Journal of the American College of Cardiology Vol. 58, No. 19, 2011 2011 by the American College of Cardiology Foundation ISSN 0735-1097/$36.00 Published by Elsevier Inc.
Heart Rhythm Disorders Cardiovascular Outcomes in theAFFIRM Trial (Atrial FibrillationFollow-Up Investigation of Rhythm Management) An Assessment of Individual Antiarrhythmic Drug TherapiesCompared With Rate Control With Propensity Score-Matched Analyses Sanjeev Saksena, MD,* April Slee, MS,* Albert L. Waldo, MD,* Nick Freemantle, PHD,*Mathew Reynolds, MD, MS,* Yves Rosenberg, MD,† Snehal Rathod, MS,* Shannon Grant MS,*Elizabeth Thomas, MS,* D. George Wyse, MD, PHD* Warren, New Jersey; and Bethesda, Maryland The impact of individual antiarrhythmic drugs (AADs) on mortality and hospital stay in atrial fibrillation (AF) was Cardiovascular (CV) outcomes in AF patients receiving pharmacologic rhythm control therapy have not been compared with rate control therapy on the basis of AAD selection.
We compared CV outcomes in the AFFIRM (Atrial Fibrillation Follow-Up Investigation of Rhythm Management) trial in subgroups defined by the initial AAD selected with propensity score matched subgroups from the rate arm (Rate).
Seven hundred twenty-nine amiodarone patients, 606 sotalol patients, and 268 Class 1C patients were matched. The composite outcome of mortality or cardiovascular hospital stays (CVH) showed better outcomes with Rate compared with amiodarone (hazard ratio [HR]: 1.18, 95% confidence interval [CI]: 1.03 to 1.36, p ⫽ 0.02), sotalol (HR: 1.32,95% CI: 1.13 to 1.54, p ⬍ 0.001), and Class 1C (HR: 1.22, 95% CI: 0.97 to 1.56, p ⫽ 0.10). There was a nonsignifi-cant increase in mortality with amiodarone (HR: 1.20, 95% CI: 0.94 to 1.53, p ⫽ 0.15) with the risk of non-CV deathbeing significantly higher with amiodarone versus Rate (HR: 1.11, 95% CI: 1.01 to 1.24, p ⫽ 0.04). First CVH eventrates at 3 years were 47% for amiodarone, 50% for sotalol, and 44% for Class 1C versus 40%, 40%, and 36%, re- spectively, for Rate (amiodarone HR: 1.20, 95% CI: 1.03 to 1.40, p ⫽ 0.02, sotalol HR: 1.364, 95% CI: 1.16 to 1.611,p ⬍ 0.001, Class 1C HR: 1.24, 95% CI: 0.96 to 1.60, p ⫽ 0.09). Time to CVH with intensive care unit stay or deathwas shorter with amiodarone (HR: 1.22, 95% CI: 1.02 to 1.46, p ⫽ 0.03).
In AFFIRM, composite mortality and CVH outcomes differed for Rate and AADs due to differences in CVH; CVH event rates during follow-up were high for all cohorts, but they were higher for all groups on AADs. Death, inten- sive care unit hospital stay, and non-CV death were more frequent with amiodarone. (Atrial Fibrillation Follow-Up Investigation of Rhythm Management; (J Am Coll Cardiol 2011;58:1975–85) 2011 by the American College of Cardiology Foundation Atrial fibrillation (AF) is the most prevalent tachyarrhyth-mia and is associated with increased mortality, stroke, and recurrent hospital stays Health care resource con- From the *Electrophysiology Research Foundation, Warren, New Jersey; and the Merck, Pfizer, Eli Lilly, Novo Nordisk, and Medtronic. Dr. Reynolds has received a †National Heart, Lung and Blood Institute, Bethesda, Maryland. Dr. Saksena is or research grant and is a consultant/advisory board member for Sanofi-Aventis. Dr. Wyse has been a consultant, investigator, and research grant recipient for the National Heart is a consultant to Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, Sanofi-Aventis, Lung and Blood Institute, Medtronic Inc., St. Jude Medical Inc., Sanofi-Aventis, Biotronik, Boston Scientific/Guidant, National Heart, Lung and Blood Institute, Duke Sorin Group, and Aryx Pharmaceuticals; and has been a Speakers' bureau member for Clinical Research Institute, European Commission, Merck, Medtronic, and Bayer, and Sanofi-Aventis. Dr. Waldo is a consultant to Sanofi-Aventis, Ortho-McNeil-Janssen, Speakers' Bureau member for Sanofi-Aventis. All other authors have reported that they Biotronik, St. Jude Medical, Daiichi, Sankyo Pharmaceuticals, Medtronic Inc., have no relationships relevant to the contents of this paper to disclose. For list of Astellas Pharma, Biosense Webster Inc., Bristol-Myers Squibb, Portola, Boehringer investigators and affiliated institutions, please see the Online Appendix.
Ingelheim, CardioInsight Technologies, Merck, AtriCure Inc., and Sanofi-Aventis; Manuscript received April 19, 2011; revised manuscript received July 18, 2011, and is a speaker for Sanofi-Aventis. Dr. Freemantle is a consultant for Sanofi-Aventis, accepted July 26, 2011.
Saksena et al.
JACC Vol. 58, No. 19, 2011 CV Outcomes of AADs in the AFFIRM Trial November 1, 2011:1975– 85 sumption due to AF, primarily (CVH). Individual components (all-cause mortality and due to hospital stay, is among the CVH) were also examined, as were subsets of both CVH highest for cardiovascular (CV) and all-cause mortality The AAD subgroups were AAD ⴝ antiarrhythmic drug diagnoses, but the patterns of compared with propensity score matched rate subgroups AF ⴝ atrial fibrillation these hospital stays and their re- (Rate) and included: 1) initial amiodarone therapy (amio- CI ⴝ confidence interval lationship to individual therapeutic darone cohort); 2) initial sotalol (sotalol cohort); and 3) CV ⴝ cardiovascular choices in AF have not been evalu- initial Class 1C drug (flecainide or propafenone, Class 1C CVH ⴝ cardiovascular ated The AFFIRM (Atrial Fi- brillation Follow-Up Investigation Propensity score matched subgroups were selected HR ⴝ hazard ratio of Rhythm Management) trial was from the rate control strategy arm (Rate) for each AAD ICU ⴝ intensive care unit conducted to examine 2 treatment cohort. The score was derived with 62 baseline patient Rate ⴝ rate control strategies for AF, namely rate con- characteristics from the AFFIRM database deemed a trol or rhythm control All- priori to potentially affect AAD selection. Two additional cause mortality, the primary out- characteristics that were determined to be important to come measure, showed a trend achieve balanced cohorts (left ventricular ejection frac- toward excess mortality in the rhythm control arm. The antiar- tion, and history of coronary artery disease) were added in rhythmic drugs (AADs) used in the rhythm arm have been cited as a potential cause of the excess mortality Despite concerns with regard to their safety, most of the AADs used in theAFFIRM trial remain widely used in clinical practice.
Relating outcomes to clinical and treatment factors. The
The impact of individual AADs on mortality and hospital severity of CVH was characterized by acuity of hospital stay stay outcomes in the AFFIRM population in relation to rate on the basis of concomitant intensive care unit (ICU) stay, control has not been available. In part, this was related to CV procedures, CV interventions, or emergency room the intent of the AFFIRM investigators to test the treat- visits. Outcomes in AAD subgroups were related to patient ment strategy hypothesis rather than individual drug ther- characteristics, underlying disease state, clinical events, and apies. In this report, we examined the impact on outcomes treatment strategy.
of the selection of amiodarone, sotalol, or a Class 1C Study Outcomes and Definitions
antiarrhythmic agent (flecainide or propafenone) as the firstAAD, compared with a rate strategy in the AFFIRM study.
The principal outcome for this analysis was a composite The AADs were selected for this analysis on the basis of outcome: the first of death from any cause or a CVH. A current widespread clinical usage. To address the nonran- CVH was defined as a hospital admission for CV reasons dom nature of drug selection in the rhythm arm, we (per investigator) or for non-CV reasons but with a CV employed propensity score matching derived from 64 base- event occurring during the same follow-up interval. Exact line patient characteristics deemed to affect antiarrhythmic dates were available for death but not for hospital admission selection. Propensity score matching has not been employed or discharge. The midpoint of the previous follow-up visit to assess individual drug outcomes in the AFFIRM trial and the follow-up visit when the hospital stay was reported We compared mortality and hospital stay outcomes in were used to estimate event time for CVH. Investigators patient subgroups defined by each type of AAD selected as recorded total number of hospital days and total number of first therapy with propensity score matched subgroups from ICU days. Visits occurred at 2 months after randomization the rate control arm.
and every 4 months thereafter. Patients who did notexperience CVH or death were censored at the last follow-up visit. For death alone, follow-up informationfrom a vital status sweep (telephone contact with all subjects Patient Selection in the AFFIRM Trial
and national death index scan) at the end of the study was The AFFIRM trial recruited consenting patients who had AF used to determine censoring date.
that was likely to be recurrent, warranted therapy, and had risk Statistical Methods and Analytical Techniques
factor(s) for stroke. Patients were candidates for at least 2 drugswithin each strategy and for anticoagulation Propensity score and establishment of matched cohorts.
The goal of development of propensity score matched
Primary Objective of Analysis
cohorts was to account for possible confounding variables Reassessment of clinical outcomes by initial AAD therapy.
that might be related to drug selection, because the patients The primary objective was to reassess clinical outcomes in were not assigned randomly to specific initial drug therapy the AF population enrolled in the AFFIRM study by initial in the AFFIRM trial.
AAD therapy with a composite principal outcome and its Selection of covariates. Propensity score was calculated
individual components. The principal outcome was a com- separately for each AAD subgroup (amiodarone, sotalol, or posite of mortality or first cardiovascular hospital stay Class 1C). Four patients received more than 1 AAD and JACC Vol. 58, No. 19, 2011 Saksena et al.
November 1, 2011:1975– 85 CV Outcomes of AADs in the AFFIRM Trial Covariates Used in Propensity Score Model Covariates Used in Propensity Score Model Primary cardiac diagnosis Coronary artery disease Year of randomization Current CCS angina class History of myocardial infarction Number of AAD failures History of pulmonary disease Failed amiodarone History of intracranial hemorrhage Failed disopyramide History of congestive heart failure, congestive heart failure on enrollment Failed flecainide History of cardiomyopathy Failed moricizine History of valvular heart disease Failed procainamide History of congenital heart disease Failed propafenone History of angina History of diabetes History of hepatic or renal disease History of symptomatic brady/atrioventricular block Previous other CV procedure History of resuscitated cardiac arrest Previous percutaneous coronary interventions History of stroke/transient ischemic attack Previous coronary artery bypass grafting History of peripheral vascular disease Previous thrombolytic therapy History of systemic embolism LV ejection fraction History of hemorrhage or coagulopathy History of thyroid disease/specific drugs—thyroid replacement History of carotid disease Symptoms constellations are 2. Diaphoresis, fatigue, panic, dizziness, syncope 4. Dyspnea, edema, orthopnea, paroxysmal nocturnal dyspnea 5. Fast heart rate, palpitations AF symptoms frequency Duration of qualifying AF episode(s) Hospitalized for qualifying episode Cardioverted for qualifying episode(s) Current ventricular/max HR during AF ⬎100 beats/min Other cardiac neurologic interaction List of covariates used in propensity score model. Please note that multiple imputation was used for body mass index (BMI) and systolic blood pressure (SBP).
AAD ⫽ antiarrhythmic drug; AF ⫽ atrial fibrillation; BMI ⫽ body mass index; CCS ⫽ Canadian Cardiovascular Society; CV ⫽ cardiovascular; FADS ⫽ first antiarrhythmic drug substudy; HR ⫽ heart rate; LV ⫽ left ventricular; max ⫽ maximum; NYHA ⫽ New York Heart Association functional class;SBP ⫽ Systolic blood pressure.
were excluded. The propensity score model used data from Model building. Proc GLIMMIX in SAS (version 9.2, SAS
AFFIRM patients randomized to rhythm control. Identical Institute, Cary, North Carolina) was used for building the baseline explanatory variables were included in each model propensity-matched cohorts. Each model considered all ex- and were prospectively determined by consensus before data planatory variables in Site was included as a fixed effect analysis This model included explanatory vari- for this step. The functional form of response was assessed for ables that might be considered by clinicians when selecting continuous variables to determine whether transformation was an AAD, including demographic data, clinical characteris- necessary Then, the model was run twice, with site as a tics of patients, treating physicians (cardiologists or other), fixed and then as a G-sided (generalized) random effect. These centers, and study design factors. Patients in the first AAD models were compared for evidence of extra binomial variabil- substudy had their first AAD randomly assigned, so partic- ity at the investigator site level. Risk score was calculated for ipation in first AAD sub-study was included as a variable each patient in the rate subgroup, and the VMATCH algo- A stepwise model reduction procedure was used to rithm (Zentrum fur Bioinformatik, Hamburg, Germany) was produce a parsimonious model for each propensity score used to construct the cohorts Matching was 1:1 between equation. After initial cohort construction, imbalances in 2 each AAD cohort and the rate cohort.
additional variables, coronary artery disease and left ventric- Descriptive reporting. Once the propensity score matched
ular ejection fraction, were identified; these items were cohorts were established, baseline demographic and clinical added to the model in a second step.
characteristics were tabulated to be consistent with the main Saksena et al.
JACC Vol. 58, No. 19, 2011 CV Outcomes of AADs in the AFFIRM Trial November 1, 2011:1975– 85 AFFIRM publication Tests for differences across patients were receiving the initially selected drug at first matched cohorts were conducted (Fisher exact or chi-square CVH. There was no increased mortality risk for sotalol and for categorical variables, analysis of variance or Wilcoxon for Class 1C cohorts, but an increase in risk was observed for amiodarone (HR: 1.20, 95% CI: 0.94 to 1.53, p ⫽ 0.15),compared with Rate, which was not statistically significant.
Time to first CVH was shorter for all AADs, compared The principal outcome analyzed was a comparison of event with Rate. First CVH event rates at 3 years were 47% for time with the log-rank test on an intention-to-treat basis, amiodarone, 50% for sotalol, and 44% for Class 1C com- similar to the primary AFFIRM analysis. Unadjusted pared with 40%, 40%, and 36%, respectively, for the Kaplan-Meier survival curves were examined for each matched Rate cohorts. The CV mortality did not differ propensity-score matched cohort pair. Proportional hazards between Rate and any of the AAD cohorts (p ⬎ 0.15 for all models were used to obtain hazard ratios (HRs) and 95% comparisons). There was an increased risk of noncardiovas- confidence intervals (CIs) and to determine the effect in cular mortality with amiodarone (HR: 1.11, 95% CI: 1.01 to clinically important subgroups.
1.24, p ⫽ 0.04) but not with sotalol or Class 1C drugs, Sensitivity analyses. To determine the impact of treatment
compared with Rate. However, deaths attributable to cancer strategy-related hospital stays, (e.g., cardioversions) and or pulmonary causes were comparable across each cohort.
further define acuity of CVH, we repeated the analysis with A composite of death or ICU hospital stays showed a composite of death and first hospital stay requiring ICU moderately increased risk with amiodarone (HR: 1.22, 95% stay. To evaluate the propensity score methodology, a Cox CI: 1.02 to 1.46, p ⫽ 0.03) but not with sotalol or Class 1C proportional hazards model with a frailty term for site was agents (HR: 1.06, 95% CI: 0.87 to 1.30, p ⫽ 0.56, and HR: 1.07, 95% CI: 0.78 to 1.46, p ⫽ 0.67, respectively),compared with Rate There was no difference intime to ICU hospital stays for sotalol and Class 1C, compared with Rate, but a nonsignificant increased risk was Patient population. Seven hundred twenty-nine AF pa-
noted for amiodarone (HR: 1.18, 95% CI: 0.95 to 1.47, p ⫽ tients initially received amiodarone therapy, 606 received 0.14) All-cause hospital stays were increased in initial sotalol therapy, and 268 received either initial flecain- amiodarone compared with Rate (HR: 1.19, 95% CI: 1.05 ide or propafenone. The clinical characteristics of these 3 to 1.35, p ⫽ 0.008) and in sotalol compared with Rate (HR: AAD cohorts on the basis of initial drug therapy selection 1.22, 95% CI: 1.06 to 1.41, p ⫽ 0.005). There was no are shown in The AAD cohorts were generally increased risk of all-cause hospital stay with Class 1C well-matched. Patients were usually elderly, predominantly compared with Rate.
male, and had recurrent AF associated with cardiac disease.
Concomitant beta-blocker therapy did not alter outcomes The amiodarone cohort had a slight excess of men, com- for either sotalol or Class 1C cohorts for either mortality or pared with its matched Rate cohort (67.4% vs. 61.3%, CVH risk (CVH for sotalol HR: 1.09, 95% CI: 0.89 to respectively). More patients in the sotalol cohort had a 1.34, for death HR: 1.15, 95% CI: 0.81 to 1.63; CVH for history of angina, compared with Rate (11.1% vs.
Class 1C HR: 0.75, 95% CI: 0.60 to 1.03), for death HR: 6.9%).There were no other significant differences. The C 0.65, 95% CI: 0.40 to 1.07). Amiodarone-Rate cohort statistic for the 3 propensity models were 0.814 for amio- patients who were concomitantly taking beta-blockers had darone, 0.837 for sotalol, and 0.837 for Class 1C subgroups.
an increased mortality risk (CVH risk for amiodarone HR: Outcomes analysis. HRs and 95% CIs for the overall com-
1.06, 95% CI: 0.90 to 1.25, for death HR: 1.53, 95% CI: parison (rhythm compared with rate) in the AFFIRM trial and 1.16 to 2.02). There was no evidence of a treatment– individual AAD subgroups with the matched rate cohort are digoxin interaction for the principal outcome. Time- shown for the composite principal outcome of mortality and dependent digoxin use was significantly associated with first CVH in All AAD cohorts had inferior CVH in the amiodarone-Rate cohorts (HR: 1.43, 95% CI: principal outcomes, compared with Rate (HR for amioda- 1.21 to 1.68) and in the Class 1C-Rate cohorts (HR: 1.36, rone: 1.18, 95% CI: 1.03 to 1.36, p ⫽ 0.02; HR for sotalol: 95% CI: 1.04 to 1.77) but not in the sotalol-Rate cohorts 1.32, 95% CI: 1.13 to 1.54, p ⬍ 0.001; and HR for Class (HR: 1.15, 95% CI: 0.96 to 1.37). After adjusting for 1C: 1.22, 95% CI: 0.97 to 1.56, p ⫽ 0.10). In the smaller time-dependent digoxin use, AADs still increased the risk Class 1C cohort, this difference did not reach statistical of CVH (HR for amiodarone: 1.34, 95% CI: 1.13 to 1.57, significance. shows the individual components of HR for sotalol: 1.40, 95% CI: 1.17 to 1.67, compared with the composite endpoint. Risk of CVH was increased for all matched rate patients; HR for Class 1C: 1.34, 95% CI: 1.03 3 AAD cohorts (amiodarone HR: 1.20, 95% CI: 1.03 to to 1.75, compared with the respective AAD rate-matched 1.40, p ⫽ 0.05; sotalol HR: 1.36, 95% CI: 1.16 to 1.61, p ⬍ patients). The increased risk of CVH or death was consis- 0.001; and Class 1C HR: 1.24, 95% CI: 0.96 to 1.64, p ⫽ tent across clinically important subgroups including coro- 0.09, compared with Rate). Ninety-one percent of amioda- nary disease, female sex, and age for amiodarone and sotalol rone patients, 88% of sotalol patients, and 78% of Class 1C patients, presence of thyroid disease only in amiodarone Baseline Patient Characteristics for Entire Rate Cohort in AFFIRM Baseline Patient Characteristics for Entire Rate Cohort in AFFIRM Ethnic minority group Predominant cardiac diagnosis Coronary artery disease (MI, angina, and so on) Dilated cardiomyopathy Valvular heart disease No apparent heart disease History of congestive heart failure Duration of qualifying AF ⱖ2 days First episode of AF (vs. recurrent episode)* Any pre-randomization failure of an antiarrhythmic drug Size of left atrium normal† Baseline CCS class Class II or greater Baseline NYHA functional class Values are mean ⫾ SD or n (%). Baseline patient characteristics for entire Rate cohort in AFFIRM (Atrial Fibrillation Follow-Up Investigation of Rhythm Management) trial (Overall Rate) and the 3 paired propensity (PS) matched cohorts for individual antiarrhythmic drugs and matched rate control groups. The size of the left atrium was unknown in 185 of 3,311 cases, and left ventricular function (where normal was defined as left ventricular ejection fraction [LVEF] ⫽ 0.50) was unknown in 279 of 3,311. Electrocardiogram information was not used in PS models. *This information was not collected on the initial version of the data form and therefore was imputed for 143 patients. †Electrocardiograms were obtained in 3,311 of 4,060.
CHF ⫽ congestive heart failure; MI ⫽ myocardial infarction; other abbreviations as in

Saksena et al.
JACC Vol. 58, No. 19, 2011 CV Outcomes of AADs in the AFFIRM Trial November 1, 2011:1975– 85 Comparison of Composite Principal Outcome: Individual AADs Versus Rate Hazard ratios (HRs) and Kaplan-Meier survival analyses comparing individual antiarrhythmic drugs (AADs) with matched rate control strategy arm (Rate) cohorts for the composite principal outcome (time to first cardiovascular hospital stay [CVH] or death). Individual panels are shown as follows: (A) HRs and 95% confidence intervals (CIs) (HR: rhythm drug/Rate); (B) propensity score matched Rate and amiodarone (Amio) subgroups; (C) propensity score matched Rate and sotalol subgroups; and (D) propensity score matched Rate and Class 1C subgroups. All AADs and matched Rate cohorts show substantial event rates for the principal outcome during follow- up, but all AADs studied had a higher risk of events during follow-up. LR ⫽ log rank.
patients but in none of the subgroups examined for the sex was associated with increased risk in sotalol and Class Class 1C patients. These results are detailed in the next 1C cohorts, compared with matched Rate cohorts, but this was not observed in the amiodarone-Rate cohort compari- CVH categorized by intensity, duration, and associated son. A history of heart failure, coronary disease, and diabetes procedures are tabulated in There were substantially at enrollment were associated with increased risk for CVH more hospital stays of ⬍3-day duration associated with in all AAD cohorts. Pulmonary disease at baseline was cardioversion in the amiodarone and sotalol cohorts than associated with increased risk of CVH with amiodarone, matched rate cohorts. Cardioversion occurred at similar and age ⬎75 years was associated with increased risk of CVH rates in the matched Class 1C and Rate cohorts (7.2%).
with sotalol. There was evidence of significant AAD– Cardiovascular hospital stays with a length of stay of ⬍3 comorbidity interactions only in the amiodarone cohort; age days with a cardioversion procedure alone (without another ⬎75 years and thyroid disease were associated with increased CV procedure, emergency room visits, or ICU stay [i.e., risk for amiodarone patients but not for their matched Rate events that might reflect adherence to AF rhythm control counterparts. A significant increased risk for CVH was main- treatment strategy only]) constituted 6.1%, 6.1%, and 4.0% tained for amiodarone and sotalol, compared with Rate, of first CVH for amiodarone, sotalol, and Class 1C, despite adjustments for age, sex, or any of these comorbidities.
respectively. The corresponding rates in the matched Rate Time-dependent changes in clinical status that increased cohorts were 1.9%, 1.6%, and 0.9%, respectively. Stroke, risk of CVH are shown in In the amiodarone embolism, and major bleeds accounted for only a minority patient cohort, relapse from sinus rhythm to AF and of first CVH in both AAD and rate cohorts increase in New York Heart Association (NYHA) func- Warfarin use at first CVH or death was slightly but not tional class by 1 or more were associated with a 1.9- and significantly higher in the rate cohorts.
1.7-fold increase in CVH risk, respectively. For sotalol, Potential risk factors for CVH. Baseline historical char-
relapse from sinus rhythm to AF, increase ⱖ1 in NYHA acteristics that increased risk of CVH with AAD, compared functional class, increase in angina class by 1 or more, and with matched Rate cohorts, are shown in Female ventricular rate increase ⱖ15 beats/min were all associated

JACC Vol. 58, No. 19, 2011 Saksena et al.
November 1, 2011:1975– 85 CV Outcomes of AADs in the AFFIRM Trial Components of Principal Outcome—First CVH and Mortality: Individual AADs Versus Rate (A) First CVH: individual AADs versus Rate. The HRs and Kaplan-Meier survival analyses comparing individual AADs with matched Rate cohorts for a component of princi- pal outcome: time to first CVH. Individual panels are shown as follows: 1) HRs and 95% CIs (HR: rhythm drug/Rate); 2) propensity score matched Rate and Amio sub- groups; 3) propensity score matched Rate and sotalol subgroups; 4) propensity score matched Rate and Class 1C subgroups. All AADs and matched Rate cohorts show substantial event rates during follow-up, but all AADs studied had a significantly higher risk of a first CVH during follow-up. (B) Mortality: individual AADs versus Rate. The HRs and Kaplan-Meier survival analyses comparing individual AADs with matched Rate cohorts for a component of principal outcome: time to death. Individual panels are shown as follows: 1) HRs and 95% CIs (HR: rhythm drug/Rate); 2) propensity score matched Rate and Amio subgroups; 3) propensity score matched Rate and sota- lol subgroups; and 4) propensity score matched Rate and Class 1C subgroups. Sotalol and Class 1C groups and matched rate cohorts show comparable event rates for risk of death during follow-up, but there is a nonsignificant increase in mortality with Amio compared with its matched Rate cohort. Abbreviations as in

Saksena et al.
JACC Vol. 58, No. 19, 2011 CV Outcomes of AADs in the AFFIRM Trial November 1, 2011:1975– 85 Comparison of Secondary Composite Outcome—ICUH or Death: Individual AADs Versus Rate (A) Secondary composite outcome (intensive care unit hospital stays [ICUH] or death): individual AADs versus Rate. The HRs and Kaplan-Meier survival analyses comparing individual AADs with matched rate cohorts for secondary composite outcome: time to first ICUH or death. Individual panels are shown as follows: 1) HRs and 95% CIs (HR: rhythm drug/Rate); 2) propensity score matched Rate and Amio subgroups; 3) propensity score matched Rate and sotalol subgroups; 4) propensity score matched Rate and Class 1C subgroups. Composite outcome shows that time to ICUH or death was shorter with Amio but not with sotalol or Class 1C versus Rate during follow-up. (B) Comparison of ICUH: individual AADs versus Rate. The HRs and Kaplan-Meier survival analyses comparing individual AADs with matched rate cohorts for secondary outcome: time to first ICUH. Individual panels are shown as follows: 1) HRs and 95% CIs (HR: rhythm drug/Rate); 2) propensity score matched Rate and Amio subgroups; 3) propensity score matched Rate and sotalol subgroups; 4) propensity score matched Rate and Class 1C subgroups. Time to ICUH was comparable for sotalol and Class 1C groups, compared with matched Rate cohorts, but a nonsignificant increased risk was seen with Amio compared with Rate during follow-up.
JACC Vol. 58, No. 19, 2011 Saksena et al.
November 1, 2011:1975– 85 CV Outcomes of AADs in the AFFIRM Trial Patient Cohorts for Individual Antiarrhythmic Drugs in the AFFIRM Trial # fatal first CVH CVH ⬍3 days ⫹ CV CVH ⬍3 days, CV, no ER/ICU ICU days first CVH Warfarin use at first CVH (% of CVH) Bleeds/stroke/embolic events (% of CVH) Warfarin use at above event (% of event) Values are n or n (%).
AFFIRM ⫽ Atrial Fibrillation Follow-Up Investigation of Rhythm Management; CV ⫽ cardiovascular event; CVH ⫽ cardiovascular hospital stay(s); ER ⫽ emergency room visit; ICU ⫽ intensive care unit stay.
with increased risk for CVH. For Class 1C, ventricular rate To evaluate these agents individually, we employed propen- increase ⱖ15 beats/min was associated with increased risk.
sity score matching to permit comparative analysis with the Higher absolute ventricular rate (in steps of 15 beats/min) rate control patients In this report, it produced highly was associated with increased risk for sotalol and Class 1C comparable Rate and AAD cohorts for demographic data, patients. Overall, a higher NYHA functional class was disease status and severity, prior interventions, and therapy associated with increased risk for all cohorts and higher angina class for amiodarone and Class 1C patients.
Major Findings of Study
Clinical outcomes, especially CVH, are affected by initial
Analyses of overall and secondary outcomes for the AF AAD selection. The present analysis demonstrates inferior
population in the AFFIRM study have suggested no overar- performance in the principal clinical outcome for the indi- ching benefit of a particular strategy There was, vidual AADs studied versus rate control for the AFFIRM however, a nonsignificant increase in mortality in the rhythm population. This difference in composite outcome was arm with an excess in pulmonary and cancer deaths largely due to excess and earlier CVH for each AAD.
This finding raised the specter of AAD therapy-related mor- Sotalol and Class 1C cohorts were comparable to Rate for tality risk. The impact of individual AAD selection on both all-cause mortality. The HR comparing amiodarone with mortality and hospital stay, compared with Rate, has not been Rate was very similar to the overall AFFIRM study result available due to the investigator-determined process for AAD for mortality risk with rhythm control, but in this small selection, which makes unbiased comparisons challenging.
matched cohort the power to see a significant difference was However, such an analysis is still relevant and potentially low (⬍30%). Initial amiodarone therapy was associated with informative, because most of these agents are currently in significantly increased risk of non-CV death and mortality widespread clinical use and still employed in clinical trials plus ICU hospital stay. The sotalol and Class 1C cohorts were similar to Rate with respect to these outcomes, Relationship Between Baseline Characteristics and Risk of CVH Relationship Between Baseline Characteristics and Risk of CVH Amiodarone-Rate Cohort Sotalol-Rate Cohort Class 1C-Rate Cohort Baseline variable 1.63 (1.4–1.91)* 1.55 (1.29–1.86)* 1.5 (1.08–2.08)† 1.08 (0.92–1.27) 1.23 (1.04–1.46)† 1.37 (1.06–1.78)† Coronary artery disease 1.83 (1.57–2.14)* 1.4 (1.18–1.65)* 1.37 (1.01–1.85)† Pulmonary disease 1.3 (1.08–1.58)‡ 1.07 (0.82–1.4) 1.23 (0.86–1.74) 1.62 (1.36–1.92)* 1.29 (1.07–1.57)‡ 1.56 (1.13–2.15)‡ 1.44 (1.16–1.79)‡ 1.12 (0.88–1.43) 1.26 (0.92–1.73) 1.14 (0.97–1.35) 1.25 (1.05–1.5)† 1.1 (0.81–1.49) Interactions with treatment Rate control ⫻ age ⬎75 yrs 0.93 (0.72–1.21) Amiodarone ⫻ age ⬎75 yrs 1.35 (1.08–1.69) Rate ⫻ thyroid disease 1.10 (0.80–1.51) Amiodarone ⫻ thyroid disease 1.92 (1.43–2.59) Values are hazard ratio (95% confidence interval). *p ⬍ 0.001; †p ⬍ 0.05; ‡p ⬍ 0.01.
Saksena et al.
JACC Vol. 58, No. 19, 2011 CV Outcomes of AADs in the AFFIRM Trial November 1, 2011:1975– 85 Relationship Between Time Dependent Changes in Clinical Status and Risk of CVH Relationship Between Time Dependent Changes in Clinical Status and Risk of CVH Amiodarone vs. Rate Class 1C vs. Rate 1.87 (1.40–2.50) 1.76 (1.29–2.41) 1.11 (0.64–1.94) NYHA functional class 1.82 (1.45–2.29) 1.35 (1.00–1.82) 2.17 (1.30–3.63) 2.28 (1.78–2.93) 1.81 (1.19–2.77) 1.95 (1.00–3.82) 3.51 (2.42–5.09) 3.72 (2.19–6.33) 4.23 (1.49–12.04) 7.44 (3.42–16.20) 15.61 (4.65–52.47) 22.45 (6.01–83.82) Increase in NYHA functional class 1.72 (1.35–2.20) 1.98 (1.39–2.83) 1.25 (0.67–2.34) 2.20 (1.68–2.89) 1.26 (0.82–1.92) 2.58 (1.28–5.19) 3.57 (2.40–5.30) 1.62 (0.87–3.01) 5.42 (2.09–14.01) 3.73 (1.61–8.64) 2.20 (0.64–7.48) 6.37 (1.19–34.01) 4.08 (1.29–12.88) 1.96 (0.45–8.61) 28.74 (3.10–266.46) 1.25 (0.87–1.80) 2.35 (1.40–3.92) 0.90 (0.36–2.22) 1.13 (1.04–1.24) 1.10 (1.00–1.21) 0.99 (0.84–1.16) Increase in VR by ⱖ15 beats/min 1.25 (0.96–1.64) 1.58 (1.20–2.07) 1.62 (1.04–2.51) Values are hazard ratio (HR) (95% confidence interval). HR for ventricular rate (VR) is the increase in risk associated with a 15-beat/min increase in VR.
CHC ⫽ Canadian Heart Association classification for angina pectoris; CVH ⫽ cardiovascular hospital stay; Sota ⫽ sotalol; SR ⫽ sinus rhythm; other abbreviations as in suggesting that the excess CVH seen with these drugs were CVH in AF are costly, with average costs less serious events than those seen with amiodarone.
estimated to exceed $12,000/AF admission in the United CVH was extremely common with AF therapies in the
States and $3 billion in annual costs Atrial fibrillation AFFIRM trial. From our data, we can estimate overall
hospital stays are widely assumed to be related to AF CVH risk for AF populations and its relation to therapy recurrences, but such an assumption has neither been selection during the period 1995 to 2001. Cardiovascular critically verified and quantified, nor has the uniformity of hospital stay incidence ranged from 36% to 50% at 3 years this risk been assessed across AF subpopulations or treat- for rate and rhythm therapies. Cardiovascular hospital stay rates in the AFFIRM Rate subgroups were similar to To date, small trials of nonpharmacologic therapies those seen in the placebo (rate control therapies only) arm and 1 large pharmacologic therapy trial have provided of the ATHENA (A placebo-controlled, double-blind, some information about CVH in AF Anal- parallel-arm Trial to assess the efficacy of dronedarone ysis of the AFFIRM database provides important addi- 400 mg BID for the prevention of cardiovascular Hospi- tional data from a large randomized controlled trial over talization or death from any cause in patiENts with Atrial a long follow-up. CVH presaged mortality, but it was fibrillation/atrial flutter) trial (36.3% at 2.5 years) unclear how these events related to treatment strategy Clinical characteristics and initial AAD selection rather
and clinical condition Given the observations with than treatment strategy influenced CVH risk. Potential
respect to ICU hospital stays, CVH are usually related to mechanisms proposed for increased CVH include hospi- serious morbidity, with treatment strategy-related hospi- tal stays related to change in AAD therapy with associ- tal stays—such as for a change of drug therapy or for ated cardioversion or possible higher warfarin discontin- cardioversion— being a relatively small component. Ex- uation rates with potential complications Ouranalysis of CVH related solely to cardioversions for the cess CVH events observed with the AADs evaluated are rhythm control strategy, although higher than in associated with age, sex, and comorbidity status. There is matched Rate cohorts, demonstrated a fairly low inci- a residual excess CVH risk even after adjustment for dence in all AAD cohorts. Stoke, embolism, and major these historical factors, which is related to AAD use.
bleeds also had a low incidence that was comparable in Additionally, CVH risk can be related to changes in the matched Rate cohorts. Longer hospital stays, ICU cardiovascular disease status longitudinally. Time- stays, and other CV procedures constituted the bulk of dependent changes that impact risk can include either AF CVH, suggesting more serious clinical conditions. Dif- relapses or worsening of major cardiovascular symptoms ferences in CVH rates persisted across clinically impor- of the underlying disease. We propose, on the basis of our tant subgroups, such as elderly persons, women, and analysis, that both baseline patient characteristics and coronary disease patients.
time-dependent changes in clinical status contribute to CVH in AF: insights from the AFFIRM trial. CVH has
CVH risk. Any heart failure or coronary disease was become a major endpoint for clinical trials. It can impact associated with increased risk in all 3 matched cohorts treatment strategy recommendations and regulatory ap- but was more common in the amiodarone and matched proval of new therapies but is rarely used in AF trials Rate cohorts. An increase in heart failure or angina class JACC Vol. 58, No. 19, 2011 Saksena et al.
November 1, 2011:1975– 85 CV Outcomes of AADs in the AFFIRM Trial by 1 or more increased risk of CVH. These findings make 5. The AFFIRM Investigators. A comparison of rate control and rhythm a strong case for baseline disease state variables and control in patients with atrial fibrillation. N Engl J Med 2002;347:1825–33.
change in clinical status leading to CVH.
6. Corley SD, Epstein AE, DiMarco JP, et al., AFFIRM Investigators.
Relapse from sinus rhythm to AF was also related to Relationships between sinus rhythm, treatment, and survival in the CVH, suggesting failure of rhythm control as a potential Atrial Fibrillation Follow-Up Investigation of Rhythm Management(AFFIRM) Study. Circulation 2004;109:1509 –13.
mechanism. Finally, specific interactions of antiarrhythmic 7. Rosenbaum PR, Rubin DB. The central role of the propensity score in agents such as amiodarone with comorbidities such as observational studies for causal effects. Biometrika 1983;41–55.
thyroid disease suggest additional mechanisms leading to 8. Freemantle N, Calvert M, Wood J, Eastaugh J, Griffin C. Composite outcomes in randomized trials— greater precision with greater uncer- hospital stay. The reasons for CVH are multiple and tainty. JAMA 2003;289:2554 –9.
multifactorial. Atrial fibrillation patients have varying risk 9. AFFIRM First Antiarrhythmic Drug Substudy Investigators. Main- for the principal outcome in this analysis on the basis of tenance of sinus rhythm in patients with atrial fibrillation: an AFFIRM sub these factors.
study of the first antiarrhythmic drug. J Am Coll Cardiol 2003;42:20 –9.
10. Akaike H. A new look at the statistical model identification. IEEE Trans on Automatic Control 1974;19:716 –23.
Propensity score matching cannot correct for erroneous 11. Sherman DG, Kim SG, Boop BS, et al., National Heart, Lung, and Blood Institute AFFIRM Investigators. Occurrence and characteris- omission or inclusion of variables that might have affected tics of stroke events in the Atrial Fibrillation Follow-up Investigation AAD selection, but it is a significant improvement over of Sinus Rhythm Management (AFFIRM) study. Arch Intern Med naïve subgroup analyses. Some of the hospital stays might 12. Wyse DG, Slee A, Epstein AE, et al. Alternative endpoints for be the result of routine patient care for rhythm control mortality in studies of patients with atrial fibrillation: the AFFIRM rather than for medical necessity, but these still occur in study experience. Heart Rhythm 2004;1:531–7.
current clinical practice. The AFFIRM study did not 13. Jenkins LS, Brodsky M, Schron E, et al. Quality of life in atrial fibrillation: the Atrial Fibrillation Follow-up Investigation of Rhythm capture detailed reasons for hospital stay or drug doses.
Management (AFFIRM) study. Am Heart J 2005;149:112–20.
Exact dates of hospital stay were not collected, which results 14. Steinberg JS, Sadaniantz A, Kron J, et al. Analysis of cause-specific in decreased precision in estimates of time to hospital stay mortality in the Atrial Fibrillation Follow-up Investigation of RhythmManagement (AFFIRM) study. Circulation 2004;109:1973– 80.
but probably not for the comparison of matched cohorts.
15. Hohnloser SH, Crijns HJ, van Eickels M, et al., ATHENA Investi- gators. Effect of dronedarone on cardiovascular events in atrial fibril- lation. N Engl J Med 2009;360:668 –78.
16. Wilber DJ, Pappone C, Neuzil P, et al., ThermoCool AF Trial CV hospitalizations were common in AFFIRM with both Investigators. Comparison of antiarrhythmic drug therapy and radio-frequency catheter ablation in patients with paroxysmal atrial fibrilla- treatment strategies but more frequent with amiodarone, tion: a randomized controlled trial. JAMA 2010;303:333– 40.
sotalol and class 1C agents. The severity of this risk varied 17. Bosco JL, Silliman RA, Thwin SS, et al. A most stubborn bias: no with the individual AAD, patient characteristics and time adjustment method fully resolves confounding by indication in obser-vational studies. J Clin Epidemiol 2010;63:64 –74.
dependent changes in clinical status, but was largely unre- 18. Wazni OM, Marrouche NF, Martin DO, et al. Radiofrequency lated to the rhythm treatment algorithm. Death, intensive ablation vs antiarrhythmic drugs as first-line treatment of symptomatic care unit hospital stay, and non-CV death were more atrial fibrillation: a randomized trial. JAMA 2005;293:2634 – 40.
frequent with amiodarone.
19. Stabile G, Bertaglia E, Senatore G, et al. Catheter ablation treatment in patients with drug-refractory atrial fibrillation: a prospective, multi-centre, randomized, controlled study (Catheter Ablation for the Cure Reprint requests and correspondence: Dr. Sanjeev Saksena,
of Atrial Fibrillation Study). Eur Heart J 2006;27:216 –21.
20. Rao BH, Saksena S. Impact of "hybrid therapy" on long-term rhythm RWJ Medical School, Department of Medicine, 161 Washington control and arrhythmia related hospitalizations in patients with drug- Valley Road, Suite 201, Warren, New Jersey 07059. E-mail: refractory persistent and permanent atrial fibrillation. J Interv Card 21. Naccarelli G, Johnston S, Lin J, Patel P, Schulman K. Cost burden of cardiovascular hospitalization and mortality in ATHENA-like pa- tients with atrial fibrillation/atrial flutter in the United States clinicalcardiology 2010;33:270 –9.
1. Benjamin EJ, Wolf PA, D'Agostino RB, Silbershatz H, Kannel WB, 22. Khairallah F, Ezzedine R, Ganz LI, London B, Saba S. Epidemiology Levy D Impact of atrial fibrillation on the risk of death: The and determinants of outcome of admissions for atrial fibrillation in the Framingham Heart Study. Circulation 1998;98:946 –52.
United States from 1996 to 2001. Am J Cardiol 2004;94:500 – 4.
2. AF Hospitalizations in the USA, June 2009. Available at: chart 3-47& 3-48. Accessed September 13, 2011.
Key Words: antiarrhythmic drugs y atrial fibrillation y cardiovascular
3. Miyasaka Y, Barnes ME, Gersh BJ, et al. Changing trends of hospital hospitalizations y cardiovascular outcomes y clinical trials y outcomes utilization in patients after their first episode of atrial fibrillation. Am J Cardiol 2008;102:568 –72.
4. The Planning and Steering Committees of the AFFIRM Study for the NHLBI AFFIRM Investigators. Atrial fibrillation follow-up investi-gation of rhythm management—the AFFIRM study design. Am J For list of investigators and affiliated institutions, Cardiol 1997;79:1198 –202.
please see the online version of this article.

Source: http://www.eprf.org/downloads/AFFIRM_final_published_paper.pdf


Trends in the Abuse of Prescription Drugs by Jane Carlisle Maxwell, Ph.D. The sale of narcotic analgesic pills is increasing, as is diversion and the non-medical use of prescription drugs. These drugs are easy to obtain and they are viewed as "safer" than street drugs. Young adults have the highest rates of lifetime use of these drugs and fewer teenagers in 2005


Preventive journalism A media and coverage of professional's risk situations avian influenza ANDI ANDI LATIN AMERICAN NETWORK UNICEF Preventive journalism A media and coverage of professional's risk situations avian influenza ANDI – BRAZILIAN NEWS AGENCY FOR PREVENTIVE JOURNALISM AND COVERIAGE OF RISK SITUATIONS