The goal of treatment for atrial fibrillation (AF) is often to control symptoms. It remains unclear whether targets for treatment such as AF rate or AF burden are correlated with AF symptom severity. Two hundred eighty-six patients completed a questionnaire of their general health and well-being, including a detailed AF symptom assessment immediately followed by a 7-day continuous monitor. AF characteristics assessed from the monitor included AF burden, AF rate, sinus rhythm rate, frequency and severity of pauses, and premature atrial contraction or premature ventricular contraction burden. Characteristics were analyzed separately for patients with paroxysmal or persistent AF. Symptom severity was assessed using the University of Toronto Atrial Fibrillation Severity Scale. Monitor characteristics were compared with AF symptom severity. The mean age of the cohort was 61.8 years and the majority of subjects were male (65.4%). Co-morbidities included hypertension (64.5%), sleep apnea (38.4%), congestive heart failure (19.6%), and diabetes (16.4%). In those with persistent or paroxysmal AF, there were no significant predictors of AF symptom severity. Specifically, heart rate in AF or sinus rhythm, AF burden, or premature atrial contraction or premature ventricular contraction burden was not predictive of AF symptom severity. After adjusting for potential cofounders (including age, gender, and co-morbidities), these findings persisted. In conclusion, there is no value in using AF monitor characteristics to predict symptoms in patients with AF.
A number of potential rhythm characteristics may be predictive of atrial fibrillation (AF) symptom severity. Rapid ventricular response may lead to worsened symptom severity. However, previous studies of the correlation of ventricular rate and AF symptom severity have been inconsistent. Burden of AF (percentage of time in AF) in those with paroxysmal AF may be predictive of symptom severity, although there are limited data considering this possibility. Burden of premature contractions (percentage of beats that are premature atrial or ventricular contractions) may correlate with symptom severity. No previous studies have comprehensively investigated these rhythm characteristics of AF and their effects on AF symptom severity using an AF-specific symptom severity scale. To further characterize the effect of rhythm characteristics on AF symptom severity, we evaluated AF symptom severity in 286 participants enrolled in the Symptom Mitigation in Atrial Fibrillation (SMART) study with paroxysmal or persistent AF who underwent a 7-day continuous Holter monitor.
Methods
The SMART study is a prospective single-center cohort study of AF symptoms and health outcomes in stable outpatients with AF. Participants were recruited through electrophysiology clinics at the University of North Carolina at Chapel Hill. Participants were eligible to participate if they had at least 1 episode of AF not attributed to a reversible cause documented by electrocardiography or continuous looping monitor.
From September 2008 to August 2012, a total of 450 participants were enrolled and completed a baseline questionnaire of their general health and well-being. Of these 450 participants, 286 underwent long-term (>1 week) continuous monitoring and were included in this substudy. The protocol was approved by the appropriate institutional review boards, and all participants provided written informed consent.
Our primary predictor variables were measured from 7-day continuous monitors (ACT III; LifeWatch Corp, Rosemont, Illinois or ZIO Patch; iRhythm Technologies, Inc., San Francisco, California). Rhythm characteristics captured from the monitor included the percentage of time spent in AF (AF burden), the mean ventricular rate of AF, the maximum ventricular rate of AF, the minimum ventricular rate of AF, the mean ventricular rate in sinus rhythm, the maximum ventricular rate in sinus rhythm, the minimum ventricular rate in sinus rhythm, and the overall ventricular rate (either AF or sinus rhythm). Mean rate was calculated over the entire monitoring period. Other irregularities in heart rhythm measured included the number of pauses >2.5 seconds, the longest pause, total number of premature atrial contractions, premature atrial contraction burden, total number of premature ventricular contractions, and premature ventricular contraction burden. These predictor variables were all stratified between patients with paroxysmal or persistent AF.
The University of Toronto Atrial Fibrillation Severity Scale (AFSS) was used as our primary outcome variable. The AFSS is an 18-item self-administered questionnaire to quantitatively describe subjective and objective ratings of AF. The 7-item AF symptom severity subscale was used in this analysis. Seven individual AF-related symptoms (palpitations, dyspnea at rest, dyspnea on exertion, exercise intolerance, tiredness at rest, light-headedness, and chest pain) generate a numerical description of symptom severity. Participants are asked to remark on how bothered they have been by these AF symptoms in the past 4 weeks. The AFSS questionnaire was administered immediately before placement of the continuous monitor. Individual symptoms attributable to AF are scored on a 5-point Likert scale such that the total AFSS severity score ranges from 0 to 35, with higher scores indicating increased AF symptom severity. The AF symptom severity score was further dichotomized into those with severe (highest quartile: AFSS ≥20) versus nonsevere (lower 3 quartiles: AFSS <20) symptoms. This dichotomized AFSS was used as our primary outcome variable.
Age, gender, medication use, and medical co-morbidities were assessed by electronic chart review or questionnaire. The pattern of AF was classified as paroxysmal if there was any period of sinus rhythm during the monitoring period (including 100% sinus rhythm). Pattern was considered persistent if AF was present during the entire monitoring period.
Univariate statistics were used to examine means, standard deviations, and shapes of distributions for continuous variables and frequencies for categorical variables in all participants. Bivariate analyses were used to assess for predictors of the dichotomous outcome of severe AF symptoms among baseline descriptive participant features using Student t test for continuous variables and chi-square analysis for categorical variables. Bivariate analyses were then used to assess for predictors of severe AF symptoms among AF monitor characteristics. For parameters with a non-normal distribution (AF burden, number and length of pauses, and number and burden of premature atrial or ventricular contractions), the nonparametric Wilcoxon rank sum test was used for bivariate analyses. Similar analyses were then performed stratifying participants into those with paroxysmal or persistent AF. An association of AF monitor characteristics with AF symptom severity as a continuous variable was assessed using Pearson correlation coefficient. Finally, multivariate linear regression was used to assess for a correlation of AF monitor characteristics with AF symptom severity adjusting for potential confounders. Potential confounders included in multivariate models included age, gender, and heart failure, which have all been previously described to be associated with cardiac symptom severity. Multivariate analysis was performed after first stratifying participants into those with paroxysmal or persistent AF. Analyses were performed with Stata, version 11 (StataCorp LP, College Station, Texas). Statistical tests were 2-tailed, with p <0.05 considered significant.
Results
Details of the patient population are included in Table 1 . Most participants were male. Significant co-morbidities included hypertension, sleep apnea, congestive heart failure, and diabetes. Medication inventory is detailed in Table 1 . Participants were more likely to have paroxysmal (n = 170, 59.6%) than persistent (n = 116, 40.4%) AF. Overall, 74 of the 286 participants (25.9%) had severe AF symptoms.
| Characteristic | Mean ± SD or Percentage | |||
|---|---|---|---|---|
| Overall (n = 286) | Severe AF Symptoms | p Value | ||
| Yes (AFSS ≥20; n = 74) | No (AFSS <20; n = 212) | |||
| Age (yrs) | 61.8 ± 13.3 | 58.6 ± 13.3 | 62.3 ± 13.3 | 0.071 |
| Men | 65 | 40 | 72 | 0.000 |
| Hypertension | 65 | 64 | 58 | NS |
| Diabetes mellitus | 16 | 18 | 16 | NS |
| Coronary artery disease | 14 | 19 | 12 | NS |
| Heart failure | 20 | 27 | 10 | 0.001 |
| Smoker | 9 | 16 | 8 | 0.048 |
| Obstructive sleep apnea | 38 | 36 | 27 | NS |
| Medication | ||||
| β Blocker | 60 | 63 | 59 | NS |
| Calcium antagonist | 25 | 22 | 25 | NS |
| ACE/ARB | 40 | 35 | 40 | NS |
| Digoxin | 12 | 17 | 11 | NS |
| Antiarrhythmic | 48 | 44 | 49 | NS |
| Statin | 41 | 44 | 40 | NS |
| Persistent (vs paroxysmal) | 40 | 44 | 40 | NS |
In bivariate analyses ( Table 1 ), older patients were less symptomatic and men were less symptomatic. Smokers and patients with congestive heart failure were more symptomatic. Neither medication use nor the pattern of AF presentation (persistent vs paroxysmal) had an effect on symptom severity.
There was a bimodal distribution of AF burden with approximately 1/3 (n = 94) of participants having 0% AF burden during the monitoring period and approximately 40% having 100% AF burden during the monitoring period. In the entire cohort (including those with paroxysmal or persistent AF), there was no AF monitor characteristic that was predictive of severe AF symptoms ( Table 2 ). This included heart rate, AF burden, premature atrial contraction or premature ventricular contraction burden, and pauses. When stratifying patients into those with paroxysmal or persistent AF, again no AF monitor characteristic predicted severe AF symptoms in patients with either paroxysmal or persistent AF. This included sinus rate measures, AF rate measures, overall rate measures, number and burden of atrial or ventricular ectopy, or AF burden.
| Variable | Mean ± SD or Median (25%–75% IQR) | ||
|---|---|---|---|
| Overall (n = 286) | Severe AF Symptoms | ||
| Yes (AFSS ≥20; n = 74) | No (AFSS <20; n = 212) | ||
| AF burden (%) | 25.0 (0–100) | 77 (0–100) | 24.6 (0–100) |
| Rate in AF mean | 95.4 ± 21.6 | 96.3 ± 22.6 | 93.6 ± 20.5 |
| Rate in AF maximal | 163.0 ± 35.0 | 160.0 ± 30.0 | 163.6 ± 36.2 |
| Rate in AF minimal | 55.8 ± 23.7 | 53.7 ± 14.6 | 55.9 ± 25.9 |
| Rate in sinus mean | 71.1 ± 11.7 | 74.1 ± 10.0 | 70.8 ± 11.6 |
| Rate in sinus maximal | 149.7 ± 43.2 | 144.5 ± 33.8 | 148.5 ± 44.6 |
| Rate in sinus minimal | 45.9 ± 17.0 | 46.9 ± 12.1 | 45.7 ± 18.9 |
| Overall rate | 77.5 ± 16.0 | 78.2 ± 16.4 | 77.3 ± 15.5 |
| No. of pauses >2.5 seconds | 0 (0–0) | 0 (0–0) | 0 (0–0) |
| Longest pause (seconds) | 0 (0–0) | 0 (0–0) | 0 (0–0) |
| Premature atrial contractions (n) | 126 (0–1,859) | 5 (0–3,067) | 141 (0–1,645) |
| Premature atrial contraction burden (%) | 0 (0–0.33) | 0 (0–0.4) | 0 (0–0.29) |
| Premature ventricular contractions (n) | 87.5 (9–1,432) | 94 (5–835) | 92 (10–1,555) |
| Premature ventricular contraction burden (%) | 0 (0–0.1) | 0 (0–0.2) | 0 (0–0.1) |
| Total recording time (days) | 7.7 ± 2.9 | 7.2 ± 3.0 | 7.7 ± 2.8 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
