Vertebral artery anomaly and injury in spinal surgery.

STUDY
DESIGN: Systematic review. STUDY RATIONALE: The purpose of this review is to further define the published literature with respect to vertebral artery (VA) anomaly and injury in patients with degenerative cervical spinal conditions.
OBJECTIVES: In adult patients with cervical spine or degenerative cervical spine disorders receiving cervical spine surgery, what is the incidence of VA injury, and among resulting VA injuries, which treatments result in a successful outcome and what percent are successfully repaired?
MATERIALS AND METHODS: A systematic review of pertinent articles published up to April 2013. Studies involving traumatic onset, fracture, infection, deformity or congenital abnormality, instability, inflammatory spinal diseases, or neoplasms were excluded. Two independent reviewers assessed the level of evidence quality using the Grades of Recommendation Assessment, Development and Evaluation criteria; disagreements were resolved by consensus.
RESULTS: From a total of 72 possible citations, the following met our inclusion criteria and formed the basis for this report. Incidence of VA injuries ranged from 0.20 to 1.96%. None of the studies reported using preoperative imaging to identify anomalous or tortuous VA. Primary repair and ligation were the most effective in treating VA injuries.
CONCLUSION: The incidence of VA injuries in degenerative cervical spinal surgery might be as high as 1.96% and is likely underreported. Direct surgical repair is the most effective treatment option. The most important preventative technique for VA injuries is preoperative magnetic resonance imaging or computed tomography angiographic imaging to detect VA anomalies. The overall strength of evidence for the conclusions is low.

Study Rationale and Context

There is a paucity of published literature addressing the topic of vertebral artery (VA) anomaly and VA injury in degenerative cervical spine surgery. The incidence and risk factors for VA injury in degenerative spinal surgical procedures have not been adequately defined. The purpose of this study is to review the existing literature and to provide an accurate description of the published literature regarding variation of the normal VA anatomy and injury to the VA during degenerative cervical spinal surgery.

Key Question (KQ) 1: In studies of adult patients with conditions warranting cervical spine surgery, what formal classification systems of VA anomalies based on imaging studies exist, and are these classification systems reliable?

Objectives

In adult patients with cervical spine or degenerative cervical spine disorders receiving cervical spine surgery:

KQ 2: What is the incidence of VA injury?

KQ 3: Among the resulting VA injuries, which treatments result in a successful outcome and what percent are successfully repaired?

Materials and Methods

Study design: Systematic review.

Search: The databases included PubMed, Cochrane, and National Guideline Clearinghouse Databases; bibliographies of key articles.

Dates searched: The data were searched through April 19, 2013.

Inclusion criteria: Patients 18 years or older with cervical spine, degenerative cervical spine disorders, or other conditions warranting cervical spine surgery.

Exclusion criteria: Studies in patients younger than 18 years, those involving traumatic onset, cervical fracture, infection, deformity or congenital abnormality, instability, inflammatory spinal diseases, or neoplasms; case reports, comparative studies with fewer than 10 patients per treatment group; cadaveric studies; nonhuman in vivo, in vitro, and biomechanical studies.

Outcomes: Classification systems of VA anomalies; incidence of VA injuries; and successful outcome of VA injury treatment (no residual neurologic or vascular symptoms, including stroke, pain, neurologic deficit, or death).

Analysis: Descriptive statistics.

Overall strength of evidence: Risk of bias for individual studies was based on using criteria set by The Journal of Bone and Joint Surgery 1 modified to delineate criteria associated with methodological quality and risk of bias based on recommendation from the Agency for Healthcare Research and Quality (AHRQ).2 3 The overall strength of evidence across studies was based on precepts outlined by the Grades of Recommendation Assessment, Development and Evaluation Working Group4 and recommendations made by the AHRQ.2 3

Details about methods can be found in online supplementary material.

Results

A total of 696 unique citations were retrieved for all three KQs. From a total of 72 citations evaluated for full-text review, 15 met the inclusion criteria for this report (Fig. 1).

Fig. 1

Flow chart showing results of literature search.

KQ 1: Formal Classification Systems of VA Anomalies

We identified three case series that presented a classification system or a definition of VA anomalies.5 6 7

Eskander et al (2010)5 classified VA anomalies for C2 to C7 into three categories: intraforminal (midline migration of the VA, including tortuosity), extraforaminal (VA not contained within transverse foramen from C6 to C2), and arterial (VA that is fenestrated, hypoplastic, or absent). Eskander et al also suggested assessing VA abnormalities on magnetic resonance imaging (MRI) rather than plain radiograph and proposed a modification to Oga's classification for C1 to C7 VA tortuosity to account for arterial abnormalities (classified as normal or abnormal).

Hong et al (2008)6 described anomalies of the V2 segment of the VA as: abnormal entrance (entrance into C4, C5, or C7 transverse foramen), V2 segment anomalies (bilateral or unilateral), and other variations of the VA (fenestrations or foraminal erosions).

Oga et al (1996)7 classified the tortuosity of the VA from C1 to C7 into four categories based on the VA's relationship to three zones of the cervical spine. The categories are as follows: type 1 (straight and within zone I), type 2 (mildly tortuous and in zone I), type 3 (loop formation and most medial portion in zone II), and type 4 (loop formation and migration in zone III). Refer to Table 1 for definitions of these zones.

Table 1

Studies reporting classification systems of VA anomalies (KQ 1)

Author (y)Study design	DemographicsDiagnosis	Classification system of VA anomalya	Method of assessing VA anomaly	Notes
Eskander et al (2010)Case series	• N = 250 patients• Age (mean): 49 y (9–88)• Male: 45%• Axial neck pain, radiculopathy, myelopathy (n = NR)	VA anomalies for C2–C7:• Intraforaminal abnormalities: midline migration with VA located either medial to or < 1.5 mm lateral to the uncovertebral joint• Extraforaminal abnormalities: VA not contained within respective transverse foramen from C6 to C2• Arterial abnormalities: fenestrated, hypoplastic, or absent	• MRI from base of skull thru T2, measurements obtained from C2 to C7• IVAD• MVAD• UJVA• VAD	Author suggests modifying Oga's classification for VA tortuosity to account for arterial abnormalities and to assess the VA on MRI, not radiograph. The following is a proposed addition to Oga's classification:• “A” for normal VA (≤ 2 mm size differential)• “B” for abnormal VA (hypoplastic/absent or fenestrated arteries)
Hong et al (2008)Case series	• N = 350 patients (700 VA)• Age (mean): 54.8 y (16–89)• Male: 39%• Included: patients who underwent CT angiography for reasons other than VA disease• Excluded: bony abnormalities (Klippel–Feil syndrome and ankylosing spondylitis) and VA aplasia	V2 segment of VA:• Abnormal entrance: entrance into C4, C5, or C7 transverse foramen• V2 segment anomalies: bilateral and unilateral• Variations of VA, including fenestrations or foraminal erosions	• CT angiography	Author also includes measurements between the extraosseous portions of the VA to surgical landmarks
Oga et al (1996)Case series	• N = 23 patients• Cervical myelopathy from cervical spondylosis or cervical disc herniation (n = 23)	VA tortuosity for C1–C7:• Type 1: VA straight and in zone I• Type 2: mildly tortuous and in zone I• Type 3: loop formation and most medial portion in zone II• Type 4: loop formation and migration in zone III	• Radiograph, MRI of C1–C7, vertebral arteriography, or MRI angiography, CT scan

Abbreviations: CT, computed tomography; KQ, key question; IVAD, intervertebral artery distance; MRI, magnetic resonance imaging; MVAD, midline VA distance; UJVA, uncovertebral joint VA distance; VA, vertebral artery; VAD, VA diameter.

V2 segment defined as being between C5 and C2 transverse process (Hong et al, 2008); zones from the anteroposterior view for VA tortuosity defined as: zone I (outside of the lateral end of Luschka joint), zone II (between the lateral and medial end of Luschka joint), and zone III (medial of Luschka joint) (Oga et al, 1996).

None of the included studies conducted a formal analysis of the reliability of the proposed classification systems.

KQ 2: Incidence of VA Injuries

Incidence of VA injuries resulting from degenerative cervical spinal surgery ranged from 0 to 1.96% as reported in 12 studies. A total of 10 VA injuries were reported (Tables 2 and 3).

Table 2

Studies reporting frequencies for and outcomes of VA injuries resulting from cervical spine surgery (KQs 2 and 3)

Author (y)Study designCoE	Demographicsa Diagnosis	Type of surgery	Incidence of VA injuriesb VA affected	Diagnosis, treatment, and outcome from VA injury	Potential risk factors for VA injury
Anterior procedures
Bilbao et al (2010)Case seriesCoE: IV(KQs 2 and 3)	• N = 71• Age (mean): NR (> 50 y: n = 48, < 50 y: n = 23)• Male: 63%• Spondylotic spinal cord compression with tetraparesis (n = 29), pain (n = 21), motor paresis (n = 12), and paresthesia (n = 9)	Corpectomy with titanium mesh or telescopic cage and autograft:• 1 level (n = 46), 2 levels (n = 25)• right-side approach (n = 71)	• 1.4% (1/71)• VA affected: left (n = 1)	Patient 1:• Intraoperative diagnosis (details NR). Treatment: embolization above and below damaged site. Outcome: multiple medullar, cerebellar, and supratentorial infarcts; at 18 mo patient is ambulatory with aid and has normal deglutition
Lu et al (2008)Retrospective cohortCoE: III(KQs 2 and 3)	• N = 51• Age (mean): 61.0 y (37–86)• Male: 75%• Multilevel degenerative cervical myelopathy with upper extremity paresis, walking and balance problems, progressive myelopathy, persistent radicular pain Aged group: n = 20, mean age 75 y (70–86) Control group: n = 31, mean age 52 y (37–68)	Corpectomy with titanium mesh and autograft	• 1.96% (1/51)• VA affected: NR	Patient 1:• Intraoperative diagnosis (details NR). Treatment: ligation. Outcome: patient recovered after blood transfusion	VA injury occurred in one patient in the aged group
Pechlivanis et al (2008)Retrospective cohortCoE: III(KQ 2 only)	• N = 90• Age (mean): 47.9 ± 8.7 y (29–69)• Male: 50%• Cervical degenerative disc disease with soft-, hard-, or soft/hard-disc pathology Soft-disc group: n = 49, mean age 47.9 y (29–69) Hard-disc group: n = 24, mean age 47.8 y (36–65) Soft/hard-disc group: n = 17, mean age 46.7 y (41–65)	Single-level uncoforaminotomy using Jho technique with Saringer modification	• 1.11% (1/90)• VA affected: NR	Patient 1:• Intraoperative diagnosis (details NR, surgery discontinued). Treatment: NR. Outcome: no occlusion documented on postoperative angiography, no neurological deficits experienced, successful reoperation at 3 mo	The following techniques were used to protect the VA: the Saringer surgical modification was used, drilling was conducted using a diamond drill, and a thin layer of the cortical bone of the lateral wall of the uncinate process was preserved.VA injury occurred in one patient in the hard-disc pathology group
Choi et al (2007)Case seriesCoE: IV(KQ 2 only)	• N = 20• Age (mean): 48.7 y (37–74)• Male: 75%• Cervical radiculopathy from soft disc herniation (n = 7), spondylotic stenosis with osteophytes (n = 3), soft disc with osteophytes (n = 10)	Modified transcorporeal anterior cervical microforaminotomy (transverse skin incision made at 1 level higher than affected disc level)	0% (0/20)	n/a	Author claims that with this technique VA artery is not exposed or endangered
Fountas et al (2007)Case seriesCoE: IV(KQ 2 only)	• N = 1,015• Age (mean): 56.3 y (28–75)• Male: 54%• Cervical radiculopathy and/or myelopathy secondary to degenerative disc disease and/or cervical spondylosis	ACDF:• Right-side approach 93.8%, left-side approach 6.2%	0% (0/1015)	n/a
Sasso et al (2007)RCTCoE: II(KQ 2 only)	• N = 115• Age (mean): 44.0 y (25.1–66.9)• Male: 54%• Single-level cervical degenerative disease causing radiculopathy or myelopathy Bryan disc group: n = 56, mean age 42.5 ± 7.8 y (25.1–64) ACDF group: n = 59, mean age 46.1 ± 7.8 y (29.4–66.9)	• Single-level cervical arthroplasty with Bryan cervical disc prosthesis (n = 56)• Single-level ACDF with allograft and plate (n = 59)	0% (0/115)	n/a
Burke et al (2005)Case seriesCoE: IV(KQs 2 and 3)	• N = 1,976 patients• Age (mean): NR• Male: % NR• Herniated or degenerative disc disease or spondylosis	Anterior cervical spine procedures (right-side approach) with VA injuries occurring during:• ACDF, 1 level (n = 1)• ACDF, 3 levels (n = 1)• Anterior cervical corpectomy + fusion reexploration (n = 1)• Anterior cervical corpectomy + fusion (n = 3)	• 0.3% (6/1,976)• left VA (n = 5), right VA (n = 1)	Patients 1 and 2:• Diagnosis: bright intraoperative arterial bleeding. Treatment: primary repair. Outcome: no sequelae.Patient 3:• Diagnosis: bright intraoperative arterial bleeding. Treatment: ligation. Outcome: no sequelae.Patient 4:• Diagnosis: bright intraoperative arterial bleeding. Treatment: tamponade with thrombin-soaked Gelfoam. Outcome: no sequelae.Patient 5:• Diagnosis: bright intraoperative arterial bleeding. Treatment: tamponade with thrombin-soaked Gelfoam (Pfizer, New York, NY, USA). Outcome: hemodynamic instability from hypovolemia resulting in intraoperative death.Patient 6:• Diagnosis: intraoperative epidural oozing. Treatment: tamponade with thrombin-soaked Gelfoam and postoperative anticoagulation with heparin. Outcome: posterior inferior cerebellar artery occlusion, lateral medullary infarct, and syndrome.	Operating microscope used in 83% (5/6) of surgeries resulting in VA injury.Intraoperative CT scan was not performed in any of the surgeries that incurred a VA injury
Shen et al (2004)Retrospective cohortCoE: III(KQ 2 only)	• N = 109 patients• Age (mean): 46.3 y (27–83)• Male: 60%• Radiculopathy secondary to herniated disc and/or spondylosis	ACDF (Smith-Robinson left-side approach):• Uncovertebral joint decompression (n = 71)• Indirect decompression by means of disc space distraction (n = 38)	0% (0/109)	n/a
Graham et al (1996)Case seriesCoE: IV(KQ 2 only)	• N = 21 patients• Age (mean): 52.4 ± 15.2 y (28–81)• Male: 81%• Degenerative spondylosis (n = 17), congenital stenosis (n = 2), trauma (n = 2), OPLL (n = 1)	Cervical arthrodesis and stabilization with lateral mass plate (anterior approach n = 20, posterior approach n = 1)	0% (0/21)	n/a	17% (5/29) screws placed in central axial zone were malpositioned, placing the VA at risk
Bertalanffy and Eggert (1989)Case seriesCoE: IV(KQ 2 only)	• N = 450• Age (mean): 50 y (25–78)• Male: 72%• Cervical degenerative disc disease with radiculopathy (61%), pure myelopathy (16%), or combined myeloradiculopathy (23%)	Anterior cervical discectomy without fusion	0% (0/450)	n/a
Busch (1978)Case seriesCoE: IV(KQ 2 only)	• N = 138• Age (mean): 50.5 y (27–74)• Male: 61%• Cervical spondylosis from: soft prolapse (n = 49), segmental deficits (n = 54), compression of spinal cord (n = 29), amyotrophic lateral sclerosis (n = 3), spinal muscular atrophy (n = 1), spastic spinal paralysis (n = 1), plexus carcinoma polyneuropathy (n = 1)	ACDF with ICBG or Kiel bone graft	• 0.7% (1/138)• VA affected: NR	Patient 1:• Diagnosis: NR. Treatment: NR. Outcome: death	Author reports that VA injury occurred in an atypical VA, but no other details were given
Posterior procedures
Katonis et al (2011)Case seriesCoE: IV(KQ 2 only)	• N = 225• Age (mean): 68 y (45–84)• Male: 53%• Cervical spondylosis with myelopathy (n = 225)	Posterior cervical fixation using screw-plate and polyaxial screw-rod implant system	0% (0/225)	n/a	Suboptimal screw placement present in 0.6% (11/1,662 screws), but did not result in VA injury

Abbreviations: ACDF, anterior cervical discectomy and fusion; CT, computed tomographic scan; ICBG, iliac crest bone graft; KQ, key question; n/a, not applicable; NR, not reported; OPLL, ossification of the posterior longitudinal ligament; RCT, randomized controlled trial; VA, vertebral artery.

Demographics for patients with 24-month follow-up (Pechlivanis et al, 2008) or minimum 12-month follow-up (Choi et al, 2007).

Author reports 0.76% incidence of VA injuries, but it appears that intraoperative findings are presented for 90 patients with 24-month follow-up (Pechlivanis et al, 2008).

Table 3

Studies reporting frequencies for VA injuries resulting from cervical spine surgery

Surgical procedure	VA injury, % (n/N)
Anterior procedures
ACDF
Fountas et al (2007)	0% (0/1015)
Burke et al (2005)a	0.10% (2/1,976)
Shen et al (2004)	0% (0/109)
Busch (1978)	0.72% (1/138)
ACDF or cervical arthroplasty
Sasso et al (2007)	0% (0/115)
Cervical arthrodesis/stabilization with lateral mass plate
Graham et al (1996)	0% (0/21)
Cervical discectomy without fusion
Bertalanffy and Eggert (1989)	0% (0/450)
Corpectomy
Bilbao et al (2010)	1.41% (1/71)
Lu et al (2008)	1.96% (1/51)
Burke et al (2005), primary or re-explorationa	0.20% (4/1,976)
Modified transcorporeal microforaminotomy
Choi et al (2007)	0% (0/20)
Uncoforaminotomy
Pechlivanis et al (2008)	1.11% (1/90)
Posterior procedures
Cervical fixation using screw-plate/polyaxial screw-rod implant
Katonis et al (2011)	0% (0/225)

Abbreviations: ACDF, anterior cervical discectomy and fusion; VA, vertebral artery.

Author does not report the types of specific anterior cervical spine procedures included in the study other than those in which a VA injury occurred (Burke et al, 2005).

One RCT (CoE II),8 three retrospective cohort studies (CoE III),9 10 11 and eight case series (CoE IV)12 13 14 15 16 17 18 19 reported the incidence of VA injuries resulting from cervical spinal surgery. One of these 12 studies involved a posterior surgical procedure.19 Populations were predominantly male and middle aged. Sample sizes ranged from 20 to 1,976. Further details on the class of evidence rating for these studies can be found in the online supplementary material.

Incidence was highest overall in corpectomy procedures in three studies (1.96,9 1.41,10 and 0.20%14) and uncoforaminotomy in one study (1.11%).10

The left VA was injured in the majority of cases (86%, 6/7) in two studies reporting this level of detail. Burke et al (2005)14 reported that five of the six VA injuries occurred in the left VA during corpectomy (n = 3) or anterior cervical discectomy and fusion (n = 2). Bilbao et al (2010)13 reported that the one reported VA injury occurred in the left VA.

The majority of VA injuries were detected intraoperatively (80%, 8/10).9 10 13 14 Only one VA injury was detected postoperatively (10%, 1/10) in one study when the patient awoke with a lateral medullary infarct; subsequent cerebral angiography demonstrated a VA injury with posterior inferior cerebellar artery occlusion.14 Another study did not report the diagnostic workup of the VA injury (10%, 1/10).15

None of the studies reported the phase of surgery resulting in the VA injury.

None of the studies reported on the use of preoperative computed tomography (CT) or MRI to identify anomalous or tortuous VAs in their studies.

KQ 3: Treatment of VA Injuries and Successful Outcomes

No studies were identified that specifically addressed successful outcomes from the treatment for VA injuries. However, three studies identified from KQ 2 addressed the treatment of VA injuries.9 13 14 Overall, 63% (5/8) of the VA injuries were successfully repaired (Table 4).

Table 4

Treatments and success rates for VA injuries resulting from cervical spine surgery

Treatment for VA injurya	No. of patients receiving treatment	% patients with successful outcomeb
Primary repair (Burke et al, 2005)	2	100%
Ligation (Lu et al, 2008; Burke et al, 2005)	2	100%
Tamponade (Burke et al, 2005)	2	50%
Tamponade with postoperative anticoagulation (Burke et al, 2005)	1	0%
Embolization (Bilbao et al, 2010)	1	0%
Anticoagulation only	0	n/a
No treatment	0	n/a

Abbreviations: n/a, not applicable; VA, vertebral artery.

Two studies reported the outcome from treatment of VA injury (successful outcome in one patient (Pechlivanis et al, 2008), death in one patient (Busch, 1978), but did not report the treatment administered.

Successful outcome defined as: no residual neurologic or vascular symptoms, including stroke, pain, neurologic deficit, or death.

Primary repair and ligation were effective in treating VA injuries.9 14

○ Two successful primary repairs were reported by Burke et al (2005).14

○ Lu et al (2008)9 and Burke et al (2005)14 each reported one case of VA injury in which ligation repair was successful.

Mixed results were reported using tamponade, resulting in only one successful repair in one study.

○ One successful VA injury repair using tamponade was reported by Burke et al (2005).14

○Burke et al also reported one case of VA injury repaired by intraoperative tamponade. This patient suffered from hemodynamic instability from hypovolemia and subsequently died.

○A third VA injury was not detected until the patient suffered a postoperative lateral medullary infarct. The patient had received intraoperative tamponade for epidural oozing. Postoperative cerebellar angiography identified a VA injury with a posterior inferior cerebellar artery occlusion. This patient then received postoperative anticoagulation therapy.

Embolization was attempted to repair one case of VA injury.13 The patient suffered multiple medullar, cerebellar, and supratentorial infarcts; at 18-month follow-up, the patient was ambulatory with aid.

Anticoagulation therapy or no treatment: no studies reported these treatments.

Clinical Guidelines

None found.

The reported risk of VA in degenerative cervical spine surgery ranged from 0 to 2%. The overall strength of evidence for determining risk is low; that is, we have low confidence that the evidence reflects the true risk. Regarding treatments for VA injuries, three studies reported 5/8 (63%) successful repairs. The overall strength of evidence for this estimate is low (Table 5).

Table 5

Evidence summary

Baseline quality: High = majority of articles level I/II; low = majority of articles level III/IV Upgrade: Large magnitude of effect (1 or 2 classes); dose response gradient (1 class) Downgrade: Inconsistency of results (1 or 2 classes); indirectness of evidence (1 or 2 classes); imprecision of effect estimates (1 or 2 classes)
	Strength of evidence	Conclusions/comments
KQ 1: In studies of adult patients with conditions warranting cervical spine surgery, what formal classification systems of VA anomalies based on imaging studies exist and are these classification systems reliable?
Reliability of classification systems	Not applicable	None of the included studies conducted a formal analysis of the reliability of the proposed classification systems.
KQ 2: In adult patients receiving cervical spinal surgery, what is the incidence of VA injuries?
Incidence of VA injuries		One RCT, three retrospective cohort studies, and eight case series reported on VA injuries during cervical spine surgery. Seven studies reported no VA injuries. Five studies reported rates of 0.10 to 1.96%, depending on the type of anterior cervical spine procedure. No VA injuries were reported in the one study comprising posterior spine procedures.
KQ 3: Among VA injuries resulting from cervical spinal surgery in adult patients, which treatments result in successful outcome and what percent are successfully repaired?
Treatments resulting in successful outcome		One retrospective cohort study and two case series reported on VA injury treatments and treatment outcome. Primary repair and ligation were effective in treating VA injuries with patients experiencing no residual neurologic or vascular symptoms. Mixed results were seen with tamponade, with only one of three cases resulting in successful outcome. Embolization was attempted in one case and did not result in a successful outcome. No studies reported on anticoagulation therapy only or no treatment.

Abbreviations: CoE, class of evidence; KQ, key question; VA, vertebral artery.

Notes: All AHRQ “required” and “additional” domainsa are assessed. Only those that influence the baseline grade are listed in the table.

Baseline strength: Risk of bias (including control of confounding) is accounted for in the individual article evaluations. High = majority of articles level I/II; low = majority of articles level III/IV.

Downgrade: Inconsistencyb of results (1 or 2); indirectness of evidence (1 or 2); imprecision of effect estimates (1 or 2); subgroup analyses not stated a priori and no test for interaction (2).

Upgrade: Large magnitude of effect (1 or 2); dose response gradient (1).

Required domains: risk of bias, consistency, directness, precision. Plausible confounding that would decrease observed effect is accounted for in our baseline risk of bias assessment through individual article evaluation. Additional domains: dose–response, strength of association, publication bias.

Single study = “consistency unknown.”

Illustrative Case

A 69-year-old man presented with a 3-month history of progressive cervical myelopathy (Fig. 2). The patient's preoperative MRI and CT studies demonstrated an abnormal posterior course of the right VA (Figs. 3 and 4). Injury of this anomalous VA segment occurred during posterior surgical exposure. Intraoperative direct repair of the injury was performed by the on-call vascular surgeon using 6.0 Prolene sutures (Ethicon, Somerville, NJ, USA) (Fig. 5). The patient had an uneventful postoperative course without clinical evidence of VA injury. He demonstrated excellent recovery from his clinical myelopathy at 6 months postoperative.

Fig. 2

Preoperative sagittal CT image of a 69-year-old man with multilevel cervical spinal stenosis, C1-2 instability, and progressive myelopathy. CT, computed tomography.

Fig. 3

Sagittal T2 MRI image showing abnormal posterior course of the right V2 VA segment. MRI, magnetic resonance imaging; VA, vertebral artery.

Fig. 4

Sagittal CT image showing abnormal posterior course of the right V2 VA segment. CT, computed tomography; VA, vertebral artery.

Fig. 5

Immediate postoperative sagittal CT angiographic image demonstrating patency across the VA injury repair site. CT, computed tomography; VA, vertebral artery.

This systematic review is limited by the following:

○The majority of studies were CoE III or IV.

○None of the three classification systems of VA anomalies found in the literature was validated.

○Some of the studies reporting on VA injuries had a relatively small sample size (< 100 patients). Given the rarity of VA injuries, larger study populations are needed to identify this surgical complication.

This review highlights several important issues with respect to VA injury in degenerative spinal surgery. Most important is that the incidence of this potentially devastating complication may be as high as 1.96% of all cases performed, which should serve to raise the level of awareness among spinal surgeons.

It is our opinion that the VA classification system described by Eskander et al (2010) is the most comprehensive and useful to surgeons for the purpose of preoperative planning. When applied to a preoperative MRI study, this classification system enables surgeons to detect and grade VA anomalies in an organized manner. Grading VA anatomy with respect to intraforaminal midline migration, extraforaminal abnormal entrance to the cervical spine, and arterial side dominance is clearly the most important preventative technique described to date and should become an integral portion of degenerative cervical spinal surgeons' preoperative routine.

While relatively few degenerative cervical surgery VA injuries have been reported, it is likely that the true incidence might be higher that what is currently depicted in the literature.

It is clearly apparent from our review that when the intraoperative injury to the VA is encountered, direct surgical repair is the treatment option that results in the best clinical outcomes.

There remains a paucity of existing literature on this topic and much more research is needed.