The Association of Memory Disorders and Chronic HIV Disease in the Antiretroviral Therapy Era

A Systematic Literature Review

E Ripamonti; M Clerici


HIV Medicine. 2020;21(1):9-20. 

In This Article


In the current literature, there does not seem to be compelling evidence of impairment of either the phonological loop or the visuo-spatial sketchpad of the working memory in HIV infection. However, there is an indication of an association of chronic HIV disease with impairment of the central executive,[12,15] which may be connected to the diminished psychomotor speed frequently reported in these patients.[5] In fMRI paradigms, increased BOLD signal in the frontal and parietal lobes in tasks involving coordination by the central executive has been interpreted in terms of the need for a greater cognitive 'reserve' in patients with HIV infection.[15]

Verbal episodic memory deficit was found in four of the 31 studies.[11,16–18] However, other authors did not report such deficit,[10,19,20] and the actual effect of HIV infection on verbal episodic memory has still to be clarified. Visual episodic memory would be unimpaired,[20,30] although the role of covariates, such as age, and stage of HIV infection is a very open issue. An effect of ageing across neurocognitive domains, including memory, was reported in the Multicenter AIDS Cohort Study,[26] as well as an effect of HIV disease severity. As to the role of the retrograde component,[33] a pattern of temporally graded memory loss, which is a hallmark of Alzheimer's disease, was not found either in patients living with HIV or in those with Huntington's disease. This pattern supports the original hypothesis that cognitive impairment in HIV/AIDS, at least in the final stage, leads to a subcortical type of dementia.[3,4] Nevertheless, conclusions in this regard were hampered by the very small sample size.[33] Episodic memory impairment in HIV infection has a clear neuroanatomical correlate in hippocampus atrophy, which occurs with ageing.[47,48] People living with HIV have also been reported to have a deficit in prospective memory, which is used to keep track of future events, such as appointments and meetings.[34,36] It still has to be ascertained, however, whether this impairment relates to memory function or is confounded by the effect of impaired executive functions.

There is currently insufficient evidence to determine whether there is an association of semantic memory and implicit memory impairment with HIV infection. The former has rarely been tested and evidence from larger cohorts is needed to discriminate a pure semantic memory deficit from impairment of lexical access. As for implicit memory, it is not formally investigated in detail during a routine neurological/neuropsychological examination, but its neural substrate could be damaged in the chronic stage of HIV infection.[39] There is indeed evidence of effects of ageing by HIV disease stage on motor function,[26] whose anatomical reference can be circumscribed in the basal ganglia, which are involved in early HIV infection.[49]

In general, ART allows patients to attain normal CD4 counts, defined as > 500 cells/μL.[7] A low proximal CD4 cell count is still (frequently) observed in clinical practice (especially among patients not treated with ART) and is a consistent predictor of non-AIDS-related comorbidity in HIV-infected patients.[50] Nevertheless, the impact of this covariate on cognitive function is unclear:[51] some studies observed an association between current CD4 count and brain volume,[40,52–57] while others did not.[58–60] In the present review, it did emerge that low current CD4 count is a risk factor for the development of memory disorders,[11,18,23,31] and this variable will play a relatively independent role from that played by other factors such as stage of HIV disease and viral load[23] or dysplipidaemia.[11] However, different results have been reported in other studies[10,13,21,22,28,29,39,43] and, while the heterogeneity reported in the literature might be attributed to methodological and sample selection differences across studies, it seems important to underline that the effect of current CD4 count on memory function should be systematically analysed in connection with the role of chronic inflammation, as originally suggested by McArthur et al..[44]

Nadir CD4 count has been reported to be associated with reduced white matter[61] and with neurocognitive disorders,[62,63] but such correlations have been questioned.[59] From the studies reviewed herein, it has emerged that nadir CD4 count is associated with episodic memory performance[10] and psychomotor speed,[45] but not with working memory performance.[12] From a neuroimaging perspective, nadir CD4 count has been found to correlate with thalamus atrophy and total brain volume,[45] but not with atrophy of the amygdala, corpus callosum or caudate nucleus.[39]

It is important to monitor HIV RNA concentrations in chronic HIV infection, as such measurements can be adopted as a proxy of ART adherence.[64] Moreover, viral load has been reported to be associated with brain atrophy[54] and neurocognitive impairment,[65] although some studies found this not to be the case.[53,59] In the studies included in the present review, an association of HIV viral load with memory disorders was consistently reported[17,20,37] and virological suppression was found to be a protective factor against memory impairment.[45] However, these results were not replicated in patients with clade C HIV infection[21] and the association of HIV viral load with subcortical damage is very much a matter of debate.[9,19,39]

It remains a challenge for future research to ascertain which (sub)components of memory function are associated with HIV infection. It also remains to be established whether cognitive 'reserve' may constitute a protective factor in chronic HIV infection.[66,67] In addition, the putative protective role of cognitive 'reserve' in chronic HIV disease could be investigated in conjunction with the impact of inflammation (being a lentivirus, HIV infiltrates the central nervous system in the early phase of infection[68] and primarily replicates in the microglia and perivascular macrophages, leading to chronic inflammation[69]), the effectiveness of penetration of antiviral drugs into the central nervous system, success in controlling HIV replication, and ageing of the immune system.

To summarize, results from the contemporary neuropsychological literature are controversial and very heterogeneous. Most of the studies are affected by poor controlling for important covariates, such as ageing, duration of HIV infection, duration of ART and stage of HIV disease; if such covariates were properly taken into account, quantitatively and qualitatively different conclusions might be drawn. We recommend that future studies should be conducted in large cohorts of people living with HIV. Although normative data collected for healthy participants are available for standardized neuropsychological tests, we suggest that a sample of matched control participants should also be routinely tested, using ad hoc criteria to select the matching variables. We also suggest the adoption in data analysis of multivariable statistical models, adjusted for variables such as nadir/current CD4 count, stage of HIV infection and viral load, alcohol/substance use, and the presence of chronic diseases such as diabetes and HCV infection. Our general conclusion is that the question of whether there is an association of HIV infection with memory function impairment is far from being answered. Given the clinical and ecological importance of this problem, we encourage new collaborative research to resolve this issue definitively.