Methylphenidate Improves HIV-1–Associated Cognitive Slowing

Despite recent advances that have led to the development of novel medication regimens for the treatment of human immunodeficiency virus–type 1 (HIV-1) infection, the majority of HIV-1–infected individuals will nonetheless develop some degree of neurocognitive and neuropsychiatric symptomatology during the course of their illness. Cognitive decline, including dementia, as well as neuropsychiatric symptoms such as depression, apathy, and anergia are common symptoms of HIV-associated central nervous system (CNS) disease.¹ Although studies of the prevalence of HIV-associated dementia vary considerably, it appears that between 6% and 20% of patients with AIDS will eventually develop dementia.² Cognitive deficits of lesser severity are estimated to occur in another 30% to 50% of HIV-infected individuals, particularly those with more advanced disease.^2,3 Mood disturbance, especially depression, is also common in HIV-1 infection, with rates of major depressive disorder greatly exceeding those found in the general population.^4–7

The characteristic neurocognitive deficits seen in HIV-1 infection include alterations in memory, attention, executive function, mood, and psychomotor speed—symptoms that are considered reflective of a subcortical dementia. Of these HIV-related neuropsychological deficits, psychomotor slowing is the cardinal symptom. Slowing of information processing has been reported in numerous studies and may be among the earliest HIV-associated cognitive changes. Computerized measures of reaction time, especially choice reaction time, have been shown to be particularly sensitive to HIV-associated cognitive slowing and may be more likely to detect the sometimes subtle nature of information processing slowing seen among a subset of HIV-infected persons.^8,9 In particular, reaction time tasks that require higher-level cognitive processing (e.g., decision-making) seem to be particularly sensitive to the deleterious effects of HIV.^9–13

Although the pharmacotherapeutic treatment of depression is a relatively mature field, drug treatment of HIV-associated neurocognitive disorder remains in its nascency. One agent that has been employed in the treatment of HIV-1 related cognitive impairment is the psychostimulant methylphenidate (MPH; Ritalin). The primary mechanism of MPH's action in the CNS is to increase the synaptic concentration of dopamine through reuptake blockade. There is mounting evidence suggesting that at least a subset of HIV-infected individuals may suffer some degree of dopaminergic dysregulation. For example, lower levels of dopamine and its metabolite, homovanillic acid, have been observed in cerebrospinal fluid samples taken from HIV-infected persons.^14,15 For this reason, increasing available dopamine through MPH administration should prove beneficial in the treatment of HIV-associated neuropsychiatric symptoms in addition to potentially improving cognitive performance. Studies have shown that either too little or too much available dopamine adversely affects working memory.¹⁶ Also, in studies with both rats and primates dopaminergic depletion has been associated with reaction time slowing,^17,18 and investigators working with human subjects have suggested that dopaminergic dysfunction is related to psychomotor slowing and dual-task deficits in other subcortical neurologic disorders such as Parkinson's disease.^19,20

Although an extensive body of literature exists with regard to the efficacy of MPH in the treatment of disorders such as childhood attention deficit disorder, depression, and narcolepsy,^21,22 to date only four studies have reported on the use of MPH in HIV-1–infected patients, and all of these were limited by small sample sizes (N≤8). Fernandez and co-workers^23,24 reported on two series of 6 subjects who were administered MPH and studied with measures of mood and cognition in an unblinded, uncontrolled fashion. Although no statistical analyses were performed in the first study, these authors reported an improvement in cognition, whereas in the second study they found no improvement in cognitive functioning on MPH; however, depressed mood did improve. Angrist et al.²⁵ used an ABAB, single-blind design and found that the cognitive performance of 7 HIV-infected patients, all of whom had AIDS and were experiencing some degree of neuropsychiatric impairment at the time, improved both on and off MPH, suggesting that improvement may have been due to practice rather than to the effects of MPH. In contrast, van Dyck et al.²⁶ did find a beneficial effect of sustained-release MPH on a composite measure of neuropsychological function among a sample of 8 substance abusers with HIV infection who were participating in an inpatient methadone treatment program. The van Dyck et al. study is the only extant placebo-controlled, double-blind crossover study of MPH among patients with HIV-1 infection.

Given the ambiguity in this literature, the present study was designed to evaluate the effectiveness of MPH in the treatment of HIV-associated cognitive slowing and depressive symptoms. This study presents data from the largest sample of HIV+ persons yet to participate in a placebo-controlled, crossover study design. In addition, sensitive computerized reaction time (RT) measures are used to assess cognitive slowing, including a dual-task RT paradigm shown to be particularly sensitive to the CNS effects of HIV-1 infection.²⁷ Finally, individual differences in MPH effectiveness, which have never been examined, are also analyzed in the current study.

We hypothesized that 1) treatment with 30 mg/day of MPH would result in significantly better performance on RT, with the largest improvement evident in the slowest HIV+ adults; 2) participants on MPH would report a subjective improvement in mood as reflected by scores on the Beck Depression Inventory, with the largest improvement in the most depressed HIV+ adults; and 3) because depression is related to slower processing speed in HIV-1,²⁸ MPH would produce the largest improvement in reaction time in the most depressed HIV+ adults.

METHODS

Participants

Sixteen HIV-1–seropositive (HIV+) individuals, recruited from the infectious disease clinics at two university-affiliated medical centers, participated in this study. Ten subjects met CDC diagnostic criteria for AIDS, two by virtue of a CD4 count <200. The mean absolute CD4 count for the sample was 362±215 (values reported as mean±SD). All subjects were on antiretroviral combination therapy at the time of testing. Mean age of the participants (15 men, 1 woman) was 42.2±7.2, mean education level was 15.2±2.6 years, and mean estimated premorbid IQ, based on the National Adult Reading Test, was 108.6±10.6. Fifty percent of participants were African American, 31% were white, and 19% were Hispanic. Prospective participants with a history of neurologic disease, including opportunistic infections/neoplasms of the CNS, significant head injury (loss of consciousness>30 minutes), major psychiatric disorder (per the Structured Clinical Interview for DSM), or current substance abuse were excluded from the study. All subjects provided informed consent and received $50.00 for participating in the study.

Procedure

A single-blind, placebo-controlled crossover design was employed. Participants were not blinded, since the stimulant properties of MPH were readily apparent to recipients. After initial neuropsychological testing was performed, subjects were randomly assigned to MPH 10 mg tid or placebo tid. After 3 days, participants were again administered the measures of mood and cognition, which are described below. Following a 3-day washout period, participants were crossed over to the other medication condition and again tested before and 3 days after starting on the second medication. Testings at Time 2 and Time 4 began one hour following afternoon dosage in an effort to ensure an optimally therapeutic level of MPH. The following measures were employed:

Single-task simple RT. This task required participants to respond with a button press (using the index finger of their dominant hand) to a 1,000-Hz tone presented through earphones. Stimulus onset was always preceded by a visual cue (an X displayed for 1,000 ms at the center of the monitor) with interstimulus interval between cue offset and stimulus onset quasi-randomly varied. Eight practice trials were administered to ensure that participants understood the task. Two experimental blocks of 60 trials were then administered.

Single-task choice RT. The choice RT task required participants to rapidly determine whether two sequentially presented polygons were identical. Exactly 1,000 ms following a visual warning cue, a complex geometric design (a nine-sided polygon), subtending approximately 11 degrees vertically and 10 degrees horizontally, was presented to the center of the monitor for a duration of 1,000 ms. After being displayed for 1,000 ms, the figure was erased. A second polygon, either identical to or slightly different from the first, was then presented 1,000 ms after offset of the first design. Participants then had to vocally respond “same” or “different” into a microphone with the response detected by voice-activated relay and computer-recorded to the nearest ms. The stimuli were identical on 50% of trials. Following 8 practice trials, two blocks of 30 trials were administered.

Dual-task RT. In the dual-task condition, participants simultaneously engaged in both the simple and the choice RT tasks described above. During each trial the participants would be required to respond to the auditory tone with a button push while concurrently completing the above visual discrimination task with a speeded vocal response. The auditory tone was quasi-randomly presented at varying temporal positions while participants were performing the choice RT task. Following the administration of 10 practice trials to ensure that the participants comprehended the task instructions, four experimental blocks consisting of 40 trials each were administered.

Beck Depression Inventory (BDI). The BDI is a 21-item self-report rating scale on which respondents report the presence and severity of symptoms of depression. The BDI was administered per standard directions at each visit, with the exception that at visits 2, 3, and 4 the participants were asked to rate depressive symptomatology since the previous visit.

Data Analyses

SPSS version 8.0 was used for all analyses. The dependent variables were median RTs on each task (simple or choice) in each condition (dual-task or single-task). For choice RT trials, only those trials on which the visual discrimination was correct were used in calculating median RTs. Anticipatory responses (RT<125 ms) and non-responses (RT>2,000 ms) were excluded from analysis.

RESULTS

The first series of analyses, which examined the efficacy of MPH among the entire cohort of HIV+ participants, employed a mixed-model analysis of variance with treatment order (MPH first vs. placebo first) as a between-subjects factor and change score (pre–post MPH vs. pre–post placebo) as the within-subjects factor. Dependent variables included simple RT, choice RT, dual-task RT (simple and choice), and BDI score. Means are presented in Table 1.

Among the entire cohort, for the simple RT, choice RT, and dual-task simple RT conditions, there were no statistically significant (at the 0.05 level) main effects or interactions (all P≥0.25). For the dual-task choice RT condition, there was a main effect for drug condition (change score), F=6.05, df=1,14, P=0.03, with choice RT being 60 ms faster in the placebo condition (relative to baseline) and 23 ms slower in the MPH condition (relative to baseline). No other analyses were significant for dual-task choice RT (P≥0.22). For the BDI, there was no order effect (P=0.73), but there was a main effect for drug condition (F=7.51, df=1,14, P=0.02), in which BDI score dropped in the placebo condition by 3.2 points relative to baseline, compared with a 0.4-point reduction in the MPH condition relative to baseline. There was also a significant interaction between treatment order and drug condition (F=10.0, df=1,14, P=0.007), indicating that the largest reduction in BDI score was in the placebo condition when placebo was the first condition.

We then tested the hypothesis that MPH might prove effective only among the more cognitively or affectively impaired participants. This hypothesis received considerable support. A series of regression analyses was performed relating cognitive slowing and level of depression at study entry with degree of subsequent cognitive improvement on MPH versus placebo (Table 2). As predicted, the beneficial effects of MPH on reaction time were found to vary, and participants with the greatest degree of cognitive slowing on entry improved the most while on MPH. Specifically, choice RT at entry was predictive of MPH-related improvement in choice RT and dual-task choice RT (Figures 1 and 2). There was no such relationship between initial cognitive slowing and subsequent change in RT while on placebo nor in the simple RT conditions. Similar to the findings with choice RT at entry, BDI score at entry was predictive of MPH-related improvement in dual-task choice RT (Figure 3), although this was not evident with single-task choice RT.

Regression analyses were also conducted to determine if MPH had a beneficial effect on level of depression as a function of baseline cognitive slowing or depression score. As can be seen in Table 2, neither cognitive slowing at entry nor BDI at entry was predictive of MPH effects on BDI score.

DISCUSSION

The results of this study, the largest to date that has investigated this research question, suggest that 30 mg/ day of MPH has a beneficial effect on cognitive speed among some, but not all, HIV-infected individuals. Those participants who entered the study with a greater degree of cognitive slowing, and perhaps greater depressive symptomatology, responded best to MPH. This was true only with the relatively more complex tasks such as choice RT and dual-task choice RT. Because simple psychomotor speed (simple RT) was never improved by MPH, nor was simple psychomotor speed at entry related to MPH effects, the findings with the choice RT tasks indicate a cognitive component—such as decision-making processes—in the modulatory mechanism of MPH effects in certain HIV-infected adults. Subjects were not blinded to drug condition, and thus to some degree expectancy effects might have influenced study results. Arguing against this possibility is the finding that subjects did not significantly improve on simple RT tasks while on MPH, but only (if at all) on the more complex tasks. Such specificity in response to MPH would not be consistent with expectancy effects.

In contrast to the beneficial effects of MPH among the slower subjects, those HIV-infected individuals who demonstrated better functioning at study entry did not differentially improve on MPH compared with placebo. In all likelihood, this is because the higher-functioning individuals had no discernable symptoms for MPH to treat and thus had limited room to improve. The use of a counterbalanced crossover design and the lack of improvement in the more impaired patients on placebo suggest that the findings were not simply a function of regression toward the mean. In addition, as can be seen in Figures 1, 2, and 3, HIV-infected adults with quicker RTs at study entry were actually slower on MPH (negative change scores indicating slower RT relative to baseline). Similar findings were recently reported by Mehta et al.²⁹ in a study of regional cerebral blood flow among normal adults who were administered MPH prior to engaging in a spatial working memory task. They too found that the beneficial effects of MPH were most pronounced among subjects with the poorest baseline performance. Our finding—that slower subjects improved on MPH while faster subjects declined on MPH—suggests that MPH might actually have adverse cognitive effects when administered to unimpaired patients. Another explanatory model may be that MPH benefited the initially slower adults because they were presumably at a low or suboptimal level of arousal. On the other hand, MPH might be producing an overaroused state in the HIV-infected adults with the faster-choice RT at entry. Although this is only a post hoc explanation, arousal theory, especially with its close relationship with stress and stimulant effects on performance,³⁰ may provide a useful construct in explaining individual differences in MPH effects on cognition in HIV-infected adults.

Surprisingly, symptoms of depression did not differentially respond to MPH, although this may in part be attributable to insufficient length of treatment. These results are inconsistent with prior studies by Angrist et al.²⁵ and Fernandez et al.,²⁴ both of whom reported improvement in depressive symptoms and signs. It seems likely that the patients in these two studies, as a group, were more neuropsychiatrically impaired than those in the current study. For example, the majority of participants in the current study scored in the mild to minimal range of depressive symptomatology on the BDI. Another possible explanation for the lack of change in BDI score is that it may be an artifact of item content. That is, the stimulant effects of MPH might have adversely affected sleep, appetite, and irritability (all items measured by the BDI), thereby offsetting any improvement in the cognitive or affective components of depression. Also, there is reason to believe that methylphenidate may be particularly successful at alleviating symptoms and signs of apathy.^31,32 Our group is currently exploring the extent to which apathy, more so than depression, may have responded to methylphenidate. In addition, although there was some limited evidence that the more depressed HIV-infected adults at entry improved in dual-task choice RT while on MPH, there was no such effect of MPH on symptoms of depression. This would argue against the hypothesis that MPH quickens reaction time by reducing levels of depression. Rather, this finding suggests that MPH likely exerts a direct influence on cognitive speed. It also bears mention that because the chronometric measures used in this study are experimental and hence lack normative data, it is difficult to know the exact clinical significance of these improvements in reaction time.

Thus, slowed information processing in HIV-1 infection appears amenable to pharmacologic intervention using the dopamine agonist MPH. However, the data suggest that clinicians using MPH as an adjunctive treatment for HIV-associated cognitive slowing should consider reserving the use of MPH for those patients with more pronounced cognitive and affective deficits. Furthermore, the results of the present study are restricted to fairly simple measures of reaction time. Processing speed has been shown to partially mediate other domains of cognitive decline in HIV-infected adults.^28,33,34 How these MPH results generalize to other, more complex cognitive activities (including non-timed measures) in HIV-infected adults remains to be explored.

ACKNOWLEDGMENTS

The authors thank Omar Mahmood and Marta Stefaniak for their research assistance. This study was supported by funds from the University of California university-wide AIDS Research Program and from the Department of Veterans Affairs. Portions of this paper were presented at the 27th annual meeting of the International Neuropsychological Society, Boston, MA, February 1999.

TABLE 1. Reaction times and BDI scores in baseline, placebo, and methylphenidate conditions (mean±SE)

Enlarge table

TABLE 2. Regression functions predicting change in RT measures and BDI due to placebo and methylphenidate

Enlarge table