A New Tool for Answering “How Much Time Is Left?”: The Corneal Reflex
In this article
Can the Final 24 Hours Be Predicted?
Study: A prospective observational study of 112 patients with advanced cancer
One of the questions most frequently asked by families in hospice and palliative care, and one of the hardest to answer, is: “How much time is left?” A new study by a hospice physician from South Korea, published in BMJ Supportive & Palliative Care, draws attention to a simple examination finding that could provide a medically clearer answer to this question: the loss of the corneal reflex. According to the study, the likelihood of death within 24 hours in patients with advanced cancer whose corneal reflex is completely absent is 5,48 times higher than in those with a preserved reflex. This finding offers an indicator that can be easily assessed at the bedside to help clinicians and families manage the final hours with greater awareness. But an important note at the outset: patients who have not lost this reflex may also be on the verge of death — this test does not mean “absent/present—wait”, but rather “if absent, probably soon”.
What Is the Corneal Reflex? Where Does It Originate, and What Does It Mean?
The corneal reflex is a neural reflex defined as the involuntary closing of the eyelid when the eye's transparent front layer is touched lightly. This seemingly simple reflex actually operates through the pons , one of the most sensitive structures of the brainstem. The corneal stimulus is detected by the fifth cranial nerve (trigeminal nerve — ophthalmic branch), reaches the center in the pons, and closes the eyelid through the seventh cranial nerve (facial nerve). A preserved reflex therefore means that brainstem functions at the level of the pons are largely preserved.
KEY CONCEPT: THE CORNEAL REFLEX AND THE BRAINSTEM
For half a century, the corneal reflex has been used in neurological practice as an indicator in the determination of brain death (brain death). In patients in a coma, the corneal reflex is assessed to check the integrity of brainstem functions. What is new in this study, however, is the systematic investigation of the corneal reflex as a tool that could be used to predict the final hours of the natural dying process in patients with cancer . The same neurophysiological principle (the degree of brainstem function) is thus being used as a clinical assessment tool, this time not in a comatose patient in intensive care, but in a patient with advanced cancer in a hospice bed.
Study Design: Simple and at the Bedside
Kang and his team enrolled 112 patients with advanced cancer at the hospice center of Gyeongsang National University Hospital. Their estimated life expectancy was 1-2 weeks, and the median age was 73,5. The study used the following protocol:
- Assessment by trained nurses 3 times a day
- Approaching from the side (to avoid visual stimulation)
- Lightly touching the cornea with a strand of sterile cotton/gauze
- Only patients with RASS -3 or lower, to avoid causing discomfort
- Assessment in three categories (intact / diminished / absent)
- An asymmetric reflex (intact on one side, diminished on the other) was coded as “diminished”
- Total: 112 patients with advanced cancer
- Median age: 73,5
- Life expectancy: 1-2 weeks
- 110 participants died within 7 days
- Conducted in a hospice setting
- Single center, South Korea
Assessing brainstem (pons) function with a simple bedside examination
Striking Results: An Absent Reflex and 24 Hours
The analysis results can be summarized as follows: in patients whose corneal reflex was completely absent, the chance of dying within 24 hours was 5,48 times higher than in those with a preserved reflex. The 24-hour mortality rate among patients with an absent reflex was calculated as 70,7%. However, as Kang emphasized, this is not an “absolute prediction”, but a strong supporting finding: “The absence of the reflex can strongly support the prediction of imminent death, but its preservation does not rule out imminent death.” A present reflex therefore does not mean the patient will not die soon — but if it is absent, the likelihood is very high.
Comparison with Existing Indicators: Where Does It Fit?
In palliative care, clinicians have monitored signs of approaching death for decades. These signs are commonly called “late signs” and include:
Death Rattle
The characteristic sound produced by secretions accumulating in the upper airways. It has high specificity but low sensitivity.
Mandibular Breathing
Breathing accompanied by up-and-down movements of the jaw — also known as “fish-mouth” breathing; it occurs in the final hours.
Peripheral Cyanosis
Bluish discoloration of the fingertips, lips, and tip of the nose — a sign of circulatory insufficiency.
RASS -4/-5
Richmond Agitation-Sedation Scale: -4 (deeply drowsy) or -5 (unresponsive). These scores were found to be the most sensitive among 13 indicators.
An important point emphasized by the researchers is that the existing signs share high specificity but low sensitivity . When these signs appear, death is very close; however, many dying patients may never display them. When these “late signs” are detected, the likelihood of death within 48 hours is around 95%. Kang's research suggests that the corneal reflex could narrow this 48-hour window to 24 hours .
The study's most practical recommendation is to use two indicators together: first identify “entry into the dying phase” using RASS, then assess a more specific time window using the corneal reflex. Among deeply sedated patients (RASS -4/-5), the rate of death within 24 hours was 71,2% in those with an absent corneal reflex, compared with 37,1% in those with a preserved reflex — a twofold difference.
An Expert Perspective from MD Anderson
Dr. David Hui, director of supportive and palliative care research at MD Anderson Cancer Center, who was not involved in the study, considers the results comparable to other “late signs” in terms of sensitivity and specificity. However, he also highlights limitations: the small sample size and potential selection bias. The study was conducted only in patients with RASS -3 or lower — a choice made to avoid causing discomfort to more conscious patients, but one that limits the generalizability of the findings.
Another question Hui wonders about is: how many patients lacked both the corneal reflex and pupillary response, and how many lacked only one? Could assessing these two together further strengthen prediction? This is a practical question for future studies to answer.
Why Is This So Important? The Cultural and Human Dimension
The starting point of Kang's study was a human observation rather than a scientific one. In his own words: “In Korea, family members place great importance on being beside the patient at the moment of death. This leads to difficult and urgent questions about how much time is left. In practice, I have often encountered limitations when trying to answer these questions using existing clinical signs.”
This cultural dimension is also extremely familiar in Turkish society. Having loved ones beside a patient in their final moments — seeking mutual forgiveness, offering prayers, recording final words, and allowing them to feel the touch of people they love — is central to both Islamic and Turkish cultural traditions. Families' question “how much time is left?” often arises not from scientific curiosity but from the wish to call a sibling who lives far away, wait for a nephew or niece arriving from abroad, or be prepared. Even a clinician saying “probably within 24 hours” instead of “within 48 hours” can make a vital difference for families.
THE FUNDAMENTAL ETHICAL PRINCIPLE OF PALLIATIVE CARE
The fundamental principle emphasized by Kang and accepted by all palliative care researchers is: “The dignity and comfort of the dying patient must always take precedence over invasive sampling or procedures.” One reason the corneal reflex test has attracted so much interest is that it causes almost no discomfort — a light touch with sterile gauze requires no invasive procedure.
Future Directions: Multicenter and Supported by Artificial Intelligence
Kang states that a multicenter follow-up study in a different population is planned. He envisages using the corneal reflex test as a confirmatory step after assessing RASS and peripheral cyanosis. Another of Kang's hopes is that artificial intelligence will help detect subtle patterns clinicians miss at the bedside — especially facial microexpressions, changes in breathing patterns, and other parameters that are difficult to assess subjectively.
As legal frameworks for euthanasia and end-of-life care change around the world (medically assisted dying practices in Canada, the Netherlands, Belgium, and a growing number of US states), demand for bedside prognostic tools is increasing. However, as Kang points out, research conducted in patients on the verge of death involves unique ethical complexities — the legitimacy of every procedure must be balanced against the comfort or information it provides for the patient.
Critical Assessment: Points to Consider
Although the results are promising, several important limitations of the study need careful consideration:
🔴 Small Sample and a Single Center
Only 112 patients, from a single hospice center in South Korea. The findings need to be replicated in different cultural and clinical settings. Demographic homogeneity is a limitation.
🔴 Selection Bias (Selection Bias)
Only patients with RASS -3 or lower (that is, very drowsy/unresponsive) were included. This is an ethically appropriate choice to avoid discomfort in more conscious patients, but it limits the generalizability of the results to patients at an earlier stage.
🟡 Limited Sensitivity
As Kang clearly states: “Preservation of the reflex does not rule out imminent death.” This is therefore not a “negative rule-out test”, but a “positive confirmation test”. A substantial proportion of patients with a preserved reflex may also die within 24-48 hours.
🟡 Factors Beyond the Patient's Dying Process
The corneal reflex can be affected by eye diseases (keratitis, corneal ulcers), brain metastases, certain medications (particularly high-dose opioids and benzodiazepines), and neurological comorbidities. These factors were not fully controlled in the study.
🟡 The “How Early Is Death?” Paradox
29% of patients with an absent corneal reflex lived not just 24 hours, but longer . This test therefore also has a margin of error, and telling a family “within 24 hours” still involves uncertainty.
Applicability in Turkish Practice
This simple, inexpensive approach that can be used at the bedside is highly suitable for hospice and palliative care practice in Turkey. Palliative care services in Turkey have developed considerably over the past fifteen years — the palliative care structure initiated by the Ministry of Health in the early 2010s led to the opening of hundreds of palliative care units nationwide. However, there are substantial differences in the quality of practice between centers.
Several practical implications follow. First, trained palliative care nurses can easily learn this test and integrate it into routine assessment — the South Korean study performed it three times a day. In Turkey's three-shift care system, assessment once per shift also makes sense. Second, the ability to give families more accurate information about how much time remains is very well suited to the cultural needs of Turkish society — families attach great importance to the cultural tradition of seeking mutual forgiveness and being present in the final moments. Third, this approach can also be used as a teaching tool — teaching it to medical students and resident physicians during palliative care rotations helps them perceive the dying process as a normal clinical process.
Caution: this test should not be used to “make certain” the time frame communicated to the family. When speaking with families, flexible, honest, and empathetic language such as “probably within the next 24 hours” should be maintained. Definitive statements such as “definitely 24 hours” can undermine the family's trust if the prediction is wrong.
🎓 DROZDOGAN Academy — Commentary and Discussion
CORNEAL REFLEX · PALLIATIVE CARE · PROGNOSTIC TOOLS · BRAINSTEM · END-OF-LIFE CARE
The Importance of This Study
The “final hours”, often peripheral to oncology practice but perhaps the most important time for patients and families, are an area insufficiently examined with scientific tools and attention. Oncology literature is filled with new treatment approvals, phase III trial results, and molecularly targeted therapies — yet the question “how can we care well for a dying patient?” remains in the shadow of this intense scientific output. Kang's study is a small but valuable part of the effort to fill precisely this gap. Showing that a neurological examination finding used for half a century to diagnose brain death may also be meaningful in predicting the final hours of the natural dying process is both conceptually elegant and practically valuable. Most importantly, this approach can be used without any invasive procedure and is simple enough for resource-limited palliative care settings anywhere in the world. The message for oncology practitioners is that the most valuable clinical tools are not always the newest or most technological.
Strengths
The study has several important strengths. First, a research question arising from one's own clinical experience — Kang noticing the loss of the corneal reflex in dying patients as a hospice physician and deciding to investigate it systematically — is one of science's most productive sources of inspiration. Second, the prospective observational design gives the results internal validity. Third, three daily assessments performed in a standardized manner by trained nurses increase the reliability of the findings. Fourth, the contribution shown when used with RASS (71,2% vs 37,1%) demonstrates that the corneal reflex is a tool to be used with other indicators, rather than alone. Fifth, a leading expert such as Dr. David Hui from MD Anderson assessing the findings and offering critical contributions shows the study's capacity to engage in scientific dialogue. Finally, the researcher's persistent emphasis on ethical responsibilities — stating that the dying patient's dignity and comfort take precedence over any research value — is an exemplary attitude in palliative care research.
Limitations and Open Questions
The study's limitations also need consideration. First, a single-center sample of 112 patients is insufficient for generalization — validation is particularly lacking outside Asia and in different ethnic and clinical contexts. Second, the inclusion only of patients with RASS -3 or lower leaves the validity of the test in more conscious patients unknown — families' request to “assess the reflex” may actually arise most often in this intermediate group, which the study does not examine. Third, sensitivity remaining “moderate” — preservation of the reflex does not rule out imminent death — is an important limitation in clinical decision-making; a family cannot simply say “the reflex is present, so we have time”. Fourth, the effects on the corneal reflex of medications heavily used in end-of-life care, such as opioids and benzodiazepines, were not fully assessed — high doses may falsely suppress the reflex. Fifth, an effect of eye diseases on the corneal reflex was not excluded (neurotrophic keratopathy can also occur in patients with brain metastases), which may affect the accuracy of the findings. Sixth, as Dr. Hui noted, evaluating other brainstem indicators such as pupillary response together might produce a stronger prognostic tool — but this combination was not tested. Finally, how a finding of this kind should be communicated to families — should one say “within 24 hours”, or is “in the coming days” more appropriate? — remains an open communication issue.
Implications for Clinical Practice
Several concrete messages can be drawn from this study for oncology and palliative care practitioners in Turkey. First, daily assessment of the corneal reflex in patients with advanced cancer is a simple approach requiring no additional cost, time, or technology and can be added to routine palliative care examinations. Second, this test should not be used alone — it should be assessed holistically alongside other signs such as RASS, peripheral cyanosis, breathing patterns, and urine output. Third, when informing families, probabilistic language rather than definitive language should be used — such as “probably within the next day”. Fourth, a short training program for palliative care nurses can add this test to the standard assessment checklist; palliative care nursing is a growing field in Turkey, and simple tools of this kind strengthen nurses' clinical assessment repertoire. Fifth, teaching family members this test for patients receiving home care is debatable — on the one hand it could help families avoid missing the final moments, but on the other it could create a care culture focused on “testing”; cultural sensitivity is needed here. Sixth, this study is a valuable example reminding us that oncology's role is not only to extend life but also to “help achieve a good death” — a good oncologist is a physician who not only knows the chemotherapy regimen but also understands the final hours and can manage them with compassion. Finally, as emphasized in our previous article on cancer-related distress, family distress is also critical in end-of-life care, and every prognostic assessment made for the patient should help the family cope with uncertainty.
Questions for the Future
The critical questions for the coming period are: when will Kang's planned multicenter validation study with a more diverse population be completed, and will its results hold in different cultural settings such as Europe, America, and Turkey? Could a combination of the corneal reflex + pupillary response + other brainstem indicators (such as the gag reflex and vestibulo-ocular reflex) produce a stronger prognostic score? Will the same finding apply to dying patients without cancer — those with end-stage heart failure, end-stage chronic obstructive pulmonary disease, or neurodegenerative diseases? How will the effects of opioids, benzodiazepines, and medications used for palliative sedation on the corneal reflex be controlled? Could AI-assisted image analysis identify microexpressions and fluctuations in the reflex that clinicians fail to detect at the bedside? What are the psychological and ethical effects of teaching family members this test in home care — does it offer practical help or create a new burden of grief? Could palliative care centers in Turkey establish a national registry in which prognostic indicators of this kind are assessed systematically, and could this increase the international visibility of Turkish palliative care research? Finally, while improving service to families, might a physician's increased ability to “predict death” also reinforce the medicalization of death — to what extent should the natural, sacred, and unpredictable aspects of human death be preserved? This philosophical and ethical question will continue in the background of all scientific developments.
Scientific Sources
- Kang JH et al. Corneal reflex as a predictor of imminent death in advanced cancer patients. BMJ Supportive & Palliative Care. Published online: April 2026.
- Medscape Medical News. A Simple Test Could Predict a Patient's Final 24 Hours. Lisa O'Mary, 14 April 2026.
- Hui D, dos Santos R, Chisholm G et al. Clinical signs of impending death in cancer patients. Oncologist 2014;19(6):681–687.
- Sessler CN et al. The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med 2002;166(10):1338–1344.
- Hui D et al. Bedside clinical signs associated with impending death in patients with advanced cancer: preliminary findings of a prospective, longitudinal cohort study. Cancer 2015;121(6):960–967.
- Slotman E, Hedman C, Fransen HP et al. Pain at the end of life in patients with cancer: a population-based study on prevalence, relief, and the role of pain assessment. Support Care Cancer 2026; doi:10.1007/s00520-026-10349-y
- Wijdicks EFM. The diagnosis of brain death. N Engl J Med 2001;344(16):1215–1221.
- Ellershaw J, Ward C. Care of the dying patient: the last hours or days of life. BMJ 2003;326(7379):30–34.
- Republic of Turkey Ministry of Health. Procedures and Principles for the Implementation of Palliative Care Services. Official Gazette, 2014 and subsequent updates.
This article was created using various editorial tools, including artificial intelligence, as part of the preparation process. It was reviewed by human editors before publication.
Was this article helpful?
Your feedback helps us improve our content.
Reader comments
Share your comment
Your name and comment are published only after administrator approval. Your email address is never shown to visitors.
Be the first to comment.