Risk of Initiating Antiarrhythmic Drug Therapy for Atrial Fibrillation in Patients Admitted to a University Hospital

Methods

Patients

The medical records of all patients with a diagnosis of atrial fibrillation or atrial flutter who were admitted to Brigham and Women's Hospital between July 1990 and June 1994 were evaluated for study inclusion. Patients were included if they 1) were older than 18 years of age, 2) had documented atrial fibrillation or atrial flutter on an electrocardiogram, and 3) underwent initiation of antiarrhythmic drug therapy for conversion of atrial fibrillation or atrial flutter to sinus rhythm or for maintenance of sinus rhythm after cardioversion. Exclusion criteria were 1) acute illness, such as acute myocardial infarction, at the time of admission; 2) cardiac surgery within 30 days of admission; 3) previous heart transplantation; or 4) amiodarone treatment within 1 year of admission. These criteria were met by 417 patients who underwent a total of 597 drug trials during a total of 550 hospitalizations (Table 1). The mean age of these patients (±SD) was 64 ± 13 years, and the median duration of atrial fibrillation before admission was 12 months (range, 0 to 292 months).

Definitions

Cardiac adverse events were predefined to include ventricular arrhythmias, including torsade de pointes (polymorphic ventricular tachycardia associated with prolongation of the QT interval); bradyarrhythmias (heart rate < 60 beats/min); prolongation of the QT interval; conduction abnormalities, including prolongation of the PR interval and QRS complex widening; rapid ventricular rate; congestive heart failure; hypotension [systolic blood pressure < 90 mm Hg]; and cerebrovascular emboli. With the exception of cerebrovascular emboli, these occurrences were considered to be cardiac adverse events only if they resulted in 1) discontinuation or dose reduction of therapy with antiarrhythmic medication or 2) medical intervention to allow continued administration of the drug (for example, pacemaker insertion or discontinuation of therapy with nodal blocking agents).

The antiarrhythmic drug dose and the decision about when to discharge a patient from the hospital were at the discretion of the treating physician. The most common initial dosing regimens were quinidine, 200 mg orally every 2 hours for 3 doses and then 324 mg orally three times daily; procainamide, 500 mg orally four times daily; propafenone, 150 mg orally three times daily; or sotalol, 80 mg orally twice daily. In general, anticoagulants were administered whenever the duration of arrhythmia exceeded 24 to 48 hours and no contraindication to anticoagulation was present. Overall, anticoagulants were given during 82% of drug trials (warfarin or heparin in 70% and aspirin alone in 12%). Electrical conversion was performed in 247 trials when patients did not convert to sinus rhythm with drug therapy alone. For each drug trial, the observation period was the time between administration of the first dose of antiarrhythmic medication and the occurrence of one of the following: discontinuation of drug therapy, a cardiac adverse event, or hospital discharge.

Statistical Analysis

Five potential risk factors for cardiac adverse events (age, sex, previous myocardial infarction, previous congestive heart failure, and structural heart disease) that were defined a priori were entered into a multivariate logistic regression model so that we could assess their relation to adverse events. Generalized estimating equations were used to adjust for multiple hospitalizations of individual patients [11]. We estimated the daily hazard rate for the first 3 days of therapy in order to assess the change in risk over time [12].

Results

Cardiac Adverse Events

Eighty cardiac adverse events occurred in 73 patients during the 597 drug trials, for an overall incidence of 13.4% (Table 2). The greatest risk for cardiac adverse effects was seen during the first 24 hours of therapy: Seven adverse events per 100 patient-days occurred during this period compared with 3.8 adverse events per 100 patient-days during the second 24-hour period and 3.3 adverse events per 100 patient-days during the third 24-hour period. Bradyarrhythmias were the most common adverse event: They occurred in 47 drug trials (7.9%) and represented 59% of all adverse events. Of the bradyarrhythmias, 32 (68%) were cases of sinus bradycardia, 12 (26%) were cases of atrial fibrillation with a slow ventricular response, and 3 (6%) were cases of junctional rhythms. Twenty-five of the 47 bradyarrhythmias (53%) were classified as severe (causing symptoms, resulting in pacemaker insertion, requiring intravenous administration of atropine or β-adrenergic agonists, or resulting in transfer of the patient to an intensive care unit). Although the remaining bradyarrhythmias were asymptomatic, therapy in these cases was altered because the treating physician was concerned that progression to more severe bradycardia might occur. A comparison between drug trials that did and did not include bradyarrhythmias showed no difference in the use of digoxin (49% compared with 44%; P > 0.2), β-blockers (36% compared with 32%; P > 0.2), or calcium-channel blockers (36% compared with 26%; P = 0.13). Similarly, no differences in the use of these drugs were found between trials that did and did not have severe bradyarrhythmias.

Ventricular arrhythmias complicated eight drug trials (1.3%). Seven of these arrhythmias were cases of nonsustained monomorphic ventricular tachycardia. Torsade de pointes occurred in only one drug trial (0.2%) after five doses of quinidine, 5 hours after cardioversion to sinus rhythm (serum potassium concentration. 3.7 mmol/L; QTc interval, 650 ms). Prolongation of the QT interval (>500 ms in all cases) caused the discontinuation of therapy in nine trials (1.5%). No other medications known to prolong the QT interval had been administered [13]. A serum potassium concentration less than 3.5 mmol/L (in 5% of drug trials) or a serum magnesium concentration less than 1.8 mmol/L (in 17% of drug trials) measured at the time of admission was not predictive of ventricular arrhythmias or adverse events as a whole. Cerebrovascular emboli, confirmed by computed tomography, occurred in two drug trials (0.3%).

Electrical Conversion

Electrical conversion was attempted in 247 of the 597 drug trials (41.4%) and successfully restored sinus rhythm in 208 (84.2%). The incidence of adverse events was similar in drug trials that did and did not use cardioversion (11.3% compared with 14.9%; P > 0.2). Of the 28 adverse events that followed electrical conversion, 14 (50.0%) occurred within 2 hours and 26 (92.9%) occurred within 8 hours of the procedure; 21 (75.0%) were bradyarrhythmias. Adverse events were more likely to occur after successful cardioversion to sinus rhythm than after unsuccessful cardioversion (13.0% compared with 2.6%; P = 0.04).

Predictors of Cardiac Adverse Events

On multivariate logistic regression analysis, patients with previous myocardial infarction had a significantly increased risk for adverse events (odds ratio, 1.90 [95% CI, 1.05 to 3.43]; P = 0.03). A trend toward increased risk was seen with older age (odds ratio, 1.29 per decade [CI, 0.97 to 1.72]; P = 0.08) but not with sex, structural heart disease, or previous heart failure.

Discussion

The 13.4% incidence of early cardiac adverse events during initiation of antiarrhythmic therapy seems to support the in-hospital initiation of antiarrhythmic therapy for conversion of atrial fibrillation to sinus rhythm in certain patients. The greatest risk for cardiac adverse events is during the first 24 hours of drug therapy and after successful cardioversion. Bradyarrhythmias are the most common life-threatening adverse events.

Previous myocardial infarction increases the risk for cardiac adverse events, as was previously reported during long-term antiarrhythmic therapy [7, 14]. In previous studies, older age [14] and congestive heart failure [7, 14] were also associated with adverse events during antiarrhythmic therapy. In our study, however, the association between older age and risk for cardiac adverse events was of borderline significance and heart failure was not a predictor of cardiac adverse events, probably because of the small numbers of elderly patients and patients with heart failure.

For certain drugs, in-hospital initiation of therapy has been advocated to detect early proarrhythmia, particularly torsade de pointes [10], the reported incidence of which varies from 1% to 8% [15, 16]. This type of arrhythmia occurred in only one drug trial (0.2%) in our study. Careful attention to electrolyte repletion and discontinuation of drug therapy for prolongation of the QT interval probably account for the low incidence of this event.

Although ours was a retrospective study done at a single site, the prespecified adverse events were not ambiguous and were easily determined. Drug administration was not randomly assigned, and the dosing regimens varied. Therefore, comparison of the risks among individual drugs could have been misleading and was not done. The observation period was restricted to the period of hospitalization, and no out-of-hospital group was included for comparison. Because 90% of the patients in our study had structural heart disease, our findings may not be applicable to patients with “lone” atrial fibrillation.

In conclusion, a significant risk for cardiac adverse events exists during the initiation of antiarrhythmic therapy for atrial fibrillation. Close observation with electrocardiographic monitoring is warranted, particularly for patients who have previously had myocardial infarction, elderly patients, and patients undergoing electrical conversion. The optimal duration of close observation during the initiation of amiodarone therapy, which has potential bradycardic effects and a long half-life, remains to be defined. An observation period of 24 to 48 hours is probably sufficient for the shorter-acting antiarrhythmic drugs.

Drs. Kuntz and Lee: The Section for Clinical Epidemiology, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115.