Reports

This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.

PHYSICIAN PERSPECTIVE Viewpoint based on the following article: Arch Neurol 2008;65(1):65-70.

January 2009

Reported by:

Scott Letendre, MD

Associate Professor of Medicine, University of California, San Diego, La Jolla, California

The introduction of combination antiretroviral therapy (CART) is credited with substantially reducing the complications of HIV infection, including death,¹ but the persistence of dementia and neurocognitive impairment has been a prominent exception.² Since the introduction of CART, neurocognitive complications initially remained stable, but more recently, appear to have increased.³ Persistence of dementia and cognitive decline in treated individuals may be attributable to irreversible neural injury that can occur prior to initiation of CART⁴ but another appears to be due to the persistence of HIV replication in the central nervous system (CNS) even when plasma viremia is well controlled.^5,6 In CART-treated patients, estimates of the prevalence of cognitive impairment and other CNS complications range from 20% to 53%.^7-9 The incidence appears to increase with duration of HIV infection,^10,11 predicting rising rates of neurocognitive complications in an aging HIV-infected population.

Measuring CNS Penetration

The imperfect correlation between suppression of HIV replication in the CNS relative to the plasma is attributed to unequal CNS penetration among different agents between and within antiretroviral drug classes. Due to the expected correlation between drug concentration and HIV suppression, CNS penetration may be a critical risk factor for the development of CNS complications, including neurocognitive impairment. Although other risk factors, such as pretreatment CNS viral loads, patient susceptibility to CNS complications, and the presence of comorbid infections, such as hepatitis C virus, may also influence the risk of CNS complications, quantification of CNS penetration is an important step toward predicting control of HIV in the CNS sanctuary and, potentially, for reducing the risk of CNS complications.

Several techniques for quantifying CNS penetration have been proposed. These include evaluating chemical properties important for crossing the blood-brain barrier (BBB), such as protein binding, lipophilicity and ionization; quantifying drug concentrations in the cerebrospinal fluid (CSF); and evaluating antiretroviral efficacy against CNS symptoms in comparative clinical studies. Since the information provided by these measures is not always consistent, CNS drug penetration was recently ranked using all three of these approaches.

Within drug classes, such as nucleoside reverse transcriptase inhibitors (NRTIs), non-NRTIs (NNRTIs) and protease inhibitors (PIs), a score of 0 was assigned for drugs with low penetration, 0.5 for drugs with intermediate penetration and 1 for drugs with high penetration. In the NRTI classes, examples of drugs with low penetration are tenofovir, didanosine and zalcitabine. Emtricitabine, lamivudine and stavudine demonstrate intermediate penetration while abacavir (ABC) and zidovudine are associated with high penetration. Among NNRTIs, no drug had low penetration, efavirenz had intermediate penetration, and delavirdine and nevirapine were estimated to have high penetration. The majority of PIs have low penetration. Exceptions include indinavir, amprenavir/ritonavir and atazanavir (with or without ritonavir), which show intermediate penetration, and lopinavir/ritonavir and indinavir/ritonavir which exhibit high penetration.

The clinical relevance of these rankings was tested in 467 HIV-infected patients who underwent venipuncture and lumbar puncture to permit quantification HIV in both plasma and CNS.¹² Based on cumulative penetration scores for the combination of antiretrovirals patients were taking, there was a strong correlation between a low penetration rank and detectable viral load in the CNS. Specifically, the odds of having a detectable CNS viral load were increased by threefold in those patients taking an antiretroviral regimen with a penetration rank of 1.5 or less. In contrast, the CNS penetration rank had no effect on likelihood of plasma HIV suppression.

The CSF Factor

Retrospective studies have already identified that improving control of HIV in the CNS can benefit neurological outcomes,¹³ but prospective, controlled studies are needed to confirm this benefit and to address other important questions relevant to day-to-day management. For example, more information is needed on relative risk of neurocognitive complications and whether it is useful to measure HIV RNA routinely in the CSF. It is not yet clear whether all HIV-infected individuals are at risk for neurocognitive decline or if additional risk factors can be identified to single out those individuals at greatest need for selecting antiretroviral regimens with high CSF penetration.

Several recent clinical studies have had important implications for treatment. One example is a multicentre collaborative study that evaluated HIV RNA levels in 112 patients with HIV infection, of whom 90% were taking CART.¹⁴ This study, like previous studies, found a potentially important disconnect between plasma HIV load and HIV load in the CSF, but it also demonstrated that HIV viral loads in the CSF did not correspond with immune function as measured by CD4+ cell levels. In this study, the NRTIs previously associated with good CNS penetration, ABC and zidovudine, were associated with lower CSF HIV loads than tenofovir, the NRTI with a relatively low penetration.

In addition, this study found that of the 40 subjects who had HIV RNA levels <50 copies/mL in both the plasma and the CSF, 17 (42%) had measurable HIV when more sensitive assays for HIV were employed. This raises concern that even patients with undetectable HIV in CSF by conventional assays may still have ongoing HIV replication that poses a risk for neurological damage.

The potential for ongoing progression of HIV-induced damage to the CNS, despite good HIV control by conventional viral load measures in the plasma, suggests the need for controlled studies to evaluate treatment goals specifically in the CNS sanctuary. Ultimately, the concentration of HIV that injures the brain may be different than the concentration which impairs immune function or which poses a threat for the emergence of resistance mutations.

In evaluating the clinical significance of employing antiretroviral drugs that penetrate the CNS, it is important to consider goals, such as treatment vs. prevention of neurocognitive decline. In one study of patients with immune recovery after advanced AIDS, the prevalence of neuropsychological impairment declined from 20% prior to initiation of effective HAART to 14% at 48 weeks and 12% at 96 weeks, suggesting some degree of reversibility.¹⁵ The authors of this study correlated the improvement on neuropsychological function with improving control of plasma viral load but not with immune function as measured with CD4+ cell levels. However, a more recent study conducted over a shorter timeframe found that nearly half of CART-treated individuals who had neuropsychological impairment did not return to normal performance after 24 weeks and those least likely to improve neuropsychological performance had the lowest CD4+ nadirs.¹⁶

Tools to Detect Neurocognitive Decline

If neurocognitive impairment is not fully reversible in all patients, the ultimate goal will be prevention; however, but demonstrating protection from neurocognitive impairment with antiretroviral regimens that offer good CNS penetration over those that offer less CNS penetration will require long follow-up with sufficient numbers of patients to compensate for other variables, including relative susceptibility to CNS decline, that might affect outcomes. Such analyses have important potential clinical significance given the important risk of dementia in aging HIV infected patients, but definitive answers about whether drug choice can affect relative risk are not imminent.

From the practical standpoint, it is reasonable to evaluate and follow neurocognitive performance in HIV-infected patients. Relatively brief neuropsychological assessment tools can be used to quantify performance as a baseline measure and over time. In patients demonstrating neurocognitive impairment prior to initiation of CART, the benefits of selecting CNS-penetrating antiretroviral agents have been incompletely documented but remain a practical approach. For patients already on treatment who have demonstrated a change in cognitive performance, the decisions are more difficult. Clinicians are justifiably reluctant to alter a therapeutic regimen that is providing good control of HIV in blood. Yet this step may be an important consideration when clinical neurocognitive deterioration is significant or progressive. Another reasonable alternative is to add a CNS-penetrating agent to the existing regimen.

Based on recent studies, measuring viral load in the CNS may eventually become a routine part of HIV management.10 Given the disparities between plasma and CSF viral loads, this step may already influence treatment in patients with symptoms of neurocognitive impairment, but CSF viral load assays are not routinely reimbursed, the risk-to-benefit ratio has not been established, and the value of this measure may improve as more information accumulates about the goal for HIV RNA load in the CSF, which may be substantially lower than the plasma target of <50 copies/mL.⁶

For the future, other methods of patient assessment, particularly neuroimaging, may have value for assessing neurocognitive status and risk. Despite substantial ongoing work in this area, neurological damage due to HIV can be difficult to distinguish from other conditions such as infection or tumour, and diagnostic neuroradiological guidelines for HIV-associated dementia have not been published. Quantification of abnormalities of white matter or other objective measures of neurological involvement are strategies that are being pursued to improve understanding of the pathogenic effect of HIV on neural structures over time. For example, such studies may be helpful in understanding whether there is a long latent period of subclinical damage before the cumulative effects produce neurocognitive impairment or to monitor the neurological response to CART. It may also help to evaluate relative susceptibility to neurological damage in relation to patient age. While the increased incidence of neurocognitive decline in older patients often appears to be a complication of chronic HIV infection, susceptibility to dementia also appears to be important to older individuals newly infected with HIV.¹⁷

Biomarkers are another potential tool for evaluating the presence of neurocognitive decline in patients with early or ambiguous symptoms. A variety of biomarkers are being evaluated, including those that signal astrocyte activation, such as MCP-1, macrophage activation, such as beta-2 microglobulin, and neuronal injury, such as neurofilament proteins.¹⁸ Early detection of neurologic injury is a particularly important approach to controlling HIV-related neurocognitive impairment if, as suspected, the opportunity to reverse damage is limited.

Dementia was a well recognized complication of AIDS prior to the introduction of CART, but the interest in understanding the neurocognitive impairments associated with otherwise well controlled HIV infection is relatively recent. There has been substantial recent progress. The evidence that there is a disparity between control of HIV in the CNS and the plasma has encouraged a closer look at whether the differences in the relative CNS penetration of antiretroviral drugs provide an opportunity to treat or prevent HIV-related neurocognitive decline.

Summary

In the aging HIV-infected population, neurocognitive decline is emerging as an important health threat that persists in patients on stable and effective antiretroviral regimens. There is a variety of evidence suggesting that the risk of neurocognitive complications increases over the duration of HIV infection. Preliminary clinical studies indicate that substituting antiretroviral agents with good CNS penetration may reduce the risk of neurocognitive decline relative to those with less penetration. Additional work is needed to determine whether early use of antiretroviral agents with a high CNS penetration can reduce the risk of neurocognitive decline. Ongoing clinical studies are expected to provide specific guidance for modifying this risk.

References

1. Palella et al. J Acquir Immune Defic Syndr 2006;43:27-34.

2. McArthur et al. J Neurovirol 2003;9:205-21.

3. Sevigny et al. Neurology 2004;63:2084-90.

4. Robertson et al. AIDS 2007;21:1915-21.

5. Reddy et al. AIDS Res Hum Retrovir 2003;19:167-76.

6. Letendre et al. 16th Conference on Retroviruses and Opportunistic Infections (CROI) 2009, Abstract 484b.

7. Langford et al. Brain Pathol 2003;13:195-210.

8. Heaton et al. Conference on Retroviruses and Opportunistic Infections (CROI) 2009, Abstract 154.

9. Robertson et al. AIDS 2007; 21:1915–21.

10. McArthur J. J Neuroimmunol 2004;157:3-10.

11. Valcour et al. Neurology 2004;63:822-7.

12. Letendre et al. Arch Neurol 2008;65:65-70.

13. Letendre et al. Ann Neurol 2004;56:416-23.

14. McClernon et al. European AIDS Conference 2007, Abstract P6.11/01.

15. McCutchan et al. Conference on Retroviruses and Opportunistic Infections (CROI) 2004, Abstract 498.

16. Letendre et al. Conference on Retroviruses and Opportunisitic Infections (CROI) 2008, Abstract 68.

17. Corder et al. Nature Med 1998;4:1182-4.

18. Brew BJ, Letendre SL. Int Rev Psychiatry 2008;20:73–88.