icon star paper   HIV Articles  
Back grey_arrow_rt.gif
HIV, Dementia & HAART
  "Evidence for a change in AIDS dementia complex in the era of highly active antiretroviral therapy and the possibility of new forms of AIDS dementia complex"
AIDS: Volume 18 Supplement 1 January 2004Brew, Bruce J
From the Departments of Neurology and HIV Medicine, St Vincent's Hospital, Sydney, NSW 2010, Australia; and National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, Darlinghurst, NSW 2010, Australia.
This review will discuss emerging evidence for the possibility that AIDS dementia complex (ADC) has changed in the era of highly active antiretroviral therapy (HAART). The consequences of not considering these possibilities at the patient level and at the level of research are considerable. Data will be discussed that are derived from epidemiological studies, neuropsychological and positron emission tomography studies, as well as analyses from the abacavir ADC trial. These will then be assessed to develop the concept that there are now different forms of ADC: an inactive form, a chronic variety and a 'transformed' variant. Whereas the latter relates to the compounding influence of a number of other processes on ADC, such as hepatitis C, particular discussion will focus upon Alzheimer's disease and whether HIV may lead to Alzheimer-like changes. It is certainly recognized that some of the concepts discussed here are highly speculative.
Recent evidence from epidemiological studies points to a change in AIDS dementia complex (ADC). In the era before the introduction of highly active antiretroviral therapy (HAART) the mean CD4 cell count at the time of the diagnosis of ADC was between 50 and 100 cells depending on which series are examined. In the HAART era, the mean CD4 cell count has significantly increased. This was first reported in an Australian cohort, in which the count had risen to 160 cells. Subsequent studies have confirmed this change. The reason for this elevation is not clear. It may be related to the failed restoration of a specific defect in immune function related to ADC, or it may signify that the duration of HIV disease is now becoming more important, or it could be a combination of the two. Previously, the CD4 cell count was a marker of disease severity as determined by the degree of immunodeficiency as well as disease duration; the two were linked in the vast majority of patients. Now, HAART seems to have severed that link. The picture may be more complicated, however. Given that the risk of ADC increases as the CD4 cell count falls, especially when it is below 200 cells, and that such a CD4 cell count may allow or facilitate autonomous brain infection by HIV, it seems that the nadir CD4 cell count may be a new important risk factor for ADC. This is also supported by the observation that most of the currently available antiretroviral drugs have limited access to the brain. If a low nadir CD4 cell count allows HIV access to the brain then the institution of HAART is unlikely to eradicate it. HAART has thus raised the CD4 cell count which is commonly seen in patients with ADC, and has introduced two potentially new risk factors: nadir CD4 cell count and disease duration.
The 'natural history' of ADC has also changed. In the pre-HAART era the mean time to death was 6 months from the time of diagnosis. This has now been considerably lengthened to 44 months. There are two consequences to this: the prevalence of ADC is increasing, and ADC patients are now more likely to be exposed to other potentially confounding conditions simply because they are living longer. Hepatitis C, testosterone deficiency and the effects of aging are now becoming important confounders, as will be discussed later.
Apart from the potentially increased importance of confounding factors, the actual cognitive deficit in ADC may be changing. Evidence for this is indirect, in that it pertains to HIV-infected patients who are neurologically asymptomatic. One study found that there is an altered pattern of neuropsychological deficits in such patients, with a tendency for more 'cortical' type abnormalities to be found. There is an increased frequency of verbal memory impairment along with a more classical distribution of impairment (fine motor incoordination), suggesting the imposition of one neuropsychological pattern upon a more classical picture.
A recently completed prospective positron emission tomography (PET)-cerebrospinal fluid (CSF) study also highlighted changes in ADC. The basal ganglia hypermetabolism that was so characteristic of ADC in the pre-HAART era and which correlated with the clinical and neuropathological changes in the basal ganglia does not seem to be as prominent. Indeed, mesial temporal lobe abnormalities now appear to be more relevant. Although it is true that temporal lobe changes were seen in ADC even in the pre-HAART era, they were not prominent and were less conspicuous compared with the alterations observed in the basal ganglia both by PET scanning and formal neuropathological assessment. Finally, standard CSF markers of ADC, such as beta-2 microglobulin and the HIV viral load in the CSF, are no longer sensitive to the presence or severity of ADC. Of the six patients who developed ADC, four did so despite CSF suppression. This is in marked contrast to the pre-HAART era in which such markers were strongly correlated with the severity of ADC.
Further evidence of a change in ADC can be seen from the close inspection of the abacavir ADC trial. A total of 105 patients with mild to moderate ADC and stable HAART for more than 8 weeks were randomly assigned to receive abacavir (600 mg twice a day) or matched placebo for 12 weeks. Patients were assessed using eight neuropsychological tests, with the primary outcome measure being a change in the summary Z score. Secondary measures included HIV-1 RNA, beta-2 microglobulin, soluble TNF receptor-2 and quinolinic acid in the plasma and CSF. The median change in the summary Z score at week 12 was +0.76 (range -2.53 to +2.69) for the abacavir group and +0.63 (range -3.89 to +3.32) for the others (p=0.735). This lack of efficacy occurred despite evidence for the activity of abacavir: at week 12 more subjects in the abacavir group had plasma HIV-1-RNA levels of 400 copies/ml or less (46 versus 13%, p=0.002) , the abacavir group tended towards greater decreases in CSF HIV-1 RNA (-0.64 versus -0.26 log10 copies/ml), and. the mean abacavir concentration reached 0.741 ug/ml 3-4 h post-dosing, exceeding a representative abacavir IC50 for HIV-1. The lack of efficacy, however, was probably related to unusual features of the ADC patients: 56% at baseline had CSF HIV-1-RNA levels less than 100 copies/ml and 83% had CSF beta-2 microglobulin levels below 3 nmol/l. This observation is completely at odds with what was found in the pre-HAART era. Indeed, CSF HIV RNA was the most sensitive marker of the severity of ADC, closely followed by CSF beta-2 microglobulin.
Evidence for different forms of the AIDS dementia complex
Because over half the patients in the abacavir ADC trial had essentially unremarkable CSF concentrations of HIV RNA and beta-2 microglobulin, there is a strong suspicion that the majority of such patients had an inactive or 'burnt out' form of ADC. This would not necessarily be surprising. They may have already responded maximally to HAART before starting the trial, and may have been left with a fixed deficit, perhaps as a result of neuronal loss. Intensification of their antiretroviral regimen would therefore not substantially improve their cognitive deficit. Although this concept remains speculative, it does seem likely. If one accepts the existence of this form of ADC, then it is also likely that individual patients may have a component of their deficit that is fixed or inactive. In other words, instead of the whole current deficit being inactive, a component might be. This also seems likely. Both these concepts have potentially profound effects on efforts directed at understanding the pathogenesis of ADC. The interpretation of the possible pathogenetic significance of certain changes in the CSF or brain in a patient with ADC could be invalidated if the presence of completely or even partly inactive disease is not appreciated. The verification of the existence and the means of identifying an inactive form of ADC are under evaluation. Nonetheless, preliminary post-hoc analyses of the abacavir ADC trial data suggest that a CSF HIV-RNA load below 100 copies/ml may be a marker. Whether there are other markers such as the duration of ADC or the lack of a raised choline peak (an indirect marker of inflammation seen in ADC patients before the commencement of HAART) on magnetic resonance spectroscopy awaits further study.
HAART may also be associated with a chronic form of ADC. The above-mentioned prospective PET-CSF study contained patients who developed ADC over several years in the face of undetectable CSF and plasma HIV viral loads at least at the time points studied. Moreover, there is indirect evidence. A recent study by Abdulle et al. found that neuroasyptomatic HIV-infected patients who had been treated with HAART for 2 years still had mildly elevated CSF concentrations of neopterin in the presence of normal concentrations of beta-2 microglobulin and HIV RNA in both compartments. This suggests that HAART cannot return all CSF parameters to normal. The elevated neopterin concentration implies that there is activation within the central nervous system of monocytes, microglia or both. Such immune activation may be secondary to productive brain infection, which is below the detection limits in the CSF, or it may be related to local unchecked immune activation from past damage. Whether this leads in the long term to brain damage and a chronic form of ADC is, however, speculative at present. Nonetheless, the lack of awareness of this possibility may also confound efforts aimed at understanding ADC pathogenesis. There is also indirect evidence from systemic HIV infection for the presence of a chronic form of ADC. It is well known that HIV can be cultured from lymph nodes in patients with plasma HIV viral loads below 20 copies/ml. It seems likely that this could also be true for the brain.
ADC may also be 'transformed' in the era of HAART. As patients live longer they are more likely to be exposed to other disorders that could affect cognition. There are four such conditions that deserve particular attention. The first is testosterone deficiency. In the pre-HAART era, patients usually did not live long enough to develop significant testosterone deficiency. Now it has been estimated that approximately 30% of patients with advanced HIV disease have testosterone deficiency. Moreover, there is evidence, albeit controversial, that testosterone deficiency can lead to a cognitive deficit at least in non-HIV-infected individuals. The extent to which this can be reversed by replacement is unknown. It is therefore possible that this may be superimposed on existing ADC-related deficits, leading to a 'hybrid' state. The second condition is that of brain mitochondrial toxicity related to HAART. Whereas it is certainly clear that some of the antiretroviral drugs may cause tissue-specific mitochondrial toxicity, for example stavudine and peripheral neuropathy, it is still unknown whether in the long term HAART-treated patients may develop brain mitochondrial toxicity. The third condition is that of hepatitis C. It is now becoming increasingly apparent that hepatitis C in conjunction with HIV may lead to more profound cognitive deficits. In an individual patient this may also lead to a 'layering' of one deficit (hepatitis C) on top of another (ADC).
Finally, and perhaps most importantly, ADC non-ADC patients may be at an increased risk of developing Alzheimer's disease. Although this is also speculative there are several reasons for suspicion: increased age, high lipids, axonal injury, and the effects of tat and quinolinic acid on the brain. HIV-infected patients in general and ADC patients in particular are now living longer. The number of patients over the age of 60 years is increasing. The mean age in a prevalence survey of neuropsychological abnormalities of an outpatient clinic at a tertiary referral hospital 10 years ago was 38.5 ± 7 years, whereas the same referral population currently has a mean age of 49 ± 8.8 years.
Similarly, the number of patients with raised cholesterol and triglyceride levels is increasing as a result of increased age, HIV disease itself, and HAART, especially the use of protease inhibitor drugs. Raised cholesterol concentrations have repeatedly been shown to be a risk factor for Alzheimer's disease. Also of importance is the finding in neuroasymptomatic patients of axonal injury, similar to that which occurs in patients with head injury. The latter group is at an increased risk of Alzheimer's disease, therefore making it likely that HIV-infected patients also have an increased risk. The HIV regulatory protein tat may play a role in increasing the risk of Alzheimer's disease. Rempel et al. found in vitro that tat inhibits the activity of the beta amyloid degrading enzyme, neprilysin. Furthermore, the macrophage/microglia-derived neurotoxin quinolinic acid may also be important. Recent evidence shows that amyloid beta 1-42 is capable of inducing microglia to produce quinolinic acid, which could then lead to neuronal death through N-methyl-d-aspartate receptor activation and lipid peroxidation. There are thus numerous reasons for suspecting that HIV patients, in particular those with ADC, are at an increased risk of Alzheimer's disease.
In conclusion, it therefore seems very likely that the introduction of HAART has changed the very nature of ADC. It is now not so strongly linked to the CD4 cell count, there may be inactive and chronic forms of ADC, and patients are living longer making them vulnerable to the possible cognitive effects of hitherto unappreciated disorders. Chief among these is the large body of evidence that points, theoretically at least, to an increased risk of Alzheimer's disease. Verification of these concepts is clearly required, but equally clearly such confirmation will take several years. In the mean time, clinicians and researchers should be open to these possibilities in order to avoid their confounding effects on understanding HIV neuropathogenesis.
  icon paper stack View Older Articles   Back to Top   www.natap.org