The natural history of untreated spinal dAVFs is generally thought to be poor. An early series found that 50% of untreated patients became severely disabled (wheel-chair-bound) within 3 years of the onset of lower extremity weakness.37

Both surgery and endovascular treatment can be effective for treatment of type I lesions. Although surgery appears to be more curative, embolization is less invasive and some authors recommend an attempt at embolization prior to surgery. In a systematic review of 20 published clinical series, 98% of patients treated with surgery were reported to have successful obliteration of their fistulae, compared to only 46% with embolization.38 Complications were reported in 1.9% of surgical patients and 3.7% of embolization patients.

Extradural dAVFs are treated almost exclusively with embolization of the arterial feeder and rarely require surgery.2,39

In 1926, Foix and Alajouanine published a 42-page report of two cases of progressive myelopathy.32 An extensive pathological analysis was carried out, and the authors implicated vascular congestion, as reflected by spinal cord vessel thickening, in the pathological process afflicting both patients. In the decades since this report, numerous authors have included spinal cord venous thrombosis as a central feature of the Foix–Alajouanine Syndrome.6,44,50–53 Indeed, both authors of this handbook were taught, during their training, that Foix–Alajouanine Syndrome is equivalent to progressive, malignant spinal cord venous thrombosis. In the actual report, however, Foix and Alajouanine emphasized that in their two cases no thrombosis was present.54 They described vessel wall thickening, without luminal narrowing or obliteration of cord vessels, and excluded the presence of vascular malformations within the cord. The inclusion of thrombosis as a feature of Foix–Alajouanine Syndrome is a myth that has been perpetuated most likely because the original report was written in French. In retrospect, it seems likely that both patients in the original report had progressive myelopathy due to type I dural AVFs,44 an entity that had not yet been recognized at the time of the publication.54

The natural history of untreated intramedullary AVMs is not clear. Progressive evolution of symptoms, by either worsening myelopathy or subsequent haemorrhages, is reported in 31–71% of patients observed over several years.12,65,70,71 Because spinal cord AVM anatomy is variable and the risk of potential complications on any procedure involving the cord is relatively high, decision making about the management of these patients is highly individualized. Patients with a cord AVM consisting of a compact, surgically accessible nidus may be good candidates for surgery. Embolization may be a useful adjunct to surgery, or, in some cases, may provide symptomatic relief without necessarily obliterating the lesion. There is a school of thought that holds that partial embolization of spinal cord AVMs, even with impermanent materials (such as PVA) may provide an (impermanent) improvement in symptoms such as pain and myelopathy.70 The notion that partial treatment of cord AVMs lowers the risk of haemorrhage, which is generally believed not to be the case with intracranial AVMs, is more controversial.

In a preliminary report of stereotactic radiosurgery for intramedullary AVMs, six of seven patients at least 3 years from treatment had a significant reduction in AVM volume, and one patient with a conus lesion was found to have complete angiographic obliteration.79

Cobb syndrome is a rare congenital disorder with a slight male predominance characterized by a combination of vascular skin naevi and a spinal vascular lesion occurring within the same metameres (i.e., the skin lesions are found in the dermatomes corresponding to the spinal levels where the spine lesion is). It was first described by Stanley Cobb, a resident of Harvey Cushing, in a report of a 8-year-old boy who presented with paraplegia.82 The child had naevi over the 9th to 12th ribs, and at surgery he was found to have an angioma of the thoracic spinal cord. The spinal vascular lesion that occurs as part of Cobb syndrome may be a type III AVM, or a less complex lesion such as a perimedullary AV fistula.83 Therefore, type III spinal vascular lesions and Cobb syndrome are not exactly synonymous.

Because type IV lesions are rare, published series are limited, and firm conclusions about treatment cannot be made. Estimations of the natural history of untreated lesions suggest progression from myelopathy to paraplegia within 5–7 years, and a high incidence of repeated haemorrhage in patients presenting with haemorrhage.12,55,84,87,88 Most authors recommend prompt diagnosis and treatment of perimedullary fistulae to minimize neurological injury.12,55,84,87 The objective of treatment should be occlusion of the fistula. The absence of a nidus facilitates surgery;55 most lesions can be effectively obliterated with either surgery or embolization, or a combination of both.3,55,86,93,94

Spinal extradural AV fistulas (aka epidural AV fistulas, extradural AVMs, or perivertebral AVMs) are rare.98–103

Spinal cord aneurysms are rare. Most are discovered as flow-related lesions associated with an intramedullary AVM, although isolated spinal cord aneurysms do occur.106–108 and can cause subarachnoid haemorrhage.108 Spinal aneurysms may explain SAH when no intracranial aneurysm is found; often the patient complains of back or neck pain in addition to and often beginning before the classic headache of SAH. Particularly large aneurysms may present with symptoms of myelopathy or radiculopathy.106 There appears to be a strong association with developmental vascular anomalies, as metameric angiomatosis was found in 43% of patients in a series of cord aneurysms.59 Spinal aneurysms are typically fusiform in shape and thus very difficult to treat directly with endovascular techniques. Surgery of spinal cord aneurysms with trapping106 and wrapping109 has been reported. Disappearance or a decrease in size of the aneurysm was observed in several patients who underwent treatment of an associated intramedullary AVM.58 Importantly, even solitary spinal cord aneurysms may spontaneously regress without treatment.108

Cavernous malformations (aka cavernomas) are distributed along the entire neuroaxis and represent 5–12% of all spinal vascular lesions.110,111 See Chap. 16 for a discussion of intracranial cavernous malformations. Rarely, spinal cavernous malformations may be found in an epidural location.112 Spinal cavernomas are histologically identical to those in the brain. They are angiographically occult and have a characteristic appearance on MRI.111 As many as 40% of patients with a spinal cavernoma have similar intracranial lesions.113 As in the brain, spinal cord cavernomas are frequently associated with a developmental venous anomaly.109 A systematic review found a female predominance with a peak age at presentation in the forth decade.114 Symptoms is highly variable and the natural history of symptomatic spinal cavernomas is far from clear. Annual rates of symptomatic rehaemorrhage range from 0%115 to 66%.116 Although surgical series report relatively good results,111,114,117 expectant management of selected patients may also be reasonable.115 A recent report of long-term results in 80 patients after resection found:118

Any part of the central nervous system may suffer ischaemic injury, and this may occur in the spinal cord. Spinal cord ischaemia in the cord is uncommon and may be misdiagnosed as other pathological processes such as transverse myelitis. Cord ischaemia is becoming increasingly recognized due increasing clinical awareness of this condition, especially in association with stent-grafting for aortic aneurysms. Brain ischaemia is discussed in Chap. 17.