Nov 14 2011

Bedside Test for Consciousness

The diagnosis of coma, specifically persistent vegetative state (PVS), is not as straightforward as it might seem. The current standard of care is to perform a thorough neurological exam in order to assess function. The limitation here, however, is that we are inferring brain functioning from what the patient can do. Some parts of the neurological exam are straightforward, like pupillary reflexes. Reflexes are a more direct way of interrogating the nervous system because the examiner is able to see the response, and therefore the integrity, of one specific pathway.  This is very useful in determining which parts of the nervous system are working and which are damaged.

Coma, however, is a condition of impaired consciousness – so we also need to determine, to the best of our ability, the exact level of consciousness of a patient. This is where it can get a little tricky. The parts of the exam that are most useful for determining level of consciousness are response to various types of stimuli and the ability to follow commands. If I say to a patient (without any non-verbal cues), “show me two fingers on your right hand,” and they then raise two fingers on their right hand, that tells me a lot about their function. They can hear, they can understand language sufficiently to understand the command, they are aware enough to make sense of the command and the context of the exam (they know I want them to do something), and they have the motor pathways necessary to move their fingers on the right hand.

The flip side of that, however, is that if the patient is unable to show me two fingers it could mean that they are deaf, aphasic (impaired language), delirious, or paralyzed in that hand. I cannot know for sure that their lack of response is due to impaired level of consciousness. We can compensate for this somewhat by giving various commands – using various body parts, language that is easiest to understand, and as loud as possible to be heard. But this is only a partial compensation. A patient who is locked-in, for example (conscious but paralyzed below the eyes), would give the same lack of response as someone who is in a PVS.

There is also the potential to overcall non-specific responses. Patients in a PVS still have sleep-wake cycles. They open their eyes and look around, and may grimace or move their limbs. They do not, however, give any specific response to their environment. It is not uncommon, however, for family members, or inexperienced health care providers, to misinterpret non-specific movements as if they were responses to the environment. By chance alone the random behaviors of the patient in a PVS may correlate with something happening in their environment, like someone speaking to them. So to be sure they are really responsive there needs to be a specific and repeatable response to stimuli.

The ability to reliably diagnoses PVS vs other conditions that may mimic PVS is important in order to give an accurate diagnosis and prognosis to family members and caretakers. There is also the possibility that if a person in an apparent PVS actually has some awareness, that it may be possible to communicate with them if we can get access to their consciousness despite their physical limitations. There is therefore the need for more sensitive and specific tests for consciousness, to supplement the neurological exam.

I have written here previously about research into this problem, mainly using two diagnostic technologies – EEG and functional MRI scanning. These are both functional tests that look at brain activity. The theory behind recent research is that perhaps a person in an apparent PVS can have a specific response to a command that we can detect with EEG or fMRI, even though they are unable to signal their awareness with physical movement. In the recent issue of The Lancet is the latest in this line of research.

Cruse et. al. used EEG monitoring on 16 patients who, by clinical exam, are in a PVS, and 12 healthy controls. They asked the subjects to imagine themselves clenching their right hand, and then separately wiggling their toes. In 9 of the 12 healthy control the researchers were able to detect specific EEG responses in their motor cortex to the two distinct commands. That is a bit disappointing in terms of the sensitivity of the test. In 75% of the healthy controls the EEG activity reflected the specific commands they were given, but in 25% the EEG could not detect this.

I find this disappointing because, at present, I feel the most useful application of such diagnostic tests to supplement the exam is to confirm that a patient is truly in a PVS, facilitating decisions about long term care. With a 25% false negative rate, however, a negative test will not sufficiently confirm PVS in order reassure families that they are, for example, making the right decision in limiting care.

In patients in the study in PVS by clinical exam, 3 could “repeatedly and reliably generate appropriate EEG responses to two distinct commands, despite being behaviourally entirely unresponsive.” This result is similar to prior research showing that some patients in apparent PVS show signs of awareness when EEG or fMRI scanning is used.

While we can know for certain what the false negative rate is from this study, because we can know for certain that the healthy controls were, in fact, conscious, we cannot know what the false positive rate is. In other words, we cannot yet be sure what these results mean. It implies that there is more going on in the brains of those three subjects than their exam would indicate, and that’s an interesting result in itself. But we cannot know if that specific brain activity is a true indicator of conscious awareness. There is a great deal of subconscious processing in the brain, and perhaps those processes are disinhibited in a patient with impaired consciousness.

In other words – this and other studies indicate specific brain activity in response to commands, but that does not necessarily mean that subjects have any meaningful awareness of that activity.

The results of this and similar studies still need further validation. It would be nice to go back to the same set of patients and see if the results are reproducible over time, by examiners blinded to the previous results. It would also be a good indicator that the results are meaningful if they correlate to some other marker or predictor of severity of damage. What do the brains of those PVS subjects who showed responsiveness look like? Are there other reasons to think their damage may not be as severe as those in a true PVS, or perhaps they have a different kind of damage.

The only pattern to emerge so far is that patients with traumatic brain injury seem to have a greater chance of showing responsiveness than those with non-traumatic brain injury. This study showed that as well, with 20% of traumatic and 9% of non-traumatic subjects showing a response (but the small number of subjects in the study makes those numbers statistically unreliable). This does make sense, however, because non-traumatic injuries (like anoxic injury – lack of oxygen to the brain) tend to cause damage to the whole brain. Traumatic injury, however, can damage certain parts of the brain while leaving others relatively unharmed, creating the potential for paralysis in excess of loss of consciousness.

Before these new techniques become standard, however, I think that researchers need to indicate that they mean something practical. Specifically the results should make some prediction about prognosis or response to rehabilitation or a specific intervention. If they do not predict anything about outcome, that would call into question the significance of the results. Perhaps the technology is just detecting small but irrelevant differences in the residual brain function of those with serious damage.

While this technology is hopeful and should be pursued, it will likely be years before we understand what the true implications of this research are.


9 responses so far