Zolpidem Treatment of Disorders of Consciousness

By: John Whyte, MD, PhD. Director, Moss Rehabilitation Research Institute

For many years it was assumed by clinicians and researchers that patients suffering from prolonged disorders of consciousness (DOC, i.e., the vegetative and minimally conscious states) were not candidates for any meaningful treatments. This pessimism was adopted by insurers as well, who typically exclude such patients from active rehabilitation programs. In recent years, however, glimmers of optimism have begun to emerge, fueled by several interrelated lines of research.

The formal definition of the minimally conscious state (MCS) in 2002,[1] while not directly altering prognosis, encouraged clinicians and researchers to scrutinize the cognitive status of patients with DOC, and highlighted the problem of misdiagnosis. It is now clear that as many as 40% of patients carrying a diagnosis of the vegetative state (VS) are actually in the MCS.[2] The problem of misdiagnosis, in turn, has highlighted the value of assessment tools and expertise in trying to lower this error rate.

At the same time, a range of research findings have suggested that even some patients who meet current behavioral definitions of VS (i.e., who are accurately diagnosed according to current criteria) may have covert evidence of consciousness, obtainable through functional imaging,[3] or electrophysiologic assessment methods.[4] Such findings raise the question of whether it is appropriate to “warehouse” patients with a diagnosis of VS in environments without the expertise to carefully assess their mental processing.

Finally, there are intriguing reports of treatments that may enhance the level of consciousness of certain patients with DOC. Although the reports, to date, are mostly in the form of single subject experiments and case reports,[5, 6] they carry more weight than other reports of that type because the probability of significant clinical change years after injury in the absence of an effective intervention is very low.

These recent research findings appropriately encourage further treatment studies, but it is unrealistic to hope that any treatment will be of benefit to all patients with DOC. The VS, in particular, can be thought of as a “floor” condition, meaning that there it has an upper boundary that differentiates it from MCS, but no lower boundary other than brain death. Thus, in principle, the clinical condition of VS may include patients with virtually no viable cerebrum and others with much more restricted damage. No pharmacologic or electrical stimulation treatment can be expected to substitute for viable neural circuitry. Thus, it seems that one of the most important priorities for research is to develop methods for identifying subgroups of VS patients who differ in their underlying neuropathology and, accordingly, who differ in their ability to respond to emerging treatments.

Zolpidem and Restoration of Consciousness

One intriguing treatment possibility for chronic DOC is zolpidem, an omega-1 GABA agonist commonly prescribed to healthy individuals for the treatment of insomnia. In 2006, Clauss and Nel reported on 3 patients in the chronic VS who regained consciousness within an hour of receiving a single dose of zolpidem, and who returned to the VS several hours later as the drug wore off.[5] The authors report that repeated drug administration could maintain consciousness in these individuals, one of whom had an anoxic injury and 2 of whom suffered traumatic accidents. Although these descriptions were persuasive, they suffer from publication bias and leave unanswered the question of how frequent such a dramatic drug effect might be. Over the last few years, case reports of other positive responses to zolpidem of impairments such as aphasia, gait disorders, and akinetic mutism also appeared, suggesting a more generalized drug effect.[7-11]

Based on these initial reports, Robin Myers and I conducted a pilot study of 15 consecutive patients with DOC of any etiology, funded by a grant from the Pennsylvania Department of Health. We tested an unselected series of 15 individuals and identified one clear drug responder, while demonstrating that the remaining 14 subjects showed no sign of clinical change on the drug.[12]

Zolpidem’s Mechanism of Action in DOC

How zolpidem exerts this restorative effect, and who it does so for some patients but not others remains unclear. This is not a crucial practical question, since a test dose of zolpidem is inexpensive and low risk. However, understanding the neural factors associated with reversible vs. irreversible DOC is a crucial research problem which will likely be relevant to selection of subjects for a wide range of treatment trials and for predicting differential natural recovery.

A few patients have been studied with functional imaging on and off zolpidem to elucidate the drug’s effects.[e.g., 7, 13] Not surprisingly, patients who are drug responders show greater brain activity on drug than off. But this doesn’t really answer the question of mechanism. Indeed, since conscious behavior is dependent on an active brain state it would be very suprising if patients’ functional imaging results were the same on and off the drug (or, while conscious vs. unconscious). This is referred to in imaging research as the “performance confound” – when behavioral performance in the scanner differs greatly between 2 groups or 2 drug conditions, one expects major differences in brain activation and one can’t tell if those differences are causes of the behavioral difference or the result of the behavioral difference.

One way around the performance confound is to examine associations between pre-treatment brain function and the results of treatment. That is, since the status of the brains of drug responders and non-responders must differ prior to receiving the drug, which particular brain differences best predict who will be a drug responder? In order to conduct such studies, however, one needs to assemble a group of drug responders and a group of non-responders in order to explore how they differ. That, in turn, returns us to the question of the frequency with which a therapeutic response to zolpidem occurs.

The Current Study

We are now funded by the National Institute on Disability and Rehabilitation Research to conduct a larger study on this topic. The current study has two distinct but interrelated aims. The first is to test the drug on a larger sample (100 subjects) in order to obtain a more accurate estimate of the frequency of a clinical response to the drug. The second, assuming a reasonable rate of drug response, is to assemble a group of drug responders and a group of non-responders who are otherwise clinically comparable, and to study their neuropathology and neurophysiology in search of explanatory differences.

The current study, being conducted nationwide, is organized into 3 phases. In the first phase, after review of medical records and abstracting of relevant clinical data, subjects are studied at home by their caregivers. Caregivers are mailed two blinded doses of the study drug (zolpidem and placebo) and are coached by telephone in conducting a 3-hour assessment after the drug dose on 2 different days. The results of their observations are reported on a structured assessment form and reviewed blindly by study personnel for a possible drug response. Those who demonstrate substantial behavioral differences on the 2 days are classified as “possible drug responders” and invited to participate in Phase II. Phase II is identical in design except that a trained clinician performs the 2 assessments using the Coma Recovery Scale-Revised to obtain a numerical performance score. Those individuals who show a substantial difference between the 2 sessions in Phase II have their drug codes broken and, if their superior performance occurred on the zolpidem day in both Phase I and Phase II, they are considered “definite drug responders.” It is from phases I and II that we will calculate the rate of clinically meaningful drug response and examine whether there are any clinically evidence predictors of that response, such as etiology, time post-injury, etc.

For Phase III, we will invite up to 10 definite drug responders and a clinically similar sample of non-responders to travel to Philadelphia and be admitted to the Clinical and Translational Research Center at the University of Pennsylvania. During their 3-day research stay, both responders and non-responders will be intensively studied primarily in the unmedicated state, to assess relevant differences in their baseline neural structure and function. These studies will include sensitive assessment of brain structure and function.  Brain structural measures will be obtained via MRI and include voxel based morphometric measures of focal and diffuse atrophy, and diffusion tensor imaging measures of white matter integrity. Brain function will be assessed with both functional MRI and event related potentials. Specifically, we will assess patterns of resting perfusion via arterial spin labeling, and regional activation in response to passive language stimulation via BOLD MRI imaging. We will also examine resting connectivity by examining spontaneous fluctuation of the BOLD signal at rest. An ERP protocol will assess response to various levels of “deviant” auditory stimuli, ranging from acoustic deviants to complex semantic deviants, and we will also gather resting EEG spectra.  Although Phase III is largely exploratory in nature, we hope to identify differences between groups in baseline brain structure and function that can then serve as a priori hypotheses to be explored in subsequent studies.


Although zolpidem is clearly beneficial to some individuals with DOC, we do not at present know how many or what kinds of patients can benefit. This information can help assess the practical utility of the drug as well as shed light on mechanistically relevant differences within the DOC population that may shed light on a wide range of treatment opportunities.


  1. Giacino, J., S. Ashwal, N. Childs, et al., The minimally conscious state: definition and diagnostic criteria.Neurology, 2002. 58: p. 349-353.
  2. Andrews, K., L. Murphy, R. Munday, et al., Misdiagnosis of the vegetative state: retrospective study in a rehabilitation unit. British Medical Journal, 1996. 313(7048): p. 13-16.
  3. Owen, A.M., M.R. Coleman, M. Boly, et al., Detecting awareness in the vegetative state. Science, 2006. 313(5792): p. 1402.
  4. Kotchoubey, B., S. Lang, G. Mezger, et al., Information processing in severe disorders of consciousness: Vegetative state and minimally conscious state. Clinical Neurophysiology, 2005. 116: p. 2441-2453.
  5. Clauss, R. and W. Nel, Drug induced arousal from the permanent vegetative state. NeuroRehabil, 2006. 21: p. 23-28.
  6. Schiff, N.D., J.T. Giacino, K. Kalmar, et al., Behavioural improvements with thalamic stimulation after severe traumatic brain injury. Nature, 2007. 448: p. 600 - 604.
  7. Brefel-Courbon, C., P. Payoux, O. Fabierine, et al., Clinical and imaging evidence of zolpidem effect in hypoxic encephalopathy. Ann Neurol, 2007. 62: p. 102-105.
  8. Clauss, R.P., M.M. Sathekge, and H.W. Nel, Transient improvement of spinocerebellar ataxia with zolpidem. New Engl J Med, 2004. 351: p. 511-512.
  9. Cohen, L., B. Chaaban, and M.O. Habert, Transient improvement of aphasia with zolpidem. New England Journal of Medicine, 2004. 350(9): p. 949-50.
  10. Jarry, C., J.P. Fontenas, A.P. Jonville-Bera, et al., Beneficial effect of zolpidem for dementia. Ann Pharmacother, 2002. 36(11): p. 1808.
  11. Shadan, F.F., J.S. Poceta, and L.E. Kline, Zolpidem for postanoxic spasticity. South Med J, 2004. 97(8): p. 791-2.
  12. Whyte, J. and R. Myers, Incidence of clinically significant responses to zolpidem among patients with disorders of consciousness: A preliminary placebo controlled trial. American Journal of Physical Medicine & Rehabilitation, 2009. 88: p. 410-418.
  13. Clauss, R.P. and W.H. Nel, Effect of zolpidem on brain injury and diaschisis as detected by 99mTc HMPAO brain SPECT in humans. Arzneim-Forsch./Drug Research, 2004. 54(10): p. 641-646.