HIV-1-associated neurocognitive disturbances include a debilitating subacute dementing disorder, known as HIV-1-associated dementia or AIDS dementia complex (ADC), and a lesser degree of cognitive impairment, the HIV-associated minor cognitive motor disorder (MCMD), that usually appears earlier in the course of the disease. ADC and MCMD are present in up to a third of HIV-infected patients. The incidence of ADC is only 0.4% during the asymptomatic phase of infection, but in patients with overt AIDS ADC may develop in up to 15%. Markers of risk factors for dementia include high plasma HIV-1 RNA levels, low CD4 cell count, anaemia, low body mass index, older age, injection drug use and female sex[2,3].
Both MCMD and ADC are characterized by neurocognitive impairment affecting attention, concentration, processing speed and motor dysfunction that can be documented by neuropsychological testing. Such measures can be used to quantify the severity of cognitive impairment, define the patterns of involvement and assess the impact of treatment. Although controlled studies of antiretroviral treatment on ADC reported positive effects on neuropsychological measures[4,5], the clinical improvement was not sustained over time. The introduction of highly active antiretroviral therapy (HAART) has resulted in an improvement in the survival rate and a reduction in the incidence and severity of opportunistic infections[7,8], but its impact on the incidence and clinical course of ADC is still unclear. Recent data from a Multicenter AIDS Cohort Study  indicate that the incidence rate for ADC, although showing a downward trend, was 14.7/1000 person year in 1996-1997, thus still representing an important central nervous system (CNS) complication of HIV-1 infection.
Protease inhibitors (PI) of HIV combined with nucleoside reverse transcriptase inhibitors (nRTI) cause profound and sustained suppression of the plasma viral load, but data on the ability of PI to cross the blood-brain barrier are conflicting[10-12], and the effects of this class of compounds on neurological signs and neuropsychological impairment are only partly known. Preliminary data concerning the effect of HAART on neurocognitive deficits  reported a significant improvement in 31% of patients, whereas 48 and 21% showed no modification or worsening, respectively, after a follow-up of 12 weeks. Moreover, a recent cross-sectional analysis  suggested that HAART could benefit neuropsychological functions through a reduction of the plasma viral load. Aggressive triple-drug regimens with one PI plus two nRTI (HAART) are extensively used, but to achieve prolonged and sustained inhibition of HIV replication, effective levels of drugs are needed in all compartments, including the CNS. Any information on the clinical evolution of cognitive impairment in patients treated with HAART is thus of critical interest. Against this background, we conducted a prospective observational study to examine the prevalence and the clinical course of ADC and MCMD in nucleoside analogue-experienced, advanced HIV-positive patients treated with HAART.
Patients and methods
The study was an open-label, prospective evaluation of the effects of HAART on HIV-1-associated neurocognitive impairment in nRTI-experienced HIV-infected patients. For this purpose from April 1996 to September 1997 116 patients with advanced HIV infection, followed at the Second Division of the L. Spallanzani Hospital of Rome, Italy, were screened for the presence of HIV-1-associated neurocognitive disturbances. The study was conducted in the context of usual clinical care according to an internal protocol. No written informed consent was required.
Patients were eligible if they had advanced HIV infection, defined as a CD4 cell count below 50/μl in at least two determinations, or two or more previous AIDS-defining events regardless of the CD4 cell count, and if they had been treated with nRTI of HIV for at least 6 months.
Ninety patients with HIV-related opportunistic infections within the previous 6 months, or with confounding neurological illness detected by clinical assessment, haematological and biochemical screens and brain computed tomography, or with non-HIV-related major neurological or psychiatric disorders such as previous stroke, head injury, active psychosis, mental retardation, or with drug abuse within the previous 12 months were excluded. Current and past medications were recorded on standardized forms, and patients who had used sedative-hypnotic agents or opioid analgesics were also excluded.
The remaining 26 patients underwent a neurological examination and comprehensive neuropsychological testing before, after 6 and 15 months of treatment with one PI plus two nRTI (HAART). The patients‚ baseline characteristics are reported in Table 1.
The PI used was indinavir in 22 patients, ritonavir in two patients and saquinavir in two patients, at the currently recommended doses. Because only three nRTI were available when the study started (zidovudine, didanosine, zalcitabine), the nRTI regimen associated with the PI was initially unchanged in 12 patients, whereas in the remaining 14 patients it contained one (10 patients) or two (four patients) new nRTI. During the study period, when lamivudine and stavudine became available, the nRTI regimen associated with the PI was modified in 18 patients because of side effects or in order to achieve a better HIV RNA response. Details on the antiretroviral regimen in the study population are reported in Table 2. The mean follow-up was 19 months (range 9-24).
Neurological and neuropsychological assessments were performed at enrolment and then repeated after 6 and 15 months. A battery of 16 standardized tests was administered by one of us (P.B.). Tests were selected to be sensitive to a wide spectrum of different cognitive domains: mental flexibility (Trail making B, Stroop colour word, controlled oral-word), concentration and speed of mental processing (Trail making A, WAIS-R digit span, WAIS-R digit symbol, Corsi cube test, Stroop word and colour), memory [Rey auditory verbal learning test, Rey complex figure (delayed), Babcook story recall test], visuospatial and constructional [Rey complex figure (copy trial)], fine motor functioning (Lafayette grooved pegboard). The scores of each test were adjusted for age and years of education. All neuropsychological examinations were assessed by the consultant neurologist and by the neuropsychologist who were blinded to patient‚s clinical parameters and antiretroviral treatment.
For data analysis, neuropsychological test scores were used in two ways. Raw scores were used to compare changes from individual patient‚s baseline performance after 6 and 15 months. Moreover subjects were classified as neuropsychologically impaired or unimpaired based on their performance relative to normative data. Cognitive impairment was defined as performing below one standard deviation from the normative mean on at least two neuropsychological tests or two standard deviations below the mean on at least one test. The same criterion was used to define a patient as impaired in a specific domain.
Medical and neurological examinations
All subjects underwent comprehensive baseline and follow-up clinical evaluations. A general physical examination was performed monthly by three of us. A general medical history was taken and current and past medications were recorded at each visit. Disease status was classified according to the Centers for Disease Control and Prevention (CDC) criteria. Adherence to HAART was assessed by pill count because antiretroviral drugs were routinely given to patients at each monthly examination. Drug abuse and the use of sedative hypnotic agents or opioid analgesics were excluded because all prescribed medications were supervised by an infectious disease specialist.
A standard neurological examination was performed by an experienced neurologist at baseline, after 6, and after 15 months of HAART. In addition, a self-reported assessment of impairment regarding the patients‚ capacity for working and performing the activities of daily living was assessed through a medical history interview.
According to the American Academy of Neurology criteria, patients were classified as ADC, MCMD or as neurologically asymptomatic.
Single photon emission computed tomography
A single photon emission computed tomography (SPECT) cerebral perfusion scan was offered to the 21 patients with baseline neuropsychological test abnormalities. SPECT was performed, as described in detail elsewhere, before HAART and was repeated after 6-15 months, in 13 subjects (one patient with ADC and 12 of the 19 patients with MCMD) who accepted it. All SPECT studies were first visually graded independently by two independent experienced observers blinded to the patient‚s neuropsychological assessment. Homogeneous brain perfusion without focal uptake defects or visible asymmetry was considered normal. Pathological patterns were classified as focal defects (one to three contiguous areas in different lobes), multifocal defects (more than three areas) and diffuse uptake defects (decreased and heterogeneous cortical uptake). In the case of disagreement between observers, a consensus approach was used.
CD4 cell counts were determined by flow cytometry.
Plasma HIV-1 RNA concentrations were detected using the NASBA kit (Organon Teknika, Boxtel, The Netherlands).
Measures of efficacy
Neuropsychological assessment was the primary outcome criterion.
For statistical evaluation we used the paired t-test, the chi-squared test and the Fischer-Tocher test.
At the baseline examination, 21 out of 26 patients (80.8%) were classified as neuropsychologically impaired and five patients (19.2%) as neuropsychologically unimpaired. According to the American Academy of Neurology criteria, ADC was diagnosed in two patients and MCMD in 19 patients. The functions involved were concentration and speed of mental processing (17 of 26: 65.4%), memory (13 of 26: 50%), visuospatial and constructional abilities (two of 26: 7.7%), fine motor functioning (three of 26: 11.5%) and mental flexibility (eight of 26: 30.8%) (Table 3). Neuropsychological test data were obtained for 24 patients at 6 months and for 23 patients at 15 months. Two patients chose not to be evaluated at the sixth month, one patient was lost to follow-up before the 15th month visit, the remaining two patients have been followed for less than 15 months at the time of writing.
After 6 and 15 months of HAART, the prevalence of patients with abnormal neuropsychological examination decreased from 80.8 % (21 of 26) to 50% (12 of 24) (P<0.05) and to 21.7% (five of 23) (P<0.001) (Table 3). We have assessed the prevalence of impairment in different cognitive areas. After 15 months of HAART, the prevalence of patients with impaired concentration and speed of mental processing decreased from 17/26 (65.4%) to five out of 23 (21.7 %) (P<0.01) and of subjects with impaired memory from 13/26 (50%) to two out of 23 (8.7%) (P<0.01), whereas changes in the prevalence of patients with visuospatial and constructional, fine motor functioning and mental flexibility impairment were not significant.
Detailed results of the mean changes from baseline for the raw score of each neuropsychological measure are reported in Table 4. A significant improvement was seen in the measure assessing concentration and the speed of mental processing (Trail making form A, WAIS-R digit span, WAIS-R digit symbol) both after 6 and 15 months of HAART.
The analysis of measures exploring mental flexibility showed significant improvements after 6 and 15 months in the Trail making B, the controlled oral-word and in the Stroop colour word tests.
Among measures exploring memory, a significant improvement was observed in the Rey complex figure (delayed) and in the Babcook story recall test. Such tests allow the analysis of learning, retention and retrieval from memory. These results indicate an improvement of learning ability. The Lafayette grooved pegboard test provides an index of fine motor functioning. When both dominant and non-dominant patients‚ hands were tested, a statistically significant improvement was seen after 6 and 15 months. Differences emerged also at the Rey complex figure test.
No correlation was seen between baseline characteristics, such as education and age (mean values in years), HIV risk (percentage of subjects with previous drug use), CDC stage (percentage of subjects with stage C disease), previous and current antiretroviral regimen, and the presence of normal or impaired neuropsychological test performance at baseline and during the follow-up (data not shown).
From a clinical point of view the two patients classified at baseline as ADC did not meet the criteria for the disorder either at the sixth or at the 15th month visit.
Single photon emission computed tomography
SPECT examination was performed in 13 patients at enrolment, and repeated one or two times after 6 to 18 months. Raw neuropsychological test scores of the 13 patients undergoing SPECT did not significantly differ from those of the eight patients not undergoing SPECT.
At the baseline visit, 10 patients had multifocal and three patients had focal uptake defects. After 6-18 months, five out of 13 patients (38.5%) showed a complete reversal of the uptake defects (P<0.05) (Table 3). Three patients showed a persistence of uptake defects with an improvement in respect to the baseline examination, whereas no changes were observed in four patients and a deterioration was observed in one patient. A regression of the neuropsychological abnormalities was observed in four of the five patients with complete reversal of the uptake defects, in two of the three subjects with improvement of SPECT findings and in two of the four patients with unchanged SPECT, whereas the patient with a deterioration of SPECT abnormalities showed a persistence of abnormal neuropsychological evaluation. However, no correlation was found between SPECT findings and neuropsychological test results, because changes from baseline for the raw scores of each neuropsychological measure showed no statistically significant difference between the five patients with complete regression of the uptake defects and the seven patients with persistence of the uptake defects.
Significant changes in CD4 cell counts, plasma HIV RNA, haemoglobin levels and body mass index were also noted (Table 3).
Because high plasma HIV-1 RNA levels, low CD4 cell count, anaemia and low body mass index have been associated with an increased risk of ADC, we evaluated whether the neuropsychological impairment was related to any of these variables. At baseline, neuropsychologically impaired and neuropsychologically normal patients did not differ in terms of plasma HIV RNA, CD4 cell count, body mass index and haemoglobin levels (Table 5). After 6 months of HAART, patients with normal neuropsychological examination, compared with neuropsychologically impaired subjects, had lower mean levels of log plasma HIV RNA (2.95 versus 3.97 log copies/ml, P<0.05), a greater drop in the log of plasma viraemia (-1.84 versus -0.83 changes in log copies/ml, P<0.05) and a higher prevalence of HIV RNA below the detection limit of 400 copies/ml (nine of 12 versus three of 12, P<0.05). Neither the absolute CD4 cell count, the CD4 cell changes from baseline, the body mass index nor the haemoglobin levels correlated to the neuropsychological test results. Moreover, probably because of the increase in plasma HIV RNA beyond the sixth month of treatment, after 15 months of HAART the correlation between plasma HIV RNA and neuropsychological performance was less evident (Table 5). Finally, no correlation was seen between baseline characteristics, such as years of education, age, HIV risk, CDC cell count stage, previous drug abuse and the improvement in neuropsychological test results (data not shown).
The nervous system is one of the primary targets of HIV. In this study, we assessed the prevalence and clinical course of neurocognitive impairment in patients with advanced HIV disease treated with HAART.
Our data indicate that HIV-1-associated neurocognitive disturbances are very frequent in advanced HIV patients with previous extensive use of nRTI, being diagnosed in 80.8% of subjects. The neuropsychological measures more impaired were among those that commonly decline in the later stages of disease, including concentration and mental flexibility, motor functioning and figural memory, with functional impairment leading to the diagnosis of ADC and MCMD in two and 19 patients, respectively (Table 1).
The introduction of HAART has resulted in an improvement in survival rates and a reduction in the incidence and severity of opportunistic infections[7,8], but its impact on the incidence and on clinical course of HIV-related neurocognitive disorders is poorly understood. Recent data  indicate that the incidence of ADC is probably decreasing, and a few observations [13,14] suggest that HAART could benefit neuropsychological functions.
Our data, obtained during a prospective evaluation of nRTI-experienced advanced HIV patients, provide a preliminary demonstration that HAART could be effective in treating cognitive impairment in advanced HIV patients. Our study has several limitations, such as the lack of a control group and the relatively large number of patients not completing the evaluations given the small sample of 26 subjects. However, our observations strongly suggest that HIV-1-related cognitive impairment may improve with HAART. The prevalence of subjects with abnormal neuropsychological examination decreased during HAART, and such therapy produced a significant improvement in concentration and the speed of mental processing, mental flexibility, memory, fine motor functioning and visuospatial and constructional abilities. This effect was evident after 6 months and was sustained up to the 15th month of HAART.
The clinical significance and the implications of the improvement in neuropsychological testing on the daily activities of the patients is difficult to assess. Such information was derived from patient history, from self-reported complaints, and from standardized neurological examination. Because self-reported information depends on subjective factors, it was not evaluated in the statistical analysis, in which only data from the neuropsychological tests were included. During the study, the two patients with ADC no longer met the criteria for the disorder, but the extent and duration of the impact of HAART on the symptoms of ADC remains to be determined.
Preliminary studies showed a relationship between increased plasma viral load and the severity of neurological disease. Moreover, a correlation was found between plasma viraemia and the levels of HIV RNA in centrifugated cerebrospinal fluid (CSF). More recently, studies demonstrated a more clear association between levels of the virus in the CSF and the onset and progression of CNS disease. In our study, although at baseline screening patients with normal or impaired neuropsychological functions did not differ in terms of plasma HIV RNA, CD4 cell count, body mass index and haemoglobin level, after 6 months of HAART subjects with a normal neuropsychological examination were more likely to have lower plasma HIV RNA levels, a greater reduction of plasma viraemia, and plasma viral load below the detection limit (Table 5). Therefore in our experience, according to Ferrando et al., both the reduction and the absolute level of plasma viral load showed a weak correlation with neuropsychological improvement. Although in our patients CSF HIV RNA was not assessed, preliminary studies reported that the reduction of CSF viral load was paralleled by a reduction of blood viraemia after the introduction of HAART. In the same way, a rapid increase of CSF and blood HIV RNA was reported after HAART interruption. However, because patients with discordant peripheral and CNS viral load response have been described, the relative value of plasma versus CSF viral load as a tool for monitoring HIV-associated neurocognitive disturbances deserves further investigation.
The improvement of the neurocognitive disturbances at the neuropsychological examination is supported by the regression of the brain blood perfusion defects at the SPECT study in five out of 13 patients. Functional neuroimaging with SPECT has revealed a high sensitivity in detecting neurocognitive disturbances. In cognitively impaired patients the prevalence of cortical and subcortical perfusion defects varies between 88 and 100%[23,24], and correlates with the degree of cognitive impairment. The complete reversal of SPECT perfusion defects is noteworthy because its spontaneous regression very rarely, if ever, occurs.
Our observation has theoretical and practical implications. First of all the reversibility of neurological dysfunction confirms that viral replication in the brain is necessary to induce clinical and pathological neurological abnormalities. The principal cells infected by HIV-1 within the brain are indeed the perivascular and parenchymal brain macrophages and microglia. Moreover, both an increase in neurotoxic activities or a concomitant decrease in neurotrophic factors may underlie ADC. How any of these factors may affect the onset and progression of ADC remains undefined; ultimately it is the neuronal injury that leads to cognitive and motor impairments.
In our experience, the reversibility of the neurological dysfunction correlated to the inhibition of viral replication (plasma viraemia) rather than to the immune recovery (CD4 cells) and to the amelioration of physical parameters (body mass index and haemoglobin).
A reduction in the incidence of ADC should be expected with the widespread use of HAART. In our patients, neuropsychological improvement lasted up to the 15th month whereas plasma viral load, after an initial decrease, showed a trend towards a new increase. The duration of the neuropsychological response to HAART, although greater than that observed with zidovudine, could thus still be only transient. Moreover, other groups have found evidence for the independent development of drug resistance in plasma and CSF compartments and, possibly, an ‚escape‚ of CNS control of HIV replication. These observations underline the need for careful monitoring of neuropsychological functions even in the HAART era.
Present opinions for the treatment of ADC recommend the use of drugs with significant CSF penetrations. In this regard, it is believed that the amount of time in which the CSF drug concentration is above the IC50 levels is important to achieve an inhibition of HIV replication. Most of our patients received indinavir plus stavudine and lamivudine. Stavudine has been shown to penetrate into the CSF in efficacious concentrations. Similarly, lamivudine can reach efficacious steady state concentrations in the CSF[11,25]. Indinavir also shows favourable CSF concentrations compared with IC50 values, because after chronic oral dosing CSF indinavir levels exceeded the IC95 for HIV clinical isolates. Our study was not designed to address the question of the optimal treatment for ADC, and further studies are needed to correlate clinical and neuropsychological results with virological and pharmacokinetical data within the CNS.
1. Bacellar H, Munoz A, Miller EN, et al
. Temporal trends in the incidence of HIV-1 related neurologic diseases: Multicenter AIDS Cohort Study, 1985-1992.Neurology
2. McArthur JC, Hoover DR, Bacellar H, et al
. Dementia in AIDS patients: incidence and risk factors.Neurology
3. Janssen RS, Nwanyanwu OC, Selik RM, Stehr-Green JK. Epidemiology of human immunodeficincy virus encephalopaty in the United States.Neurology
4. Schmitt FA, Bigley JW, McKinnis R, Logue PE, Evans RW, Drucker JL. Neuropsychological outcome of zidovudine (AZT) treatment of patients with AIDS and AIDS-related complex.N Engl J Med
5. Sidtis JJ, Gatsonis C, Price RW, et al
. Zidovudine for the treatment of the AIDS dementia complex: results of a placebo-controlled trial.Ann Neurol
6. Tozzi V, Narciso P, Galgani S, et al
. Effects of zidovudine in 30 patients with mild to end-stage AIDS dementia complex.AIDS
7. Hammer S, Squires KE, Hughes MD, et al
. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less.N Engl J Med
8. Palella FJ, Delaney KM, Moorman MO, et al
. Declining mortality among patients with advanced human immunodeficiency virus infection.N Engl J Med
9. Saacktor NC, Lyles RH, McFarlane G, et al
. The changing incidence of HIV-1-related neurological disease: 1990-1997.
In: Program and Abstract of the 6th Conference on Retrovirusis and Opportunistic Infections
, Chicago, 31 January-4 February, 1999 [Abstract 410].
10. Moyle GJ, Sadler M, Hawkins D, Buss N. Pharmacokinetics of saquinavir at steady state in CSF and plasma: correlation between plasma and CSF viral load in patients on saquinavir containing regimens.
In: Program and Abstract of the Infectious Diseases Society of America 35th Annual Meeting
, San Francisco, 13-16 September, 1997 [Abstract 115].
11. Enting RH, Hoetelmans MW, Lange JMA, Burger DM, Beijnen JH, Portegies P. Antiretroviral drugs and the central nervous system.AIDS
12. Flexner C. HIV-protease inhibitors.N Engl J Med
13. Letendre S, Ellis R, Heaton RK, Atkinson JH, Nelson J, Grant I, McCutchan JA. Change in CSF RNA level correlates with the effects of antiretroviral therapy on HIV-1 associated neurocognitive disorder [Abstract].J Neurovirol
14. Ferrando S, van Gorp W, McElhiney M, Goggin K, Sewell M, Rabkin J. Highly active antiretroviral treatment in HIV infection: benefits for neuropsychological function. AIDS
15. Centers for Disease Control and Prevention. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR
16. American Academy of Neurology. Nomenclature and research case definitions for neurologic manifestations of human immunodeficiency virus-type 1 (HIV-1) infection: report of a Working Group of the American Academy of Neurology AIDS Task Force. Neurology
17. Pau FM, Rosci MA, Bernabei A, et al
. 99mTc-HMPAO SPECT in asymptomatic HIV-1 infected patients. Preliminary results.J Nucl Med Allied Sci
18. Robertson K, Fiscus S, Wilkins J, van der Horst C, Hall C. Viral load and neuropsychological functioning in HIV seropositive individuals: a preliminary descriptive study.J NeuroAIDS
19. McArthur JC, McClernon D, Cronin M, Nance-Sproson T, Saah A, St Clair M, Lanier R. Relationship between HIV-associated dementia and viral load in cerebrospinal fluid and brain.Ann Neurol
20. Robertson K, Fiscus S, Kapoor C, et al
. CSF, plasma viral load and HIV associated dementia.J Neurovirol
21. Eggers C, Van Lunzen J, Stellbrink HJ. Rapid decay of HIV-RNA in the cerebrospinal fluid during combination antiretroviral therapy [Abstract].J Neurovirol
22. Cunningham P, Smith D, Satchell C, Cooper DA, Brew BJ. Evidence for independent development of RT inhibitor (RTI) resistance patterns in cerebrospinal fluid (CSF) compartment.
In: Program and Abstract of the 12th World AIDS Conference
, Geneva, Switzerland, 28 June-3 July, 1998 [Abstract 563/32284].
23. Sacktor N, Prohovnik R, Van Heerturn R, et al
. Cerebral SPECT abnormalities in HIV-1 infected gay man without cognitive impairment.J Nucl Med
24. Maini CL, Pigorini F, Pau FM, et al
. Cortical cerebral blood flow in HIV-1-related dementia complex. Nucl Med Commun
25. Brew BJ, Portegies P. Antiretroviral drugs within the CSF.
In: The pathogenesis and treatment of HIV CNS infection.
London: International Medical Press; 1997:14-15.
26. Letendre SL, Caparelli E, Ellis RJ, Dur D, McCuttan A. Levels of serum and cerebrospinal fluid (CSF) Indinavir (IDV) and HIV-RNA in HIV-infected individuals.
In: Program and Abstract of the 6th Conference on Retrovirusis and Opportunistic Infections
, Chicago, 31 January-4 February, 1999 [Abstract 407].