The way researchers assess and identify HIV-associated neurocognitive disorders is not fit for purpose, a viewpoint article in Clinical Infectious Diseases suggests. Studies typically find that around half of people with HIV meet the criteria for a disorder, which does not fit with the modern clinical experience that relatively few people living with HIV have cognitive impairment affecting their daily lives.
The authors argue that future reports of cognitive impairment should focus on people with symptoms and more clearly describe the contribution of HIV brain disease. Any new criteria need to be appropriate for use in diverse settings, say the group of experts from universities in South Africa, Sweden, Zambia, the United States and the United Kingdom.
Background
In the absence of antiretroviral treatment, significant numbers of people with HIV may experience severe cognitive impairment, caused by longstanding HIV infection’s damage to the brain. This is known as HIV-associated dementia and mostly affects people with a very low CD4 count and an AIDS diagnosis. It typically causes problems with concentration, motor skills (e.g. difficulty walking or doing up buttons) and depressive feelings.
With the availability of antiretrovirals, HIV-associated dementia became less common and there was concern that more subtle cognitive problems might be missed. In 2007, a group of experts developed the ‘Frascati criteria’ to define the broader concept of HIV-associated neurocognitive disorder (HAND).
HAND is based on the results of cognitive tests. They assess attention and working memory; language and verbal fluency; executive function and abstraction; speed of processing; memory (learning and recall); and sensory-perceptual and motor skills. Scores are compared with those in a control population, matched for age, sex, ethnicity and education.
People who have low scores are described as having HIV-associated neurocognitive disorder (HAND) if the impairment cannot be explained by other conditions, such as head trauma, substance abuse, vascular dementia or Alzheimer disease. There are three levels within HAND: asymptomatic neurocognitive impairment (a low score + no symptoms); mild neurocognitive disorder (a low score + problems observed in daily life); and HIV-associated dementia (a much lower score + more severe problems observed).
Last year, a review of 123 studies found that 43% of people with HIV included in the studies had been classified as having HAND, including 24% with asymptomatic neurocognitive impairment, 13% with mild neurocognitive disorder and 5% with HIV-associated dementia.
However the authors say this does not reflect the reality seen in clinics in recent years, where people present with cognitive disorders relatively infrequently (for example, 7.5% of HIV-positive patients in six years at a London hospital). Those who do present are very often diagnosed with HIV at a late stage or have substantial co-morbidities.
Moving on from HAND
Dr Sam Nightingale and colleagues highlight a number of issues with the way HAND is defined and suggest a way forward.
Firstly, by definition, HAND is attributed to the direct effect of HIV on the brain. There are multiple mechanisms by which HIV can damage the brain, including the presence of detectable virus in cerebrospinal fluid when it may be undetectable elsewhere and the inflammation this causes.
“A label of HAND assumes that low performance on cognitive tests in a person with HIV is caused by HIV.”
However, there are other reasons why a person with HIV may perform poorly on cognitive tests. Some, like a severe head injury or alcoholism, may be quite clear and would exclude the person from being diagnosed with HAND. Under the Frascati criteria, other health issues, such as depression, diabetes and hepatitis C, are usually described as “contributing” factors, along with HIV disease.
“A label of HAND assumes that low performance on cognitive tests in a person with HIV is caused by HIV, at least in part,” say the authors. “In reality some is entirely caused by HIV, some is due to a combination of HIV and comorbid factors, and in some people HIV brain pathology may not be contributing at all.”
Given that cognitive impairment is often the result of a combination of factors, the authors suggest that a new definition needs to include cognitive impairment from any cause in a person with HIV, rather than attempting to reflect only that which is a direct effect of HIV on the brain. As prompt use of HIV treatment becomes more common, impairment caused by HIV is likely to be less frequent.
Crucially, rather than being based primarily on scores from cognitive testing, they argue that people described as having cognitive impairment should have symptoms. Under the Frascati criteria, a majority of the people who are currently described as having HAND have no symptoms. They are classified as having ‘asymptomatic neurocognitive impairment’ – there has been no discernible impact of their ‘impairment’ in their daily lives. The clinical significance of this is unclear; researchers do not agree on whether this group are more likely to progress to the next stage of mild impairment.
Nightingale and colleagues say that they wish to avoid people with low performance on cognitive tests and no symptoms being labelled as having a cognitive disorder. This group should simply be described as having “low performance on cognitive tests” and not be included in those described as having cognitive impairment.
They believe a diagnosis of cognitive impairment must be based on clinical history. The person with HIV, someone close to them or a healthcare professional needs to have noticed a decline in cognitive function or difficulties with daily activities. At the same time, there should be objective evidence of impairment from tests, for example showing decline over time.
Problems with cognitive tests
One of the issues with relying on cognitive testing to diagnose people is that results must be compared to something defined as ‘normal’, with scores below a certain cut-off signifying impairment. These cut-offs are somewhat arbitrary and there is a high rate of false-positive results – over 20% of cognitively normal HIV-negative people taking part in studies are defined as impaired.
The Frascati criteria suggest the control population (the group against which people with HIV are compared) should be matched on the basis of age, sex, ethnicity and years of education. But this does not fully take into account all the social and economic factors that affect how well people do on these tests, especially considering that HIV disproportionately affects marginalised groups. Lower test performance is associated with poverty, economic hardship, stress and lower socioeconomic status. One person’s test results may also vary from day to day, due to mood, lack of sleep, pain and other transient issues.
Culture affects test performance but most available tests were developed in North America. Few have been culturally adapted and translated for use in countries with a high prevalence of HIV. There is a lack of data on normal results in most resource-limited settings and the results that are available suggest wide variation between countries and between sites in the same country.
As attention shifts away from cognitive tests, researchers and clinicians need to pay closer attention to biomarkers of HIV and inflammation in the cerebrospinal fluid as well as a range of brain scanning techniques. This will help them identify impairments whose root cause is HIV brain disease, be precise about the mechanisms and distinguish them from impairment with other causes.
There are challenges though. There aren’t yet any biomarkers which are validated to identify HIV brain disease or predict its worsening. Taking a sample of cerebrospinal fluid to test for biomarkers requires a lumbar puncture (spinal tap), which can cause headaches for several days afterwards. (Identification of biomarkers in blood would be helpful.) In resource-limited settings, brain scans are even less available than cognitive testing.
Conclusion
“Our assertion that the HAND criteria risk overestimating the extent of cognitive disorders in people with HIV should not be mistaken for a view that we do not believe HIV brain pathology and cognitive impairment in people with HIV to be important or widespread,” say the authors. The fact that cognitive impairment in people with HIV is multifactorial does not detract from its impact on the individual or the importance of developing interventions to reduce its impact, they add.