HIV brain impairment: who gets it, and why?

A number of papers at the 19th Conference on Retroviruses and Opportunistic Infections presented advances in research on HIV-related brain impairment and neurocognitive problems.

A couple of problems have dogged research into HIV-related brain impairment. One is how to differentiate between degrees of severity. Psychological tests can detect barely noticeable slowing of performance but the high prevalence of this asymptomatic neurocognitive impairment (ANI) overestimates the proportion of people with HIV who have real difficulty with working or daily life. We also do not know to what extent ANI is a predictor of the development of significant impairment, because some people improve, especially with antiretroviral treatment (ART).

The other problem is how to link neurological findings uncovered by brain scanning with the results of psychological tests: some studies, while uncovering both changes in cognitive performance and evidence of inflammation in the brain in HIV-positive people, have found little statistical linkage between the two.

The cause of brain impairment also remains uncertain: it is thought to be to do with HIV-induced inflammation of the microglia, the cells that function as the brain’s immune system, but exact causes remain elusive and recent studies have found that diabetes and fat accumulation are risk factors too. Finally, it remains unknown to what extent HIV-positive people on ART will still be more likely than HIV-negative people to develop dementia.

ANI is very common: in the largest ongoing study of HIV and neurocognitive problems, CHARTER, 52% of patients were found to have some degree of brain impairment, but the majority was asymptomatic.

Does asymptomatic impairment matter?

One study from CHARTER presented at CROI (Heaton) introduced new criteria to distinguish between ANI and mild neurocognitive disorder (MND), which is impairment serious enough to interfere with everyday living. In addition to the psychological tests that diagnose ANI, the diagnosis of MND is based upon four additional tests: the patient’s own assessment of their mental abilities and of their ability to perform household tasks, and two assessments of work- and medication-based practical tasks – the latter important because one of the most worrying implications of brain impairment in people with HIV is the impact it may have on adherence.

The study looked at 387 CHARTER patients and compared test results taken three years apart. At baseline, 64% were defined as having normal functioning, 22% ANI and 15% MND.

People with either ANI or MND were more likely to be female and non-white and people found to have ANI were significantly more highly educated than patients either with normal functioning or MND – was this because they were better able to prevent slowness in performance turning into practical difficulties? There was little else to distinguish between people with ANI and MND except that the latter were much more likely to have co-morbidities, meaning anything from heart disease to hepatitis C.

At follow-up, both the ANI and MND groups had similar declines in cognitive function (an 11% and 13% decline in test scores respectively) but the normal-functioning group had no decline. The CHARTER researchers comment that “It has been suggested the ANI diagnosis is not valid, does not predict clinical outcomes, and may be a statistical artefact” but that their studies find that “ANI and MND are significant and comparable risk factors for cognitive worsening over 18 to 42 months.”

Early infection and brain impairment

A pair of papers from researchers from Yale, UCSF and North Carolina Universities contributed complementary perspectives on neurocognitive processes in people in early HIV infection. They found there is some evidence of brain damage from early infection and, in some domains at least, this slowly worsens over time in people not taking ART but can be reversed by antiretroviral therapy.

One study (Peterson) looked at 70 HIV-positive men who had been infected for an average of four months and gave them the usual psychological tests at that point and regularly, on average, for the next fifteen months. It found that overall functioning and in particular motor skills were poorer than in the general population, though not by much: 20% of the patient group had scores more than one standard deviation below the mean, when in the general population 16% would be expected.

Motor co-ordination and global functioning continued to get slowly poorer in men not taking HIV therapy during follow-up, though even off ART performance in other domains like memory improved, suggesting that primary infection may cause brain damage in itself, as it does to the immune cells of the gut. Motor performance stabilised in patients on ART and other domains markedly improved. However, the researchers warn that “in HIV-uninfected subjects, repeated testing might be expected to associate with improved performance due to practice effects. These findings suggest that some neurological dysfunction accrues prior to starting ART and may persist in the setting of treatment, with implications for the impact of earlier initiation of ART.”

This study was paralleled by one that put people from the same cohort – though not the same people – into an MRI scanner (Young). It measured the ratios of different metabolites (brain chemicals) in 53 recently infected HIV-positive patients, 18 chronically infected HIV subjects and 19 HIV-negative ones. It found a slow increase over 2.5 years post-infection in the ratio of choline and myoinositol, two markers of brain inflammation, to other brain chemicals in the HIV-positive subjects off treatment. Choline in the brain is evidence of invasion by scavenging white cells (macrophages) from the rest of the body while myoinositol is evidence of inflammation in the microglia. This increase stopped as soon as subjects were put on ART.

It found increases in inflammation especially in the frontal white matter, the ‘wiring’ that connects together the nerve cells used in the frontal lobes, which are where the brain does its highest-level processing: decision-making, abstract thinking and understanding other people.

Brain inflammation on long-term ART

Not all people on ART improve and there is evidence of ongoing inflammation in some parts of the brain in people on ART. In a study by Imperial College in London (Garvey) seven people with chronic HIV infection (average time from infection nine years) and undetectable viral loads, who had no neurological symptoms, were compared with nine age- and ethnicity-matched HIV-negative controls by putting them into a PET scanner; this detects a protein called PK11195, which specifically attaches to inflamed microglia. The positive and negative subjects were also given the same battery of psychological tests as above.

In this study, scans found, on average, significantly greater signs of inflammation in a number of brain areas in the HIV-positive cases versus the HIV-negative controls, but specifically in the corpus callosum, which is the bridge of nerve fibres connecting the two halves of the brain, the temporal and frontal lobes, and an area called the cingulate cortex. The temporal lobe is part of the system for laying down memories and the cingulate cortex is thought to be one of the seats of consciousness: it connects the older, emotional parts of the brain with the higher-order, thinking ones and governs the faculty called saliency, which is the ability to notice when objects, thoughts or events are of particular significance and need attention. These higher levels of inflammation were correlated with deficits in executive function, the ability to juggle multiple inputs and prioritise, though not with other psychological domains.

Cardiovascular risk and brain impairment

Finally, in addition to a poster (McCutchan) re-running the CHARTER findings that diabetes and waist circumference are strongly linked to neocoginitive decline, an Italian study (Colafigli) added more weight to the evidence that cardiovascular disease is a risk factor for brain impairment.

In this study, 269 people with HIV, 94% of them on ART, Were given the usual psychological tests and also had a number of risks for cardiovascular health measured, including tests for intima media thickness (IMT, a measurement of the fatty plaque on arterial walls), and high-density cholesterol (HDL-C or ‘good’ cholesterol). They then had the same tests done a year later.

At baseline, 36% of subjects had some degree of neurocognitive impairment, 33% had high IMT measurements and 35% had a very high risk, according to the Framingham risk equation, of having a cardiovascular event in the next ten years. A year later, 12% no longer had cognitive impairment but 8% now had it who did not have it at baseline. Globally, cognitive scores had actually gone up but long-term memory recall scores had gone down. During the year there had been various changes, good and bad, to the patients’ cardiovascular risk. On the one hand 7% had given up smoking and the group had gained, on average, 67 more CD4 cells per mm³, but their average HDL cholesterol had also fallen. After adjustment for CD4 count, age and some other factors the only factor significantly associated with worse neurocognitive performance was HDL-cholesterol.

What should we make of all this? At a press conference Dr John Mellors, the CROI conference co-chair, commented: “My fear is that in ten years’ time we’ll be able to eliminate HIV in the rest of the body but will see an epidemic of dementia because our new drugs won’t get to the HIV in the microglia.” We are getting a clearer picture of what is going on in the brain but need drugs that can penetrate the blood-brain barrier and dampen down the hypersensitive immune responses in the brain and central nervous system.