ADVERTISEMENT

Issue Navigator

Volume 10 No. 02
Earn CME
Accepted Papers
Classifieds







Scientific Investigations

Effect of Armodafinil on Cortical Activity and Working Memory in Patients with Residual Excessive Sleepiness Associated with CPAP-Treated OSA: A Multicenter fMRI Study

http://dx.doi.org/10.5664/jcsm.3440

Douglas N. Greve, Ph.D.1; Stephen P. Duntley, M.D., F.A.A.S.M.2; Linda Larson-Prior, Ph.D.3; Andrew D. Krystal, M.D., F.A.A.S.M.4; Michele T. Diaz, Ph.D.5; Sean P. A. Drummond, Ph.D.6; Stephen G. Thein, Ph.D.7; Clete A. Kushida, M.D., Ph.D., F.A.A.S.M.8; Ronghua Yang, Ph.D.9; Robert J. Thomas, M.D., F.A.A.S.M.10
1Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Charlestown, MA; 2Washington University School of Medicine, Department of Neurology, St. Louis, MO; 3Washington University School of Medicine, Department of Radiology, St. Louis, MO; 4Duke University Medical Center, Durham, NC; 5Duke University Medical Center, Brain Imaging and Analysis Center, Durham, NC; 6University of California and Veterans Affairs San Diego Healthcare System, San Diego, CA; 7Pacific Research Network, Inc., San Diego, CA; 8Stanford School of Medicine, Stanford Sleep Medicine Center, Redwood City, CA; 9Teva Branded Pharmaceutical Products R&D, Inc., Frazer, PA; 10Beth Israel Deaconess Medical Center, Boston, MA

ABSTRACT

Study Objective:

To assess the effect of armodafinil on task-related prefrontal cortex activation using functional magnetic resonance imaging (fMRI) in patients with obstructive sleep apnea (OSA) and excessive sleepiness despite continuous positive airway pressure (CPAP) therapy.

Methods:

This 2-week, multicenter, prospective, randomized, double-blind, placebo-controlled, parallel-group study was conducted at five neuroimaging sites and four collaborating clinical study centers in the United States. Patients were 40 right-handed or ambidextrous men and women aged between 18 and 60 years, with OSA and persistent sleepiness, as determined by multiple sleep latency and Epworth Sleepiness Scale scores, despite effective, stable use of CPAP. Treatment was randomized (1:1) to once-daily armodafinil 200 mg or placebo. The primary efficacy outcome was a change from baseline at week 2 in the volume of activation meeting the predefined threshold in the dorsolateral prefrontal cortex during a 2-back working memory task. The key secondary measure was the change in task response latency.

Results:

No significant differences were observed between treatment groups in the primary or key secondary outcomes. Armodafinil was generally well tolerated. The most common adverse events (occurring in more than one patient [5%]) were headache (19%), nasopharyngitis (14%), and diarrhea (10%).

Conclusions:

Armodafinil did not improve fMRI-measured functional brain activation in CPAP-treated patients with OSA and excessive sleepiness.

Study Registration:

Double-Blind, Placebo-Controlled, Functional Neuroimaging Study of Armodafinil (200 mg/Day) on Prefrontal Cortical Activation in Patients With Residual Excessive Sleepiness Associated With Obstructive Sleep Apnea/Hypopnea. ClinicalTrials.gov Identifier: NCT00711516. http://www.clinicaltrials.gov/ct2/show/study/NCT00711516

Citation:

Greve DN; Duntley SP; Larson-Prior L; Krystal AD; Diaz MT; Drummond SP; Thein SG; Kushida CA; Yang R; Thomas RJ. Effect of armodafinil on cortical activity and working memory in patients with residual excessive sleepiness associated with CPAP-treated OSA: a multicenter fMRI study. J Clin Sleep Med 2014;10(2):143-153.


Impaired attention/vigilance, executive function, and mood are features of obstructive sleep apnea (OSA), a condition associated with excessive sleepiness, loss of sleep time, sleep fragmentation, intermittent nocturnal hypoxia, blood pressure non-dipping, and impaired cytokine/metabolic regulation.15 Continuous positive airway pressure (CPAP) is effective treatment, yet a significant minority of patients experience persistent sleepiness or cognitive impairment.6,7 While inadequate use, complex sleep apnea, technical factors such as mask leak, and comorbid illnesses such as depression and circadian phase delay may explain some of the residual symptoms, poor recovery of brain function despite excellent treatment compliance is a well-recognized clinical outcome.7,8 Impaired recovery could reflect effects of true brain injury that are not reversible.812

BRIEF SUMMARY

Current Knowledge/Study Rationale: Impaired attention/vigilance, executive function, and mood are features of obstructive sleep apnea (OSA), and a vulnerability of the prefrontal cortex has been consistently demonstrated. The purpose of this study was to assess the effect of armodafinil on task-related prefrontal cortex activation using functional magnetic resonance imaging (fMRI) in patients with OSA and excessive sleepiness despite continuous positive airway pressure (CPAP) therapy.

Study Impact: Armodafinil has consistently been shown to reduce subjective and objective sleepiness in patients with narcolepsy, sleep apnea, shift work sleep disorder, and jet lag, yet effects on brain activation have not been previously described. The findings of a non-response to armodafinil treatment may support the hypothesis that the brain in those with persistent sleepiness despite good sleep apnea treatment shows features of permanent injury and impaired neurocircuitry. Further investigation into this possibility is warranted.

Stimulant and wakefulness-promoting medications have been used with some success in patients with residual sleepiness following sleep apnea treatment. Besides self-treatment with caffeine-containing beverages, these treatments include the amphetamines and their derivatives, as well as modafinil and armodafinil.1317 The brain mechanisms mediating known positive treatment effects are not well understood, but likely involve at least enhancement of dopaminergic systems.18,19 The neuroanatomical substrates and/or functional network correlates underlying the effects of such interventions are not known.

Sleep apnea is associated with altered activation in task-specific neural networks as assessed by functional magnetic resonance imaging (fMRI), with most but not all reports suggesting a reduced activation.2022 While different investigators have used different imaging protocols, the majority have used verbal working memory tasks, and a vulnerability of the prefrontal cortex has been consistently demonstrated.20 The reversibility of anatomical or functional changes with treatment is being demonstrated,23,24 but the reversibility may not be complete. Thus, imaging biomarkers that focus on the prefrontal cortex seem reasonable choices for assessing sleep apnea related pathology and the impact of therapies.

Stimulants have complex effects on activation in relation to task difficulty and sleep deprivation. For example, modafinil has been associated with preservation of cortical and subcortical activation following sleep deprivation,25 with the greatest impact at moderate task difficulty. The amphetamines have shown a “U” shaped effect that interacts with dopamine availability and catecholamine metabolism genotypes.26,27 Our aim was to evaluate task-related fMRI as imaging biomarkers of drug effects. Specifically, we hypothesized that improved wakefulness following use of armodafinil in patients with OSA and residual sleepiness would be associated with increased activation in the executive function network using fMRI. We chose the volume of activation as the primary activation measure. Percent signal change in the activated voxels was the secondary imaging measure. Clinical and neuropsychological outcomes were also assessed as secondary measures. A key secondary outcome was mean response time for the 2-back working memory task that was used during fMRI.

The study evaluated, in a blinded, randomized, prospective, placebo-controlled design, the effect of brain activation during a working memory task, following the use of armodafinil in sleep apnea patients with severe persistent sleepiness despite absence of comorbidities and demonstrating effective use of CPAP therapy.

METHODS

Overview and Subject Selection

This was a 2-week, multicenter, randomized, double-blind, placebo-controlled, parallel-group study with a 1- to 3-week screening period (Figure 1). This study received approval from the appropriate health authorities and the independent ethics committee/institutional review board at each site, and patients provided written informed consent after full explanation of the procedure. The study was conducted in full accordance with the Good Clinical Practice: Consolidated Guidance approved by the International Conference on Harmonisation28 and any applicable national and local laws and regulations. Right-handed or ambidextrous men or women between the ages of 18 and 60 years with a current diagnosis of OSA (apnea-hypopnea index > 15/h of sleep) and a complaint of excessive sleepiness despite effective, stable CPAP usage defined as ≥ 4 h/night on ≥ 70% of nights (objectively verified using data from the CPAP device), were enrolled.

Study design.

jcsm.10.2.143a.jpg

jcsm.10.2.143a.jpg
Figure 1

Study design.

(more ...)

To be eligible, CPAP use must have been stable for ≥ 4 weeks prior to the beginning of the study. Other entry criteria included a mean sleep latency < 8 min on the multiple sleep latency test (MSLT),29 an Epworth Sleepiness Scale (ESS) score ≥ 10 at the initial screening visit,30 and an accuracy ≥ 80% on the 2-back working memory task during the screening period. Exclusion criteria included a history of or current use of central nervous system—active prescription medications, nicotine, chronic caffeine usage averaging more than ∼ 400 mg/d (≥ 4 cups of coffee daily), National Adult Reading Test IQ < 90,31 clinically significant depression or other psychiatric illness, uncontrolled medical condition, diagnosed sleep disorder other than OSA, positive urinary drug screen without medical justification, past or present seizure disorder, history of head trauma, and abnormal physical examination or laboratory parameters. The pre-treatment testing was administered within a week prior to starting armodafinil, and the post-treatment assessments were done on the final day of participation, which included the final scanning.

Drug Treatment

Armodafinil at a 200-mg dose was administered orally, once daily in the morning at or before 08:00, approximately 30 min before the first meal of the day. There was an up-titration period in which armodafinil 50 mg or matching placebo was given on day 1, increased to 100 mg/d on day 2, 150 mg/d on day 5, and 200 mg/d on day 8, which was the dosage that was then continued for the remainder of the double-blind treatment period for a total of 14 days.

Polysomnogram and MSLT

Polysomnographic and 5-session MSLT assessments were performed according to standard guidelines.32,33 Stages, arousals, and respiratory events were scored by polysomnography to confirm the efficacy of CPAP therapy, which was defined as an apnea-hypopnea index ≤ 10 per hour of sleep. Hypopneas were scored when associated with a 3% oxygen desaturation and/or an electroencephalographic arousal. The mean sleep latency was the latency to the first epoch of sleep. Polysomnographic data were used only to confirm inclusion criteria but were not otherwise collected or assessed during the study.

Neuropsychological Assessment

An abbreviated version of the Cambridge Neuropsychological Test Automated Battery (CANTAB) was used.34 Executive function, processing speed, and episodic memory as measured by change from baseline in the performance of selected tasks (Motor Control Task, Reaction Time, Pattern Recognition Memory, and One Touch Stockings of Cambridge [OTS]) from the CANTAB were assessed at week 2 (or last post-baseline observation).

Functional Neuroimaging

Imaging was performed approximately 2 h after the study drug was taken, but no later than 11:00, to minimize circadian effects. Prior to each scanning session, patients were asked to rate their subjective sleepiness on a visual analog scale. Depending upon the recruitment site, imaging was performed with one of six 3 Tesla MRI scanners (4 General Electric and 2 Siemens) using the Functional Biomedical Research Network35 protocol in which the scanning parameters were matched as closely as possible across sites.36,37 Patients' baseline and final assessments were performed at the same site. Anatomical scanning was performed followed by functional scans (4 active-task runs). Whole-brain T1-weighted anatomical scans were acquired at each site for each subject. During anatomic scanning (and before functional runs when anatomic scanning was not performed), a modified continuous 10-min sustained attention task, similar to the psychomotor vigilance test (PVT)38 but without performance feedback, was run to obtain a measure of vigilance in the scanner. For this task, the “+” symbol appeared on a screen at random (mean inter-trial interval of 5 s, range 2-10 s) but disappeared when a button was pressed. The attention tasks were run to obtain a measure of vigilance in the scanner as well as to keep patients awake while functional data were not being collected. For the task paradigm during the functional scanning runs, a block-design protocol was used, with 60 s “on-task” alternating with 30 s “off-task.” Four whole-brain fMRI task runs were acquired for each patient at each visit. Scanning parameters included echo planar imaging, repetition time = 2000 ms, time to echo = 30 ms, flip angle = 77 degrees, bandwidth = 2298 Hz/pixel, 30 slices, slice thickness = 4 mm with a 1-mm gap, 64 × 64 matrix at 220 × 220 mm field of view, 285 time points, 9.5 min, sequential slice order.

Imaging Tasks

A 2-back working memory task was interleaved with a vigilance task. During the 60-s “on-task” periods, random letters were presented in the center of the visual field every 4 s, 15 for each 60-s block, and each stimulus lasted 500 ms. There were a total of 6 active-task blocks. During the 30-s “off-task” periods, the sustained attention task was performed that mirrored the 10-min sustained attention task. The task requires sustained vigilance and is not the same as a 0-back condition, for instance. The need to provide ongoing responses for this attention task also allowed the investigators to be certain that runs used did not contain sleep data—if a subject fell asleep, responses would stop and the scanning would also be stopped for that run.

Imaging Analysis

Anatomical image analysis was performed with Free-Surfer.3941 Specifically, a T1-weighted MP-RAGE (repetition time = 2300 ms, time to echo = 2.94 ms, inversion time = 1100 ms, flip angle = 9 deg, 0.86 × 0.86 × 1.2 mm3) volume was collected for each subject and analyzed in FreeSurfer to create a mesh model of the cortical surface and generate subject-specific cortical42,43 and subcortical44 regions of interest (ROIs), including the anterior cingulate cortex (ACC), posterior parietal cortex (PPC), and thalamus. The dorsolateral prefrontal cortex (DLPFC) was defined as the middle frontal gyrus and sulcus and the inferior frontal sulcus using software-based automatic identification.45 The surface-based ROIs were converted into volume-based ROIs by mapping them back into the anatomical volume and expanding them to fill the cortical ribbon.

The fMRI data were analyzed in the native fMRI volume using the FSL FEAT program in the FMRIB Software Library.46,47 During preprocessing, each fMRI run was 3-dimensionally motion corrected to the middle time point and spatially smoothed by a 5-mm, full-width/half-maximum Gaussian filter for individual patient analysis. Statistical correlation maps were generated using a general linear model, in which the hemodynamic response to the 2-back task was modeled using the FEAT default gamma function and its derivative convolved with a 60-s boxcar. The vigilance periods were modeled as baseline. The FEAT analysis included temporal prewhitening to account for temporal autocorrelation. Significance maps testing 2-back > sustained attention task contrast were computed and thresholded at a voxel-wise uncorrected p < 0.05. The functional maps as well as the middle time point used for motion correction were registered to the subject's anatomical scan.48 This allowed the thresholded maps to be converted into the anatomical volume space where the Free-Surfer volume-based ROIs were defined as described above. The number of voxels with p < 0.05 was counted in each ROI to provide the change in the primary efficacy measure (activation volume in mm3 of the DLPFC) for each subject, visit, and run. A voxel count measure can often be less reliable due to variations in noise.4951 For this reason, we also computed the average percent signal change within the active voxels of the ROI as a secondary measure as this would be less sensitive to noise.52

Approach to Minimize Variability from Multi-institution and Multi-vendor fMRI

Collecting data from different scanning sites introduces the possibility of inconsistency both in the MRI acquisition and in the way the subject is handled. Several steps were taken to reduce these effects as suggested by the fBIRN group53: (1) the field strength (all 3 Tesla) and parameters for the MRI protocols were matched as closely as possible; (2) each site collected data on a standardized phantom which was then used to measure scanner stability and noise36; (3) a single human was scanned at all sites using the complete study protocol just as if he were a real subject, and this person made suggestions as to how the subject experience could be made more consistent across sites; (4) the structural and fMRI data for this person were analyzed to show that similar results were obtained for the same subject across all sites; and (5) the MRI parameters for this data were also checked to assure they conformed to the study design before actual subject scanning began.

Study Measures

The primary efficacy measure was change in activation volume of the DLPFC from baseline to final visit, as measured by the activation volume measured by the number of voxels meeting the predefined threshold in the DLPFC on fMRI. The key secondary measure was mean performance speed, as measured by response latency, at the final visit for each group, on the 2-back working memory task performed during fMRI scanning.

Other secondary measures included (1) percent blood oxygenation level dependent (BOLD) signal intensity change in the DLPFC, ACC, PPC, and thalamus; (2) change from baseline to final visit in subjective efficacy measures, including the CANTAB, a modified continuous 10-min sustained attention task, ESS, Clinical Global Impression of Change (CGI-C),54 and the Medical Outcomes Study 6-Item Cognitive Functioning Scale (MOS-CF6)55; and (3) correlations between the 2-back working memory task response latency and fMRI activation volume and percent signal change in the DLPFC, ACC, PPC, and thalamus.

Drug Tolerability

Tolerability was assessed by the occurrence of adverse events (AEs), concomitant medication usage, and change in vital signs.

Statistical Analysis

Sample Size and Power Considerations

The primary efficacy variable is the change from baseline in the number of voxels meeting predefined threshold in DLPFC on fMRI assessed at endpoint (week 2 or last post-baseline observation). The sample size calculation was based on the study results in Thomas and Kwong.25 In their study, 8 healthy men were enrolled for a single-dose crossover study. Subjects were randomly assigned to 1 of 4 conditions (sleep deprived + a single dose of modafinil or placebo at 06:00, rested + modafinil or placebo). The sleep-deprived condition involved approximately 28 h of continuous wakefulness, and the rested condition included 8 h of scheduled sleep opportunity. Each subject was tested 4 times; all subjects were studied in each condition (32 separate experiments), counter-balanced with a washout period of a minimum of 1 week. The mean difference in the number of voxels was 14,231, and the standard deviation (SD) of the placebo treatment group was 11,107 (which was larger than that of the modafinil treatment group). The sample size calculation assumed an SD of 11,107 and a mean difference of 12,218. The assumption of the mean difference took into account the estimated enrollment rate (about 2 patients per center per month) and the time-line to complete the study while providing adequate power. With a standardized difference of 1.10 (12,218/11,107), 28 evaluable patients (≥ 14 per treatment group) are required to provide 80% power while controlling the 2-sided, type I error rate at 0.05. With an expected (estimated) 25% attrition rate, a total of 38 patients (19 per group) were planned to be enrolled in this study.

Analysis

The primary and key secondary variables were analyzed using analysis of covariance (ANCOVA), with treatment group, center, and sex as the fixed effects, and the baseline value as a potential covariate. The model assumptions for ANCOVA were tested, and if they could not be established, Wilcoxon rank sum test was used. Regarding demographic and baseline characteristics, treatment groups were compared for continuous variables using an analysis of variance with treatment groups as fixed factors. Categorical variables were summarized using descriptive statistics. Treatment groups were compared for all categorical variables using Fisher's exact test (because of small sample sizes).

RESULTS

Baseline Characteristics

A total of 170 patients with OSA were screened for the study; 40 were enrolled and assigned to randomized, double-blind treatment with either armodafinil (n = 21) or placebo (n = 19) (Figure 2). All 40 enrolled patients received at least one dose of study drug and were evaluated for safety; 36 (90%) patients completed the study, were included in the final analysis, and evaluated for efficacy, including 20 patients in the armodafinil group and 16 in the placebo group (Figure 2).

Patient disposition.

jcsm.10.2.143b.jpg

jcsm.10.2.143b.jpg
Figure 2

Patient disposition.

(more ...)

Demographic characteristics were similar between the armodafinil and placebo groups (Table 1). Baseline clinical characteristics were also generally similar, due in part to the study inclusion criteria (Table 1). On the Clinical Global Impression of Severity (CGI-S) scale, the majority of patients were rated as either moderately ill (57% and 42%) or markedly ill (33% and 42%) in the armodafinil and placebo groups, respectively. Mean (standard deviation [SD]) CPAP usage during screening was 6.7 (1.17) h per night for the armodafinil group (n = 19) and 6.1 (1.48) h per night for the placebo group (n = 18). No alcohol use was reported by 48% of patients in the armodafinil group and 68% of those in the placebo group. While no patients in the armodafinil group reported caffeine use during the 48 h prior to baseline, 3 (16%) patients in the placebo group reported consuming between 1 and 199 mg of caffeine during that period (Table 1).

Baseline demographic and clinical characteristics

jcsm.10.2.143.t01.jpg

table icon
Table 1

Baseline demographic and clinical characteristics

(more ...)

Compliance with Treatment

Compliance was assessed again during the run-in period leading to study participation. All subjects were highly compliant, averaging 6.4 h of CPAP use per night during a 1-week period. During the study, no patients discontinued due to noncompliance to either the study medication or procedures.

Polysomnography and MSLT

At baseline, all patients had an apnea-hypopnea index of ≤ 10/h of sleep. Mean (SD) MSLT was 5.4 (1.56) min in the armodafinil group and 6.1 (1.32) min in the placebo group.

Imaging Results

Figure 3 shows the general areas activated by the task. These activation maps were a random-effect group analysis56 (all subjects, all visits) of the 2-back vs. sustained attention task contrast projected onto the inflated cortical surface. The maps were thresholded at p < 0.05, cluster-wise corrected for multiple comparisons (cluster forming threshold at p < 0.01).57 Yellow indicates that the 2-back hemodynamic response amplitude was greater than that of sustained attention task; blue indicates the opposite. The green line indicates the boundary for DLPFC used in the study. In both hemispheres, there was strong task-positive activation in the DLPFC. In addition, there was bilateral task positive activation in the following areas: calcarine sulcus, superior parietal, anterior insula, and the medial wall of the superior frontal area. There was also task-positive activation in the left hemisphere in the motor area (subjects responded with their right hand). There were wide areas of task-negative activation in expected default mode areas: inferior parietal, posterior cingulate, precuneus, and medial frontal, as well as lateral occipital and insula. These results indicated that the brain is well activated in areas expected for this task. Figure 4 shows the same activation maps broken down by group and visit (uncorrected p values thresholded at p < 0.01).

Significance maps of the 2-back vs. sustained attention task contrast (all subjects, all visits) projected onto the inflated cortical surface and thresholded at p < 0.05, cluster-wise corrected for multiple comparisons.

Yellow indicates that the amplitude of the 2-back hemodynamic response was greater than that of sustained attention task; blue/cyan indicates the opposite polarity. The green line is the definition of dorsolateral prefrontal cortex for this study.

jcsm.10.2.143c.jpg

jcsm.10.2.143c.jpg
Figure 3

Significance maps of the 2-back vs. sustained attention task contrast (all subjects, all visits) projected onto the inflated cortical surface and thresholded at p

(more ...)

Significance maps of the 2-back vs. sustained attention task contrast at baseline and final visit for each group.

(A) Group hemodynamic response: right hemisphere lateral view. (B) Group hemodynamic response: right mid-sagittal view. The maps are thresholded at p < 0.01 (uncorrected for multiple comparisons). Red/ yellow indicates that the amplitude of the 2-back hemodynamic response was greater than that of sustained attention task; blue/cyan indicates the opposite polarity.

jcsm.10.2.143d.jpg

jcsm.10.2.143d.jpg
Figure 4

Significance maps of the 2-back vs. sustained attention task contrast at baseline and final visit for each group.(A) Group hemodynamic response: right hemisphere lateral view. (B) Group hemodynamic response: right mid-sagittal view. The maps are thresholded at p

(more ...)

Efficacy

Primary Imaging Efficacy Measure

The primary efficacy measure, activation volume of the DLPFC, was reduced from baseline to final visit in both treatment groups; however, the changes were not significantly different between groups (p = 0.74, Wilcoxon rank sum test; Figure 5). To further explore this finding, a post hoc analysis of image quality was performed by computing signal-to-fluctuation-noise ratio (SFNR) as the mean intensity divided by the time series residual SD.58 An average SFNR value was calculated for the DLPFC. This analysis indicated a reduction in the SFNR from baseline to final visit in both groups (placebo baseline SFNR = 203, final SFNR = 185; armodafinil baseline SFNR = 179, final SFNR = 168) perhaps due to an increase in motion artifact at the final visit.

Mean (SEM [standard error of the mean]) change from baseline in functional activation volume (1 mm3) meeting threshold in DLPFC (dorsolateral prefrontal cortex) on fMRI (functional magnetic resonance imaging).

Change from baseline with armodafinil vs. placebo, p = 0.74, Wilcoxon rank sum test.

jcsm.10.2.143e.jpg

jcsm.10.2.143e.jpg
Figure 5

Mean (SEM [standard error of the mean]) change from baseline in functional activation volume (1 mm3) meeting threshold in DLPFC (dorsolateral prefrontal cortex) on fMRI (functional magnetic resonance imaging).Change from baseline with armodafinil vs. placebo, p = 0.74, Wilcoxon rank sum test.

(more ...)

Secondary Imaging Efficacy Measures

Analysis of the key secondary variable, change from baseline in mean response latency for the 2-back working memory task at the final visit, showed that the mean (SD) changes in latency of 2.3 (78.94) ms for the armodafinil group and -59.0 (112.69) ms for the placebo group were not statistically different (p = 0.17, Wilcoxon rank sum test). A decline in activation volume was observed for the ACC, PPC, and thalamus in each group. The decrease was significant within each group (p < 0.01), but was not significantly different between groups (p < 0.30).

With regard to other secondary variables, no change from baseline was observed in the difference in the BOLD signal between the working memory and the sustained attention task blocks in the DLPFC, ACC, PPC, or thalamus. Exploratory spatial (e.g., voxel-wise) analyses were also conducted. These showed no statistically significant differences, including direct contrast between groups, between visits, or for the interaction of group and visit.

Neuropsychological Testing

In CANTAB measures of cognitive function, there were no substantial changes from baseline to final visit for most tasks in either treatment group (Table 2). For mean “latency to correct” on the OTS task (“difficult” tasks, defined as tasks requiring a 4- to 6-move average), the placebo group had a larger median improvement than the armodafinil group (-8058.0 ms vs. -227.3 ms, respectively; p = 0.02, Wilcoxon rank sum test).

CANTAB results

jcsm.10.2.143.t02.jpg

table icon
Table 2

CANTAB results

(more ...)

Sustained Attention and 2-Back Working Memory Task

Results for the modified continuous 10-min sustained attention task showed that patients in the armodafinil group had numerically faster (although not statistically different) responses. From mean (SD) baseline mean latency values of 343.1 (50.2) ms and 373.3 (73.8) ms for the armodafinil and placebo groups, respectively, mean times decreased to 332.1 (55.3) ms for the armodafinil group while increasing to 377.1 (74.9) ms in the placebo group. At final visit, the mean decrease in time was -17.7 (39.6) ms for the armodafinil group, compared with a mean increase of 4.6 (41.4) ms in the placebo group (p = 0.13, ANCOVA). However, mean response latencies for the 2-back working memory task increased in the armodafinil group from 790.1 (223.6) ms at baseline to 792.4 (214.9) ms at final visit compared with a decrease from 966.5 (322.0) ms at baseline to 907.5 (306.7) ms at final visit for the placebo group. Mean differences were 2.3 (78.9) ms in the armodafinil group and -59.0 (112.7) ms in the placebo group.

Subjective Measures

Among subjective efficacy measures, the CGI-C ratings for excessive sleepiness showed that 65% of patients in the armodafinil group were classified as responders (rated as “minimally,” “much,” or “very much” improved), compared with 56% of the placebo group (p = 0.73, Fisher exact test). The armodafinil group showed greater reduction in excessive sleepiness with a mean (SD) change in ESS score from baseline to final visit of -5.2 (4.53), compared with -3.4 (4.53) for placebo (p = 0.0499, ANCOVA). In addition, patient-reported cognitive functioning as measured by the MOS-CF6 also reflected a mean improvement in the armodafinil group from baseline to final visit of 9.2 (14.82) points compared with a decline of -0.8 (6.86) points for the placebo group, although this difference was not significant (p = 0.12, ANCOVA).

Activation-Performance Correlations

Analysis of correlations between fMRI results in the DLPFC, ACC, PPC, and thalamus and response latency for 2-back working memory task generally showed high variability and few statistically significant correlations (Table 3). In the armodafinil group, however, activation volumes were positively correlated with response latency for all of these regions, both at baseline and final visit. In the placebo group, in contrast, activation volumes were positively correlated at baseline and negatively correlated at final visit.

Correlation coefficients a for fMRI variables and performance on the 2-back working memory task

jcsm.10.2.143.t03.jpg

table icon
Table 3

Correlation coefficients a for fMRI variables and performance on the 2-back working memory task

(more ...)

Tolerability

Armodafinil was found to be generally well tolerated during this trial. During the double-blind treatment period, 13 (62%) patients in the armodafinil group and three (16%) patients in the placebo group reported at least one AE (Table 4). The corresponding number of treatment-related AEs (as judged by the investigator) was nine (43%) for patients in the armodafinil group and three (16%) in the placebo group. Most AEs were mild in severity, and no severe AEs, deaths, or other serious AEs were reported during the study. Two (5%) patients were withdrawn from the study because of AEs, one (5%) in the armodafinil group due to headache and nausea (both of which were considered by the investigator to be of moderate severity and probably related to the study drug) and one (5%) in the placebo group due to urticaria and anxiety (both of which were mild in severity and considered by the investigator to be probably related to the study drug). The most frequently occurring AE in the armodafinil group was headache (Table 4). The only other AEs reported in more than one patient in the armodafinil group were nasopharyngitis and diarrhea. No AEs were reported for more than one patient in the placebo group.

Adverse events occurring in at least 5% of the armodafinil group and at a higher rate than in the placebo group

jcsm.10.2.143.t04.jpg

table icon
Table 4

Adverse events occurring in at least 5% of the armodafinil group and at a higher rate than in the placebo group

(more ...)

DISCUSSION

The key result of our study was that armodafinil treatment for 14 days did not change functional brain activation or improve cognitive performance, in a variety of primary and secondary measures, in treated patients with OSA and persistent excessive sleepiness. The subjects for this study were highly selected, excellent users of CPAP, and free of comorbid illnesses such as depression and circadian rhythm disorders. There was an unexplained reduction in signal/noise during the second scanning session in both treatment and placebo groups, and a reduction in activation volumes in both groups during the second scanning session. These two phenomena are probably related because a reduction in the SFNR will cause a drop in activation volume that exceeds a given p-value threshold. As these were patients well treated with CPAP without associated cardiopulmonary morbidity, it is unlikely that undetected hypercapnia altered the imaging findings.

Armodafinil has consistently been shown to provide clinical benefits in those with narcolepsy, sleep apnea, shift work sleep disorder, and jet lag.1517,5961 Lack of a response to both objective and subjective measures has several possible explanations. The study may have been underpowered given the inter-individual variability in sensitivity to sleepiness seen in OSA patients (for given degrees of sleep apnea severity) and experimental sleep deprivation.62,63 Armodafinil did not show adequate clinical effectiveness in this subset of patients (CGI-C scores were not significantly changed), and the imaging findings may have reflected that fact.

The design of the baseline imaging task may also have contributed to the negative result. The contrast was a 2-back vs. an active attention baseline that closely mimicked the classic PVT.38 It is possible that the cognitive effort required to adequately perform the PVT-like task was so high in these patients with severe persistent sleepiness that they did not have the cognitive resources/reserve to perform the 2-back. That is, both the baseline and 2-back condition may have been about equally difficult. Different results may have been obtained if the baseline was a 0-back or fixation condition.

The unexplained signal/noise factor is an additional possibility. The duration of treatment may have been too short, as it takes several days for armodafinil to reach steady state.64 Although prior imaging work in healthy individuals with modafinil does suggest rapid effects,65,66 patients with OSA may be different. A sluggish BOLD response in sleep apnea from abnormal cerebrovascular reactivity may also be considered, but the imaging findings were consistent with the minimal noted subjective benefits.

The findings of non-response to treatment may support the hypothesis that the brain in those with persistent sleepiness despite good sleep apnea treatment shows features of permanent injury and impaired neurocircuitry.8,67,68 Gray or white matter dysfunction and recovery has implications in adults and children for cognitive outcomes such as age-related cognitive change, cognitive development, affective disorders, and learning/ memory, besides just subjective sleepiness. Additionally, the reversibility of sleep apnea's effects on the body is an important issue, given the spectrum of cardiovascular, metabolic, and neurological abnormalities demonstrable in untreated patients.

Patients in this study were long-term users of CPAP treatment yet were very sleepy subjectively and objectively. The degree of objective sleepiness exhibited was in the range seen in narcolepsy.33 This may very well be an extreme group in a pathological sense, especially given high compliance and polysomnographic treatment efficacy. However, the threshold apnea-hypopnea index of 10, though including the arousal criterion, may have some residual disease and the standard deviations of CPAP use (6.4 ± 1.35), which indicates that about 15% of patients used CPAP for 5 hours or less. Real-life contributions to residual sleepiness include some unprotected (from sleep apnea) sleep time, depression, and behavioral sleep restriction. While both the armodafinil and placebo treatment groups showed a reduction from baseline to final visit in the primary efficacy measure, DLPFC activation volume, no substantial changes were observed in either group in cognitive function. Working memory task performance times in this study were relatively slow, a finding similar to those of an earlier fMRI study20 in which patients with OSA had significantly longer reaction times than healthy controls (908 ms vs. 596 ms; p < 0.02). In that study, although rigorous treatment with CPAP therapy completely resolved subjective sleepiness, there was no significant effect on working memory task performance, suggesting that there may be dissociations between respiratory and cortical recovery in OSA.20

There are numerous experimental data in animal models showing mechanisms of neuronal injury from intermittent hypoxia.11,12,6971 Sleep apnea in humans is associated with several abnormalities that have the potential to cause gray and white matter injury.10,7375 These include nocturnal hypertension/non-dipping blood pressure patterns prior to diagnosis, hypoxia-related injury, metabolic syndrome, and cytokine dysregulation.4,5,10 Sleep deprivation and sleep fragmentation impair hippocampal neurogenesis and long-term potentiation.76,77

Reversibility of brain function or structural impairment/ change is an important clinical issue in sleep apnea. Structural and fMRI and magnetic resonance spectroscopy have been used to assess reversibility. At least partial reversibility has been demonstrated in those who have clinical response to treatment,78 but there is a real concern that full reversibility may not occur. If our study subjects are at one end of the spectrum, the neurological effects of sleep apnea on the brain can be thought of as a continuum, starting at purely functional with reversible deficits, progressing to clinical reversibility but demonstrable persistent neuroanatomical or functional alterations, to a final state of loss of compensatory activity with poor response to stimulants associated with permanent neurological injury. Sleep apnea thus may have the potential to cause lasting brain injury and this population of highly treatment compliant patients with severe objective persistent sleepiness without other comorbid factors may reflect this outcome.

CONCLUSION

Armodafinil did not improve fMRI measures of prefrontal cortical activation during a working memory task or mean performance speed in the 2-back working memory task, compared with placebo, in patients with OSA who were experiencing persistent excessive sleepiness despite effective, stable CPAP treatment. Armodafinil was generally well tolerated.

DISCLOSURE STATEMENT

This study was sponsored by Cephalon, which is now a wholly owned subsidiary of Teva Branded Pharmaceutical Products R&D, Inc., Frazer, Pennsylvania. Funding for editorial support was provided by Teva Branded Pharmaceutical Products R&D, Inc., to The Curry Rockefeller Group, LLC, Tarrytown, New York. Dr. Greve received consulting fees from Cephalon (Teva Branded Pharmaceutical Products R&D, Inc.) for analyzing the data, interpreting results, and helping with the writing of this manuscript. Dr. Krystal has received grants/research support from NIH, Sanofi-Aventis, Cephalon (Teva Branded Pharmaceutical Products R&D, Inc.), GlaxoSmithKline, Merck, Neurocrine, Pfizer, Sunovion/Sepracor, Takeda, Transcept, Phillips-Respironics, Neurogen, Evotec, Astellas, Abbott, and Neuronetics; and has consulted for Abbott, Actelion, Arena, Astellas, Axiom, AstraZeneca, BMS, Cephalon (Teva Branded Pharmaceutical Products R&D, Inc.), Eisai, Eli Lilly, GlaxoSmithKline, Jazz, Johnson and Johnson, King, Merck, Neurocrine, Neurogen, Neuronetics, Novartis, Organon, Ortho-McNeil-Janssen, Pfizer, Respironics, Roche, Sanofi-Aventis, Somnus, Sunovion/Sepracor, Somaxon, Takeda, Transcept, and Kingsdown, Inc. Dr. Drummond serves on the advisory board for and has received investigator-initiated grants from Actelion Pharmaceuticals Ltd. Dr. Thein owns and operates a for-profit clinical trials research clinic; has conducted trials for BMS, Cephalon (Teva Branded Pharmaceutical Products R&D, Inc.), GSK, Medivation, Merck, Martek, MAP, Otsuka, Genentech, Pfizer, Sonexa, XenoPort, Janssen, Lilly, Astellas, and Abbott; and is a paid consultant to Lilly and a paid speaker for United BioSource Corp., Gerson Lehrman Group, PIR, LLC, Guidepoint Global Advisors, KLJ Associates, and Health Research, Inc. Dr. Kushida indicated that his research activities were supported by research contracts between Stanford University and Respironics, ResMed, and Cephalon (Teva Branded Pharmaceutical Products R&D, Inc.); and he has received royalties from patents licensed to Respironics. Dr. Thomas is a co-inventor of the ECG-spectrogram technique that provides sleep and sleep breathing phenotypes based on the analysis of the ECG, and the Positive Airway Pressure Gas Modulator, a device for treating central and complex sleep apnea. Dr. Yang is an employee of Cephalon (Teva Branded Pharmaceutical Products R&D, Inc.). Drs. Duntley, Larson-Prior, and Diaz reported no other potential conflicts of interest.

ACKNOWLEDGMENTS

The authors acknowledge Ryan Dammerman, M.D., Ph.D. (formerly of Cephalon, Inc.) for contributions to study design and data interpretation. Aji Nair, Ph.D., of Teva Branded Pharmaceutical Products R&D, Inc., provided editorial support during the development of this manuscript.

This research was conducted at several sites: Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, MA; Washington University School of Medicine, St. Louis, MO; Duke University Medical Center, Durham, NC; University of California and Veterans Affairs San Diego Healthcare System, San Diego, CA; Pacific Research Network, Inc., San Diego, CA; Stanford School of Medicine, Stanford Sleep Medicine Center, Redwood City, CA; and Beth Israel Deaconess Medical Center, Boston, MA.

REFERENCES

1 

Beebe DW, Groesz L, Wells C, Nichols A, McGee K, authors. The neuropsychological effects of obstructive sleep apnea: a meta-analysis of norm-referenced and case-controlled data. Sleep. 2003;26:298–307. [PubMed]

2 

Malhotra A, author. Obstructive sleep apnoea. Lancet. 2002;360:237–45. [PubMed]

3 

Pankow W, Nabe B, Lies A, Becker H, Köhler U, Kohl FV, et al., authors. Influence of sleep apnea on 24-hour blood pressure. Chest. 1997;112:1253–8. [PubMed]

4 

McNicholas WT, Bonsignore MR, authors. Management Committee of EU Cost Action B26. Sleep apnoea as an independent risk factor for cardiovascular disease: current evidence, basic mechanisms and research priorities. Eur Respir J. 2007;29:156–78. [PubMed]

5 

Sahlman J, Miettinen K, Peuhkurinen K, et al., authors. Kuopio Sleep Apnoea Group. The activation of the inflammatory cytokines in overweight patients with mild obstructive sleep apnea. J Sleep Res. 2010;19:341–8. [PubMed]

6 

Black J, author. Sleepiness and residual sleepiness in adults with obstructive sleep apnea. Respir Physiol Neurobiol. 2003;136:211–20. [PubMed]

7 

Pépin JL, Viot-Blanc V, Escourrou P, et al., authors. Prevalence of residual excessive sleepiness in CPAP-treated apnoea patients: the French multicentre study. Eur Respir J. 2009;33:1062–7. [PubMed]

8 

Santamaria J, Iranzo A, Montserrat M, de Pablo J, authors. Persistent sleepiness in CPAP treated obstructive sleep apnea patients: evaluation and treatment. Sleep Med Rev. 2007;11:195–207. [PubMed]

9 

Beebe DW, Gozal D, authors. Obstructive sleep apnea and the prefrontal cortex: towards a comprehensive model linking nocturnal upper airway obstruction to daytime cognitive and behavioral deficits. J Sleep Res. 2002;11:1–16

10 

Simmons MS, Clark GK, authors. The potentially harmful consequences of untreated sleep-disordered breathing: the evidence supporting brain damage. J Am Dent Assoc. 2009;140:536–42. [PubMed]

11 

Veasey SC, Davis CW, Fenik P, et al., authors. Long-term intermittent hypoxia in mice: protracted hypersomnolence with oxidative injury to sleep-wake brain regions. Sleep. 2004;27:194–201. [PubMed]

12 

Veasey SC, Zhan G, Fenik P, Pratico D, authors. Long-term intermittent hypoxia: reduced excitatory hypoglossal nerve output. Am J Respir Crit Care Med. 2004;170:665–72. [PubMed]

13 

Black JE, Hirshkowitz M, authors. Modafinil for treatment of residual excessive sleepiness in nasal continuous positive airway pressure-treated obstructive sleep apnea/ hypopnea syndrome. Sleep. 2005;28:464–71. [PubMed]

14 

Hirshkowitz M, Black J, authors. Effect of adjunctive modafinil on wakefulness and quality of life in patients with excessive sleepiness-associated obstructive sleep apnoea/hypopnoea syndrome: a 12-month, open-label extension study. CNS Drugs. 2007;21:407–16. [PubMed]

15 

Roth T, White D, Schmidt-Nowara W, et al., authors. Effects of armodafinil in the treatment of residual excessive sleepiness associated with obstructive sleep apnea/hypopnea syndrome: a 12-week, multicenter, double-blind, randomized, placebo-controlled study in nCPAP-adherent adults. Clin Ther. 2006;28:689–706. [PubMed]

16 

Hirshkowitz M, Black JE, Wesnes K, Niebler G, Arora S, Roth T, authors. Adjunct armodafinil improves wakefulness and memory in obstructive sleep apnea/ hypopnea syndrome. Respir Med. 2007;101:616–27. [PubMed]

17 

Roth T, Rippon GA, Arora S, authors. Armodafinil improves wakefulness and long-term episodic memory in nCPAP-adherent patients with excessive sleepiness associated with obstructive sleep apnea. Sleep Breath. 2008;12:53–62. [PubMed Central][PubMed]

18 

Qu WM, Huang ZL, Xu XH, Matsumoto N, Urada Y, authors. Dopaminergic D1 and D2 receptors are essential for the arousal effect of modafinil. J Neurosci. 2008;28:8462–9. [PubMed]

19 

Zolkowska D, Jain R, Rothman RB, et al., authors. Evidence for the involvement of dopamine transporters in behavioral stimulant effects of modafinil. J Pharmacol Exp Ther. 2009;329:738–46. [PubMed]

20 

Thomas RJ, Rosen BR, Stern CE, Weiss JW, Kwong KK, authors. Functional imaging of working memory in obstructive sleep-disordered breathing. J Appl Physiol. 2005;98:2226–34. [PubMed]

21 

Ayalon L, Ancoli-Israel S, Aka AA, McKenna BS, Drummond SP, authors. Relationship between obstructive sleep apnea severity and brain activation during a sustained attention task. Sleep. 2009;32:373–81. [PubMed]

22 

Ayalon L, Ancoli-Israel S, Drummond SP, authors. Altered brain activation during response inhibition in obstructive sleep apnea. J Sleep Res. 2009;18:204–8. [PubMed Central][PubMed]

23 

Castronovo V, Canessa N, Strambi LF, et al., authors. Brain activation changes before and after PAP treatment in obstructive sleep apnea. Sleep. 2009;32:1161–72. [PubMed]

24 

Canessa N, Castronovo V, Cappa SF, et al., authors. Obstructive sleep apnea: brain structural changes and neurocognitive function before and after treatment. Am J Respir Crit Care Med. 2011;183:1419–26. [PubMed]

25 

Thomas RJ, Kwong K, authors. Modafinil activates cortical and subcortical sites in the sleep-deprived state. Sleep. 2006;29:1471–81. [PubMed]

26 

Mattay VS, Goldberg TE, Fera F, et al., authors. Catechol O-methyltransferase val158-met genotype and individual variation in the brain response to amphetamine. Proc Natl Acad Sci U S A. 2003;100:6186–91. [PubMed Central][PubMed]

27 

Huotari M, Garciá-Horsman JA, Karayiorgou M, Gogos JA, Männistö PT, authors. D-amphetamine responses in catechol-O-methyltransferase (COMT) disrupted mice. Psychopharmacology (Berl). 2004;172:1–10.

28 

International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) adopts Consolidated Guideline on Good Clinical Practice in the Conduct of Clinical Trials on Medicinal Products for Human Use. Int Dig Health Legis. 1997;48:231–34. [PubMed]

29 

Carskadon MA, Dement WC, Mitler MM, Roth T, Westbrook PR, Keenan S, authors. Guidelines for the multiple sleep latency test (MSLT): a standard measure of sleepiness. Sleep. 1986;9:519–24. [PubMed]

30 

Johns MW, author. A new method for measuring daytime sleepiness: the Epworth Sleepiness Scale. Sleep. 1991;14:540–5. [PubMed]

31 

Nelson HE, author. The National Adult Reading Test (NART): test manual. Windsor, UK: NFER-Nelson, 1982. [PubMed]

32 

Kushida CA, Littner MR, Morgenthaler T, et al., authors. Practice parameters for the indications for polysomnography and related procedures: an update for 2005. Sleep. 2005;28:499–521. [PubMed]

33 

Thorpy MJ, author. The clinical use of the multiple sleep latency test. Sleep. 1992;15:268–75. [PubMed]

34 

Robbins TW, James M, Owen JM, Sahakian BJ, McInnes L, Rabitt P, authors. Cambridge neuropsychological test automated battery (CANTAB): a factor analytic study of a large sample of normal elderly volunteers. Dementia. 1994;5:266–81. [PubMed]

35 

Biomedical Informatics Research Network. Accessed April 27, 2012. http://www.nbirn.net.

36 

Greve DN, Mueller BA, Liu T, et al., authors. A novel method for quantifying scanner instability in fMRI. Magn Reson Med. 2011;65:1053–61. [PubMed Central][PubMed]

37 

Brown GG, Mathalon DH, Stern H, et al., authors. Multisite reliability of cognitive BOLD data. Neuroimage. 2011;54:2163–75. [PubMed Central][PubMed]

38 

Dinges DF, Powell JW, authors. Microcomputer analyses of performance on a portable, simple visual RT task during sustained operations. Behav Res Methods Instrum Comput. 1985;17:652–5.

39 

Dale AM, Fischl B, Sereno MI, authors. Cortical surface-based analysis. I. Segmentation and surface reconstruction. Neuroimage. 1999;9:179–94. [PubMed]

40 

Fischl B, Sereno MI, Dale AM, authors. Cortical surface-based analysis. II: inflation, flattening, and a surface-based coordinate system. Neuroimage. 1999;9:195–207. [PubMed]

41 

FreeSurfer software. Athinoula A = Martinos Center for Biomedical Imaging. Accessed April 27, 2012. http://surfer.nmr.mgh.harvard.edu.

42 

Fischl B, van der Kouwe A, Destrieux C, et al., authors. Automatically parcellating the human cerebral cortex. Cereb Cortex. 2004;14:11–22. [PubMed]

43 

Desikan RS, Segonne F, Fischl B, et al., authors. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage. 2006;31:968–80. [PubMed]

44 

Fischl B, Salat DH, Busa E, et al., authors. Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron. 2002;33:341–55. [PubMed]

45 

Destrieux C, Fischl B, Dale A, Halgren E, authors. Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature. Neuroimage. 2010;53:1–15. [PubMed Central][PubMed]

46 

Woolrich MW, Ripley BD, Brady M, Smith SM, authors. Temporal autocorrelation in univariate linear modeling of FMRI data. Neuroimage. 2001;14:1370–86. [PubMed]

47 

FSL 4.1 software. FMRIB Software Library. Accessed April 27, 2012. http://www.fmrib.ox.ac.uk/fsl.

48 

Greve DN, Fischl B, authors. Accurate and robust brain image alignment using boundary-based registration. Neuroimage. 2009;48:63–72. [PubMed Central][PubMed]

49 

McGonigle DJ, Howseman AM, Athwal BS, Friston KJ, Frackowiak RS, Holmes AP, authors. Variability in fMRI: an examination of intersession differences. Neuroimage. 2000;11(6 Pt 1):708–34. [PubMed]

50 

Cohen MS, DuBois RM, authors. Stability, repeatability, and the expression of signal magnitude in functional magnetic resonance imaging. J Magn Reson Imaging. 1999;10:33–40. [PubMed]

51 

Kimberley TJ, Khandekar G, Borich M, authors. fMRI reliability in subjects with stroke. Exp Brain Res. 2008;186:183–90. [PubMed]

52 

Smith SM, Beckmann CF, Ramnani N, et al., authors. Variability in fMRI: a re-examination of inter-session differences. Hum Brain Mapp. 2005;24:248–57. [PubMed]

53 

Glover GH, Mueller BA, Turner JA, et al., authors. Function biomedical informatics research network recommendations for prospective multicenter fMRI studies. J Magn Reson Imaging. 2012;36:39–54. [PubMed Central][PubMed]

54 

National Institute of Mental Health: 028 CGI. Guy W, editor. Clinical Global Impressions. ECDEU assessment manual for psychopharmacology. Rockville: National Institute of Mental Health, 1976: p. 217–22.

55 

Ware JE Jr, Sherbourne CD, authors. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;30:473–83. [PubMed]

56 

Friston KJ, Holmes AP, Worsley KJ, authors. How many subjects constitute a study? Neuroimage. 1999;10:1–5. [PubMed]

57 

Hagler DJ Jr, Saygin AP, Sereno MI, authors. Smoothing and cluster thresholding for cortical surface-based group analysis of fMRI data. Neuroimage. 2006;33:1093–103. [PubMed Central][PubMed]

58 

Friedman L, Glover GH, authors. FBRIN Consortium. Reducing interscanner variability of activation in a multicenter fMRI study: controlling for signal-to-fluctuation-noise-ratio (SFNR) differences. Neuroimage. 2006;33:471–81. [PubMed]

59 

Harsh JR, Hayduk R, Rosenberg R, et al., authors. The efficacy and safety of armodafinil as treatment for adults with excessive sleepiness associated with narcolepsy. Curr Med Res Opin. 2006;22:761–74. [PubMed]

60 

Czeisler CA, Walsh JK, Wesnes KA, Arora S, Roth T, authors. Armodafinil for treatment of excessive sleepiness associated with shift work disorder: a randomized, controlled study. Mayo Clin Proc. 2009;84:958–72. [PubMed]

61 

Rosenberg RP, Bogan RK, Tiller JM, et al., authors. A phase 3, double-blind, randomized, placebo-controlled study of armodafinil for excessive sleepiness associated with jet lag disorder. Mayo Clin Proc. 2010;85:630–8. [PubMed]

62 

Koutsourelakis I, Perraki E, Bonakis A, Vagiakis E, Roussos C, Zakynthinos S, authors. Determinants of subjective sleepiness in suspected obstructive sleep apnoea. J Sleep Res. 2008;17:437–43. [PubMed]

63 

Van Dongen HP, Vitellaro KM, Dinges DF, authors. Individual differences in adult human sleep and wakefulness: leitmotif for a research agenda. Sleep. 2005;28:479–96. [PubMed]

64 

Darwish M, Kirby M, Hellriegel ET, Yang R, Robertson P Jr, authors. Pharmacokinetic profile of armodafinil in healthy subjects: pooled analysis of data from three randomized studies. Clin Drug Investig. 2009;29:87–100. [PubMed]

65 

Volkow ND, Fowler JS, Logan J, et al., authors. Effects of modafinil on dopamine and dopamine transporters in the male human brain: clinical implications. JAMA. 2009;301:1148–54. [PubMed Central][PubMed]

66 

Minzenberg MJ, Watrous AJ, Yoon JH, Ursu S, Carter CS, authors. Modafinil shifts human locus coeruleus to low-tonic, high-phasic activity during fMRI. Science. 2008;322:1700–2. [PubMed]

67 

Antczak J, Popp R, Hajak G, Zulley J, Marienhagen J, Geisler P, authors. Positron emission tomography findings in obstructive sleep apnea patients with residual sleepiness treated with continuous positive airway pressure. J Physiol Pharmacol. 2007;58 Suppl 5:25–35. [PubMed]

68 

Tonon C, Vetrugno R, Lodi R, et al., authors. Proton magnetic resonance spectroscopy study of brain metabolism in obstructive sleep apnoea syndrome before and after continuous positive airway pressure treatment. Sleep. 2007;30:305–11. [PubMed]

69 

Gozal D, Daniel JM, Dohanich GP, authors. Behavioral and anatomical correlates of chronic episodic hypoxia during sleep in the rat. J Neurosci. 2001;21:2442–50. [PubMed]

70 

Row BW, Kheirandish L, Neville JJ, Gozal D, authors. Impaired spatial learning and hyperactivity in developing rats exposed to intermittent hypoxia. Pediatr Res. 2002;52:449–53. [PubMed]

71 

Pae EK, Chien P, Harper RM, authors. Intermittent hypoxia damages cerebellar cortex and deep nuclei. Neurosci Lett. 2005;375:123–8. [PubMed]

72 

Row BW, Kheirandish L, Cheng Y, Rowell PP, Gozal D, authors. Impaired spatial working memory and altered choline acetyltransferase (CHAT) immunoreactivity and nicotinic receptor binding in rats exposed to intermittent hypoxia during sleep. Behav Brain Res. 2007;177:308–14. [PubMed Central][PubMed]

73 

Kamba M, Inoue Y, Higami S, Suto Y, Ogawa T, Chen W, authors. Cerebral metabolic impairment in patients with obstructive sleep apnoea: an independent association of obstructive sleep apnea with white matter change. J Neurol Neurosurg Psychiatry. 2001;71:334–9. [PubMed]

74 

Macey PM, Henderson LA, Macey KE, et al., authors. Brain morphology associated with obstructive sleep apnea. Am J Respir Crit Care Med. 2002;166:1382–7. [PubMed]

75 

Joo EY, Tae WS, Lee MJ, et al., authors. Reduced brain gray matter concentration in patients with obstructive sleep apnea syndrome. Sleep. 2010;33:235–41. [PubMed]

76 

Guzman-Marin R, Bashir T, Suntsova N, Szymusiak R, McGinty D, authors. Adult hippocampal neurogenesis is reduced by sleep fragmentation in the adult rat. Neuroscience. 2007;148:325–33. [PubMed Central][PubMed]

77 

Mueller AD, Pollock MS, Lieblich SE, Epp JR, Galea LAM, Mistlberger RE, authors. Sleep deprivation can inhibit adult hippocampal neurogenesis independent of adrenal stress hormones. Am J Physiol Regul Integr Comp Physiol. 2008;294:R1693–703. [PubMed]

78 

Ferini-Strambi L, Baietto C, Di Gioia MR, et al., authors. Cognitive dysfunction in patients with obstructive sleep apnea (OSA): partial reversibility after continuous positive airway pressure (CPAP). Brain Res Bull. 2003;61:87–92. [PubMed]