Working memory performance and cognitive flexibility after dexamethasone or hydrocortisone administration in healthy volunteers

Objective

Several studies have shown that glucocorticoids can impair declarative memory retrieval and working memory (WM) performance. The aim of the present study was to investigate the impact of a high dose of hydrocortisone on WM, as well as to examine the effects of cortisol suppression via treatment with a high dose of dexamethasone (DEX). We hypothesized that hydrocortisone treatment results in an impaired cognitive function compared with placebo. We further expected that dexamethasone treatment is also followed by cognitive impairment, due to the hypothesis that very low levels of cortisol are also associated with alterations in memory performance.

Methods

In a placebo-controlled study with a within-subject design, 16 healthy volunteers received placebo or 120?mg of hydrocortisone (two boluses of 60?mg) directly before neuropsychological testing or 4?mg of DEX the day before testing.

Results

We did not find any effect of hydrocortisone on WM and cognitive flexibility, even though cortisol levels were high at the time of testing. Furthermore, we did not find any effect of DEX treatment on WM and reaction time in a cognitive flexibility test. However, cognitive flexibility was negatively correlated with adrenocorticotropin (ACTH) in the DEX condition.

Conclusions

Our results found no clear effect of hydrocortisone and dexamethasone treatment on WM. These results emphasize the need for further research on the association between hypothalamic?Cpituitary?Cadrenal axis activity and cognition. These studies should investigate the hypotheses of dose-dependent associations in more detail and should also include analyses on ACTH and cognition.     

Hydrocortisone impairs working memory in healthy humans, but not in patients with major depressive disorder

Objective

Several studies have shown that stress or the administration of glucocorticoids can impair hippocampus-based declarative memory retrieval and prefrontal dependent working memory performance in healthy subjects. Major Depressive Disorder (MDD) is often characterized by memory impairment and increased cortisol secretion. Studies indicate that the impairing effects of glucocorticoids on declarative memory performance are missing in patients with MDD. The purpose of our study was to investigate whether the finding of missing effects of acute cortisol administration on memory performance in MDD is also seen when examining prefrontal-based working memory.

Methods

In a placebo-controlled study, 57 patients with MDD and 56 sex- and age-matched healthy control subjects received either placebo or 10?mg of hydrocortisone orally before memory testing. To test the verbal modality of working memory, the Word Suppression Test was applied with one negative and one neutral test part.

Results

After hydrocortisone intake, healthy subjects showed a significantly poorer working memory performance compared to placebo treatment when negative interference words were administered. In contrast, memory performance of MDD patients was not affected by hydrocortisone treatment.

Conclusions

The missing effects of glucocorticoid administration on working memory in MDD might be interpreted in the context of reduced central glucocorticoid receptor function.     

Effects of transdermal nicotine on attention and memory in healthy elderly non-smokers

RATIONALE: Nicotine has been found to improve cognitive functions in patients with Alzheimer's disease, but little is known about its effects in the healthy non-smoking elderly. OBJECTIVES: This study aimed to investigate the effects of nicotine on cognitive function in healthy non-smoking or nicotine-na?ve elderly subjects. METHODS: A transdermal patch containing either 5 mg nicotine or placebo was applied on the back of 63 healthy nicotine-na?ve or non-smoking elderly Koreans. Cognitive functions were evaluated with the Short Blessed Test, Rey-Kim Memory Test, and digit span test of the Korean-WAIS, both before and 5.5 h after nicotine administration. The plasma level of nicotine after testing was measured using gas chromatography. RESULTS: The subjects' memory functions in trial 5 of the Rey-Kim Memory Tests improved significantly. Furthermore, the effect on memory slope was significantly correlated with the higher plasma level of nicotine. However, the other tests did not reveal any correlation to a significant degree. CONCLUSIONS: These results suggest that nicotine of lower plasma level can improve short-term verbal memory functions in non-smoking or nicotine-na?ve healthy elderly people and that some effects are dependent on nicotine plasma levels.     

Effects of repeated cortisol administration on brain potential correlates of episodic memory retrieval

RATIONALE: Neuropsychological impairments in depressive illness may be secondary to hypercortisolaemia. Cortisol administration to healthy subjects impairs episodic memory, though how this is mediated is unknown. OBJECTIVES: To examine the effects of 1 week's administration of cortisol on the neural correlates of episodic memory in healthy subjects. METHODS: Fourteen healthy men were treated with oral cortisol (hydrocortisone 20 mg) or placebo twice daily for 1 week, in a double blind, crossover fashion. Event related potentials (ERPs) were recorded during a well-validated source memory task. Subjects listened to words spoken in a male or female voice. At test, old and new words were presented visually; subjects judged whether words were old or new, and if old, the gender of the voice at study. RESULTS: Response times were significantly speeded by cortisol. A significant reduction in recognition accuracy with cortisol was found for the second study occasion. ERP recordings with placebo showed greater positivity over left parietal and right frontal scalp areas for ERPs to items given correct source judgements versus correctly rejected new items. In comparison, cortisol increased ERP voltage between 500 and 1400 ms post-stimulus and this effect interacted with item type and electrode site, being diffusely distributed for correct rejections but of a lesser magnitude frontally for old items accorded a correct source judgement. CONCLUSIONS: Repeated cortisol administration leads to a qualitative change in the neural correlates of episodic memory retrieval in healthy subjects. This change may contribute to cognitive impairments seen in illnesses characterised by hypercortisolaemia.     

The effects of 3-week estrogen hormone replacement on cognition in elderly healthy females

Rationale: Estrogen concentrations decline with age and menopause is often followed by an acceleration of the age effects on cognition. It is suggested that replacement of estrogen would reinstate, at least in part, cognitive abilities. Effects of estrogens on memory have been reported in studies with women in a clinical setting who either needed or wished to have the estrogen replacement and are mostly in the perimenopausal age-band. Objective: The present study investigated the effects of estradiol on memory and on frontal lobe function in elderly female subjects who did not suffer any of the postmenopausal symptoms and had never taken estrogen hormone replacement (EHR) previously. Methods: EHR (Progynova TS, transdermal estradiol; n=19) or placebo (n=18) was given for a period of 3 weeks to elderly healthy female subjects. Memory, frontal lobe functions (inhibition and planning) and visuospatial abilities (mental rotation) were tested before and after treatment. Estrogen plasma levels were measured to confirm the result of EHR. Cortisol plasma levels were also measured before and after cognitive performance in order to evaluate the effects of EHR on the sensitivity of the hypothalamo- pituitary-adrenal (HPA) axis to mild mental stress. Results: Plasma estradiol levels in the drug group increased to levels equivalent to that of a fertile woman (0.21± 0.5 nmol/l). Memory function as well as visuospatial abilities as measured by a mental rotation task improved significantly with EHR. However, there was no effect of EHR on frontal lobe functions. The cognitive effects were not dependent on an improvement in mood or general well-being as may be the case with EHR in women at peri- or post-menopausal stage. EHR was found to increase the HPA response to task-induced stress, as indicated by an increase in cortisol plasma levels. Conclusions: The present study has provided evidence of a beneficial effect of EHR on cognitive abilities given for first time to healthy elderly women. Furthermore, the present study has demonstrated a differential effect of EHR on memory, visuospatial abilities and frontal lobe function. Received: 14 June 1999 / Final version: 7 October 1999     

Mineralocorticoid Receptor Stimulation Improves Cognitive Function and Decreases Cortisol Secretion in Depressed Patients and Healthy Individuals

Memory and executive function are often impaired in patients with major depression, while cortisol secretion is increased. Mineralocorticoid receptors (MR) are abundantly expressed in the hippocampus and in the prefrontal cortex, brain areas critical for memory, executive function, and cortisol inhibition. Here, we investigated whether MR stimulation with fludrocortisone (1) improves memory and executive function and (2) decreases cortisol secretion in depressed patients and healthy individuals. Twenty-four depressed patients without medication and 24 age-, sex-, and education-matched healthy participants received fludrocortisone (0.4 mg) or placebo in a randomized, double-blind, within-subject cross-over design. We measured verbal memory, visuospatial memory, executive function, psychomotor speed, and salivary cortisol secretion during cognitive testing between 1400 and 1700 hours. For verbal memory and executive function, we found better performance after fludrocortisone compared with placebo across groups. No treatment effect on other cognitive domains emerged. Depressed patients performed worse than healthy individuals in psychomotor speed and executive function. No group effect or group × treatment interaction emerged on other cognitive domains. Fludrocortisone decreased cortisol secretion across groups and there was a significant correlation between cortisol inhibition and verbal memory performance. Our data suggest a crucial role of MR in verbal memory and executive function and demonstrate the possibility to improve cognition in depressed patients and healthy individuals through MR stimulation.     

The effects of nicotine on cognition are dependent on baseline performance

Since cholinergic neurotransmission plays a major role in cognition, stimulation of the nicotinic acetylcholine receptor may be a target for cognitive enhancement. While nicotine improves performance on several cognitive domains, results of individual studies vary. A possible explanation for these findings is that the effect of nicotine administration may be dependent on baseline cognitive function, where subjects with a suboptimal cognitive performance may benefit from nicotine, while subjects who already perform optimally may show a decline in performance after nicotinic stimulation. We conducted a double-blind randomised placebo-controlled crossover trial, examining the effects of placebo, 1, and 2 mg of nicotine on cognition in young (n=16, age 18–30 years) and healthy elderly (n=16, age 60–75 years) subjects. We hypothesised that the elderly would benefit more from nicotine compared to young subjects, as normal ageing is associated with decreases in cognitive function. Attention, working memory, visual memory, information-processing speed, psychomotor function, stereotypy, and emotion recognition were assessed. Compared to the young volunteers, the elderly performed significantly worse on psychomotor function and emotion recognition in the placebo condition. Nicotine had no effect in the young volunteers and decreased performance on working memory and visual memory in the elderly. Contrary to our hypothesis, the effect of nicotine was dependent on baseline performance in both the groups, with subjects with lower baseline performance benefiting from nicotine administration, while those with higher baseline performance performed worse after nicotine administration. This suggests that subjects with lower cognitive performance, irrespective of age, may benefit from nicotine.     

A comparison of the effects of olanzapine, haloperidol and placebo on cognitive and psychomotor functions in healthy elderly volunteers.

The cognitive and psychomotor effects of olanzapine (3 mg) were compared with haloperidol (3 mg) and placebo in a double-blind, cross-over study. Fourteen healthy elderly volunteers (>65 years) were randomized to receive once daily medication for 4 days with a 16-day interval between treatment periods. Assessments of attention, memory and motor control were made prior to dosing on each day, at 2, 4, 6 and 8 h after dosing on days 1 and 4, and at 24 and 48 h following the last dose. On day 1, detectable impairment was observed at all time points in both groups. On day 4, haloperidol treated subjects showed increased impairment compared with day 1 and this was sustained throughout the 48 h of testing. Olanzapine treated subjects showed reduced day 4 deficit (compared with day 1), with no significant difference from placebo beyond 6 h post dose. These results suggest that both haloperidol and olanzapine have a measurable initial effect on cognitive and psychomotor function in elderly volunteers. However, acute effects associated with olanzapine decrease with repeated dosing and show substantial adaptation within 4 days. In contrast, effects seen with haloperidol are sustained and increase with repeated dosing over the same period.     

Blockade of the mineralocorticoid receptor in healthy men: effects on experimentally induced panic symptoms, stress hormones, and cognition.

Animal studies have shown that blockade of central mineralocorticoid receptors (MR) has anxiolytic effects and impairs several aspects of cognitive function. No study to date assessed the effects of MR blockade on anxiety and cognitive function in humans. In the present study, 16 healthy young men were treated either with placebo or with 300 mg spironolactone, a MR-antagonist, at 1100, 1330, and 1630 hours in a balanced cross-over design with the two study conditions being 1 week apart. At 1500 hours, the panic symptoms provoking compound cholecystokinin-tetrapeptide (CCK-4) was administered i.v. on both occasions and panic symptoms were assessed. We measured plasma ACTH and cortisol between 1300 and 1900 hours and assessed cognitive function between 1800 and 1900 hours. CCK-4 elicited panic symptoms and increased ACTH and cortisol secretion in both conditions. Intensity of panic symptoms after CCK-4 was not different between spironolactone and placebo. Spironolactone significantly impaired selective attention and delayed recall of visuospatial memory, and diminished set shifting/mental flexibility on a trend level. Pretreatment with spironolactone led to higher baseline cortisol levels compared to placebo whereas no differences in stimulated cortisol, baseline ACTH, and stimulated ACTH emerged. Blockade of MR with spironolactone increases baseline cortisol secretion and impairs cognitive function but has no effect on experimentally induced panic symptoms in humans, for the study design and dosage of spironolactone used. The domains of cognitive function that are impaired after blockade of MR in men, that is, selective attention, visuospatial memory, and mental flexibility/set shifting appear to be remarkably similar to those described in animal studies.     

The cognitive effects of modulating the glycine site of the NMDA receptor with high-dose glycine in healthy controls

N-methyl-D-aspartate (NMDA) receptors play an important role in learning and memory. Targeting the glycine modulatory site of the NMDA receptor has been suggested as a therapeutic strategy to improve cognition, although findings have not been convincing. We used the Cognitive Drug Research computerised assessment system to examine the effects of high-dose glycine on a number of cognitive processes in healthy young subjects. The study was a randomised placebo controlled repeated measures design in which each subject received acute placebo or glycine (0.8 g/kg) orally, with treatment conditions separated by a 5-day washout period. No significant effects of glycine were found on measures of working memory, declarative memory, attention or perceptual processing. These findings, together with those of previous studies, cast doubt over the ability of acute high-dose glycine to improve cognitive function in healthy subjects and suggest that the optimum dose of glycine for improving cognition may vary between different populations, possibly due to differences in endogenous glycine levels and the functional status of NMDA receptors.     

Scopolamine effects on memory,language, visuospatial praxis and psychomotor speed

Scopolamine hydrobromide was administered by subcutaneous injection to 30 young subjects in a dose of 0.22 mg/70 kg, 0.43 mg/70 kg, or 0.65 mg/70 kg. Treatment effects were compared to placebo on an extensive cognitive assessment battery. Almost all tests in the battery had been previously administered to Alzheimer's disease patients and nondemented elderly subjects. Scopolamine produced deficits on tests of verbal recall, visuospatial recall, visual recognition memory, visuospatial praxis, visuoperceptual function, and psychomotor speed. Immediate memory, language function, object sorting, and frequency of intrusion errors were unaffected. The low dose of scopolamine produced some peripheral anticholinergic signs but did not affect the cognitive measures. The results support the conclusion reached in previous studies that the cognitive profile of scopolamine-injected young subjects is more similar to that of the nondemented elderly than to that of Alzheimer's disease patients.     

Effects of a single dose of cortisol on the neural correlates of episodic memory and error processing in healthy volunteers

Rationale. Neuropsychological impairments seen in depression may be secondary to hypercortisolaemia. Repeated cortisol administration impairs episodic memory with an alteration in event-related potentials (ERPs) recorded during information retrieval. It is unclear whether such ERP effects are specific to episodic memory, or whether repeated cortisol administration is required. Objective. To investigate the effect of a single dose of hydrocortisone on the neural correlates of episodic memory and error detection. Methods. Twenty healthy subjects were treated with hydrocortisone (100 mg) or placebo orally, in a double-blind, two-way crossover study. ERPs were recorded during an episodic memory and a Stroop task, 1–3 h following the medication. Results. Cortisol increased error rates during the Stroop task but had no effect on episodic memory. The magnitude of ERPs associated with incorrect response in the Stroop task between −250 ms and +500 ms post-response was increased by cortisol, with no effect on correct-response ERPs. There was no effect of cortisol on episodic memory-retrieval-dependent ERPs. Conclusions. Cortisol can impair not only episodic memory but also processes involved in error detection. In contrast to repeated cortisol administration, a single dose of cortisol does not alter the behavioural performance or the electrophysiological correlates of episodic memory. However, it increases error rates in a choice response task with associated quantitative changes in incorrect-response ERPs. This probably reflects an alteration in anterior cingulate cortex activity. Such changes may contribute to the neuropsychological impairment seen in depression. This study also demonstrates the utility of ERPs for investigating the effect of neuroendocrine manipulations on the neural correlates of neuropsychological function. Electronic Publication     

Effect of repeated gaboxadol administration on night sleep and next-day performance in healthy elderly subjects.

Aging is associated with dramatic reductions in sleep continuity and sleep intensity. Since gaboxadol, a selective GABA(A) receptor agonist, has been demonstrated to improve sleep consolidation and promote deep sleep, it may be an effective hypnotic, particularly for elderly patients with insomnia. In the present study, we investigated the effects of subchronic gaboxadol administration on nocturnal sleep and its residual effects during the next days in elderly subjects. This was a randomized, double-blind, placebo-controlled, balanced crossover study in 10 healthy elderly subjects without sleep complaints. The subjects were administered either placebo or 15 mg gaboxadol hydrochloride at bedtime on three consecutive nights. Sleep was recorded during each night from 2300 to 0700 h and tests assessing attention (target detection, stroop test) and memory function (visual form recognition, immediate word recall, digit span) were applied at 0900, 1400, and 1700 h during the following days. Compared with placebo, gaboxadol significantly shortened subjective sleep onset latency and increased self-rated sleep intensity and quality. Polysomnographic recordings showed that it significantly decreased the number of awakenings, the amount of intermittent wakefulness, and stage 1, and increased slow wave sleep and stage 2. These effects were stable over the three nights. None of the subjects reported side effects. Next-day cognitive performance was not affected by gaboxadol. Gaboxadol persistently improved subjective and objective sleep quality and was devoid of residual effects. Thus, at the employed dose, it seems an effective hypnotic in elderly subjects.     

Effects of a combined extract of Ginkgo biloba and Bacopa monniera on cognitive function in healthy humans

Extracts of Ginkgo biloba and Bacopa monniera have been shown to produce positive effects on cognitive function in healthy subjects. While the exact mechanisms are not known, it has been suggested that antioxidant properties and cholinergic modulation may play a role. In the current study the sub-chronic (2 weeks) and chronic (4 weeks) effects of an extract containing Ginkgo biloba (120 mg) and Bacopa monniera (300 mg) (Blackmores Ginkgo Brahmi) on cognitive function were examined. The study was a randomized, double-blind, placebo-controlled, independent group design in which 85 healthy subjects were allocated to one of two treatment conditions (placebo or combined Ginkgo biloba and Bacopa monniera extract). Testing was conducted at baseline and 2 and 4 weeks post treatment. The results showed that the combined extract relative to placebo did not demonstrate any significant effects on tests investigating a range of cognitive processes including attention, short-term and working memory, verbal learning, memory consolidation, executive processes, planning and problem solving, information processing speed, motor responsiveness and decision making. These findings suggest that at least within the current treatment duration and doses, an extract containing Ginkgo biloba and Bacopa monniera had no cognitive enhancing effects in healthy subjects.     

The acute effects of an extract of Bacopa monniera (Brahmi) on cognitive function in healthy normal subjects

The Ayurvedic medicine Bacopa monniera (Brahmi) has been shown to exert cognitive enhancing effects in animals. The current study examined the acute effects of an extract of Bacopa monniera on cognitive function in normal healthy human subjects. The study was a double-blind, placebo-controlled independent group design in which subjects were randomly allocated to one of two treatment conditions, Bacopa monniera (300 mg) (n = 18) or placebo (n = 20). Neuropsychological testing was conducted before and 2 h after drug administration. No significant changes were found on any of the tests. The findings suggest that Bacopa monniera, at least for the dose administered, has no acute effects on cognitive functioning in normal healthy subjects. Copyright 2001 John Wiley & Sons, Ltd.     

The effects of chronic administration of hydrocortisone on cognitive function in normal male volunteers

Rationale: Corticosteroids are elevated in certain neuropsychiatric disorders and this may contribute to the neuropsychological impairments reported in these disorders. Objective: To examine the effects of hydrocortisone on learning, memory and executive function. Methods: Hydrocortisone 20 mg was administered twice daily for 10 days to normal male volunteers in a randomized, placebo control, crossover, within-subject design. Learning, memory and executive function were measured using selected subtests from the Cambridge Neuropsychological Test Automated Battery. Results: Hydrocortisone caused impairments of visuo-spatial memory. These included increased within search errors and impaired use of strategies on the spatial working memory subtest. In addition, administration of hydrocortisone was associated with more errors in the paired associate learning subtest, although no effect was found on the Tower of London. Hydrocortisone speeded response latencies in certain tests (pattern and spatial recognition memory). Conclusion: These results indicate that chronic administration of hydrocortisone leads to deficits in certain tests of cognitive function sensitive to frontal lobe dysfunction and may contribute to the cognitive impairment reported in certain neuropsychiatric disorders. Received: 27 July 1998 / Final version: 9 February 1999     

Hypersensitivity to scopolamine in the elderly

Scopolamine hydrobromide, 0.43 mg/70 kg, was administered by subcutaneous injection to ten young and ten elderly subjects. A comprehensive neuropsychological test battery was used to assess the effects of scopolamine, as compared to placebo, on cognitive function. As previously reported for this group of young subjects, scopolamine significantly impaired performance on tests of recent memory and visuospatial praxis. The same effects were observed in the elderly subjects, but the magnitude of the effects was much larger. The scopolamine injections produced significant psychomotor slowing in the elderly, whereas higher doses of the drug are required to produce this effect in young subjects. In both young and old subjects scopolamine failed to affect immediate memory, language function, object sorting, and the frequency of intrusion errors (although trends toward an effect were more apparent in the elderly). Remote memory, tested in the elderly only, was also unaffected. The results suggest that scopolamine's cognitive effects are quantitatively more pronounced in elderly subjects than young subjects, but that they are qualitatively similar and do not constitute a valid model for the cognitive dysfunction associated with Alzheimer's disease.     

Modulation of memory and visuospatial processes by biperiden and rivastigmine in elderly healthy subjects

Rationale The central cholinergic system is implicated in cognitive functioning. The dysfunction of this system is expressed in many diseases like Alzheimer's disease, dementia of Lewy body, Parkinson's disease and vascular dementia. In recent animal studies, it was found that selective cholinergic modulation affects visuospatial processes even more than memory function. Objective In the current study, we tried to replicate those findings. In order to investigate the acute effects of cholinergic drugs on memory and visuospatial functions, a selective anticholinergic drug, biperiden, was compared to a selective acetylcholinesterase-inhibiting drug, rivastigmine, in healthy elderly subjects. Methods A double-blind, placebo-controlled, randomised, cross-over study was performed in 16 healthy, elderly volunteers (eight men, eight women; mean age 66.1, SD 4.46 years). All subjects received biperiden (2 mg), rivastigmine (3 mg) and placebo with an interval of 7 days between them. Testing took place 1 h after drug intake (which was around T _max for both drugs). Subjects were presented with tests for episodic memory (wordlist and picture memory), working memory tasks (N-back, symbol recall) and motor learning (maze task, pursuit rotor). Visuospatial abilities were assessed by tests with high visual scanning components (tangled lines and Symbol Digit Substitution Test). Results Episodic memory was impaired by biperiden. Rivastigmine impaired recognition parts of the episodic memory performance. Working memory was non-significantly impaired by biperiden and not affected by rivastigmine. Motor learning as well as visuospatial processes were impaired by biperiden and improved by rivastigmine. Conclusions These results implicate acetylcholine as a modulator not only of memory but also of visuospatial abilities.     

The effects of single doses of CL 284,846, lorazepam,and placebo on psychomotor and memory function in normal male volunteers

Summary The effects of single doses of CL284,846 (20 mg), lorazepam (2 mg) and placebo on psychomotor performance, memory function and subjective feelings were assessed in 12 normal, healthy male volunteers. Each subject received each treatment in balanced order and a minimum of 6 days was left between treatments.The subjects performance on a comprehensive battery of tests of psychomotor performance, memory function and subjective ratings was assessed pre-treatment and at 1, 3 and 5 h post-treatment.In general, the effects of CL284,846 on memory were similar to those of lorazepam at 1 h post-treatment but, recovery was rapid with CL284,846. Impairments induced by lorazepam persisted throughout the post-drug testing sessions.This pattern of effects was repeated across most of the variables tested. However, at 1 h, CL284,846 produced less marked psychomotor impairment than lorazepam.The results of this study suggest that CL284,846 is a safe, rapid acting and effective sedative with some clear advantages over lorazepam with respect to unwanted cognitive and psychomotor impairments.     

Interdose elevation in plasma cortisol during chronic treatment with alprazolam but not lorazepam in the elderly.

Benzodiazepines (BZPs) have been shown to reduce hypothalamic-pituitary-adrenal (HPA) axis activity acutely in normal humans. In contrast, the effects of chronic BZP treatment on the HPA axis have not been well studied, especially in the geriatric population. This study examined the acute and chronic effects (3 weeks) of alprazolam and lorazepam on plasma cortisol in 68 subjects (60-83 years) who received 0.25 or 0.50 mg b.i.d. alprazolam, or 0.50 or 1.0 mg b.i.d. lorazepam, or placebo orally according to a randomized, double-blind, placebo-controlled parallel design. Memory assessment and blood samples for plasma cortisol were obtained prior to the morning dose on days 0, 7, 14, and 21, and at 1, 2.5, and 5 h postdrug on days 0 and 21. Assessments of anxiety and depression were carried out at days 0, 7, 14, and 21 before drug administration. Plasma cortisol was affected compared to placebo only by the 0.5 mg alprazolam dose. During the first and the last day of treatment, there was a significant drop in cortisol at 2.5 h after alprazolam compared to placebo. The predose cortisol levels increased significantly during chronic alprazolam treatment, and correlations were found between these cortisol changes and changes in depression, anxiety, and memory scores. These findings suggest that even a short period of chronic treatment with alprazolam, but not lorazepam, may result in interdose HPA axis activation in the elderly, consistent with drug withdrawal. If confirmed, this effect may contribute to an increased risk for drug escalation and dependence during chronic alprazolam treatment.