Rate control of atrial fibrillation (AF) is a main treatment modality. However, data are scarce on the relative efficacy of calcium channel blockers and β blockers or between drugs within each class. The purpose of the present study was to compare the effect of 4 rate-reducing, once-daily drug regimens on the ventricular heart rate and arrhythmia-related symptoms in patients with permanent AF. We included 60 patients (mean age 71 ± 9 years, 18 women) with permanent AF in an investigator-blind cross-over study. Diltiazem 360 mg/day, verapamil 240 mg/day, metoprolol 100 mg/day, and carvedilol 25 mg/day were administered for 3 weeks in a randomized sequence. The 24-hour heart rate was measured using Holter monitoring, and arrhythmia-related symptoms were assessed using the Symptom Checklist questionnaire before randomization and on the last day of each treatment period. The 24-hour mean heart rate was 96 ± 12 beats/min at baseline (no treatment), 75 ± 10 beats/min with diltiazem, 81 ± 11 beats/min with verapamil, 82 ± 11 beats/min with metoprolol, and 84 ± 11 beats/min with carvedilol. All drugs reduced the heart rate compared to baseline (p <0.001 for all). The 24-hour heart rate was significantly lower with diltiazem than with any other drug tested (p <0.001 for all). Compared to baseline, diltiazem significantly reduced both the frequency (p <0.001) and the severity (p = 0.005) of symptoms. In contrast, verapamil reduced symptom frequency only (p = 0.012). In conclusion, diltiazem 360 mg/day was the most effective drug regimen for reducing the heart rate in patients with permanent AF. Arrhythmia-related symptoms were reduced by treatment with the calcium channel blockers diltiazem and verapamil, but not by the β blockers.
The guidelines state that β blockers or nondihydropyridine calcium channel blockers alone or combined with digitalis are the drugs of choice for rate control in patients with atrial fibrillation (AF). Despite the huge number of patients treated with pharmacologic rate control, information is lacking about these drugs regarding their rate-reducing effects, their effects on patients’ symptoms, and how well they are tolerated. Most of the earlier studies that compared the effectiveness of different rate-controlling drugs were small, included few women, and often compared old substances with multiple dose regimens combined with digitalis. The present study compared the effects of 4 once-daily drug regimens on heart rate and arrhythmia-related symptoms, in both men and women with permanent, rapidly conducted AF.
Methods
The RATe control in Atrial Fibrillation (RATAF) study was a prospective, randomized, investigator-blind, cross-over study designed to compare 4 drug regimens used to reduce the ventricular heart rate in patients with permanent AF. Eligible patients were >18 years old, with permanent AF of ≥3 months’ duration and had a heart rate at rest of ≥80 beats/min or an average heart rate of ≥100 beats/min during the day. The main exclusion criteria were congestive heart failure or ischemic heart disease with the need for concomitant treatment with β blockers, hypotension, treatment with class I or III antiarrhythmic drugs, severe renal or hepatic failure, and pregnancy. Most patients were recruited from the AF outpatient clinic at Baerum Hospital (Baerum, Norway). A few were recruited from other hospitals in the area or through an advertisement in the local newspaper. The patients were randomized from May 2006 to June 2010.
The patients who used rate-reducing drugs before inclusion had a 2-week washout period before starting the first study drug. Those taking digitalis had this drug discontinued and did not start the washout period until the digitalis was undetectable in the serum. The participants then received all the following drug regimens in a randomized cross-over design: (1) metoprolol slow-release tablets 100 mg/day (AstraZeneca, London, Great Britain), (2) diltiazem sustained-release capsules 360 mg/day (Pfizer, New York, New York), (3) verapamil modified-release tablets 240 mg/day (Abbott, Abbott Park, Illinois), and (4) carvedilol immediate-release tablets 25 mg/day (Roche, Basel, Switzerland/Hexal, Holzkirchen, Germany). Each drug was given for ≥3 weeks to ensure an adequate period of washout of the previous treatment and steady-state plasma concentrations. The investigator was unaware of the study drug sequence, although for practical reasons, the participants were aware of the drug assigned. Compliance with the drug regimen was assessed by pill count after each drug period.
Before starting the first treatment and at the last day of each of the 4 treatment periods, the heart rate at rest and the 24-hour heart rate were obtained, and the patients completed questionnaires regarding any perceived arrhythmia-related symptoms.
The heart rate at rest was assessed from the 12-lead electrocardiogram recorded with the patient in the supine position after 10 minutes’ rest (10-second printouts with 50 mm/s and 10 mm/mV on a Schiller AT-101 machine, Schiller, Baar, Switzerland). All electrocardiograms were obtained between 9:00 a.m. and 11:00 a.m. The electrocardiograms were analyzed automatically and reviewed by 1 of the investigators. The 24-hour heart rate was measured from 3-channel electrocardiograms (Holter monitoring) performed with Schiller MT-101 recorders and analyzed with Schiller MT-200 software, version 2.72. The recordings were started between 9:00 a.m. and 11:00 a.m. The results of the automatic analyses were reviewed manually by the primary investigator. Whenever possible, all visits for each patient were scheduled at the same time of day. All recordings were performed on weekdays. The patients were encouraged to maintain their normal level of physical activity during the 24-hour recordings.
Arrhythmia-related symptoms were assessed using a self-administered questionnaire: the Symptom Checklist–Frequency and Severity (SCL) in Norwegian translation. The SCL questionnaire rates the frequency (from 0 to 4) and severity (from 1 to 3) of 16 symptoms potentially associated with AF, thereby generating frequency and severity scores ranging from 0 to 64 and 0 to 48, respectively, with higher scores representing worse symptoms.
The RATAF study was funded by the South-Eastern Norway Regional Health Authority and the Medical Research Foundation, Baerum Hospital. All study drugs were paid for by the study group, and none of the manufacturers were involved in any part of the study planning, execution or interpretation. The regional ethics committee and the Norwegian medicines agency approved the study, which was registered at www.clinicaltrials.gov (clinical trial no. NCT00313157 ) and conducted in accordance with the Helsinki Declaration. Each patient provided written informed consent before any study-related procedures were performed.
The sample size was targeted to permit detection of a difference in efficacy on the heart rate of 50% among the different treatment regimens. On the basis of a conservative estimate, a patient number of 40 to 71 would be needed to discriminate between a relative reduction in the mean heart rate of 14% to 22%. A de facto power analysis confirmed an adequate sample size. The categorical variables are presented as frequencies and percentages and continuous variables as the mean ± SD for normally distributed variables or median and ranges for non-normally distributed variables. The different treatment regimens (including a baseline of no drug intervention) were compared using a linear mixed model for repeated measurements with a random intercept for each patient. Possible carryover effects were assessed with an interaction term between treatment regimens and study periods. If this interaction term was not statistically significant, it was removed from the final statistical model. All data management and analysis were performed using the Statistical Package for Social Sciences, version 18.0 software (SPSS, Chicago, Illinois).
Results
We assessed 563 patients with AF for eligibility; 473 potential participants did not the meet inclusion criteria, had exclusion criteria, or were unwilling or unable to participate. The most common reasons for not being included were ischemic heart disease, systolic heart failure, paroxysmal AF, or AF persistent for <3 months. Of the 90 patients who were deemed eligible, 80 were finally randomized and included in the present trial. We excluded 4 patients from additional participation before starting the first study drug either because of episodes of ventricular tachycardia or a normal heart rate and thus no need for a rate-reducing drug or symptoms of heart failure (2 patients). Also, 12 patients withdrew from 1 of the treatment periods because of side effects: 3 receiving diltiazem, 1, verapamil, 5, metoprolol, and 3 receiving carvedilol. Dizziness, dyspnea, and fatigue were the most common reasons for withdrawal. Four patients were excluded from the study before completing all drug regimens because of unrelated events. These events included 1 death from intracerebral bleeding (the patient had received anticoagulation with warfarin and had an international normalized ratio within the therapeutic range at his death), 1 renal failure due to another drug, 1 stroke, and 1 spontaneous conversion to sinus rhythm. Finally, 60 patients completed ≥10 consecutive days of the full dose of all 4 study drug treatments, permitting evaluation of all treatment periods ( Figure 1 ). The final analysis included these 60 patients (42 men and 18 women). No serious adverse events were recorded that could be ascribed to the study drugs. The baseline characteristics are listed in Table 1 .
| Variable | Value |
|---|---|
| Age (yrs) | 71 ± 9 |
| Gender | |
| Women | 18 |
| Men | 42 |
| Body mass index (kg/m 2 ) | 27 ± 4 |
| Duration of permanent atrial fibrillation (mo) | 11 (2–121) |
| CHADS 2 score ∗ | 1.2 ± 1.1 |
| Hypertension | 25 (41.7%) |
| Stroke or transitory ischemic attack | 7 (11.7%) |
| Diabetes mellitus | 3 (5.0%) |
| Chronic obstructive pulmonary disease | 3 (5.0%) |
| Current cigarette smokers | 3 (5.0%) |
| Alcohol intake (drinks/wk) | 3.5 (0–35) |
| Systolic blood pressure (mm Hg) | 141 ± 18 |
| Diastolic blood pressure (mm Hg) | 91 ± 10 |
| Heart rate at rest (beats/min) | 95 ± 15 |
| Left atrial diameter (long-axis view) (mm) | 50.4 ± 6.6 |
| Left ventricular ejection fraction (%) | 61.4 ± 7.5 |
| Medication | |
| Warfarin | 56 (93.3%) |
| Aspirin | 4 (6.7%) |
| Angiotensin receptor blocker or angiotensin-converting enzyme inhibitor | 22 (36.7%) |
| Diuretics | 9 (15.0%) |
| Statins | 12 (20.0%) |
| Rate-controlling medication at study entry, before washout period | |
| Metoprolol | 34 (56.7%) |
| Carvedilol | 2 (3.3%) |
| Verapamil | 11 (18.3%) |
| Diltiazem | 1 (1.7%) |
| Digitoxin | 8 (13.3%) |
∗ A measure of stroke risk in patients with AF (score range 0–6, higher scores indicate greater risk): congestive heart failure (1), hypertension (blood pressure consistently >140/90 mm Hg or hypertension treated with medication [1]), age ≥75 yrs (1), diabetes mellitus (1), previous stroke, transient ischemic attack, or thromboembolism (2).
Regardless of the study drug regimen, the heart rate showed a distinct circadian pattern with faster rates during the day than at night ( Figure 2 ). No carryover effects were found between the treatments and the interaction term between the treatment regimens; thus, the periods were omitted in all linear mixed models. The overall compliance with the study drugs was 97.4%.
The 24-hour mean heart rate was significantly reduced from baseline with all 4 treatments (p <0.001 for all). Both the 24-hour mean heart rate and the nocturnal mean heart rate were lower with diltiazem 360 mg/day than with the other 3 treatment drugs (p <0.001 for all). However, diltiazem and verapamil were equally effective in reducing the daytime heart rate, performing better than the β blockers (p <0.001 for all). The median number of minutes with the heart rate >110 beats/min during 24 hours was 266 minutes at baseline (no treatment), 32 minutes with diltiazem, 66 minutes with verapamil, 63 minutes with metoprolol, and 35 minutes with carvedilol. The heart rate at rest correlated significantly with the 24-hour mean heart rate for all treatments (r = 0.63 to 0.68, p <0.001 for all; Table 2 ). The mean duration of the Holter recordings was 23 hours, 44 minutes.
| Treatment | HR at Rest (beats/min) | 24-h HR (beats/min) | Daytime HR (beats/min) | Nighttime HR (beats/min) |
|---|---|---|---|---|
| Baseline | 95 ± 15 | 96 ± 12 | 106 ± 14 | 79 ± 12 |
| Diltiazem | 77 ± 13 | 75 ± 10 | 80 ± 12 | 66 ± 9 |
| Verapamil | 82 ± 16 | 81 ± 11 | 82 ± 13 | 76 ± 12 |
| Metoprolol | 81 ± 15 | 82 ± 11 | 88 ± 13 | 72 ± 10 |
| Carvedilol | 78 ± 11 | 84 ± 11 | 89 ± 12 | 76 ± 10 |
All 60 patients (100%) achieved a heart rate at rest according to guidelines (lenient rate control, heart rate <110 beats/min) during treatment with diltiazem, 56 (93%) when treated with verapamil, 58 (97%) with metoprolol, and 60 (100%) with carvedilol. The number of patients satisfying the criteria for strict rate control (heart rate <80 beats/min) was 34 (57%) with diltiazem, 29 (48%) with verapamil, 34 (57%) with metoprolol, and 35 (58%) with carvedilol.
Arrhythmia-related symptoms were assessed using the SCL questionnaire ( Table 3 ). The mean score at baseline was 13 for symptom frequency and 8 for symptom severity. Compared to baseline, treatment with diltiazem significantly reduced both symptom frequency (p <0.001) and severity (p = 0.005). However, treatment with verapamil reduced symptom frequency only (p = 0.012). Treatment with metoprolol and carvedilol improved neither the frequency nor the severity of symptoms compared to baseline. Overall, a weak positive correlation was seen between the achieved heart rate and the frequency (r = 0.27, p <0.001) and severity (r = 0.21, p <0.001) of symptoms ( Figure 3 ). Similarly, a weak correlation was seen between the period with the heart rate >110 beats/min and the frequency (r = 0.19, p = 0.001) and severity (r = 0.16, p = 0.006) of symptoms. Women reported more frequent and more severe symptoms than men, both at baseline and during all drug regimens ( Figure 4 ).
