How this instrument works
The ICH score sums five independent predictors of death within 30 days of a spontaneous intracerebral hemorrhage: level of consciousness, hemorrhage volume, whether blood has entered the ventricles, whether the bleed sits below the tentorium, and age. Glasgow Coma Scale contributes 0 points for a score of 13-15, 1 point for 5-12, or 2 points for 3-4. A CT-measured volume of 30 cubic centimeters or more adds a point. Intraventricular hemorrhage adds a point. An infratentorial origin — brainstem or cerebellum rather than the cerebral hemispheres — adds a point. Age 80 or older adds a final point. The total runs from 0 to 6.
Hemphill and colleagues built the scale in 2001 from 152 patients treated at a single hospital, then checked that it held up in a separate group. Every input was chosen because it can be read off a head CT and a bedside neurological exam within the first hour a patient arrives — nothing on the list waits on a lab result, an MRI slot, or a specialist's note. That was deliberate: the scale exists to hand a family and a treating team a rough, early figure to anchor a difficult conversation, not a diagnosis that needs a day of workup to produce.
The honest caveat concerns how that figure gets used afterward. Because a high score correlates with poor outcomes, it can tempt a team toward earlier withdrawal of aggressive care — and once care is withdrawn, the patient's actual trajectory can never be observed, which risks the score becoming a self-fulfilling prophecy rather than a genuine forecast. A 2001 Neurology study found that in a cohort with unusually low rates of early withdrawal, patients with moderate scores survived far more often than the original mortality figures suggested. The scale was validated to open a conversation, not to end one.
- Set Glasgow Coma Scale to the band matching the exam: 13-15, 5-12, or 3-4.
- Toggle ICH volume ≥ 30 cm³ once a CT-based estimate exists — this site's ICH Volume (ABC/2) instrument produces that figure from three CT measurements.
- Toggle Intraventricular hemorrhage present if blood is visible in the ventricular system on CT.
- Toggle Infratentorial origin if the hemorrhage centers in the brainstem or cerebellum rather than the cerebral hemispheres.
- Toggle Age 80 or older, then read the ICH score, 0 to 6, as the sum of all five entries.
Worked example — GCS 9, volume 34 cm³, IVH present
A patient arrives with a Glasgow Coma Scale of 9, in the 5-12 band, worth 1 point. CT shows a hemorrhage volume of 34 cm³, over the 30 cm³ line, adding 1 more. Blood is visible in the ventricles — intraventricular hemorrhage — adding a third point. The bleed sits above the tentorium and the patient is 67, under 80, so both remaining entries add 0. The sum is 1 + 1 + 1 + 0 + 0 = 3, roughly the middle of the 0-6 range.
The two extremes show the scale's shape more starkly. Every entry at its best — GCS 13-15, volume under 30 cm³, no IVH, supratentorial, under 80 — sums to 0, and in Hemphill's original cohort not one patient scoring 0 died within 30 days. Every entry at its worst sums to 6; no patient in that cohort actually reached a 6, but every patient who scored 5 or higher died within the month, which is why a 6 carries essentially the same grave outlook by extension.
Questions
Does an ICH score of 0 guarantee survival?
No — it means the original 152-patient cohort behind the scale recorded zero deaths at that score within 30 days, which is reassuring but not a guarantee for any individual. A score of 0 still describes a patient with a small, supratentorial bleed, near-normal consciousness, and no ventricular extension, all genuinely favorable signs. Treat it as a strong prior, not a certainty, and keep watching the patient rather than the number.
Why does the scale use only information from the first CT scan?
Because it was built for a conversation that often has to happen within hours, not days. Every one of the five inputs — coma scale, hemorrhage volume, ventricular extension, location, age — comes from the initial exam and imaging, before any lab result, follow-up scan, or specialist consult arrives. That immediacy is the entire design goal: a number a treating physician can produce and discuss with a family on the day of admission.
Should a high ICH score be used to justify withdrawing care?
Not alone, and this is the scale's most cited limitation. A 2001 Neurology study by Becker and colleagues found that in a hospital cohort with an unusually low rate of early care withdrawal, observed mortality at moderate scores ran well below the original figures — evidence that aggressive early treatment itself changes the outcome the score is trying to predict. Hemphill's own later guideline work echoes this: use the score to inform discussion, not to substitute for full clinical assessment over time.
Where does the hemorrhage volume in this calculator come from?
Almost always from the ABC/2 method — three linear measurements taken off the CT slice with the largest hemorrhage area, multiplied and halved to approximate an ellipsoid's volume. This site's separate ICH Volume calculator produces that figure directly from those three measurements; feed the result here as the volume≥30 cm³ toggle.
Is the ICH score still the standard tool in 2026?
It remains the most widely validated and most widely taught grading scale for spontaneous intracerebral hemorrhage, and it appears in the 2015 American Heart Association / American Stroke Association management guideline. Newer variants exist — the max ICH score and the FUNC score among them — aimed at refining specific subgroups, but none has displaced the original as the common bedside reference.
Does the ICH score apply to a bleed caused by trauma or a ruptured aneurysm?
No. It was derived and validated for spontaneous, primary intracerebral hemorrhage — bleeding that starts within the brain tissue itself, typically from small-vessel disease or hypertension. Traumatic hemorrhage and aneurysmal subarachnoid hemorrhage follow different courses and are scored with entirely different tools; applying this scale to either would extend it well past the population it was built on.
References
- Hemphill JC 3rd et al. 2001, Stroke — original ICH score derivation (PubMed)
- Hemphill JC 3rd et al. 2015, Stroke — AHA/ASA spontaneous ICH management guideline (PubMed)
- Becker KJ et al. 2001, Neurology — withdrawal of support and self-fulfilling prophecies (PubMed)
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.