KEY TEACHING POINTS
Variations in coronary artery dominance and the relationship to the coronary sinus
can lead to unexpected injury to a coronary artery during coronary sinus ablation.
Maintaining alertness during coronary sinus radiofrequency application is critical.
Monitoring the relevant surface electrocardiographic leads during ablation at a sweep
speed of 25 mm/s in addition to monitoring of the 12-lead electrocardiogram immediately
after ablation are 2 simple measures that can prevent dramatic complications.
Performing a coronary angiogram on every patient undergoing coronary sinus ablation
is excessive, tipping the risk–benefit ratio the other direction.
Catheter ablation of atrioventricular nodal reentrant tachycardia (AVNRT) and atrioventricular
reentrant tachycardia is an established procedure with a high success rate, but procedure-related
complications are not rare.
One of the feared complications is coronary injury, the frequency of which is estimated
to be ≤1%. We have recently shown that the risk of coronary artery injury is especially
high when ablating in the coronary sinus (CS) branches, such as the middle cardiac
vein (MCV), and the risk correlates inversely with the distance from the ideal ablation
site in the CS to the coronary artery. If this distance is <2 mm, the risk of coronary
artery injury is as high as 50%.
Understanding the relationship between coronary arteries and the CS is essential to
avoid this complication. We describe 2 cases in which ablation in a presumed safe
area resulted in coronary injury.
A 17-year-old white male presented to our institution with severely symptomatic tachycardia
refractory to medical therapy. During electrophysiology study, AVNRT was induced and
targeted for ablation. A magnetic navigation catheter (Stereotaxis, St. Louis, MO)
in an SR-0 sheath was used to create a CARTO map (Biosense Webster; Johnson and Johnson,
New Brunswick, NJ) and to perform the ablation. A CS venogram was not performed.
In all, 3 separate radiofrequency (RF) applications at 25 W were delivered. The lesions
started at the inferior triangle of Koch and continued by gradually pulling the catheter
toward the CS ostium, at a level of the middle CS ostium. Junctional beats were seen
during the RF applications closer to the CS ostium. During the third RF application,
the ablation catheter suddenly moved inferiorly, possibly into the MCV. RF energy
was stopped soon after catheter movement was noted. As seen in his echocardiogram
(ECG), ST elevation developed in the inferior leads (aVF, III) shortly after this
RF application (Figure 1A). Reduction of his clinical arrhythmia was not attempted
secondary to the presence of ST elevation. Of note, during ablation, only leads I,
II, and V1 were monitored at a paper speed of 200 mm/s and ST elevation was not appreciated
during RF application.
The patient was immediately transferred to the cardiac catheterization laboratory.
Access was obtained via the radial artery. The right coronary artery (RCA) was angiographically
dominant, and the RCA and right posterior descending artery were both normal. At the
ostium of the right posterior-lateral, there was a complete occlusion, as shown in
Figure 1B. Aspiration thrombectomy did not retrieve significant thrombus, and a 2
× 18-mm bare metal stent was deployed. The final angiographic result was excellent,
with TIMI-3 flow, as seen in Figure 1C. A transthoracic echocardiogram performed the
day after ablation revealed normal systolic function and mild hypokinesis of the inferior
myocardium. The patient was discharged on acetylsalicylic acid for life and clopidogrel
for 30 days. At 14 months of follow-up, he has had no clinical tachycardia recurrence
or chest discomfort.
A 39-year-old white female with preexcitation presented to our institution after 2
previous unsuccessful ablations. A CS angiogram, performed at the start of our case,
did not reveal any abnormalities. Through mapping of the right atrium during ventricular
pacing, the accessory pathway (AP) was identified near the roof of the CS, approximately
1.5 cm from the ostium, which is generally thought not to be a high-risk area for
coronary artery injury. We monitored antegrade preexcitation during RF application,
and ablation at this site eliminated conduction in 1 second. Considering the patient’s
history of 2 previous failed ablations and the pathway slant, we elected to perform
additional RF applications. 2 additional lesions were site. placed more distal and
proximal to the original A fourth application was applied to the ventricular end of
the AP, at the level of mid CS ostium. Shortly after the third ablation, ST elevation
was noted in the inferior leads (Figure 2A), and a coronary angiogram was performed
via the radial artery. Of note, during ablation, only leads I, II, and V1 were monitored
at a paper speed of 200 mm/s, and ST elevation was not appreciated.
An angiogram showed the RCA, left main coronary artery, and left anterior descending
artery to be normal. The left circumflex was dominant. At the ostium of the second
obtuse marginal, there was a 100% occlusion, as shown in Figure 2B. Because aspiration
thrombectomy was nonproductive, a balloon and a bare metal stent resulted in TIMI-3
flow, as shown in Figure 2C. A transthoracic echocardiogram revealed normal systolic
function with hypokinesis of the basal inferolateral and inferior myocardium. The
patient was discharged on acetylsalicylic acid for life and ticagrelor for 1 year.
At 15 months of follow-up, she has had no tachycardia recurrence or chest pain.
Understanding the relationship of the coronary arteries to the CS is critical during
ablation. The left circumflex artery is located on the epicardial surface of the atrioventricular
groove, near the CS. The CS is usually situated more atrial to the atrioventricular
groove, as shown in Figure 3, with only 16% of cases in the “normal” atrioventricular
(AV) groove position. Figure 3, Figure 3 show this relationship with a venogram and
arteriogram in the right anterior oblique projection. Figure 3, Figure 3 show this
in the left anterior oblique projection. The more atrial CS path varies: 1–3 mm above
the AV groove is seen in 12% of patients, a moderate elevation (4–7 mm) in 50%, and
an extreme elevation (8–15 mm) in 22%. As the left cardiac chambers and mitral annulus
dilate, the CS shifts toward the ventricular part of the mitral valve annulus.
Two previous studies searched for coronary artery injury and found the incidence of
ablation-related coronary injury to be 1%.4, 5 Retrospective and prospective registries
have reported coronary artery injury from ablation as low as 0.06%–0.1% in adults
and 0.03% in children.6, 7, 8 The underlying mechanism for injury is not completely
understood. Transient thermal irritability resulting in coronary spasm appears to
be the primary mechanism, and this stenosis can be relieved with intracoronary glycerine
trinitrate in 100% of patients, as described in a recent study.
However, an additional inflammatory component may exist. As noted in reports of animal
experiments, the inflammatory component may result in delayed medial necrosis and
intimal hyperplasia causing late stenosis.
Unfortunately, the risk factors that predispose patients to coronary artery injury
are not completely defined. One proposed hypothesis is that vessels <3 mm in diameter
do not have the protection of the heat-sink effect, making them more vulnerable to
RF heat and therefore injury. Certain procedural situations, by virtue of the targeted
site for ablation, may also increase this risk. These scenarios include linear ablation
within the CS, epicardial posteroseptal APs, and ablation of AVNRT.
Linear ablation within the CS to create an LA isthmus line may lead to circumflex
artery injury. Longstanding persistent atrial fibrillation with perimitral flutter
is a common form of LA macroreentry. Procedural success usually requires additional
ablation within the CS to either terminate the tachycardia or create an LA isthmus
block, often necessitating extension ablation. Wong et al
found that 28% of their patients had circumflex artery angiographic changes post ablation,
when compared with preprocedure coronary angiograms. Of note, 33% of these patients
had a significant response to intracoronary glycerine trinitrate. Makimoto et al
reported a case of delayed incessant ventricular tachycardia >40 hours after the creation
of an LA isthmus line in the coronary venous system.
Coronary injury may result from ablation of epicardial posteroseptal APs that used
part of the CS muscle for conduction. The ablation target, where the ventricular end
of the epicardial AP is located, is often within a tributary of the CS, the MCV, or
the neck of a CS diverticulum. Stavrakis et al
reviewed results of 240 patients with such pathways and their preablation coronary
angiograms. They found an inverse correlation between the risk of coronary artery
injury and the distance from the ablation site, with a 50% risk if the target was
within 2 mm of the artery.
Ablation of AVNRT can also lead to coronary artery injury. The atrioventricular node
(AVN) artery is the primary blood supply to the AVN as well as the His bundle. The
posterior descending and posterolateral left ventricular branches also supply the
inferior aspect of the interventricular septum. Arterial supply varies with dominance
of the coronary circulation. In approximately 85% of patients in whom the circulation
is right dominant, the AVN artery is supplied by a branch of the RCA. In patients
who have left-dominant circulation, the AVN artery originates from the left circumflex
The AVN artery usually ends as single vessel, with the remainder being shaped like
a fork or as a double-stranded vessel. Lin et al
studied the risk of AV block during slow pathway ablation for AVNRT using coronary
angiography either before or after ablation. Irrespective of other common electrophysiology
signals, they found that an ablation distance of <2 mm to the distal end of the AV
nodal artery almost always caused suprahisian block.
This case series exposes the inherent dangers to the coronary circulation when RF
energy is delivered within or near the CS, especially when the targeted area is thought
to be safe. Multiple ablation procedures may increase the risk of damage to coronary
arteries, but this relationship has not been studied specifically. In our institute,
coronary angiography is performed in patients when the ablation target may be near
the coronary artery, especially if it is located in the MCV or at the floor of the
proximal CS. However, coronary angiograms are not performed on every patient when
the ablation target is located at the CS ostium, such as with AVNRT.
Performing preablation angiograms on every patient in such cases seems excessive.
Diagnostic cardiac catheterization is safe, with a complication rate from death, myocardial
infarction, or major embolization that is well below 1%.
In centers that perform high-volume radial access, vascular complications are reduced
Although there are no data specifically addressing preablation angiograms, or even
cardiac computed tomography scans, when ablation within the CS is contemplated, specific
pros (noninvasive, well-defined anatomy) and cons (contrast load, increased radiation
exposure) of each approach will need to be evaluated by the physician for each patient.
Regardless, an understanding of coronary artery variances in relation to the CS is
essential, especially when ablation within the coronary sinus may be required.
Keeping a high degree of suspicion for coronary artery injury with CS RF application
is highly recommended. Checking relevant ECG leads for ST changes at a sweep speed
of 25–100 mm/s during an ablation may not be feasible in all patients, such as when
AV conduction is being monitored. However, checking a 12-lead ECG immediately after
ablation is a critical step that can diagnose ST changes and possibly prevent dramatic
complications. Performing imaging of the coronary circulation on every patient undergoing
CS ablation, whether with a cardiac computed tomography scan or with a diagnostic
catheterization, may be unnecessary and should be individualized for each case.