key: cord-0814526-nnuewh7m
authors: Nakamura, Rena; Oda, Atsuhito; Tachibana, Shinichi; Sudo, Koji; Shigeta, Takatoshi; Sagawa, Yuichiro; Kurabayashi, Manabu; Goya, Masahiko; Okishige, Kaoru; Sasano, Tetsuo; Yamauchi, Yasuteru
title: Prone‐position computed tomography in the late phase for detecting intracardiac thrombi in the left atrial appendage before catheter ablation for atrial fibrillation
date: 2021-05-16
journal: J Cardiovasc Electrophysiol
DOI: 10.1111/jce.15062
sha: 371cd72149765fcbba0b6703734c0f8966d85117
doc_id: 814526
cord_uid: nnuewh7m

BACKGROUND: Contrast computed tomography (CT) is a useful tool for the detection of intracardiac thrombi. We aimed to assess the accuracy of the late‐phase prone‐position contrast CT (late‐pCT) for thrombus detection in patients with persistent or long‐standing persistent atrial fibrillation (AF). METHODS: Early and late‐phase pCT were performed in 300 patients with persistent or long‐standing AF. If late‐pCT did not show an intracardiac contrast defect (CD), catheter ablation (CA) was performed. Immediately before CA, intracardiac echocardiography (ICE) from the left atrium was performed to confirm thrombus absence and the estimation of the blood velocity of the left atrial appendage (LAA). For patients with CDs on late‐pCT, CA performance was delayed, and late‐pCT was performed again after several months following oral anticoagulant alterations or dosage increases. RESULTS: Of the 40 patients who exhibited CDs in the early phase of pCT, six showed persistent CDs on late‐pCT. In the remaining 294 patients without CDs on late‐pCT, the absence of a thrombus was confirmed by ICE during CA. In all six patients with CD‐positivity on late‐pCT, the CDs vanished under the same CT conditions after subsequent anticoagulation therapy, and CA was successfully performed. Furthermore, the presence of residual contrast medium in the LAA on late‐pCT suggested a decreased blood velocity in the LAA ( ≤ 15 cm/s) (sensitivity = 0.900 and specificity = 0.621). CONCLUSIONS: Late‐pCT is a valuable tool for the assessment of intracardiac thrombi and LAA dysfunction in patients with persistent or long‐standing persistent AF before CA.

Atrial fibrillation (AF) is among the major risk factors for intracardiac thrombi that can lead to cerebral or systemic embolism. 1 The majority of intracardiac thrombi are formed inside the left atrial appendage (LAA) due to turbulence and lower blood flow rates, particularly during AF. 2, 3 Meanwhile, catheter ablation (CA) has become the first AF treatment choice in the last few decades. Before CA, when a thrombus is present in the LAA, safety evaluations must be performed, particularly in patients with persistent or longstanding persistent AF. Although the gold standard for the detection of intracardiac thrombi is transesophageal echocardiography (TEE), this method is associated with occasional technical and appraisal difficulties.

Contrast computed tomography (CT) is another standard method used for the detection of intracardiac thrombi and has the potential to replace TEE. Compared with TEE, CT is easier to perform and is related to lower levels of discomfort among physicians. In addition, the level of difference observed in the evaluation of CT images is much lower than that in TEE. However, similar to TEE, CT can still show false-positive results under conditions of low blood flow or stasis in the LAA. In these situations, the contrast medium hardly fills the whole LAA, particularly the edge of LAA, at the time of CT. Some reports have shown that CT performed in the prone position is more useful in the differentiation of thrombi from stasis than scans performed in the supine position, as gravity can increase the spread of contrast dye into the top area of the LAA. 4, 5 It has been suggested that contrast CT conducted in the delayed or late phase yields improved accuracy in the detection of thrombi than that performed in the early phase. [6] [7] [8] However, neither method is completely exclusive of false-positive results, and there is currently no known method that combines late-phase contrast CT and proneposition contrast CT (late-pCT) in this setting. Therefore, with the aim of minimizing the rate of false-positive results, we sought to investigate the diagnostic accuracy of late-pCT in the detection of thrombi in the LAA among patients with persistent or long-standing persistent AF before CA and re-evaluated the LAA by intracardiac echocardiography (ICE) during CA.

Patients with persistent or long-standing persistent AF who were scheduled for CA and enrolled at the Japan Red Cross Yokohama City Bay Hospital from January 2018 to July 2020 were included in this study. Patients were excluded if they had an allergy to the contrast agent, renal dysfunction, or severe mitral valve disease, or were intolerant to oral anticoagulants. All patients were prescribed oral anticoagulants for at least 4 weeks before the CA procedure.

The study protocol was approved by the ethics committee of the Japan Red Cross Yokohama City Bay Hospital, and written informed consent was obtained from all the participants before study initiation.

CT was performed within the 3 weeks before CA. Early and latephase cardiac CT was conducted in the prone position using a 128-section helical CT scanner (Somatom Definition AS, Siemens Healthcare GmbH). Contrast material (Iopamiro®, 370 mg of iodine/ml, Bracco Imaging) was injected intravenously at a dose of 25.9 mg/kg body weight/s for 12-14 s. Early-phase prone-position CT (early-pCT) was initiated 5 s after the region of interest in the ascending aorta reached 160 Hounsfield units. Late-pCT was started 30 s later. Once the contrast was visualized, a breath-hold instruction was provided to the patient, and CT was initiated. After CT, the patients were observed for at least 15 min for the detection of any contrast medium-related side effects, such as a rash, nausea, or abnormal blood pressure. Three-dimensional CT images were reconstructed for the left atrium (LA) and LAA. All CT images were reviewed by two experienced cardiologists and one radiologist.

If no contrast defect (CD) was observed in the LAA by late-pCT, CA was performed as planned. If CDs were detected only during early pCT but not late-pCT, CA was performed on schedule. The performance of CA was postponed in patients with CDs who were found on late-pCT; in these patients, we changed or increased the anticoagulant dose and performed a careful follow-up at our outpatient clinic. CT angiography was repeated after several months of altered anticoagulation therapy. If the subsequent late-pCT revealed the absence of CDs in the LAA, CA was performed.

All the patients were administered oral anticoagulants more than 3 weeks before CA. Novel oral anticoagulants other than dabigatran were changed to dabigatran on the day of surgery and the day following CA. If the anticoagulants used were dabigatran or warfarin, their intake was continued throughout the perioperative period. CA was performed under general anesthesia. Activated clotting time (ACT) was measured using blood samples obtained venously every 15-30 min during the procedure; this value was maintained at 300-350 s using unfractionated heparin. An ICE probe (ViewFlex Xtra ICE catheter, Abbott Laboratories or ACUSON AcuNav, Johnson and Johnson) was inserted from the right or left femoral vein through a long sheath (SL0, Abbott Laboratories). After the transseptal approach, we placed the ICE probe in the LA, and the probe was turned to scan the entire LA and LAA before CA performance in the LA. The LAA was visualized from the body of the LA. If there was no thrombus in the LAA, CA was performed. Meanwhile, if a thrombus was suspected on ICE, CA in the LA was abandoned. The thrombi evaluated by ICE were defined as circumscribed and uniformly echo-dense intracavity masses distinct from the underlying endocardium and pectinate muscles.

Using the ICE probe, the velocity of blood flow was estimated inside the LAA. These data were obtained using pulse-Doppler imaging. Sample volumes were positioned within 1 cm of the orifice of the LAA and the velocities were assessed during AF. The peak emptying velocities of the LAA were averaged over a minimum of five consecutive cardiac cycles (with each R-R interval). 9 Three-dimensional CT (3DCT) of the LA was made of images of early-pCT for LA reconstruction. These LA images were merged with a 3D mapping system during CA.

We also assessed the presence of symptomatic stroke events during the perioperative period.

Data on the following parameters were analyzed: baseline clinical characteristics, the type of oral anticoagulants used, congestive heart failure, hypertension, diabetes, previous stroke/transient ischemic attack-vascular disease, age, sex, and the CHA 2 DS 2 -VASc score. The morphology of the LAA obtained from 3DCT was classified into the following four types: cactus, cauliflower, chicken-wing, and windsock, as previously reported. 10 

Data are presented as mean ± standard deviation. The two-tailed Welch's t-test was used for the comparison of nonparametric variables between the groups. Fisher's exact test was used to evaluate the differences in the categorical variables between the groups.

A p value < .05 was considered significant.

During the study period, 300 patients with persistent or longstanding persistent AF were enrolled. Table 1 presents their baseline characteristics. Among the 294 patients without a CD on late-pCT, CA was performed on schedule. The absence of thrombus was reconfirmed using ICE performed from the LA ( Table 2 ). In one patient, the LAA could not be visualized clearly from the body of the LA. The ICE was inserted into the left superior pulmonary vein, which enabled to observe the whole LAA.

CDs existed in both the early and late phases of pCT in the remaining six patients (2%, Figure 2A ,B). None of the patients showed CDs only on late-pCT. The CDs were located at the entrance of the LAA in 1, outer edge in 2, inner edge in 2, and the residual LAA portion after LAA clipping in 1. In these six patients, CA was canceled, and the type of anticoagulant was changed or the dose was increased in the case of warfarin use. Follow-up CT was performed after 3-7 months. We observed CD disappearance on subsequent late-pCT under the same CT conditions ( Figure 2C ); following this, (11) 0 CD only in late-pCT 0 CD in early and late-pCT 6 (2) 0 0

Note: In total, 263 patients exhibited no contrast defect (CD) by prone-position contrast computed tomography (pCT), and catheter ablation for atrial fibrillation was subsequently performed. At that time, no intracardiac thrombus was found on intracardiac echocardiography (ICE). A CD observed only during pCT in the early phase was found in 34 patients, but not during pCT in the late one; these patients showed no thrombi due to ICE during ablation. The remaining six patients had CD in both the early and late phases of pCT. They did not undergo catheter ablation, and pCT was conducted after several months of altered or increased oral anticoagulation (OAC) therapy. Thereafter, all six patients showed no CD on pCT in the late phase, and no thrombus was observed using ICE. In the whole population, six patients had also undergone CT in the supine position (sCT) within 3 days before pCT and CD presence was suspected. These CDs disappeared on late-pCT in three patients ( Figure 4 ) and no thrombus was detected on ICE during CA, as mentioned previously.

ICE imaging was performed during the AF rhythm in all patients. The average duration between first CT scanning to catheter ablation was 9.4 ± 7.5 days. The first subsequent CT in patients with CDs was performed in median 68 (38-131) days, and for one patient, another CT was needed for the disappearance of a contrast defect 58 days after seconds CT scanning ( Figure 3D) . None of the patients had complications related to CT, and no symptomatic stroke events were observed during the perioperative period of CA.

Of the 300 patients, 48 (16%) showed residual contrast medium in the LAA during late-pCT ( Figures 1B,C and 4D) . Table 3 There was no significant difference in the distribution of the LAA morphology types between the groups. The presence of residual contrast was significantly associated with a higher CHAD 2 S2-VASc score, an enlarged left atrial diameter, a decreased left ventricular ejection fraction, an increased brain natriuretic peptide level, and the presence of CDs, as detected on early or late-pCT. In addition, the blood velocity in the LAA during AF was much lower in the group with residual contrast medium. 

In the present study conducted among patients with persistent or long-standing persistent AF: (1) the negative predictive value for thrombi using late-pCT was 100% as a reference of ICE, (2) all the and specificity of 99%-100%. 11, 12 However, the performance of TEE can prove difficult or may be impossible in some patients, such as those who experience pain on probe insertion even under anesthesia or those with anatomical abnormalities. The overall TEE complication rate is 0.18%-2.8%, inclusive of a mortality value of 0.01%-0.02%, laryngospasm rate of 0.14%, dysphagia rate of 1.8%, hoarseness rate of 12%, and dental injury rate of 0.1%. 13 In addition, patients who have experienced difficulties in association with TEE may be unwilling to undergo the procedure again. Moreover, TEE has been shown to be associated with a heightened risk of severe acute respiratory syndrome coronavirus 2 transmissions during the ongoing novel coronavirus disease pandemic. Therefore, there is a need for another detection tool with high accuracy. Cardiac CT performed in the delayed phase has been established as a reliable alternative to TEE in the detection of thrombi, with a sensitivity of 96% and specificity of 92%. 6 In the absence of contraindications to the contrast medium used, CT is minimally invasive and associated with a lower level of differences in the evaluation of results and rarely leaves a bad impression on patients.

Although the detection of thrombi using contrast CT exhibits excellent negative predictive values (99%-100%), the positive predictive value is still low, at 13%-31%. [14] [15] [16] This may be attributed predominantly to difficulties in distinguishing real thrombi from circulatory stasis in the LAA, particularly at a low blood velocity. The discrimination between thrombi and circulatory stasis is more crucial in patients with persistent or long-standing AF than in those with paroxysmal AF; in the former population, 28% of patients showed CDs on sCT in the early phase and as a reference of TEE or ICE, the positive and negative predictive values were 40.7% and 100%, respectively. 17 On the other hand, in patients with paroxysmal AF, the assessment of thrombi using either TEE or CT is not difficult, owing to a high blood flow rate during the sinus rhythm and lower degree of LAA dysfunction. Although delayed or late-phase CT reportedly improves the accuracy of the detection of thrombi and significantly increases the positive predictive value to up to 92% from 41%, 8 

The prone position tends to lead to the congestion of the contrast medium due to the influence of gravity. Therefore, in patients with impaired LAA function, the contrast agent may not be completely removed from the LAA even in the late phase of contrast CT. The presence of residual contrast medium in the LAA during late-pCT indicated the presence of LAA dysfunction. Patients with AF and an LAA blood flow velocity less than 20 cm/s, as estimated by TEE, have 2.6 times the risk of ischemic stroke compared with those with a higher LAA velocity. 25 Moreover, the presence of spontaneous echo contrast (SEC) in the LA or LAA on TEE is related to thrombus formation and LAA velocities < 20-25 cm/s. 25, 26 Although CT cannot aid in the measurement of the exact LAA velocity or SEC in the LA or LAA, our study suggested that the presence of residual contrast medium in the LAA during late-pCT was associated with an LAA velocity less than or equal to 15 cm/s. This cutoff value was lower than that for TEE. Therefore, the presence of residual contrast medium on late-pCT implies a higher risk for thrombotic events.

Our study has several limitations that should be noted. First, it did not have a randomized, single-center design. Second, CT and ICE were not performed on the same day; therefore, the presence of changes in the thrombus or blood circulation status cannot be ruled out. Third, ICE could be performed only in patients without CDs on late-pCT. In addition, further examination of thrombi using TEE was performed in four of six patients with CDs. Forth, the 3DCT model from the CA procedure was reconstructed from CT images obtained in the prone position. Although 3DCT did not cause any issues throughout the CA procedure, there may be differences in the 3DCT images of the LA between those obtained in the prone and supine positions.

Late-pCT was highly useful in the assessment of intracardiac thrombi in patients with persistent or long-standing AF before CA. Compared to TEE, CT was noninvasive and reproducible and associated with a significantly lower degree of differences in the evaluation performed by physicians. The presence of residual contrast medium in the LAA, as detected on late-pCT, is indicative of a low blood velocity.

Impact of atrial fibrillation on the risk of death: the Framingham Heart Study

Left atrial appendage dysfunction: a cause of thrombus? Evidence by transesophageal echocardiography-Doppler sturdies

Inflammation and thrombosis in atrial fibrillation

Circulatory stasis or thrombus in left atrial appendage, an easy diagnostic solution

Real-time surveillance of left atrial appendage thrombus during contrast computed tomography imaging for catheter ablation: THe Reliability of cOMputed tomography Beyond UltraSound in THROMBUS detection (THROMBUS) study

Usefulness of cardiac computed tomographic delayed contrast enhancement of the left atrial appendage before pulmonary vein ablation

Left atrial appendage thrombi in stroke patients: detection with two-phase cardiac CT angiography versus transesophageal echocardiography

Detection of left atrial appendage thrombus by cardiac computed tomography in patients with atrial fibrillation: a meta-analysis

Relation of left atrial appendage function to the duration and reversibility of nonvalvular atrial fibrillation

Anatomical characteristics of the left atrial appendage in cardiogenic stroke with low CHADS2 scores

Diagnostic accuracy of transesophageal echocardiography for detecting left atrial thrombi in patients with rheumatic heart disease having undergone mitral valve operations

Accuracy of transesophageal echocardiography for identifying left atrial thrombi. A prospective, intraoperative study

Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American Society of Echocardiography and the Society of Cardiovascular Anesthesiologists

Left atrial appendage filling defects identified by multidetector computed tomography in patients undergoing radiofrequency pulmonary vein antral isolation: a comparison with transesophageal echocardiography

Utility of nongated multidetector computed tomography for detection of left atrial thrombus in patients undergoing catheter ablation of atrial fibrillation

CTA detection of left atrial stasis and thrombus in patients with atrial fibrillation

Identification of left atrial appendage thrombi in patients with persistent and long-standing persistent atrial fibrillation using intra-cardiac echocardiography and cardiac computed tomography

Prone patient positioning to exclude left atrial appendage thrombus using cardiac CT

Usefulness of electron beam tomography in the prone position for detecting atrial thrombi in chronic atrial fibrillation

Resolusion of left atrial appendage thrombus by 64-defector CT scan

Comparison of intracardiac echocardiography and transesophageal echocardiography for imaging of the right and left atrial appendages

Detection of left atrial thrombus by intracardiac echocardiography in patients undergoing ablation of atrial fibrillation

Is transesophageal echocardiogram mandatory in patients undergoing ablation of atrial fibrillation with uninterrupted novel oral anticoagulants? Results from a prospective multicenter registry

Intracardiac echocardiography for detection of thrombus in the left atrial appendage: comparison with transesophageal echocardiography in patients undergoing ablation for atrial fibrillation: the Action-ICE study

Echocardiographic assessment of the left atrial appendage

Assessment of left atrial appendage function by biplane transesophageal echocardiography in patients with nonrheumatic atrial fibrillation: identification of a subgroup of patients at increased embolic risk

Prone-position computed tomography in the late phase for detecting intracardiac thrombi in the left atrial appendage before catheter ablation for atrial fibrillation

All the authors declare that there are no conflict of interests.

The data that support the findings of this study are available from the corresponding author upon reasonable request.