Varicella zoster virus (VZV) vasculopathy

History
Evidence that varicella zoster virus (VZV) could invade the central nervous system began to accumulate in the late 1800s. In 1876, Hardy reported a patient who had shingles of the trunk that later developed bilateral belt-like pain, followed by weakness in the legs that ascended, eventually leading to respiratory paralysis, attributed to ascending myelitis. Hardy and Muselier subsequently reported cases in which patients developed weakness remote from the site of shingles that progressed over time, and reported that on autopsy there was evidence of inflammation entering into the spinal cord.

Involvement of the brain itself appears to first have been described in 1896, when French physician and pathologist Édouard Brissaud described three patients with cutaneous shingles who subsequently developed neurological symptoms. In one, a 54 year old man developed L facial shingles, and 3 months later suddenly fell and developed a complete left 3rd nerve palsy as well as dysarthria, which gradually improved in the subsequent months, and was hypothesized to be a left superomedial midbrain syndrome (Weber's syndrome). The second case was a 44 year old man who developed a severe left-sided headache followed by facial rash which lasted for two weeks who one week later suddenly developed right hemiplegia, right hemisensory loss, and aphasia. In the last case, a 56 year man developed severe facial shingles which caused left eye ptosis and several months later developed weakness on the right side of his body. Brissaud attributed these to central lesions of possible vascular origin, with the first and last case both suggestive of Weber's syndrome. Many years later Cope et al. doubted that these cases represented true hemiplegia from ophthalmic zoster, and instead attributed the first case to Dumary in 1896, which was reported in 1919 by Baudouin and Lantuejel.

The first evidence that VZV could cause vascular inflammation in the nervous system came in 1903, when Ernest Hedinger noted perivascular inflammatory changes in the spinal cord and nerves of a patient who died after having a zoster infection. In 1924, Thalhimer as well as Lhermitte and Nicolas separately described patients with cutaneous zoster and neurological symptoms who were noted to have perivascular inflammatory changes in the medulla and spinal cord on autopsy,  Several years later in 1929, Faure Beaulieu and Lhermitte described a 74 year old man who developed zoster in T11-L1 distribution on the right, who subsequently developed right face/arm/leg weakness and confusion. Spinal inflammation was seen on autopsy, but the brain was not autopsied.

Debates throughout the 1940s-1960s continued regarding the etiology of CNS involvement of VZV. Most suggested some diffuse tissue invasion, including through the vessels, with the virus possibly traveling via the vessels as they were preformed pathways. In 1945 Gordon and Tucker suggested a "vascular" etiology but did not describe further details.

In 1958, Anastasopoulos et al. hypothesized that VZV may start as a ganglionitis but then locally invade nearby vessels. Then, in 1968, Kolodny et al. described a 21 year old with cutaneous VZV in the left T3-T4 dermatome who subsequently had left-sided weakness and was found on angiography to have decreased small vessel filling in the right MCA territory. Pathology showed granulomatous inflammation. The patient seemed to stabilize transiently on dexamethasone but deteriorated after it was discontinued due to severe side effects. In 1971 in Japan, Sato et al. described the first angiogram showing large vessel changes. In 1972, Rosenblum and Hadfield described the association of VZV and vasculopathy in two patients immunosuppressed due to lymphoma and chemotherapy, who developed mental status changes after cutaneous VZV infection, eventually died, and were found to have cerebral granulomatous vasculitis and infarcts with no evidence of meningitis. One year later in 1973, Walker et al. reported a 7-year old boy with right hemiplegia six months after left trigeminal VZV, with angiography showing clear narrowing of the left proximal MCA and ACA vessels and normal vessels on the right side. This was the first evidence of a large vessel vasculopathy associated with VZV, and was quickly followed one year later by a similar report but with stenosis in the carotid siphon, who had improvement with steroid treatment. In 1977, Pratesi et al. published a case with right V1 VZV who subsequently developed mental status changes and had angiography showing bilateral MCA narrowing, showing that this could be a more diffuse process and not necessarily limited to the side of initial VZV involvement.

Early theories suggested a possible link between VZV and "granulomatous arteritis," commonly referred to as "temporal arteritis" or "giant cell arteritis." In one report of a temporal artery biopsy in a patient with evidence of focal CNS disease consistent with prior reports, a temporal artery biopsy was done and showed evidence of vasculitis that was not definitively granulomatous, but importantly showed no evidence of viral particles. However, in 1980 Linnemann and Alvira published a case of a 20-year-old with Hodgkin's lymphoma who was admitted with right trigeminal VZV with subsequent dissemination to the entire body. As the skin lesions were resolving he developed altered mental status, confusion, tremor, and jerking movements of his limbs, after which he suddenly died on the sixth hospital day. The brain had significant inflammation and leptomeningeal opacification, with infarcts in sections of the pons, medulla, and upper cervical cord, shown on pathology to be due to granulomatous vasculitis. Electron microscopy showed evidence of herpes-like viral intranuclear particles inside smooth muscle cells in the arteries. However, subsequent evaluations of patients with giant cell arteritis showed no evidence of VZV in temporal artery biopsy specimens of those with giant cell arteritis.

In 1981, Mackenzie et al. published a series of four patients who developed focal neurological deficits 6-8 weeks after ophthalmic VZV and showed changes on angiography intracranially only. They suggested that VZV vasculopathy is characterized by single or multiple smooth-tapered segmental narrowings in either the A2 segment of the proximal pericallosal artery, the M1 segment of the MCA, or the parasellar carotid siphon, without involvement of the external carotids. However in 1983 Bourdette et al. published a case series which included a patient suspected of having cerebellar involvement, expanding the list of possible arteries that could be involved. Later in 1983, Eda et al. reported that primary VZV infection (chickenpox) was also associated with a delayed hemiplegia. Then, in 1985, Tabbaa and Selhorst expanded the disease spectrum further by publishing a case of a patient with infarction limited to small vessel disease of the lenticulostriate arteries after ophthalmic VZV. Filloux and Townsend then reported the first case of ipsilateral cerebellar (SCA) infarction after ophthalmic VZV. In 1986, Powers reported the first case of V2 involvement of VZV with subsequent PCA vasculitis and infarction. In 1988, Snow and Simcock described patient who developed R paramedian pontine infarction 4 weeks after right cervical (C2) VZV zoster, and suggested that this was due to innervation of the vertebral artery from the C2 supply, but no vessel imaging was done. That same year, Joy et al. described a patient with left otic VZV zoster (Ramsay-Hunt syndrome) who developed a right lenticulostriate infarct, demonstrating that focal VZV could affect more remote vessels, even in immunocompetent patients.

These initial studies were later followed by numerous studies showing an association between VZV infection and stroke. The first case-control study was done in 1999 and showed a significant association between primary VZV infection (chickenpox) and cerebral infarction within 9 months, a result which was confirmed two years later.

Pathophysiology
VZV appears to travel along sensory neurons innervation arteries (vasa nervorum), leading to vasculopathy in specific vessels. The anatomical innervation very frequently correlates with affected vessels.

Relevant anatomy
Using horseradish peroxidase staining in cats, Mayberg et al. showed that there was ipsilateral sensory innervation of V1 to the proximal > distal MCAs, and middle meningeal arteries, as well as bilateral innervation of V1 to the superior sagittal sinus, but other vessels weren't studied. Other experiments have suggested trigeminal innervation of predominantly the MCAs, but also the intracranial ICAs, ACAs, PCAs, Pcomms, SCA, and rostral basilar artery. The rostral basilar artery may be innervated by a recurrent branch from the cavernous plexus of trigeminal origin that attaches to the abducens nerve briefly. One study suggested ipsilateral innervation of the AICA via the trigeminal ganglion as well, while another did not. The VAs do not appear to be innervated by the trigeminal, and the PICAs have not been evaluated  Extracranially, the trigeminal ganglion seems to innervate the superficial temporal, external maxillary, internal maxillary, and lingual arteries. Similar findings have been noted in primates although there appear to be some contributions of V2 as well in some animals.

The SCA, rostral and caudal basilar artery, AICA, and vertebral arteries also have some innervation from the C1-C3 dorsal root ganglia. There are accordingly reports of vertebrobasilar infarction in associated with cervical (C2) VZV zoster as well as CN 7 VZV zoster (Ramsay-Hunt syndrome). However, there are also reports of patients with VZV rash very far from the vessels involved. This is hypothesized to possibly be due to concurrent VZV reactivation in multiple nerves, manifesting cutaneously in one but not cutaneously in another that leads to effects on the vessels.

There is also ipsilateral sympathetic innervation from the superior cervical ganglion to the proximal > distal MCAs and middle meningeal arteries, and bilateral innervation to the superior sagittal sinus. The MCA also appears to have parasympathetic innervation from the sphenopalatine ganglion bilaterally, and the ipsilateral > contralateral otic ganglion.

Pathophysiological process
In disseminated VZV infection, the virus may gain access to the vasculature systemically through the meninges, CNS, or vessels themselves.

In more focal VZV zoster infection (i.e. shingles), VZV likely travels along the nerve to its associated innervated vessels (see table above), where it infects them. As the nerves start in the adventitia, the virus seems to infect from the adventitia and then work its way inward to the media and the intima. This causes inflammation and vascular remodeling via various mechanisms that remain poorly elucidated, that ultimately lead to vascular injury. There are probably multiple ways that VZV can gain access to the cerebral vasculature. In disseminated VZV infection, the virus may gain access to the meninges, CNS, or vessels themselves, cause meningoencephalitis, and thereby infect the vessels. This may explain why bilateral vascular involvement seems to be more frequently seen in patients with disseminated VZV or those who are immunosuppressed.

For more focal VZV zoster infection, the best hypothesis is that VZV travels along the trigeminal sensory innervation of the intracranial vasculature (or along C1-C3 innervation of intracranial vasculature) and then invades into the vascular wall. The vaso nervorum lies within the adventitia, and early infection seems to start here and then invade more deeply into the tunica media and then the intima. It has been suggested that cerebral vasculature may be affected more than other vessels because cerebral arteries do not have an external elastic lamina to prevent spread of virus from the adventitia into the media. Electron microscopy in several cases has shown evidence of herpes-like viral intranuclear particles inside smooth muscle cells in the tunica media in such cases. Anti-VZV antibody immunoperoxidase staining has confirmed presence of VZV antigens within the tunica media in several cases as well, and VZV DNA has also been isolated from extracts of affected vessels.

After VZV invasion into the vessel wall, neutrophil infiltration appears to occurs in the adventitia. CD4+ and CD8+ T-cells, CD68+ macrophages, and occasional CD20+ B-cells have been noted in the adventitia and intima but not in the media. Through an unknown process, the virus and the inflammation subsequent lead to vessel wall remodeling. VZV itself may spread by inducing the release of microparticles from brain adventitial vascular fibroblasts, and may also induce these adventitial fibroblasts to transform into myofibroblasts causing vascular remodeling.

CSF of patients with VZV vasculopathy shows elevations in proinflammatory cytokines such as IL-8 and IL-6, as well as elevated MMP-2. How these relate to the pathophysiology of the disorder remains unclear. VZV does appear to downregulate programmed death ligand-1 (PD-L1), a protein which normally suppresses the immune system through activation with its receptor, programmed cell death protein 1 (PD-1). As VZV seems to downregulate PD-L1, this could promote persistent inflammation of the infected arteries and explain how immune cells can persist for months after diagnosis.

Pediatrics
VZV accounts for a substantial portion of childhood arterial ischemic stroke.

The risk of stroke is increased 4-fold after chicken pox.

Varicella vaccination appears to be safe, with no increased risk of stroke. In children, VZV may account for 7-31% of arterial ischemic stroke, and stroke may occur in 1 of 15,000-26,000 varicella cases. Children have a 4-fold increased risk of stroke in the 6 months after having chicken pox.

Chickenpox varicella vaccination appears to be safe, with no evidence of an increased risk of ischemic stroke after varicella vaccination.

Adults
The risk of stroke is 2-fold in the first month after VZV zoster infection in any/all locations, with the risk decreasing thereafter but persisting for at least one year and probably longer. The risk is higher in young patients. Ophthalmic VZV in particular follows a similar risk pattern, but the risk is substantially higher (2 to 4.5 fold increased risk within 1 year). In a mixed zoster population, this risk appears to be very high in the first month after onset of VZV zoster (incident risk ratio of 1.5-2.3) and tapers thereafter, but most studies suggest the risk may remain elevated for at least one year and probably longer, with only one study suggesting that risk only remains significantly elevated for 3 months. The risk of TIA or stroke after VZV infection seems to be highest in young patients (especially under age 40), with the risk decreasing with advancing age but still remaining significantly increased compared with non-VZV controls. The risk of hemorrhagic stroke may also be increased as well in the first 3 months (IRR 1.53, 95% CI 1.11-2.11).

A 2016 meta-analysis noted a relative risk of stroke of 2.36 (95% CI 2.17-2.56) for the first two weeks, then 1.56 (95% CI 1.46-1.55) for the first month, then 1.17 (95% CI 1.13-1.22) for the first year, and then 1.089 (95% CI 1.02-1.16) for more than a year. Other meta-analyses from 2016-2019 have generally shown similar findings.

After chickenpox, one study suggested that adults have have a 2-fold increased risk of stroke in the subsequent 6 months. The highest risk is clearly with ophthalmic VZV, with studies showing a 2 to 4.5-fold increased risk of ischemic stroke within 1 year. The risk of stroke after ophthalmic VZV may be highest in the first month.

Signs and symptoms
VZV often involves both small and large arteries. It is frequently multifocal/bilateral in patients with more systemic VZV infection or immunosuppression (who often have coexisting signs of meningoencephalitis), but more commonly unilateral in those with focal VZV recurrent infection (i.e. shingles). There are two general clinical forms of VZV vasculopathy, focal and multifocal. Focal most commonly occurs in vessels associated with the location of a VZV zoster rash, but a rash is not necessary. Multifocal occurs more commonly in patients with disseminated disease, often in those who are immunocompromised.

VZV vasculopathy characteristically involves large and small arteries. In one study of 23 patients, 50% of cases had both small and large artery involvement, with 37% having purse small artery involvement and 13% showing only large artery disease.

Focal VZV vasculopathy
The disorder most commonly occurs with V1 VZV zoster infection. The timing of neurological symptoms has been reported anywhere from concurrently with a rash to 6 months later. However, there are numerous reports of patients with VZV vasculopathy without any evidence of prior rash. , which may occur in as many as 37% of diagnosed cases. In those who have a preceding rash, the average time from rash to neurological symptoms is approximately 4 months.

Most commonly patients developed hemiparesis, +/- aphasia or encephalopathy, but they may have any neurological symptoms depending on the location of the vessels involved. There are several reports of it occurring after primary VZV (chickenpox) infection,  and it has been shown to be a risk factor for childhood stroke. It has also been seen in disseminated zoster, focal VZV zoster in the V2 territory, or focal VZV zoster in other areas of the body. In multiple reports, patients with cervical (C2) VZV zoster, developed pontine infarction several weeks later.

Multifocal VZV vasculopathy
VZV often affects multifocal vessels in immunocompromised patients, and often coexists with signs of VZV meningoencephalitis include headache, fever, mental status changes, and focal deficits. However, it can rarely occur in immunocompetent patients.

Rare findings
Although ischemic stroke is most common, there are a variety of other vascular abnormalities and disorders that can occur in VZV vasculopathy.
 * Cerebral aneurysm  . In one report, 9 aneurysms developed over a 2 month period and these improved with medical treatment (2 aneurysms even resolved).
 * Intracranial dolichoectasia
 * Intraparenchymal hemorrhage and microhemorrhages
 * Subarachnoid hemorrhage
 * Spinal cord infarction, usually after thoracic VZV zoster
 * Cervical artery dissection
 * Venous sinus thrombosis
 * Moyamoya syndrome

Differential diagnosis

 * Primary angiitis of the nervous system
 * Giant cell (temporal) arteritis
 * Neurosarcoidosis
 * Neurosyphilis
 * Tuberculosis CNS infection
 * Fungal CNS infection

Imaging
Essentially any vessel, large or small, can be involved. Imaging often reveals infarcts in both deep and superficial territories, often at the grey-white junction. Narrowing of vessels with vasculopathic changes are often seen, but may be absent in 70% of cases (as they may have small vessel disease only). Vessels can show contrast-enhancement in the vessel wall.

Although not yet well-studied, vessel wall imaging may be useful for diagnosis and for monitoring disease progression/improvement. Vessels involved have including intracranial ICAs, M1 segments of the MCA, A1 and A2 segments of the ACA, lenticulostriate arteries, anterior choroidal artery, PCAs , the  retinal arteries, Pcomm and associated thalamic branches, brainstem, cerebellum, and spinal cord. Contrast-enhancement may be seen in the vessel wall suggesting vasculitis. Vessel wall imaging may prove to be useful in the future, as findings show improvement with treatment, so it may be useful to monitor the disease.

Vessels involved are usually unilateral, but can be bilateral especially in immunocompromised patients.

The disorder commonly leads to infarctions in both gray and white matter, and frequently at the gray-white junction. Imaging is abnormal in essentially all cases, except for those with vasculopathic changes limited to the eye and optic nerve.

Vascular abnormalities are clearly seen in approximately 70% of cases on imaging, so an absence of vascular abnormalities cannot be used to rule out the disease.

Basic CSF studies
Patients often have a pleocytosis and elevated protein with elevated oligoclonal bands, but none of these have to be present.
 * Pleocytosis (lymphocytic) is common, but may not be present in as many as 33% of cases.
 * Protein: usually elevated
 * Glucose: normal
 * Oligoclonal bands: often present, and contain VZV-specific IgG

CSF antibody testing
CSF VZV IgG (and IgM) are the tests of choice, as they are far more sensitive than PCR testing. However, PCR testing is usually sent concurrently as there are rare cases that are VZV PCR positive but antibody negative (more likely to happen in immunosuppressed patients). In 1981, Kuroiwa and Furukawa first reported on the presence of anti-VZV antibodies in the CSF (1:32) and in serum (1:256) of a patient with left MCA involvement of VZV vasculitis, which decreased to 1:2 and 1:16 in serum four months later. A year later, Vilchez-Padilla et al. reported again on the presence of high levels of anti-VZV antibodies in the CSF (1:128) and serum (1:32) of a patient with Hodgkin's lymphoma with angiography-confirmed vasculitic changes, but with no prior VZV rash. CSF VZV antibody testing appears to be more sensitive than PCR testing.

CSF VZV IgG antibodies are positive in a number of patients with VZV infection without clear evidence of vasculopathy (in one study 10 of 44 patients sampled). The time course of CSF VZV IgG in patients without clear CNS symptoms is that it may begin to be presented from 1-7 days after onset, but is very frequently present 8-30 days after onset. After 1 month, CSF VZV IgG antibody titers may or may not return to normal.

If CSF VZV IgG antibody titers are higher than those in the serum, this indicates more intrathecal IgG production than systemically, and is strong evidence of active viral infection.

Up to 93% of patients with VZV vasculopathy have positive CSF VZV IgG antibodies and a reduced serum/CSF ratio of VZV IgG indicated intrathecal VZV IgG production.

This test is available through ARUP and Mayo Clinic laboratories.

CSF VZV PCR testing
This is a specific but very insensitive test for the disease, and should never be sent in isolation to rule out this diagnosis -- it must be sent concurrently with CSF VZV antibody testing. While this may be present in patients, it is not very sensitive as patients can have VZV vasculopathy with a negative CSF VZV PCR. CSF VZV PCR is only positive in ~30% of patients and is more frequently positive in immunocompromised patients.

CSF VZV PCR is positive in a number of patients with VZV infection without clear evidence of vasculopathy (in one study 10 of 42 patients sampled), although in 90% of these cases VZV DNA was no longer detectable two weeks after onset of the VZV rash. Due to its low sensitivity, it is not a useful test for this diagnosis.

Brain biopsy
Brain biopsy should not be necessary. in most cases. Biopsy is less useful as viral DNA and antigens are restricted to the cerebral arteries and may not be captured in the sample.

Antiviral agents
Although there is no trial data, most recommend IV acyclovir of 10-15 mg/kg IV q8h for 10-14 days (similar to meningitis dosing). For cases that do not improve, oral valacyclovir 1g q8h can be considered for an additional 1-2 months. Reports of acyclovir use for the condition began to emerge in 1983 and this became an accepted treatment over the years. A commonly used dose is 10 mg/kg IV q8h. In a mixed adult and child population of 30 patients, 66% of patients treated with acyclovir alone improved or stabilized, while 75% treated with acyclovir and steroids improved. Therefore adding steroids may be of utility. In a large study of patients with VZV zoster without vasculopathy, the risk of stroke was numerically (but not statistically) less in those treated with oral antiviral medications. In cases who did not have stabilization or improvement after IV acyclovir treatment, experts have recommended oral valacyclovir 1g q8h for an additional 1-2 months. One case report suggested benefit of treatment even after two years of progression of disease.

Steroids
This may be of benefit when used for a short course, at doses of prednisone 1 mg/kg daily for 5 days. Long courses have not been shown to be beneficial and could increase the risk of viral recurrence. In a mixed adult and child population of 30 patients, 66% of patients treated with acyclovir alone improved or stabilized, while 75% treated with acyclovir and steroids improved. One commonly used regimen is oral prednisone 1 mg/kg daily for 5 days. It is reasonable to discontinue steroids within one week, due to the risk of viral infection with long-term steroid treatment.

Antithrombotics
Antiplatelets can be considered but there is little evidence. Full-dose anticoagulation is usually avoided du eot the risk of hemorrhagic complications. In pediatric patients treated with antithrombotic agents only (aspirin or heparin), approximately one third of patients developed recurrent ischemic stroke or TIA within 33 weeks after presentation. Its efficacy remains unclear.

Surgery
There is one report of a patient who developed a moyamoya-like pattern who improved after undergoing EC/IC bypass surgery.

Other approaches
Moskowitz and Henrikson suggested in 1985 that colchicine may block axonal transport of VZV so that it could not spread to the cerebral vessels. However, this has not been studied in more detail.

Prognosis
In pediatric patients with post-varicella VZV vasculopathy without antiviral treatment, the disease was monophasic, with progression for up to 6 months followed by regression for as long as 48 months in one study In a mixed adult and child population of 30 patients, 66% of patients treated with acyclovir alone improved or stabilized, while 75% treated with acyclovir and steroids improved. Treatment can even show benefit when started very late after the onset of disease, so it should probably not be withheld from anyone.