Spontaneous posterior spinal artery infarction

Spinal cord infarctions (SCIs) predominantly affect the anterior spinal artery (ASA) territory and manifest with acute weakness and loss of pain and temperature sensation. In comparison, posterior spinal artery (PSA) infarctions are more rare,¹ possibly due to paired PSAs with increased collateral arterial supply; the diagnosis may also be underrecognized.^1–5 Reports on clinical and radiologic features and outcomes of PSA infarcts are scant. We sought to describe all PSA infarcts seen at Mayo Clinic (1997–2017) to improve the understanding and awareness of this uncommon myelopathy. All patients consented to use of their medical records for research.

Of 133 spontaneous SCIs (i.e., excluding periprocedural infarcts) diagnosed at our facility, 15 (11%) were PSA infarctions. Clinical and diagnostic details of the 15 included patients are summarized in the table. Fourteen patients (93%) had vascular risk factors: hyperlipidemia, 11; tobacco use, 6; hypertension, 6; diabetes mellitus, 1; and atrial fibrillation, 1. Diagnostic evaluations for alternative etiologies were unrevealing, including vitamin B₁₂ (n = 14), rheumatologic/vasculitic markers (n = 13), infection (n = 12), aquaporin-4 immunoglobulin G (n = 11), hypercoagulability (n = 8), copper (n = 8), and EMG/nerve conduction studies (n = 4).

Table.

Clinical and diagnostic features of posterior spinal artery infarctions

Median time to spine MRI was 1 day (range 0–38 days). Five patients had an initial normal MRI (day 0–1); all subsequently showed posterior spinal cord T2 hyperintensity (figure) on the second MRI (≥1 day later). Other MRI findings are summarized in the table. Ten patients had vascular imaging performed (MR/CT/conventional angiography); all were normal.

Figure. MRI findings in posterior spinal artery infarctions.

(A) A patient with hyperacute onset left-sided sensory loss and gait ataxia with MRI demonstrating a dorsolateral T2-hyperintense cervical spinal cord lesion (A.a, A.d) with diffusion restriction confirmed by diffusion-weighted imaging and apparent diffusion coefficient (A.b, A.c). (B) A patient with multiple abrupt episodes of leg numbness and weakness within 24 hours, demonstrating subtle dorsal spinal cord T2 hyperintensity on sagittal view (B.a, arrowhead) better seen with axial views (B.b, arrow) and an accompanying vertebral body infarction (B.a, arrows). (C) A patient with sudden onset sensory gait ataxia with long noncontiguous posterior T2 hyperintensity on sagittal view (C.a, arrows) accompanied by cystic myelomalacia with similar signal properties to CSF on axial view (C.b, arrow). A patient with sudden arm and torso sensory loss with rapid progression to leg weakness with residual focal linear T2 hyperintensity representing myelomalacia on sagittal (D.a, arrow) and axial views (D.b, arrow).

Eleven patients (73%) received immunotherapy for suspected myelitis or Guillain-Barré syndrome (corticosteroids, 10; IV immunoglobulin [IVIg], 4); no complications were reported. One patient received IV tissue plasminogen activator 4 hours after onset with no benefit acutely but delayed recovery over 2 months allowed ambulation with an ankle-foot orthosis. Once SCI was recognized, vascular risk factor modification and antiplatelet treatment was undertaken. Thirteen patients reported improvement and 2 did not (sensory-predominant myelopathy without improvement, 1; <1 month follow-up, 1). At last follow-up (median 8 months; range 1–71 months), the ambulatory status was ambulatory without assistive device, 8; gait-aid dependent, 6; and nonambulatory, 1; none had developed new neurologic episodes.

Discussion

This is the largest series of PSA infarctions to date. All 15 patients presented with an abrupt sensory disturbance, often accompanied by pain and sensory ataxia. Many patients had impairment beyond the posterior 1/3 of the spinal cord (e.g., weakness),^1,3–5 possibly from overlapping/heterogeneous spinal arterial territories, initial coexisting ASA ischemia, or edema. An inflammatory etiology was suspected in most and notably 4 received IVIg, which is potentially harmful with prothrombotic properties.⁶

Overlapping features of PSA and ASA infarcts include deficit nadir within hours, pain at onset, motor and sensory deficits, older age, and presence of vascular risk factors.¹ In contrast to ASA ischemia, PSA infarcts have predominantly posterior column dysfunction and better outcomes (93% ambulatory).⁷ Cervical location was more frequent than reported in the literature, which may partially reflect exclusion of periprocedural SCI (e.g., aortic repair) that is thoracic predominant.⁷

Initial MRI spine may be normal, but ultimately a dorsal T2 hyperintense cord lesion is demonstrated. Diffusion sequences and vertebral body infarctions (figure) can be diagnostic; vessel imaging for dissection or occlusion should be considered, especially in the cervical spine.² Follow-up (≥1 month) MRI showing cystic myelomalacia favors infarction over inflammation. Development of severe deficits within hours, noninflammatory CSF, MRI features consistent with SCI, and exclusion of alternative etiologies supported SCI diagnosis in all our patients; these features and evaluation for specific findings (e.g., diffusion restriction) should be sought when suspecting SCI.

No clear ischemic mechanism was found in our cases despite extensive evaluations but most had traditional vascular risk factors. Treatment for spontaneous SCI is supportive; potentially harmful therapies should be avoided.⁶ Rehabilitation is important, as good outcomes can occur.⁷ Previous reports showed atherosclerosis or dissection as likely mechanisms, suggesting antiplatelet treatment and risk factor modification may be appropriate.¹

Author contributions

Nicholas Zalewski: lead author, study and concept design, data acquisition, analysis and interpretation. Alejandro Rabinstein: data acquisition, critical revision of manuscript. Eelco Wijdicks: critical revision of manuscript. George W. Petty: critical revision of manuscript. Sean J. Pittock: critical revision of manuscript. William Mantyh: data acquisition, critical revision of manuscript. Eoin Flanagan: study and concept design, analysis and interpretation, critical revision of manuscript, study supervision.

Study funding

No targeted funding reported.

Disclosure

N. Zalewski, A. Rabinstein, E. Wijdicks, and G. Petty report no disclosures relevant to the manuscript. S. Sean Pittock has received no royalties to date but may accrue revenue for patents relating to AQP4 antibodies for diagnosis of neuromyelitis optica and AQP4 autoantibody as a cancer marker. He receives research support from the Guthy-Jackson Charitable Foundation, Alexion Pharmaceuticals, Inc., and the NIH (R01 NS065829). W. Mantyh and E. Flanagan report no disclosures relevant to the manuscript. Go to Neurology.org/N for full disclosures.