Transthoracic Echocardiography to Rule Out Paroxysmal Atrial Fibrillation as a Cause of Stroke or Transient Ischemic Attack
Background and Purpose—We assessed whether echocardiography can predict paroxysmal atrial fibrillation (PAF) in patients with cerebral ischemia presenting in sinus rhythm.
Methods—Within the prospective Find-AF cohort, 193 consecutive patients with cerebral ischemia and sinus rhythm on presentation had evaluation of echocardiographic parameters of left atrial size and function. PAF was diagnosed by 7-day Holter monitoring.
Results—In 26 patients with PAF, late diastolic Doppler (A) and tissue Doppler (a′) velocities were lower whereas left atrial diameter, left atrial volume index (LAVI), LAVI/A, and LAVI/a′ were larger (P<0.05 for all) than they were in 167 patients without PAF. In multivariate models A, a′, LAVI/A, and LAVI/a′ predicted the presence of PAF. Area under the receiver operating characteristic curve to diagnose PAF was highest for LAVI/a′ (0.813 [0.738; 0.889]). A previously suggested cut-off of LAVI/a′ <2.3 had 92% sensitivity, 55.8% specificity, and 98% negative predictive value for PAF.
Conclusions—LAVI/a′ <2.3 can effectively rule out PAF in patients with cerebral ischemia.
Atrial fibrillation is a frequent cause of ischemic stroke.1 Diagnosing atrial fibrillation is relevant, as specific secondary prevention is highly effective,2 but can be challenging when the arrhythmia is paroxysmal (PAF) and not present on admission. We have recently shown that prolonged Holter monitoring for up to 7 days triples the detection rate of PAF.3 Toh et al have recently shown in a highly experimental setting that an index indicative of left atrial contractility can discriminate hypertensive patients with PAF from matched controls without the arrhythmia.4 The aim of the current analysis was to evaluate this index in a contemporary cohort of patients with cerebral ischemia.
Echocardiographic parameters were prospectively collected according to current guidelines.5 The average of septal and lateral late diastolic peak tissue Doppler velocity, a′, was used in this study. Left ventricular mass was calculated by the Devereux formula6 and indexed to body surface area. Left atrial volume was calculated by the ellipsoid formula5 and indexed to body surface (LAVI). LAVI/a′ was calculated as described by Toh et al.4 Echocardiographers were blinded to results of 7-day Holter monitoring.
Differences between groups were tested as indicated. Multivariate analysis was performed by logistic regression correcting for age, sex, body mass index, systolic blood pressure, and history of heart failure. Receiver operating characteristics curves were constructed to assess discrimination between PAF and non-PAF.
Statistical tests were performed with PASW Statistics 18.0 software (SPSS, Inc).
Of 281 consecutive patients included in Find-AF, 193 patients with sinus rhythm on admission and with echocardiography in sinus rhythm were included in this analysis. Of these, 26 patients had PAF, identified by medical history (n=2), 7-day Holter monitoring (n=23), or both (n=1; Table 1).
All atrial parameters differed between patients with or without PAF. In multivariate analyses, all parameters of atrial function emerged as independently associated with the presence of PAF: A (P=0.006), LAVI/A (P=0.013), a′ (P=0.002), and LAVI/a′ (P=0.009). Isolated measures of left atrial size missed significance: left atrial diameter (P=0.084) and LAVI (P=0.069). Discrimination between PAF and non-PAF, as assessed by receiver operating characteristic curve analysis (Figure) resulted in areas under the curve ranging from 0.638 for A to 0.813 for LAVI/a′ (diagnostic properties at the Youden point in Table 2). LAVI/a′ <2.3, a cut-off previously reported,4 had 92% sensitivity and 56% specificity with a resultant 98% negative predictive value for PAF in our cohort. Using this cut-off, 90 patients without PAF (55%) would have been excluded from additional 7-day Holter monitoring, with 2 of 25 PAF-patients (8%) being missed.
The present analysis is the first to investigate echocardiographic indicators of left atrial function to detect PAF in patients with cerebral ischemia.
We find that parameters incorporating left atrial function differ between patients with or without PAF, discrimination by a′ and LAVI/a′ is reasonable, and the negative predictive value of LAVI/a′ using the previously suggested cut-off <2.3 is confirmed to be very high.
Increased left atrial size and decreased atrial systolic function have previously been reported in subjects with PAF.7 Areas under the curve for atrial echo parameters are lower than were previously reported,4 but differences in study cohorts (stable hypertensive outpatients versus hospitalized stroke patients), unblinded4 compared with blinded assessment in our study as well as matching controls for age and sex, may explain these differences.
We propose to use LAVI/a′ to rule out the presence of PAF in patients presenting with cerebral ischemia to reduce the number of those undergoing prolonged electrocardiographic monitoring. In Find-AF, applying the cut-off of 2.3 would have reduced the number of patients to undergo 7-day Holter monitoring from 193 to 101 patients. With this approach, only 2 patients with PAF (8%) would have been missed. Interestingly, both patients had but 1 short episode of atrial fibrillation <1 minute.
Our analysis conceptionally represents a late step in evaluating a novel diagnostic marker by validating its usefulness in a real-life cohort.
Our analysis was performed post hoc, and echocardiography was not the focus of Find-AF (although parameters were prospectively collected), so additional validation is mandatory. More recent echocardiographic techniques, eg, strain rate, might further improve diagnostic accuracy.
LAVI/a′ appears to effectively rule out PAF. We suggest to use this parameter routinely and to apply 7-day-Holter-monitoring selectively to those with LAVI/a′ >2.3.
- Received July 18, 2011.
- Accepted July 29, 2011.
- © 2011 American Heart Association, Inc.
- Wolf PA,
- Abbott RD,
- Kannel WB
- Crystal E,
- Connolly SJ
- Stahrenberg R,
- Weber-Krüger M,
- Seegers J,
- Edelmann F,
- Lahno R,
- Haase B,
- et al
- Toh N,
- Kanzaki H,
- Nakatani S,
- Ohara T,
- Kim J,
- Kusano KF,
- et al
- Lang RM,
- Bierig M,
- Devereux RB,
- Flachskampf FA,
- Foster E,
- Pellikka PA,
- et al