Studies confirm that drugs that penetrate brain control HIV better and improve symptoms of brain impairment

This article is more than 11 years old. Click here for more recent articles on this topic

A number of different studies presented at the 17th Conference on Retroviruses and Opportunistic Infections (CROI) confirmed that antiretrovirals (ARVs) that penetrate the blood-brain barrier more fully are better at suppressing HIV replication in the cerebrospinal fluid (CSF) and help to improve symptoms of neurological impairment.

It is clear from previous studies that being on suppressive HAART (highly active antiretroviral therapy) improves neurocognitive symptoms but it has not been clear whether drugs with better brain penetration produce greater improvements in measures of cognitive function: see this report for a summary of some previous findings. The consensus of studies presented at this conference was that they do.

A number of studies made use of what is called the CNS Penetration Effectiveness (CPE) scoring system, a way of ranking ARVs for the levels they reach in the CSF. The total CPE score is achieved by adding up the scores for the individual drugs the patient is on.


cerebrospinal fluid (CSF)

The liquid surrounding the brain and spinal cord.


The fluid portion of the blood.

multivariate analysis

An extension of multivariable analysis that is used to model two or more outcomes at the same time.

detectable viral load

When viral load is detectable, this indicates that HIV is replicating in the body. If the person is taking HIV treatment but their viral load is detectable, the treatment is not working properly. There may still be a risk of HIV transmission to sexual partners.

virological suppression

Halting of the function or replication of a virus. In HIV, optimal viral suppression is measured as the reduction of viral load (HIV RNA) to undetectable levels and is the goal of antiretroviral therapy.

A new CPE system was issued this year, adding some new drugs and giving drugs a penetration score from one (poor) to four (best), where there had previously only been three ranks. The two commonly used drugs in the highest rank are nevirapine (Viramune) and AZT (zidovudine, Retrovir, and the most commonly used drugs in the lowest rank are tenofovir (Viread) and boosted saquinavir (Invirase. Some of the studies presented at CROI used the old system and some the new, but results are broadly similar.

Findings from CHARTER

Scott Letendre, principal investigator in the largest study investigating neurocognitive impairment, the CHARTER study, looked at factors associated with a detectable CSF viral load in a cross-sectional study comparing viral loads in 1221 simultaneously taken CSF and blood samples.

Thirty-one per cent of the patients were not taking HIV therapy. Three-quarters of these patients had detectable CSF viral loads compared with 16% of those taking HAART.

In multivariate analysis the factors significantly associated with CSF viral load on HAART were high plasma viral load, white ethnicity, non-adherence, and lower CPE score. Twenty per cent of white patients had detectable viral loads compared with 9% of non-white patients. Some of the association with white ethnicity was explained by age, but not all.

Factors associated with a higher CSF viral load off HAART included viral load and CD4 count as well as older age, and a trend towards association with white ethnicity and male sex. In multivariate analysis only blood plasma viral load and older age remained significant.

Detectable CSF viral load (over 50 copies/ml) was not associated with poor performance on neuro-psychological (NP) tests, but having a CSF viral load higher than plasma viral load was. In this study 15% of subjects off HAART, and 4% of subjects on HAART had a higher CSF viral load than they did in plasma. Mean NP test scores were not significantly different with these patients, but in patients off treatment the proportion with very poor scores was.

Letendre commented that in another study the same had been shown of patients on HAART when a more sensitive viral load test was used that could detect CSF viraemia down to two copies/ml.

In a second presentation, Letendre looked at factors associated with treatment failure in the CHARTER study. Patients who initially had undetectable viral loads in plasma (225 patients) or CSF (346 patients – many were suppressed in both plasma and CSF) were followed and time to loss of virological response (TLOVR) was measured. Patients were followed-up for an average time of 18 months for plasma testing and 20 months for CSF testing, unless they lost viral suppression first.

During the study 82 patients (36%) lost viral suppression in plasma, with a mean TLOVR of eleven months and 67 (19%) lost viral suppression in CSF, with a mean TLOVR of nine months.

The factors associated at baseline with loss of virological response in plasma were a CD4 count under 200 cells/mm3, black ethnicity, and neuropsychological impairment. These were also associated with loss of suppression in CSF but so were age under 45 and taking a protease inhibitor-based regimen. A regimen with a low CPE score was only associated with loss of response in patients under 45, and patients over 45 with a high CPE score maintained a better virological response for up to five years. Multivariate analysis did not alter the significance of these predictors.

Two more factors, a baseline plasma viral load over 50 copies/ml and non-use of antidepressants, were significantly associated with loss of response in CSF in univariate analysis but were not in multivariate analysis.

Other studies

One more study looked at loss of viral response in CSF. A team from Sweden looked at 63 patients with plasma viral loads under 50 copies/ml and found that seven (11%) had detectable HIV in their CSF.

Five of these were on an efavirenz-based regimen out of 24 taking the drug (21%) and two on an atazanavir-based regimen out of 13 taking it (15%); in terms of their nucleoside backbone regimen three out of 17 on abacavir failed (18%) and four out of 27 on tenofovir (15%). None of the patients taking boosted lopinavir (Kaletra) or AZT had CSF virological failure. The difference in the proportion failing on efavirenz and the lack of failure on lopinavir was statistically significant (p = 0.02).

Two more studies looked at the effect of ARVs on psychological performance.

In the first, psychologists from Cotugno Hospital in Italy administered a battery of psychological tests to 45 patients without being told what ARV regimen the patients were taking. All patients had to have viral loads under 50 copies/ml in plasma and to have been on the same HAART regimen for at least six months. Patients at high risk of impairment were excluded. These exclusions were age over 60, poor education, diagnosed depression or other psychiatric or neurological conditions, and drug use.

Thirty of the patients were on regimens with a high CPE score and the remaining 15 on ones with a low CPE score. Patients with a high CPE score showed better performance in language performance, namely tests measuring verbal memory, verbal fluency, and ability to manipulate symbols. Other psychological domains were unaffected.

In the second study, a team from the University of New South Wales conducted a substudy of a larger study called ALTAIR, which compared cerebral function and physiological indicators of neuronal damage in patients on a randomised controlled study of three different regimens: tenofovir/FTC (Truvada) plus either efavirenz, boosted atazanavir, or an unconventional quadruple-NRTI regimen of Truvada plus abacavir and AZT. The CPE scores of these three regimens are 7, 6 and 11 respectively.

Patients taking the quadruple-NRTI regimen had an improved psychological test performance over 48 weeks whereas patients on the efavirenz-based regimen showed more signs of recovery from neuronal damage.

HIV drugs and CNS toxicity

Finally, one in vitro study reminded us that some HIV drugs – in particular the NRTI drugs and efavirenz – are associated with neurotoxicity themselves, with efavirenz causing psychological symptoms and the NRTIs peripheral neuropathy.

A team from the University of North Carolina cultured rat cortical neurones in solutions of various HIV drugs and found that four drugs – ddI, FTC, tenofovir and AZT – caused neurone damage, with a maximum loss of cells ranging from 32% to 52%. Damage to the mitochondria, the energy-producing components of cells, was observed with all of these drugs except ddI. The widely used combination of tenofovir/FTC/efavirenz (Atripla) produced an amplified, but slowed, nerve response in these neurones.

These are laboratory experiments with as yet unknown clinical significance, but the fact that some of the drugs that best penetrate the brain are also amongst those most toxic to nerve cells may explain the paradoxical effects seen in some studies of the effect of HAART on neuropsychological performance.

Further information

You can view abstract 172, abstract 430, abstract 432, abstract 433, abstract 434, and abstract 435 on the official conference website.

You can also view a webcast and slides of this session on the official conference website.


Letendre S et al. Correlates of CSF viral loads in 1221 volunteers in the CHARTER Cohort. Seventeenth Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 172, 2010.

Letendre S et al. Correlates of time-to-loss-of-viral-response in CSF and plasma in the CHARTER Cohort. Seventeenth Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 430, 2010.

Edén A et al. CSF escape is uncommon in HIV-1 infected patients on stable ART. Seventeenth Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 432, 2010.

Starace F et al. Neuropsychological performance is better in HIV-infected subjects treated with neuroactive HAART. Seventeenth CROI Conference, San Francisco. Abstract 433. 2010.

Winston A et al (presenter Puls R). cART alters changes in cerebral function testing after 48 weeks in treatment-naive, HIV-1-infected subject commencing ART. Seventeenth Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 434, 2010.

Liner J et al (presenter Robertson K). CNS toxicity of antiretroviral drugs. Seventeenth Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 435, 2010.