Transradial Uterine Artery Embolization Complicated by Stroke

Over the last few decades, transradial access (TRA) has become an increasingly popular approach for angiography and intervention. Interventional radiologists (IRs) trained in the last decade are more likely to have gained experience with TRA during training and observed the advantages inherent to this technique including faster ambulation postprocedure, increased patient comfort, and decreased bleeding and vascular complications. ^{1 2 3 4 5} However, some IRs remain reluctant to use TRA rather than transfemoral access (TFA), due to the learning curve and potential risks. ⁶ With standard anatomy and proper TRA technique, the wire and catheter cross from the left subclavian into the descending aorta. In cases with challenging anatomy, additional manipulation may be necessary in the aortic arch, posing a rare but devastating risk of stroke. Herein, we report on a pre-menopausal female who underwent elective, outpatient uterine artery embolization (UAE) through TRA, which was complicated by a postprocedural stroke. Data on stroke rates during catheter-based intervention are reviewed, we discuss the existing literature comparing TRA versus TFA, and suggest approaches to management of stroke in this setting.

Case Presentation

A 52-year-old woman, G3P2013, with heavy menstrual bleeding was referred to IR clinic for UAE evaluation. Her chief complaints were severe cramping, heavy bleeding with menstrual cycles, and urinary frequency, refractory to medroxyprogesterone therapy. Her past medical history was significant for uterine fibroids, iron deficiency anemia treated with iron supplementation, and hypertension treated with hydrochlorothiazide 25 mg daily. She had no history of stroke, diabetes, or cardiovascular disease. Magnetic resonance imaging (MRI) revealed severe diffuse adenomyosis and an anterior, intramural fibroid compressing the bladder and undergoing partial myxoid degeneration ( Fig. 1a, b ).

Fig. 1.

A 52-year-old woman with urinary frequency and heavy menstrual bleeding causing iron deficiency anemia underwent uterine artery embolization via a transradial arterial approach. ( a ) Coronal T2-weighted magnetic resonance image reveals a subserosal fibroid (white arrow) undergoing degeneration and causing mass effect on the bladder (star). ( b ) Sagittal T1 post-contrast image with fat saturation reveals that the fibroid (white arrow) enhances similarly to the remainder of the uterus.

Preprocedurally, intravenous cefazolin (2 g), intravenous ondansetron (8 mg), oral hydrocodone/paracetamol (10–325 mg), and oral cyclobenzaprine (10 mg) were administered. In the IR suite, a Barbeau B waveform was noted, and the left radial artery measured 1.9 mm. Under moderate conscious sedation, the left radial artery was accessed under ultrasound with a micropuncture needle, and a 5-Fr Glide-slender sheath (Terumo, Somerset, NJ) was placed. A “radial artery cocktail” of verapamil (2.5 mg), nitroglycerin (200 μg), and heparin (3,000 units) was slowly injected through the side arm of the sheath before connecting to a continuous flush. A 5 Fr × 125 cm angled catheter (Merit Medical, South Jordan, UT) was introduced and over a 0.035-inch angled Glidewire (Terumo). The catheter was advanced, without prolonged manipulation in the aortic arch, into the abdominal aorta, and used to select the right internal iliac artery. A 2.8-Fr Progreat microcatheter (Terumo) was introduced coaxially and a 0.018-inch Glidewire (Terumo) was used to select the right uterine artery. The right uterine artery was embolized with two vials of 300 to 500 μm particles (Embospheres, Merit Medical; Fig. 2a ). The contralateral side was approached similarly, and the left uterine artery selected and then embolized using two vials of 300 to 500 μm particles ( Fig. 2b ). The wires and catheters were removed uneventfully, and hemostasis was achieved using a TR band (Terumo). Total sedative and pain medications used during the procedure included intravenous midazolam (3 mg), fentanyl (150 μg), and ketorolac (60 mg). The patient tolerated the procedure well and was transferred to the postprocedure care area awake and in satisfactory condition.

Fig. 2.

Same patient as in Fig. 1 . The patient underwent transradial uterine artery embolization. ( a ) The right uterine artery is tortuous, and embolization was performed beyond the horizontal portion. ( b ) The left uterine artery is tortuous, and embolization was performed beyond the horizontal portion.

Shortly after arriving in the postprocedure unit, the patient became unresponsive. Vital signs and oxygen saturation levels were normal. After administration of 0.4 mg of flumazenil and 0.4 mg of naloxone, a limited neurologic exam demonstrated right hemiparesis, gaze fixation to the left, and a right Babinski sign. A “code stroke” was activated and the patient was transported to the emergency department where a noncontrast head computed tomography (CT) was unrevealing. Following discussions with a neurologist and the patient's family, 100 mg of tissue plasminogen activator was administered intravenously. CT angiography (CTA) of the head and neck revealed a large left middle cerebral artery stroke ( Fig. 3 ). Embolectomy was deferred given the size of the infarct. The patient began vomiting and grew more unresponsive requiring intubation. MRI the next day revealed additional infarcts in the left anterior cerebral artery territory and the left middle/posterior watershed regions ( Fig. 4a–c ).

Fig. 3.

Same patient as in Fig. 1. The patient developed right hemiparesis post transradial uterine artery embolization. Computed tomography angiography of the head was performed and this coronal multiplanar reformatted image reveals relative paucity of vascular filling and enhancement in the left hemisphere (*) compared with the right. The patient received tissue plasminogen activator.

Fig. 4.

Same patient as in Fig. 1. One day after suffering a postprocedural stroke, the patient underwent brain magnetic resonance imaging. ( a and b ) Axial diffusion-weighted images reveal a large area of increased signal primarily in the middle cerebral artery territory extending anteriorly to watershed areas. ( c ) Coronal diffusion-weighted image similar in position to Fig. 3 confirms that the paucity of blood flow has corresponding high signal reflecting restricted diffusion and an infarct.

Over the course of her 12-day hospitalization, the patient underwent evaluation including telemetry, transthoracic echocardiography with bubble study, transesophageal echocardiography, carotid Doppler ultrasonography, and thrombophilia workup all showing no abnormalities. At discharge to acute care rehabilitation, she endorsed expressive aphasia and right-side weakness. She continued to convalesce and gradually regained motor function and became communicative. She had begun to ambulate but had residual right-sided weakness and occasionally required a walker. Symptoms of menorrhagia and pain, however, have resolved.

Discussion

This case demonstrates the rare instance of stroke during UAE performed by TRA and its consequences on patient's well-being. The distribution of the stroke is consistent with the left middle cerebral artery, which is unusual given that manipulation of the wire and catheter in TRA typically occurs in the left subclavian artery and would most likely affect the left vertebral artery distribution. Possible explanations include aortic insufficiency, anatomic variation, or carotid stenosis, but echocardiography, head and neck CTA, and carotid ultrasonography excluded those causes. Our multicenter, high-volume practice performs hundreds of TRA interventions annually including for most patients undergoing UAE, prostate artery embolization, and intra-arterial liver-directed therapies. A similar complication occurred in the more expected posterior distribution with another patient. In the second instance, a 37-year-old female undergoing TRA UAE also experienced symptoms postprocedurally including right-sided hemiparesis, numbness, tingling, and reduced right hemifield vision ( Fig. 5a, b ). This patient did not receive tissue plasminogen activator or thrombectomy and over time her symptoms have been improving with rehabilitative therapies. Beyond hypertension, there were no other risk factors suggesting these patients would be predisposed to a stroke nor did subsequent workup reveal any etiologies. We review the expected rate of stroke during endovascular procedures, compare TFA and TRA, and present approaches to management of a postprocedural stroke.

Fig. 5.

A 37-year-old woman undergoing transradial uterine artery embolization developed right-sided hemiparesis, numbness, tingling, and reduced right hemifield vision. Magnetic resonance imaging of the brain was performed. Axial diffusion weight images ( a and b ) reveal multiple foci of increased signal in the left posterior cerebral artery territory.

While we do not know how often patient's may suffer a subclinical stroke or transient ischemic attack, stroke as a consequence of a catheter-based intervention is a rare occurrence, but can result in severe morbidity and is independently associated with in-hospital death. ⁷ Estimates of stroke complication rates range from 0.1 to 1% in cerebral angiography ^{8 9 10 11} and 0.2 to 0.4% in left heart catheterizations. ⁷ The primary mechanism of stroke during catheter-based intervention is postulated to be cerebral microembolism from air embolism, thrombus formation on the catheter, or aortic atheromatous fragments. ^{12 13 14 15} This is relevant given that the TRA necessarily traverses this pertinent vasculature. Several studies on predisposing factors of periprocedural stroke have demonstrated that higher risk patient characteristics include greater severity of coronary artery disease, diabetes mellitus, hypertension, older age, female sex, history of cerebrovascular disease, and poor renal function. Procedural factors include higher numbers of catheters used per case, greater contrast volumes, larger guide catheter calibers, and the use of rotational atherectomy devices; a higher level of experience with TRA has been shown to be weakly associated with a reduced risk of stroke. ^{16 17 18 19} Interventionalists can potentially use these factors to reduce the risk of periprocedural stroke, and to judiciously choose the access route based on patient factors and their own TRA learning curve.

Interventional cardiologists were early adopters of TRA and have the most robust data on its safety and efficacy. Several major meta-analyses have compared TRA and TFA approaches in percutaneous coronary interventions for patients with ST-segment elevation myocardial infarctions. These studies have consistently demonstrated that TRA is associated with fewer complications such as periprocedural bleeding; decrease in length of hospital stay; improvement of clinical outcomes; and comparable rates of major adverse events. ^{20 21 22 23} In 2018, the American Heart Association recommended a “radial-first” strategy in the United States for patients with acute coronary syndromes. ²⁴ Early observational studies and the more recent ACCOAST study suggested a higher frequency of stroke with TRA, ^{25 26} though the aforementioned meta-analyses also examined more recent rates of stroke and found no statistical difference. ^{21 22 23} The data from these analyses, however, were not granular or well powered. Singh et al specifically noted that their study was unable to detect statistically significant differences in stroke incidence, as only half of the included trials reported stroke outcomes and the incidence was quite low (0.48–0.59%). ²¹ Similarly, meta-analyses by Jolly et al and Patel et al specifically examined the impact of access site on stroke and/or bleeding events yielded promising but inadequately powered results. ^{27 28} In 2016, Sirker et al conducted a meta-analysis using pooled data from more than 24,000 patients in 15 randomized controlled trials (RCT) and more than 475,000 patients from 21 observational studies. They found no significant differences in risk ratios (RRs) of stroke using the RCT data (RR: 0.87, 95% CI: 0.58–1.29) and a significant difference favoring TRA using the observational data (RR: 0.71, 95% CI: 0.52–0.98). They calculated an overall stroke rate of 0.52%. ²⁹ While not definitive, these data suggest that TRA is at least comparable to TFA in regard to stroke incidence. Regarding stroke in UAE cases, however, the evidence is sparse. A systematic review by Himiniuc et al compared the safety and efficacy of TRA to other approaches for UAE and consisted of 10 relevant studies. The technical success rates of TRA were between 95 and 100% and local complication rates were comparable to or lower than TFA. None of the studies mentioned any occurrence of thromboembolic complications such as stroke. ³⁰ Furthermore, 8 of the 10 studies were retrospective and none were randomized, suggesting that these data are insufficient to draw definitive conclusions.

In the event of a postprocedure stroke, there have been reports and algorithms to offer guidance. If the event occurs intraprocedurally and the provider is comfortable with selective cerebral angiography, it is possible to directly visualize and treat immediately. ^{7 31} If the event occurs postprocedure, management is similar to strokes of other etiologies: assess vital signs, airway, breathing, and circulation; conduct a history and physical examination; and consult for a full neurologic evaluation. Additionally, expeditious imaging with a noncontrast head CT is necessary to rule out a hemorrhagic process particularly if the patients received anticoagulation, antiplatelet agents, or tissue plasminogen activator. A brain MRI is often needed if the clinical suspicion remains high as it affords visualization of acute processes earlier than CT. Magnetic resonance angiography or CTA may also be helpful. Next, it should be determined if the patient is a candidate for thrombolysis or thrombectomy or also if decompressive hemicraniectomy is indicated. Following the acute phase, care should be taken to promptly initiate physical, occupational, and speech therapy, to engage in early mobilization if permissible, and to prevent complications such as intracranial hypertension, aspiration, or thromboembolism. ^{7 32}

There are several learning points that can be drawn here. The first is the fact that periprocedural stroke is a rare event that providers should be aware of, given the potentially devastating effects. Secondly, while current evidence suggests that TRA is comparable or superior to TFA regarding stroke rates, the data are inconclusive. As such, care should be taken to minimize the likelihood of an event in cases of high-risk patients, potentially by enlisting providers familiar with TRA. Lastly, the current gaps in knowledge require further research beyond the cardiology literature to determine preventative guidelines on periprocedural stroke.

Conclusion

While transradial approaches are compelling, more data are required to draw firm conclusions on its comparisons to TFA regarding stroke rates, as periprocedural stroke from any catheter-based intervention remains uncommon. Despite the low incidence, the associated morbidity of stroke warrants that each patient be objectively evaluated to determine which access is most appropriate for patient outcomes and safety.