Rating the incidence of iatrogenic vascular injuries in thoracic and lumbar spine surgery as regards the approach: A PRISMA-based literature review

To assess the rate, timing of diagnosis, and repairing strategies of vascular injuries in thoracic and lumbar spine surgery as their relationship to the approach. PubMed, Medline, and Embase databases were utilized for a comprehensive literature search based on keywords and mesh terms to find articles reporting iatrogenic vascular injury during thoracic and lumbar spine surgery. English articles published in the last ten years were selected. The search was refined based on best match and relevance. Fifty-six articles were eligible, for a cumulative volume of 261 lesions. Vascular injuries occurred in 82% of instrumented procedures and in 59% during anterior approaches. The common iliac vein (CIV) was the most involved vessel, injured in 49% of anterior lumbar approaches. Common iliac artery, CIV, and aorta were affected in 40%, 28%, and 28% of posterior approaches, respectively. Segmental arteries were injured in 68% of lateral approaches. Direct vessel laceration occurred in 81% of cases and recognized intraoperatively in 39% of cases. Incidence of iatrogenic vascular injuries during thoracic and lumbar spine surgery is low but associated with an overall mortality rate up to 65%, of which less than 1% for anterior approaches and more than 50% for posterior ones. Anterior approaches for instrumented procedures are at risk of direct avulsion of CIV. Posterior instrumented fusions are at risk for injuries of iliac vessels and aorta. Lateral routes are frequently associated with lesions of segmental vessels. Suture repair and endovascular techniques are useful in the management of these severe complications.


Introduction
Vascular injuries are rare but potentially life-threatening complications in thoracic and lumbar spinal surgery, having a reported incidence of 0.01% to 1%, and variable overall mortality that ranges between 15 and 65% [1][2][3].
The widening of the spectrum of the surgical corridors for the treatment of spinal diseases, frequently involving the dissection and mobilization of the major splanchnic vessels, has increased the risk of vascular injuries [4][5][6][7]. The occurrence of vascular lesions is related to the proximity of the thoracic and lumbar column to the retroperitoneal vascular structures, separated from the latter uniquely by the anterior longitudinal ligament [8,9].
Clinical presentation is not always easily recognized, especially when the injury is related to a posterior approach. Several factors as the type of approach, vessel's caliper, and mechanism of injury complicate the patient's presentation and recognition of the injury. Direct damages, as arterial or venous laceration, generally lead to an acute retroperitoneal hemorrhage with abdominal distention, hypotension, and hematocrit drop [10,11]. Conversely, delayed injuries can occur days, weeks, or even months following the procedure. They are characterized by a widening of the arterial pressure and tachycardia, which are possibly related formation of thrombi, pseudoaneurysms, or arteriovenous fistulas [12][13][14][15][16]. One difficulty to recognize the cause of delayed vascular injury relates to progressive damage to the vessel wall by prominent hardware [17][18][19][20][21].
A prompt diagnosis, adequate intraoperative management, and early repair are paramount to prevent devastating sequelae.
Nevertheless, literature about the incidence, recognition, and treatment of vascular injuries during thoracic and lumbar spine approaches is still limited, which contributes to a potential underestimation of the overall complication rate of this type of surgery.
The present study consists of a literature review on vascular complications of thoracolumbar spine surgery aimed to infer whether or not potential links do exist between the type of surgical approach, mechanism of injury, and type of injured vessel. Repair strategies and overall outcomes are also discussed.
Clinical trials, reviews, cohort studies, and case reports in English or translated to English language and published in the last 10 years were selected. The relevance of each article about the primary endpoint of the study was established by three different independent researchers based on the match with the search terms. Apart from the match, relevance, and years, no further restriction criteria were applied. Overall data about the vascular injuries were extracted, and a descriptive analysis was performed.

Literature volume
The literature search initially returned a total of 134 articles. After the removal of duplicates, 74 articles were assessed for eligibility. The application of the exclusion criteria led to the selection of 56 relevant articles for quantitative and qualitative analysis. Non-English articles excluded from the review amounted to 12. The overall level of evidence of the literature volume according to Strength of Recommendation Taxonomy (SORT) [22] has been low and consisting of 38 case reports (68%) and 18 retrospective case series (32%).
Studies concerning the anterior, posterior lateral approaches to thoracic and lumbar spinal pathologies were 38, 11, and 7, respectively. Figure 1 reports the PRISMA flowchart of the study (Fig. 1).

Demographic and clinical data
A total of 261 patients from 56 studies were included. Patients' average age in the anterior, posterior, and lateral groups was 52.6 ± 11, 48 ± 15, and 57.5 ± 12 years old, respectively. The fifth decade was the more affected.
Tables 1, 2, and 3 summarize the main demographic and clinical data of the anterior, posterior, and lateral groups, respectively (Tables 1, 2, and 3).

Type of procedures
In 208 cases (82%), an instrumented thoracic or lumbar spinal surgery was performed. In the remaining cases, a discectomy, corpectomy, or osteophytectomy were carried out.

Time to diagnosis
A direct inspective intraoperative diagnosis was achieved in 102 cases (39%). In the remaining 55 patients where the data were reported, computed tomography angiography (CTA), digital subtraction angiography (DSA), and computed tomography (CT) led to the diagnosis in 62%, 29%, and 9% of cases, respectively, for both arterial and venous injuries.

Repairing strategies
Direct repair with sutures of the artery or vein was possible in 113 patients. Endovascular techniques allowed to treat the vascular damage in 56 cases; within these, stentgraft, thrombectomy, and embolization were utilized in 21%, 4%, and 4% of cases, respectively. In surgical repairing techniques, clipping, anastomosis, and bypass were performed in 4%, 3%, and 1% of patients, respectively. The release of topical hemostatic agents was sufficient in only 4 cases (2%).

Overall outcome, morbidity, and mortality rate
Eighty-four patients (32%) were discharged without symptoms. Only for posterior approaches, one case of paraplegia, one of disseminated intravascular coagulation, one of edema of the lower extremity, and one of postphlebitic syndrome were recorded, with a morbidity rate of 1%. The overall mortality calculated on the pooled cohort was 2%, of which the 90% (7 patients) for the posterior approach.

Discussion
The present literature review is aimed at an overview of the rate of iatrogenic approach-related vascular complications in thoracic and lumbar spine surgery. Questions about a putative correlation between the type of surgical corridor, and rate and type of injury have been addressed, in the obvious assumption that prompt recognition of these harmful complications is paramount for appropriate and effective management.

Anterior approaches
Anterior approaches to the thoracic and lumbar spine provide better exposure of the ventral surface of vertebral discs, with the main advantage of avoiding manipulation of the paravertebral muscles and ligaments, and a consequent lesser postoperative pain and risk of mechanical instability, Fig. 3 Bar graph reporting the operative levels as well as direct access to the disc and release of the anterior longitudinal ligament (ALL), placement of cages with a wider footprint and possibility of achieving better restoration of lumbar lordosis [23]. During anterior approaches, the thoracoabdominal aorta, vena cava, iliac vein, and segmental vessels are mobilized and retracted to reach the anterior surface of the spine, for levels above L5-S1. Accordingly, the incidence of damages of the abdominal vessels unavoidably increased, resulting in a higher relative risk of vascular injuries related to the anterior approaches [4,24]. Concerning the transthoracic anterior approaches, Ballard et al. reviewed 617 anterior thoracic and lumbar spine exposures, reporting 13 (2%) vascular injuries [25]. Intercostal arteries and thoracic aorta were the most affected in their series.
Clip ligation of the intercostal arteries, lung deflation, mobilization of the diaphragm, and removal of the inferior ribs were the strategies suggested facilitating disc exposure and at the same time reducing the risk of thoracic vessels' lesions [25].
About the lumbar and lumbosacral spine, Garg et al. estimated the incidence of vascular injuries in 212 patients that underwent ALIF [26], of which 6.1% of cases were intraoperative vascular damages and a direct laceration of left CIV in 61.5% of cases. Notably, L5-S1 was the level affected in 70% of these procedures, where usually there is minimal mobilization of the vasculature, given that the procedure is carried out below the bifurcation and in between the iliac vessels.
In 2012, Zahradnik et al. published a retrospective review on 269 thoracic and lumbar anterior surgical procedures where they found a rate of vascular injuries of 13.8%, along with the involvement of the CIV in 52.5% of cases [6]. On a further series of 1262 left-sided anterior thoracic and lumbar surgeries by Klezl and colleagues, the overall rate of vascular injuries was 1.1%, with the CIV interested in 65% By the literature, our review highlighted a higher incidence of vascular lesions during anterior approaches (59%) in comparison with the posterior (33%) and lateral ones (8%). Not surprisingly, this rate was even higher for instrumented procedures because the positioning of pointed pedicle screws could easily injure vascular structures. At L5-S1, CIV was the most affected vessel (49%). Conversely, IVC and iliolumbar veins were mostly injured during procedures targeting the L4-L5 and L3-L4 disc space, respectively. The most frequent mechanism of damage during anterior approaches is the direct avulsion of a vein encountered along the surgical route. The rationale explanation lies in the greater fragility of the venous wall, consisting of a minor elastic component, making them more vulnerable than the arteries. On the other hand, despite an increased risk of vascular injury, anterior access allows for an easier exposure of the damaged vessel, direct control of the bleeding, and repairing of the vessel wall, which is achieved by means of a direct, often straightforward, suture in most cases. Vein repairing strategy usually entails the employment of monofilament nonabsorbable stiches.
Concerning the arterial counterpart, the CIA was found to be the most damaged during anterior procedures. CIA avulsion is a fearsome complication of the anterior lumbar retroperitoneal approach that can lead to claudication, ischemia, and lifetime paraplegia. In 2011, Hans et al. reported 4 cases of CIA injuries, consisting of thrombosis in 3 cases and laceration in one case [29,30]. Retraction and mobilization of the great arteries are thought to potentially cause a small dissection of the intimal wall, ultimately resulting in thrombosis and vessel occlusion. In anterior lumbar procedures, bilateral pulse oximetry has been suggested as a useful tool to detect intraoperatively the arterial vessel damage at an early stage, allowing its immediate treatment [31]. Arterial repairing techniques generally encompass an open arteriotomy, angioplasty, or endovascular stent placement [32][33][34][35].

Posterior approaches
Posterior approaches are related to a significantly lower risk of vascular complications, negligible in the present review (0.05%-0.08%). The rationale lies in the anatomy of the surgical route, which does not require exposure and mobilization of the retroperitoneal vessels [36,37].
Some case reports have discussed injuries to the thoracic aorta during posterior instrumented surgery, almost all sustained by impingement or avulsion of the artery leading to a delayed hemothorax [38][39][40]. Parker et al. retrospectively examined 964 patients who underwent thoracic and lumbar screw fixation procedures, finding 10 thoracic pedicle screws that encroached the thoracic aorta. Four lumbar screws damaged the CIV and 1 the IIV [20]. Riedemann-Wistuba et al. reported 4 cases of vascular injuries that occurred during PLIF, involving the CIA and CIV [41]. Kwinta et al. recently reported 2 further cases of iliac vessel injuries during PLIF. The first was a direct avulsion of the CIA, which was treated by means of stent placement, whereas the second consisted of a laceration of iliac vessels, leading to the formation of an arteriovenous fistula. The latter was repaired by a temporary femorofemoral bypass, followed by stent implantation in the right CIA [42]. Concerning the non-instrumented lumbar procedures, the risk of vascular injuries during lumbar disc surgery is low but should not be underestimated as this is a potentially catastrophic complication, to the point to be defined by some authors as the nightmare of this procedure [3,[43][44][45][46][47][48][49][50][51].
Our study demonstrated that 33% of vascular injuries occurred during posterior approaches, of which only 24% happened during instrumented surgery.
In posterior instrumented surgery, the vascular injuries are theoretically attributable to the incorrect placement of the screw tip in the proximity to the vessel wall. Conversely, a possible explanation for the paradoxically even higher rate of vascular injuries in posterior non-instrumented surgery lies in the different aim of most of the posterior non-instrumented approaches where, for example in lumbar diskectomy, it is required to manipulate the disc with a consequent higher probability that curettes or rongeurs may accidentally be deepened beyond the anterior longitudinal ligament injuring the retroperitoneal vessels. This aspect makes complex also repairs of these injuries that are burdened by significant mortality as reported, for example, in lumbar diskectomy [36,49,53].
CIA, CIV, and aorta were involved in 40%, 28%, and 28% of cases, respectively. Laceration of CIA causes a wide retroperitoneal hemorrhage, which is not always identifiable intraoperatively, and that has to be always suspected in the light of a sudden intraoperative or postoperative hypotension.
Contrary to the anterior approaches, the main difficulties in vascular repairing of a retroperitoneal major vessel during the posterior routes lie in the urgent need for turning the patient supine to perform an anterior, retro-or transperitoneal approach. Secondary injuries, caused by the impingement of hardware or cages to the vessel wall, are generally managed by a redo surgery or implant removal.
The formation of pseudoaneurysms and arteriovenous fistulas has also been described as delayed consequences of a chronic mechanical insult to the vessel wall. The presence of a pseudoaneurysm should be suspected in the presence of a pulsating abdominal mass with distension and pain, whereas arteriovenous fistulas may cause cardiopulmonary disturbances, tachycardia, intermittent claudication, and lower limbs edema [36,44,52,54].
Endovascular techniques, such as stent-graft placement, embolization, and thrombectomy, have proven to be useful in the re-establishment of blood flow [55,56].

Lateral approaches
In 2006, Ozgur et al. firstly described the lateral approach to the lumbar spine, namely the LLIF or XLIF [57]. It provides access to the lumbar vertebral discs through a retroperitoneal transpsoas route. Later, the so-called anterior to psoas approach, also known as OLIF, was introduced to avoid the mechanical consequences of the psoas muscles imbalance [58,59].
Sandhu et al., in 2013, reported 2 cases of screw penetration of the thoracic aorta during lateral transthoracic approaches, both treated by open repair [60]. In 2015, Kueper et al. estimated at 0.056% the incidence of vascular injuries in 900 patients who underwent LLIF [61]. In their series, 4 patients incurred segmental vessel damage and 1 sustained a direct laceration of the abdominal aorta. All injuries were repaired by direct ligature or suture, and no patients suffered long-term sequelae [61].
In 2016, Abe et al. conducted a retrospective review of 155 patients who underwent an OLIF procedure, where 6 segmental vessel injuries (3% of the overall series) were reported [62].
Zeng et al., in 2018, analyzed 235 OLIFs, reporting 7 cases of vascular damages of segmental arteries in 4 cases, iliac vessels in 2 cases, and ovarian vein in the remaining case [63].
The present review overall found 22 vascular injuries (8%) during lateral approaches. The most frequent mechanism of damage was the direct focal laceration of the segmental vessels (54% of cases), because of the need for mobilization and ligation during LLIF and OLIF. Segmental vessel avulsion can be promptly controlled by direct pressure and topical hemostatic agents.

Management tips
Regardless of the surgical corridor, some important aspects emerged from this study, which should be intended as tips to decrease the risk of iatrogenic vascular injuries during spine surgery, as well as to adequately manage them during or after surgery.
The vascular anatomy should be understood preoperatively to identify potential variants, especially when planning anterior or lateral thoracic and lumbar spine approaches.
Preoperative planning is crucial also for cervical spine surgery, where the detection of the carotid and vertebral artery courses and eventual anatomical variants is fundamental.
Iatrogenic carotid artery injuries during anterior cervical spine surgeries are rare and not extensively reported in the literature. There is neither evidence that the prolonged carotid artery sheath traction during anterior cervical approaches results in the reduction of carotid blood flow and cerebrovascular accidents [64][65][66]. Vertebral artery damages due to anterior cervical spine procedures are infrequent, with an incidence of 0.1-0.5% [55,67,68].
As regards posterior cervical approaches, especially for upper cervical spine surgeries, the course of both vertebral arteries must be investigated preoperatively. Its laceration is a potentially fatal complication with an incidence rate ranging between 0.07 and 8% [69,70].
Diagnostic tools to be considered are CT or MRI angiography, and in selected cases, additional vascular studies can be performed as needed [71][72][73].
In addition, the depth of the vertebral disc to the retroperitoneum should be measured on axial imaging studies for a safer placement of the surgical instruments [46,54,74].
Surgical timeout must include the presence of blood products and immediate needs tools such as large bore suckers, topical hemostats, clips, ligature, and vascular sutures [30,75].
Adequate intraoperative hemodynamic parameters monitoring must be ensured by the anesthesiology team, with immediate detection of changes in pressure, hematocrit, and arterial pulse [4,37].
After surgery, in case of suspected vascular injury, it is mandatory to perform CTA or, if the patient is hemodynamically stable, and angiography. Furthermore, dye-based methods of intraoperative visualization of the blood flow are useful to early detect the source of bleeding and also to perform a timely treatment by means of techniques coming from neurovascular surgery and described elsewhere by our group [76,77].

Limitations
This literature review has some limitations. They are the heterogeneity as well as retrospective nature of the studies included of which more than half results from case reports (68%) and involved a relatively limited number of patients. These aspects inevitably rise the likelihood of biases. A further limit to be taken into account lies in the overall experience of the surgical teams, which affects both the rate of vascular injury and the promptness and efficacy of the repairing techniques.

Conclusion
The literature reported an overall incidence of iatrogenic vascular injuries during thoracic and lumbar spine surgery which is not negligible. The mortality rate of these complications varies widely among the series included in the review, ranging between 15 and 65%, also considering the prevalence of case reports (68%) and the low level of evidence.
Anterior approaches to the thoracic and lumbar spine are burdened by a higher risk of direct venous avulsion, especially CIV and IVC. Posterior instrumented fusion has been associated with a greater risk of damage to iliac vessels and aorta. The more recent lateral corridors have to face the handling of the segmental vessels. Repairing strategies involve direct suture and endovascular techniques, both having a primary role in decreasing the overall mortality of these rare but severe complications.
Funding Open access funding provided by Università degli Studi di Pavia within the CRUI-CARE Agreement.

Conflict of interest
The authors declare that there is no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.