
APT at discharge
1.12 (0.88-1.42)
.34
Age
1.03 (1.02-1.05)
, .0001
Smoking
1.28 (0.94-1.74)
.11
HAS-BLED score
1.28 (1.10-1.49)
.001
CHA2DS2-VASc score
0.88 (0.79-0.99)
.03
VKA at discharge
1.36 (1.05-1.76)
.02
Aspirin
0.88 (0.60-1.30)
.53
Clopidogrel
1.19 (0.88-1.60)
.25
Patients treated with APT
See Table 1 legend for expansion of abbreviations.
AF treated with VKAs.8 The current study may explain some of the results of this previous observation because we found an independent excess bleeding risk in actively smoking patients with AF treated with VKAs.8 Warfarin is eliminated through hepatic metabolism, and some experiments in tissue, animal, and human models have indicated that active smoking is associated with enhanced drug clearance through hepatic microsomal enzyme induction.12,13 A meta-analysis by Nathisuwan et al14 suggested that smoking may cause significant interactions with warfarin by increasing warfarin clearance, which leads to reduced warfarin effects. They found a significant increase in the warfarin dosage requirement in patients who were active smokers when compared with nonsmokers.14,15 Moreover, previous studies demonstrated that chronic smoking cessation was associated with increased INR values.16 This increased INR value after smoking cessation could explain the higher number of hemorrhagic events in active smokers in our cohort. Indeed, one of our hypotheses to explain this excess bleeding risk 495
TABLE 3
] Cox Regression Analysis for Prediction of Major BARC Bleeding Events
Predictors of Outcomes
OR (95% CI)
P Value
1.00 (0.99-1.01)
.96
Major BARC bleeding (248 events) Age Smoking
1.40 (1.02-1.90)
.03
HAS-BLED score
1.40 (1.20-1.64)
, .0001
CHA2DS2-VASc score
0.98 (0.88-1.10)
.77
VKA at discharge
1.72 (1.27-2.31)
.0004
APT at discharge
1.27 (0.95-1.74)
.11
1.05 (1.04-1.06)
, .0001
Smoking
1.39 (1.17-1.64)
.0002
HAS-BLED score
1.10 (1.01-1.20)
.02
CHA2DS2-VASc score
1.12 (1.06-1.19)
.0001
VKA at discharge
0.58 (0.50-0.66)
, .0001
APT at discharge
0.70 (0.60-0.82)
, .0001
1.05 (1.04-1.06)
, .0001
Smoking
1.41 (1.14-1.75)
.0002
HAS-BLED score
1.13 (1.01-1.26)
.03
CHA2DS2-VASc score
1.16 (1.07-1.24)
.0001
VKA at discharge
0.61 (0.51-0.73)
, .0001
APT at discharge
0.74 (0.61-0.90)
.003
All-cause death (989 events) Age
Cardiovascular death (595 events) Age
BARC 5 Bleeding Academic Research Consortium. See Table 1 legend for expansion of other abbreviations.
among smokers in our population of patients with AF treated with VKAs is that smoking induces INR instability, which is known to increase bleeding risk. (Of note, INR instability while taking VKAs was a component of the HAS-BLED risk score.4) Indeed, Nathisuwan et al14 recommended close INR monitoring between 1 week and 3 months after smoking cessation. Unfortunately, smoking cessation was not mentioned in our cohort, so we cannot measure its impact on bleeding risk. On the other hand, previous studies found no significant relationship between smoking and nontherapeutic INR (sub- and supratherapeutic INR).17,18 The pharmacologic interaction between smoking and antithrombotic effect is not linked to the importance of tobacco intoxication, but to active smoking. Other plausible hypotheses may also explain our findings. Active smokers may tend to pay less attention to alcohol, diet, vegetables, herbal products, or other environmental factors, which may interfere 496 Original Research
with warfarin response. Thus, the excess risk may be explained partly by INR instability in active smokers. The Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial analyzed the risk of major bleeding in patients with acute coronary syndrome according to smoking status and found that smokers had a higher risk of bleeding complications.19,20 Tobacco smoking results in increased platelet aggregability, and some authors have suggested that low-dose aspirin cannot sufficiently inhibit the platelet aggregability in smokers with coronary artery disease.21,22 In our work, we found no excess risk of thromboembolic events in smokers treated with aspirin. Of note, Zhao et al23 suggested in a recent metaanalysis that current smokers have significantly lower on-clopidogrel platelet reactivity. This could explain this higher risk of bleeding complications in active smokers treated with clopidogrel. Because of the small number of patients taking clopidogrel in our study population, the influence of this drug on bleeding risk in patients with AF cannot be evaluated. Importantly, whereas smoking may be a risk factor for GI bleeding, the excess bleeding risk in active smokers compared with nonsmokers in our cohort did not result from an increase in GI bleeding. The “smoker’s paradox” refers to the observation of a favorable prognosis in current smokers following an acute myocardial infarction.24 Several hypotheses have been proposed to explain it. Some involve a modification of the response to APT in smokers, especially for clopidogrel.6,23 In the current study, we did not show such a paradox or a higher risk of bleeding complications in patients who were active smokers treated with clopidogrel, because they did not have fewer thromboembolisms or more bleeding events. Indeed, we did not find an increased thromboembolic risk in active smokers in our cohort, consistent with the relevance of the CHA2DS2-VASc score, which was similar in both groups. In a recent study with a smaller number of patients, Nakagawa et al25 found more fatal strokes, higher all-cause mortality, and more intracranial bleeding in patients who smoked, even when controlling for risk factors. However, persistent smoking did not independently predict the risk of stroke, whereas it predicted intracranial bleeding independent of age, and these results are consistent with our findings. Study Limitations
Our work was an observational study, not a randomized trial. Data were obtained retrospectively from a search
[
148#2 CHEST AUGUST 2015
]
in the hospital discharge records, with the limitations of diagnostic coding and case ascertainment. It was also a monocentric study, and our results should be interpreted with caution when applied to the general population. Although we had information on major episodes of labile INR for some patients, data regarding a precise TTR were are not available, which is a limitation for the analysis of bleeding risk. Finally, we did not have data regarding the intensity and regularity of tobacco use or the possibility of smoking cessation during follow-up.
Acknowledgments Author contributions: L. F. is the guarantor of the study. D. A., T. B.-A., and L. F. contributed to the study conception and design; D. A., O. V., and L. F. contributed to the analyses and production of the initial manuscript; L. F. contributed to the data collection; and D. A., O. V., T. B.-A., F. I., C. S. E., G. Y. H. L., and L. F. contributed to the interpretation of the results, critical revision of the manuscript for important intellectual content, and approval of the final manuscript. Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Angoulvant has received funding for conference travel and educational symposia from AstraZeneca plc, Eli Lilly and Company, Novartis, Bayer AG, MSD, Amgen, and Pfizer Inc. Dr Lip has served as a consultant for Bayer AG, Astellas Pharma Inc, Merck & Co Inc, AstraZeneca plc, Sanofi SA, Biotronik SE & CoKG, Bristol-Myers Squibb Co/Pfizer Inc, and Boehringher Ingelheim GmbH and has been on the speakers bureau for Bayer AG, Bristol-Myers Squibb Co/Pfizer Inc, Boehringher Ingelheim GmbH, and Sanofi Aventis LLC. Dr Fauchier has served as a consultant for Bayer AG, Bristol-Myers Squibb Co/Pfizer Inc, Boehringher Ingelheim GmbH, Medtronic Inc, Novartis, and Sanofi Aventis LLC and has received funding for conference travel and educational symposia from Bayer AG, Bristol-Myers Squibb Co/Pfizer Inc, Boehringher Ingelheim GmbH, Boston Scientific, Medtronic Inc, Novartis, Sorin, and Sanofi Aventis LLC. Drs Villejoubert, Bejan-Angoulvant, Ivanes, and Saint Etienne have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
References 1. Camm AJ, Lip GY, De Caterina R, et al; ESC Committee for Practice Guidelines (CPG). 2012 focused update of the ESC guidelines for the management of atrial fibrillation: an update of the 2010 ESC guidelines for the management of atrial fibrillation. Developed with the special contribution of the European Heart Rhythm Association. Eur Heart J. 2012;33(21):2719-2747.
journal.publications.chestnet.org
Conclusions In our large cohort of consecutive patients with AF, smoking was independently associated with a higher risk of severe bleeding, particularly in patients treated with VKAs. However, smoking was not associated with an excess of thromboembolic risk. Whether our observation is driven mainly by pharmacologic interaction(s) between tobacco and VKA and/or by the direct effect of tobacco on bleeding risk remains to be investigated.
2. Olesen JB, Torp-Pedersen C, Hansen ML, Lip GYH. The value of the CHA2DS2VASc score for refining stroke risk stratification in patients with atrial fibrillation with a CHADS2 score 0-1: a nationwide cohort study. Thromb Haemost. 2012; 107(6):1172-1179. 3. Lip GYH, Tse HF, Lane DA. Atrial fibrillation. Lancet. 2012;379(9816):648-661. 4. Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJGM, Lip GYH. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138(5):1093-1100. 5. Apostolakis S, Sullivan RM, Olshansky B, Lip GYH. Factors affecting quality of anticoagulation control among patients with atrial fibrillation on warfarin: the SAMe-TT2R2 score. Chest. 2013;144(5): 1555-1563. 6. Gagne JJ, Bykov K, Choudhry NK, Toomey TJ, Connolly JG, Avorn J. Effect of smoking on comparative efficacy of antiplatelet agents: systematic review, meta-analysis, and indirect comparison. BMJ. 2013;347:f5307. 7. Mehran R, Rao SV, Bhatt DL, et al. Standardized bleeding definitions for cardiovascular clinical trials: a consensus report from the Bleeding Academic Research Consortium. Circulation. 2011; 123(23):2736-2747. 8. Lip GYH, Haguenoer K, Saint-Etienne C, Fauchier L. Relationship of the SAMe-TT2R2 score to poor-quality anticoagulation, stroke, clinically relevant bleeding and mortality in patients with atrial fibrillation. Chest. 2014;146(3): 719-726. 9. Gorin L, Fauchier L, Nonin E, Charbonnier B, Babuty D, Lip GYH. Prognosis and guideline-adherent antithrombotic treatment in patients with atrial fibrillation and atrial flutter: implications of undertreatment and overtreatment in real-life clinical practice; the Loire Valley Atrial Fibrillation Project. Chest. 2011;140(4):911-917. 10. De Caterina R, Husted S, Wallentin L, et al; European Society of Cardiology Working Group on Thrombosis Task Force on Anticoagulants in Heart Disease. General mechanisms of coagulation and targets of anticoagulants (section I). Position paper of the ESC Working Group on Thrombosis—Task Force on
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Anticoagulants in Heart Disease. Thromb Haemost. 2013;109(4):569-579. De Caterina R, Husted S, Wallentin L, et al. Vitamin K antagonists in heart disease: current status and perspectives (section III). Position paper of the ESC Working Group on Thrombosis—Task Force on Anticoagulants in Heart Disease. Thromb Haemost. 2013;110(6): 1087-1107. Holbrook AM, Pereira JA, Labiris R, et al. Systematic overview of warfarin and its drug and food interactions. Arch Intern Med. 2005;165(10):1095-1106. Zevin S, Benowitz NL. Drug interactions with tobacco smoking. An update. Clin Pharmacokinet. 1999;36(6):425-438. Nathisuwan S, Dilokthornsakul P, Chaiyakunapruk N, Morarai T, Yodting T, Piriyachananusorn N. Assessing evidence of interaction between smoking and warfarin: a systematic review and meta-analysis. Chest. 2011;139(5): 1130-1139. Bachmann K, Shapiro R, Fulton R, Carroll FT, Sullivan TJ. Smoking and warfarin disposition. Clin Pharmacol Ther. 1979;25(3):309-315. Colucci VJ, Knapp JF. Increase in international normalized ratio associated with smoking cessation. Ann Pharmacother. 2001;35(3):385-386. McGriff-Lee NJ, Csako G, Chen JT, et al; Anticoagulation Working Group. Search for predictors of nontherapeutic INR results with warfarin therapy. Ann Pharmacother. 2005;39(12): 1996-2002. Pamboukian SV, Nisar I, Patel S, et al. Factors associated with non-adherence to therapy with warfarin in a population of chronic heart failure patients. Clin Cardiol. 2008;31(1):30-34. Berger JS, Bhatt DL, Steinhubl SR, et al; CHARISMA Investigators. Smoking, clopidogrel, and mortality in patients with established cardiovascular disease. Circulation. 2009;120(23):2337-2344. Hirschl MM. Smoking status and the effects of antiplatelet drugs. BMJ. 2013;347:f5909. Li W-J, Zhang H-Y, Miao C-L, et al. Cigarette smoking inhibits the antiplatelet activity of aspirin in patients with coronary heart disease. Chin Med J (Engl). 2011;124(10):1569-1572.
497
22. Ikonomidis I, Lekakis J, Vamvakou G, Andreotti F, Nihoyannopoulos P. Cigarette smoking is associated with increased circulating proinflammatory and procoagulant markers in patients with chronic coronary artery disease: effects of aspirin treatment. Am Heart J. 2005;149(5):832-839.
498 Original Research
23. Zhao Z-G, Chen M, Peng Y, et al. The impact of smoking on clinical efficacy and pharmacodynamic effects of clopidogrel: a systematic review and meta-analysis. Heart. 2014;100(3): 192-199. 24. Aune E, Røislien J, Mathisen M, Thelle DS, Otterstad JE. The “smoker’s paradox”
in patients with acute coronary syndrome: a systematic review. BMC Med. 2011; 9:97. 25. Nakagawa K, Hirai T, Ohara K, et al. Impact of persistent smoking on longterm outcomes in patients with nonvalvular atrial fibrillation. J Cardiol. 2015; 65(5):429-433.
[
148#2 CHEST AUGUST 2015
]
