*Article* **Outcome after Interdisciplinary Treatment for Aneurysmal Subarachnoid Hemorrhage—A Single Center Experience**

**Benjamin Voellger 1,\* , Rosita Rupa <sup>1</sup> , Christian Arndt 2, Barbara Carl <sup>1</sup> and Christopher Nimsky <sup>1</sup>**


Received: 28 September 2019; Accepted: 28 October 2019; Published: 1 November 2019

**Abstract:** *Background and Objectives:* To identify predictors of outcome after aneurysmal subarachnoid hemorrhage (aSAH) in our interdisciplinary setting. *Materials and Methods:* 176 patients who had been treated for aSAH by a team of neurosurgeons and neuroradiologists between 2009 and 2017 were analyzed retrospectively. Age, gender, clinical presentation according to the Hunt and Hess (H&H) grading on admission, overall clot burden, aneurysm localization, modality of aneurysm obliteration, early deterioration (ED), occurrence of vasospasm in transcranial Doppler ultrasonography, delayed cerebral ischemia (DCI), spasmolysis, decompressive craniectomy (DC), cerebrospinal fluid (CSF) shunt placement, deep vein thrombosis (DVT), pulmonary embolism (PE), severe cardiac events (SCE), mortality on Days 14, and 30 after admission, and outcome at one year after the hemorrhage according to the Glasgow Outcome Scale (GOS) were recorded. Chi square, Fisher's exact, Welch's t, and Wilcoxon rank sum served as statistical tests. Generalized linear models were fitted, and ordered logistic regression was performed. *Results*: SCE (*p* = 0.049) were a significant predictor of mortality at 14 days after aSAH, but not later during the first year after the hemorrhage. Clipping as opposed to coiling (*p* = 0.049) of ruptured aneurysms was a significant predictor of survival on Day 30 after aSAH, but not later during the first year after the hemorrhage, while coiling as opposed to clipping of ruptured aneurysms was significantly related to a lower frequency of DVT during hospitalization (*p* = 0.024). Aneurysms of the anterior circulation were significantly more often clipped, while aneurysms of the posterior circulation were significantly more often coiled (*p* < 0.001). Age over 70 years (*p* = 0.049), H&H grade on admission (*p* = 0.022), overall clot burden (*p* = 0.035), ED (*p* = 0.009), DCI (*p* = 0.013), DC (*p* = 0.0005), and CSF shunt placement (*p* = 0.038) proved to be predictive of long-term outcome after aSAH. *Conclusion:* Long-term results after clipping and coiling of ruptured aneurysms appear equal in an interdisciplinary setting that takes aneurysm localization, available staff, and equipment into account.

**Keywords:** aneurysmal subarachnoid hemorrhage; outcome; interdisciplinary setting

#### **1. Introduction**

The high-level evidence provided by the International Subarachnoid Aneurysm Trial (ISAT) [1] led to favor coiling, a technique first described by Guglielmi [2], over clipping for the obliteration of ruptured cerebral aneurysms. Nonetheless, interdisciplinary settings were tailored to the local particularities at many neurovascular centers over time, due to the indisputable weaknesses of ISAT, namely the disproportionately high number of coiled aneurysms of the anterior communicating artery, and the comparison of the results of aneurysm obliteration performed by highly specialized

neuroradiologists with those performed by averagely trained neurosurgeons in ISAT. The Barrow Ruptured Aneurysm Trial (BRAT) [3] initially seemed to confirm the results of ISAT, while data from the BRAT study at six years after the hemorrhage [4] suggested that the modality of aneurysm obliteration of ruptured aneurysms of the anterior circulation would not affect long-term results as long as treatment remained in experienced hands. Furthermore, data from the BRAT study at 10 years after the hemorrhage demonstrated that rates of complete aneurysm obliteration and rates of retreatment actually favored clipping over coiling [5].

There is, however, only a limited number of publications on results after single-center interdisciplinary treatment for aSAH. Güresir et al. [6] identified intraparenchymal hemorrhage as prognostically unfavorable in 585 patients treated interdisciplinarily for aSAH. Proust et al. [7] found a decrease of verbal memory capabilities in 50 patients after clipping versus coiling of ruptured aneurysms of the anterior communicating artery, while other neuropsychological deficiencies and quality of life did not differ significantly between treatment groups. The same group [8] reported on the results of interdisciplinary treatment at six months after aSAH, according to the modified Rankin Scale (mRS) [9] in 64 patients who were 70 years of age or older at the time of bleeding; an unfavorable correlation of initially poor clinical presentation and delayed cerebral ischemia with outcome was found. Schöller et al. [10] described an initially good clinical presentation and an age of less than 70 years on admission as prognostically favorable factors. In a small series of patients with ruptured aneurysms of the posterior inferior cerebellar artery, Sejkorova et al. [11] identified an initially high Hunt and Hess (H&H) [12] grade as unfavorable for the outcome after interdisciplinary treatment. Schwartz et al. [13] found young age at the time of admission and absence of cerebral ischemia to yield a favorable prognosis in 106 cases of interdisciplinary treated ruptured cerebral aneurysms. AlMatter et al. [14] identified age, initial clinical presentation, re-rupture of the aneurysm, intraparenchymal hemorrhage, and ruptured aneurysms of the middle cerebral artery as relevant prognostic factors; they described a trend towards unfavorable outcomes after vasospasm, intraventricular hemorrhage, and rupture of large aneurysms.

In our study, we aimed to retrospectively identify predictive factors after treatment for aSAH in a single center series of 176 cases.

#### **2. Materials and Methods**

#### *2.1. Patients*

In this single center retrospective study, 176 patients admitted to our university hospital between 2009 and 2017 were included. The patients fulfilled each of the following inclusion criteria: subarachnoid hemorrhage diagnosed after cranial computed tomography (CCT) or lumbar puncture, detection of at least one cerebral aneurysm in digital subtraction angiography (DSA) or computed tomography angiography (CTA), obliteration of the ruptured aneurysm by coiling or clipping within 24 h after admission. Patients with a history of severe cognitive impairment prior to the hemorrhage, such as progressive dementia, and patients with H&H grade 5 hemorrhages who did not benefit from external ventricular drainage (EVD) insertion were not included.

Age, gender, aneurysm localization, blood distribution according to the modified Fisher scale [15], and clinical findings according to the H&H scale on admission were recorded.

Necessity of EVD insertion, modality of aneurysm obliteration, detection of cerebral vasospasm, frequencies of spasmolysis, delayed cerebral ischemia (DCI), decompressive craniectomy (DC), cerebrospinal fluid (CSF) shunt dependency, deep vein thrombosis (DVT), pulmonary embolism (PE), and severe cardiac events (SCE), i.e., incidents requiring electrical cardioversion or cardiopulmonary resuscitation, were recorded.

Survival at Days 14 and 30 after SAH was recorded. Follow-up at one year after the hemorrhage was recorded according to the Glasgow Outcome Scale (GOS) [16]. Favorable outcomes were defined as GOS 4 or 5.

#### *2.2. Interdisciplinary Setting*

Urgent EVD insertion was performed in cases with symptomatic hydrocephalus. Obliteration of the ruptured aneurysm was achieved within 24 h after admission. In all but one patient with a ruptured aneurysm of the middle cerebral artery (MCA), the aneurysm was clipped, while, in one patient with a ruptured aneurysm of the MCA, the aneurysm was coiled. In the remaining cases, an interdisciplinary decision as to the modality of aneurysm obliteration was made, and the ruptured aneurysm was treated accordingly. In cases with multiple aneurysms, sequence and modality of aneurysm obliteration were determined interdisciplinarily. All patients received neurosurgical intensive care.

#### *2.3. Statistical Analysis*

Statistical analysis was conducted using OpenOffice 4.1.3 and R 3.5.1 with R Studio 1.1.383 on a Mac OS X 10.14.4. Figures were created with R and R Studio. Chi square, Fisher's exact, Welch's t, and Wilcoxon rank sum served as statistical tests. To assess the impact of predictors on outcome variables, generalized linear models were fitted, and a proportional odds logistic regression was performed. Statistical significance was assumed with *p* values less than 0.05.

#### *2.4. Ethical Approval*

Upon our request in March 2018, the local ethics committee at the University Hospital Marburg considered an ethical approval unnecessary for this pseudonymized retrospective analysis.

#### **3. Results**

Mean age on admission was 56 years (range: 22–90 years). On admission, 63 patients (35.8%) were 60 years of age or older, and 9 patients (5.1%) were 80 years of age or older. One hundred six patients (60.2%) were female.

Overall, 167 of 176 patients (94.9%) presented with symptomatic hydrocephalus on admission and urgently received an EVD.

Clinical findings according to the H&H scale [12] on admission are given in Table 1. Information on blood distribution according to the modified Fisher scale in the initial CCT is provided in Table 2.


**Table 1.** Clinical presentation according to the Hunt and Hess (H&H) scale [12] on admission in 176 patients with aneurysmal subarachnoid hemorrhage.

**Table 2.** Blood distribution in the initial cranial computerized tomography (CCT) scan according to the modified Fisher scale [15] in 176 patients with aneurysmal subarachnoid hemorrhage.


Clipped and coiled aneurysms by location are tabulated in Table 3. In 42 patients (23.9%), multiple cerebral aneurysms were detected. Ruptured aneurysms of the anterior circulation were significantly more often clipped, while ruptured aneurysms of the posterior circulation were significantly more often coiled (chi square test, *p* < 0.001).


**Table 3.** Clipped and coiled aneurysms by location in 176 patients with aneurysmal subarachnoid hemorrhage.

Abbreviations: MCA, middle cerebral artery; ACA, anterior cerebral artery; Acomm, anterior communicating artery; ICA, internal carotid artery; Pcomm, posterior communicating artery; SCA, superior cerebellar artery; PICA, posterior inferior cerebellar artery. Superscript 1, anterior circulation; superscript 2, posterior circulation. \* Statistically significant finding.

Events of clinical significance during hospitalization are listed in Table 4. We found a significantly higher probability of DVT in patients who underwent clipping as opposed to coiling of ruptured cerebral aneurysms (Fisher's exact test, *p* = 0.024).

**Table 4.** Events of clinical significance during hospitalization in 176 patients with aneurysmal subarachnoid hemorrhage.


Abbreviations: EVD, external ventricular drainage; SCE, severe cardiac event; TCD, transcranial Doppler sonography; DCI, delayed cerebral ischemia; DC, decompressive craniectomy; CSF, cerebrospinal fluid; DVT, deep vein thrombosis; PE, pulmonary embolism. \* Statistically significant finding.

Information on survival during the first month after the hemorrhage is provided in Table 5. At 30 days after the hemorrhage, we found a significantly higher probability of survival in patients who underwent clipping as opposed to coiling of ruptured cerebral aneurysms (generalized linear modeling, *p* = 0.0495). Outcome according to the GOS at one year after the hemorrhage is given in Table 6. Clinical data at 14 and 30 days after the hemorrhage were available in all patients, while follow-up at one year was obtained in 133 of 176 patients (75.6%).


**Table 5.** Survival during the first month in 176 patients with aneurysmal subarachnoid hemorrhage.

\* Statistically significant finding.

**Table 6.** Outcome according to the Glasgow Outcome Scale (GOS) [16] at one year in 176 patients with aneurysmal subarachnoid hemorrhage.


The impact of potentially predictive variables on outcome after aSAH is illustrated in Figure 1.

**Figure 1.** Statistical significance of potential predictors of outcome in 176 patients with aneurysmal subarachnoid hemorrhage as estimated from generalized linear models for dichotomous outcome variables (survival and fav. outcome) and proportional odds logistic regression for an ordinal outcome variable (GOS) after aneurysmal subarachnoid hemorrhage. Information on survival at Days 14 and 30 after the hemorrhage was available in all patients, while information on outcome at one year after the hemorrhage was obtained in 133 patients (75.6%). Abbreviations: H&H, Hunt and Hess; TCD, transcranial Doppler sonography; DCI, delayed cerebral ischemia; DC, decompressive craniectomy; CSF, cerebrospinal fluid; DVT, deep vein thrombosis; PE, pulmonary embolism; SCE, severe cardiac event; fav., favorable; GOS, Glasgow Outcome Scale; (\*) true (dichotomous variables) or increasing (ordered values) values predictive of survival, favorable outcome, or high GOS score; (+) true (dichotomous variables) or increasing (ordered values) values predictive of mortality, unfavorable outcome or low GOS score.

#### **4. Discussion**

Various authors reported an initially poor H&H grade to predict an unfavorable outcome after aSAH [1,8,10,11,14]. This finding was confirmed in our study: according to our data, an initially poor H&H grade significantly predicted mortality at Days 14 and 30 after the hemorrhage, mortality at one year after the hemorrhage and a less favorable outcome according to the GOS at one year after the hemorrhage. We regularly refrain from obliterating ruptured cerebral aneurysms in patients who, prior to the onset of aSAH, have a history of severe cognitive impairment, such as progressive dementia, since we consider the potential benefit of aneurysm obliteration to these patients highly questionable.

Age as a predictor of outcome after aSAH has been reported before [10,13,14]. In our study, age of over 70 years on admission, being a cut-off age within the range of those of several other publications, was a strong predictor of mortality at 14 days after the hemorrhage, an unfavorable outcome and a less favorable outcome according to the GOS at one year after the hemorrhage. We share, however, the view of other authors that age alone should not be an objection as to the diagnosis and treatment of cerebrovascular diseases [17].

In other studies on aSAH, a high overall clot burden has been described as a predictive factor [18]. We found the extent of overall clot burden in the initial CCT, as recorded according to the modified

Fisher scale, to be significantly associated with mortality at 30 days after the hemorrhage, and with a less favorable outcome according to the GOS at one year after the hemorrhage.

Clipping as opposed to coiling of ruptured aneurysms was a significant predictor of survival at 30 days after aSAH, but not later during the first year after the hemorrhage. As far as we know, a similar finding has not been reported in other studies before. We suppose that the reduction of overall clot burden, which is part of the standard clipping procedure as opposed to the standard coiling procedure after aSAH, leads to a temporary relief from spasmogenic stimuli in the subarachnoid space, which may explain our finding.

Furthermore, our single center experience has shown that particularly in patients with poor H&H grades, clipping of a ruptured aneurysm is often accompanied by DC in the same session. By contrast, in patients with coiled aneurysms usually a clinical deterioration somewhat later in the course of the disease gives rise to DC. The earlier onset of the effect of DC may lead to temporary recovery in clinically poor patients who undergo clipping of ruptured aneurysms, which may contribute to the statistically significant effect of clipping on survival at 30 days after the hemorrhage as observed in our study.

Coiling as opposed to clipping of ruptured aneurysms was significantly correlated with a lower frequency of DVT, while the frequency of PE did not significantly depend on the modality of aneurysm obliteration in our patients. One may assume that acetylsalicylic acid (ASS) and heparin in doses with therapeutic effects, as regularly administered after aneurysm coiling, help to prevent DVT, while larger patient cohorts need to be analyzed to prove a potential impact of ASS and heparin on the frequency of PE after aSAH.

The fact that aneurysms of the anterior circulation were significantly more often clipped, while aneurysms of the posterior circulation were significantly more often coiled, is primarily attributable to particularities of our interdisciplinary approach.

ED has been reported to predict an unfavorable outcome after aSAH [19], which was confirmed in our study: ED was significantly related to an unfavorable outcome at 12 months after the hemorrhage.

Other authors have reported SCE as well as elevated Troponin levels, electrocardiographic or echocardiographic abnormalities, to be linked to unfavorable outcomes after aSAH [20–22]. In our patients, SCE were a significant predictor of mortality at 14 days after the hemorrhage, but not later during the first year after the hemorrhage. This finding may become relevant when clinical decisions have to be consented with the patient's next of kin: in patients with aSAH, a severely unstable cardiovascular situation during hospitalization after the hemorrhage occasionally tempts family members to fear lack of recovery and to demand intensive care not to be extended. Our study may provide a rationale to continue curative treatment in these cases.

We found that DCI was a significant predictor of mortality on Day 30 and of a less favorable overall outcome according to the GOS at one year after the hemorrhage, which is a finding other authors have reported before [14].

In our study, DC was a significant predictor of mortality, of an unfavorable outcome (i.e., GOS < 4) and of a less favorable outcome (i.e., a lower GOS) at one year after the hemorrhage. We found, however, that more than two out of three patients who required DC survived the first year after the hemorrhage. It therefore seems well warranted to indicate DC generously in patients who deteriorate due to malignant brain swelling after aSAH.

In our study, CSF shunt placement was a significant predictor of survival on days 14 and 30, and at one year after aSAH as well as of a less favorable outcome according to the GOS at one year after aSAH. This finding should be interpreted cautiously: of 97 patients who did not receive a CSF shunt, 14 (respectively, 17 and 21) patients were deceased at 14 (respectively, 30 and 365) days after the hemorrhage, while, of 79 patients who received a CSF shunt, 0 (respectively, 0 and 4) patients were deceased at 14 (respectively, 30 and 365) days after the hemorrhage. This observation is in accordance with the clinical experience that most patients with initially poor H&H grades simply do not survive long enough to receive a CSF shunt after aSAH.

#### *Limitations of Our Study*

When interpreting our results, it should be kept in mind that this work is a single center retrospective study with an incomplete follow-up of 75.6% at one year after the hemorrhage. These facts set limitations to any generalized conclusion one would want to derive from our data. Our patients were treated at a university hospital in a country with an overall high standard of medical care.

#### **5. Conclusions**

SCE were predictive of mortality at 14 days after aSAH but not later during the first year after the hemorrhage.

Clipping as opposed to coiling of ruptured aneurysms was a significant predictor of survival at 30 days but not later during the first year after aSAH, while coiling as opposed to clipping of ruptured aneurysms was significantly related to a lower frequency of DVT during hospitalization.

Age over 70 years, H&H grade on admission, overall clot burden, ED, DCI, DC, and CSF shunt placement proved to be predictive of long-term outcome after aSAH.

Long-term results after clipping and coiling of ruptured aneurysms appear equal in an interdisciplinary setting that takes aneurysm localization, available staff, and equipment into account.

**Author Contributions:** Conceptualization: B.V. and C.N.; methodology: B.V. and C.N.; validation: B.V.; formal analysis: B.V. and R.R.; investigation: R.R.; resources: C.N.; data curation: B.V. and R.R.; writing—original draft preparation: B.V.; writing—review and editing: C.N., B.C., R.R., and C.A.; visualization: B.V.; and supervision: C.N.

**Funding:** This research received no external funding.

**Acknowledgments:** We are very grateful to med. Maximilian Schulze, Department of Neuroradiology, University Hospital Marburg, for the careful analysis of CCT scans in order to detect radiographic signs of DCI in our patients.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Case Report* **Thromboembolic Events Following Atrial Fibrillation Cardioversion and Ablation: What's the Culprit?**

**Francesco De Sensi 1,\*, Gennaro Miracapillo 1, Luigi Addonisio 1, Marco Breschi 1, Alberto Cresti 1, Pasquale Baratta 1, Francesco Paneni <sup>2</sup> and Ugo Limbruno <sup>1</sup>**


Received: 25 June 2019; Accepted: 14 August 2019; Published: 20 August 2019

**Abstract:** Stroke is a rare but possible complication after atrial fibrillation (AF) ablation. However, its etiopathogenesis is far from being completely characterized. Here we report a case of stroke, with recurrent peripheral embolism after AF ablation procedure. In our patient, an in situ femoral vein thrombosis and iatrogenic atrial septal defect were simultaneously detected. A comprehensive review of multiple pathophysiological mechanisms of stroke in this context is provided. The case underlines the importance of a global evaluation of patients undergoing AF ablation.

**Keywords:** atrial fibrillation ablation; stroke; iatrogenic interatrial septum defect; paradoxical embolism; anticoagulant interruption

#### **1. Case Report**

A gentleman, 76 years old, was scheduled for catheter ablation of atrial fibrillation (AF) and atypical left atrial flutter in the context of symptomatic left ventricular dysfunction. He reported fatigue and exertional dyspnea, and presented persistent AF on EKG. He had a weight of 68 kg, and a height of 170 cm (BMI = 23 kg/m2), with high estimated thromboembolic risk (CHA2DS2VASc = 4). He was previously prescribed with anticoagulation (Dabigatran 110 mg bid), beta-blocker (bisoprolol 5 mg od), ACE-inhibitor (ramipril 5 mg od), diuretic (furosemide 50 mg) therapy. A 2D-echocardiogram documented left ventricle dilation (LVEDD (end diastolic diameter): 61 mm) with systolic dysfunction (EF (ejection fraction): 38%). A 2D-transesophageal echocardiogram (TEE) showed absence of images referable to atrial and auricular thrombosis. Single-lobe left appendage displayed reduced function with velocity peaks of 25 cm/sec. The left atrial area was 28 cm2. No relevant atherosclerotic plaques were found in the thoracic aorta. Written informed consent was obtained and the patient underwent radiofrequency electrical pulmonary veins isolation plus roof and mitral isthmus ablation lines during systemic intraprocedural heparinization (activation clotting time (ACT)-target: 300–350 s). Electrical cardioversion was also performed due to presence of persistent AF. The total procedural time was 180 min. Dabigatran was temporarily interrupted for 36 h across the procedure and the patient was discharged the next day. After one week he was admitted to the emergency department for sudden dyspnea, being hospitalized for acute heart failure. At admission the EKG showed sinus tachycardia, while chest X-ray depicted bilateral alveolar edema. During hospitalization, after achieving hemodynamic stabilization, the patient suffered aphasia and space-time disorientation with near loss of consciousness. The Angio-CT (computational tomography) showed hypodense lesions in the left cortico-subcortical temporo-occipital area and in the left cerebellar hemisphere as showed in Figure 1. Carotid and vertebral arteries were free from hemodynamic atherosclerotic plaques. Symptoms completely disappeared after two days and at the 24 h CT scan control, the lesions were stable, in the absence of hemorrhagic transformation. After a few days, the patient complained

left limb pain and an acute distal embolism was diagnosed. A new transthoracic echocardiogram revealed a further deterioration of left ventricular ejection fraction (EF: 30%) with no evidence of intraventricular thrombosis and a clearly discernable interatrial septal defect with left-to-right shunt, this was likely attributable to the trans-septal puncture performed during the ablation (Figure 2). Ultrasonography of the groin region documented in situ not compressible left femoral vein thrombosis (Figure 3). Non fractioned heparin infusion was administered with complete resolution of both the arterial embolic occlusion and venous thrombosis. After a few days, oral anticoagulation with apixaban was initiated and the patient was discharged. At the six months follow-up, he presented with mild cognitive impairment, which persisted overtime till the last visit.

**Figure 1.** Angio-CT (computational tomography) brain scan. The exam showed an acute ischemic lesion in the left cortico-subcortical temporo-occipital area and in the left cerebellar hemisphere (last one marked with red arrow).

**Figure 2.** Transthoracic echocardiogram (subxiphoid view). The exam showed a clearly discernable interatrial septal defect with left-to-right shunt identified, at rest, with color doppler.

**Figure 3.** Ultrasonographic femoral scan. The exam showed in situ thrombosis of the left femoral vein which was not compressible with the probe.

#### **2. Case Discussion**

The case illustrates an uncommon complication after atrial fibrillation (AF) ablation manifested with recurrent embolic events: A stroke and a leg embolism. Although stroke is a well-known described complication after AF ablation, the etiopathogenetic mechanisms underlying this complication are yet to be completely characterized. Here, we summarize and discuss all the potential factors involved in this undesirable complication.

#### *2.1. Radiofrequency Lesion Set and Ablation "Per Se"*

Evidence from non-randomized studies has shown that AF catheter ablation may reduce stroke risk, when successful. Among 361,913 patients with AF of the Swedish Patient Registry, catheter ablation was associated with a lower risk of stroke (HR = 0.69) and mortality (HR = 0.50). These results were even more significant in patients with CHA2DS2-VASc score ≥2 (HR = 0.39) [1]. Especially in patients with CHA2DS2-VASc score of ≥2 (83% of 3953 patients) Saliba and colleagues found a reduction in stroke rate in the ablation group compared to the non-ablated group (HR = 0.61) [2]. On this ground, Hunter et al. demonstrated, in an international multicenter registry of 1273 patients, that freedom from AF was associated with stroke-free survival (HR = 0.30) [3]. However, when discussing the possibility of a catheter ablation procedure for AF treatment, physicians should clearly make their patients aware about a periprocedural stroke risk which is approximately 0.5–1% [4]. Thromboembolic risk is directly related to the amount of radiofrequency lesions applied in the left atrial cavity. In fact, radiofrequency produces colliquative necrosis, thus leading to endothelial dysfunction and activation of the Virchow triad. Hence, during ablation, tissue involvement is directly related to an increased embolic risk [5]. An approach adding linear or complex lesion sets to pulmonary vein isolation (PVI) did not demonstrate an increase in freedom from AF recurrences, thus the standard endpoint during the first procedure should be PVI alone [6]. In our case, extensive left atrial ablation was performed with PVI plus tracing of two ablation lines along the roof and the mitral isthmus. Such ablation strategy was due to the presence of atypical left atrial flutter as well as of persistent atrial fibrillation. Although stroke is considered an uncommon complication after AF ablation, a growing body of evidence is consistently reporting asymptomatic or subclinical ischemic lesions in up to 41% of patients [7,8]. An elegant Italian study by Gaita and colleagues analyzed postprocedural brain magnetic resonance imaging (MRI) of 232 consecutive patients with paroxysmal or persistent atrial fibrillation who underwent radiofrequency left atrial catheter ablation. Techniques used were PVI or PVI plus linear lesions plus

atrial defragmentation. A clinical cerebrovascular accident occurred in only 1 patient. However, brain MRI returned positive for new embolic lesions in 33 patients. Cardioversion (CV) during the procedure was associated with an increased risk of 2.75 (95 confidence interval, 1.29–5.89; *p* = 0.009) [9]. Our patient underwent electrical CV during the ablation due to the presence of persistent AF at the beginning of the procedure. It has been recognized that CV is related to thromboembolic events "per se", independently by the ablation procedure. In patients undergoing TEE-guided cardioversion, patients on direct oral anticoagulants (DOACs), such as dabigatran and apixaban, experienced low incidence of thromboembolic events during follow-ups (0.6% and 1.1%, respectively), similar to warfarin, with a favorable trend of bleeding safety profile [10,11]. The highest risk period after CV is the following week, which would be a suitable timeline for our patient considering the stroke and the peripheral embolism.

Finally, it seems that techniques other than radiofrequency, such as cryoballoon based one-shot ablation and duty-cycled phased radiofrequency ablation (PVAC) are not free from silent cerebral embolisms, suggesting other mechanisms (like air embolism) could play a pivotal role in the physiopathology of these subclinical findings [12–14].

#### *2.2. Management of Anticoagulant Therapy in the Periprocedural Period*

Ablation was performed in January 2017. The patient had been prescribed Dabigatran 6 months before. We decided to perform TEE due to the patient's high thromboembolic risk (CHA2DS2VASc = 4). Indeed, despite optimal oral anticoagulation with DOACs, left atrial (LA) thrombus was detected in the left appendage (LAA) in >3.6% of AF patients undergoing catheter ablation in the real world. In this setting higher CHA2DS2VASc (*p* = 0.02), but not the type of DOAC, significantly predicted the presence of LA thrombus [15].

Dabigatran has been available in Italy since 2014. At the time of ablation there were no clear guidelines on appropriate periprocedural ablation management of such a new kind of drug. On the contrary, evidence available on uninterrupted warfarin showed reduction in bleeding and thromboembolic complications [16]. Since we used all the available tools in order to reduce bleeding complications (i.e., ultrasound guided femoral veins puncture, intracardiac echocardiography, contact force sensing catheters) [17,18], we felt confident to minimize dabigatran interruption. In fact, the last assumption was in the morning of the day before, and first retake was in the evening of the day of the procedure (36 h, total interruption time). Despite this short interruption and the use of heparinization during the procedure (target activation clotting time ACT = 300–350 s), we should consider this anticoagulation break as a putative factor implicated in the patient's recurrent embolic events. Indeed, the 2017 expert consensus statement on AF Ablation provide a Class I recommendation for performing the procedure with uninterrupted dabigatran (Class I, LOE A) or rivaroxaban (Class I, LOE B-R), and a 2A recommendation for the other Xa inhibitors for which specific clinical studies had not been performed at the time [5]. These recommendations derived from the results of the RE-CIRCUIT trial which was a head-to-head comparison between uninterrupted dabigatran and uninterrupted warfarin in patients undergoing AF ablation. The incidence of major bleeding was significantly lower with dabigatran than with warfarin (5 patients (1.6%) vs 22 patients (6.9%)). No strokes/TIA (transient ischemic attack) occurred in the dabigatran arm, while there was one TIA in the warfarin group. Idarucizumab, the specific reversal agent, was never used during the study [19]. Two years before, Cappato and colleagues published the results of the VENTURE-AF trial, comparing uninterrupted rivaroxaban vs uninterrupted warfarin. Complications (a major bleeding event, one ischemic stroke, and one vascular death) occurred only in the warfarin group [20]. More recently consistent evidences were provided also for apixaban and edoxaban. The AEIOU trial, published in 2018, randomized 300 patients undergoing AF ablation to uninterrupted versus minimally interrupted (holding 1 dose) periprocedural apixaban. A retrospective cohort of patients treated with uninterrupted warfarin at the same centers was matched to the apixaban-treated subjects for comparison. There were no stroke or SE events observed in all groups. The rates of clinically significant, major bleeding were similar for all apixaban patients compared with the matched warfarin group [21]. Finally, in 2019 Hohnloser

et al. published results from the ELIMINATE-AF trial, which confirmed the safety and efficacy of uninterrupted edoxaban vs vitamin K antagonists (VKAs) in the same setting. Among 553 patients undergoing AF ablation, brain magnetic resonance imaging was performed in 177 subjects to assess silent cerebral infarcts. There was one ischaemic and one haemorrhagic stroke, both in patients on edoxaban. Cerebral microemboli were detected in 13.8% (16) of patients who received edoxaban and 9.6% (5) of patients in the VKA group (*p*= ns) [22]. Based on these clinical trials, it is now clear that a strategy of performing AF ablation on patients receiving uninterrupted anticoagulation can be performed safely and will minimize the risk of thromboembolic events. Finally, international guidelines state in the absence of controlled trial data, anticoagulation management after AF ablation should follow general recommendations (i.e., on the basis of CHA2DS2-VASc score), regardless of the presumed rhythm outcome [23].

#### *2.3. Iatrogenic Interatrial Septal Defect, In Situ Thrombosis and Paradoxical Embolism*

The diagnosis in our patient, of simultaneous iatrogenic interatrial septal defect (IASD) and in-situ thrombosis, is rather unique. These are two well characterized phenomena that have rarely been discovered together in this setting. Real incidence of IASD after AF ablation is under debate. Older studies, using transesophageal echocardiography (TEE), reported up to 19% rate during follow-up [24–27].

More recently, other rates have been described (5.6% following a first procedure and 2.2% following a second procedure) [28].

The risk of persistent IASD is in part related to the tools, technologies, and approaches used for catheter ablation. For example, the incidence of IASD at 1-year follow-up following cryoballoon ablation procedure for PVI is significantly higher in front of radiofrequency procedures [29–31]. After a single-puncture, using the robotic navigation system, an IASD was detected in 38 of 40 (95%) patients one day after the ablation. At 6-months follow up, the IASDs were closed only in 30 of 39 (78.9%) patients. The authors also addressed that persistent IASDs are not associated with an increased rate of paradoxical embolism or with relevant shunting [32].

On the other side, the real incidence of in situ asymptomatic femoral thrombosis after AF ablation is unknown. Asymptomatic deep venous thrombosis (DVT) formation, following sheath placement for electrophysiological studies (EPS) in general, were detected in up to 16–44% of patients. In contrast, symptomatic DVTs are much lower (0.5–0.8%) [33]. In 2004, Chen and colleagues reported a significant incidence (17.6%) of non-occlusive DVT after multiple sheath placements for EPS. Nonetheless, in the study, all venous thrombi were non-occlusive and asymptomatic. None of the femoral veins developed occlusive DVT [34]. Although there are weak supporting data, it is reasonable to conclude that limiting the number and the size of femoral vein sheaths on the same side can minimize thrombosis risk. Despite the fact that there are no large prospective or randomized trials, prophylactic heparin administration during the procedure may be considered on an individual basis for right chamber ablations, particularly for longer procedures, or in high-risk patients [35]. Large emboli migrated from leg veins can lodge in the right ventricle [36], whereas smaller emboli are likely to pass unimpeded to the pulmonary arteries. The occurrence of pulmonary embolism following EP procedures has previously been reported, especially in patients with a thrombophilic state [37]. Moreover, two cases of floating atrial thrombi following EP studies were successfully treated with thrombolysis in asymptomatic patients [38].

To the best of our knowledge, there are no reported cases of paradoxical embolism following AF ablation where in situ thrombosis and iatrogenic atrial septal defect are detected simultaneously. Indeed, DVT developed despite fully systemic heparinization during the procedure and minimal oral anticoagulation interruption.

#### **3. Conclusions**

In conclusion, we report a case of stroke and peripheral embolism after atrial fibrillation ablation procedure. In our patient an in situ femoral vein thrombosis and iatrogenic atrial septal defect were simultaneously detected. We highlighted and discussed each etiopathogenetic mechanism underlying this clinical condition. The case encourages a critical clinical and instrumental evaluation in the management of such undesirable complications.

**Author Contributions:** Conceptualization, F.D.S. and A.C.; methodology, F.D.S. and L.A. and A.C.; software, M.B.; validation, F.P., U.L. and G.M.; formal analysis, F.P. and U.L.; investigation, G.M. and L.A.; data curation, A.C. and P.B.; writing—original draft preparation, F.D.S. and P.B. and F.P.; writing—review & editing, F.D.S. and F.P. and G.M.; supervision, U.L.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to thank all the nurses working in the Electrophysiology Lab of Misericordia Hospital, Grosseto, Italy.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Medicina* Editorial Office E-mail: medicina@mdpi.com www.mdpi.com/journal/medicina

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18