Despite the availability of combination antiretroviral therapy (cART), human immunodeficiency virus (HIV) sensory neuropathy and HIV-1- associated dementia (HAD, formally acquired immune deficiency syndrome [AIDS] dementia complex) remain common problems in people with HIV infection. The diagnosis and management of these conditions are discussed in this section. Other less common neurological diseases, including vacuolar myelopathy and mononeuritis multiplex, are also addressed.
HIV-associated spinal cord diseases
White matter vacuolisation of the spinal cord, known as vacuolar myelopathy (VM), was clinically evident in 25% of patients with HIV and found in 40% of patients at post-mortem in the pre-cART era. A productive HIV myelitis occasionally accompanies VM, or occurs alone. In VM, pathological changes in the spinal cord are strikingly similar to those found in vitamin B12 deficiency and, indeed, patients with HIV infection with VM have been shown to have abnormalities in their vitamin B12-dependent transmethylation pathway which leads to reduced levels of s-adenosyl-methionine. The prevalence of HIV myelopathy in the cART era is not well defined.
Patients with VM and HIV myelitis typically present with features of a spastic paraparesis (Table 1). The antemortem diagnosis of either condition is presumptive and relies on the exclusion of other causes of spinal cord pathology. In particular, patients require: investigations to exclude syphilis, vitamin B12 or folate deficiency; imaging of the spinal cord to exclude a mass lesion; and lumbar puncture to exclude central nervous system (CNS) infection with cytomegalovirus (CMV), varicella-zoster virus, herpes simplex virus or human T-lymphotropic virus I or II. A placebo controlled trial comparing methionine to placebo in patients with HIV infection with VM did not show any improvement in cord conduction times, or clinical disease state.
|HIV myelitis||Vacuolar myelopathy|
Variable proprioceptive loss
Prominent sensory ataxia
|Onset||Late(CD4+ Tcellcountusually<50cells/μL)||Late(CD4+ Tcellcountusually<50cells/μL)|
|Spastic paraparesis occurring earlier in the course of HIV disease is probably due to another cause.|
CSF viral load may be high
Exclude other causes
CSF viral load may below
Exclude other causes
|MRI findings||The main role of imaging is to exclude a mass lesion||
Normal, cordatrophy or increased white matter signal
Vacuolation of posterolateral white matter
Little inﬂammation seen
|Management||May respond to cART||TrialofcARTrecommended|
|HIV is thought to be important in the pathogenesis of both conditions, and there is no reliable test to diﬀerentiate between them. Consequently, although controlleddata are lacking, the use of cART (including agents known to penetrate the central nervous system)is recommended in both cases.Both conditions require ongoing supportive therapy, including physiotherapy, occupational therapy, management of incontinence and walking aids.|
|CSF: cerebrospinal ﬂuid; cART: combination antiretroviral therapy.|
HIV myelitis Diseases of the peripheral nervous system in HIV
A symmetrical, predominantly sensory, peripheral neuropathy is the most common neurological complication of HIV infection. Sensory neuropathy (SN) due to HIV itself (HIV-SN) is typically a complication of advanced HIV infection, and was clinically apparent in more than 30% of people with AIDS in the pre-cART era. HIV-SN is characterised by distal, symmetrical numbness, hyperaesthesia and often burning pain. Neurological examination may be normal, but ankle reflexes are often reduced or absent. The pathology of HIV-SN includes axonal degeneration and macrophage infiltration of the nerve. Cytokine dysregulation is thought to play a role in the pathogenesis of this condition, as it does in VM and HAD.
A SN may also occur as a complication of the use of nucleoside reverse transcriptase inhibitors (NRTIs) – stavudine, didanosine or zalcitabine. This condition is variably called nucleoside or antiretroviral-toxic SN. The condition is clinically and pathologically similar to HIV-SN, but is temporally related to therapy with one or more of the NRTIs mentioned above. SN related to the use of NRTIs may occur at any stage of HIV disease, but people with more advanced HIV disease, a previous diagnosis of HIV-SN or risk factors for peripheral neuropathy of any kind (such as diabetes or nutritional deficiencies) may be at increased risk. Increasing height and host genetics have also been implicated. The pathogenesis of SN related to the use of NRTIs is thought to relate to NRTI-induced mitochondrial dysfunction. Australian data have demonstrated a significant increase in the prevalence of SN since the introduction of cART (from 13% in 1993 to 44% in 2001). This increase has occurred despite an overall improvement in the morbidity of people with HIV infection. A major risk factor for SN in 2001 was the prior prescription of didanosine or stavudine, suggesting that nucleoside and antiretroviral-toxic SN had become an important cause of the HIV-SN seen in clinical practice. However, despite a significant reduction in the use of stavudine and didanosine in Australia since 2001 and the removal of zalcitabine from clinical use, the prevalence of HIV-SN has remained unchanged. There are some recent data suggesting an epidemiological link between exposure to protease inhibitors and HIV-SN. Protease inhibitors are not known to be neurotoxic. It is unclear at this time whether this association is causal and, if so, whether this represents a direct drug toxicity or a secondary complication of the metabolic effects of protease inhibitors (such as glucose intolerance and hypertriglyceridaemia).
There is also an observed association between impairment in certain neuropsychological domains and HIV-SN, irrespective of whether the patients’ symptoms are in the lower limbs alone or involve both upper and lower limbs. Whether this neuropsychological impairment is peripheral in nature (due to the neuropathy) or there is an association between HIV-SN and HIV-associated neurocognitive disorder (as a parallel development due to disease progression) needs further elucidation.
Sensory neuropathy is a clinical diagnosis. The Brief Peripheral Neuropathy Screen (BPNS) is a simple and useful tool to assess patients and compare the severity of symptoms. Nerve conduction tests may confirm the diagnosis and show axonal neuropathy pattern. However, nerve conduction tests are often normal in HIV-SN and are not routinely necessary. Skin biopsy is rarely necessary, but often shows a significant decrease in epidermal nerve fibre density (ENFD), especially in those with high HIV viral load or more severe neuropathic pain symptoms.
For most patients with established SN, the management is symptomatic (Table 2) Therapeutic agents used for other causes of painful neuropathy, such as tricyclic antidepressants and anti-convulsive drugs, achieve only variable success in HIV-SN. In a systematic review comparing the therapeutic efficacy from randomised control trials,only smoked cannabis, recombinant nerve growth factor and high-dose capsaicin patch showed a superior analgesic effect compared to placebo. While smoked cannabis for pain management is still illegal in Australia, the use of recombinant nerve growth factor is for research purpose only. High-dose capsaicin (8%) patch causes initial depolarisation with subsequent dysfunctionalisation and reduction in epidermal nerve fibre density. It provides a prolonged pain control for at least 12 weeks after each application without significant adverse event in a recent trial. Currently, this patch serves as an off-label treatment for HIV-SN. As in other forms of neuropathy, symptom control is difficult and often incomplete.
|Table 2 The diagnostic and therapeutic approach to patients with HIV sensory neuropathy|
Exclude diabetes and heavy alcohol consumption
Exclude deﬁcienciesofthiamine,B12 andfolate
Remove any other neurotoxins (including isoniazid,dapsone,thalidomide,high-dose pyridoxine or metronidazole)
|Remove the underlying cause||
HIV-SN -may improve when viral replication is controlled
ATSN-ceaseorreduce the dose of the causative NRTI where practical (symptoms may worsen initially)
(relief rarely complete)
Simple or compound analgesics (mildpain):
Paracetamolor NSAIDs plus orminuscodeine
Trial of oralmorphine foreﬃcacy,then switch to long-acting alternative
|Pain modiﬁcation (in combination with analgesiaf or more severe pain or when relief is incomplete)||
Tricyclic antidepressants (start with low dose, e.g.amitriptyline1 0mg at night, and increase gradually according to side eﬀects and eﬃcacy)
New antidepressants may be useful when tricyclics are not tolerated
Sodium valproate (start 200mg three times daily, increase gradually)
Carbamazepine (leukopeniamay limit use)
Phenytoin(drug interactions may limit use)
Gabapentin (not funded under thePharmaceuticalBeneﬁts Scheme for this indication, but available at some tertiary referral centres);lack of interactions with antiretroviral agents useful; start 300 mg at night andtitrate according to relief ands edation
Pregabalin (similar mechanism as gabapentin but usually less side effect and better absorption profile; start 75mg twice a day and titrate accordingly)
8% capsaicin patch (apply the patch to the painful site for 30 minutes; oral analgesia may be necessary for initial burning pain on application)
|Other treatments||Anecdotal reports support the use of various complementary therapies to relieve the symptoms of sensory neuropathy, including acupuncture and massage|
|Treat any coexisting depression||Depression is common with chronic pain, andrequires adequate assessment and treatment|
|ATSN:antiretroviral-toxic sensoryneuropathy;NRTI: nucleoside reverse transcriptase inhibitor; NSAID:non-steroidal anti-inﬂammatory drug|
Acute demyelinating and chronic inflammatory demyelinating polyneuropathy
Acute demyelinating polyneuropathy, which closely resembles Guillain-Barré syndrome, may occur at HIV seroconversion, early in the course of HIV disease or as an immune reconstitution disorder. Chronic inflammatory demyelinating polyneuropathy may occur during the early or later stages of HIV disease.
Acute demyelinating polyneuropathy and chronic inflammatory demyelinating polyneuropathy are both characterised by progressive, ascending weakness with early loss of reflexes. The pathogenesis of both entities is undetermined, but may have an autoimmune basis.
Patients should be investigated with nerve conduction studies (confirming a demyelinating process) and lumbar puncture. Cerebrospinal fluid (CSF) examination typically shows a mild, mononuclear pleocytosis and raised protein, rather than the CSF findings of classical non-HIV-associated Guillain-Barré syndrome in which the CSF is acellular, but the protein concentration is elevated. The response to immune-based therapies (including plasmapheresis or intravenous immunoglobulin) is similar to that seen in non- HIV-associated acute demyelinating polyneuropathy and chronic inflammatory demyelinating polyneuropathy.
Mononeuritis multiplex is characterised by the acute onset of one or more nerve palsies. The likely underlying cause of mononeuritis multiplex varies at different stages of HIV disease. A relatively benign form of mononeuritis multiplex may occur early in the course of HIV infection, and commonly resolves without specific treatment. In patients with moderate immunosuppression (CD4 T-lymphocyte [CD4] cell count 200-500 cells/μL), mononeuritis multiplex may be related to immune complex disease secondary to hepatitis B virus (HBV) or hepatitis C virus (HCV) infection, to vasculitis, or to an underlying infiltrating malignancy. In advanced HIV disease, mononeuritis multiplex is more likely to be due to nervous system involvement with an opportunistic infection, notably CMV.
A presumptive diagnosis of mononeuritis multiplex can be made on clinical findings, but investigations should include serology tests for HBV and HCV infection, CSF cytology and polymerase chain reaction (PCR) assays on the CSF for CMV, herpes simplex virus and varicella-zoster virus.
Around 80% of patients with CMV-associated mononeuritis multiplex will respond to treatment, and a trial of therapy with ganciclovir or foscarnet is warranted.
HIV-associated neurocognitive disorders
Neurocognitive disorders represent a significant problem for people living with HIV infection. The spectrum of severity of neurocognitive impairment ranges from having none to having a dementing illness, HAD. However the progression of disease may vary wherein up to 20% of patients may fluctuate between being impaired and normal.
HIV-associated asymptomatic neurocognitive impairment
HIV-associated asymptomatic neurocognitive impairment (ANI) refers to those patients who are fully functional without obvious neurocognitive impairment, but who are found to have neurocognitive abnormalities upon formal neuropsychological testing, defined as 1 standard deviation below the mean in two cognitive domains.
HIV-1-associated mild neurocognitive disorder
HIV-1-associated mild neurocognitive disorder (MND) refers to those patients with clinical and neuropsychological evidence of neurocognitive impairment (1 standard deviation below the mean in two cognitive domains) and impairment in their daily functioning.
HIV-1-associated dementia (HAD) patients have notable disability in their daily living. Neuropsychologically, it is defined as at least 2 standard deviation below the mean in two cognitive domains. The availability of cART has been associated with a significant decrease in the incidence of HAD from 7% to approximately 3% per year in immunocompromised patients. Furthermore, Dore et al. have shown that since 1996 there has been a significant increase in the median survival of patients with HAD in Australia from 11.9 to 48.2 months. However, cART has also been associated with a shift in the natural history of HAD. Since 1996 there has been an increase in the proportion of patients presenting with HAD as their AIDS-defining illness in Australia and, overall, patients appear to be developing HAD with higher CD4 cell counts, notably within the 200-350 cells/μL range HIV-1 subtype C is associated with less severe HAD.
Changing characteristics of HAND in the cART era:
Although the incidence of HAD is decreasing, the incidence of the less severe form of HAD (ANI and MND) remains static, ranging from 20-50% in various studies. It has become the predominant subtype of HAD in the cART era. Indeed, a further shift in various aspects of HAD is anticipated as the Australian population with HIV infection ages.
Metabolic risk factor profile
Metabolic risk factors for HAD include older age, insulin resistance, diabetes, host genotype, hypertension and raised cholesterol. Increased cerebrovascular small-vessel disease was recently noted to be more prevalent in older patients and in those with hypertension. Together with aging, a major concern is the possible development of vascular cognitive impairment superimposed upon HAD.
The commonest presentation of HAD in the pre-cART era is progressive subcortical dementia. Some of the noted clinical risk factors for the development of HAD include the presence of minor cognitive and motor disorders, extant cognitive deficits and depression, mania, older age at presentation and low educational level. Symptoms included inattention, forgetfulness, slowed thinking, gait unsteadiness, impaired handwriting and social withdrawal sometimes with irritability. Involvement of cortical function has been increasingly reported in the cART era. Executive function and learning ability are now frequently affected domains in HAD patients.
Overlapping with other neurodegenerative disease
HAD is a subacute and subcortical dementia that principally involves the deep white matter and the basal ganglia. The chief neuropathological findings associated with HAD are cerebral atrophy, myelin pallor and the presence of multinucleated giant cells, predominately within the deep white matter and the basal ganglia. Overall, there is reasonable, albeit incomplete, correlation between the presence of HIV in the brain and the clinical and neuropathological severity of HAD. However, there is growing evidence suggesting that HAD also shares similarity with degenerative disorders such as Parkinson disease and Alzheimer disease in terms of pathogenesis, histopathological findings, genetic screening and CSF studies. Whether HAD shares a similar pathogenesis with neurodegenerative disease, or whether HIV CNS infection accelerates underlying disease development as a form of accelerated aging requires further clarification.
Laboratory parameters in cART era
HAD occurs in the setting of moderate-to-severe immune suppression. In the pre-cART era, the median CD4 cell count in Australian patients presenting with HAD as their AIDS-defining illness was 70 cells/μL, whereas the median CD4 cell count in the cART era in the same population is 120 cells/μL (p = 0.012). Both current plasma HIV viral load and CD4 cell count are important predictors for developing HAD but this association is no longer true in the cART era. Thus far, only the nadir CD4 cell count serves as an important predicting factor for HAD across the treatment eras. Ongoing studies related to the benefit of early commencement of cART may clarify whether nadir CD4 cell poses a threshold effect on development of HAD.
Another well observed presentation of HAD in the cART era is its development in patients who have successful systemic HIV virological suppression. The observation of CSF viral escape or discordance, which means detectable HIV in the CSF and undetectable HIV in blood has become a growing concern in the last few years and is reported in about 10% of people with HIV infection. The presence of CSF discordance typically denotes that the HIV that is detectable in the CSF is resistant to the patient’s current cART regimen. It occurs in both asymptomatic and HAD patients. Other presentations include headache, sensory disturbance and encephalopathy. Improvement of symptoms was reported in most of these CSF discordant cases after switching to a cART regimen with better CNS penetration (see Management below) and achievement of undetectable HIV viral load in CSF.
No single clinical, laboratory or neuroradiological finding is pathognomonic of HAD; instead, the diagnosis of HAD rests on a composite of findings wherein other major differential diagnoses have been excluded. Screening tests such as the HIV dementia scale (HDS) provide good sensitivity and specificity for detecting HAD in the clinic setting. A revised cut-off score of 14 points on the HDS (out of a maximum of 16 points) still serves as a useful screening tool for milder forms of HAD in the cART era. An example of the diagnostic process and differential diagnoses for HAD is found in Table 3.
|Differential diagnoses to be excluded||Comments|
|Depression||In patients with a CD4+ T cell count nadir below 200 cells/μL who have been diagnosed with depression, repeat neuropsychological testing should be performed once the depression has been treated, to exclude residual, underlying HAD|
|Cerebral toxoplasmosis, cerebral lymphoma and progressive multifocal leukoencephalopathy (PML)||The lesions of cerebral toxoplasmosis, PML and lymphoma should be evident upon neuroimaging of the brain|
|Cryptococcal meningitis||If the serum cryptococcal antigen is positive, then CSF cryptococcal antigen, CSF India ink stain and fungal culture are recommended to exclude cryptococcal meningitis|
|Neurosyphilis||In patients with positive serum Treponema pallidum antibodies, a diagnosis of neurosyphilis should be considered and the CSF should be tested for T. pallidum antibodies and the Venereal Disease Research Laboratory test should be performed. In a patient with documented loss of T. pallidum antibodies as a result of advancing HIV disease, an empiric course of 15 days of intravenous penicillin should be considered|
|Other uncommon causes of dementia||In individual patients, consideration may be given to the possibility of other causes of dementia, including Alzheimer disease, multi-infarct dementia, Pick disease, Creutzfeldt-Jacob disease, hypothyroidism and heavy metal poisoning|
|MRI and MRS brain||
Cerebral atrophy usually seen. T2-weighted hyperintensities seen in deep white matter and periventricular areas.
MRS shows increased choline and myo-inositol, and decreased n-acetyl aspartate.
Note: none of these ﬁndings is speciﬁc to HAD
|Neuropsychological testing||Findings include decreased concentration, reduced psychomotor speed, reduced verbal and visual memory, and reduced executive function|
|General CSF analysis||Protein, glucose, microscopy, culture and sensitivity should be performed. A mononuclear pleocytosis is compatible with, but not speciﬁc for, a diagnosis of HAD|
|Plasma HIV viral load||
Development of cognitive impairment during complete plasma viral suppression is not uncommon in cART era
|CSF HIV viral load||
There is no cut-oﬀ above which the CSF viral load is pathognomonic for HAD; rather, theCSF HIV viral load correlates best with the severity of HAD in patients with CD4+ T cell counts below 200 cells/μL in the pre-cART era.
A study suggested that a baseline CSF HIV viral load above 400 copies/mL in neurologically intact patients signiﬁcantly predicts progression to neurocognitive impairment,but larger studies are needed to conﬁrm this.
|Plasma and CSF HIV genotype if patient is currently using antiretroviral agents||Resistance to NRTIs, NNRTIs and PIs has been reported in HIV isolated from the CSF,and discordance between resistance patterns in CSF and plasma has also been reported|
Combination antiretroviral therapy has been shown to improve the neurocognitive performance of patients with HIV-associated neurocognitive impairment. Some data suggest a greater cognitive improvement can be achieved by a cART regimen that has better central nervous system penetration. Enting et al. give a detailed discussion of the pharmacodynamics of antiretroviral drugs and the CNS. Indeed, the issue of how to determine the efficacy of individual antiretroviral agents within CNS is complex, which includes the consideration of their chemical properties, CSF concentration and effectiveness in achieving reduction in CSF viral load from the available literature.
To investigate this issue, Letendre et al. introduced a CNS penetration-effectiveness (CPE) score system of individual antiretroviral agents based on the considerations mentioned above. The overall CPE of a cART regimen is calculated by adding up the CPE numbers of each drug used. The original version of CPE was published in 2007 and has been revised (see Table 4). To date, studies have confirmed that a high CPE score, defined as a regimen with total CPE > 7, leads to higher chance of undetectable HIV viral load in CSF. However, the benefit of CPE on neurocognitive impairment is less clear. While studies have demonstrated that better neuropsychological performance or greater neuropsychological improvement are associated with a high CPE regimen, there are also reports failing to demonstrate superiority. While HIV replication in CNS is undoubtedly a key issue for HAD pathogenesis, the CPE system may not be a perfect model to account for all the contributing factors. For example, the effect of antiretroviral agents on astrocytosis and prolonged inflammatory response despite undetectable HIV RNA in CSF (as reflected in radiological and CSF markers studies) is not taken into account in the current CPE system.
An expert consensus was published in 2013 accounting for the usage of CPE in managing HAD patients. Practically, HAD patients who fail to achieve virological suppression in serum or CSF should be given a high CPE regimen. For those who have undetectable HIV viral load in CSF, a more sensitive HIV RNA assay with a lower limit of detection of 1-2.5 copies/mL should be considered. If HIV RNA is detectable using a high sensitivity assay, a high CPE regimen should be employed where possible. It is also reasonable to start a high CPE regimen in case the more sensitive assay is not available.
|Table 4 The revised central nervous system penetration-effectiveness (CPE) score for different antiretroviral agents.|
Central nervous system penetration scale
(Increasing penetration with increasing score)
NRTI: nucleoside analogue reverse transcriptase inhibitor; NNRTI: non-nucleoside reverse transcriptase inhibitor; PI = protease inhibitor; PI/r: ritonavir boosted protease inhibitor Note: Used with permission
Table 5 Symptoms candsigns of HIV-1-associated neurocognitive disorder (HAND)
Mainly subcortical symptoms: with Inattention, forgetfulness, slowed thinking, gait unsteadiness, impaired handwriting and social withdrawal sometimes with irritability
Cortical deficit are increasingly reported
Extrapyramidal features similar to Parkinson’s disease may occur
Patient’s family or partner may notice these changes first
Initially, patients may complain of increased effort in maintaining working performance or difficulties in retaining information from reading or watching TV
These changes generally take place over several months
Progression of HAD leads to profound dementia,
mutism, incontinence and paraparesis
Patients may demonstrate one or more of these signs
Focal sensory or motor signs, or cranial nerve palsies are not a feature of HAND
HAD: HIV-1-associated dementia; CSF: cerebrospinal ﬂuid; MRI: magnetic resonance imaging; MRS: magnetic resonance spectroscopy; NRTIs: nucleoside analogue reverse transcriptase inhibitors; NNRTIs: non-nucleoside reverse transcriptase inhibitors; PIs: protease inhibitors.
Monitoring response to treatment of HIV-1- associated dementia
A reasonable approach for monitoring would be to recommend that the patient undergo a repeat lumbar puncture with analysis of CSF HIV viral load, repeat neuropsychological testing, and repeat MRI or magnetic resonance spectroscopy (MRS) after 12 weeks of the new antiretroviral drug regimen. Although this approach remains unproven in clinical trials, it allows for the autonomous nature of CNS HIV infection, which has different and slower decay kinetics than that of plasma HIV infection. Treatment response may be appreciated by partial normalisation of N-acetylaspartate (NAA) level in MRS, together with cognitive improvement.
One of the key endpoints of cART for HAD is an undetectable CSF HIV viral load. In practice, once the CSF HIV viral load becomes undetectable, a repeat CSF analysis is not necessary if the patient’s condition is stable or improving. However, it is reasonable to continue to perform neurological and neuropsychological follow-up every 6 months, and to repeat a lumbar puncture if there is evidence of a relapse of the symptoms of HAD. It usually takes at least a month and often much longer to observe a significant neuropsychological improvement after commencement of cART. Indeed, stability may be delayed for up to 18 months.