IPM Take
In intensive care, prognosis is never just a number.
It shapes rehabilitation plans, family conversations, clinical confidence and the decisions made when nobody can yet know what a damaged brain may recover. A new pupil-response measure may offer another window into that uncertainty. But the point is not to replace one imperfect prediction with a more sophisticated one.
It is to stop pretending that a bedside assessment made today can always see the future clearly.
Executive Summary
Researchers from Copenhagen University Hospital Rigshospitalet and the Technical University of Denmark presented prospective cohort findings at the 2026 European Academy of Neurology Congress suggesting that a little-used phase of the pupil response, the late light-off response, may predict improvement in consciousness seven days later after acute brain injury.
The study included 250 ICU patients with impaired consciousness after traumatic or non-traumatic brain injury, alongside 30 age- and sex-matched healthy controls. Patients underwent daily automated pupillometry and serial neurological examinations for up to 20 ICU days.
Late light-off response latency independently predicted improvement in Full Outline of UnResponsiveness scores seven days later after adjustment for baseline neurological status, ICU day, time since injury, sedation and injury cause. By contrast, standard measures including the Neurological Pupil Index and conventional pupillary light-reflex latency did not predict later gains in consciousness in this cohort.
The findings come from a conference abstract and require larger multicentre validation before they can support routine prognostic use.
Why it matters
- Patients / advocates: A better estimate of recovery potential could matter deeply for families living through the uncertainty of severe brain injury. But no single measure should be allowed to close down hope, rehabilitation or careful reassessment.
- Clinicians: The signal is intriguing because it may reveal something not visible in routine examinations. It is not yet a tool for definitive prognosis or treatment-limitation decisions.
- Hospitals / providers: The technology is already available in many ICUs and takes around 13 seconds per eye. Implementation, however, would require standardisation, clinical training and safeguards against over-interpreting an early biomarker.
A pupil test can look almost trivial.
A handheld device. A short flash of light. A number on a screen. But in intensive care, small measurements can carry enormous weight when the question is whether a patient may regain consciousness after severe brain injury.
The Danish team focused on what happens after the light goes off. Their analysis separated the pupil’s immediate reaction from a later dilation phase known as the late light-off response. In 250 patients with acute disorders of consciousness, the timing of this later response predicted improvement in consciousness seven days later, even after accounting for baseline neurological status, sedation and injury type.
That matters because the measures clinicians usually see on the screen may not be telling the whole story. In this cohort, the Neurological Pupil Index and standard pupillary light-reflex latency did not predict subsequent gains in consciousness. The late response was also not associated with a patient’s responsiveness on the same day, raising the possibility that it captures recovery potential that routine bedside assessment cannot yet see.
But this is exactly where medicine has to stay disciplined.
A prognostic signal is not a verdict. It is not proof that a person will recover, nor a licence to treat uncertainty as certainty because a machine makes it look precise. The subgroup findings were exploratory, the work comes from a single research setting and the authors themselves call for larger multicentre studies.
Still, the political question is already clear. When the consequences of prognostic error are so human, health systems need tools that make clinical judgement more humble, not more absolute.
