Prevention and management of complications is a cornerstone of surgery. As endoscopic endonasal surgery gains popularity and is becoming the standard of care for sellar, parasellar, and medial middle fossa skull base access, its complications have been well studied with strategies that have improved over the last decade. Internal carotid artery (ICA) injury is a potentially devastating injury. Management algorithms and training paradigms have evolved for these injuries. Preoperative planning is paramount and may include a balloon test occlusion of the ICA. In this chapter we review the modern literature as well as the University of Pittsburgh series. It is clear that experienced teams with premeditated ICA injury protocols can provide effective management with potentially good outcomes. This chapter serves to review key elements in understanding vascular and endoscopic challenges, preoperative planning, landmarks, rescue steps as well as postoperative management and complication avoidance in the setting of middle fossa ICA injury. These principles are important to incorporate into training of all levels.
Key wordsICA injury – internal carotid – middle fossa – endoscopic endonasal – multidisciplinary approach
10 Dealing with Major Intraoperative Vascular Injury
10.1 Key Learning Points
It is rare that internal carotid artery (ICA) injury is the result of a single error; multiple factors lead to poor outcomes.
Preoperative checklist, imaging, realistic goals of surgery, and possibly balloon test occlusion are necessary to consider.
Dual surgeons with dynamic endoscopy and bimanual dissection are key to avoiding and managing injury.
Bipolar cautery, clip-based repair, and packing strategies should all be available options for repair of ICA injury.
Muscle packing over an area of injury is often the most reliable technique and should be considered the first option for any sizeable or difficult to control injury.
Cerebral perfusion is the end goal and should not be compromised. Indirect measurement with neurophysiologic monitoring can assist in decision-making.
A multidisciplinary team is important for postoperative management.
Digital subtraction angiography is necessary to diagnose and treat immediate and late complications.
Although never the first option, if hemorrhage becomes truly life threatening, carotid sacrifice should be considered. A majority of patients will tolerate sacrifice of a single ICA.
Older patients may tolerate sacrifice less well due to loss of collateral.
Endoscopic endonasal surgery (EES) for treating paranasal and skull base pathology has grown in popularity over the last two decades, largely replacing microscopic approaches. 1 EES has led to decreased morbidity and improved patient outcomes secondary to the numerous advantages afforded by improved visualization, midline corridor with decreased neurovascular manipulation, avoidance of skin incision, and reduced postoperative pain with shorter hospital stays. 2 , 3 , 4 The evolution of EES has gone well beyond midline pituitary surgery to encompass an almost unrestricted access to the ventral skull base. Despite multiple advances in approach, technique, and equipment, much fear remains for the inexperienced operator when dealing with some of the potential complications, especially those of a vascular nature.
Encountering the ICA is extremely common during ESS. The lateral walls of the sella have a close relationship with the ICA. Protrusion of the ICA into the sphenoid sinus due to pneumatization has been reported to occur in about 70% of patients and frank dehiscence within the sphenoid sinus occurs in about 22% of patients. 5 Furthermore, advanced pathologies including sinonasal, middle, and posterior cranial fossa lesions may require exposure and/or mobilization of the ICA. This includes lesions of the petrous apex, cavernous sinus, clivus, infratemporal fossa, and parapharyngeal space all of which may have increased risk of injury.
ICA injury may result in varying degrees of morbidity and even death. Early complications include overwhelming hemorrhage/blood loss or carotid occlusion. Delayed manifestations include formation of pseudoaneurysm, vessel spasm, thrombosis, embolization, or formation of a caroticocavernous fistula which can all further complicate repair or treatment of the underlying pathology. In addition, repair or interventions aimed at treating large vessel injury have independent risk profiles that each carry additional risks and benefits. Therefore, the prevention and management of such injuries remain important considerations for EES operators in both the short- and long-term manifestations of ICA injury.
Injury to the ICA most often occurs within the middle fossa likely owing to the majority of pathologies suitable for EES occurring centrally (i.e., pituitary tumors). The incidence of accidental injury of the ICA during traditional skull base surgery has been reported to range as high as 3 to 8%. 6 Although the rate of ICA injury during EES pituitary surgery is low by comparison (less than 1%) 3 and has been documented largely in case reports, the incidence during expanded EES (those cases with exposure beyond the face of the sella) has a higher risk profile. These estimations are subject to publication bias, as systematic accounting of ICA injuries in the literature are lacking. A recent survey of experienced surgeons while attending the University of Pittsburgh Medical Center (UPMC) Skull Base Surgery Course reported that 20% of participants had encountered a carotid injury within the preceding 12 months. 7 Expanded approaches require wide exposure with the goal of further reach and more extensive resection with reports citing injury rates as high as 4 to 9%. 8 , 9 With the growing popularity of endoscopic endonasal cases ICA injury is in need of additional study.
Large vessel injury is a difficult to treat complication and has limited the novice from enjoying the maximal benefits of EES. A learning curve exists to successfully deploying EES across a broad set of pathologies with advantages that outweigh this potential injury. Training paradigms have been proposed to avoid and address potential large vessel injury and allow for stepwise progression of comfort and skill for operators. 1 , 10 In this chapter, the literature reporting ICA injury is reviewed, and our team’s own experience and a paradigm for complication avoidance and management of injury are presented.
10.3 Vascular Challenge
10.3.1 Anatomy of the Middle Fossa ICA
Within the middle fossa and skull base, the ICA has a three-dimensional course with consistent relationships to other anatomical landmarks. However, these may be distorted by pathology, prior surgery, irradiation, and anatomical variation. Understanding the impact of these factors and careful study of appropriate preoperative imaging studies to understand the relationship of each unique case is valuable.
Many classifications have been proposed, including the Fischer classification (1938) with a focus on describing the extracranial and intracranial segments, and the more modern and widely used classification schemes of Gibo et al and Rhoton as well as Bouthillier et al. 11 , 12 , 13 , 14 , 15 Yet, the key transcranial relationships do not fully address the cranial base from a ventral viewpoint. Alfieri and Jho reported a meaningful attempt to classify the segments of the ICA from an endoscopic endonasal perspective; however, their classification has multiple limitations. 16 Namely, it excludes segments of the ICA proximal to the foramen lacerum, and relies on sphenoid sinus pneumatization, a feature that is highly variable, for identifying landmarks of the ICA. Labib et al introduced a comprehensive classification scheme that is applicable to the expanded endonasal approaches. 17 These segments correspond to consistent anatomic landmarks that can be identified. These landmarks have been refined and ICA segments correlated to create a more modern nomenclature for endonasal operators (▶ Table 10.1). Wide variability in nomenclature is still found in the neurosurgical literature which can create uncertainty when converting between systems of transcranial, endovascular, or radiology-based naming schema. Preoperative imaging of vessels and knowledge of anatomical landmarks can be used in conjunction with Doppler ultrasound, indocyanine green fluoroscopy, optical/electromagnetic navigation, and contralateral anatomical symmetry to define patient-specific ICA anatomy along with pathology.
10.3.2 Endoscopic Aspects/Challenges
ICA injury management is technically challenging given the fact that it is a large, high-flow (estimated to be around 200–300 mL/min), high-pressure artery that if injured immediately compromises the surgical field. 18 This is amplified when using endoscopic visualization in a narrow corridor. Massive bleeding leads to a loss of orientation and obscures identification of the injury point. Restoration of visualization requires stepwise maneuvers to regain control of the field. Moreover, the lack of direct endonasal vascular suture repair excludes traditional treatment options. In addition, for novices, the optical characteristics of the endoscope, which are vastly different from those of the microscope, may make endoscopic anatomy particularly challenging. Specifically, the two-dimensional view of the endoscope interferes with traditional depth perception and has to be overcome with dynamic endoscopy relying on instrument relationships. These factors together can alter the perspective as well as depth perception of the ICA (Fig. 10.1) and, combined with the nasal corridor, create a perceived lack of control. 17
10.4 Injury Avoidance
“Intellectuals solve problems, geniuses prevent them” (Albert Einstein)
Each clinical scenario, with unique anatomical variations and pathological distortions, needs to be thoroughly evaluated preoperatively. This starts with comprehensive neurological and rhinological assessments (in addition to anesthesia, endocrinology, ophthalmology, or others as deemed necessary). It is worthwhile to evaluate any high risk or complex case in a multidisciplinary clinic and tumor board setting (neuroradiology and even oncology as needed) where surgical planning is discussed between the primary surgeons to develop the operative plan including surgical challenges, overall surgical goal, reconstruction plan, and risk stratification. This plan is then further discussed with the operating room (OR) staff and anesthesia team for operative day readiness. Furthermore, these risks and potential complications are discussed with the patient on more than one occasion. Even the rare complication of large vessel injury is discussed in detailed with each patient as part of our informed consent process.
Computed tomography angiography (CTA) and magnetic resonance imaging (MRI) are part of the standard imaging protocol. Both modalities are important to highlight bony anatomy in relation to vascular anatomy as well as soft tissue and pathological involvement by MRI assessment. Devoted MR sequences are tailored to relevant anatomy with special skull base sequences (fast-spin echo, steady-state free procession, fine cut T2) or pituitary sequences (dynamic) depending on the clinical scenario. Unexpected pathology should always be considered when approaching sellar pathology (e.g., poorly pneumatized bone, dehiscence, ectatic/kissing carotids, carotid-clinoid ring, tumor invasion into adventitia, etc.). Typically, preoperative imaging can delineate vascular abnormalities such as an aneurysm, pseudoaneurysm, or carotid-cavernous fistulae. If vascular abnormalities exist or cranial circulation is unclear, they can be further delineated with digital subtraction angiography.
Case selection depends on the experience of the surgical team. High-risk cases should be considered for a balloon test occlusion (BTO) of the ICA to assess collateral cerebral blood flow and evaluate for tolerance of sacrifice. BTO, however, can be associated with a 5 to 10% false-negative result. Imaging adjunct with perfusion CT, MR single-photon emission computed tomography (SPECT), transcranial Doppler, or other methods can be used to confirm/improve the result or be utilized if patient factors prohibit preoperative endovascular testing. BTO prior to high-risk ICA cases reduces the incidence of postoperative stroke compared with indiscriminate ICA sacrifice. 19 , 20 This topic is discussed in detail in Chapter 3.
In general, good surgical technique is also key in avoiding ICA injury; ensuring excellent visualization, orientation, keeping a clean surgical field, and proper working instruments are important factors. Experience and teamwork between the two surgical teams and an operative plan will allow seamless execution and avoid off-course surgery. Subcapsular tumor dissection and proper skeletonization of key anatomy also play a role so that normal anatomy can be maintained with delicate manipulation of sensitive structures. Gross ballistic movements or aggressive biting of bone should also be avoided as safer access can be achieved with diligent drilling or “egg shelling” of delicate anatomy, including the ICA. Exposure should be adequate to ensure proximal control in the event of an injury.
Neurophysiology is of critical importance as it serves as the only measure of cerebral perfusion while under anesthesia. Somatosensory evoked potentials (SSEPs) remain the mainstay of hemispheric perfusion which is important when employing interventions. In addition to guiding decisions on which interventions can be deployed, it is useful to guide management by the anesthesia team as further discussed in the following. The routine use of SSEPs should be considered on all endoscopic endonasal cases. Electroencephalography (EEG) can also be added in high-risk cases to aid in intraoperative decision-making if perfusion is compromised. Chapter 21 thoroughly discusses neurophysiological considerations associated with skull base vascular concerns.
10.4.2 Anatomical Risk
The relationship of the ICA to the sphenoid sinus has been morphometrically detailed in the neurosurgical and skull base literature. 21 , 22 Large vessel course and pneumatization patterns can be identified on preoperative imaging. Dehiscent ICA in the sphenoid sinus has been reported to occur in 4 to 22% of cases with 8 to 70% of patients having an ICA prominence into the sphenoid sinus. 23 The average intercarotid distance of the parasellar segment ICAs has been reported as 12 mm but can be less than 4 mm. These facts alone are not risk factors and if properly incorporated into a surgical plan can aid in early identification. 22 Bony septations within the sphenoid sinus can attach in a number of patterns, including midline, on the ICA, optic nerve, or multiple locations. However, it is important to realize that almost all septations (89%) extend or attach to bone over the ICA at some point. 24 These septations should be removed with caution, typically with the use of a high-speed drill, to avoid fracturing into the encased soft tissue at the distal aspect of the bone.
10.5 Impact of Pathologies
Ectasia or distortion of the anatomical course of the ICA by tumor compression or invasion may place the vessel at risk. This may or may not include adventitia invasion that can be seen with aggressive tumors (Fig. 10.2). Invasion or flow-limiting stenosis of the ICA increases the risk and, depending on the pathology and goals of surgery, may necessitate a BTO in preparation for sacrifice or potential injury.
Specific pathologies have been suggested to pose additional risk for ICA injury. Tumors that have intimate contact with the ICA can disturb tissue planes which can increase risk of inadvertent injury. Additionally, lesions that extend beyond a single segment of the ICA (≥2), requiring exposure or mobilization of the carotids, is a cause to consider the case at an elevated risk as proposed by Al Qahtani et al. 25 , 26 Therefore, middle fossa and posterior fossa origin or growth extension pose a significant risk in comparison to anterior cranial fossa tumors. Encasement of the ICA by more than 120 degrees has been proposed as a risk factor for potential injury as it suggests loss of tissue planes, and early identification of surrounding anatomy proves more difficult. If the pathology is within the chondroid tumor family (chordomas and chondrosarcomas), there is an elevated risk profile for ICA injury; 3 this may also be related to the fact that these tumors are best treated with radical resection as they typically abut if not encase the ICA crossing periosteal planes. Hence, radical resection with curative intent is in itself a risk factor for ICA injury.
Finally, tumors may secrete factors that can affect the major vasculature. A common example is growth hormone (GH) secreting pituitary adenomas which are known to cause vascular ectasia with increased risk of ICA injury. 27
Utilizing the training paradigm proposed by Snyderman et al on the acquisition of surgical skills for endonasal skull base surgery, surgeons should be well prepared before tackling cases where ICA injury is at a more than moderate risk. 1 All endonasal surgery teams should be prepared and have forethought about a potential vascular injury; but, as case complexity increases and anatomical reach brings the likelihood of ICA injury to the forefront, management should be within a surgical team’s skill set before embarking. This is also true for trainees who are supervised by senior surgeons; stepwise progression of skill and judicious supervision are required to achieve a balance of training the next generation of surgeon while ensuring a safe and effective outcome. Preparation can be furthered by mentorship with a more senior surgeon, training courses, cadaveric dissection, and training models of ICA injury/repair. 28
Surgical planning for cases that involve tumor invasion or distortion of the ICA requires planning a wide corridor that allows maximal working space for free bimanual dissection. Careful sharp dissection is used in areas of high risk with small careful movements, which can potentially allow for repair if a minor injury occurs. Dual surgeons with dynamic endoscopy (i.e., not using a static endoscope holding device) can play an immense but difficult to quantify role. Constant movement of the endoscope allows for an ideal view and ensures that endoscope placement does not impede dissection movement. In addition, constant cross-checking between two experienced surgeons working side by side prevents inevitable single operator error. Instrumentation choice or lack of proper instrumentation can also play a role in ICA injury. Typically, the drill is used to thin bone until it can be “flaked” away and is used with a “coarse” diamond bit (≤4 mm) and copious irrigation to prevent thermal injury and increase visualization. This bit is used initially on areas other than close to the ICA which serves to slightly dull it before it is used to blue-line over the ICA. We have also found great success with the occasional use of a “minimally invasive” drill attachment (MIS curved 13 or 16 cm) that has a curved shaft and protected drill bit shaft that can improve reach and visualization. Other powered instruments such as ultrasonic aspirators can be of value for focal bone removal in close proximity to the ICA but should be used with a great deal of caution in combination with frequent assessment with Doppler ultrasound, image guidance, and even indocyanine green (ICG) fluoroendoscopy. Microdebriders and monopolar electrocautery should only be used outside of the sphenoid sinus. Use of Kerrison or pituitary rongeurs can also be high-risk maneuvers as they allow for pulling or biting of poorly visualized tissue. In addition, all necessary instrumentation should be present in case of injury, for example, hemostatic agents or bipolar cautery devices. It is essential to have knowledgeable scrub nurses that understand the surgical methods and the proper set-up and handling of instrumentation.
Additional factors that add risk but are hard to predict include the degree of adherence of the lesion to the ICA which can be influenced by previous surgery, irradiation, or bromocriptine therapy. 29 These have all been reported to be risk factors in ICA injury. 25 , 30 Given that these factors will likely require real-time operative assessment, the goals of surgery should be considered if manipulation of nearby adherent structures proves to add significant risk without clear clinical benefit to the patient. Even though one proximate cause of injury can be identified, multiple factors are at play. A disastrous complication is seldom the result of an isolated circumstance and is likely multifactorial.
10.6 Case Example
A recent review of the running case series at the University of Pittsburgh has identified a total of 18 ICA injuries, representing an event rate of 0.46% (n = 18/3889). 17/18 cases involved tumors that had grown beyond the sella and all were considered Level III, IV, or V in complexity (as described by Snyderman et al) meaning that all cases were of an advanced degree of difficulty on the learning curve for endoscopic endonasal skull base surgery. The most frequent location was in the cavernous segment (n = 7, 39%) followed by the paraclival segment (n = 5, 28%). Injury most commonly occurred during tumor dissection (n = 10, 56%). Pathologies included adenoma (n = 5, 28%), chordoma (n = 5, 28%), meningioma (n = 5, 28%), and three others (17%). Five patients (28%) had prior surgery with three (17%) having undergone prior irradiation. Bipolar electrocautery was attempted in all cases of ICA injury but most required other treatments. Aneurysm clips were used in nine cases (50%) and packing (muscle and/or cotton) in the remainder. Clipping or packing was combined with a muscle patch in six (33%) cases. All cases underwent immediate postoperative digital subtraction angiography (DSA) which in 14 cases (78%) required no subsequent intervention. Of the four remaining cases, two (11%) were treated with coil embolization, one (6%) with stent placement, and one (6%) with thrombectomy. At 1-month follow-up, pseudoaneurysm formation was detected in three (17%) cases, one treated with observation and two with stent placement. Two (11%) of the ICA injuries resulted in death with the remainder having no neurological deficits at 1-month postoperative.
We further highlight a video case that reviewed elements that contributed to injury (Video 10.1).
10.7 Management Strategy/Management Algorithm
The first step of ICA injury management is realization that an arterial injury has occurred. To the novice, minor arterial bleeding may be mistaken as venous bleeding and managed inappropriately with injection of hemostatic material. The cavernous sinus can bleed profusely and in a pulsatile fashion; typically, this is significantly less than that of the ICA or a perforator and the blood of the ICA has a distinctive arterial color. Whenever bleeding is encountered from the medial cavernous sinus, the question of origin should be addressed and never assumed. Further, when an ICA injury occurs, the lens of the endoscope is often splashed with blood and loss of visualization occurs either from direct contamination or when the surgical field fills with blood.
Once it has been identified that a large caliber vessel has been injured, the surgeon should notify the OR team. In an ideal situation, this should activate a readily available protocol of actions (Fig. 10.3). The core of this includes the surgeons, the anesthesia team, and the OR staff (scrub and circulating nurse, neurophysiologist). The subsequent steps need to occur in an effective manner and often simultaneously.
10.7.1 The Surgeons
The surgeons should regain control of the surgical field with suction to clear the blood and identify the bleeding point for further assessment. This is best done with two surgeons working to dynamically clear the blood and maintain view. Four hands allow for the endoscope, two suctions, and a working instrument to be deployed in concert, which may be necessary as single suction may be insufficient. The dual suctions should be on separate lines so that there is always one functioning suction.
Direct pressure on the injury site should be the next action to reduce blood loss and deploy the plan. The area of pressure should be as focused as possible, narrowing down to using a single cottonoid, or focal packing material to cover the injury point with suction so that subsequent salvage techniques can be attempted. Indiscriminate packing should be avoided to avoid injury to adjacent structures and minimize risk of blood tracking intracranially through a dural defect. If all else fails, packing should be judiciously employed to stop exsanguination for ensuing emergent neuroendovascular intervention. Packing is the bail-out option if the clinical scenario deteriorates (Fig. 10.4). Packing options can include muscle (preferred first layer on the artery), resorbable (gelfoam, Nasopore, Posisep) or nonresorbable (cotton based) materials, Merocels, and balloons (Foley) depending on the clinical situation.
If the bleeding point can be controlled and visualized, primary attempts at sealing the vessel injury should be undertaken. Bipolar cautery can be attempted first depending on the size of the injury. Small side wall injuries or perforator avulsions can be sealed, as previously described by Kassam et al 31 who illustrated longitudinal violation of the side wall being repaired with cautery. This is done by placing a suction directly onto a segment of the artery which allows for bipolar cautery to create a seal. The suction and cautery are then moved together along the injury. Given the size of the ICA, this ideally preserves flow and distal perfusion with minimal stenosis. Cautery should be an early line attempt with the goal of salvage while maintaining patency but can also be used if needed as a method of ligation to seal off the artery in case of an irreparable injury.
10.7.3 Suture/Clip Repair/Ligation
If bleeding cannot be easily controlled or to provide more distinct control, additional access to more proximal and distal points on the artery should be exposed in order to attempt more advanced salvage techniques. This can provide sites of temporary or permanent ligation. If properly planned, proximal and distal control may allow for trapping of the artery. If not, consideration of gaining such control will depend on the site of injury and/or tumor involvement of proximal or distal segments. For example, injury of the lacerum segment of the ICA has little hope for endonasal proximal control and wide cavernous encasement may prevent direct identification of proximal or distal control points. For proximal control in such cases, pre-planning with a small cervical cutdown to the ICA can provide for rapid and definitive proximal control (external carotid artery [ECA]-ICA distal collaterals notwithstanding) (Fig. 10.5). Compression of the ipsilateral ICA in the neck should always be performed to decrease flow and occasionally provide some degree of proximal control. This must be performed by a physician who is confidant in the location of the cervical ICA and familiar with the degree of compression required. Proximal endovascular control may be an option but is time consuming and most endovascular interventions are difficult or impossible using only C-arm fluoroscopy available in an OR. A hybrid suite may overcome this limitation but has its own challenges even if available.
Primary clip repair presents an attractive salvage technique. Straight or angled clips can be applied directly to the side wall to “pinch off” the area of injury and leave a reduced lumen with patency. If repair of the vessel wall defect cannot be achieved, then clip ligation is a reasonable option. However, this can present multiple dilemmas as it may affect cerebral perfusion and definitively closes off any further endovascular access to the site of injury. Furthermore, proximal occlusion will not control retrograde bleeding. SSEPs and mean arterial pressure (MAP) should be judiciously assessed. If test occlusion is performed, relative hypotension should be included to see if SSEPs show any pressure dependence.
Multiple commercial endoscopic (single shaft) clip appliers (Mizuho, Lazic) are available that allow for placement of a variety of clips at different working angles for precise application of clips with the intent of repair or ligation (Fig. 10.6). Sundt-Keyes clips can also be applied in an attempt to secure the injury but are often large and may not line up well with the tortuous course of the carotid siphon. Familiarization with your institution’s resources should be routinely undertaken with the team.
Direct suturing is another extraordinarily advanced technique that can be attempted; however, it has the drawback of being technically demanding and time consuming. Special commercially available needle holders and tying instruments are available but technically demanding to use (Fig. 10.6). This is not recommended unless the surgical team has dedicated significant time in a cadaver lab or other model to ensure that they can reliably apply this technique. In addition, it requires definitive proximal and distal control as ongoing bleeding would make this technique virtually impossible.