Back HIV-Related Conditions Neurocognitive CROI 2010: Risk Factors for Neurocognitive Impairment among People with HIV

CROI 2010: Risk Factors for Neurocognitive Impairment among People with HIV


People with higher blood and cerebrospinal fluid viral load and lower CD4 cell count are more likely to develop neurocognitive impairment, according to studies presented in February at the 17th Conference on Retroviruses and Opportunistic Infections (CROI 2010) in San Francisco. In addition, researchers with the SMART study reported that cardiovascular risk factors predict poorer neurocognitive performance. But conflicting findings emphasize the need for further research. 

As people with HIV live longer thanks to effective ART, neurocognitive impairment continues to be a concern even among those with suppressed blood plasma viral load. The specific risk factors and mechanisms underlying cognitive problems in HIV positive people are not fully understood and studies continue to produce conflicting data.

CSF Viral Load

Scott Letendre and colleagues (abstract 172) presented the latest data from CHARTER (CNS HIV Antiretroviral Therapy Effects Research), an ongoing multi-center observational study of neurocognitive function in people with HIV.

In a cross-sectional analysis of 1221 participants, the researchers measured HIV viral load in the cerebrospinal fluid (CSF) surrounding the brain and spinal cord, using an assay with a lower limit of quantitation of 50 copies/mL.

About three-quarters of the participants were men, the median age was 43 years, about two-thirds were on ART and the rest were treatment-naive. The median CD4 count was about 415 cells/mm3, but the nadir (lowest-ever) level was 175 cells/mm3 and about 60% had a prior AIDS diagnosis.

The median CSF viral load was 50 copies/mL, with 34% of test values showing detectable HIV RNA above this level. CSF viral load was strongly influenced by ART: 76% of 379 untreated patients had detectable CSF viral load, compared with 16% of 842 people on ART. Furthermore, 13% of untreated patients had CSF viral load equal to or higher than plasma viral load.

In a multivariate analysis of untreated patients, higher CSF viral load was associated with higher blood plasma viral load and older age. Among patients on ART, detectable CSF viral load was significantly associated with higher blood viral load, white race/ethnicity, poor ART adherence, and lower CPE score, a measure of how well an individual's antiretroviral drugs penetrate the protective blood-brain barrier.

Among participants with a regimen CPE score of 3 or lower, 39% had detectable CSF viral load, compared to 18% of those with a score of 6, and 15% of those with a score of 9 or higher. Overall, higher CSF viral load did not predict worse neurocognitive performance, but people with CSF viral load equal to or exceeding plasma viral load showed impairment.

Plasma viral load was the strongest correlate of CSF viral load, the researchers concluded, emphasizing the importance of systemic HIV suppression for control of HIV in the nervous system. They added that the mechanisms by which age and race/ethnicity influence CSF viral load are unknown (although older white men on average had a longer duration of HIV infection and more advanced disease).

CD4 Count

Ronald Ellis and colleagues (abstract 429) also presented data from the CHARTER cohort. This retrospective analysis included 1526 participants; demographic and HIV-related characteristics were similar to those described above, though a larger proportion (about 70%) were on ART.

Study participants received comprehensive medical and neuropsychological evaluations. HIV-associated neurocognitive disorder (HAND) was diagnosed according to published criteria. Participants were classified as being either neurocognitively impaired or unimpaired.

About 52% of participants were classified as having neurocognitive impairment. Those with more severe impairment were more likely to have other co-existing conditions. Among those without other contributing conditions, a majority (71%) had asymptomatic impairment; fewer than 5% had HIV dementia.

People with lower nadir CD4 cell levels had a higher likelihood of neurocognitive impairment, while those whose CD4 count never fell below 350 cells/mm3 were at lower risk (odds ratio 0.62 vs CD4 nadir < 50 cells/mm3). The same association was seen in an analysis restricted to people with HIV RNA < 50 copies/mL.

Nadir CD4 count was the only significant predictor of cognitive impairment, and it remained so after adjusting for potential confounding factors including plasma viral load, age, sex, race/ethnicity, and duration of HIV infection. Current CD4 cell count, in contrast, was not significantly associated with impaired neurocognitive function.

These findings underline the benefits of earlier ART initiation in reducing the risk of neurocognitive impairment. Preventing impairment is important, the researchers noted, since once HAND occurs, many individuals remain impaired despite effective viral suppression on ART.

Cardiovascular Risk Factors

Finally, Edwina Wright and fellow investigators with the INSIGHT SMART Study Group (abstract 415) presented findings from a neurology sub-study of the large SMART treatment interruption trial.

Cardiovascular disease and cardiovascular risk factors are associated with neurocognitive impairment in the HIV negative general population, and the researchers sought to determine whether this was true for HIV positive people as well, given that people with HIV have higher rates of both cardiovascular disease and neurocognitive impairment.

Briefly, the SMART trial enrolled more than 5000 HIV positive adults starting in 2000. Participants entered the study with a CD4 count above 350 cells/mm3, and were randomly assigned either to stay on continuous antiretroviral therapy (ART) or to suspend treatment when their CD4 cell count was above this level, resuming when it fell below 250 cells/mm3.

The treatment interruption arm was discontinued ahead of schedule in January 2006 after a preliminary analysis showed that participants in that group had higher rates of death and both AIDS-defining opportunistic illness and non-AIDS conditions including cardiovascular, liver, and kidney disease.

The sub-study included 292 participants (102 from the U.S. and Australia; 45 from Brazil, and 145 from Thailand). About 60% were men, the median age was 40 years, 92% were on ART, the median current CD4 count was 536 cells/mm3, and the median nadir CD4 cell level was 225 cells/mm3. Only 3% had prior cardiovascular disease, but about 10% were taking medications to manage elevated lipids and a similar percentage to control high blood pressure.

Participants completed a set of 5 neurocognitive tests at baseline and after 6 month, yielding a combined adjusted z-score dubbed the QNPZ-5 score. The mean QNPZ-5 score was -0.7, indicating below average neurocognitive performance compared with the general population; 14% had more severe neurocognitive impairment.

Prior cardiovascular disease significantly increased the likelihood of cognitive impairment (odds ratio 6.1, or about a 6-fold high risk compared to those without cardiovascular disease). Elevated total cholesterol and use of blood pressure-lowering drugs predicted neurocognitive impairment to a lesser but still significant extent. Baseline cardiovascular risk factors were not significantly associated with changes in QNPZ-5 scores over 6 months, but the researchers suggested this might be due to the short follow-up time.

In this study -- unlike the CHARTER analyses -- HIV-related factors including current CD4 count, nadir CD4 cell level, viral load suppression, prior AIDS diagnosis, and ART regimen CPE score were not statistically significant predictors of impaired cognitive function.

Based on these findings, the SMART investigators suggested, "in HIV-infected persons with high CD4 cell counts, cardiovascular risk factors and disease may be more detrimental to neurological functioning than those more directly attributed to HIV itself."

The mechanisms by which cardiovascular risk factors contribute to HIV risk are not well understood, and these conflicting findings emphasize the need for further research in this area.

Abstract 172: Univ of California, San Diego, CA; Washington Univ, St Louis, MO; Univ of Washington, Seattle, WA; Univ of Texas Med Branch, Galveston, TX; Johns Hopkins Univ, Baltimore, MD.

Abstract 429: Univ of California, San Diego, CA; Washington Univ, St Louis, MO; Univ of Texas Med Branch, Galveston, TX; Mt Sinai Sch of Med, New York, NY.

Abstract 415: Monash Univ, Alfred Hosp and Burnet Inst, Melbourne, Australia; Univ of Minnesota, Minneapolis, MN; Univ of North Carolina at Chapel Hill, NC; St Vincent's Hosp and Univ of New South Wales, Sydney, Australia; Natl Ctr in HIV Epi and Clin Res, Univ of New South Wales, Sydney, Australia; JICA-NIH Project, Ministry of Publ Hlth, Thailand; Denver Public Health, Denver, CO; Inst de Infectologia Emilio Ribas, Sao Paulo, Brazil; Univ of California, San Francisco, CA.



S Letendre, C FitzSimons, R Ellis, and others (CHARTER Group). Correlates of CSF Viral Loads in 1221 Volunteers of the CHARTER Cohort. 17th Conference on Retroviruses & Opportunistic Infections (CROI 2010). San Francisco. February 16-19, 2010. Abstract 172.

R Ellis, R Heaton, S Letendre, and others (CHARTER Group). Higher CD4 Nadir Is Associated with Reduced Rates of HIV-associated Neurocognitive Disorders in the CHARTER Study: Potential Implications for Early Treatment Initiation. 17th Conference on Retroviruses & Opportunistic Infections (CROI 2010). San Francisco. February 16-19, 2010. Abstract 429.

E Wright, B Grund, K Robertson, and others (INSIGHT SMART Study Group). CVD and CVD Risk Factors Are Associated with Lower Baseline Neurocognitive Performance in the SMART Neurology Substudy. 17th Conference on Retroviruses & Opportunistic Infections (CROI 2010). San Francisco. February 16-19, 2010. Abstract 415.