Propafenone Treatment of Symptomatic Paroxysmal Supraventricular Arrhythmias A Randomized, Placebo-Controlled, Crossover Trial in Patients Tolerating Oral Therapy Edward L.C. Pritchett, MD; Elizabeth A. McCarthy, RN; and William E. Wilkinson, PhD
Objective: To test the hypothesis that propafenone, administered orally, prevents symptomatic paroxysmal supraventricular arrhythmias. Design: A 6-month, open-label, dose-finding phase followed by a randomized, double-blind, placebo-controlled, crossover phase, with each treatment period lasting up to 60 days. Setting: An outpatient clinic. Patients: Thirty-three patients with either paroxysmal supraventricular tachycardia (n = 16) or paroxysmal atrial fibrillation (n = 17) were enrolled. Their arrhythmias were documented by electrocardiogram before enrollment. Twenty-three patients (14 with paroxysmal supraventricular tachycardia and 9 with paroxysmal atrial fibrillation) were randomized and the data obtained from these patients were used in the efficacy analysis. Intervention: Propafenone (300 mg three times daily in 19 patients, 300 mg twice daily in 3 patients, and 150 mg twice daily in 1 patient) and matching placebo tablets were administered in a randomized sequence. Measurements: Symptomatic arrhythmia was documented by telephone transmission of the electrocardiogram. Main Results: The time to first recurrence was prolonged for the overall group of 23 patients while they received propafenone (P = 0.004). The recurrence rate of arrhythmia during treatment with propafenone was estimated to be approximately one fifth of the recurrence rate during treatment with placebo. Conclusions: Propafenone is effective in reducing symptomatic paroxysmal supraventricular arrhythmias. Annals of Internal Medicine. 1991;114:539-544. From the Duke University Medical Center, Durham, North Carolina. For current author addresses, see end of text.
Advances in antiarrhythmic drug therapy for treating paroxysmal supraventricular arrhythmias have been impeded by the difficulty of establishing efficacy in adequately controlled clinical trials. A new antiarrhythmic drug, propafenone hydrochloride (hereafter referred to as propafenone), was approved recently by the United States Food and Drug Administration for oral administration to treat ventricular arrhythmias (1). The results of several uncontrolled studies have suggested that oral propafenone also may be useful in the management of patients with paroxysmal supraventricular arrhythmias
(2-7); however, only one placebo-controlled clinical trial (conducted in patients with paroxysmal atrial fibrillation) has been reported (8). Our rigorous, placebo-controlled trial was intended to test the efficacy of oral propafenone in the prophylaxis of recurrent symptomatic arrhythmias in patients with paroxysmal supraventricular tachycardia or paroxysmal atrial fibrillation.
Patients and Methods Patient Sample Patients with symptomatic paroxysmal supraventricular arrhythmias were recruited from the investigators' clinic between 1984 and 1989, because their arrhythmia was incompletely controlled by their current therapy. At enrollment, each patient's arrhythmia was classified on the basis of the electrocardiograms that were available at the time of referral. Paroxysmal supraventricular tachycardia was diagnosed if the electrocardiograms showed a mean rate greater than 120 beats/min, QRS-complex morphology during tachycardia that either was normal or indicated a functional bundle-branch block, less than 0.02 s of variation in successive cardiac cycles, no evidence of atrioventricular dissociation, and episodic occurrence. Paroxysmal atrial fibrillation was diagnosed if the electrocardiograms showed a mean rate greater than 100 beats/min, QRS-complex morphology during tachycardia that either was normal or indicated a functional bundle-branch block, a grossly irregular ventricular rhythm, the absence of P waves or the presence of fibrillatory waves in the baseline, and episodic occurrence. The arrhythmia was documented by an electrocardiogram that was recorded during the manifestation of symptoms in all patients. Patients who were diagnosed as having paroxysmal atrial fibrillation were allowed to take digoxin during the study if the dose was stable for 3 weeks before treatment began and did not vary during the study; all other antiarrhythmic drugs, beta adrenergic receptor blockers, and calcium-channel blockers were prohibited. Patients were excluded if paroxysms of arrhythmia precipitated angina, pulmonary edema, or neurologic symptoms. Patients with paroxysmal atrial fibrillation were excluded if they had the Wolff-Parkinson-White syndrome. Methods Recording Methods Throughout the study, patients used a small, battery-operated electrocardiogram recorder (Cardiobeeper Memory Monitor, Survival Technology, Bethesda, Maryland) to document their rhythm when symptoms were manifested and to play back the electrocardiogram to the investigators using a toll-free automatic telephone answering system. In addition, patients were contacted by telephone at 2-week intervals to encourage compliance with the study and to verify equipment function by transmitting an electrocardiogram during sinus rhythm. ©1991 American College of Physicians
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
539
Dose-Finding Phase The overall study design followed principles suggested for the study of symptomatic paroxysmal arrhythmias (9). Patients with one of the qualifying arrhythmias began the dose-finding phase of the study by taking propafenone, 300 mg twice daily. The dose of propafenone was increased or decreased depending on the recurrence of arrhythmias or the occurrence of side effects. If the recurrence of an arrhythmia was documented using the telephone monitor system, the propafenone dose was increased to 300 mg three times daily. If no recurrence of arrhythmia was documented within 90 days of beginning therapy with propafenone, 300 mg twice daily, the dose was arbitrarily increased to 300 mg three times daily. The purpose of this arbitrary dose increase was to find the highest tolerated dose in each patient during the dose-finding phase. To compare the efficacy of taking 300 mg twice daily with that of taking 300 mg three times daily in the dose-finding phase, we used only data obtained from those patients whose plasma drug concentrations were presumed to have reached steady-state (after 3 days of therapy) on both regimens. For each patient, the dose-finding phase could last as many as 180 days. Patients returned to a research arrhythmia clinic every 2 weeks for the first 3 months and at monthly intervals thereafter. At these visits, a physical examination was done, electrocardiograms were recorded, and blood samples were drawn for laboratory screening and for measuring plasma propafenone concentrations. Randomized Phase Patients who successfully completed the dose-finding phase entered a randomized, double-blind, placebo-controlled, crossover phase. At the beginning of this phase, patients were provided with double-blinded medication labeled "drug A." The tablets looked like the medication they had used at the end of the dose-finding phase, and they contained either placebo or propafenone. The first observation period began 3 days after beginning treatment with drug A to allow propafenone to wash out during placebo treatment periods. Patients were then followed until they either had their first recurrence of arrhythmia while receiving drug A or had completed an additional 60 days of therapy with drug A, whichever came first. Patients were then treated in an open-label fashion with propafenone for 3 to 7 days. Next, they were provided in a blinded fashion with an additional medication labeled "drug B" whose formulation was the opposite of drug A's. As with drug A, the observation period for drug B began 3 days after beginning treatment with drug B. The randomized phase concluded at the patient's first recurrence of arrhythmia while receiving drug B or after 60 days of therapy with drug B; the maximum possible length of the randomized phase, therefore, was 4 months. Patients were seen in the clinic at monthly intervals and between treatments with drugs A and B. Adverse Experiences Throughout the dose-finding and randomized phases of the study, both cardiac and noncardiac adverse experiences were tabulated, regardless of whether they were thought to be due to propafenone. These experiences were tabulated by drug dose even if the patient's plasma concentrations had not reached steady-state while the patient received that dose. Particular attention was given to adverse experiences that caused patients to discontinue their participation in the study prematurely. Plasma Drug Concentrations During the dose-finding phase of the study, blood samples were drawn from patients to measure plasma propafenone concentrations using the method of Harapat and Kates (10). The clock time and the time since the administration of the preceding dose of propafenone were recorded for each sample; in general, blood samples were not drawn at trough times. Long-Term Follow-up Patients who successfully completed the randomized portion of the trial were allowed to continue taking medication in an open-label fashion in an extension of the protocol until pro540
pafenone was marketed. Patients were seen in the clinic every 3 months during this phase of the study. Long-term follow-up was discontinued for all patients by June 1990. Data Analysis We compared demographic data obtained from patients who qualified with paroxysmal supraventricular tachycardia with data obtained from patients who qualified with paroxysmal atrial fibrillation; we also compared data obtained from patients included in with data obtained from patients excluded from the randomized phase. These comparisons were done using the Fisher exact test (two-sided) for dichotomous variables and the Wilcoxon rank-sum test (two-sided) for age. We used the Kaplan-Meier product-limit method to estimate and to illustrate graphically the proportion of patients remaining free of arrhythmia on each day of the various follow-up periods (11). The formal statistical comparison of propafenone treatment with placebo was done using data from the randomized phase; we compared the time to the first recurrence of arrhythmia while patients received placebo with that while patients received propafenone using a one-sided randomization test based on the paired Prentice-Wilcoxon statistic for censored data (12). Assuming that the successive occurrences of arrhythmia constitute a Poisson process (13), the rate of occurrence of these attacks is constant for each patient and the effect of propafenone on this rate can be estimated using the proportional hazards model. Using this model, the hazard function hi (t) for the ith subject was modeled as ajh0 (t) for an arbitrary, unspecified hazard function h0 (t) so that the hazard functions among subjects differed at most by multiplicative constants (14). In the Poisson model, the hazard ratio is simply the ratio of the rates of occurrence in the underlying Poisson processes; an estimate of the hazard ratio is therefore an estimate of the effect of propafenone on the rate at which symptomatic episodes of paroxysmal supraventricular arrhythmia occur. We also compared the heart rates during attacks in each treatment period using the Wilcoxon signed-rank test and a 95% confidence interval (CI) for the median difference based on the Wilcoxon statistic (15). Results Patient Sample Thirty-three patients with qualifying arrhythmias (16 with paroxysmal supraventricular tachycardia and 17 with paroxysmal atrial fibrillation) entered the study. The qualifying arrhythmia was documented by 12-lead electrocardiography in 32 patients and by transtelephonic monitoring in the other patient. Although the patients who had paroxysmal atrial fibrillation as their qualifying arrhythmia were somewhat older and more likely to have heart disease than were the patients who had paroxysmal supraventricular tachycardia, none of the comparisons of demographic variables between the two samples (P > 0.05) were significantly different (Table 1). Treatment with digoxin was continued in 5 patients with paroxysmal atrial fibrillation (2 of whom entered the randomized phase) who were taking it at the time of referral. Of the 33 patients who qualified and entered the dose-finding phase of the study, 23 completed the dose-finding phase and entered the randomized phase. The 10 patients who did not enter the randomized phase were somewhat older and more likely to have heart disease than were the 23 other patients, but none of the differences was significant (P > 0.05) (Table 2). Patients with paroxysmal supraventricular tachycardia were marginally more likely to enter the randomized phase than were patients with paroxysmal atrial fibril-
1 April 1991 • Annals of Internal Medicine • Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
Table 1. Characteristics of Patients with Paroxysmal Supraventricular Tachycardia and Patients with Paroxysmal Atrial Fibrillation* Characteristic
Paroxysmal Supraventricular Tachycardia (n = 16)
Paroxysmal Atrial Fibrillation (n = 17)
n(%) Race White Black Sex Men Women Heart disease Lung disease Randomized
14 (87.5) 2 (12.5)
17 (100.0) 0 (0.0)
8 (50.0) 8 (50.0) 5(31.2) 4 (25.0) 14 (87.5)
11 (64.7) 6 (35.3) 9 (52.9) 2(11.8) 9(47.1)
* For patients with paroxysmal supraventricular tachycardia, the median age was 51.8 years (interquartile range, 12.9 years); for patients with paroxysmal atrial fibrillation, the median age was; 63.1 years (interquartile range, 21.4 years).
lation (14 of 16 patients compared with 9 of 17 patients; P = 0.06). Dose-Finding Phase Thirty-two patients began the dose-finding phase of the study receiving propafenone, 300 mg twice daily, and one patient (with paroxysmal atrial fibrillation) began this phase receiving propafenone, 300 mg three times daily. Of the 32 patients who began the study receiving 300 mg twice daily (13 with paroxysmal supraventricular tachycardia and 10 with paroxysmal atrial fibrillation), 23 also completed at least 3 days of therapy with 300 mg three times daily. The data obtained from these 23 patients were used to compare the efficacy of these two dosages (Figure 1). Figure 1 shows the time to the first recurrence of arrhythmia for these 23 patients while they received each of the two dosages. The time to the first recurrence of arrhythmia appeared to be longer with the higher dosage; however using a statistical test was not considered to be appropriate, because the study was not designed to compare the efficacy of these two dosages. Randomized Phase
recurrence of arrhythmia is graphically illustrated in Figure 2. On the basis of the proportional hazards model, the hazard ratio was estimated to be 0.21; (95% CI, 0.08 to 0.58). Assuming that successive occurrences of arrhythmia constitute a Poisson process, the rate of recurrence of arrhythmia during treatment with propafenone is estimated to be 0.21 times (or approximately one-fifth) the recurrence rate during treatment with placebo. For the subgroup of patients with paroxysmal supraventricular tachycardia (n = 14), the difference between propafenone and placebo was similar to the difference found with the total group; however, the significance of this comparison was weakened by the smaller number of patients (P = 0.03). In the even smaller subgroup with paroxysmal atrial fibrillation (/? = 9), the difference between propafenone and placebo again was similar to the difference found with the total group, but the significance of this comparison was still less (P = 0.06). For the 10 patients who had a recurrence of their arrhythmia while receiving propafenone and while receiving placebo in the randomized phase, the median difference in heart rate was not different from 0 (P = 0.2); the median difference (heart rate while receiving placebo minus heart rate while receiving propafenone) was 12 beats/min (CI, - 6 to 29 beats/min). Adverse Experiences Among the 32 patients who were exposed to propafenone, 300 mg twice daily, during the dose-finding phase, the most common noncardiac adverse experiences reported were bitter taste (7 patients), nausea (7 patients), anorexia (2 patients), and weakness (2 patients). Among the 27 patients exposed to the propafenone, 300 mg three times daily, the same experiences were noted by 4, 5, 0, and 2 patients, respectively. No other noncardiac adverse experience was reported by more than 1 patient receiving either dosage. Two patients discontinued participation in the study prematurely, because they had noncardiac adverse experiences: One patient had paresthesia of the Table 2. Characteristics of Patients Included in and Patients Excluded from the Randomized Phase* Characteristic
The propafenone dosage that was used in the randomized phase was the final dosage used for the 23 patients successfully completing the dose-finding phase: 300 mg three times daily in 19 patients, 300 mg twice daily in 3 patients, and 150 mg twice daily in 1 patient. One patient discontinued participation in the study after taking drug A (placebo) and before starting drug B (propafenone) because of a prolonged attack of paroxysmal atrial fibrillation. This patient was included in the efficacy analysis by conservatively assuming that he had a recurrence of arrhythmia on the first day of the propafenone observation period. Compared with placebo, propafenone caused an increase in the time to the first recurrence of arrhythmia (P = 0.004). The prolongation of the time to the first
Excluded Patients (/i = 10)
Included Patients (n = 23) n(%)
Race White
Black Sex Men Women Heart disease Lung disease The Wolflf-Parkinson-White syndrome
21 (91.3) 2 (8.7)
10 (100.0) 0 (0.0)
11 (47.8) 12 (52.2) 8 (34.8) 5(21.7)
8 (80.0) 2 (20.0) 6 (60.0) 1 (10.0)
2 (8.7)
0 (0.0)
* For patients included ini the analysis, the median age was 49.0 years (interquartile range, 16.6 years); for patients excluded from the analysis, the median age wasi 63.9 years (interquartile range, 14.4 years).
1 April 1991 • Annals of Internal Medicine
• Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
541
Long-Term Follow-up Of the 23 patients who entered the randomized phase, 21 also received long-term propafenone treatment in an open-label fashion. These 21 patients had been followed for a median of 2.9 years (range, 98 days to 5.6 years) when follow-up was discontinued in June 1990. Twenty of the twenty-one patients had at least one recurrence of symptomatic arrhythmia during long-term follow-up. Discussion Efficacy
Figure 1. The time to the first recurrence of symptomatic paroxysmal supraventricular arrhythmia in patients receiving propafenone, 300 mg twice daily ( ), and in patients receiving propafenone, 300 mg three times daily ( ). These data were obtained from 13 patients with paroxysmal supraventricular tachycardia and 10 patients with paroxysmal atrial fibrillation who received both dosages in the dose-finding phase of the study. BID = twice daily, TID = three times daily.
arms, and the other had anorexia and nausea. Both of these patients had paroxysmal supraventricular tachycardia. Cardiac adverse experiences were less common but resulted in nine premature discontinuations, all of them in patients with paroxysmal atrial fibrillation. Of nine patients, five (two receiving 300 mg twice daily and three receiving 300 mg three times daily) were thought to have had a prolonged episode of atrial fibrillation while receiving propafenone and discontinued their participation in the study prematurely for this reason. Another patient had atrial flutter with a mean ventricular rate of 263 beats/min recorded using the telephone monitor. One patient had chest pain, dyspnea, and presyncope, but was unable to transmit a satisfactory rhythm strip with his electrocardiogram monitor. Another patient had a nonsustained regular tachycardia with a wide QRS-complex morphology, and propafenone was withdrawn; this same arrhythmia was recorded months later when he was receiving no antiarrhythmic medication. Sinus pauses with ventricular asystole lasting 3.08 s caused 1 additional patient with a history of sinus node dysfunction to discontinue participation in the study. A total of 11 patients discontinued participation in the study prematurely due to adverse experiences; all were receiving propafenone in an open-label fashion when they withdrew. Of the 11 patients, 9 had paroxysmal atrialfibrillation,and 2 had paroxysmal supraventricular tachycardia.
Placebo-controlled clinical trials of oral antiarrhythmic therapy to reduce the rate of recurrence of paroxysmal supraventricular arrhythmias are a relatively recent innovation in the study of arrhythmias (9). To date, these trials have shown the beneficial effects of verapamil for treating paroxysmal supraventricular tachycardia (16-18), of propafenone for treating paroxysmal atrial fibrillation (8), of flecainide for treating paroxysmal atrialfibrillation(19), and of flecainide for treating paroxysmal supraventricular tachycardia (20). In our rigorously controlled clinical trial, propafenone (compared with a placebo) reduced the rate of symptomatic recurrence of paroxysmal supraventricular arrhythmias (P = 0.004); the treatment effect was similar in the subgroups with paroxysmal supraventricular tachycardia and paroxysmal atrialfibrillation.The rate of recurrence while patients received propafenone was estimated to be approximately one-fifth the rate seen while patients received placebo. Propafenone therefore appears to be effective in treating these frustrating clinical problems. In addition to reducing the rate of attacks, propafenone may also reduce the heart rate during the recurrence of symptomatic supraventricular arrhythmia. Although the median reduction in heart rate while patients received propafenone was not significantly different from zero (P > 0.05), the number of-patients
Plasma Propafenone Concentrations Not surprisingly, a wide range of plasma propafenone concentrations was found in the samples drawn from patients at follow-up visits. Figure 3 shows concentrations and the time since the preceding dose was administered for all samples with concentrations in the detectable range. 542
Figure 2. The time to the first recurrence of symptomatic paroxysmal supraventricular arrhythmia in patients receiving placebo ( ) or propafenone ( ). These data were obtained from 14 patients with paroxysmal supraventricular tachycardia and 9 patients with paroxysmal atrial fibrillation who were studied in the randomized phase of the trial. Plac. = placebo, Prop. = propafenone.
1 April 1991 • Annals of Internal Medicine • Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
Figure 3. Plasma propafenone concentrations for patients taking propafenone, 300 mg twice daily (top) and 300 mg three times daily (bottom). The range of values is very wide, and there is substantial overlap among the concentrations measured with the two dosages.
having recurrences while receiving propafenone and while receiving placebo was small (n = 10). The CI for the median difference in the rate of attacks between propafenone and placebo periods was consistent with a minor propafenone effect: The upper limit of the CI corresponds to a 15% reduction in heart rate with propafenone. Adverse Experiences A substantial number of patients (11 of 33 patients) discontinued participation in our study prematurely because of adverse experiences. A disproportionate number of these premature discontinuations (9 of the 11) occurred in patients with paroxysmal atrial fibrillation (17 of 33 patients enrolled). Most premature discontinuations were due to cardiac adverse events, particularly arrhythmias. Several patients thought that propafenone lengthened their attacks of atrial fibrillation, and other rhythms were recorded, including tachycardias and bradycardias. Because our study did not use a parallel placebo group during dose-finding, we cannot know whether these events were drug effects or whether they were changes in the patients' arrhythmia behavior. For example, 1 of our patients who withdrew prematurely because of the occurrence of a tachycardia with wide
QRS-complex morphology was later found to have the same arrhythmia when he was receiving no antiarrhythmic drug at all. In addition, Weiner and colleagues (21) showed that other tachycardias were commonly documented by the telephone monitor in patients with paroxysmal atrialfibrillation.In summary, we are reluctant to ascribe all the cardiac adverse experiences to propafenone, because some of them may simply have been fortuitous changes in arrhythmia behavior. Our study design, in which some patients discontinued participation before receiving a placebo, limits the interpretation of adverse experiences (9). The sample of patients going through the randomized phase of the study is ' 'enriched'' by the elimination of patients who have intolerable side effects or who refuse to continue because the drug appears to fail during dose-finding. This study design contrasts with a parallel design in which patients are randomly assigned to receive placebo or one of several drug dosages. Compared with a parallel design, the design that we used probably overestimates efficacy at the price of exaggerating adverse experiences. Such a design may be particularly useful when treatment effects are weak or confined to a small subgroup of patients. The strength of the propafenone treatment effect that we found in this study was, in fact, much greater than we had expected on the basis of previous work with verapamil in this type of study (16). Verapamil reduced the rate of recurrent attacks of symptomatic paroxysmal supraventricular tachycardia to approximately one-half the rate achieved with placebo; in our current study, propafenone reduced the rate to one-fifth the rate achieved with placebo. Related Reports A large body of clinical literature consisting of uncontrolled clinical experiences suggests that oral propafenone is useful for treating supraventricular arrhythmias (2-7). Only one other controlled clinical trial, by Connolly and Hoffert (8), however, was done in patients with paroxysmal atrial fibrillation. In that study, patients kept a diary of days on which they had atrial fibrillation and sent transtelephonic rhythm strips to the investigators to confirm the diagnosis. Like our study, the Connolly and Hoffert study consisted of a dosefinding phase followed by a randomized, double-blind, placebo-controlled, crossover phase. The investigators found a significant reduction in the mean percentage of days with atrialfibrillation(from 51% ± 34% with placebo to 27% ± 34% with propafenone; P < 0.01). They also found that studying patients with paroxysmal atrial fibrillation was difficult: Of 20 patients who entered the study, 2 were later found to have been ineligible; 3 withdrew from the study during the dose-finding phase because of poor therapeutic response; 1 withdrew during the dose-finding phase because of poor compliance; 2 withdrew during the dose-finding phase because of intolerable side effects; 2 withdrew during the randomized phase because of poor therapeutic response; and 1 died during the randomized phase while receiving placebo. Only 9 of 20 patients completed the entire study, quite similar to our own experience in which only 8 of 17 patients with paroxysmal atrialfibrillationcompleted
1 April 1991 • Annals of Internal Medicine • Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
543
the study. Because placebo-controlled trials of paroxysmal atrial fibrillation are rare, whether the many premature withdrawals in this group were related to propafenone or to the type of patient who has paroxysmal atrial fibrillation was uncertain. In conclusion, the results of our trial confirm the efficacy of propafenone for treating both paroxysmal supraventricular tachycardia and paroxysmal atrial fibrillation in patients who tolerate therapy. The results of additional studies, accumulating clinical experience, and cost will be important factors in establishing the precise role for propafenone in antiarrhythmic practice. Grant Support: In part by a grant from Knoll Pharmaceutical Company and by grant ROl HL40392 from the National Heart, Lung and Blood Institute, Bethesda, Maryland, and grant M01 RR30 from the National Center for Research Resources, National Institutes of Health, Bethesda, Maryland. Current Author Addresses: Dr. Pritchett and Ms. McCarthy: Box 3477, Duke University Medical Center, Durham, NC 27710. Dr. Wilkinson: Box 2914, Duke University Medical Center, Durham, NC 27710. References 1. Funck-Brentano C, Kroemer HK, Lee JT, Roden DM. Propafenone. N Engl J Med. 1990;322:518-25. 2. Breithardt G, Borggrefe M, Wiebringhaus E, Seipel L. Effect of propafenone in the Wolff-Parkinson-White syndrome: electrophysiologic findings and long-term follow-up. Am J Cardiol. 1984;54:29D39D. 3. Ludmer PL, McGowan NE, Antman EM, Friedman PL. Efficacy of propafenone in Wolff-Parkinson-White syndrome: electrophysiologic findings and long-term follow-up. J Am Coll Cardiol. 1987;9:135763. 4. Kerr CR, Klein GJ, Axelson JE, Cooper JC. Propafenone for prevention of recurrent atrial fibrillation. Am J Cardiol. 1988;61:914-6. 5. Antman EM, Beamer AD, Cantillon C, McGowan N, Goldman L, Friedman PL. Long-term oral propafenone therapy for suppression of refractory symptomatic atrial fibrillation and atrial flutter. J Am Coll Cardiol. 1988;12:1005-11.
544
1 April 1991 • Annals
of Internal
Medicine
6. Hammill SC, Wood DL, Gersh BJ, Osborn MJ, Holmes DR Jr. Propafenone for paroxysmal atrial fibrillation. Am J Cardiol. 1988; 61:473-4. 7. Allen BJ, Brodsky MA, Doria R, Luckett CR, Thomas R, Henry WL. Oral propafenone therapy for patients with paroxysmal supraventricular tachyarrhythmia. Chest. 1988;94:853-4. 8. Connolly SJ, Hoffert DL. Usefulness of propafenone for recurrent paroxysmal atrial fibrillation. Am J Cardiol. 1989;63:817-9. 9. Pritchett EL, Lee KL. Designing clinical trials for paroxysmal atrial tachycardia and other paroxysmal arrhythmias. J Clin Epidemiol. 1988;41;851-8. 10. Harapat SR, Kates RE. High-performance liquid chromatographic analysis of propafenone in human plasma samples. J Chromatogr. 1982;230:448-53. 11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Statist Assoc. 1958;53:457-81. 12. O'Brien PC, Fleming TC. A paired Prentice-Wilcoxon test for censored paired data. Biometrics. 1987;43:169-80. 13. Pritchett EL, Smith MS, McCarthy EA, Lee KL. Observations on the spontaneous occurrence of paroxysmal supraventricular tachycardia. Circulation. 1984;70:1-6. 14. Holt JD, Prentice RL. Survival analyses in twin studies and matched pair experiments. Biometrika. 1974;61:17-30. 15. Conover WJ. Practical nonparametric statistics. 2d ed. New York: John Wiley & Sons; 1980. 16. Pritchett EL, Hammill SC, Reiter MJ, et al. Life-table methods for evaluating antiarrhythmic drug efficacy in patients with paroxysmal atrial tachycardia. Am J Cardiol. 1983;52:1007-12. 17. Midtbo K. Verapamil in the prophylactic treatment of paroxysmal supraventricular tachycardia. Curr Ther Res. 1981;30:372-7. 18. Mauritson DR, Winniford MD, Walker WS, Rude RE, Cary JR, HUlis LD. Oral verapamil for paroxysmal supraventricular tachycardia: a long-term, double-blind randomized trial. Ann Intern Med. 1982;96:409-12. 19. Anderson JL, Gilbert EM, Alpert BL, et al. Prevention of symptomatic recurrences of paroxysmal atrial fibrillation in patients initially tolerating antiarrhythmic therapy. A multicenter, double-blind, crossover study of flecainide and placebo with transtelephonic monitoring. Flecainide Supraventricular Tachycardia Study Group. Circulation. 1989;80:1557-70. 20. Henthorn RW, Waldo AL, Anderson JL, et al. Flecainide acetate prevents recurrence of paroxysmal supraventricular tachycardia. Circulation 1991;83:119-25. 21. Weiner HL, McCarthy EA, Pritchett EL. Regular tachycardias in patients with symptomatic atrial fibrillation [Abstract]. J Am Coll Cardiol. 1990;193A.
• Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
Objective: To test the hypothesis that propafenone, administered orally, prevents symptomatic paroxysmal supraventricular arrhythmias. Design: A 6-month, open-label, dose-finding phase followed by a randomized, double-blind, placebo-controlled, crossover phase, with each treatment period lasting up to 60 days. Setting: An outpatient clinic. Patients: Thirty-three patients with either paroxysmal supraventricular tachycardia (n = 16) or paroxysmal atrial fibrillation (n = 17) were enrolled. Their arrhythmias were documented by electrocardiogram before enrollment. Twenty-three patients (14 with paroxysmal supraventricular tachycardia and 9 with paroxysmal atrial fibrillation) were randomized and the data obtained from these patients were used in the efficacy analysis. Intervention: Propafenone (300 mg three times daily in 19 patients, 300 mg twice daily in 3 patients, and 150 mg twice daily in 1 patient) and matching placebo tablets were administered in a randomized sequence. Measurements: Symptomatic arrhythmia was documented by telephone transmission of the electrocardiogram. Main Results: The time to first recurrence was prolonged for the overall group of 23 patients while they received propafenone (P = 0.004). The recurrence rate of arrhythmia during treatment with propafenone was estimated to be approximately one fifth of the recurrence rate during treatment with placebo. Conclusions: Propafenone is effective in reducing symptomatic paroxysmal supraventricular arrhythmias. Annals of Internal Medicine. 1991;114:539-544. From the Duke University Medical Center, Durham, North Carolina. For current author addresses, see end of text.
Advances in antiarrhythmic drug therapy for treating paroxysmal supraventricular arrhythmias have been impeded by the difficulty of establishing efficacy in adequately controlled clinical trials. A new antiarrhythmic drug, propafenone hydrochloride (hereafter referred to as propafenone), was approved recently by the United States Food and Drug Administration for oral administration to treat ventricular arrhythmias (1). The results of several uncontrolled studies have suggested that oral propafenone also may be useful in the management of patients with paroxysmal supraventricular arrhythmias
(2-7); however, only one placebo-controlled clinical trial (conducted in patients with paroxysmal atrial fibrillation) has been reported (8). Our rigorous, placebo-controlled trial was intended to test the efficacy of oral propafenone in the prophylaxis of recurrent symptomatic arrhythmias in patients with paroxysmal supraventricular tachycardia or paroxysmal atrial fibrillation.
Patients and Methods Patient Sample Patients with symptomatic paroxysmal supraventricular arrhythmias were recruited from the investigators' clinic between 1984 and 1989, because their arrhythmia was incompletely controlled by their current therapy. At enrollment, each patient's arrhythmia was classified on the basis of the electrocardiograms that were available at the time of referral. Paroxysmal supraventricular tachycardia was diagnosed if the electrocardiograms showed a mean rate greater than 120 beats/min, QRS-complex morphology during tachycardia that either was normal or indicated a functional bundle-branch block, less than 0.02 s of variation in successive cardiac cycles, no evidence of atrioventricular dissociation, and episodic occurrence. Paroxysmal atrial fibrillation was diagnosed if the electrocardiograms showed a mean rate greater than 100 beats/min, QRS-complex morphology during tachycardia that either was normal or indicated a functional bundle-branch block, a grossly irregular ventricular rhythm, the absence of P waves or the presence of fibrillatory waves in the baseline, and episodic occurrence. The arrhythmia was documented by an electrocardiogram that was recorded during the manifestation of symptoms in all patients. Patients who were diagnosed as having paroxysmal atrial fibrillation were allowed to take digoxin during the study if the dose was stable for 3 weeks before treatment began and did not vary during the study; all other antiarrhythmic drugs, beta adrenergic receptor blockers, and calcium-channel blockers were prohibited. Patients were excluded if paroxysms of arrhythmia precipitated angina, pulmonary edema, or neurologic symptoms. Patients with paroxysmal atrial fibrillation were excluded if they had the Wolff-Parkinson-White syndrome. Methods Recording Methods Throughout the study, patients used a small, battery-operated electrocardiogram recorder (Cardiobeeper Memory Monitor, Survival Technology, Bethesda, Maryland) to document their rhythm when symptoms were manifested and to play back the electrocardiogram to the investigators using a toll-free automatic telephone answering system. In addition, patients were contacted by telephone at 2-week intervals to encourage compliance with the study and to verify equipment function by transmitting an electrocardiogram during sinus rhythm. ©1991 American College of Physicians
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
539
Dose-Finding Phase The overall study design followed principles suggested for the study of symptomatic paroxysmal arrhythmias (9). Patients with one of the qualifying arrhythmias began the dose-finding phase of the study by taking propafenone, 300 mg twice daily. The dose of propafenone was increased or decreased depending on the recurrence of arrhythmias or the occurrence of side effects. If the recurrence of an arrhythmia was documented using the telephone monitor system, the propafenone dose was increased to 300 mg three times daily. If no recurrence of arrhythmia was documented within 90 days of beginning therapy with propafenone, 300 mg twice daily, the dose was arbitrarily increased to 300 mg three times daily. The purpose of this arbitrary dose increase was to find the highest tolerated dose in each patient during the dose-finding phase. To compare the efficacy of taking 300 mg twice daily with that of taking 300 mg three times daily in the dose-finding phase, we used only data obtained from those patients whose plasma drug concentrations were presumed to have reached steady-state (after 3 days of therapy) on both regimens. For each patient, the dose-finding phase could last as many as 180 days. Patients returned to a research arrhythmia clinic every 2 weeks for the first 3 months and at monthly intervals thereafter. At these visits, a physical examination was done, electrocardiograms were recorded, and blood samples were drawn for laboratory screening and for measuring plasma propafenone concentrations. Randomized Phase Patients who successfully completed the dose-finding phase entered a randomized, double-blind, placebo-controlled, crossover phase. At the beginning of this phase, patients were provided with double-blinded medication labeled "drug A." The tablets looked like the medication they had used at the end of the dose-finding phase, and they contained either placebo or propafenone. The first observation period began 3 days after beginning treatment with drug A to allow propafenone to wash out during placebo treatment periods. Patients were then followed until they either had their first recurrence of arrhythmia while receiving drug A or had completed an additional 60 days of therapy with drug A, whichever came first. Patients were then treated in an open-label fashion with propafenone for 3 to 7 days. Next, they were provided in a blinded fashion with an additional medication labeled "drug B" whose formulation was the opposite of drug A's. As with drug A, the observation period for drug B began 3 days after beginning treatment with drug B. The randomized phase concluded at the patient's first recurrence of arrhythmia while receiving drug B or after 60 days of therapy with drug B; the maximum possible length of the randomized phase, therefore, was 4 months. Patients were seen in the clinic at monthly intervals and between treatments with drugs A and B. Adverse Experiences Throughout the dose-finding and randomized phases of the study, both cardiac and noncardiac adverse experiences were tabulated, regardless of whether they were thought to be due to propafenone. These experiences were tabulated by drug dose even if the patient's plasma concentrations had not reached steady-state while the patient received that dose. Particular attention was given to adverse experiences that caused patients to discontinue their participation in the study prematurely. Plasma Drug Concentrations During the dose-finding phase of the study, blood samples were drawn from patients to measure plasma propafenone concentrations using the method of Harapat and Kates (10). The clock time and the time since the administration of the preceding dose of propafenone were recorded for each sample; in general, blood samples were not drawn at trough times. Long-Term Follow-up Patients who successfully completed the randomized portion of the trial were allowed to continue taking medication in an open-label fashion in an extension of the protocol until pro540
pafenone was marketed. Patients were seen in the clinic every 3 months during this phase of the study. Long-term follow-up was discontinued for all patients by June 1990. Data Analysis We compared demographic data obtained from patients who qualified with paroxysmal supraventricular tachycardia with data obtained from patients who qualified with paroxysmal atrial fibrillation; we also compared data obtained from patients included in with data obtained from patients excluded from the randomized phase. These comparisons were done using the Fisher exact test (two-sided) for dichotomous variables and the Wilcoxon rank-sum test (two-sided) for age. We used the Kaplan-Meier product-limit method to estimate and to illustrate graphically the proportion of patients remaining free of arrhythmia on each day of the various follow-up periods (11). The formal statistical comparison of propafenone treatment with placebo was done using data from the randomized phase; we compared the time to the first recurrence of arrhythmia while patients received placebo with that while patients received propafenone using a one-sided randomization test based on the paired Prentice-Wilcoxon statistic for censored data (12). Assuming that the successive occurrences of arrhythmia constitute a Poisson process (13), the rate of occurrence of these attacks is constant for each patient and the effect of propafenone on this rate can be estimated using the proportional hazards model. Using this model, the hazard function hi (t) for the ith subject was modeled as ajh0 (t) for an arbitrary, unspecified hazard function h0 (t) so that the hazard functions among subjects differed at most by multiplicative constants (14). In the Poisson model, the hazard ratio is simply the ratio of the rates of occurrence in the underlying Poisson processes; an estimate of the hazard ratio is therefore an estimate of the effect of propafenone on the rate at which symptomatic episodes of paroxysmal supraventricular arrhythmia occur. We also compared the heart rates during attacks in each treatment period using the Wilcoxon signed-rank test and a 95% confidence interval (CI) for the median difference based on the Wilcoxon statistic (15). Results Patient Sample Thirty-three patients with qualifying arrhythmias (16 with paroxysmal supraventricular tachycardia and 17 with paroxysmal atrial fibrillation) entered the study. The qualifying arrhythmia was documented by 12-lead electrocardiography in 32 patients and by transtelephonic monitoring in the other patient. Although the patients who had paroxysmal atrial fibrillation as their qualifying arrhythmia were somewhat older and more likely to have heart disease than were the patients who had paroxysmal supraventricular tachycardia, none of the comparisons of demographic variables between the two samples (P > 0.05) were significantly different (Table 1). Treatment with digoxin was continued in 5 patients with paroxysmal atrial fibrillation (2 of whom entered the randomized phase) who were taking it at the time of referral. Of the 33 patients who qualified and entered the dose-finding phase of the study, 23 completed the dose-finding phase and entered the randomized phase. The 10 patients who did not enter the randomized phase were somewhat older and more likely to have heart disease than were the 23 other patients, but none of the differences was significant (P > 0.05) (Table 2). Patients with paroxysmal supraventricular tachycardia were marginally more likely to enter the randomized phase than were patients with paroxysmal atrial fibril-
1 April 1991 • Annals of Internal Medicine • Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
Table 1. Characteristics of Patients with Paroxysmal Supraventricular Tachycardia and Patients with Paroxysmal Atrial Fibrillation* Characteristic
Paroxysmal Supraventricular Tachycardia (n = 16)
Paroxysmal Atrial Fibrillation (n = 17)
n(%) Race White Black Sex Men Women Heart disease Lung disease Randomized
14 (87.5) 2 (12.5)
17 (100.0) 0 (0.0)
8 (50.0) 8 (50.0) 5(31.2) 4 (25.0) 14 (87.5)
11 (64.7) 6 (35.3) 9 (52.9) 2(11.8) 9(47.1)
* For patients with paroxysmal supraventricular tachycardia, the median age was 51.8 years (interquartile range, 12.9 years); for patients with paroxysmal atrial fibrillation, the median age was; 63.1 years (interquartile range, 21.4 years).
lation (14 of 16 patients compared with 9 of 17 patients; P = 0.06). Dose-Finding Phase Thirty-two patients began the dose-finding phase of the study receiving propafenone, 300 mg twice daily, and one patient (with paroxysmal atrial fibrillation) began this phase receiving propafenone, 300 mg three times daily. Of the 32 patients who began the study receiving 300 mg twice daily (13 with paroxysmal supraventricular tachycardia and 10 with paroxysmal atrial fibrillation), 23 also completed at least 3 days of therapy with 300 mg three times daily. The data obtained from these 23 patients were used to compare the efficacy of these two dosages (Figure 1). Figure 1 shows the time to the first recurrence of arrhythmia for these 23 patients while they received each of the two dosages. The time to the first recurrence of arrhythmia appeared to be longer with the higher dosage; however using a statistical test was not considered to be appropriate, because the study was not designed to compare the efficacy of these two dosages. Randomized Phase
recurrence of arrhythmia is graphically illustrated in Figure 2. On the basis of the proportional hazards model, the hazard ratio was estimated to be 0.21; (95% CI, 0.08 to 0.58). Assuming that successive occurrences of arrhythmia constitute a Poisson process, the rate of recurrence of arrhythmia during treatment with propafenone is estimated to be 0.21 times (or approximately one-fifth) the recurrence rate during treatment with placebo. For the subgroup of patients with paroxysmal supraventricular tachycardia (n = 14), the difference between propafenone and placebo was similar to the difference found with the total group; however, the significance of this comparison was weakened by the smaller number of patients (P = 0.03). In the even smaller subgroup with paroxysmal atrial fibrillation (/? = 9), the difference between propafenone and placebo again was similar to the difference found with the total group, but the significance of this comparison was still less (P = 0.06). For the 10 patients who had a recurrence of their arrhythmia while receiving propafenone and while receiving placebo in the randomized phase, the median difference in heart rate was not different from 0 (P = 0.2); the median difference (heart rate while receiving placebo minus heart rate while receiving propafenone) was 12 beats/min (CI, - 6 to 29 beats/min). Adverse Experiences Among the 32 patients who were exposed to propafenone, 300 mg twice daily, during the dose-finding phase, the most common noncardiac adverse experiences reported were bitter taste (7 patients), nausea (7 patients), anorexia (2 patients), and weakness (2 patients). Among the 27 patients exposed to the propafenone, 300 mg three times daily, the same experiences were noted by 4, 5, 0, and 2 patients, respectively. No other noncardiac adverse experience was reported by more than 1 patient receiving either dosage. Two patients discontinued participation in the study prematurely, because they had noncardiac adverse experiences: One patient had paresthesia of the Table 2. Characteristics of Patients Included in and Patients Excluded from the Randomized Phase* Characteristic
The propafenone dosage that was used in the randomized phase was the final dosage used for the 23 patients successfully completing the dose-finding phase: 300 mg three times daily in 19 patients, 300 mg twice daily in 3 patients, and 150 mg twice daily in 1 patient. One patient discontinued participation in the study after taking drug A (placebo) and before starting drug B (propafenone) because of a prolonged attack of paroxysmal atrial fibrillation. This patient was included in the efficacy analysis by conservatively assuming that he had a recurrence of arrhythmia on the first day of the propafenone observation period. Compared with placebo, propafenone caused an increase in the time to the first recurrence of arrhythmia (P = 0.004). The prolongation of the time to the first
Excluded Patients (/i = 10)
Included Patients (n = 23) n(%)
Race White
Black Sex Men Women Heart disease Lung disease The Wolflf-Parkinson-White syndrome
21 (91.3) 2 (8.7)
10 (100.0) 0 (0.0)
11 (47.8) 12 (52.2) 8 (34.8) 5(21.7)
8 (80.0) 2 (20.0) 6 (60.0) 1 (10.0)
2 (8.7)
0 (0.0)
* For patients included ini the analysis, the median age was 49.0 years (interquartile range, 16.6 years); for patients excluded from the analysis, the median age wasi 63.9 years (interquartile range, 14.4 years).
1 April 1991 • Annals of Internal Medicine
• Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
541
Long-Term Follow-up Of the 23 patients who entered the randomized phase, 21 also received long-term propafenone treatment in an open-label fashion. These 21 patients had been followed for a median of 2.9 years (range, 98 days to 5.6 years) when follow-up was discontinued in June 1990. Twenty of the twenty-one patients had at least one recurrence of symptomatic arrhythmia during long-term follow-up. Discussion Efficacy
Figure 1. The time to the first recurrence of symptomatic paroxysmal supraventricular arrhythmia in patients receiving propafenone, 300 mg twice daily ( ), and in patients receiving propafenone, 300 mg three times daily ( ). These data were obtained from 13 patients with paroxysmal supraventricular tachycardia and 10 patients with paroxysmal atrial fibrillation who received both dosages in the dose-finding phase of the study. BID = twice daily, TID = three times daily.
arms, and the other had anorexia and nausea. Both of these patients had paroxysmal supraventricular tachycardia. Cardiac adverse experiences were less common but resulted in nine premature discontinuations, all of them in patients with paroxysmal atrial fibrillation. Of nine patients, five (two receiving 300 mg twice daily and three receiving 300 mg three times daily) were thought to have had a prolonged episode of atrial fibrillation while receiving propafenone and discontinued their participation in the study prematurely for this reason. Another patient had atrial flutter with a mean ventricular rate of 263 beats/min recorded using the telephone monitor. One patient had chest pain, dyspnea, and presyncope, but was unable to transmit a satisfactory rhythm strip with his electrocardiogram monitor. Another patient had a nonsustained regular tachycardia with a wide QRS-complex morphology, and propafenone was withdrawn; this same arrhythmia was recorded months later when he was receiving no antiarrhythmic medication. Sinus pauses with ventricular asystole lasting 3.08 s caused 1 additional patient with a history of sinus node dysfunction to discontinue participation in the study. A total of 11 patients discontinued participation in the study prematurely due to adverse experiences; all were receiving propafenone in an open-label fashion when they withdrew. Of the 11 patients, 9 had paroxysmal atrialfibrillation,and 2 had paroxysmal supraventricular tachycardia.
Placebo-controlled clinical trials of oral antiarrhythmic therapy to reduce the rate of recurrence of paroxysmal supraventricular arrhythmias are a relatively recent innovation in the study of arrhythmias (9). To date, these trials have shown the beneficial effects of verapamil for treating paroxysmal supraventricular tachycardia (16-18), of propafenone for treating paroxysmal atrial fibrillation (8), of flecainide for treating paroxysmal atrialfibrillation(19), and of flecainide for treating paroxysmal supraventricular tachycardia (20). In our rigorously controlled clinical trial, propafenone (compared with a placebo) reduced the rate of symptomatic recurrence of paroxysmal supraventricular arrhythmias (P = 0.004); the treatment effect was similar in the subgroups with paroxysmal supraventricular tachycardia and paroxysmal atrialfibrillation.The rate of recurrence while patients received propafenone was estimated to be approximately one-fifth the rate seen while patients received placebo. Propafenone therefore appears to be effective in treating these frustrating clinical problems. In addition to reducing the rate of attacks, propafenone may also reduce the heart rate during the recurrence of symptomatic supraventricular arrhythmia. Although the median reduction in heart rate while patients received propafenone was not significantly different from zero (P > 0.05), the number of-patients
Plasma Propafenone Concentrations Not surprisingly, a wide range of plasma propafenone concentrations was found in the samples drawn from patients at follow-up visits. Figure 3 shows concentrations and the time since the preceding dose was administered for all samples with concentrations in the detectable range. 542
Figure 2. The time to the first recurrence of symptomatic paroxysmal supraventricular arrhythmia in patients receiving placebo ( ) or propafenone ( ). These data were obtained from 14 patients with paroxysmal supraventricular tachycardia and 9 patients with paroxysmal atrial fibrillation who were studied in the randomized phase of the trial. Plac. = placebo, Prop. = propafenone.
1 April 1991 • Annals of Internal Medicine • Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
Figure 3. Plasma propafenone concentrations for patients taking propafenone, 300 mg twice daily (top) and 300 mg three times daily (bottom). The range of values is very wide, and there is substantial overlap among the concentrations measured with the two dosages.
having recurrences while receiving propafenone and while receiving placebo was small (n = 10). The CI for the median difference in the rate of attacks between propafenone and placebo periods was consistent with a minor propafenone effect: The upper limit of the CI corresponds to a 15% reduction in heart rate with propafenone. Adverse Experiences A substantial number of patients (11 of 33 patients) discontinued participation in our study prematurely because of adverse experiences. A disproportionate number of these premature discontinuations (9 of the 11) occurred in patients with paroxysmal atrial fibrillation (17 of 33 patients enrolled). Most premature discontinuations were due to cardiac adverse events, particularly arrhythmias. Several patients thought that propafenone lengthened their attacks of atrial fibrillation, and other rhythms were recorded, including tachycardias and bradycardias. Because our study did not use a parallel placebo group during dose-finding, we cannot know whether these events were drug effects or whether they were changes in the patients' arrhythmia behavior. For example, 1 of our patients who withdrew prematurely because of the occurrence of a tachycardia with wide
QRS-complex morphology was later found to have the same arrhythmia when he was receiving no antiarrhythmic drug at all. In addition, Weiner and colleagues (21) showed that other tachycardias were commonly documented by the telephone monitor in patients with paroxysmal atrialfibrillation.In summary, we are reluctant to ascribe all the cardiac adverse experiences to propafenone, because some of them may simply have been fortuitous changes in arrhythmia behavior. Our study design, in which some patients discontinued participation before receiving a placebo, limits the interpretation of adverse experiences (9). The sample of patients going through the randomized phase of the study is ' 'enriched'' by the elimination of patients who have intolerable side effects or who refuse to continue because the drug appears to fail during dose-finding. This study design contrasts with a parallel design in which patients are randomly assigned to receive placebo or one of several drug dosages. Compared with a parallel design, the design that we used probably overestimates efficacy at the price of exaggerating adverse experiences. Such a design may be particularly useful when treatment effects are weak or confined to a small subgroup of patients. The strength of the propafenone treatment effect that we found in this study was, in fact, much greater than we had expected on the basis of previous work with verapamil in this type of study (16). Verapamil reduced the rate of recurrent attacks of symptomatic paroxysmal supraventricular tachycardia to approximately one-half the rate achieved with placebo; in our current study, propafenone reduced the rate to one-fifth the rate achieved with placebo. Related Reports A large body of clinical literature consisting of uncontrolled clinical experiences suggests that oral propafenone is useful for treating supraventricular arrhythmias (2-7). Only one other controlled clinical trial, by Connolly and Hoffert (8), however, was done in patients with paroxysmal atrial fibrillation. In that study, patients kept a diary of days on which they had atrial fibrillation and sent transtelephonic rhythm strips to the investigators to confirm the diagnosis. Like our study, the Connolly and Hoffert study consisted of a dosefinding phase followed by a randomized, double-blind, placebo-controlled, crossover phase. The investigators found a significant reduction in the mean percentage of days with atrialfibrillation(from 51% ± 34% with placebo to 27% ± 34% with propafenone; P < 0.01). They also found that studying patients with paroxysmal atrial fibrillation was difficult: Of 20 patients who entered the study, 2 were later found to have been ineligible; 3 withdrew from the study during the dose-finding phase because of poor therapeutic response; 1 withdrew during the dose-finding phase because of poor compliance; 2 withdrew during the dose-finding phase because of intolerable side effects; 2 withdrew during the randomized phase because of poor therapeutic response; and 1 died during the randomized phase while receiving placebo. Only 9 of 20 patients completed the entire study, quite similar to our own experience in which only 8 of 17 patients with paroxysmal atrialfibrillationcompleted
1 April 1991 • Annals of Internal Medicine • Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
543
the study. Because placebo-controlled trials of paroxysmal atrial fibrillation are rare, whether the many premature withdrawals in this group were related to propafenone or to the type of patient who has paroxysmal atrial fibrillation was uncertain. In conclusion, the results of our trial confirm the efficacy of propafenone for treating both paroxysmal supraventricular tachycardia and paroxysmal atrial fibrillation in patients who tolerate therapy. The results of additional studies, accumulating clinical experience, and cost will be important factors in establishing the precise role for propafenone in antiarrhythmic practice. Grant Support: In part by a grant from Knoll Pharmaceutical Company and by grant ROl HL40392 from the National Heart, Lung and Blood Institute, Bethesda, Maryland, and grant M01 RR30 from the National Center for Research Resources, National Institutes of Health, Bethesda, Maryland. Current Author Addresses: Dr. Pritchett and Ms. McCarthy: Box 3477, Duke University Medical Center, Durham, NC 27710. Dr. Wilkinson: Box 2914, Duke University Medical Center, Durham, NC 27710. References 1. Funck-Brentano C, Kroemer HK, Lee JT, Roden DM. Propafenone. N Engl J Med. 1990;322:518-25. 2. Breithardt G, Borggrefe M, Wiebringhaus E, Seipel L. Effect of propafenone in the Wolff-Parkinson-White syndrome: electrophysiologic findings and long-term follow-up. Am J Cardiol. 1984;54:29D39D. 3. Ludmer PL, McGowan NE, Antman EM, Friedman PL. Efficacy of propafenone in Wolff-Parkinson-White syndrome: electrophysiologic findings and long-term follow-up. J Am Coll Cardiol. 1987;9:135763. 4. Kerr CR, Klein GJ, Axelson JE, Cooper JC. Propafenone for prevention of recurrent atrial fibrillation. Am J Cardiol. 1988;61:914-6. 5. Antman EM, Beamer AD, Cantillon C, McGowan N, Goldman L, Friedman PL. Long-term oral propafenone therapy for suppression of refractory symptomatic atrial fibrillation and atrial flutter. J Am Coll Cardiol. 1988;12:1005-11.
544
1 April 1991 • Annals
of Internal
Medicine
6. Hammill SC, Wood DL, Gersh BJ, Osborn MJ, Holmes DR Jr. Propafenone for paroxysmal atrial fibrillation. Am J Cardiol. 1988; 61:473-4. 7. Allen BJ, Brodsky MA, Doria R, Luckett CR, Thomas R, Henry WL. Oral propafenone therapy for patients with paroxysmal supraventricular tachyarrhythmia. Chest. 1988;94:853-4. 8. Connolly SJ, Hoffert DL. Usefulness of propafenone for recurrent paroxysmal atrial fibrillation. Am J Cardiol. 1989;63:817-9. 9. Pritchett EL, Lee KL. Designing clinical trials for paroxysmal atrial tachycardia and other paroxysmal arrhythmias. J Clin Epidemiol. 1988;41;851-8. 10. Harapat SR, Kates RE. High-performance liquid chromatographic analysis of propafenone in human plasma samples. J Chromatogr. 1982;230:448-53. 11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Statist Assoc. 1958;53:457-81. 12. O'Brien PC, Fleming TC. A paired Prentice-Wilcoxon test for censored paired data. Biometrics. 1987;43:169-80. 13. Pritchett EL, Smith MS, McCarthy EA, Lee KL. Observations on the spontaneous occurrence of paroxysmal supraventricular tachycardia. Circulation. 1984;70:1-6. 14. Holt JD, Prentice RL. Survival analyses in twin studies and matched pair experiments. Biometrika. 1974;61:17-30. 15. Conover WJ. Practical nonparametric statistics. 2d ed. New York: John Wiley & Sons; 1980. 16. Pritchett EL, Hammill SC, Reiter MJ, et al. Life-table methods for evaluating antiarrhythmic drug efficacy in patients with paroxysmal atrial tachycardia. Am J Cardiol. 1983;52:1007-12. 17. Midtbo K. Verapamil in the prophylactic treatment of paroxysmal supraventricular tachycardia. Curr Ther Res. 1981;30:372-7. 18. Mauritson DR, Winniford MD, Walker WS, Rude RE, Cary JR, HUlis LD. Oral verapamil for paroxysmal supraventricular tachycardia: a long-term, double-blind randomized trial. Ann Intern Med. 1982;96:409-12. 19. Anderson JL, Gilbert EM, Alpert BL, et al. Prevention of symptomatic recurrences of paroxysmal atrial fibrillation in patients initially tolerating antiarrhythmic therapy. A multicenter, double-blind, crossover study of flecainide and placebo with transtelephonic monitoring. Flecainide Supraventricular Tachycardia Study Group. Circulation. 1989;80:1557-70. 20. Henthorn RW, Waldo AL, Anderson JL, et al. Flecainide acetate prevents recurrence of paroxysmal supraventricular tachycardia. Circulation 1991;83:119-25. 21. Weiner HL, McCarthy EA, Pritchett EL. Regular tachycardias in patients with symptomatic atrial fibrillation [Abstract]. J Am Coll Cardiol. 1990;193A.
• Volume 114 • Number 7
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19725/ by a Universite Laval Biblioteque User on 07/31/2017
