The AIDS Dementia Complex (ADC) is one of the most common and clinically important CNS complications of late HIV-1 infection. It is a source of great morbidity and, when severe, is associated with limited survival. While its pathogenesis remains enigmatic in several important aspects, ADC is generally thought to be caused by HIV-1 itself, rather than to another opportunistic infection.(1-3) This chapter offers a general review of ADC.

ADC was first identified early in the AIDS epidemic as a common and novel CNS syndrome.(4,5) The three components of the term, AIDS dementia complex embody central features of the condition. AIDS emphasizes its morbidity and poor prognosis, particularly when its severity is at stage 2 or greater (see Table 1), a severity comparable to other clinical AIDS-defining complications of HIV-1 infection. Dementia designates the acquired and persistent cognitive decline with preserved alertness that usually dominates the clinical presentation and determines its principal disability. Complex emphasizes that this disease not only impairs the intellect, but also concomitantly alters motor performance and, at times, behavior. This involvement of the nervous system beyond cognition is evidence of a wider involvement of the CNS than occurs in some other types of dementia such as Alzheimer's disease. Additionally, myelopathy may be an important, indeed predominating, aspect of ADC, and organic psychosis may also be a feature in a subset of patients (see Rheumatologic and Musculoskeletal Manifestations of HIV). These manifestations are therefore also encompassed within this term. By contrast, neither neuropathy nor functional psychiatric disturbance are included in ADC.

An empirically derived, five-step ADC Staging System (Table 1)(6,7) is applied once the diagnosis of ADC is made. The designation is based on the degree of functional incapacity in cognitive and motor activities of work and daily living and ranges from stage 0.5, which describes patients in whom there are neurologic symptoms or signs without functional impairment (subclinical disease) and in whom mild impairment may possibly relate to another condition that cannot be clearly distinguished (equivocal ADC), to stages 4, which indicates severe/end-stage dysfunction.

Committees sponsored by the World Health Organization (WHO)(8) and American Academy of Neurology (AAN)(9) have proposed alternative terminology for ADC. Disadvantages to their use include the awkwardness of the general term chosen for the full spectrum of severity - HIV-1-associated cognitive/motor complex, omission of a term equivalent to stage 0.5 ADC, and designation of stage 1 as HIV-1-associated minor cognitive/motor disorder. Those suffering stage 1 ADC usually do not consider their condition to be "minor." Additionally, abbreviations may be difficult, because HAM is already used for HTLV-I-associated myelopathy. I prefer the simpler term "ADC," designating a subset as suffering a myelopathic variant when their signs and symptoms indicate isolated spinal cord dysfunction. Fortunately, the WHO/AAN terminology can be "translated" to the ADC Staging on which it is founded.(10)

ADC develops principally in the context of late HIV-1 infection and associated severe immunosuppression. Its prevalence accordingly varies depending on the characteristics of the population sampled. Prospective studies have provided some clarification of the epidemiology and natural history of ADC (stage 2 or higher), at least with respect to its more severe forms, in the era before combination highly active antiretroviral therapy (HAART). In the Multicentered AIDS Cohort Study (MACS), which followed a selected group of gay men, the incidence rate of ADC over a 5-year period was 7.34 cases per 100 person years for subjects with CD4+ counts < 100, 3.04 cases in those with counts of 101 to 200, 1.31 for counts of 201 to 350, 1.75 for counts of 351 to 500, and 0.46 for counts > 500.(11) Diagnostically, these results show that more severe ADC is principally a disease found in advanced HIV-1 infection, although uncommonly it may develop in those with relatively preserved helper lymphocyte counts. Pathogenetically, these data suggest that severe immunosuppression has a strong "permissive" effect on the development of ADC, but alone is neither sufficient (since many severely immunosuppressed patients do not develop the disorder) nor absolutely necessary for ADC to manifest.

Data from the Community Programs for Clinical Research on AIDS (CPCRA) following AIDS patients in a series of treatment protocols showed this same association of incidence with advanced stage of HIV disease and additionally indicated that the development of stage 2 or greater ADC was associated with limited survival.(12) The 6-month cumulative mortality of 97 ADC patients among an overall group of 3,382 HIV-1-infected subjects followed in this program was 67%, which was nearly three times the mortality rate for Pneumocystis carinii pneumonia (PCP) and closer to the 6-month mortality rates of other neurologic diseases (nearly 85% for PML, 70% for primary CNS lymphoma, and 51% for cerebral toxoplasmosis). This poor prognosis for survival likely relates in part to the late stage of HIV-1 infection in which ADC develops, the concomitant susceptibility to other lethal complications of immunosuppression, limited effectiveness of antiviral treatments available at the time of the study, and the vulnerability of neurologic debility. The tendency for relatives and care givers to "give up" on such patients in the face of severe neurologic impairment may also influence this short survival.

Evidence supports a central role for HIV-1 in causing ADC, particularly in the subset of patients with more severe brain dysfunction (see reviews[3,13-15]). Numerous studies have documented HIV-1 productive infection of macrophages and related microglia and nonproductive infection of a wider variety of cells, including particularly astrocytes.(16-19) Because of the seeming discrepancy between the severity of clinical deficits and the pathologic and virologic features, indirect pathogenetic processes relating infection and brain injury have been proposed (see reviews[13-15,20,21]). Investigators have hypothesized that macrophage and microglial infection drives a chain of pathologic processes that eventuate in neuronal dysfunction. Some of these effects are exerted directly on neurons while others involve intermediary cells to transduce and amplify these signals to eventuate in neurotoxicity. Particularly important in these sequences are activation of cytokine circuits, largely in macrophages and perhaps also astrocytes. Pathogenic cytokine activation and resulting neuropathologic sequelae can be considered as immunopathologic reactions that link infection to neurotoxicity. The importance of understanding the individual mechanisms involved in these activation and neurotoxic reactions relates to the possibility that they may provide therapeutic targets.

Although the severity and relative prominence of some symptoms and signs compared to others may vary among individual patients, the general character of ADC involves three functional categories: cognition, motor performance, and behavior.(5) Table 2 provides an outline of some of the early and late manifestations. Of the three categories, cognitive and motor dysfunction are the most helpful in characterizing patients and in defining diagnosis; it is for this reason that they provide the basis of ADC Staging, which omits behavioral criteria. When approaching diagnosis, it is useful to separately consider milder and more severe affliction.

Stage 0.5 and 1 ADC

Cognitive impairment usually underlies patients' earliest symptoms. Mildly afflicted patients most often have difficulty attending to more complex tasks at work or at home. They need to make lists, sometimes very detailed, of the day's activities. They lose track of actions (e.g., leave the water boiling, get up to go to another room and then forget why they did so) or of conversations in mid-sentence ("What was I saying?"). Processing unrelated or complex thoughts becomes slower and less facile. While similar lapses can trouble many normal people especially in the face of fatigue or generalized illness, lapses in ADC patients intrude on daily function to a greater degree. Multi-staged tasks become difficult; e.g., the waiter can no longer keep verbal orders straight when he arrives at the kitchen or the avid reader needs to reread paragraphs or pages. When such dysfunction is mild, it may be difficult to substantiate the basis for these complaints by bedside examination, and it is important to apply tests that are sensitive to these abnormalities, including particularly tests requiring concentration, change of sets, and timed performance. Because it was constructed for other conditions, the standard Mini-Mental Status(22) may not be sufficiently sensitive at this point; however, when ADC patients do perform abnormally, it is usually on reversals (reversing a five-letter word like "world," or subtracting from 100 by 7's), complex sequential tasks (placing the right thumb on the left ear and sticking out the tongue), or remembering three objects. Although motor symptoms are far less common during this early phase, individuals relying on rapid or fine coordination may note a change. For example, the guitarist may no longer be able to keep up with a difficult piece or the athlete may be slowed to below a competitive level. An inquiring history may discover a change in handwriting or, less commonly, clumsiness in tying shoes or buttoning a shirt. Moreover, even in those without overt symptoms, motor signs may be detected on examination, including slowing of attempts at rapid opposition of the thumb and forefinger, rotation of the wrist, or tapping of the toe. While the gait may be generally steady, it is often slow, and rapid turns may be interrupted by an extra step or performed hesitantly. Reflexes are also often abnormal. The deep tendon stretch reflexes, including importantly the jaw jerk, are frequently hyperactive, although the ankle jerks may be relatively less active when there is concomitant polyneuropathy. Babinski signs may be present and other "pathologic" release signs may also be detected; of these, the snout response is relatively frequent and particularly helpful when present in young patients. The time course and onset of milder ADC is variable. It may begin insidiously or abruptly and progress more rapidly to a higher stage, or it may continue to evolve slowly or even remain static for some period. In patients with mild ADC, missed diagnosis usually results from overlooking the important aspects of the history. Functional difficulties may erroneously be considered as a "normal" part of systemic illness by patients and their caregivers despite the fact that most AIDS patients without neurologic conditions perform quite normally, even in the late stage of HIV-1 infection. Particularly important is the distinction of ADC from clinical depression, which can produce similar complaints but carries distinct therapeutic implications (see Chapter 5.15). Hypochondriasis and anxiety in those understandably worried about body function may also lead to similar complaints.

Stage 2-4 ADC

In patients with more advanced disease, the principal exercise is in being certain that abnormal function relates to ADC rather than to another CNS disease. Cognitive function in these subjects is clearly abnormal and obviously impairs functional status, although in patients with stage 2 or even 3 ADC who can maintain the civilities of casual conversation and personal interaction, it too can be missed by a cursory history and examination. With more careful questioning of both the patient and associates, however, it is clear that stage 2 patients are too slow or forgetful to work, manage the household or, importantly, even maintain their own medications. They may get lost walking or driving and cannot be relied on to prepare meals, much less to balance the checkbook. On examination, a broader array of cognitive "domains" are afflicted. The bedside Mini-Mental Status is now often abnormal. The components mentioned earlier, assessments of attention and concentration, are frequently impaired, and patients have trouble recalling three objects after 5 or 10 minutes. Timed activities are further slowed and attempts to draw complex objects may produce only simplified representations. With increased severity, there is frank disorientation to time and place. Motor abnormalities may also become more clearly symptomatic and obvious to the patient's family and associates. Walking may be sufficiently unsteady to require a cane or someone alongside to prevent falling. Hyper-reflexia and pathologic reflexes are now almost always present, and gait instability and slowness is more clearly evident not just on turning, but even on the straightaway. With further progression, ambulation constantly requires someone to balance and support the patient or is entirely precluded (stage 3 or 4). Thinking and speaking also becomes slower and the content more impoverished. Concomitant behavioral changes may become more evident. Patients appear duller and less vivacious. If left alone, they may sit still without spontaneously offering conversation, but only answering briefly in response to questions. This poverty of output and apathy may be mistaken for depression, but in most of these patients, dysphoria is absent, and disinterest and lack of initiative are the predominating aspects of behavior without sadness. A striking variant that manifests in a small minority of patients includes agitation with features of mania. These patients usually exhibit a background of confusion that remains even as the hyperactivity is controlled by medication.

Vacuolar Myelopathy

The myelopathic variant of ADC has a distinct pathologic substrate, known as vacuolar myelopathy.(23) This condition can often be distinguished clinically when the patient's gait dysfunction is disproportionally affected in comparison to the intellect. Some of these patients may become wheelchair bound with normal or near normal cognition. In others, the myelopathy is combined with cognitive difficulty. While the lower extremities are more severely affected than the arms, there is no distinct segmental level of spinal cord dysfunction. Rather, there is a gradual caudal increase in abnormality: knee tendon reflexes are more active than those in the arms, and gait or toe tapping are not performed as well as hand coordination tests or rapid finger movements. Mild sensory loss is common, and is usually worse distally in the feet with impaired vibration and position sense most common; sensation is usually normal over the trunk and in the upper extremities. Because myelopathy is frequently combined with neuropathy, the cause of sensory loss in individual patients may not always be clear, and one often relies on the ankle tendon jerks to indicate the relative contribution of myelopathy (increased ankle jerks) or neuropathy (decreased ankle jerks) in the presence of increased patellar stretch reflexes. Clinical presentation usually includes an ataxic and, sometimes, spastic gait.

Pediatric AIDS Patients

While this chapter deals with adult ADC, a similar syndrome complicates pediatric AIDS. This disorder in children may be proportionally more important in the sense that it both occurs more commonly than in adults and results in even greater morbidity.(24) It can manifest as a progressive encephalopathy with reversal of pediatric developmental milestones and subsequent microcephaly or can be static with developmental delay. Clinical presentation is with combined cognitive and motor dysfunction. Its etiology and pathogenesis are likely similar to the adult condition, although altered somewhat by involvement of the developing brain. Treatment also follows a similar approach.

Diagnosis of stage 2-4 ADC is both an inclusionary and exclusionary exercise. It is important to understand that the distinct features of the syndrome described earlier in the chapter when occurring in the setting of late HIV-1 infection allow a positive diagnosis. The combination of the characteristic cognitive dysfunction and symmetric motor abnormalities usually readily allows tentative diagnosis on the basis of the history and examination alone. There are a few other conditions that mimic the typical and uncomplicated presentation; some patients with primary CNS lymphoma located deep in the frontal white matter near the lateral ventricles, particularly when bilateral, may manifest a similar progressive dementia accompanied by motor slowing, apathy, impoverished speech output, and minimal lateralization. Hydrocephalus can also produce a similar picture. In contrast, toxic and metabolic encephalopathies usually lead to a reduced level of arousal that is parallel to the cognitive deficiency rather than altering the latter in the face of full wakefulness as in ADC. These disorders may coexist with and exacerbate ADC, however, and the combination can confuse diagnosis. ADC patients appear to be more sensitive to the side effects of neuroleptic agents, and subclinical ADC may explain the sensitivity of AIDS patients more generally to the extrapyramidal side effects of these drugs.(25)

Cytomegalovirus (CMV) encephalitis is one of the diagnostically more difficult conditions complicating late HIV-1 infections and may be difficult to distinguish from ADC.(26,27) While its features may vary, however, CMV encephalitis is more often accompanied by blunted arousal than ADC and probably more frequently causes seizures, particularly therapeutically intractable seizures, than does ADC. Minor focal findings, including cranial nerve palsies and ataxia, may also accompany CMV encephalitis. Hyponatremia has also been noted more frequently in CMV encephalitis. An uncommon form of cerebral toxoplasmosis, the so-called encephalitic form,(28-30) may be similarly confused with ADC as well as with CMV encephalitis. This condition is caused by the acute development of multiple small toxoplasma abscesses throughout the brain, and may present with subacute global encephalopathy with few or no focal abnormalities. Once again, however, consciousness is usually depressed. Neurosyphilis is commonly considered in the differential diagnosis, although how often it clinically mimics ADC is uncertain.(31) Wernicke's encephalopathy may also occur in patients with AIDS as a result of nutritional deficiency, although altered consciousness usually occurs during the acute phase along with ocular and other distinct abnormalities.(32)

The differential diagnosis of ADC-related myelopathy is also relatively narrow in persons with HIV-1 infection. HTLV-I and, as more recently reported, HTLV-II can cause a clinically similar myelopathy that is symmetric and without segmental focality. While the epidemiologies of HIV-1 and these other retroviruses overlap, particularly HTLV-I and HIV-1 in the Caribbean, southern United States, and parts of Africa, there is no evidence that one of these infections exacerbates or accelerates the other. Viral serology, CSF profile, the presence of immunosuppression, and other ancillary findings usually allows distinction. In contrast to these conditions, other myelopathies in AIDS patients have a more focal and segmental presentation, including the myelopathy of varicella-zoster virus (VZV), which frequently, although not always, complicates temporally proximate herpes zoster with its characteristic rash and develops at or near the spinal cord segment corresponding to the involved dermatome.(33) The rare cases of spinal cord toxoplasmosis and lymphoma, including both primary CNS lymphoma (intramedullary focus in the spinal cord) and metastatic systemic lymphoma (located in the epidural or meningeal spaces with compression or invasion of the cord), also cause segmental myelopathies.

Clinical laboratory studies reveal the characteristic abnormalities noted commonly in ADC and, perhaps even more importantly, detect other conditions in the differential diagnosis. The most helpful tests in this setting are neuroimaging, cerebrospinal fluid (CSF) examination, and formal neuropsychological testing.

Neuroimaging

Neuroimaging is usually an essential component of the evaluation of AIDS patients with CNS dysfunction, including those with suspected ADC. Principally, it is used to detect evidence of other diagnoses such as the mass lesions of primary CNS lymphoma or the ependymal signal changes of CMV encephalitis. It can also detect abnormalities associated with ADC. Anatomical imaging, including both computed tomographic (CT) scanning and magnetic resonance imaging (MRI), usually shows evidence of cerebral atrophy with widened cortical sulci and enlarged ventricles in ADC patients, particularly in those individuals with stages 2-4.(5,34,35) Basal ganglia are also reduced in volume.(36) Additionally, MRI in some patients detects T2-weighted abnormalities in the hemispheric white matter and, less commonly, the basal ganglia or thalamus. These signal changes can be patchy or "fluffy" in appearance; the white matter may have a more homogeneous "ground-glass" appearance. Although these signal changes are sometimes referred to as "HIV encephalopathy," neither these imaging abnormalities nor the cerebral atrophy are diagnostic of ADC. Indeed, one needs to be cautious in equating neuroimaging findings with the clinical syndrome. In children with AIDS, mineralization of the basal ganglia is often prominent.(37) Despite its sometimes striking pathology, vacuolar myelopathy is usually not detected by spinal MRI.

CSF Analysis

Routine CSF examination is most useful in differential diagnosis rather than in directly supporting ADC since findings in these patients are unspecific.(5,38,39) The protein may be normal but is usually mildly elevated, while the white blood cell counts, likewise, may be normal or show mild mononuclear pleocytosis; these elevations in cell count and protein do not distinguish these patients from neurologically normal, HIV-1 infected individuals.(40) Similarly, culture for HIV-1 is not useful because of the high background of culture positivity in asymptomatic patients. HIV-1 p24 can also be detected in the CSF of severely affected patients, but in most of these cases, the diagnosis is readily made on clinical grounds.(41) More recently, there has been interest in the issue of whether quantitative assay of HIV-1 RNA ("viral load") in the CSF might be useful diagnostically.(42-44) Unfortunately, studies to date suggest that measurements of the CSF viral load are not sufficiently specific to use as a diagnostic marker.(45) Thus, whereas CSF viral loads are often elevated in more severe ADC (stages 2-4) and, indeed, may be very high in some (> 105 copies per milliliter), they may also be high in asymptomatic patients. While one can show group differences between those with and without ADC, the overlap is such that this test is not diagnostically useful in the individual case. Treatment with HAART results in marked reduction in CSF viral load just as in the plasma. It is not yet certain whether monitoring this reduction is useful in assessing ADC therapy. Earlier studies assessing the utility of measuring the concentrations of markers of immune activation in the CSF suggest that such testing might be useful in some clinical situations. These markers include beta2-microglobulin and 13:05 8/15/2002 neopterin, which can be readily measured in most institutions, and quinolinic acid, which requires more specialized assay.(46-49) Although elevation of these markers is nonspecific and occurs in CNS opportunistic infections and primary CNS lymphoma, in the absence of these other conditions, increased concentrations support a diagnosis of ADC. For practical purposes, these determinations are most useful in differentiating ADC from psychiatric disease or coincident neurodegenerative diseases.

Neuropsychological Testing

Formal neuropsychological testing may be viewed as a quantitative neurologic examination, and for this reason, it has provided the principal endpoint measure for clinical trials of ADC treatment.(50) While performances on such tests are not diagnostically specific and individual test performance is variable and can be affected by a variety of factors (age, education, prior drug or alcohol use, head injury, etc.), if compared to appropriate norms and carefully interpreted in the particular clinical context, formal examination can help to determine whether symptoms truly reflect abnormal neurologic function, and whether the character of such dysfunction conforms to that of ADC.

Strategies to treat ADC have followed the guidelines suggested by concepts of pathogenesis.(13) Thus, if brain injury results from the linking of brain infection to endogenous cytokine-linked neuropathic processes, then treatment efforts might attempt to interrupt these processes at various vulnerable points. The most effective approach involves antiretroviral therapy, because HIV infection seems to be the prime-mover of ADC pathogenesis.

Unfortunately, the optimum regimen for ADC is not established, largely because this issue has not yet been addressed with respect to contemporary combination therapy. In fact, evidence of antiviral efficacy derives principally from the experience with zidovudine monotherapy, which has been shown in a number of adult and pediatric studies to prevent and reverse clinically symptomatic ADC and also to reduce the incidence of brain infection.(51-57) One therefore needs to extrapolate from these data, which demonstrate that zidovudine monotherapy is helpful, to the suggestion that combination therapy likely would be even more effective, as it is in systemic HIV-1 infection. This conclusion seems reasonable even if not directly proven or supported by controlled observations.(58)

An additional issue is the importance and extent of antiviral drug penetration into the brain, across the blood-brain barrier, about which uncertainty remains. Although it seems reasonable that antiviral drugs should have to reach the site of brain infection to be effective, there are reports of neurologic improvement of patients treated with protease inhibitors that penetrate the blood-brain barrier poorly. Given this uncertainty as well as the limited information on penetration of some of the antiviral drugs, I recommend the following empiric approach:

• ADC patients should be treated with aggressive antiretroviral therapy.

• Combinations of three, four, or more drugs should usually be used.

• These drugs should be chosen on the basis of whether or not they are likely to be effective in suppressing systemic infection in the individual patient (particularly that the patient's predominating viral quasispecies is unlikely to be resistant to the component drugs), with consideration of how they will be practically tolerated by the patient.

• To include, if possible, two drugs with appreciable penetration of the blood-brain barrier.

Among the nucleoside reverse transcriptase inhibitors (RTI), zidovudine, stavudine, and abacavir likely have the best penetration, and lamivudine to a lesser extent.(59,60) Nevirapine, a non-nucleoside RTI, also has favorable penetration. Among the protease inhibitors, only indinavir has been reported to appreciably penetrate into CSF.(61) More precise definition of the penetration of these and other drugs is likely to be available in the near future.

Additional approaches to treatment may be classified as adjuvant therapies because they target processes beyond the infection.(62) While these are now under investigation, there is no clear evidence that they are effective. By contrast, judicious use of medicines to control some symptoms and of supportive measures to soften some of the suffering of these patients can be invaluable(63) (see also Chapter 5.15).