Cognitive testing in early-phase clinical trials: Development of a rapid computerized test battery and application in a simulated Phase I study

https://doi.org/10.1016/j.cct.2006.10.010Get rights and content

Abstract

Background

Inclusion of cognitive assessment in Phase I trials of novel pharmaceutical agents may help identify subtle yet meaningful CNS effects early in clinical development, and lead to a greater understanding of the pharmacokinetic/pharmacodynamic relationship prior to entering pivotal late-phase trials.

Aims

To examine issues surrounding the inclusion of a computerised cognitive test battery in Phase I clinical trials.

Methods

A 12-minute battery of five computerized cognitive tasks was administered to 28 healthy males in a double-blind, single ascending dose study using three doses of midazolam (0.6 mg, 1.75 mg and 5.25 mg) with placebo insertion. Subjects were enrolled and assessed at two Phase I units. Statistical analyses sought to determine the sensitivity of the test battery to sedation-related cognitive dysfunction, any between-site differences in outcome, and also the effects of repeated test administration (i.e., practice or learning effects).

Results

There were no significant differences in data collected between sites. All standard safety measurements were completed. No substantial technical issues were noted. No learning effects were observed on four of the five cognitive tasks. ANOVA comparing baseline to post-baseline results revealed significant cognitive deterioration on all five cognitive tasks 1 h following administration of 5.25 mg midazolam. The magnitude of these changes were very large according to conventional statistical criteria. Smaller but significant changes were observed on a subset of memory and learning tasks at 1 h post-dosing in 1.75 mg condition, and at 2 h post-dosing in the 5.25 mg condition.

Conclusions

The cognitive test battery was well tolerated by subjects and research unit staff. The tests demonstrated minimal learning effects, were unaffected by language and cultural differences between sites, and were sensitive to the sedative effects of midazolam. Inclusion of this cognitive test battery in future studies may allow identification of cognitive impairment or enhancement early in the clinical development cycle.

Introduction

For drug therapies that penetrate the Central Nervous System (CNS), cognitive effects are most often evaluated in Phase II to IV clinical trials conducted in target patient groups. While cognitive assessments are not routinely administered in first-in-human Phase I trials, a number of manuscripts describe cognitive outcome in such trials [1], [2], [3], [4], [5]. These studies typically employ a series of standard ‘paper-and-pencil’ tests assessing a range of cognitive functions. These may include measures of motor function (e.g., finger tapping, simple reaction time), attention (e.g., choice reaction time, digit-symbol substitution), memory (e.g., auditory verbal learning task) and executive function (e.g., Stroop task). Inclusion of cognitive testing early in clinical drug development may allow early identification of clinically meaningful CNS effects (which in principle may be adverse or beneficial) and a greater understanding of the pharmacokinetic/pharmacodynamic relationship prior to entering pivotal later-phase trials [1].

Other potential benefits are also evident if the cognitive tests administered have metric properties that are adequate for making statistical decisions about cognitive changes in individuals or small groups of subjects (i.e., no range restriction, interval level outcome data, normal distribution, high reliability, minimal practice effects; [6], [7], [8]). If a test is responsive to cognitive change, it is possible to examine the time-course of drug effects and determine the magnitude of cognitive change at Cmax for the compound under investigation at various doses (e.g., [7]). If the data derived from a test includes estimates of mean and variability in an individual’s performance, it is possible to determine cognitive response to a drug in individual subjects (e.g., [8]). This is important in Phase I trials that routinely enroll only a small number of subjects, and may allow classification of individuals as cognitive ‘responders’ or ‘non-responders’. Further, a responsive test yielding appropriate data allows comparison of cognitive outcome with the results of adverse event recordings, leading to a greater understanding of the safety and tolerability of a new drug.

There are unique aspects of Phase I clinical trials that make the application of cognitive testing challenging in this environment. For example, many conventional neuropsychological tests do not have equivalent alternate forms, resulting in substantial practice effects when these tests are administered serially [9]. These tests may also suffer from other limitations (e.g., range restriction, skewed data distributions, interval level data) that hinder their ability to identify subtle changes in individuals, and therefore limit their use in trials involving only a small number of subjects [8]. The specialist nature of these tasks may also mean that trial staff must undergo substantial training in test administration procedures. The time available in an early-phase trial between blood sampling and safety measures is often limited. Most such trials are now tightly scheduled meaning that introduction of cognitive pharmacodynamic tests has a cost (e.g., larger staffing required per trial) as well as a potential benefit. In addition, for accurate inferences to be made between exposure and cognitive effect during administration and subsequent clearance of a pharmaceutical agent, the time taken for cognitive evaluation must be restricted. Thus, there is a clear need for a cognitive test battery that can be rapidly performed. A typical ‘paper–pencil’ cognitive test battery may require 30 to 60 min to administer (e.g., [2]). Cognitive tests with such administration times may not be easily applied at multiple time-points throughout the trial, restricting the inferences that can be drawn on the basis of data derived from these tests. Also, the paper-based nature of many neuropsychological tasks means that data derived from these tests cannot easily be integrated with electronic data capture (EDC) systems, introducing the potential for transcription error and time delays in performing real-time monitoring of such data or upon study completion.

An area that is expected to benefit from application of cognitive testing early in clinical development will be the development of compounds for the treatment of neuropathic pain. For example, patients with neuropathic pain who are prescribed gabapentin commonly complain of somnolence, dizziness and confusion, and these appear commonly as adverse events in clinical trials with agents of this type. In choosing a test agent to probe the effectiveness of a cognitive test battery, we have noted that effects on psychomotor function have been detected with the benzodiazepine midazolam (‘Hypnovel’). This agent has well-known sedative properties that result in CNS side-effects including drowsiness, confusion, amnesia and fatigue. Midazolam has also been shown to affect performance on cognitive tests [10], [11]. For example, an oral dose of 0.075 mg/kg results in significant decrement in performance of a computerized maze learning task between 30 and 60 min post-dosing (Lawendy et al, manuscript in preparation).

This study aimed to identify the practical implications of including a 12-minute computerized cognitive test battery in a Phase I clinical trial, including issues of compliance and examination of between-site differences. Psychometric issues such as learning effects and repeatability/stability were also examined. A further aim was to examine the sensitivity of the test battery cognitive change following administration of the sedative-hypnotic midazolam.

Section snippets

Subjects

Twenty-eight healthy male volunteers between the ages of 18 and 55 were recruited for the study. Subjects were recruited from two different sites in Brussels (Site 1; N = 12) and Singapore (Site 2; N = 16). The age, height, weight and body mass index of the group is described in Table 1. All subjects were in good health as determined by medical history, physical examination, vital signs, ECG, and clinical laboratory measurements, and all had a body weight of greater than 50 kg, and a body mass

Results

All 28 subjects completed all stages of the study. During the study, one subject left the research unit to attend the funeral of a relative. However, this individual subsequently completed all study assessments. All planned pharmacokinetic and safety measurements were completed. A total of 56 study drug related adverse events were noted throughout the trial. These generally occurred in a dose-dependent manner with the majority of these events involving either fatigue (N = 12) or somnolence (N = 

Discussion

Inclusion of cognitive outcomes in early phase clinical trials may enhance understanding of the efficacy and/or safety of new chemical entities in clinical development. However, there are numerous practical and scientific issues surrounding the application of cognitive testing in early phase trials (e.g., long administration times, sensitivity to subtle changes, availability of alternate forms of tests, integration with electronic data capture systems). For the current study a 12-minute series

References (24)

  • P.J. Snyder et al.

    A method for determining the magnitude of change across different cognitive functions in clinical trials: the effects of acute administration of two different doses of alprazolam

    Hum Psychopharm

    (2005)
  • C.M. Mollica et al.

    Development of a statistical approach to classifying treatment response in individual children with ADHD

    Hum Psychopharm

    (2004)
  • Cited by (67)

    • The Vitamins in Psychosis Study: A Randomized, Double-Blind, Placebo-Controlled Trial of the Effects of Vitamins B<inf>12</inf>, B<inf>6</inf>, and Folic Acid on Symptoms and Neurocognition in First-Episode Psychosis

      2019, Biological Psychiatry
      Citation Excerpt :

      The Global Assessment of Functioning scale (21) was used to determine global functioning. The 11-test neurocognition battery included several tasks from the computerized CogState test battery (34–36) and a range of paper-and-pencil tasks. Six neurocognitive domains were assessed: speed of processing [Trail Making Test–Part A (37) and Simple Reaction Time], attention/vigilance (Identification Test and Monitoring Test), working memory [One-Back Test and Digit Sequencing from the Brief Assessment of Cognition in Schizophrenia (38)], executive function (Trail Making Test–Part B (37) and Mazes subtest from the Neuropsychology Assessment Battery (39)], visual learning and memory [Brief Visuospatial Memory Test–Revised (40) and Associate Learning], and verbal learning and memory [California Verbal Learning Test-II (41) total 1–5 trials].

    • The effects of vortioxetine on cognitive performance in working patients with major depressive disorder: A short-term, randomized, double-blind, exploratory study

      2018, Journal of Affective Disorders
      Citation Excerpt :

      Both measures were assessed at baseline and at week 8; the DSST in addition at week 1. Secondary efficacy assessments comprised the following cognitive performance tests obtained at baseline and at weeks 1 and 8: the Trail Making Test-A and B (TMT-A, TMT-B) (Lezak, 1983), the Stroop Color Naming Test (STROOP) Congruent and Incongruent (Stroop, 1992), Simple Reaction Time (SRT), and Choice Reaction Time (CRT) (Collie et al., 2007). All cognitive tests were administered using paper and pencil except the computerized tests of SRT and CRT.

    View all citing articles on Scopus

    This research was funded by Pfizer Inc. The authors would like to thank Caroline Wooldridge and Dik Ng for expert operational support and the Erasme and Singapore Pfizer Phase I Unit staff for performing this study. For the duration of this study, Dr Collie and Dr Maruff were staff scientists at CogState Ltd, a cognitive test development company that maintains several contractual arrangements with Pfizer Inc, for the provision of their clinical technologies. All other authors were staff scientists at Pfizer Inc throughout the design and conduct of this study.

    View full text