CMV, neurologic



  • CMV disease afflicts up to 40% of untreated AIDS.
    • CMV seroprevalence 50-80%, >90% in active MSM.
    • CNS disease is seen only in immune-suppressed, HIV CD4 < 50-100.
  • CMV CNS disease in HIV/AIDS: relatively uncommon
    • Pathological evidence in >30% (some autopsy series), but < 2% with clinical neurological disorders.
    • Typically a progressive encephalopathy, usually in pt w/ hx of prior CMV disease (e.g., retinitis).
      • Acute onset, rapid progression helps distinguish from HIV encephalitis or PML.
      • CMV encephalitis subsets:
        • Diffuse (decreased memory with dementia-like presentation usually non-acute, attention, motor/sensory/CN deficits, ataxia +/- fever and often confused with HIV-related dementia)
        • Ventriculo-encephalitis (radiculopathy, CN deficits, nystagmus less neurocognitive features presenting in more aggressive fashion)
        • Mass lesion (focal deficits relating to mass)
  • CMV polyradiculitis: presents as back pain, sciatica, paresthesia, sphincter dysfunction (urinary retention), distal sensory loss, ascending paralysis.
    • May appear Guillain-Barré-like with lower extremity weakness or include urinary retention/loss of bowel function. Transverse myelitis presentation w/ para- or quadriplegia, sensory deficits.
      • Tip off to CMV presentation = neutrophilic CSF pleocytosis, typically 100-200 PMNs/μL.
    • Ddx includes HSV, VZV.


  • Neuroimaging: MRI preferred over CT.
  • CMV encephalitis: periventricular enhancement (ventriculitis) or diffuse hyperintense T2 images on MRI.
    • Dx based on clinical presentation, imaging studies, and usually by CSF PCR (sensitivity 62-100%, specificity 89-100%).
    • Brain biopsy: CMV inclusions ("owl’s eye") or CMV+ stains. A biopsy is not typically performed given the usual brainstem or periventricular locations of the disease.
  • CMV polyradiculomyelitis: MRI may show enhancement of cauda equina or meninges. Positive CSF CMV PCR.
  • CSF studies:
    • Findings often nonspecific, expect lymphocytic pleocytosis for CNS disease.
      • CSF may yield low glucose and/or PMN pleocytosis seen with encephalitis (ventriculo-encephalitis > diffuse) and polyradiculomyelitis (100-200 PMNs/ml on average).
    • CSF CMV PCR: a preferred method for diagnosing CMV neurological involvement.
    • CSF CMV viral cx: insensitive, but 100% specific if positive.
    • CMV PCR (blood, quantitative): role less clear, and not recommended because of poor predictive value.
      • Viremia may exist without evidence of end-organ disease.
      • High/rising levels may correlate with increased incidence of end-organ disease.
    • Serology: CMV-specific serology only helpful if negative CMV IgG, which would make CMV infection unlikely.
  • CMV retinitis may occur concomitantly in up to 30% of HIV CMV neurological patients. Ophthalmological examination of all pts recommended.


CMV encephalitis/polyradiculomyelitis: primary induction

  • Preferred: given poor outcomes, many prefer to use combination CMV therapy, although weak data support its use and toxicities are significant.
    • Ganciclovir 5 mg/kg IV q12h PLUS foscarnet 90 mg/kg IV q12h
    • An open-label trial showed 94 d median survival vs. 42 d historical control using monotherapy with a combination regimen poorly tolerated[8].
    • Monitor at least twice weekly CBC, chemistries including phosphorous during induction, once weekly during maintenance.
  • Alternatives:
    • Monotherapy with either ganciclovir or foscarnet.
    • Cidofovir 5 mg/kg IV q wk x 2wks, then dose q2wks (Note: CSF penetration of cidofovir not well studied).
  • The decision to use ganciclovir and/or foscarnet often in part decided upon hematologic and renal aspects.
  • Duration, induction therapy: 2 wks or until resolution of neurologic symptoms (not well-established)
    • If using combination ganciclovir + foscarnet, upon reaching stabilization of neurologic symptoms may switch to monotherapy.
  • Immune reconstitution with ART should be the concomitant goal.
    • Potential exists for IRIS worsening CMV neurologic infections but is likely rare (~0.04 per person-year).
    • Most experts do not delay ART more than two weeks after starting anti-CMV therapy.

CMV encephalitis/radiculomyelitis: post-primary induction

  • Preferred:valganciclovir 900 mg PO twice-daily
  • Alternatives:
    • Ganciclovir 5 mg/kg/d IV.
    • Foscarnet 90-120 mg/kg/d IV
    • Combination therapy: valganciclovir 900 mg PO twice-daily + foscarnet 90 mg/kg/d IV.
  • Usual duration of therapy: 3-6 wks if immune reconstitution occurs.
    • Otherwise, long-term maintenance therapy recommended with profound immune suppression to reduce the risk of relapse.
    • Regimen for maintenance unclear but valganciclovir 900 mg once- or twice-daily preferred due to oral route.
  • Oral ganciclovir rarely used because valganciclovir has superior bioavailability.

Selected Drug Comments




May have some in vitro and in vivo activity against CMV, but not indicated treatment since more active drugs now available.


Second-tier choice; probably effective, though renal toxicity and lack of data regarding CSF penetration preclude routine consideration.


Alternative choice for induction, though significant renal dysfunction will preclude use. Intermittent shortages may complicate use.


Favored agent for induction therapy, although patients with severe hematological problems despite transfusions/G-CSF etc. may do better with foscarnet.


Would not use for induction therapy due to multiple treatment failures in non-retinitis CMV infection. Not studied for CNS disease, but probably effective as oral maintenance therapy given drug bioavailability means acceptable ganciclovir levels.



  • CMV progression despite monotherapy: consider combination therapy ganciclovir + foscarnet (above dosing).
    • Progression despite combination therapy: consider cidofovir.
    • Role of oral valganciclovir is not established for neurological disease as maintenance, awaiting sufficient immune reconstitution (see CMV retinitis).
  • Ganciclovir resistance well-described, and may rarely occur even in previously untreated pts.
  • Generally poor prognosis for pts w/ AIDS and clinical CNS CMV. Neurological sequelae common, and existing deficits at the time of treatment initiation may not reverse.


  • Radiculopathy tends to improve within 2-3 wks.
  • In severe cases, primary induction/maintenance dosing may need to be continued beyond 6 wks until sufficient clinical response noted. Role of serial CMV PCR testing unclear as to guiding intensity/duration of therapy.


  • Typical CMV encephalitis pt has had prior CMV disease (e.g., retinitis, GI), lapsed maintenance therapy, advanced HIV w/ low CD4, ventricular/white matter lesions on MRI, and (+) CSF CMV PCR studies.
    • Treatment and dosage recommendations based on studies of CMV retinitis and colitis given the lack of controlled studies on CNS CMV infection.
  • Clinical disease usually a consequence of infection reactivation.
  • CMV viremia may occur in absence of end-organ disease.
  • Drug resistance was primarily seen in those receiving long-term anti-CMV therapy.
    • Low-level resistance: CMV UL97 (phosphotransferase) gene mutations.
    • High-level resistance: usually due to mutations in both CMV UL97 and UL54 (DNA polymerase) genes.
    • Determine resistance by CMV UL97 gene sequencing.
    • CMV UL97 mutants usually respond to foscarnet.

Pathogen Specific Therapy


1st Line Agent

2nd Line Agent




Basis for recommendation

  1. Guidelines for the Prevention and Treatment of Opportunistic Infections in HIV-Infected Adults and Adolescents, 2015 (last updated 11/4/15, last reviewed 6/26/2019, accessed 5/20/21).

    Comment: CMV section last updated in 2015. The guideline gives moderately weighted recommendations for combination therapy at outset, but also acknowledges significant toxicities with this approach and limited evidence. An update said to be in progress (2019).


  1. Palma P, Costa A, Duro R, et al. Mononeuritis multiplex: an uncommon neurological manifestation of cytomegalovirus reactivation in an HIV-infected patient. BMC Infect Dis. 2018;18(1):554.  [PMID:30419847]

    Comment: Authors use guideline-recommended treatment; however, emphasize the selection of ART drugs which higher CSF penetration given that distinguishing between HIV-induced encephalopathy and CMV encephalitis may be difficult.

  2. Adjei AA, Armah HB, Gbagbo F, et al. Seroprevalence of HHV-8, CMV, and EBV among the general population in Ghana, West Africa. BMC Infect Dis. 2008;8:111.  [PMID:18706107]

    Comment: A study from Ghana found that among HIV-seronegative blood donors, the seroprevalence of CMV was 77.6% and among HIV/AIDS patients was 59.2%. CMV (along with HHV-8 and EBV) are endemic in this region.

  3. Martínez PA, Díaz R, González D, et al. The effect of highly active antiretroviral therapy on outcome of central nervous system herpesviruses infection in Cuban human immunodeficiency virus-infected individuals. J Neurovirol. 2007;13(5):446-51.  [PMID:17994429]

    Comment: Interesting report from Cuba where ART was only used starting in 2001. Central lab analyzing CSF from HIV+ pts found CMV (44%), EBV (28%), and dual herpesvirus infections (16%) infections as the most common important agents identified. The use of ART, not surprisingly, correlated with lessened mortality.

  4. Portegies P, Solod L, Cinque P, et al. Guidelines for the diagnosis and management of neurological complications of HIV infection. Eur J Neurol. 2004;11(5):297-304.  [PMID:15142222]

    Comment: Uses Cinque 1998 and Anduze-Faris 2000 to primarily guide recommendations.
    Rating: Important

  5. Griffiths P. Cytomegalovirus infection of the central nervous system. Herpes. 2004;11 Suppl 2:95A-104A.  [PMID:15319096]

    Comment: The author suggests reserving foscarnet only for ganciclovir-resistant cases due to nephrotoxicity concerns or ganciclovir intolerance.

  6. Maschke M, Kastrup O, Esser S, et al. Incidence and prevalence of neurological disorders associated with HIV since the introduction of highly active antiretroviral therapy (HAART). J Neurol Neurosurg Psychiatry. 2000;69(3):376-80.  [PMID:10945813]

    Comment: Study documents prevalence of HIV -associated dementia and polyneuropathy as significantly lower in 1997-8 (p=0.02), with the incidence of Toxoplasma encephalitis, decreased from 5.7% in 1995-6 to 2.2% in 1997-8 (p=0.015). However, because of small numbers, authors could reach no such similar conclusions regarding CMV CNS infection.

  7. Anduze-Faris BM, Fillet AM, Gozlan J, et al. Induction and maintenance therapy of cytomegalovirus central nervous system infection in HIV-infected patients. AIDS. 2000;14(5):517-24.  [PMID:10780714]

    Comment: Pre-HAART open-label European study examining combination therapy with ganciclovir and foscarnet in 31 pts. with acute CMV encephalitis (CMVe) (n = 17) or CMV myelitis (CMVm) (n = 14). Induction therapy = "foscarnet" 90 mg/kg + ganciclovir 5 mg/kg twice-daily followed by maintenance dosing. Improvement or stabilization achieved in 74% (23/31 pts). Among 23 pts. on maintenance, CMV progressed in 10 (first relapse of 126 days; range 64-264 days). Median survival time 3 mos.
    Rating: Important

  8. Whitley RJ, Jacobson MA, Friedberg DN, et al. Guidelines for the treatment of cytomegalovirus diseases in patients with AIDS in the era of potent antiretroviral therapy: recommendations of an international panel. International AIDS Society-USA. Arch Intern Med. 1998;158(9):957-69.  [PMID:9588429]

    Comment: The document serves as the basis for suggested therapies mainly from the pre-ART era. Though little data other than case series exist for neurological CMV disease, extrapolation from CMV retinitis and GI studies not unreasonable given excellent penetration of both ganciclovir (24-67% of serum level) and foscarnet (13-68%) into the CSF.

  9. Cinque P, Cleator GM, Weber T, et al. Diagnosis and clinical management of neurological disorders caused by cytomegalovirus in AIDS patients. European Union Concerted Action on Virus Meningitis and Encephalitis. J Neurovirol. 1998;4(1):120-32.  [PMID:9531020]

    Comment: Authors review data from literature and trials demonstrating that virus isolation or antigen detection in CSF not useful for HIV-related CMV encephalitis. Utility of CMV CSF-PCR appears to be the gold standard short of brain Bx: sensitivities 62-100% with 89-100% specificity.
    Rating: Important

  10. Arribas JR, Clifford DB, Fichtenbaum CJ, et al. Level of cytomegalovirus (CMV) DNA in cerebrospinal fluid of subjects with AIDS and CMV infection of the central nervous system. J Infect Dis. 1995;172(2):527-31.  [PMID:7622897]

    Comment: Autopsy series with proven CNS CMV found 12/13 had positive CSF CMV PCR while none of the control patients positive. Higher log CSF PCR correlated with findings of more severe disease, suggesting that quantitative studies could be useful correlates of infection extant.
    Rating: Important

  11. Kim YS, Hollander H. Polyradiculopathy due to cytomegalovirus: report of two cases in which improvement occurred after prolonged therapy and review of the literature. Clin Infect Dis. 1993;17(1):32-7.  [PMID:8394748]

    Comment: In this series, typical presentation is the subacute onset of leg weakness, numbness progressing to paraparesis or paraplegia. Bladder dysfunction is common. Authors note that improvement within 2-3 wks is typical for radiculopathy, but may take mos.

  12. de Gans J, Portegies P. Neurological complications of infection with human immunodeficiency virus type 1. A review of literature and 241 cases. Clin Neurol Neurosurg. 1989;91(3):199-219.  [PMID:2548785]

    Comment: An older study examining literature, including autopsy series suggesting that silent CMV infection of the brain very common in those who die of HIV (~33%) but only a few have clinically suspected neurological disease antemortem (2%).

  13. Silva CA, Oliveira AC, Vilas-Boas L, et al. Neurologic cytomegalovirus complications in patients with AIDS: retrospective review of 13 cases and review of the literature. Rev Inst Med Trop Sao Paulo. 2010;52(6):305-10.  [PMID:21225213]

    Comment: A review of 13 pts Dx’d between 2004-2008 found that only 4 (31%) were on ART but 8 patients (62%) had CMV neurological disease as presenting OI. In this group, most presented with diffuse encephalitis (7, 62%), then polyradiculopathy (7), rhomboencephalitis (1), and ventriculo-encephalitis, although some presentations included multiple presentations types. The mean CD4 count was 13.and overall mortality was 38%. A variety of treatment regimens used, but no conclusions could be reached regarding their use. Similar to the pre-ART era, 30% of pts present with concomitant CMV retinitis. Authors recommend routine opthamological exams in all patients with CMV neurological disease. No pts experienced IRIS.

Last updated: May 30, 2021