- Email: [email protected]
Comparison of Frequency and Outcome of Major Gastrointestinal Hemorrhage in Patients With Atrial Fibrillation on Versus Not Receiving Warfarin Therapy (from the ATRIA and ATRIA-CVRN Cohorts)
Comparison of Frequency and Outcome of Major Gastrointestinal Hemorrhage in Patients With Atrial Fibrillation on Versus Not Receiving Warfarin Therapy (from the ATRIA and ATRIA-CVRN Cohorts) Jeffrey M. Ashburner, MPHa,b,*, Alan S. Go, MDc, Kristi Reynolds, PhD, MPHd, Yuchiao Chang, PhDa, Margaret C. Fang, MD, MPHe, Lisa Fredman, PhD, MSPHb, Katie M. Applebaum, ScD, MSPHf,g, and Daniel E. Singer, MDa To date, there have been few studies evaluating outcomes of patients with atrial fibrillation (AF) who have experienced gastrointestinal (GI) hemorrhages. We examined short- and long-term mortality of major GI hemorrhage in patients with AF on and off warfarin in recent clinical care. We evaluated this association in the large Anticoagulation and Risk Factors in Atrial fibrillation (ATRIA) and ATRIA-Cardiovascular Research Network (CVRN) California community-based cohorts of patients with AF (study years 1996 to 2003 and 2006 to 2009, respectively), where all events were clinician adjudicated. We used proportional hazards regression with propensity score adjustment to estimate the short- (30 days) and long-term (>30 days for 1 year) mortality rate ratio for patients using warfarin compared with those who were not using warfarin at the time of GI hemorrhage. In the 414 ATRIA participants with major GI hemorrhage, 54% were taking warfarin at the time of the hemorrhage; in the 361 ATRIA-CVRN participants with major GI hemorrhage, 58% were taking warfarin. Warfarin use at the time of GI hemorrhage was not associated with 30-day mortality in the ATRIA cohort but was associated with significantly reduced 30-day mortality in the ATRIA-CVRN cohort (adjusted mortality rate ratio [95% confidence interval], ATRIA 0.97 [0.54 to 1.74]; ATRIA-CVRN 0.38 [0.17 to 0.83]). There was a modest suggestion of lower mortality on warfarin after 30 days in both cohorts. In conclusion, our study demonstrates that GI hemorrhages on warfarin are certainly no worse and may be less life threatening than those occurring off warfarin. These findings are in stark contrast to the deleterious effect of warfarin on mortality from intracranial hemorrhage and add another factor favoring anticoagulation in clinical decision making for patients with AF. Ó 2014 Elsevier Inc. All rights reserved. (Am J Cardiol 2014;-:-e-) To date, there have been few studies evaluating outcomes of patients with atrial fibrillation (AF) who have experienced gastrointestinal (GI) hemorrhages, the most common extracranial hemorrhage site. Although intracranial hemorrhages are more severe, extracranial hemorrhages are more common and can be fatal.1e3 In warfarin users, intracranial hemorrhages are much more likely to result in death than GI hemorrhages.2 Although warfarin exposure significantly worsens outcomes from intracranial hemorrhage, it is not known whether anticoagulation
similarly affects mortality from GI hemorrhage. This is a particularly relevant question for patients with AF as it may further inform the decision to use warfarin therapy. These considerations apply as well to the novel anticoagulants for which GI hemorrhage remains an important concern.4 The present study aimed to assess the association between warfarin use at the time of GI hemorrhage and the rate of short-term (30 days) and long-term (1 year of follow-up after surviving 30 days) mortality in patients with AF in clinical care. We studied 2 large community-based cohorts
a Division of General Internal Medicine, Massachusetts General Hospital, Boston, Massachusetts; bEpidemiology Department, Boston University School of Public Health, Boston, Massachusetts; cDivision of Research, Kaiser Permanente Northern California, Oakland, California; dDepartment of Research and Evaluation, Kaiser Permanente Southern California, Pasadena, California; eDepartment of Medicine, University of California, San Francisco, San Francisco, California; fDepartment of Environmental and Occupational Health, School of Public Health and Health Services, The George Washington University, Washington, District of Columbia; and g Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts. Manuscript received August 25, 2014; revised manuscript received and accepted October 7, 2014.
See page 7 for disclosure information. This study was supported by the National Institute on Aging (R01 AG15478 and K23 AG028978), the National Heart, Lung, and Blood Institute (U19 HL91179 and RC2HL101589), and the Eliot B. and Edith C. Shoolman fund of the Massachusetts General Hospital (Boston, MA). The funding sources had no role in study design, data collection, data analysis, data interpretation, or preparation of this article. Jeffrey Ashburner, MPH, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. *Corresponding author: Tel: (617) 724-3828; fax: (617) 726-4120. E-mail address: [email protected] (J.M. Ashburner).
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.10.006
www.ajconline.org
2
The American Journal of Cardiology (www.ajconline.org)
with comprehensive follow-up and clinical adjudication of hemorrhages. Methods Patients included in these analyses were members of 2 cohort studies of patients with AF from Kaiser Permanente (KP) Northern and Southern California, 2 large integrated health care delivery systems. The Anticoagulation and Risk Factors in Atrial fibrillation (ATRIA) cohort includes 13,559 adults aged 18 years and older with diagnosed nonvalvular AF who received care with KP Northern California.5 The ATRIA-Cardiovascular Research Network (CVRN) cohort includes 33,247 adults aged 18 years and older with incident nonvalvular AF who were enrolled in either KP Northern or KP Southern California. In each sample, cohort members were identified by searching electronic inpatient, outpatient, and electrocardiographic databases for physician-assigned International Classification of Diseases, Ninth Revision, Clinical Modification, diagnostic codes of AF (427.31, 427.32). Patients from the ATRIA cohort were identified between July 1996 and December 1997 and followed through September 2003, and patients from the ATRIA-CVRN cohort were identified between January 2006 and June 2009, with follow-up through June 2009. Because we were interested in nontransient nonvalvular AF, we excluded patients with AF with diagnosed mitral stenosis (ATRIA only), valvular repair or replacement (ATRIA only), transient postoperative AF, or concurrent hyperthyroidism. Unlike the ATRIA cohort, the ATRIA-CVRN cohort did not exclude patients with mitral stenosis or a history of a valve replacement; such patients accounted for only 1.5% of the ATRIA-CVRN cohort. We focused on patients who were hospitalized for major GI hemorrhage during the observation period of the full ATRIA and ATRIA-CVRN studies. GI hemorrhages were identified by searching comprehensive hospitalization and billing databases for primary discharge diagnoses (Supplementary Material A). Major GI hemorrhages were defined as fatal (death during hospitalization) or requiring transfusion of 2 or more units of packed red blood cells. GI hemorrhages not leading to hospitalization were excluded. Because KP is an insurer and a health care provider, cohort members admitted to facilities outside the network were still identified by our search strategy. Each potential GI hemorrhagic event was reviewed by an Outcomes Review Committee composed of physicians. This resulted in identification of 712 patients in the ATRIA cohort who experienced a hospitalization for a validated GI hemorrhage, with 414 (58%) classified as major hemorrhages. In the ATRIA-CVRN cohort, 601 patients experienced a validated GI hemorrhage, with 361 (60%) classified as major hemorrhages. Warfarin use at the time of the GI hemorrhagic event was assessed through manual review of the admission medical record. Patients were considered to be using warfarin if the admission medical record indicated that they were taking warfarin within 5 days before admission for GI hemorrhage. We recorded the international normalized ratio (INR) value at admission for each patient before reversal of warfarin effect. For the ATRIA cohort, we recorded the number of units of blood that were transfused and whether fresh frozen
plasma (FFP) was used. For the ATRIA-CVRN cohort, we recorded whether the number of transfused units of blood was 2 and did not record use of FFP. Deaths were determined through reviewing medical charts, health plan databases, Social Security Administration vital status file, and the comprehensive California State death certificate registry. Patient characteristics were obtained from administrative databases. Low educational attainment status (
Arrhythmias and Conduction Disturbances/Death Following Gastrointestinal Hemorrhage On and Off Warfarin
3
Table 1 Baseline patient characteristics by use of warfarin at the time of gastrointestinal hemorrhage: among 414 patients from the ATRIA study and 361 patients from the ATRIA-CVRN study ATRIA
Variable Age (years), mean (SD) Asian/Pacific Islander Black Hispanic ethnicity Other/Unknown White Women Low educational attainment Low annual household income Current Smoker Unknown History of mechanical falls Prior GI hemorrhage Prior ischemic stroke Diabetes mellitus Coronary disease Hypertension Chronic heart failure Dementia Chronic liver disease Kidney dysfunction (eGFR < 30 ml/min/1.73 m2) Cancer† Anemiaz Lung disease Transient ischemic attack Medications at Admission Aspirin Other antiplatelet therapies Lipid-lowering agents ACE inhibitors Angiotensin II receptor blockers Stroke/Bleeding Risk CHADS2, mean (SD) CHA2DS2-VASc, mean (SD) ATRIA stroke risk score, mean (SD) HAS-BLED, mean (SD)* ATRIA hemorrhage risk score, mean (SD)
Warfarin use at GI hemorrhage (n ¼ 223)
No Warfarin at GI hemorrhage (n ¼ 191)
77.4 (7.1) 5.8% 5.4% 7.2% 1.4% 80.3% 43.5% 5.8% 9.9% 4.9% 14.8% 10.8% 13.9% 15.7% 30.5% 45.3% 66.8% 57.4% 7.6% 4.0% 21.1% 22.0% 47.1% n/a n/a
79.9 (9.3) 6.3% 7.9% 6.8% 0.5% 78.5% 45.0% 3.7% 7.9% 3.1% 14.1% 11.5% 20.9% 12.0% 31.4% 43.5% 70.7% 53.9% 13.1% 4.2% 21.5% 22.5% 49.2% n/a n/a
13.9% 0.0% 12.1% 44.4% 4.0%
56.5% 1.6% 6.3% 35.1% 4.7%
2.6 4.4 7.6 2.3 4.3
(1.1) (1.4) (2.3) (1.0) (2.4)
2.6 4.5 7.7 2.8 4.6
(1.2) (1.6) (2.5) (1.0) (2.3)
ATRIA-CVRN P-Value
0.002 0.78
0.75 0.31 0.47 0.63 0.81 0.06 0.30 0.84 0.71 0.40 0.48 0.07 0.94 0.92 0.90 0.67
<0.001 0.10 0.04 0.05 0.74 0.93 0.89 0.67 <0.001 0.12
Warfarin use at GI hemorrhage (n ¼ 210)
No Warfarin at GI hemorrhage (n ¼ 151)
P-Value
76.0 (8.0) 9.5% 8.1% 15.7% 1.0% 65.7% 40.5% 27.1% 15.7% 4.8% 6.2% 12.4% 14.8% 9.5% 41.4% 42.4% 91.4% 48.6% 3.3% 3.3% 12.4% 22.9% 55.2% 47.6% 7.1%
77.9 (9.1) 9.3% 13.3% 9.9% 0.7% 66.9% 48.3% 29.1% 13.9% 7.3% 5.3% 21.2% 20.5% 8.0% 45.0% 39.1% 94.7% 49.7% 7.3% 5.3% 13.9% 23.2% 65.6% 44.4% 6.0%
0.03 0.33
0.02 0.15 0.60 0.49 0.52 0.24 0.84 0.09 0.36 0.67 0.94 0.05 0.54 0.66
38.6% 10.5% 63.3% 38.6% 11.0%
63.6% 12.6% 51.0% 40.4% 10.6%
<0.001 0.53 0.02 0.73 0.91
2.6 4.4 7.1 2.6 4.4
(2.5) (1.5) (2.3) (0.9) (2.0)
2.7 4.6 7.5 3.0 5.0
(2.5) (1.6) (2.4) (0.9) (2.1)
0.14 0.68 0.63 0.57
0.46 0.28 0.07 <0.001 0.01
* The HAS-BLED score presented here has a maximum of 7 points since data on alcohol usage was not available and time in therapeutic range was not included. If adding 1 point for time in therapeutic range < 60%, there is no significant difference between those using and not using warfarin. † All cancer excluding non-melanomatous skin cancer. z Female: hemoglobin < 13.5 g/L; Male: hemoglobin < 12.0 g/L.
therefore, decided to examine the cohorts separately rather than as a combined data set (Supplementary Material B). We conducted descriptive and stratified analyses to examine the distribution of potential confounders in those using and not using warfarin. Cox proportional hazards regression was used to estimate the crude and adjusted mortality rate ratio (mRR) for the association between warfarin use at the time of GI hemorrhage and mortality (SAS, version 9.2; SAS Institute, Cary, North Carolina). Additionally, we examined models stratified by age group (<85 and 85 years). To control for baseline confounding, we used Cox proportional hazard regression models adjusted for propensity
score. Propensity scores were generated from logistic regression models to estimate the predicted probability of using warfarin at the time of GI hemorrhage versus not using warfarin.13 Propensity scores were generated for 4 separate analyses: (1) ATRIA sample: 30-day mortality, (2) ATRIA-CVRN sample: 30-day mortality, (3) ATRIA sample: long-term mortality, and (4) ATRIA-CVRN sample: long-term mortality. For each of these 4 analyses, we fit a preliminary proportional hazards model predicting death that included all potential confounders to determine the strength of the association of each covariable with either short- or long-term
4
The American Journal of Cardiology (www.ajconline.org)
Table 2 Association between warfarin usage at the time of major gastrointestinal hemorrhage and mortality in the ATRIA and ATRIA-CVRN cohorts Variable
ATRIA Warfarin use at GI hemorrhage
30-day Mortality Deaths Person-years Rate Ratio (unadjusted) Rate Ratio (adjusted)* Long-term mortality (>30 days following hospital admission for hemorrhage through 1-year following admission) Deaths Person-years Rate Ratio (unadjusted) Rate Ratio (adjusted)†
(n ¼ 223) 25 (11%) 16.8 0.84 (0.48-1.46) 0.97 (0.54-1.74) (n ¼ 194)
46 (24%) 153.2 0.66 (0.44-0.97) 0.70 (0.47-1.05)
ATRIA-CVRN No Warfarin at GI hemorrhage
Warfarin use at GI hemorrhage
No Warfarin at GI hemorrhage
(n ¼ 191) 25 (13%) 14.0 -
(n ¼ 210) 10 (5%) 16.3 0.33 (0.16-0.70) 0.37 (0.17-0.81) (n ¼ 176)
(n ¼ 151) 21 (14%) 11.2 -
(n ¼ 160)
55 (34%) 119.3 -
23 (13%) 117.8 0.90 (0.48-1.68) 0.83 (0.42-1.66)
(n ¼ 128)
17 (13%) 75.7 -
* Variables included in propensity score for short-term analyses, ATRIA: age, race/ethnicity, low educational attainment, low annual household income, smoking status, history of falls, prior ischemic stroke, diabetes mellitus, coronary disease, hypertension, chronic heart failure, dementia, kidney dysfunction, cancer, use of other antiplatelet therapies, lipid-lowering agents, and ACE inhibitors; ATRIA-CVRN: age, race/ethnicity, gender, low educational attainment, low annual household income, smoking status, history of GI hemorrhage, prior ischemic stroke, diabetes mellitus, coronary disease, hypertension, chronic heart failure, dementia, chronic liver disease, kidney dysfunction, anemia, lung disease, transient ischemic attack, use of antiplatelets, lipid-lowering agents, ACE inhibitors, angiotensin II receptor blockers. † Variables included in propensity score for long-term analyses, ATRIA: age, race, gender, low educational attainment, smoking status, history of coronary disease, chronic heart failure, dementia, kidney dysfunction, cancer, anemia, use of other antiplatelet therapies, lipid-lowering agents, ACE inhibitors, and angiotensin II receptor blockers; ATRIA-CVRN: age, race/ethnicity, gender, low educational attainment, smoking status, history of falls, prior GI hemorrhage, ischemic stroke, diabetes mellitus, coronary disease, hypertension, chronic heart failure, dementia, chronic liver disease, kidney dysfunction, cancer, anemia, lung disease, use of aspirin, antiplatelets, lipid-lowering agents, ACE inhibitors, and angiotensin II receptor blockers.
mortality. To avoid including nonconfounding predictors of warfarin use in the propensity score model, only those variables with a hazard ratio (or 1/hazard ratio) of at least 1.2 from the preliminary outcome model were used to create the propensity scores. After generating the propensity scores, we trimmed 2.5% of patients on both extremes of the propensity score distribution to improve comparability of the warfarinuse and nonewarfarin-use groups. These observations were outside the primary area of overlap of the propensity scores and increase residual confounding.14 The Cox proportional hazards models were adjusted for propensity score quintile (Supplementary Material C). We also performed additional analyses where the propensity score stratification was more finely divided (8 and 10 strata, respectively). Results In the 414 patients in ATRIA with major GI hemorrhages, 54% were taking warfarin. Major GI hemorrhages included 205 (49%) upper GI bleeds, 94 (23%) lower GI bleeds, and 115 (28%) where the location was undetermined by the time of hospital discharge. At the time of GI hemorrhage, patients using warfarin were younger, much less likely to be using aspirin, less likely to have diagnosed dementia or a history of GI hemorrhage, and were more likely to have a history of ischemic stroke and to be using lipid-lowering agents and angiotensin-converting enzyme inhibitors, compared with patients not using warfarin. There was no statistically
significant difference in CHADS2, CHA2DS2-VASc, or ATRIA stroke risk scores between patients using or not using warfarin at the time of their bleeding event, whereas warfarin users had clearly lower HAS-BLED and ATRIA hemorrhage risk scores (Table 1). Characteristics of warfarin users and nonusers within propensity score quintiles were quite similar (Supplementary Material C). Overall, 398 (96%) patients were transfused with a median number of 3.0 (interquartile range [IQR]) units of blood. In warfarin users, the median INR at presentation was 3.0 (IQR). FFP was used in 56% of warfarin users. In the 361 patients in ATRIA-CVRN with major GI hemorrhages, 58% were taking warfarin. The distribution of anatomical sites was very similar to that of the ATRIA cohort with 176 (49%) upper GI bleeds, 99 (27%) lower GI bleeds, and 86 (24%) undetermined. Similar to those in the ATRIA cohort, patients with GI hemorrhage in the ATRIA-CVRN using warfarin were younger, less likely to be women, less likely to have a history of falls, GI hemorrhage, or dementia, and much less likely to be using aspirin. In the more recent ATRIA-CVRN cohort, the use of lipid-lowering agents was far more common and patients with GI bleed on warfarin were prescribed such agents more frequently than those not taking warfarin. Patients with GI bleed on warfarin were less likely to have preexisting anemia. As with the ATRIA experience, there was no statistically significant difference in CHADS2, CHA2DS2-VASc, or ATRIA stroke scores between patients using or not using warfarin at the time of their
Arrhythmias and Conduction Disturbances/Death Following Gastrointestinal Hemorrhage On and Off Warfarin
5
Figure 1. (A) Adjusted survival function for the association between warfarin and long-term mortality in the ATRIA cohort (1996 to 2003). The y-axis (survival probability) extends from 0.5 to 1.0. (B) Adjusted survival function for the association between warfarin and long-term mortality in the ATRIA-CVRN cohort (2006 to 2009). The y-axis (survival probability) extends from 0.5 to 1.
bleeding event, whereas warfarin users had lower HASBLED and ATRIA hemorrhage risk scores (Table 1). Comparison of the characteristics of warfarin users and nonusers in each of the 5 propensity score quintiles showed similar distributions for covariates included in the propensity score model (Supplementary Material C). Overall, 357 (99%) patients were transfused with 2.0 (IQR) units of blood. In warfarin users, the median INR at presentation was 3.2 (IQR).
In the ATRIA cohort, warfarin use at the time of GI hemorrhage was not associated with 30-day mortality in unadjusted or adjusted analyses. In contrast, in ATRIACVRN, warfarin use at the time of GI hemorrhage was associated with a much lower 30-day mortality rate (Table 2). Warfarin use appeared to be more strongly associated with lower 30-day mortality in the oldest group of patients. In the ATRIA cohort, the adjusted mRR (95% confidence interval) was 0.75 (0.28 to 2.03) for patients
6
The American Journal of Cardiology (www.ajconline.org)
85 years (n ¼ 94) versus 1.16 (0.56 to 2.41) for those <85 years (n ¼ 320). In ATRIA-CVRN, warfarin appeared to be associated with reduced mortality in all ages, but, as in ATRIA, the largest observed association again was seen in older patients (adjusted mRR [95% confidence interval] for 85 years [n ¼ 62] 0.15 [0.02 to 1.27]; for <85 years [n ¼ 299] 0.50 [0.20 to 1.23]). A substantial number of patients resumed using warfarin in the year after GI hemorrhage (ATRIA 25%; ATRIA-CVRN 52%), whereas few nonwarfarin users began using warfarin during this time period (ATRIA 1%; ATRIA-CVRN 6%). After excluding deaths and follow-up time during the first 30 days after major GI hemorrhage, warfarin use was associated with a small, but not statistically significant, reduction in mortality over 1 year of follow-up after GI hemorrhage in both ATRIA and ATRIA-CVRN (Table 2 and Figure 1). Discussion We compared the mortality rates of patients sustaining a major GI hemorrhage on- versus off-warfarin therapy in 2 very large community-based cohorts of patients with AF. Patients using warfarin at the time of the GI hemorrhage had a reduced 30-day mortality rate in the more contemporary ATRIA-CVRN cohort. In the older ATRIA cohort, the 30day mortality rate was essentially the same regardless of whether the patient had been using warfarin or not. When stratified by age, warfarin use was associated with better outcome in the oldest patients (85 years) in both cohorts, although the age-stratified results were conducted in small groups and were not statistically significant. For long-term mortality, the data from both cohorts were consistent with a reduced mortality rate in patients whose GI hemorrhage occurred while using warfarin, although the confidence intervals for the mRRs included the null value of 1.0. Our finding that warfarin users had the same or lower 30day mortality risk after major GI hemorrhage clearly contrasts with warfarin’s effect on intracranial hemorrhage where anticoagulation markedly worsens outcome.15,16 In previous ATRIA analyses, warfarin use at the time of intracranial hemorrhage was independently associated with a 62% increased odds of 30-day mortality. The lessened mortality risk with GI hemorrhage may be the result of less severe lesions bleeding on warfarin and/or reversibility of warfarin’s hemorrhagic effect.17 In contrast, patients who experience intracranial hemorrhages while on warfarin often face grim outcomes before the effect of warfarin can be reversed.18 Part of the observed mortality effect in GI hemorrhage may be related to residual confounding by underlying co-morbidity despite our propensity score analysis. Indeed, the lower long-term post 30-day mortality we observed in patients taking warfarin at the time of GI bleed is consistent with preferential use of warfarin in healthier patients because the impact of a resolved GI hemorrhage must be small by 30 days. More generally, the anticoagulation decision for patients with AF depends on the expected net clinical benefit of warfarin, which summarizes the difference in thromboembolic events (primarily ischemic stroke) prevented and hemorrhagic events induced.7e11,19 Net clinical benefit depends on both the rates of events on and off warfarin and the
severity of the different types of events. Extracranial hemorrhages are included in some,20,21 but not all,22 proposed calculations of net clinical benefit. The current analyses, and our previous work on intracranial hemorrhages,15 make clear that there is a different distribution of severity for both thromboembolic and hemorrhagic events depending on whether the patient was taking warfarin at the time of the event. Our current work demonstrates that GI hemorrhages on warfarin are certainly no worse and may be less life threatening than those occurring off warfarin. Although confidence intervals were wide, reduced GI hemorrhage mortality appeared greatest for the oldest patients, a group at the highest risk of hemorrhage and least likely to be prescribed warfarin.22e26 In all, our results add to the expected net clinical benefit of warfarin, particularly in older patients with AF. This study was strengthened by including a large number of validated major GI hemorrhages from 2 separate large cohorts from KP Northern and Southern California. These cohorts were community based with comprehensive followup, with the result that the findings are likely to be broadly generalizable. All events were adjudicated by physicians or nurses using a standardized medical record review protocol. Such an approach minimizes error in case ascertainment and determination of warfarin exposure, a problem faced by studies using only administrative data. Additionally, we used propensity score modeling tailored to each analysis to control for confounding. This approach ensured that comparisons were made between patients who had similar distributions of measured confounders, thereby reducing confounding.14 We observed a much larger association between warfarin and reduced short-term mortality in the more contemporary ATRIA-CVRN cohort. This difference was mainly because of a lower mortality rate in warfarin users in this cohort, whereas mortality rates were similar in nonwarfarin users in the 2 cohorts. It is not clear why the GI hemorrhage mortality rate was lower in warfarin users in the ATRIA-CVRN cohort, although part of the difference may be explained by the play of chance. Even with these differences across cohorts, our results make clear, at the least, that GI hemorrhages occurring in patients taking warfarin pose no increase in mortality risk compared with GI hemorrhages occurring in patients not taking warfarin. Our results are indirectly applicable to the use of novel anticoagulants for AF because GI hemorrhages remain a major concern with these agents.4 There are very limited data on outcomes of GI hemorrhages comparing patients on novel anticoagulants versus patients not taking anticoagulants. It does appear that the case fatality rate for extracranial hemorrhages on novel agents is the same or lower than for extracranial hemorrhages on warfarin.27 As such, the reassuring results of our analysis should apply to GI hemorrhages on novel agents. This study also has several potential limitations. It is possible that there is residual confounding by indication because of underlying co-morbidity, that is, patients not receiving warfarin may have been “sicker.” These patients may be at particularly high mortality risk and may not be anticoagulated for this reason. We aggressively addressed confounding by propensity score approaches. Furthermore, we also included terms in our regression models for
Arrhythmias and Conduction Disturbances/Death Following Gastrointestinal Hemorrhage On and Off Warfarin
variables that were not balanced (p <0.20) in at least 4 quintiles of the propensity score, and the mRRs were similar to models only adjusted for propensity score quintile. We also performed additional analyses where the propensity score stratification was more finely divided (8 and 10 strata), and the results were similar (data not shown). Despite the advantages of propensity score methods, they do not account for unmeasured confounders. It is likely that residual confounding persists in our comparison of warfarin users versus nonusers. However, it is unlikely that any residual confounding is large enough to reverse the association observed for short-term mortality in the ATRIA-CVRN cohort. Disclosures Dr. Singer serves as a consultant/advisory board member for Boehringer Ingelheim, Bristol-Myers Squibb, and Johnson and Johnson related to novel anticoagulants to prevent stroke in AF and with CSL Behring related to reversal of warfarin anticoagulation. Additionally, Dr. Singer has research contracts with Medtronic, Inc., related to AF and risk of stroke and with Bristol-Myers Squibb related to AF and risk of stroke. Dr. Go has received a research grant from CSL Behring related to reversal of warfarin anticoagulation. No other authors have relation with industry. Supplementary Data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. amjcard.2014.10.006. 1. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomized controlled trials. Arch Intern Med 1994;154:1449e1457. 2. Fang MC, Go AS, Chang Y, Hylek EM, Henault LE, Jensvold NG, Singer DE. Death and disability from warfarin-associated intracranial and extracranial hemorrhages. Am J Med 2007;120:700e705. 3. Linkins LA, Choi PT, Douketis JD. Clinical impact of bleeding in patients taking oral anticoagulant therapy for venous thromboembolism: a meta-analysis. Ann Intern Med 2003;139:893e900. 4. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitz MD, Camm AJ, Weitz JI, Lewis BS, Parkhomenko A, Yamashita T, Antman EM. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials. Lancet 2014;383:955e962. 5. Go AS, Hylek EM, Chang Y, Phillips KA, Henault LE, Capra AM, Jensvold NG, Selby JV, Singer DE. Anticoagulation therapy for stroke prevention in atrial fibrillation: how well do randomized trials translate into clinical practice? JAMA 2003;290:2685e2692. 6. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD EPI. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604e612. 7. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864e2870. 8. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. 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 2010;137:263e272. 9. Singer DE, Chang Y, Borowsky LH, Fang MC, Pomernacki NK, Udaltsova N, Reynolds K, Go AS. A new risk scheme to predict ischemic stroke and other thromboembolism in atrial fibrillation: the ATRIA study stroke risk score. J Am Heart Assoc 2013;2:e000250.
7
10. Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. 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;138:1093e1100. 11. Fang MC, Go AS, Chang Y, Borowsky LH, Pomernacki NK, Udaltsova N, Singer DE. A new risk scheme to predict warfarinassociated hemorrhage: the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) Study. J Am Coll Cardiol 2011;58:395e401. 12. Suissa S. Immortal time bias in observational studies of drug effects. Pharmacoepidemiol Drug Saf 2007;16:241e249. 13. Rosenbaum P, Rubin D. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41e55. 14. Sturmer T, Rothman KJ, Avorn J, Glynn RJ. Treatment effects in the presence of unmeasured confounding: dealing with observations in the tails of the propensity score distribution—a simulation study. Am J Epidemiol 2010;172:843e854. 15. Fang MC, Go AS, Chang Y, Borowsky LH, Pomernacki NK, Udaltsova N, Singer DE. Thirty-day mortality after ischemic stroke and intracranial hemorrhage in patients with atrial fibrillation on and off anticoagulants. Stroke 2012;43:1795e1799. 16. Rosand J, Eckman MH, Knudsen KA, Singer DE, Greenberg SM. The effect of warfarin and intensity of anticoagulation on outcome of intracerebral hemorrhage. Arch Intern Med 2004;164:880e884. 17. Schulman S, Beyth RJ, Kearon C, Levine MN; American College of Chest Physicians. Hemorrhagic complications of anticoagulant and thrombolytic treatment: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133:257Se298S. 18. Goldstein JN, Thomas SH, Frontiero V, Joseph A, Engel C, Snider R, Smith EE, Greenberg SM, Rosand J. Timing of fresh frozen plasma administration and rapid correction of coagulopathy in warfarin-related intracerebral hemorrhage. Stroke 2006;37:151e155. 19. Gage BF, Yan Y, Milligan PE, Waterman AD, Culverhouse R, Rich MW, Radford MJ. Clinical classification schemes for predicting hemorrhage: results from the National Registry of Atrial Fibrillation (NRAF). Am Heart J 2006;151:713e719. 20. Connolly SJ, Eikelboom JW, Ng J, Hirsh J, Yusuf S, Pogue J, de Caterina R, Hohnloser S, Hart RG; ACTIVE (Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events) Steering Committee and Investigators. Net clinical benefit of adding clopidogrel to aspirin therapy in patients with atrial fibrillation for whom vitamin K antagonists are unsuitable. Ann Intern Med 2011;155: 579e586. 21. Eikelboom JW, Connolly SJ, Hart RG, Wallentin L, Reilly P, Oldgren J, Yang S, Yusuf S. Balancing the benefits and risks of 2 doses of dabigatran compared with warfarin in atrial fibrillation. J Am Coll Cardiol 2013;62:900e908. 22. Singer DE, Chang Y, Fang MC, Borowsky LH, Pomernacki NK, Udaltsova N, Go AS. The net clinical benefit of warfarin anticoagulation in atrial fibrillation. Ann Intern Med 2009;151:297e305. 23. Bungard TJ, Ghali WA, McAlister FA, Buchan AM, Cave AJ, Hamilton PG, Mitchell LB, Shuaib A, Teo KK, Tsuyuki RT. The relative importance of barriers to the prescription of warfarin for nonvalvular atrial fibrillation. Can J Cardiol 2003;19:280e284. 24. Casciano JP, Singer DE, Kwong WJ, Fox ES, Martin BC. Anticoagulation therapy for patients with non-valvular atrial fibrillation: comparison of decision analytic model recommendations and realworld warfarin prescription use. Am J Cardiovasc Drugs 2012;12: 313e323. 25. Fang MC, Chang Y, Hylek EM, Rosand J, Greenberg SM, Go AS, Singer DE. Advanced age, anticoagulation intensity, and risk for intracranial hemorrhage among patients taking warfarin for atrial fibrillation. Ann Intern Med 2004;141:745e752. 26. Gage BF, Boechler M, Doggette AL, Fortune G, Flaker GC, Rich MW, Radford MJ. Adverse outcomes and predictors of underuse of antithrombotic therapy in Medicare beneficiaries with chronic atrial fibrillation. Stroke 2000;31:822e827. 27. Hylek EM, Held C, Alexander JH, Lopes RD, De Caterina R, Wojdyla DM, Huber K, Jansky P, Steg PG, Hanna M, Thomas L, Wallentin L, Granger CB. Major bleeding in patients with atrial fibrillation receiving apixaban or warfarin: the ARISTOTLE Trial (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation): predictors, characteristics, and clinical outcomes. J Am Coll Cardiol 2014;63:2141e2147.
similarly affects mortality from GI hemorrhage. This is a particularly relevant question for patients with AF as it may further inform the decision to use warfarin therapy. These considerations apply as well to the novel anticoagulants for which GI hemorrhage remains an important concern.4 The present study aimed to assess the association between warfarin use at the time of GI hemorrhage and the rate of short-term (30 days) and long-term (1 year of follow-up after surviving 30 days) mortality in patients with AF in clinical care. We studied 2 large community-based cohorts
a Division of General Internal Medicine, Massachusetts General Hospital, Boston, Massachusetts; bEpidemiology Department, Boston University School of Public Health, Boston, Massachusetts; cDivision of Research, Kaiser Permanente Northern California, Oakland, California; dDepartment of Research and Evaluation, Kaiser Permanente Southern California, Pasadena, California; eDepartment of Medicine, University of California, San Francisco, San Francisco, California; fDepartment of Environmental and Occupational Health, School of Public Health and Health Services, The George Washington University, Washington, District of Columbia; and g Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts. Manuscript received August 25, 2014; revised manuscript received and accepted October 7, 2014.
See page 7 for disclosure information. This study was supported by the National Institute on Aging (R01 AG15478 and K23 AG028978), the National Heart, Lung, and Blood Institute (U19 HL91179 and RC2HL101589), and the Eliot B. and Edith C. Shoolman fund of the Massachusetts General Hospital (Boston, MA). The funding sources had no role in study design, data collection, data analysis, data interpretation, or preparation of this article. Jeffrey Ashburner, MPH, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. *Corresponding author: Tel: (617) 724-3828; fax: (617) 726-4120. E-mail address: [email protected] (J.M. Ashburner).
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.10.006
www.ajconline.org
2
The American Journal of Cardiology (www.ajconline.org)
with comprehensive follow-up and clinical adjudication of hemorrhages. Methods Patients included in these analyses were members of 2 cohort studies of patients with AF from Kaiser Permanente (KP) Northern and Southern California, 2 large integrated health care delivery systems. The Anticoagulation and Risk Factors in Atrial fibrillation (ATRIA) cohort includes 13,559 adults aged 18 years and older with diagnosed nonvalvular AF who received care with KP Northern California.5 The ATRIA-Cardiovascular Research Network (CVRN) cohort includes 33,247 adults aged 18 years and older with incident nonvalvular AF who were enrolled in either KP Northern or KP Southern California. In each sample, cohort members were identified by searching electronic inpatient, outpatient, and electrocardiographic databases for physician-assigned International Classification of Diseases, Ninth Revision, Clinical Modification, diagnostic codes of AF (427.31, 427.32). Patients from the ATRIA cohort were identified between July 1996 and December 1997 and followed through September 2003, and patients from the ATRIA-CVRN cohort were identified between January 2006 and June 2009, with follow-up through June 2009. Because we were interested in nontransient nonvalvular AF, we excluded patients with AF with diagnosed mitral stenosis (ATRIA only), valvular repair or replacement (ATRIA only), transient postoperative AF, or concurrent hyperthyroidism. Unlike the ATRIA cohort, the ATRIA-CVRN cohort did not exclude patients with mitral stenosis or a history of a valve replacement; such patients accounted for only 1.5% of the ATRIA-CVRN cohort. We focused on patients who were hospitalized for major GI hemorrhage during the observation period of the full ATRIA and ATRIA-CVRN studies. GI hemorrhages were identified by searching comprehensive hospitalization and billing databases for primary discharge diagnoses (Supplementary Material A). Major GI hemorrhages were defined as fatal (death during hospitalization) or requiring transfusion of 2 or more units of packed red blood cells. GI hemorrhages not leading to hospitalization were excluded. Because KP is an insurer and a health care provider, cohort members admitted to facilities outside the network were still identified by our search strategy. Each potential GI hemorrhagic event was reviewed by an Outcomes Review Committee composed of physicians. This resulted in identification of 712 patients in the ATRIA cohort who experienced a hospitalization for a validated GI hemorrhage, with 414 (58%) classified as major hemorrhages. In the ATRIA-CVRN cohort, 601 patients experienced a validated GI hemorrhage, with 361 (60%) classified as major hemorrhages. Warfarin use at the time of the GI hemorrhagic event was assessed through manual review of the admission medical record. Patients were considered to be using warfarin if the admission medical record indicated that they were taking warfarin within 5 days before admission for GI hemorrhage. We recorded the international normalized ratio (INR) value at admission for each patient before reversal of warfarin effect. For the ATRIA cohort, we recorded the number of units of blood that were transfused and whether fresh frozen
plasma (FFP) was used. For the ATRIA-CVRN cohort, we recorded whether the number of transfused units of blood was 2 and did not record use of FFP. Deaths were determined through reviewing medical charts, health plan databases, Social Security Administration vital status file, and the comprehensive California State death certificate registry. Patient characteristics were obtained from administrative databases. Low educational attainment status (
Arrhythmias and Conduction Disturbances/Death Following Gastrointestinal Hemorrhage On and Off Warfarin
3
Table 1 Baseline patient characteristics by use of warfarin at the time of gastrointestinal hemorrhage: among 414 patients from the ATRIA study and 361 patients from the ATRIA-CVRN study ATRIA
Variable Age (years), mean (SD) Asian/Pacific Islander Black Hispanic ethnicity Other/Unknown White Women Low educational attainment Low annual household income Current Smoker Unknown History of mechanical falls Prior GI hemorrhage Prior ischemic stroke Diabetes mellitus Coronary disease Hypertension Chronic heart failure Dementia Chronic liver disease Kidney dysfunction (eGFR < 30 ml/min/1.73 m2) Cancer† Anemiaz Lung disease Transient ischemic attack Medications at Admission Aspirin Other antiplatelet therapies Lipid-lowering agents ACE inhibitors Angiotensin II receptor blockers Stroke/Bleeding Risk CHADS2, mean (SD) CHA2DS2-VASc, mean (SD) ATRIA stroke risk score, mean (SD) HAS-BLED, mean (SD)* ATRIA hemorrhage risk score, mean (SD)
Warfarin use at GI hemorrhage (n ¼ 223)
No Warfarin at GI hemorrhage (n ¼ 191)
77.4 (7.1) 5.8% 5.4% 7.2% 1.4% 80.3% 43.5% 5.8% 9.9% 4.9% 14.8% 10.8% 13.9% 15.7% 30.5% 45.3% 66.8% 57.4% 7.6% 4.0% 21.1% 22.0% 47.1% n/a n/a
79.9 (9.3) 6.3% 7.9% 6.8% 0.5% 78.5% 45.0% 3.7% 7.9% 3.1% 14.1% 11.5% 20.9% 12.0% 31.4% 43.5% 70.7% 53.9% 13.1% 4.2% 21.5% 22.5% 49.2% n/a n/a
13.9% 0.0% 12.1% 44.4% 4.0%
56.5% 1.6% 6.3% 35.1% 4.7%
2.6 4.4 7.6 2.3 4.3
(1.1) (1.4) (2.3) (1.0) (2.4)
2.6 4.5 7.7 2.8 4.6
(1.2) (1.6) (2.5) (1.0) (2.3)
ATRIA-CVRN P-Value
0.002 0.78
0.75 0.31 0.47 0.63 0.81 0.06 0.30 0.84 0.71 0.40 0.48 0.07 0.94 0.92 0.90 0.67
<0.001 0.10 0.04 0.05 0.74 0.93 0.89 0.67 <0.001 0.12
Warfarin use at GI hemorrhage (n ¼ 210)
No Warfarin at GI hemorrhage (n ¼ 151)
P-Value
76.0 (8.0) 9.5% 8.1% 15.7% 1.0% 65.7% 40.5% 27.1% 15.7% 4.8% 6.2% 12.4% 14.8% 9.5% 41.4% 42.4% 91.4% 48.6% 3.3% 3.3% 12.4% 22.9% 55.2% 47.6% 7.1%
77.9 (9.1) 9.3% 13.3% 9.9% 0.7% 66.9% 48.3% 29.1% 13.9% 7.3% 5.3% 21.2% 20.5% 8.0% 45.0% 39.1% 94.7% 49.7% 7.3% 5.3% 13.9% 23.2% 65.6% 44.4% 6.0%
0.03 0.33
0.02 0.15 0.60 0.49 0.52 0.24 0.84 0.09 0.36 0.67 0.94 0.05 0.54 0.66
38.6% 10.5% 63.3% 38.6% 11.0%
63.6% 12.6% 51.0% 40.4% 10.6%
<0.001 0.53 0.02 0.73 0.91
2.6 4.4 7.1 2.6 4.4
(2.5) (1.5) (2.3) (0.9) (2.0)
2.7 4.6 7.5 3.0 5.0
(2.5) (1.6) (2.4) (0.9) (2.1)
0.14 0.68 0.63 0.57
0.46 0.28 0.07 <0.001 0.01
* The HAS-BLED score presented here has a maximum of 7 points since data on alcohol usage was not available and time in therapeutic range was not included. If adding 1 point for time in therapeutic range < 60%, there is no significant difference between those using and not using warfarin. † All cancer excluding non-melanomatous skin cancer. z Female: hemoglobin < 13.5 g/L; Male: hemoglobin < 12.0 g/L.
therefore, decided to examine the cohorts separately rather than as a combined data set (Supplementary Material B). We conducted descriptive and stratified analyses to examine the distribution of potential confounders in those using and not using warfarin. Cox proportional hazards regression was used to estimate the crude and adjusted mortality rate ratio (mRR) for the association between warfarin use at the time of GI hemorrhage and mortality (SAS, version 9.2; SAS Institute, Cary, North Carolina). Additionally, we examined models stratified by age group (<85 and 85 years). To control for baseline confounding, we used Cox proportional hazard regression models adjusted for propensity
score. Propensity scores were generated from logistic regression models to estimate the predicted probability of using warfarin at the time of GI hemorrhage versus not using warfarin.13 Propensity scores were generated for 4 separate analyses: (1) ATRIA sample: 30-day mortality, (2) ATRIA-CVRN sample: 30-day mortality, (3) ATRIA sample: long-term mortality, and (4) ATRIA-CVRN sample: long-term mortality. For each of these 4 analyses, we fit a preliminary proportional hazards model predicting death that included all potential confounders to determine the strength of the association of each covariable with either short- or long-term
4
The American Journal of Cardiology (www.ajconline.org)
Table 2 Association between warfarin usage at the time of major gastrointestinal hemorrhage and mortality in the ATRIA and ATRIA-CVRN cohorts Variable
ATRIA Warfarin use at GI hemorrhage
30-day Mortality Deaths Person-years Rate Ratio (unadjusted) Rate Ratio (adjusted)* Long-term mortality (>30 days following hospital admission for hemorrhage through 1-year following admission) Deaths Person-years Rate Ratio (unadjusted) Rate Ratio (adjusted)†
(n ¼ 223) 25 (11%) 16.8 0.84 (0.48-1.46) 0.97 (0.54-1.74) (n ¼ 194)
46 (24%) 153.2 0.66 (0.44-0.97) 0.70 (0.47-1.05)
ATRIA-CVRN No Warfarin at GI hemorrhage
Warfarin use at GI hemorrhage
No Warfarin at GI hemorrhage
(n ¼ 191) 25 (13%) 14.0 -
(n ¼ 210) 10 (5%) 16.3 0.33 (0.16-0.70) 0.37 (0.17-0.81) (n ¼ 176)
(n ¼ 151) 21 (14%) 11.2 -
(n ¼ 160)
55 (34%) 119.3 -
23 (13%) 117.8 0.90 (0.48-1.68) 0.83 (0.42-1.66)
(n ¼ 128)
17 (13%) 75.7 -
* Variables included in propensity score for short-term analyses, ATRIA: age, race/ethnicity, low educational attainment, low annual household income, smoking status, history of falls, prior ischemic stroke, diabetes mellitus, coronary disease, hypertension, chronic heart failure, dementia, kidney dysfunction, cancer, use of other antiplatelet therapies, lipid-lowering agents, and ACE inhibitors; ATRIA-CVRN: age, race/ethnicity, gender, low educational attainment, low annual household income, smoking status, history of GI hemorrhage, prior ischemic stroke, diabetes mellitus, coronary disease, hypertension, chronic heart failure, dementia, chronic liver disease, kidney dysfunction, anemia, lung disease, transient ischemic attack, use of antiplatelets, lipid-lowering agents, ACE inhibitors, angiotensin II receptor blockers. † Variables included in propensity score for long-term analyses, ATRIA: age, race, gender, low educational attainment, smoking status, history of coronary disease, chronic heart failure, dementia, kidney dysfunction, cancer, anemia, use of other antiplatelet therapies, lipid-lowering agents, ACE inhibitors, and angiotensin II receptor blockers; ATRIA-CVRN: age, race/ethnicity, gender, low educational attainment, smoking status, history of falls, prior GI hemorrhage, ischemic stroke, diabetes mellitus, coronary disease, hypertension, chronic heart failure, dementia, chronic liver disease, kidney dysfunction, cancer, anemia, lung disease, use of aspirin, antiplatelets, lipid-lowering agents, ACE inhibitors, and angiotensin II receptor blockers.
mortality. To avoid including nonconfounding predictors of warfarin use in the propensity score model, only those variables with a hazard ratio (or 1/hazard ratio) of at least 1.2 from the preliminary outcome model were used to create the propensity scores. After generating the propensity scores, we trimmed 2.5% of patients on both extremes of the propensity score distribution to improve comparability of the warfarinuse and nonewarfarin-use groups. These observations were outside the primary area of overlap of the propensity scores and increase residual confounding.14 The Cox proportional hazards models were adjusted for propensity score quintile (Supplementary Material C). We also performed additional analyses where the propensity score stratification was more finely divided (8 and 10 strata, respectively). Results In the 414 patients in ATRIA with major GI hemorrhages, 54% were taking warfarin. Major GI hemorrhages included 205 (49%) upper GI bleeds, 94 (23%) lower GI bleeds, and 115 (28%) where the location was undetermined by the time of hospital discharge. At the time of GI hemorrhage, patients using warfarin were younger, much less likely to be using aspirin, less likely to have diagnosed dementia or a history of GI hemorrhage, and were more likely to have a history of ischemic stroke and to be using lipid-lowering agents and angiotensin-converting enzyme inhibitors, compared with patients not using warfarin. There was no statistically
significant difference in CHADS2, CHA2DS2-VASc, or ATRIA stroke risk scores between patients using or not using warfarin at the time of their bleeding event, whereas warfarin users had clearly lower HAS-BLED and ATRIA hemorrhage risk scores (Table 1). Characteristics of warfarin users and nonusers within propensity score quintiles were quite similar (Supplementary Material C). Overall, 398 (96%) patients were transfused with a median number of 3.0 (interquartile range [IQR]) units of blood. In warfarin users, the median INR at presentation was 3.0 (IQR). FFP was used in 56% of warfarin users. In the 361 patients in ATRIA-CVRN with major GI hemorrhages, 58% were taking warfarin. The distribution of anatomical sites was very similar to that of the ATRIA cohort with 176 (49%) upper GI bleeds, 99 (27%) lower GI bleeds, and 86 (24%) undetermined. Similar to those in the ATRIA cohort, patients with GI hemorrhage in the ATRIA-CVRN using warfarin were younger, less likely to be women, less likely to have a history of falls, GI hemorrhage, or dementia, and much less likely to be using aspirin. In the more recent ATRIA-CVRN cohort, the use of lipid-lowering agents was far more common and patients with GI bleed on warfarin were prescribed such agents more frequently than those not taking warfarin. Patients with GI bleed on warfarin were less likely to have preexisting anemia. As with the ATRIA experience, there was no statistically significant difference in CHADS2, CHA2DS2-VASc, or ATRIA stroke scores between patients using or not using warfarin at the time of their
Arrhythmias and Conduction Disturbances/Death Following Gastrointestinal Hemorrhage On and Off Warfarin
5
Figure 1. (A) Adjusted survival function for the association between warfarin and long-term mortality in the ATRIA cohort (1996 to 2003). The y-axis (survival probability) extends from 0.5 to 1.0. (B) Adjusted survival function for the association between warfarin and long-term mortality in the ATRIA-CVRN cohort (2006 to 2009). The y-axis (survival probability) extends from 0.5 to 1.
bleeding event, whereas warfarin users had lower HASBLED and ATRIA hemorrhage risk scores (Table 1). Comparison of the characteristics of warfarin users and nonusers in each of the 5 propensity score quintiles showed similar distributions for covariates included in the propensity score model (Supplementary Material C). Overall, 357 (99%) patients were transfused with 2.0 (IQR) units of blood. In warfarin users, the median INR at presentation was 3.2 (IQR).
In the ATRIA cohort, warfarin use at the time of GI hemorrhage was not associated with 30-day mortality in unadjusted or adjusted analyses. In contrast, in ATRIACVRN, warfarin use at the time of GI hemorrhage was associated with a much lower 30-day mortality rate (Table 2). Warfarin use appeared to be more strongly associated with lower 30-day mortality in the oldest group of patients. In the ATRIA cohort, the adjusted mRR (95% confidence interval) was 0.75 (0.28 to 2.03) for patients
6
The American Journal of Cardiology (www.ajconline.org)
85 years (n ¼ 94) versus 1.16 (0.56 to 2.41) for those <85 years (n ¼ 320). In ATRIA-CVRN, warfarin appeared to be associated with reduced mortality in all ages, but, as in ATRIA, the largest observed association again was seen in older patients (adjusted mRR [95% confidence interval] for 85 years [n ¼ 62] 0.15 [0.02 to 1.27]; for <85 years [n ¼ 299] 0.50 [0.20 to 1.23]). A substantial number of patients resumed using warfarin in the year after GI hemorrhage (ATRIA 25%; ATRIA-CVRN 52%), whereas few nonwarfarin users began using warfarin during this time period (ATRIA 1%; ATRIA-CVRN 6%). After excluding deaths and follow-up time during the first 30 days after major GI hemorrhage, warfarin use was associated with a small, but not statistically significant, reduction in mortality over 1 year of follow-up after GI hemorrhage in both ATRIA and ATRIA-CVRN (Table 2 and Figure 1). Discussion We compared the mortality rates of patients sustaining a major GI hemorrhage on- versus off-warfarin therapy in 2 very large community-based cohorts of patients with AF. Patients using warfarin at the time of the GI hemorrhage had a reduced 30-day mortality rate in the more contemporary ATRIA-CVRN cohort. In the older ATRIA cohort, the 30day mortality rate was essentially the same regardless of whether the patient had been using warfarin or not. When stratified by age, warfarin use was associated with better outcome in the oldest patients (85 years) in both cohorts, although the age-stratified results were conducted in small groups and were not statistically significant. For long-term mortality, the data from both cohorts were consistent with a reduced mortality rate in patients whose GI hemorrhage occurred while using warfarin, although the confidence intervals for the mRRs included the null value of 1.0. Our finding that warfarin users had the same or lower 30day mortality risk after major GI hemorrhage clearly contrasts with warfarin’s effect on intracranial hemorrhage where anticoagulation markedly worsens outcome.15,16 In previous ATRIA analyses, warfarin use at the time of intracranial hemorrhage was independently associated with a 62% increased odds of 30-day mortality. The lessened mortality risk with GI hemorrhage may be the result of less severe lesions bleeding on warfarin and/or reversibility of warfarin’s hemorrhagic effect.17 In contrast, patients who experience intracranial hemorrhages while on warfarin often face grim outcomes before the effect of warfarin can be reversed.18 Part of the observed mortality effect in GI hemorrhage may be related to residual confounding by underlying co-morbidity despite our propensity score analysis. Indeed, the lower long-term post 30-day mortality we observed in patients taking warfarin at the time of GI bleed is consistent with preferential use of warfarin in healthier patients because the impact of a resolved GI hemorrhage must be small by 30 days. More generally, the anticoagulation decision for patients with AF depends on the expected net clinical benefit of warfarin, which summarizes the difference in thromboembolic events (primarily ischemic stroke) prevented and hemorrhagic events induced.7e11,19 Net clinical benefit depends on both the rates of events on and off warfarin and the
severity of the different types of events. Extracranial hemorrhages are included in some,20,21 but not all,22 proposed calculations of net clinical benefit. The current analyses, and our previous work on intracranial hemorrhages,15 make clear that there is a different distribution of severity for both thromboembolic and hemorrhagic events depending on whether the patient was taking warfarin at the time of the event. Our current work demonstrates that GI hemorrhages on warfarin are certainly no worse and may be less life threatening than those occurring off warfarin. Although confidence intervals were wide, reduced GI hemorrhage mortality appeared greatest for the oldest patients, a group at the highest risk of hemorrhage and least likely to be prescribed warfarin.22e26 In all, our results add to the expected net clinical benefit of warfarin, particularly in older patients with AF. This study was strengthened by including a large number of validated major GI hemorrhages from 2 separate large cohorts from KP Northern and Southern California. These cohorts were community based with comprehensive followup, with the result that the findings are likely to be broadly generalizable. All events were adjudicated by physicians or nurses using a standardized medical record review protocol. Such an approach minimizes error in case ascertainment and determination of warfarin exposure, a problem faced by studies using only administrative data. Additionally, we used propensity score modeling tailored to each analysis to control for confounding. This approach ensured that comparisons were made between patients who had similar distributions of measured confounders, thereby reducing confounding.14 We observed a much larger association between warfarin and reduced short-term mortality in the more contemporary ATRIA-CVRN cohort. This difference was mainly because of a lower mortality rate in warfarin users in this cohort, whereas mortality rates were similar in nonwarfarin users in the 2 cohorts. It is not clear why the GI hemorrhage mortality rate was lower in warfarin users in the ATRIA-CVRN cohort, although part of the difference may be explained by the play of chance. Even with these differences across cohorts, our results make clear, at the least, that GI hemorrhages occurring in patients taking warfarin pose no increase in mortality risk compared with GI hemorrhages occurring in patients not taking warfarin. Our results are indirectly applicable to the use of novel anticoagulants for AF because GI hemorrhages remain a major concern with these agents.4 There are very limited data on outcomes of GI hemorrhages comparing patients on novel anticoagulants versus patients not taking anticoagulants. It does appear that the case fatality rate for extracranial hemorrhages on novel agents is the same or lower than for extracranial hemorrhages on warfarin.27 As such, the reassuring results of our analysis should apply to GI hemorrhages on novel agents. This study also has several potential limitations. It is possible that there is residual confounding by indication because of underlying co-morbidity, that is, patients not receiving warfarin may have been “sicker.” These patients may be at particularly high mortality risk and may not be anticoagulated for this reason. We aggressively addressed confounding by propensity score approaches. Furthermore, we also included terms in our regression models for
Arrhythmias and Conduction Disturbances/Death Following Gastrointestinal Hemorrhage On and Off Warfarin
variables that were not balanced (p <0.20) in at least 4 quintiles of the propensity score, and the mRRs were similar to models only adjusted for propensity score quintile. We also performed additional analyses where the propensity score stratification was more finely divided (8 and 10 strata), and the results were similar (data not shown). Despite the advantages of propensity score methods, they do not account for unmeasured confounders. It is likely that residual confounding persists in our comparison of warfarin users versus nonusers. However, it is unlikely that any residual confounding is large enough to reverse the association observed for short-term mortality in the ATRIA-CVRN cohort. Disclosures Dr. Singer serves as a consultant/advisory board member for Boehringer Ingelheim, Bristol-Myers Squibb, and Johnson and Johnson related to novel anticoagulants to prevent stroke in AF and with CSL Behring related to reversal of warfarin anticoagulation. Additionally, Dr. Singer has research contracts with Medtronic, Inc., related to AF and risk of stroke and with Bristol-Myers Squibb related to AF and risk of stroke. Dr. Go has received a research grant from CSL Behring related to reversal of warfarin anticoagulation. No other authors have relation with industry. Supplementary Data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. amjcard.2014.10.006. 1. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomized controlled trials. Arch Intern Med 1994;154:1449e1457. 2. Fang MC, Go AS, Chang Y, Hylek EM, Henault LE, Jensvold NG, Singer DE. Death and disability from warfarin-associated intracranial and extracranial hemorrhages. Am J Med 2007;120:700e705. 3. Linkins LA, Choi PT, Douketis JD. Clinical impact of bleeding in patients taking oral anticoagulant therapy for venous thromboembolism: a meta-analysis. Ann Intern Med 2003;139:893e900. 4. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitz MD, Camm AJ, Weitz JI, Lewis BS, Parkhomenko A, Yamashita T, Antman EM. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials. Lancet 2014;383:955e962. 5. Go AS, Hylek EM, Chang Y, Phillips KA, Henault LE, Capra AM, Jensvold NG, Selby JV, Singer DE. Anticoagulation therapy for stroke prevention in atrial fibrillation: how well do randomized trials translate into clinical practice? JAMA 2003;290:2685e2692. 6. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD EPI. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604e612. 7. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864e2870. 8. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. 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 2010;137:263e272. 9. Singer DE, Chang Y, Borowsky LH, Fang MC, Pomernacki NK, Udaltsova N, Reynolds K, Go AS. A new risk scheme to predict ischemic stroke and other thromboembolism in atrial fibrillation: the ATRIA study stroke risk score. J Am Heart Assoc 2013;2:e000250.
7
10. Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. 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;138:1093e1100. 11. Fang MC, Go AS, Chang Y, Borowsky LH, Pomernacki NK, Udaltsova N, Singer DE. A new risk scheme to predict warfarinassociated hemorrhage: the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) Study. J Am Coll Cardiol 2011;58:395e401. 12. Suissa S. Immortal time bias in observational studies of drug effects. Pharmacoepidemiol Drug Saf 2007;16:241e249. 13. Rosenbaum P, Rubin D. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41e55. 14. Sturmer T, Rothman KJ, Avorn J, Glynn RJ. Treatment effects in the presence of unmeasured confounding: dealing with observations in the tails of the propensity score distribution—a simulation study. Am J Epidemiol 2010;172:843e854. 15. Fang MC, Go AS, Chang Y, Borowsky LH, Pomernacki NK, Udaltsova N, Singer DE. Thirty-day mortality after ischemic stroke and intracranial hemorrhage in patients with atrial fibrillation on and off anticoagulants. Stroke 2012;43:1795e1799. 16. Rosand J, Eckman MH, Knudsen KA, Singer DE, Greenberg SM. The effect of warfarin and intensity of anticoagulation on outcome of intracerebral hemorrhage. Arch Intern Med 2004;164:880e884. 17. Schulman S, Beyth RJ, Kearon C, Levine MN; American College of Chest Physicians. Hemorrhagic complications of anticoagulant and thrombolytic treatment: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133:257Se298S. 18. Goldstein JN, Thomas SH, Frontiero V, Joseph A, Engel C, Snider R, Smith EE, Greenberg SM, Rosand J. Timing of fresh frozen plasma administration and rapid correction of coagulopathy in warfarin-related intracerebral hemorrhage. Stroke 2006;37:151e155. 19. Gage BF, Yan Y, Milligan PE, Waterman AD, Culverhouse R, Rich MW, Radford MJ. Clinical classification schemes for predicting hemorrhage: results from the National Registry of Atrial Fibrillation (NRAF). Am Heart J 2006;151:713e719. 20. Connolly SJ, Eikelboom JW, Ng J, Hirsh J, Yusuf S, Pogue J, de Caterina R, Hohnloser S, Hart RG; ACTIVE (Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events) Steering Committee and Investigators. Net clinical benefit of adding clopidogrel to aspirin therapy in patients with atrial fibrillation for whom vitamin K antagonists are unsuitable. Ann Intern Med 2011;155: 579e586. 21. Eikelboom JW, Connolly SJ, Hart RG, Wallentin L, Reilly P, Oldgren J, Yang S, Yusuf S. Balancing the benefits and risks of 2 doses of dabigatran compared with warfarin in atrial fibrillation. J Am Coll Cardiol 2013;62:900e908. 22. Singer DE, Chang Y, Fang MC, Borowsky LH, Pomernacki NK, Udaltsova N, Go AS. The net clinical benefit of warfarin anticoagulation in atrial fibrillation. Ann Intern Med 2009;151:297e305. 23. Bungard TJ, Ghali WA, McAlister FA, Buchan AM, Cave AJ, Hamilton PG, Mitchell LB, Shuaib A, Teo KK, Tsuyuki RT. The relative importance of barriers to the prescription of warfarin for nonvalvular atrial fibrillation. Can J Cardiol 2003;19:280e284. 24. Casciano JP, Singer DE, Kwong WJ, Fox ES, Martin BC. Anticoagulation therapy for patients with non-valvular atrial fibrillation: comparison of decision analytic model recommendations and realworld warfarin prescription use. Am J Cardiovasc Drugs 2012;12: 313e323. 25. Fang MC, Chang Y, Hylek EM, Rosand J, Greenberg SM, Go AS, Singer DE. Advanced age, anticoagulation intensity, and risk for intracranial hemorrhage among patients taking warfarin for atrial fibrillation. Ann Intern Med 2004;141:745e752. 26. Gage BF, Boechler M, Doggette AL, Fortune G, Flaker GC, Rich MW, Radford MJ. Adverse outcomes and predictors of underuse of antithrombotic therapy in Medicare beneficiaries with chronic atrial fibrillation. Stroke 2000;31:822e827. 27. Hylek EM, Held C, Alexander JH, Lopes RD, De Caterina R, Wojdyla DM, Huber K, Jansky P, Steg PG, Hanna M, Thomas L, Wallentin L, Granger CB. Major bleeding in patients with atrial fibrillation receiving apixaban or warfarin: the ARISTOTLE Trial (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation): predictors, characteristics, and clinical outcomes. J Am Coll Cardiol 2014;63:2141e2147.