Clinical Presentation.

In the United States, up to 6% of HIV-infected individuals have a history of syphilis. Treponema pallidum infection can also occur at any time during HIV-1 infection and mimics neurologic complications of AIDS. Indeed, both conditions may cause acute or chronic meningitis, myelopathy, cranial or peripheral neuropathies, cerebrovascular disease, and dementia. Therefore, it is paramount to distinguish neurosyphilis from latent disease with positive serology and normal physical examination in HIV-infected patients because it has a direct impact on treatment. Neurosyphilis can occur as early as 1 year or as late as 30 years after initial infection. In this chapter, we discuss only syphilitic meningitis; other neurologic complications of this disease are covered in Chapter 239.

Laboratory Investigations.

Elevation of protein concentration and leukocyte counts can be found in the CSF of approximately 15% of patients with primary syphilis and up to 40% of patients with secondary syphilis. Some of these patients eventually experience spontaneous cure of this earlier CNS infection, but the persistence of asymptomatic CSF abnormalities for more than 5 years in the untreated patient is highly predictive of the development of clinical neurosyphilis.

Persistent mononuclear pleocytosis and elevated protein concentration can be found in the CSF of HIV-1-infected patients with neurosyphilis,¹³ as well as an elevated immunoglobulin G synthesis rate and oligoclonal bands. This is of no help in establishing the diagnosis of neurosyphilis in the context of HIV-1 infection because these findings occur as well in asymptomatic HIV-1 seropositive patients, especially when they have detectable HIV-1 RNA in their CSF. Using either the criterion of elevated CSF protein of greater than 50 mg/dL or white blood cells greater than 10 cells/µL in HIV-positive patients with reactive rapid plasma reagin may lead to overdiagnosis of neurosyphilis in the absence of clinical symptoms. Follow-up lumbar puncture after 12 months showed persistence of CSF abnormalities in 62% of cases.¹⁴ A positive CSF Venereal Disease Research Laboratories (VDRL) test result establishes the diagnosis of neurosyphilis if the tap is not bloody. However, this test may be negative in HIV-1 infection.^15,16 A reactive CSF fluorescent treponemal antibody absorbed test increases the likelihood of T. pallidum infection but is less specific because it can result from treated neurosyphilis or from contamination of the CSF with small amounts of blood containing antibody at the time of the lumbar puncture.

Acute symptomatic syphilitic meningitis is usually the earliest clinical manifestation of neurosyphilis, which occurs within the first year of infection and can be associated with cranial nerve palsies, including isolated eighth nerve palsy, and signs and symptoms of hydrocephalus. CSF abnormalities are similar to those of asymptomatic neurosyphilis, except that the CSF VDRL test result is nearly always positive (also see Chapter 239).

Treatment.

HIV-1–infected individuals with a positive serum rapid plasma reagin (RPR) test result, unexplained CSF pleocytosis, and elevated protein concentration, as well as symptoms consistent with neurosyphilis, should be treated with intravenous penicillin even in the absence of a positive VDRL test result in the CSF. Treatment consists of intravenous penicillin G (3 to 4 million units IV every 4 hours for 10 to 14 days). In case of allergy to penicillin, ceftriaxone (2 g IV once daily for at least 14 days) is another option. A careful follow-up of these patients is necessary, and a repeat spinal tap 1 month after onset of treatment should show a normalization of the CSF cellularity and protein concentration. Normalization of serum RPR predicted normalization of CSF and clinical abnormalities 13 months after treatment in more than 90% of a cohort composed mostly of HIV-infected men. This was more frequent in those receiving ART than in untreated patients.¹⁷ Some patients have a transient increase in CSF HIV-1 viral load to more than 100,000 copies/mL associated with neurosyphilis, which subsides after antibiotic treatment. Neurosyphilis may amplify intrathecal HIV replication, possibly through immune activation that persists even after syphilis treatment.¹⁸ As is the case with other opportunistic infections of the brain, this viral burden is likely carried by activated circulating lymphocytes and monocytes coming to combat the CNS infection and should not be interpreted as a manifestation of HIV-1 encephalitis.

Clinical Presentation.

This is the most common opportunistic meningitis in AIDS, which affected 10% of patients before the ART era¹⁹ but has decreased in incidence since then (also see Chapter 264).²⁰ Cryptococcal meningitis differs from aseptic meningitis in that meningismus is present in less than 40% of cases, and patients may present with only fever and headache, which become progressively more debilitating. Confusion, blindness, or altered state of consciousness occurs in severe cases.

Laboratory Investigations.

The detection of Cryptococcus neoformans antigen titers by enzyme immunoassay in the CSF provides a rapid diagnosis, which is confirmed subsequently by CSF culture. Because false-positive results occur with all the antigen tests, confirmation by culture is essential to establish the diagnosis of cryptococcal meningitis. Other CSF findings include markedly elevated opening pressure, mononuclear pleocytosis, elevated protein and decreased glucose concentration in 50% of the cases, and direct detection of the organism by India ink staining in 80% of the cases. Serum cryptococcal antigen is usually detected in cryptococcal meningitis in AIDS patients, and blood and urine cultures may also be positive.

Brain imaging is usually negative unless an associated cryptococcoma or hydrocephalus is present. Poor prognosis factors include altered mental status, absence of CSF pleocytosis, CSF antigen titer greater than 1 : 1024, a positive blood culture, and hyponatremia.

Treatment.

Treatment should begin with amphotericin B (0.7 mg/kg) with flucytosine (100 mg/kg PO in four divided doses per day) until the patient is clearly responding and for not less than 2 weeks, which is associated with significantly increased rates of yeast clearance from CSF and decreased mortality compared with amphotericin B alone. Combination of fluconazole with amphotericin B showed no survival benefit.²¹ Use of fluconazole alone as initial therapy is associated with much slower culture conversion and is not recommended.²² Consolidation therapy then consists of fluconazole (400 mg PO once or twice daily) administered for 8 weeks or until CSF cultures are sterile. Lipid formulations of amphotericin B may be used for patients with renal insufficiency, and itraconazole may be substituted for fluconazole if patients can tolerate fluconazole but is clearly inferior to the latter. Lifelong maintenance therapy using fluconazole 200 mg/day has been previously recommended to prevent relapses.²³ Updated recommendations state that maintenance therapy can be discontinued if a patient has received a minimum of 12 months of antifungal treatment, CD4⁺ counts reach greater than 100 cells/µL, and there is a sustained undetectable or very low HIV RNA level for ≥3 months.^23a

The outcome is generally favorable within 2 weeks, but early mortality still reaches 6%.²⁶ In a study done in Botswana, the in-hospital mortality was significantly lower among those patients who were on ART at the time of diagnosis of cryptococcal meningitis compared with untreated patients (8% vs. 21%).²⁷ Complications such as increased intracranial pressure greater than 200 mm H₂O occur in nearly all patients. Such a complication should be recognized and treated aggressively with mechanical drainage, including repeat lumbar punctures, temporary external lumbar drainage, or intraventricular shunts.^28,29 Corticosteroids, acetazolamide, and mannitol have not been shown to be effective. Neurologic deterioration after commencing ART occurs in 26% of patients and is caused by cryptococcosis-associated immune reconstitution inflammatory syndrome (C-IRIS).³⁰ It is more frequent if ART is given 7 days after compared with 28 days after onset of antifungal therapy.³¹ Persistent CSF cryptococcal growth at ART initiation and poor CD4⁺ T-cell increases on ART are strong predictors of C-IRIS.³⁰ Management of C-IRIS includes continuation of ART, antifungal therapy, and a course of corticosteroids.³²

This common CNS complication of HIV-1 infection occurs in 15% of patients with AIDS and can be the first manifestation of the disease in 3% to 10%.³⁸ In the past, this condition has also been named AIDS dementia complex, HIV-1 encephalopathy, or HIV-associated major cognitive disorder. A less severe entity named HIV-1-associated minor cognitive/motor disorder (MND) occurs in an additional 20% to 25%, and asymptomatic neurocognitive impairment is used to categorize individuals with subclinical impairment.^39,40 Risk factors include an AIDS-defining illness, increased age and survival duration, lower nadir of CD4⁺ T-cell counts, and higher baseline HIV viral loads.⁴¹

The clinical characteristics of this disorder can be subdivided into three main categories: cognitive, behavioral, and motor (Table 127-1).

The mental status examination reveals minor psychomotor slowing, inattention, decreased short-term memory, inability to perform simple calculations, and frontal release signs. Ocular movement testing shows saccadic pursuit. Other frequent findings are brisk reflexes, mild postural tremor, slowing of rapid alternating movements, and gait instability when performing half-turns. If untreated, dementia becomes more global, profoundly impairing orientation, memory, and cognition. Confusional or psychotic episodes have been reported, but seizures are a rare occurrence. Despite the extent of the cerebral involvement, there is usually no aphasia, apraxia, or other signs of discrete cortical dysfunction, except in terminal stages. Therefore, this syndrome has been classified as a frontal-subcortical dementia. A rapid bedside evaluation can be performed with the HIV Dementia Scale.⁴² However, this screening test is not as sensitive as a more detailed neuropsychological evaluation, which should include tests of attention, memory, and psychomotor speed such as trail making, digit span, verbal fluency, grooved pegboard, symbol digit modalities, and Rey auditory verbal learning tests.⁴³ An international dementia scale can be used to identify individuals at risk of HIV-associated dementia despite language barriers.^44,45

Unfortunately, cognitive dysfunction persists despite ART. A cross-sectional study of 1555 HIV-infected adults in the United States showed neuropsychological impairment in 52%. Prevalence estimates for specific HIV-associated neurocognitive disorder (HAND) diagnoses were 33% for asymptomatic neurocognitive impairment, 12% for MND, and 2% for HIV-associated dementia.⁴⁶

Laboratory Investigations.

Numerous groups have detected early neurologic dysfunction in HIV-1-infected asymptomatic individuals.^47–49 Subtle electrophysiologic abnormalities can be found in early HIV-1 infection (on electroencephalography, evoked potentials, nerve conduction studies), but they do not seem to have a predictive value for the later onset of AIDS dementia, which occurs generally when the CD4⁺ T-lymphocyte counts are less than 200 cells/µL. CSF analysis shows mild lymphocytic pleocytosis in 25% and elevated protein in 55%,⁵⁰ which can also be found in nondemented patients. Increased CSF levels of nonspecific markers of immune activation or neuronal injury have been reported, but their use in patients’ management is limited because these markers are also elevated during opportunistic infections of the CNS and because these assays are not readily available in the clinical setting.⁵¹

Measurement of the HIV-1 CSF viral load has been evaluated as a surrogate marker in HIV-1 infected individuals with cognitive dysfunction. A wide overlap was found between CSF HIV-1 viral load values of demented and nondemented individuals, and concomitant opportunistic infections of the CNS must be ruled out because they also contribute to transient elevation of HIV-1 viral burden in the CSF.⁵² These findings suggest that there may be two phases of HIV-1 infection in the CSF. (1) A transitory infection of the CNS may occur early in the disease via trafficking lymphocytes. These viral strains have a lymphotropic phenotype (using CXCR4 co-receptor) and may respond to therapy in parallel to plasma virus load. (2) A more autonomous infection of CNS monocytes/macrophages may take place at a later stage; with macrophage tropic strains (using CCR5 co-receptor).⁵³ Such patients may need higher drug penetration in the CNS. Whether measurement of the HIV-1 CSF viral load provides an accurate representation of the viral replication occurring in the CNS is unclear. However, subsets of patients with higher CSF than plasma viral loads have been recognized in whom intrathecal viral replication correlates with neurologic deficits. Because HIV-1–associated dementia complex is a diagnosis of exclusion, CSF examination is indicated for bacterial, fungal, and acid-fast bacteria cultures; cryptococcal antigen; VDRL test; and cytology. In addition, CSF polymerase chain reaction (PCR) for JC virus (JCV), the agent of progressive multifocal leukoencephalopathy (PML), is indicated because this condition is often difficult to distinguish from HIV-1 encephalopathy on brain imaging studies (see later and also see Chapter 147).

Imaging Studies.

Computed tomography (CT) and magnetic resonance imaging (MRI) may show greater subcortical atrophy than cortical atrophy, which is not proportional to the degree of dementia. MRI may also demonstrate multiple hyperintense signals in T2-weighted images, which are nonenhancing, poorly demarcated, and localized bilaterally in the subcortical white matter (Fig. 127-1). MRI is superior to CT for distinguishing these abnormalities from confounding illnesses such as PML. Unlike lesions of PML, there is no associated hypointensity in T1-weighted images. The severity of global brain atrophy and signal changes in basal ganglia correlates with cognitive impairment.⁵⁴

There is no indication to perform a brain biopsy in patients with HAND, unless imaging studies suggest the presence of another process. Postmortem examination reveals encephalitis with multinucleated giant cells and microglial nodules, as well as astrocytosis and perivascular mononuclear cell infiltrates. HIV-1 has been found in microglial nodules, perivascular macrophages, and multinucleated giant cells. The latter are the result of the fusion of infected macrophages. The virus can also infect astrocytes.⁵⁵ HIV-1 has also been found in endothelial cells,⁵⁶ and abnormalities of the cerebral microcirculation are characterized by increased cellularity and pleomorphism of endothelial cells, as well as prominent perivascular aggregates of HIV-infected macrophages. These findings are sometimes related to microinfarcts and are postulated to give rise to increased vascular permeability.⁵⁷ Indeed, early enhancement has been detected in the basal ganglia of patients with HAND, suggesting a disruption of the blood-brain barrier (BBB) in these regions.⁵⁸ HIV-1 does not infect glial cells in adults, whereas a limited expression of viral regulatory gene products has been demonstrated in glial cells from children with AIDS.⁵⁹ White matter pallor involving primarily the centrum semiovale is a hallmark of HIV-1 encephalopathy in adults, which may be caused by seepage of macromolecules and edema fluid in the brain parenchyma through the permeable BBB and may cause injury to the myelinated fibers. Finally, HIV-1 does not infect neurons. Therefore, quantitative neuronal loss, decrease in cell size, or dendritic injury found in the cortex of demented patients⁶⁰ may only be secondary to infection of other cells and associated immune activation.⁶¹

Treatment.

Although the pathophysiology of HAND is not entirely elucidated, HIV-1 remains the main triggering factor, and it is therefore reasonable to prevent viral replication in the CNS. However, antiretroviral drugs have variable penetration through the BBB, and evaluation of CNS concentration and pharmacokinetics of these medications is incomplete. On the basis of available pharmacologic data, antiretrovirals can be classified according to their penetration through the BBB. A CSF penetration effectiveness (CPE) score⁶² is given to each medication going from 1 (low) up to 4 (high) (Table 127-2). The CPE score of an ART regimen is calculated by summing the score of each individual antiretroviral. Patients with higher penetration-effectiveness scores tended to have lower HIV RNA levels in the CSF. Drugs with the highest CPE scores include zidovudine, nevirapine, and ritonavir-boosted indinavir.

Since 1996, the availability of protease inhibitors (PIs) and the use of three or more drugs in ART regimens have made a major impact on the incidence of neurologic disease in HIV-1-infected patients.^63–65 This is somewhat surprising; PIs in general have a very low CSF-to-plasma ratio because they are highly protein bound. Nevertheless, the incidence of HAND, which was estimated at 6.49 per 1000 person-years pre-1997, has decreased to 0.66 by 2003 to 2006, an approximately 10-fold decrease.^63,66 Although ART is now the treatment of choice for advanced HIV-1 disease, the optimal drug combination for HAND has not been established. Whether regimens containing several drugs with good penetration through the BBB were advantageous has been investigated. In ART-experienced patients with CD4⁺ T-cell counts of less than 200 cells/µL, those who received multiple CSF penetrating agents had a greater reduction of CSF HIV-1 RNA.^67,68 These results have been confirmed in recent studies, showing that higher CPE scores of long-term ART are associated with better viral suppression in CSF^61–63and decreased risk of cognitive impairment.^69,70 However, other studies have shown that discontinuation of efavirenz was associated with cognitive improvement in scheduled treatment interruption regimens⁷¹ and in otherwise asymptomatic HIV-infected individuals,⁷² suggesting that some antiretrovirals might have neurotoxic properties, which was also demonstrated in vitro.⁷³

The hypothesis that discordant mutations are present in blood and CSF has been confirmed by the analysis of HIV-1 genotypes in paired samples. Mutations of the reverse transcriptase and protease genes of CSF strains, but not in the corresponding blood isolates, were detected in 10 of 23 subjects (43%). Interestingly, mutations conferring resistance to low or high CSF-penetrating agents occurred at similar frequencies.⁷⁴ Reverse-transcriptase mutation patterns differed in 14 of 21 (67%) paired samples from plasma and CSF in one study, and HIV RNA responses and reverse-transcriptase genotypes were discordant between CSF and plasma in some subjects.^75,76 In a study of HIV-infected children starting ART, 8 of 11 had identical resistance patterns in both CSF and plasma at baseline, whereas at week 48 of treatment, only 1 of 9 children had similar genotypes, suggesting discordant viral evolution in these two compartments.⁷⁷ CSF viral escape was reported recently in 10 patients with progressive neurologic dysfunction despite well controlled plasma viral load and good immunologic response on ART. In general, these patients had a higher CSF viral load compared with plasma. The patients also had CSF pleocytosis and elevated protein concentration, MRI abnormalities, and unique CSF resistance mutations compared with plasma and improved clinically after optimization of ART.⁷⁸ Long-term follow-up of a patient who developed HAND and elevated CSF viral load despite low plasma viral load and high CD4 counts confirmed the importance of concomitant genotype testing in CSF and plasma. Subsequent change of ART targeting CSF strains was associated with suppression of CSF HIV replication, as well as a marked clinical and radiologic improvement.⁷⁸ Failure to recognize the occurrence of different genotypic resistance patterns in plasma and CSF may lead to uncontrolled viral replication in the CNS and worsening neurologic symptoms despite evidence of viral clearance in the periphery. These resistant isolates may, in turn, spill back into the circulation and contribute to the progression of HIV-1 disease. Therefore, genotypic analysis of CSF HIV-1 strains is indicated in patients with cognitive dysfunction who have good virologic response to ART in the plasma but persistence of an elevated HIV-1 CSF viral load.

In addition to antiretroviral treatment, other experimental compounds have been tested for the treatment of HAND, but without evidence of clear benefit so far. These include a calcium channel blocker (nimodipine); antioxidants (vitamin E, thioctic acid, CPI-1189); a tumor necrosis factor-α antagonist (pentoxifylline); a noncompetitive N-methyl-d-aspartate inhibitor (memantine); peptide T; selegiline; and, more recently, minocycline.⁷⁹