We conducted a search on Pubmed and Google Scholar for relevant articles published over the past 30 years (1993–2023), using the keywords ‘posterior spinal artery’, ‘posterior spinal artery infarct’, ‘posterior spinal artery infarction’, ‘posterior spinal artery syndrome’, and ‘spinal cord infarction’. The duration of 30 years was arbitrarily chosen to ensure the availability of sufficient data for analysis, whilst making sure that said data is current enough to impact present-day practice. The eligibility of suitable articles was reviewed by two neurologists prior to their inclusion. Non-human research, anatomical studies, clinical guidelines, review articles, editorials, commentaries, book chapters, and articles written in languages other than English were omitted. Records with incomplete clinico-radiological information, as determined by the two aforementioned neurologists, were duly excluded from analysis. Additionally, cases with extensive infarctions involving the anterior spinal arterial territory within the same region were also omitted. The clinical features of identified cases and their ancillary test results were then extracted and retrospectively analysed, after which statistical analysis was performed using Mann-Whitney U test, unpaired t-test, Pearson’s x2-test (Yate’s correction when relevant), and Fisher’s exact test as appropriate with a two-tailed alpha of 0.05.
Results (Table 1; Appendices A & B)
We identified 67 relevant studies published between 1993 and 2003 (Appendix B), of which 34 were excluded due to the following reasons: 18 were written in languages other than English, 9 involved patients with extensive the anterior spinal artery infarctions at the same cord level, and 7 had incomplete clinico-radiological data which precluded meaningful analysis. A total of 40 patients from the remaining 33 studies were available for analysis [4–34]. Their demographic data and clinical features were summarised in Table 1, and described in greater detail in Appendix A. There were 23 males and 17 females (male to female ratio of 1.35), and their median age was 55 years (range 19–84; the exact ages of Patients 2, 5, and 6 were unknown). At least half (21/40, 53%) had pre-existing cardiovascular risk factors, and about a third (13/40, 33%) had multiple. Hypertension predominated (16/21, 76%) amongst these risk factors, followed by smoking of tobacco (8/21, 38%), hyperlipidaemia or dyslipidaemia (6/21, 29%), diabetes mellitus (5/21, 24%), and strokes (1/21, 5%). In those without cardiovascular risk factors, three had known vertebral artery dissections (VADs; 3/21, 14%) prior to their PSAIs.
Table 1
Summary of the clinical features and outcomes of patients with posterior spinal artery infarctions.
Ref | Pt | Age (yr) | M/F | PMHx | Preceding event | Pain | Neurological symptoms | Cord infarcta (& other sites) | Cause | Treatment | Ambulant (time-pt) |
[4] | 1 | 55 | M | Nil | Neck trauma | Y | R-sided weakness & numbness | R, C1-3 | Trauma | APT, statin | N (3m) |
[5] | 2 | 60sb | M | Htn | Nil | N | L-sided numbness | L, C1 | ATH | APT, statin | Y (3m) |
[6] | 3 | 41 | F | NA | Nil | N | L-sided numbness, incl. face | L, C1 | VAD | NA | NA |
[7] | 4 | 63 | M | Htn, DM | Lifting boxes | N | Bil numbness of LLs | Bil, C4-5 (& vert) | FCE | APT | Y (pre-dc) |
[8] | 5 | NA | F | Nil | Nil | Y | Dizziness, diplopia; L-sided numbness | L, C1 (& MO, CB) | VAD | NA | Y (2 year) |
[8] | 6 | NA | M | Htn | Nil | Y | Unsteadiness, facial numbness | L, C1 (& MO) | Nil | NA | Y (6m) |
[9] | 7 | 66 | F | NA | EE of VAA | Y | Tetraplegia, L-sided numbness | R, C3-4 | ET Cx | High-dose stds | Y (3w) |
[10] | 8 | 44 | F | Nil | Nil | N | Dizziness, nausea, bil numbness | Bil, C1 (& MO) | VAD | APT | Y (1m) |
[10] | 9 | 34 | M | Hld | Nil | N | L-sided weakness & numbness | L, C1-3 | VAD | APT, statin | Y (2m) |
[11] | 10 | 56 | M | Htn, smoker | Nil | N | Tetraparesis, numbness, dystonia | Bil, C2-7 | Nil | IVMP | Y (1m) |
[12] | 11 | 22 | M | Nil | Trauma | N | L UL weakness & sens. Disturbancesc | L, C2-7 | Trauma | NA | NA |
[13] | 12 | 70 | M | Htn, Hld, CVA | AG of VAS | N | Unsteadiness, L-sided numbness | L, C1 | ET Cx | APT | Y (3m) |
[14] | 13 | 56 | F | Htn, Hld, C spond., VAD | ET (SACE) of VAA | N | R UL & chest sens. disturbances | L, C2 | ET Cx | NA | Y (at dc) |
[14] | 14 | 50 | F | VAD | ET(PO) of VAD | N | L-sided hemiparesis, sens. disturbances | L, C2 | ET Cx | NA | Y (at dc) |
[14] | 15 | 51 | F | VAD | ET(PO) of VAD | N | R-sided hemiparesis, sens. disturbances | R, C1 | ET Cx | NA | Y (NA) |
[15] | 16 | 70 | F | RA | Face-washing | N | L-sided hemiparesis & s & sens. disturbances | L, C1-2, (& MO, CB) | ATH | APT | Y (NA) |
[16] | 17 | 45 | M | Nil | Nil | N | Bil LL paraesthesia, voiding & gait difficulties; UI | Bil, T7-8 | MV strands | DOAC | Y (2w) |
[17] | 18 | 79 | F | Htn, smoker | Startled | N | Bil LL numbness, paraparesis; UI | Bil, T10-12 | Nil | AC, APT | Y (5m) |
[18] | 19 | 30 | M | Migraine | Nil | N | R-sided numbness & UL weakness | R, C1 (& CB) | VAD | AC, APT | Y (9m) |
[19] | 20 | 58 | M | NA | Lifting shutters | Y | Bil LL numbness, paraparesis; UI | Bil, T7-11 (& vert) | FCE | NA | Y (8w) |
[20] | 21 | 84 | F | NA | Fall | Y | Bil numbness, tetraparesis; UI | Bil, C & T | Trauma | IVMP | Deathd (6w) |
[21] | 22 | 41 | M | Htn, Hld | RTA | N | L-sided numbness (incl. face), | L, C1 | Trauma | APT | NA |
[22] | 23 | 37 | M | Hld | Nil | Y | R-sided weakness & numbness; UI | R, C1-2 | VAS | APT | Y (5w) |
[23] | 24 | 77 | M | IHD, DM | Nil | Y | L hip weakness, bil LL numbness; UI | Bil, T6/7 (& vert.) | Nil | APT | NA |
[23] | 25 | 83 | F | Htn, DM | Nil | N | Tetraparesis, numbness of all limbs; UI | Bil, C4-6 | Nil | APT | Deathd (3w) |
[23] | 26 | 70 | F | Htn, smoker, COPD | Lifting load | Y | L LL weakness & numbness | L, T8-10 | Nil | NA | N (2 year) |
[23] | 27 | 43 | M | Smoker | Nil | Y | Paraparesis, bil LL numbness | Bil, T8-12 | Nil | APT | Y (6w) |
[24] | 28 | 63 | F | Htn, smoker | Neck extension | Y | Bil shoulder & arm numbness, paraparesis | Bil, C4-C7 | VAD | APT | Y (4m) |
[25] | 29 | 60 | F | Htn | Moderate activity | Y | R LL weakness & numbness; UI | R, T5-7 | PFO | APT | Y (1 year) |
[26] | 30 | 58 | M | NA | Nil | Y | LL weakness & numbness | NA, T9-10 | Nil | NA | Y (NA) |
[27] | 31 | 70 | F | Nil | Head trauma | Y | L-sided numbness (incl. face), ataxia; UI | L, C1-4 (& MO) | VAT | AC | Y (40d) |
[28] | 32 | 28 | M | Nil | EE, ICdAVF | Y | L-sided weakness & numbness | L, high C cord (& MO) | ET Cx | High-dose stds | Y (1 year) |
[29] | 33 | 52 | M | Nil | Clearing snow | N | Paraparesis, bil LL numbness; UI | Bil, T9-12 (& vert) | Nil | AC, IVMP | Y (3m) |
[30] | 34 | 64 | F | Htn, Hld, DM | Nil | Y | Vertigo, nystagmus; R ataxia, weakness & numbness; perioral paraesthesia | R, C1-2 (& MO, CB) | VAS | AC | Deathd (NA) |
[31] | 35 | 48 | M | IHD, smoker | Nil | N | L hand numbness & clumsiness | L, C2-3 | ATH | APT | Y (NA) |
[32] | 36 | 52 | M | Htn, DM, smoker | Nil | N | Vertigo, R-sided weakness; R hand & face, and L-sided numbness; UI | R, C1-3 (& MO) | VAS | APT | N (NA) |
[33] | 37 | 47 | M | Nil | Decorating home | Y | L-sided numbness | L, C1 | VAD | AC | Y (3m) |
[34] | 38 | 73 | M | Nil | ET of SpdAVF | N | Paraparesis; bil LL ataxia & numbness | Bil, T10-11 | ET Cx | NA | Y (4m) |
[35] | 39 | 19 | M | Smoker | Neck extension | Y | Numbness of neck, trunk & bil limbs | Bil, C2-4 | VAD | AC | NA |
[36] | 40 | 49 | F | Htn | Climbing stairs | Y | L LL weakness & numbness; UI | L, low T cord | Nil | NA | NA |
a: Cord infarcts were demonstrated either by spinal magnetic resonance imaging or computed tomography.
b: Patient 2’s exact age is unknown
c: Patient 11’s neurological deficits were due to brachial plexus injury
d: The deaths of Patients 21, 25, and 34 were from respiratory failure, pulmonary embolism, and cardio-respiratory arrest respectively.
Abbreviations: AC, anticoagulant; AG, angioplasty; APT, antiplatelet; ATH, atherosclerosis; Bil, bilateral; C, cervical; CB, cerebellum; COPD, chronic obstructive pulmonary disease; Cx, complications; d, day; dc, discharge; DM, diabetes mellitus; DOAC, direct oral anticoagulant; EE, endovascular embolizationl ET, endovascular treatment; F, female; FCE, fibrocartilaginous embolism; HLD, hyperlipidaemia; HTN, hypertension; ICdAVF, intracranial dural arteriovenous fistula; incl., including; incont., incontinence; IVMP, intravenous methylprednisolone; L, left; LL, lower limb; M, male; mn, month; MO, medulla oblongata; MV, mitral valve; NA, not available; PO, proximal occlusion; R, right; RA, rheumatoid arthritis; RTA, road traffic accident; sens., sensory; SACE, stent-assisted coil embolization; SpdAVF: spinal dural arteriovenous fistula; spond., spondylosis; stds, steroids; T, thoracic; thromb., thrombosis; time-pt, timepoint; UI, urinary incontinence; UL, upper limb; VAA, vertebral artery aneurysm; VAD, vertebral artery dissection; VAS, vertebral artery stenosis; VAT, vertebral artery thrombosis; w, week; Y, yes; yr, year.
Clinical symptoms and signs
Pain was a common (19/40, 48%) symptom and tended to occur early (15/19, 79%). Sensory deficits of the limbs were present in all but one (38/39, 97%; Patient 11 was omitted as his neurological deficits were likely due to his brachial plexopathy), and deficits of dorsal column and spinothalamic sensory modalities were respectively present in at least 28 (74%) and 13 patients (34%; ipsilateral in 1, contralateral in 5, bilateral in 7). Facial sensory symptoms were reported in only a few (5/39, 13%), amongst whom all had infarctions of the C1 cord and three had additional involvement of the medulla oblongata. Motor deficits of the limbs were frequent (27/39, 69%; Patient 11 was again omitted for the same aforementioned reasons), and mobility was impaired in most (27/39, 68%; Patients 11 and 19 were omitted due to incomplete data on their mobility). Symptoms of bladder dysfunction were present in at least a third (13/40, 33%). Neurological symptoms consistent Brown-Sequard syndrome were present in only fourteen (14/39, 36%; Patient 30 was omitted due to insufficient data), of which nearly all were partial variants (12/14, 86%).
Level of infarction of the cord and the surrounding areas
A clear majority of PSAIs affected the cervical cord (29/40, 73%), of which all invariably involved the high cervical segments of C1-4 (28/28, 100%; excluded Patient 21 due to insufficient data). Infarctions of the C1 segment predominated (17/27, 61%; excluded Patients 21 and 32 due to insufficient data), followed in descending order of frequency by C2 (13/27, 48%), C3 (9/27, 33%), C4 (8/27, 30%), C5 (5/27, 19%), C6 (4/27, 15%) and C7 (3/27, 11%) cord segments. Comparatively, thoracic infarctions occurred in only a third (12/37, 32%), amongst whom only Patient 21 had simultaneous infarctions of the cervical and upper thoracic cord. Most infarctions were unilateral (25/39, 64%), occurring twice as frequently in the left (17/25, 68%) as the right half of the cord (8/25, 32%). While most infarcts were short in length, PSAIs in at least sixteen (16/38, 40%; Patients 32 and 40 were omitted due to insufficient data) involved three or more segments. Concomitant infarctions of the surrounding structure were infrequently observed – cerebellum in four, medulla oblongata in nine, and the vertebral body in three.
Precipitating events, and aetiologies
Precipitating events, defined as preceding incidents which respective authors deemed were temporally relevant to the patient’s PSAI, were unknown in nearly half (17/40; 43%). Of the remaining twenty-three, preceding trauma was documented in more than a third (8/23, 35%), ranging from carrying heavy boxes and lifting shutters, to road-traffic accidents and falls with head injuries. Amongst the rest, non-traumatic precipitants included endovascular treatment (7/23, 30%) of aneurysms, dural arteriovenous fistulae (dAVF), or stenoses involving the vertebral, posterior inferior cerebellar, or spinal arteries. Other precipitating events, however, varied significantly from face-washing and climbing stairs, to clearing snow and prolonged neck extension.
Despite the many investigations each patient was subjected to, the aetiologies of PSAIs were unknown or left undescribed in ten (10/40, 25%). PSAIs were attributed primarily to VADs in eight patients (8/40, 20%), and to endovascular treatment complications in seven (7/40, 18%; endovascular treatment of VADs in three, dural arteriovenous fistulae in two, vertebral artery [VA] stenosis in one, and VA aneurysm in another). Interestingly, embolic complications of fibrocartilage, mitral valve strands, and patent foramen ovale were separately described in Patients 4, 17 and 29 respectively. VA-related events (including dissections, stenoses, thrombosis, atherosclerosis, and endovascular treatment-related complications) invariably resulted in infarctions of the high cervical cord, reflective of the underlying vascular anatomy.
Treatment and outcomes
Treatment measures were undescribed in 12 patients. Of the remaining 28, the use of antiplatelet agents, anticoagulants, and high-dose steroids were respectively reported in eighteen (18/28, 64%), eight (29%), and five patients (18%). Anticoagulants used included intravenous (IV) heparin, IV urokinase, subcutaneous low-molecular weight heparin, and oral warfarin or rivaroxaban. Five patients (5/28, 18%) were treated with high-dose steroids, of which at least three received methylprednisolone intravenously. Unfortunately, clinical outcomes were incompletely and inconsistently reported across the articles, and validated measures of function and disability were used only sporadically. Furthermore, the patients were assessed across inconsistent time-points, ranging from the day of discharge to as long as two years. Despite these important deficiencies, the overall clinical outcome of PSAIs appeared fair. Deaths were uncommon (3/34, 9%), and most were ambulant (28/34, 82%; unreported in Patients 3, 11, 22, 24, 39, and 40) at their respective points of assessment. However, over a quarter amongst them required assistance when walking (8/28, 29%).
Factors influencing ambulatory outcomes (Table 2)
Table 2
Factors influencing ambulatory outcomes in patients with posterior spinal artery infarctions.
| Non-ambulant (n = 6) | Ambulant (n = 28) | p-value |
Age, y, median (range)* | 67 (52–84) | 56 (28–79) | 0.055 |
Male, n (%) | 2 (33) | 16 (57) | NS |
With CVSRF, n (%) | 4 (67) | 13 (46) | NS |
Clinical features: |
Preceding trauma, n (%) | 3 (50) | 3 (11) | 0.053 |
Partial or complete BSS, n (%)* | 3 (50) | 10 (37) | NS |
Presence of pain, n (%) | 4 (67) | 12 (43) | NS |
With weakness, n (%) | 6 (100) | 18 (64) | NS |
Level and extent of cord infarction: |
Cervical cord infarction, n (%) | 4 (67) | 20 (71) | NS |
Thoracic cord infarction, n (%) | 2 (33) | 8 (29) | NS |
Bilateral cord infarction, n (%)* | 2 (33) | 10 (37) | NS |
Infarctions of 3 or more cord segments, n (%)* | 5 (83) | 9 (33) | 0.062 |
With medulla oblongata or cerebellar infarctions, n (%) | 2 (33) | 7 (25) | NS |
With cerebellar infarction, n (%) | 1 (17) | 3 (11) | NS |
With medulla oblongata infarction, n (%) | 2 (33) | 6 (21) | NS |
Aetiology: |
Unknown, n (%) | 2 (33) | 6 (21) | NS |
Vertebral artery dissection, n (%) | 0 (0) | 6 (21) | NS |
Endovascular treatment complications, n (%) | 0 (0) | 7 (25) | NS |
*: missing data |
Abbreviations: BSS, Brown-Sequard Syndrome; CVSRF, cardiovascular risk factors; NS, not significant. |
Incomplete information precludes meaningful analyses of the factors affecting ambulatory independence. Attempts at studying how different treatments affected ambulatory outcomes were similarly hindered by the lack of data. Despite these, statistical analyses appropriate for our small study size preliminarily suggested that patients who were older (median age 67 years [range 52–84] vs. 56 [28–79], p = 0.055), who experienced preceding trauma (50% vs. 11%, p = 0.053), or have PSAIs longer than three or more cord segments (83% vs. 33%, p = 0.062) may be likelier to experience non-ambulatory outcomes, that is, being unable to walk at all. Conversely, their genders, cardiovascular risk factors, level of infarction, and the presence of pain, weakness, or concomitant brainstem or cerebellar infarcts did not appear to influence their ability to walk.