Neuropsychiatric Manifestations of HIV-1 Infection and AIDS

Robert A. Stern, Diana O. Perkins, and Dwight L. Evans

By 1993, 172,000 individuals in the United States had died from acquired immunodeficiency syndrome (AIDS) and another 1,000,000 were estimated to be infected by human immunodeficiency virus type 1 (HIV-1), the retrovirus that causes AIDS. Currently, most individuals infected with HIV-1 develop AIDS after an estimated median latency period of 10 years. After developing AIDS, the average length of survival is approximately 11 months, and the maximum length of survival has been reported to be 9 years.

Until 1993, the Centers for Disease Control (CDC) defined AIDS as the development of at least one of 23 specific infectious, neoplastic, general systemic, or nervous system diseases that are rare in the absence of immunodeficiency. In 1993, the CDC case definition was modified to include individuals with CD4+ T-lymphocyte cell counts (CD4+) less than 200/mL or a CD4+ percentage of total lymphocytes of less than 14 (14). This expanded surveillance case definition for AIDS and classification system for HIV-1 infection also included additional clinical conditions seen in women, such as invasive cervical cancer.

Since the earliest descriptions of AIDS, neuropsychiatric and neurologic signs and symptoms have been reported in a subgroup of persons with this disease. The neuropsychiatric manifestations contained in these early reports include progressive dementia, depression with pronounced apathy and psychomotor slowing, and other psychiatric presentations including manic symptoms and atypical psychosis. Initially, these AIDS-related mental disturbances were attributed to psychological reactions to a systemic illness, the effects of psychosocial stressors associated with the disease, or the consequences of opportunistic infections or tumors within the central nervous system (CNS).

It is now estimated that 40–70% of patients with AIDS develop clinical neurologic abnormalities. Upon autopsy, 75% of AIDS patients exhibit neuropathologic changes, with 30% having multiple CNS lesions (37). In as many as 20% of HIV-1-infected individuals, neurologic or neuropsychiatric symptoms may be the presenting features, prior to other medical symptoms of AIDS (7). The most severe form of HIV-1-associated neuropsychiatric disorders, HIV-1 encephalopathy or HIV-1-associated dementia complex, affects at least 7% of persons with AIDS and is one of the most common causes of dementia in individuals age 20–59 in the United States (39).

CNS involvement in AIDS results from a variety of etiologies, including (a) the direct or primary effects of HIV-1 on nervous tissue, (b) the consequences of secondary viral and nonviral opportunistic infections, tumors, and cerebrovascular disease, and (c) the complications of systemic therapies for AIDS and associated disorders (Table 1).

Early evidence for the neuropathogenicity of HIV-1 included (a) the presence of HIV-1 in the cerebrospinal fluid (CSF) of patients with AIDS, (b) abnormal neuroimaging findings in neurologically impaired AIDS patients, (c) the high frequency of peripheral neuropathies in AIDS patients, and (d) the presence of HIV-1 in tissue obtained by brain biopsy. It is now known that HIV-1 penetrates the blood–brain barrier early in the course of infection and is replicated in brain tissue using mono- and multinucleated macrophages as hosts (44). Once in the brain, the virus can have neurotoxic effects via direct or indirect mechanisms. For example, the HIV-1 envelope protein, gp120, elevates intracellular free calcium, leading to neuronal damage (17). In addition, patients with HIV-1 encephalitis have significantly higher levels of the excitatory amino acid and endogenous N-methyl-D-aspartate receptor agonist, quinolinic acid, in CSF and in brain tissue than do controls (1). Although these levels are not correlated with encephalitis severity, quinolinic acid may be an etiologic agent in HIV-1-related neuropathology. Prion Diseases details the current knowledge of the neurotoxic process of HIV-1.

In addition to the direct or primary effects of HIV-1 on the CNS, there are numerous and significant secondary effects of HIV-1 (Table 2). One of the most important with regard to CNS effects is the papovavirus, JC virus, because it leads to progressive multifocal leukoencephalopathy (PML), a severe demyelinating disease for which there is no proven treatment. Although PML is an infrequent complication of immunosuppressive therapy, it is highly prevalent among AIDS patients, with an incidence of at least 4–7% in this group (6). Among the nonviral secondary infections affecting the CNS, toxoplasmosis is by far the most significant and one of the most treatable, with up to 50% of HIV-1-infected patients experiencing this disorder (48).

The neuropathological changes associated with HIV-1 infection are the result of both the primary effects of the virus on the CNS and the secondary effects of opportunistic infections, tumors, and cerebrovascular disease. The lesions are both focal and diffuse, with much variability across patients and within a single patient over time.

Primary neuropathological effects of the virus include (a) the presence of multinucleated giant cells, (b) the infiltration of both subcortical white and gray matter by macrophages, (c) myelin pallor, and (d) gliosis (12). HIV-1-specific neuropathological appearances that predominate in subcortical areas include: (a) HIV-1 encephalitis, characterized by perivascular, focal lesions with damage to axons and myelin, large, reactive astrocytes, focal vacuolization, and mono- and multinucleated macrophages; (b) HIV-1 leukoencephalopathy, a triad of myelin loss, infiltration by mono- and multinucleated macrophages/microglia, and reactive astrogliosis; and (c) vacuolar leukoencephalopathy, involving vacuolar myelin swellings in deep white matter. However, the neuropathology of HIV-1 is not restricted to subcortical regions; cortical gray matter changes have also been observed, including cell loss, synaptic loss, and gliosis.

Various neuroimaging techniques have been used for in vivo observations of HIV-1 neuropathology. Although computerized tomography (CT) has been employed to assess the structural lesions associated with HIV-1, magnetic resonance imaging (MRI) has proven to be the most revealing technique to date in identifying HIV-1 structural neuropathology. Common MRI findings in HIV-1infected patients have involved the presence of cerebral atrophy, focal high signal intensities in subcortical white and gray matter (observed primarily using T2-weighted images), confluent high-signal areas, and solitary high-signal areas (15, 42). Abnormal MRI findings are mostly observed in patients with HIV-1 encephalopathy (Fig. 1) or in patients with focal neurologic lesions resulting from opportunistic infections such as toxoplasmosis (Fig. 2) or lymphoma. However, standard MRI interpretation is not as useful in depicting abnormalities in neurologically asymptomatic HIV-1-infected patients. If MRI abnormalities are detected in these patients, they tend to involve focal white-matter lesions only (76). More recently, quantitative morphometric analyses of MRIs have suggested that symptomatic HIV-1-infected individuals who do not have any neurologic signs do indeed exhibit cortical and central atrophy (41).

Functional neuroimaging techniques have also proven to be sensitive to the neuropathology associated with both asymptomatic and symptomatic HIV-1-infected patients. In fact, functional assessments of cerebral perfusion [using single photon emission computerized tomography (SPECT)], metabolism [using positron emission tomography (PET)], and biochemical abnormalities [using magnetic resonance spectroscopy (MRS)], appear to be more sensitive than structural techniques such as CT and MRI. For example, a SPECT study of HIV-1-infected men without AIDS or HIV-1-associated dementia revealed abnormal perfusion in approximately two-thirds of the asymptomatic and three-quarters of the symptomatic patients (88). The high prevalence of perfusion abnormalities in asymptomatic subjects have been reported by other investigators as well (95), providing further evidence that cerebral abnormalities can be detected early in the course of HIV-1 disease. In patients with AIDS and HIV-1associated dementia, multifocal perfusion defects have been found (54). However, the relationship between SPECT findings and severity of neurocognitive impairment in either demented or nondemented patients is unclear.

Although not as widely employed, PET has been useful in detecting functional cerebral abnormalities throughout the course of HIV-1. For example, focal metabolic disturbance of the basal ganglia, thalamus, and temporal lobes has been reported (96). In this study, hypermetabolism of the basal ganglia and thalamus was observed in nondemented patients, with additional hypometabolism of the temporal lobes evidenced in demented patients. MRS has been the focus of several recent investigations (e.g., see ref. 40) and appears to be both sensitive and specific to the neurological and neuropsychiatric changes seen in both asymptomatic and symptomatic HIV-1-infected patients.

Early case reports suggested that patients with AIDS experience a variety of neuropsychiatric symptoms including mental slowing, lethargy, forgetfulness, apathy, and social withdrawal. The term, "AIDS dementia complex" (ADC) was coined, referring to the changes in cognition, motor function, and behavior associated with AIDS (61, 77, 78). However, the accuracy of this term has been heavily discussed and refined. Given that HIV-1 is the etiological agent responsible for the disorder, AIDS may not be an appropriate term. "Dementia" may also not be an adequate descriptor because early presentations of the disorder did not necessarily involve decline in intellectual functioning. Lastly, not all patients exhibit the full "complex" of cognitive, motor, and behavioral dysfunction, and may instead have only one or two of these disturbances.

In 1991, a working group of the American Academy of Neurology AIDS Task Force (38) developed consensus nomenclature and case definitions for HIV-1-associated neurological conditions to be used for research purposes. A major outcome of this working group was the replacement of the term "AIDS dementia complex" with the more descriptive term "HIV-1-associated cognitive/motor complex." The latter term was then divided into two categories: (i) a more severe form, "HIV-1-associated dementia complex" or "HIV-1-associated myelopathy" (the latter term reserved for patients whose myelopathic dysfunction is worse than their neurocognitive impairment), and (ii) a less severe form, "HIV-1-associated minor cognitive/motor disorder." The classification "HIV-1-associated dementia complex" can be further delineated based on the presence of behavioral and motor abnormalities. Therefore, a patient with cognitive and motor impairment but no behavioral dysfunction would be considered to have "HIV-1-associated dementia complex (motor)." A patient with cognitive and behavioral dysfunction but no motor involvement would have "HIV-1-associated dementia complex (behavior)."

The major difference between "HIV-1-associated dementia complex" and "HIV-1-associated minor cognitive/ motor disorder" is the severity of impairment in activities of daily living. That is, by definition, dementia must have cognitive impairment severe enough to interfere with occupational or social functioning. In "HIV-1-associated minor cognitive/motor disorder," activities of daily living are generally intact with the possible exception of mild difficulties in the most demanding activities.

Differences in terminology aside, HIV-1-associated dementia arises in one-third of the AIDS patient population and is the most common neurological or neuropsychiatric disorder associated with HIV-1 infection (16). Although the dementia usually develops in patients who already show other AIDS manifestations, up to 25% of HIV-1-infected patients may develop it as their initial, AIDS-defining illness (62). An additional 15% develop dementia concurrent with other features of AIDS. HIV-1-associated dementia is a rapidly progressing process, with the time of onset to severe dementia averaging only several months (11).

The initial features of HIV-1-associated dementia include an overall slowing in cognition (i.e., bradyphrenia) and movement (i.e., bradykinesia) as well as difficulties in motor dexterity and coordination, forgetfulness, poor concentration, and marked apathy (Table 3). Although dysphoric mood is not a common feature, the pronounced slowing and apathy may appear as if the patient is depressed. Furthermore, assessing other aspects of depression (e.g., weight loss, cognitive disturbance, insomnia) is difficult for patients with this disorder due to shared symptomatology that may be indistinguishable from the psychiatric symptoms.

Later in the course of HIV-1-associated dementia, the patient may exhibit myoclonus, bowel and bladder incontinence, and, eventually, mutism and a vegetative state. Once these advanced features are present, death is typically imminent.

Because the above initial symptoms are similar to those seen in other patient groups with subcortical impairment (e.g., Parkinson's disease, progressive supranuclear palsy, multiple sclerosis) and because of the neuroimaging findings of subcortical neuropathology, HIV-1-associated dementia was originally described as a subcortical dementia. However, in light of the more recent findings of cortical atrophy and higher cortical function deficits in AIDS patients, this characterization may not fully describe the spectrum of neuropsychiatric deficits associated with HIV-1 infection and AIDS.

Although it is clear that neuropsychiatric symptoms and neuropathological changes are common in AIDS, the presence and severity of the neurocognitive changes during the early asymptomatic phase of HIV-1 infection remain unclear. In an early study addressing this issue, Grant et al. (28) reported that 44% of asymptomatic HIV-1-infected homosexual men exhibited abnormal neuropsychological performance, compared to only 9% of HIV-1-seronegative controls. This report received wide attention in the lay press, which, in turn, led to policy change in the military such that personnel infected with HIV-1 were removed from sensitive or stressful roles, such as aviation. Since that original report, scores of divergent studies have been published that have either supported the presence of early neurocognitive dysfunction or have argued against it (29). The most common deficits reported by the former studies include fine motor slowing, decreased dexterity, slowing of information processing, and memory difficulties mainly involving diminished free recall.

Between-study differences in sample composition may partially account for the discrepant findings described above. Subjects in these studies differed in terms of several factors that may have a major impact on neuropsychological test performance, such as past head trauma, alcohol and other substance abuse, premorbid psychiatric illness, learning or developmental disability, and previous CNS disorders. However, the actual impact of these possible confounds has also been the subject of debate. For example, Pakesch et al. (64) found that long-term drug abuse may account for the cognitive/motor deficits observed in HIV-1-infected subjects. Conversely, Bornstein et al. (8) found that alcohol use, despite its negative effect on neuropsychological performance, does not account for performance differences between HIV-1-seronegative and HIV-1-seropositive homosexual men (8).

Researchers have also focused on the impact of head injury on the neurocognitive status of HIV-1-infected asymptomatic individuals. In a study of homosexual and bisexual men, Bornstein et al. (9) found no difference in neuropsychological performance between asymptomatic HIV-1-infected subjects with a history of minor head injury and matched controls (HIV-1-seropositive and HIV-1-seronegative) without any history of head injury. In contrast, in a study of intravenous drug users, Marder et al. (51) found that previous head injury was associated with significantly worse neuropsychological performance in the HIV-1-infected group.

In a report addressing the effect of the above variables on the cognitive and motor functioning of HIV-1-infected patients, Wilkins et al. (98) found that confound severity was positively correlated with the degree of neurocognitive dysfunction. Although the impact of confounds on performance was the most striking for the asymptomatic group, this finding was also true for symptomatic individuals.

A variety of possible coexisting factors may account for the neurobehavioral impairment found in asymptomatic HIV-1-infected individuals. Beach et al. (5) have suggested that diminished vitamin B12 levels in some HIV-1-infected individuals may serve as one such potential cofactor because B12 deficiency has been associated with a variety of psychiatric, neurological, and neurocognitive dysfunction. However, the largest study of B12–neuropsychological/neurological interactions to date in HIV-1 spectrum disease found no relationship between B12 levels and neuropsychological, psychiatric, neurophysiological, or neurological dysfunction at any stage of HIV-1 infection (85).

Depression has been proposed as another possible cofactor for early neuropsychological impairment (45). Almost all studies addressing this issue, however, have found that the presence and severity of depressive symptoms are not associated with the early neurocognitive disturbance in HIV-1 infection (e.g., see refs. 30 and 32).

In a recent study by Satz et al. (87), low education (i.e., less than 12 years) was associated with neuropsychological impairment in early HIV-1 infection. In this study, 38% of the subjects with low education exhibited cognitive abnormalities, compared to less than 17% of the high education group. These authors suggest that low education indirectly reflects a marker of lower "reserve capacity" that increases the risk of developing cognitive/motor impairment in neurological damage, such as that which may occur in early HIV-1 infection.

It has also been suggested that immunosuppression precedes neurobehavioral dysfunction (26). Other investigators, however, have found that neurocognitive and motor alterations in early HIV-1 infection are independent of degree of immunosuppression as measured by CD4+ count (97). More recently, Martin et al. (53) have suggested that typical markers of immunosuppression, such as CD4+ count, are not valid or reliable measures of CNS status. Rather, these authors indicated that increased levels of quinolinic acid may play a salient role in the development of early HIV-1-related cognitive/motor impairment. Additional promising work assessing the relationship between neurobehavioral functioning and quinolinic acid levels is ongoing.

Stern et al. (93) conducted a study of the presence and extent of neurocognitive impairment in early HIV-1 infection in subjects who met extensive inclusion criteria aimed at excluding individuals with the previously detailed confounding variables. In addition, several potential cofactors, such as vitamin B12 deficiency, depression, limited education, and immunocompromise (i.e., low CD4+ count), were controlled for statistically. Prior to controlling for the various potential cofactors, the seropositive group was found to have significantly worse performance in attention, information processing, and motor functioning than the group of HIV-1-seronegative homosexual men. However, once the influence of these cofactors was removed, the asymptomatic HIV-1-infected participants had significantly worse performance only on tasks sensitive to motor slowing and dexterity. These group differences were not clinically significant, however, because the asymptomatic group's performance was not in the impaired range compared to normative data. The results of this study suggest two important findings: (i) in a nonconfounded group of asymptomatic HIV-1-infected homosexual men, neuropsychological dysfunction appears limited to mild, subclinical motor disturbance; and (ii) in order to assess the role of HIV-1 in the development of cognitive/motor disturbance during the asymptomatic stage, it is imperative to account for the possible impact of previously implicated cofactors.

Even with continued investigation of the impact of confounding variables and possible cofactors, two major questions remain to be addressed regarding early cognitive and motor disturbance in HIV-1: (i) Is the severity of the "impairment" clinically meaningful? That is, statistically significant between group differences on sensitive neuropsychological tests do not necessarily translate into noticeable impairments in occupational and/or daily living tasks. This issue of the "ecological validity" of early HIV-1 neuropsychological findings has yet to be addressed adequately. (ii) Is the presence of early cognitive/motor dysfunction indicative of either future presence or course of dementia, or of future disease course in general? With regard to the former, very little is known. With regard to prediction of disease course, an intriguing study by Mayeux et al. (55) suggests that the presence of impaired neuropsychological test performance in both asymptomatic and symptomatic HIV-1-infected individuals is associated with a significantly increased risk of mortality over a 36-month period.

Zidovudine (i.e., azidothymidine, AZT, Retrovir), an anti-retroviral, continues to be the mainstay in treatment of HIV-1 spectrum disease. It has been found to temporarily improve the prognosis in patients with AIDS and to reduce disease progression in asymptomatic HIV-1infected individuals. However, neurotoxicity has been associated with chronic doses of zidovudine, including headache, seizures, Wernicke's encephalopathy, coma, and mania (83). In spite of these reports of neurotoxicity, there is also evidence that zidovudine significantly improves HIV-1-associated cognitive/motor abnormalities in symptomatic HIV-1-infected patients and in persons with AIDS (90). Recent data suggest that zidovudine may reverse, albeit transiently, the severity of HIV-1-associated dementia (94). However, the positive effects of zidovudine on neuropsychological function may be limited to patients with advanced HIV infection in light of a recent finding that the drug had no significant beneficial neuropsychological effects on HIV-1-infected individuals without AIDS (27).

Another anti-retroviral, 2¢,3¢-dideoxyinosine (ddI), now used in the treatment of AIDS has also been shown to be effective in improving HIV-1-associated cognitive impairment in a small number of cases (100). However, controlled studies of the neurocognitive efficacy of ddI have yet to be completed.

Because of the bradyphrenia and attentional deficits exhibited by HIV-1-infected patients, the use of psychostimulants has been suggested. Results of anecdotal case reports and uncontrolled case series of methylphenidate and dextroamphetamine have indicated that these agents may be useful in diminishing the neurocognitive effects of HIV-1 (25). Once again, prospective, controlled investigations have not yet been conducted.

The psychiatric sequelae of HIV-1 infection and AIDS are numerous and have etiologies that involve neurobiological and psychosocial factors. These include the natural and expected grief response to being diagnosed with a terminal illness, later reactions to disability and illness, exacerbation of preexisting psychiatric illness, development of new primary psychiatric symptoms and syndromes, and the neuropsychiatric manifestations of HIV-1-associated neurological illness.

Several investigators have examined the psychological impact of HIV testing. In a longitudinal study of individuals with a variety of HIV-1 risk factors, Perry and colleagues (36, 74, 75) found that all subjects reported moderate-to-high levels of depression and anxiety prior to receiving HIV test results. Subjects who received a negative HIV test result reported a significant reduction from pretest levels of depression and anxiety symptoms. This reduction was sustained at both 2- and 10-week follow-ups. Subjects who received a positive HIV test result reported no significant change from their pretest level of distress at 2-week follow-up. However, by the 10-week follow-up, these HIV-1-infected subjects also reported a significant reduction of depression and anxiety symptoms. Furthermore, by 1-year follow-up, both the HIV-1-infected and HIV-1-uninfected individuals reported similar levels of depression and anxiety. It is important to note that, compared to routine HIV counseling, counseling that included stress reduction techniques significantly reduced depression and anxiety symptoms in the HIV-1-infected individuals (72).

The findings by Perry and colleagues are supported by other studies. For example, Moulton et al. (59) found that 2 weeks after HIV test notification there were similar levels of negative mood, hopelessness, and anxiety in 66 individuals notified of positive results and in 41 notified of negative results. In this study most (78%) HIV-1positive and about half (43%) HIV-1-negative subjects correctly anticipated their test result. Thus, similar levels of dysphoric mood may have reflected anticipatory adjustment of their presumed HIV-1 status.

It is understandable that individuals who receive notification of positive HIV test results will be emotionally distressed as they adjust to the knowledge of their HIV-1 serostatus. The severity of the acute distress will vary from individual to individual. Whereas some individuals may react with little distress, others may be at increased risk of suicide. Rundell et al. (86) found that 15 out of 826 (<2%) HIV-1-infected active duty members of the U.S. Air Force had attempted suicide, with 47% of the attempts occurring within 3 months of serostatus notification and 67% within 1 year. In contrast, Perry et al. (73) found that suicidal ideation remained at pretest levels among those who received notification of positive HIV test results. In addition, these levels dropped significantly by 2 months after positive test result notification, and, after comparing individuals who received a positive HIV test result with those who received a negative one, the authors found that the proportion of individuals with suicidal ideation was similar in both groups, suggesting little, if any, relationship between suicidal ideation and HIV test results. It is important to note, however, that the low level of suicidal ideation in Perry's HIV-1-infected subjects may have resulted from the extensive HIV counseling provided to participants in this study. Such counseling may serve to combat suicidal thoughts or intent that have been found to be associated with greater levels of depression in both HIV-infected and HIV-uninfected individuals.

Several controlled studies have shown that asymptomatic HIV-1 infection is not associated with increased depression or anxiety symptoms. For example, Perkins et al. (70) found similar levels of depression and anxiety symptoms in HIV-1-infected and HIV-1-uninfected homosexual men. Thus, it appears that although individuals are often distraught after receiving positive HIV-1 test results, after an adjustment period lasting weeks to a few months, most will cope well and will show a reduction in anxiety and depressive symptoms. Consequently, it appears that symptoms of depression and anxiety should not be considered "normal" in asymptomatic HIV-1 infection. Rather, significant symptoms should warrant careful clinical evaluation.

Controlled studies have shown that the prevalence of major depression and other mood disorders is higher in asymptomatic HIV-1-infected homosexual men than in the general population (82, 84), but is similar to that in HIV-1-seronegative homosexual men (3, 70, 99). Several studies have shown that 4–9% of both HIV-1-infected and HIV-1-uninfected homosexual men report a major depression in the month prior to study evaluation. In one study, Perkins et al. (70) found that 6% of both HIV-1-infected and HIV-1-uninfected subjects developed a major depression during a 6-month follow-up period. There is also evidence that similar proportions (ranging from 0% to 5%) of HIV-1-infected and HIV-1-uninfected individuals meet DSM-III-R criteria for current anxiety disorders (70). Once again, these findings underscore the issue that mood disorders should not be considered a "normal" phenomenon in HIV-1-infected individuals. Rather, they should be assessed carefully and treated appropriately.

Differential diagnosis of major depression in HIV-1-infected patients is complicated because several symptoms of major depression (i.e., fatigue, sleep disturbance, and weight loss) are also common symptoms of HIV-1 disease progression (e.g., see refs. 58 and 63). Recent reports, however, have indicated that although complaints of fatigue and insomnia in asymptomatic HIV-1-infected homosexual men are significantly associated with severity of depressed mood and other symptoms of major depression, they are not associated with low CD4+ counts or decreased neuropsychological functioning (4, 69). Thus, complaints of fatigue and insomnia in otherwise asymptomatic HIV-1-infected patients are highly suggestive of an underlying mood disorder and indicate that patients with such complaints should be assessed routinely for major depression.

Although factors that influence risk for development of mood symptoms in HIV-1-infected individuals have not yet been well-studied, a recent investigation found a relationship between major depression in asymptomatic HIV-1-infected homosexual men and a prior history of major depression (70). This study, however, found no relationship between major depression and neuropsychological functioning, suggesting that the depressive disorder is not related to the CNS effects of HIV-1. Therefore, in individuals infected by HIV-1, past history of a major depression may be a risk factor for subsequent development of major depression.

The strategies that an individual uses to cope with the threat of AIDS may be related to overall level of dysphoric mood (60). Leserman et al. (46) reported that asymptomatic HIV-1-infected men who used active coping strategies to deal with the threat of AIDS (e.g., fighting spirit, reframing stress to maximize personal growth, planning a course of action, seeking social support) had lower levels of depressed and anxious mood than those who used passive coping strategies (denial or feeling helpless). Importantly, there is also evidence that personality disorder is common in HIV-1-infected patients, and that these patients experience greater dysphoria and cope less well with the threat of AIDS (66). Therefore, the way in which an individual deals with his or her seropositive HIV status, vis-à-vis coping strategies, may influence the development of depression or anxiety.

Thus, the available data suggest that asymptomatic HIV-1-infected individuals have no greater dysphoric symptoms or clinical mood disorders than do uninfected control subjects. As has been reported of individuals with other life-threatening diseases, HIV-1-seropositive individuals are usually able to adjust successfully to their infection, and the majority are able to maintain hope over time. Furthermore, the available data suggest that by creating preventive intervention programs targeted at HIV-1-positive individuals at risk for developing depressive or anxious symptomatology (e.g., those with a history of clinical depression) and by fostering the development of active coping strategies in these individuals, the rates of depression and anxiety in this population may even be lowered.

The effects of HIV-1 on the CNS may result in a variety of psychiatric symptoms in the latter stages of the illness. In addition, individuals infected with HIV-1 may be at further risk for developing psychiatric symptoms due to diseases secondary to AIDS that also have CNS effects, as well as due to medications used to treat HIV-1. Furthermore, psychiatric symptoms in HIV-1-infected individuals in later stages of the illness may also represent new-onset psychiatric disorders. The direct CNS effects of HIV-1, HIV-1-related CNS disturbances, and CNS effects of medications used in the treatment of AIDS may each influence the development of mood disorders and psychosis. Several ongoing longitudinal cohort studies are currently examining psychiatric symptoms across the course of HIV-1 infection and should ultimately provide important data to address the relative contribution of each of these factors to the development of mood and related disorders.

There is some available evidence, however, that HIV-1 may cause organic mood disturbance. In a 17-month retrospective chart review of patients with AIDS, Lyketsos et al. (49) examined associated historical and clinical features in an attempt to separate organic and functional symptoms. They used family history of mood disorder as a "marker" for functional mood disorders. They further assumed that coexisting dementia and low CD4+ count were "markers" for HIV-1-related mood disorders. They found that none of the patients with a personal or family history of mood disorder had coexistent dementia, and that all but one of the patients without a personal or family history of mood disorder had coexistent dementia. In addition, among the 8% of the patients who experienced manic episodes, CD4+ count was significantly higher in those individuals without a personal history of mood disorder. Although these findings are not based on controlled studies, they do suggest that mania may be a consequence of the effect of HIV-1 on the brain.

Vitamin B12 deficiency may also place HIV-1-infected patients at risk for organic mood disturbance. Between 20% and 30% of patients with AIDS and 7% of asymptomatic HIV-1-infected patients have been reported to have a vitamin B12 deficiency. Furthermore, vitamin B12 deficiency has previously been shown to be associated with depression, and can occur in the absence of hematologic or neurologic signs (18). Although the relationship between vitamin B12 level and depressive symptomatology in HIV-1-infected individuals is not clear (85), it is prudent that the medical evaluation of depressive symptoms in HIV-1-infected individuals include assessment of serum B12 levels.

Psychosis may also result from CNS HIV-1 involvement. Numerous case studies of symptomatic HIV-1infected individuals have reported psychotic symptoms, including delusions, bizarre behavior, and hallucinations. Mood disturbances, including euphoria, irritability, and labile or flat affect, have often accompanied psychotic symptoms in these patients. Similarly, anxiety and agitation were reported in almost half of the reported cases. In addition, there is some evidence that psychosis may be a symptom of the terminal stages of AIDS, because half of the patients described had a progressively worsening course, with dementia or death occurring within a few months after the onset of the psychotic symptoms.

Psychosis may be more frequently found in patients with significant AIDS-related neurocognitive impairment than in patients in earlier stages of the disease. In one retrospective chart review of 46 patients identified with HIV-1-associated dementia, Navia and Price (62) found that 15% had developed psychotic symptoms. Relatedly, data from an ongoing, prospective study suggests that HIV-1-infected patients with psychosis have greater neurocognitive impairment than do nonpsychotic HIV infected controls (Jeste, personal communication).

Other sequelae of HIV-1 infection may also result in psychiatric symptomatology. For example, CNS mass-occupying lesions and generalized metabolic disturbance may lead to depressive symptomatology. Furthermore, almost all of the medications used to treat HIV-1-related illnesses have psychiatric symptoms as potential side effects, and zidovudine has been reported to cause mania (20). Thus, a complete organic work-up should be considered for HIV-1-infected individuals with disturbance of mood or psychosis in order to determine if the etiology is CNS-based or idiopathic in nature.

Several epidemiological studies suggest that AIDS patients are at increased risk of death by suicide. The relative prevalence is estimated to range from 7 to 36 times the rate in demographically similar control populations. Other studies, however, have not found patients with AIDS to have higher suicidal ideation, especially when comparing persons with AIDS to other medically or neuropsychiatrically ill patients. For example, McKegney and O'Dowd (56) found that in hospitalized medical patients seen in psychiatric consultation, patients with AIDS were significantly less likely to have suicidal ideation than were asymptomatic HIV-1-infected patients and patients with unknown HIV-1.

Rather than finding a correlation between current suicidal ideation and AIDS, studies have found an association between past history of suicidal ideation or past psychiatric treatment and current level of suicidal ideation. Thus, the relationship between suicide risk and AIDS remains unclear. However, suicidal ideation in patients with AIDS should never be regarded as part of an "expected" or temporary grief reaction; rather, it should warrant careful evaluation for major depression and appropriate treatment (see also Neuroendcrinology of Mood Disorders).

Available evidence suggests that mood symptoms and syndromes in the asymptomatic stage of HIV-1 infection are not secondary to the effects of HIV-1 on the brain, and should be evaluated and treated as in the general population. No controlled antidepressant clinical trials in patients with HIV-1 infection have been published, though several are currently underway and other open trials have been reported. Imipramine and fluoxetine have been shown to be effective in the treatment of major depression in asymptomatic HIV-1-infected patients (47, 67, 80). Fernandez and Levy (24) found that treatment with either bupropion or fluoxetine resulted in a moderate-to-marked improvement of major depressive symptoms in 81% of patients with AIDS and found that bupropion was superior to fluoxetine in activating anhedonic patients. Rabkin and Harrison (80) have found imipramine, fluoxetine, and sertraline to be as effective for the treatment of major depression in HIV-1-seropositive patients as for the treatment of depression in patients with no medical disease. They have also conducted the first double-blind, placebo-controlled trial of imipramine in HIV-1 and found a 74% response rate to imipramine. Although there is evidence that desipramine decreases natural immunity (natural killer cell activity) in vitro (57), Rabkin and Harrison (80) found no in vivo effects of imipramine on CD4+ cell counts. There is additional in vitro evidence that neither desipramine nor lithium enhances HIV-1 replication in cultured cells (21). Thus, there is no available evidence to suggest that antidepressants cause adverse immunological effects in the treatment of depression in HIV-1-infected individuals. In the selection of antidepressants, it appears that those with higher anticholinergic properties should be avoided; reports suggest that symptomatic HIV-1-infected patients may be more vulnerable to the anticholinergic side effects of tricyclic antidepressants than asymptomatic HIV-1infected or HIV-1-uninfected patients (24).

Data on the treatment of mania in HIV-infected individuals have been limited to clinical anecdote and retrospective chart review. The few available findings suggest that while lithium may be effective, it is associated with significant adverse effects including neurotoxicity and gastrointestinal side effects; additional work suggests that anticonvulsants, including valproic acid, carbamazepine, and clonazepam, are effective alternatives.

Psychostimulants, such as methylphenidate and dextroamphetamine, have also been reported to improve depressive symptoms in symptomatic HIV-1-infected patients (33). Psychostimulant medication may be particularly effective in improving apathy and lethargy associated with the later stages of AIDS or with the more generalized HIV-1-associated dementia complex (79). In addition, there is also evidence from an open trial that testosterone is effective in treating low mood, anergy, and anhedonia in late-stage HIV-1 disease (Rabkin, personal communication).

To date, there have been no published studies of the treatment of anxiety symptoms in HIV-1-infected individuals. However, clinical experience suggests that general treatment guidelines apply to this patient population. Benzodiazepines have been efficacious and safe in treating anxiety symptoms in HIV-1-infected patients. As with other medically ill populations, benzodiazepine dosage may need to be decreased in patients with AIDS in order to avoid CNS depressant effects. Furthermore, anecdotal clinical evidence suggests that the intermediate acting agent, lorazepam, is effective and well-tolerated in this population.

Although there have been no published controlled clinical trials of the treatment of psychosis in HIV-1-infected patients to date, several case studies have suggested that neuroleptics are generally effective (Perry, personal communication). Patients with psychosis who are in the symptomatic stages of HIV-1 infection, however, have consistently been reported to require very low doses of neuroleptics. Controlled studies are currently being conducted on the treatment of HIV-1-related delirium and HIV-1-associated psychosis. Brietbart (10) has found both haloperidol and chlorpromazine in low doses (40–80 mg chlorpromazine equivalents, mean daily dose) to be highly efficacious in the treatment of delirium. Lorazepam was ineffective as a single agent and was poorly tolerated. Both neuroleptics were well-tolerated at these low doses, and there was no evidence of significant extrapyramidal symptoms. Jeste and colleagues are conducting a controlled study of new onset psychosis in HIV-1-infected individuals. Results from this ongoing study suggest that mood symptoms, usually dysphoria, are extremely common, and there is evidence of greater neuropsychological impairment compared to nonpsychotic controls. Standard neuroleptics (haloperidol and thioridazine) are being compared for efficacy and adverse effects (Jeste, personal communication). Both neuroleptics are effective at low doses (124 mg chlorpromazine equivalents, mean daily dose). Thus, it appears that HIV-1-infected individuals are especially sensitive to neuroleptic side effects. Although a retrospective chart review demonstrated no significant increase in extrapyramidal symptoms in AIDS patients treated with dopamine antagonists as compared to non-HIV-1-infected control patients (34), the small sample size and the general methodology of this study would only allow for the detection of large differences in prevalence of extrapyramidal symptoms; moderate, but clinically relevant, differences could not be detected. Therefore, at this point, when treating psychotic symptoms in symptomatic HIV-1-infected patients it would appear clinically prudent to slowly titrate neuroleptic medications, minimize drug dose, and carefully monitor for side effects.

Although individuals with HIV-1 infection may exhibit psychiatric symptoms secondary to, or associated with, underlying CNS dysfunction, they also exhibit mood and anxiety symptoms as a result of psychosocial factors, such as loss of health, financial problems, social stigma, and other similar stressors associated with the progression of HIV-1 disease. Therefore, interventions that provide emotional support and enhance coping are important components of the treatment of functional mood symptoms in HIV-1-infected patients. Controlled studies are few, but stress prevention training has been associated with significantly lower depression and anxiety symptoms in individuals undergoing HIV testing (72). Interpersonal psychotherapy has also been found to be an effective treatment of major depression in patients with HIV-1 infection. Markowitz et al. (52) found that interpersonal psychotherapy resulted in significant improvement of symptoms in 20 out of 23 HIV-1-infected individuals. Collectively, these results suggest that individual and group psychotherapies as well as support groups may act to improve quality of life, including the ability to cope with AIDS, a factor that has been found to be associated with less depression and anxiety in HIV-1-infected individuals (46).

There is increasing evidence that the CNS influences the immune system, and there is considerable popular enthusiasm for the concept that stress and depression influence the course of physical diseases such as cancer and HIV-1 infection (31, 68). Although there are many positive studies, there is still no consistent clinical evidence supporting this relationship. However, decreases in natural immunity as measured by reductions in the number and function of natural killer cells is the most consistent and reproducible finding to date (19, 89, 92). The potential clinical relevance of these findings are suggested by studies of cancer patients who have been reported to have increased natural killer cell number and function (23), as well as increased survival time, in association with a well-controlled psychoeducational group psychotherapy intervention (22, 23, 91). A number of studies have begun to address the question as to whether stress and depression are associated with immune suppression and disease progression in HIV-1 infection.

Rabkin et al. (81) found no relationship between psychosocial and psychiatric factors such as depressive disorders, distress, and stressors, on the one hand, and measures of HIV-1 disease progression, including CD4+ and CD8+ cell counts, on the other hand. However, they found a trend for depressive disorders, distress, and stressors to be associated with number of AIDS-related symptoms. Similarly, over a 12-month follow-up period, Perry et al. (71) did not find a relationship between social support, stressful life events, and a variety of psychological factors and CD4+ and CD8+ counts. They did find, however, that severity of physical symptoms was associated with greater levels of distress in HIV-1-infected subjects without AIDS.

There is evidence that the stress of HIV-1 testing is significant enough to alter both neuroendocrine and immune system function. In a group of individuals who ultimately were shown to be HIV-1-infected, Antoni et al. (2) found that abnormal elevations in serum cortisol decreased significantly after notification of a negative HIV test. Furthermore, higher-baseline cortisol was associated with lower lymphocyte functioning, suggesting that the stress of HIV testing impacted on immune system function. Lastly, use of denial as a coping mechanism was inversely associated with lymphocyte response to stimulation, suggesting that coping strategy may in part mediate the body's reaction to the stress of HIV testing in at-risk individuals.

In a negative study, Kertzner et al. (43) found no consistent evidence of stress-associated alterations of the hypothalamic–pituitary–adrenal axis in seropositive subjects and found no relationship between cortisol and CD4 cell counts. However, other measures of immune function may be more sensitive to the effects of emotional factors. Ironson et al. (35) have found alterations in mitogen stimulation of lymphocytes in subjects before and after HIV antibody testing. Extensive data from our laboratory also suggest that life event stress is greater in HIV-1seropositive subjects than in HIV-1-seronegative controls and is associated with detriments in the number of natural killer cells and cytotoxic T cells (Evans et al., submitted). In a similar comprehensive assessment of stress and depression, Patterson et al. (65) have reported that stress and depressive symptoms predict a 6-month change in immunity as measured by percentage of CD4+ cells and serum b2-microglobulin. These data suggest that stress and depression may accelerate the course of immunologic decline in HIV-1 infection, but are preliminary and require further longitudinal confirmation.

Recent data from population-based studies of HIV-1-infected individuals have yielded conflicting findings with respect to depression and immunity (13, 50). In one investigation, Burack et al. (13) found depressive symptoms to predict a more rapid decline in CD4+ counts, while Lyketsos et al. (50) found no such relationship. Unfortunately, no available studies have followed cohorts through the course of HIV-1 infection using (a) comprehensive clinical examinations of stress and depression, (b) comprehensive endocrine and immune assessments, and (c) clinical measures of HIV-1 disease progression. Consequently, the relationship between psychological, psychiatric, and psychosocial factors and course of disease in HIV-1 infection has not been resolved and warrants continued exploration (see Interactions Between the Nervous System and the Immune System: Implications for Psychopharmacology, for more background).

Great strides have been made over the last decade in both describing the neuropsychiatric functioning associated with HIV-1 spectrum disease and increasing our understanding of the underlying neuropathological mechanisms. It is now clear that the virus enters the CNS early in the course of the disease and has both primary and secondary effects on neural tissue. Subtle abnormalities can be detected in pathology, neuroimaging, and neuropsychological studies prior to the onset of AIDS-defining illnesses, though the clinical significance of these findings remains unclear. In symptomatic AIDS, neuropsychiatric and neurological complications are highly prevalent and often can be the first manifestations of AIDS.

Although most individuals infected with HIV-1 cope well, major depression is highly prevalent in both HIV-1-seropositive and HIV-1-seronegative homosexual men, compared to epidemiologically based estimates of the general population. The interrelationship among the CNS, the endocrine system, and the immune system is only beginning to be understood in HIV-1 infection. Moreover, the effects of stress and depression on endocrine and immune function, as well as the potential effects on HIV-1 disease progression, will require comprehensive, longitudinal investigation. Future study is also necessary to understand the neuropsychiatric manifestations of HIV-1 infection in women, as well as the special effects on neurodevelopment. Cross-cultural comparisons are also needed.

Finally, early psychopharmacological treatment studies have yielded promising results for the alleviation of depression and psychotic symptoms, as well as for the improvement of neurocognitive impairment. Future neuropsychopharmacological approaches will certainly focus on both the direct and indirect effects of HIV-1 on the brain in order to develop interventions that could alter the course of disease, as well as focus on symptomatic treatments in order to improve clinical outcome and quality of life.

This chapter was supported in part by a grant from the National Institute of Mental Health, MH R01 5-47486. The authors thank John R. Z. Abela and Isabell C. Leshko for editorial assistance.

published 2000