Acute cerebral ischemias are most often caused by either atherosclerotic plaques, cardiac, or arterio-arterial embolism. The situation in the spinal cord is different. Here main causes are pathologies and interventions related to the aorta. Furthermore, infection leading to epidural thrombosis with secondary spinal cord infarction, hypercoagulable state, vasculitis (panarteritis nodosa, antiphospholipid antibody syndrome, systemic lupus erythematosus), epidural transforaminal injections, and cocaine consumption can induce spinal cord ischemia. Degenerative spine disease, for example, a vertebral disk, can directly compress a radicular artery [7]. Alternatively, material from the disk can enter the vascular system and cause so-called fibrocartilage embolism, which has been confirmed histologically. In children and young adults, spinal ischemia following scoliosis surgery has been reported. There it is most likely to be caused by traction movement of the spine. A confirmed cause is surgical occlusion of segmental arteries in the course of scoliosis surgery [8]. Nevertheless, the incidence seems to be low despite frequent occlusion of segmental arteries during respective surgeries. A retrospective study of 1090 patients undergoing corrective spinal deformity surgery reported only four cases with neurological deficits, which were attributed to vascular insufficiency due to vessel ligation [9]. Hyperextension of the thoracic spinal cord has been proposed as a mechanism to cause compression with consecutive occlusion and/or vasospasm and as a result spinal cord infarction in so-called surfer’s myelopathy. Quite a number of cases have been reported, typically novice surfers lying and paddling on the board in a supine position for prolonged periods of time [10].

Severe hypotension, cardiac arrest, or, rarely, spinal venous pathology preceding transient ischemic attacks affecting the spinal cord are rare.

Most of spinal cord ischemias are related to therapeutic interventions – either open repair of descending thoracic aortic pathologies or thoracic endovascular aortic repair (TEVAR) – to invasively repair aortic dissection or aneurysm. The incidence of spinal cord ischemia appears to be slightly lower with open repair strategies (between 3 and 15 %), most likely promoted by aortic cross-clamping, reperfusion injury, and acute hemodynamic changes [11]. The incidence of spinal cord following TEVAR varies widely. A recent retrospective monocentric analysis [12] from an established center reports an incidence of spinal cord ischemia following TEVAR as high as 31 % (total of 72 patients). Looking only at cases, where proper peri- and postoperative preventive measures were implemented (maintaining sufficient blood pressure, CSF drainage), the incidence decreased to almost 24 %. In contrast, in a meta-analysis of more than 4000 patients undergoing TEVAR, the rate of spinal cord ischemia was reported as low as 3.9 % [13]. A retrospective analysis of 424 patients undergoing TEVAR revealed a total of 12 patients (2.8 %) suffering from spinal cord ischemia [11]. Onset of ischemia was usually delayed by 10.6 h with a single case showing a paraparesis roughly 10 days after the intervention. The delayed onset of spinal cord ischemia has been attributed to postoperative hypotension, thrombosis, hematoma, embolization, and elevated CSF pressure. Of note, half of the spinal cord ischemia patients had previous open or endovascular aortic repair. As independent risk factors to develop ischemia, chronic renal insufficiency and extent C endovascular coverage (entire descending thoracic aorta from the left subclavian artery to the diaphragm) were identified. Means to raise spinal cord perfusion, before or after symptoms of spinal cord ischemia became apparent (raising arterial blood pressure, lumbar drainage to reduce cerebrospinal fluid pressure), were suggested to ameliorate neurological deficits and to contribute to more substantial recovery in the long term. Surprisingly, 9 out of 12 patients completely recovered from paraparesis (incomplete sensorimotor deficits), whereas 3 out of 12 patient at least incompletely recovered from paraplegia. Only one patient did not recover at all.

Based on the pattern of spinal cord infarction, defined etiologies have been suggested in a case series of 27 spontaneously occurring spinal cord ischemias [14]. Their mean age was 56 years with 11 men and 16 women. They were divided in an anterior, posterior spinal artery pattern, central or transverse manifestations. Concomitant infarction of the vertebral body was observed in one patient. The manifestation typically occurred within minutes up to several hours and was preceded only in two cases by transient ischemic attack (TIA)-like symptoms. Back or neck pain was observed in two thirds of the patients. The authors propose that arterial hypotension causes a central all the way up to a transverse spinal cord lesion pattern. On the other hand, anterior or posterior unilateral or bilateral infarctions are likely caused by mechanical affection of their corresponding vessel – the radicular arteries. This was based on the coincidence of “mechanical factors” referred to as “spine disease.” 75 % of the patients with anterior or posterior infarcts displayed disk pathology (protrusion) coinciding frequently with the infarct region. Respective mechanical factors can be aggravated by movement of the spine, preceding the occurrence of infarction in a number of patients with anterior/posterior infarction pattern. Of course, fibrocartilage embolism, as described by a number of case reports, and confirmed postmortem histologically cannot be excluded as an underlying mechanism. Overall, recovery in this series of spontaneous spinal cord ischemia cases was remarkably positive with complete or incomplete recovery in 70 % of the patients. In comparison, retrospective studies with spinal cord ischemia patients induced by aortic pathology (with or without surgery) revealed that at least 50 % of the patients showed an unfavorable outcome with permanent wheelchair dependency [5].

MRI represents the gold standard to visualize ischemic changes. Nevertheless, MRI of the spinal cord can be rather challenging, which is attributed to the need for strong gradients, the small size of the spinal cord, and flow artifacts among others. Axial and sagittal T1-, T2-, STIR-, and diffusion-weighted images are recommended. Contrast enhancement is absent in the initial stage and can therefore help to delineate the pathology from inflammatory or neoplastic causes. Subsequently, contrast enhancement can be observed in the majority of spinal cord ischemia cases – typically more than 2 days after disease onset [15]. In the acute stage, restriction of diffusion and hyperintense signal changes in T2 and STIR sequences can be observed. However, depending also on the quality of the scan, MRI can be without any relevant changes. According to a longitudinal analysis with serial MR scans over time, signal changes in T2-weighted images are typically observed within 2 days from disease onset; the earliest respective change was seen already 14 h after disease onset [15]. Occlusions of the anterior spinal artery can cause predominant infarction of the anterior horn and surrounding white matter uni- or bilaterally, whereas occlusion of the posterolateral artery affects the dorsal horn and the dorsal columns. Concomitant infarction of the vertebral body – associated with hyperintense signal changes in the vertebra and the adjacent disk – is due to the shared vascularization of the spinal cord and the vertebrae [14].

As pointed out above, MRI can be unremarkable in the early phase after spinal cord ischemia. In this case, an infectious/inflammatory cause of spinal cord disease has to be addressed requiring a CSF workup. The general conception is that the CSF in spinal cord ischemia is pretty much normal except for a moderate protein increase. Systematic findings about CSF results in spinal cord ischemia are sparse. In a case series of 13 patients with spontaneous spinal cord ischemia, a mean cell count of 35.4 cells/mm³ (range 1–160) and a mean protein level of 0.72 g/l (range 0.4–1.39) have been reported [16]. Accordingly, moderate CSF pleocytosis does not rule out spinal cord ischemia entirely.

In respect to evidence-based treatment for spontaneously occurring spinal cord ischemia, no randomized controlled clinical studies exist to date. Rule number one is to identify and treat the underlying cause if possible. Prophylactic treatment with a platelet inhibitory drug such as acetylsalicylic acid is commonly recommended. Thrombolytic treatment with recombinant tissue plasminogen activator (rt-PA) is not established in spinal cord ischemia. Few case reports describe a favorable outcome, which is not sufficient to recommend this therapeutic approach [17, 18]. It is challenging to confirm the diagnosis spinal cord ischemia before the treatment window for rt-PA closes. Due to a missing rationale and evidence, steroids should not be administered.

In respect to treatment and prophylaxis following surgical aortic repair, numerous studies have been performed. It is widely accepted that blood pressure needs to be stabilized immediately. At the same time, spinal cord pressure should be lowered by means of CSF drainage. A meta-analysis investigating 46 studies with a total of 4936 patients, who underwent TEVAR, did not reveal a significant benefit for routine or selective prophylactic lumbar CSF drainage [13]. In contrast, for open aortic repair strategies, the benefit of routine CSF drainage has been confirmed in meta-analyses and a randomized controlled trial [19].

Segmental artery occlusion in the course of anterior spine surgery – mostly related to scoliosis surgery – can induce spinal cord ischemia in rare instances [9]. Preoperative spinal angiography might be useful in determining the exact location of the Adamkiewicz artery and thus allows exact planning of the surgical approach to prevent a lesion of this very important artery. In 100 preoperative angiographies, the Adamkiewicz artery was located between the segments T8 and L3 (in 50 % at the T8/T9 level) and in 75 % on the left side. This information led to ten side changings and three modifications of surgical technique with segmental vessel preservation. In none of the 100 patients, neurological deficits suggesting spinal cord ischemia were observed postoperatively [20].