20.1 Intracerebral Hematoma (03.D.1)

Spontaneous intracerebral hemorrhage (ICH)—bleeding into brain parenchyma—causes about 10–15% of strokes and a much larger share of stroke deaths. Thirty-day case fatality is often near 40%, and only about one in five survivors is independent at six months. Blueprint item 03.D.1 asks you to localize the clot, estimate volume, apply the ICH score conceptually, stop hematoma expansion, reverse coagulopathy, and know when hemispheric surgery helps—and when trial evidence says it does not.

Deep hypertensive versus lobar disease

Classic hypertensive ICH ruptures small penetrating arteries damaged by lipohyalinosis in deep gray matter: putamen, caudate, thalamus, pons, and the cerebellar dentate nucleus. Cerebral amyloid angiopathy (CAA) produces lobar, often cortical–subcortical, hematomas in older adults and may be accompanied by convexity subarachnoid blood or cortical superficial siderosis. Younger patients, isolated lobar bleeds without chronic hypertension, or atypical locations need a vascular search (CTA or MRA, then catheter angiography when noninvasive studies are inconclusive) because arteriovenous malformation, aneurysm, dural fistula, or cerebral venous thrombosis changes both the operation and the reversal plan. Anticoagulant-associated ICH can occupy any territory and expands more often than hypertensive deep bleeds of the same size.

Noncontrast head CT is the first test. Volume is estimated with the ABC/2 ellipsoid method: A is the largest axial diameter, B the perpendicular diameter on the same slice, and C the number of slices that contain hematoma multiplied by slice thickness in centimeters. ABC/2 is accurate enough for triage and ICH scoring; it can overestimate irregular or anticoagulant clots. Most hematoma expansion occurs in the first 3–6 hours and becomes uncommon after 24 hours, which is why the first hours of blood-pressure control and reversal actually change anatomy.

ICH score, used conceptually

The original ICH score (Hemphill and colleagues, Stroke 2001) is a 0–6 integer built from five admission findings. It estimates 30-day mortality in the derivation cohort. It is not a mandate to withdraw care. Early do-not-resuscitate orders create a self-fulfilling prophecy that inflates predicted death, especially at intermediate scores.

In the original cohort, 30-day mortality rose from 0% at score 0 to 13%, 26%, 72%, and 97% at scores 1–4, with scores 5–6 near 100%. Later series often show lower mortality at scores 3–4 because intensive care improved. Use the integer to communicate severity and to risk-stratify, not to skip a reversible coagulopathy or a cerebellar decompression.

Worked example: an 82-year-old with GCS 10, a 40 mL putaminal hematoma, and IVH scores 1 (age) + 1 (GCS 5–12) + 1 (volume) + 1 (IVH) + 0 (supratentorial) = 4. That is a high predicted mortality—and still not a reason to withhold nicardipine or 4-factor PCC.

CTA spot sign

The spot sign is one or more punctate foci of contrast within the hematoma on CTA, representing active extravasation. Pooled series show moderate sensitivity (about 62%) and high specificity (about 88%) for later growth, with expansion in more than half of spot-positive scans versus about 10% of spot-negative scans. A positive spot sign should speed reversal, tighten blood-pressure control, and prompt an early stability CT. It is not, by itself, an indication for craniotomy.

Blood pressure: INTERACT2, ATACH-2, and the 2022 AHA/ASA ICH guideline

Higher presenting systolic pressure tracks with expansion. INTERACT2 (Anderson et al., NEJM 2013; 2,839 patients within 6 hours) compared intensive lowering to systolic <140 mm Hg within 1 hour versus a then-usual target <180 mm Hg. Death or major disability (modified Rankin Scale 3–6) at 90 days was 52.0% versus 55.6% (odds ratio 0.87, 95% CI 0.75–1.01, p = 0.06)—not significant on the primary endpoint. An ordinal analysis of the full Rankin scale suggested a modest shift toward better function (p = 0.04). Achieved minima in the intensive arm clustered near 150 mm Hg, not 120.

ATACH-2 (Qureshi et al., NEJM 2016; 1,000 patients) pushed further with intravenous nicardipine to 110–139 mm Hg versus 140–179 mm Hg. Death or disability (mRS 4–6) at 3 months was 38.7% versus 37.7% (no benefit). Renal adverse events within 7 days were 9.0% versus 4.0%. Mean minimum systolic pressure in the intensive arm was about 129 mm Hg. The practical harm is overshoot: driving systolic pressure into the 110–120 mm Hg band (ATACH-2’s intensive floor) trades expansion risk for hypoperfusion and kidney injury without improving disability.

The 2022 AHA/ASA spontaneous ICH guideline (Greenberg et al., Stroke 2022) synthesizes those trials for typical mild-to-moderate ICH with presenting systolic pressure 150–220 mm Hg: start treatment promptly (ideally within 2 hours), lower toward 140 mm Hg, keep systolic pressure about 130–140 mm Hg with smooth, low-variability control, and treat lowering <130 mm Hg as potentially harmful. For presenting systolic pressure >220 mm Hg, more gradual reduction is reasonable; those patients were under-represented in INTERACT2. Use a titratable intravenous agent (nicardipine or clevidipine are common) rather than intermittent boluses that sawtooth the tracing.

INTERACT3 care bundle

INTERACT3 (Ma et al., Lancet 2023; 7,036 patients, stepped-wedge cluster design) tested a 7-day care bundle, not blood pressure in isolation:

1. Systolic <140 mm Hg within 1 hour of starting treatment, with 130 mm Hg as the stop threshold.
2. Glucose 6.1–7.8 mmol/L (110–140 mg/dL) without diabetes and 7.8–10.0 mmol/L (140–180 mg/dL) with diabetes.
3. Temperature <37.5 °C within 1 hour.
4. Warfarin reversal to INR <1.5 within 1 hour using prothrombin complex concentrate or plasma.

The bundle produced a favorable ordinal mRS shift (odds ratio 0.86, p = 0.015) versus usual care. Teach it as coordinated ICU hygiene plus blood pressure—not as a license to overshoot systolic pressure below 130 mm Hg.

Reverse anticoagulants at the ICH bedside

Reversal pharmacology is covered in depth in Chapter 4 (Reversal of Anticoagulation After Acute Hemorrhage). The hematoma exam still requires you to pick the correct agent immediately, without waiting for every laboratory result when the exposure is known.

PATCH showed that platelet transfusion worsens outcome in antiplatelet-associated ICH outside emergency surgery or severe thrombocytopenia—do not “cover” aspirin with a unit of platelets. Tranexamic acid (TICH-2) reduced expansion slightly without a functional-outcome win and is not standard ICH rescue.

Surgery for hemispheric ICH: STICH, STICH II, and ENRICH

STICH (Mendelow et al., Lancet 2005; 1,033 patients) compared early surgery with initial conservative care for spontaneous supratentorial ICH when the surgeon was uncertain. Favorable 6-month outcome was 26% versus 24% (odds ratio 0.89, p = 0.414). STICH II (Lancet 2013; 601 patients) restricted enrollment to lobar clots 10–100 mL within 1 cm of the cortex and without IVH. Early evacuation within 12 hours of randomization produced a favorable outcome in 41% versus 38% (odds ratio 0.86, p = 0.367)—still not a significant primary-endpoint win. Mortality trended lower (18% versus 24%). Exam translation: routine early open evacuation of lobar ICH is not generally beneficial.

MISTIE III used stereotactic catheter drainage plus alteplase for large intraparenchymal clots; the primary functional endpoint was not met, although residual volume of 15 mL or less associated with better outcomes in secondary analyses. That catheter-plus-thrombolytic approach is not the same operation as ENRICH.

ENRICH (Pradilla et al., NEJM 2024) tested early minimally invasive parafascicular surgery (imaging-guided trans-sulcal corridor with a tubular retractor and automated resection device) within 24 hours plus medical care versus medical care alone in 300 patients with 30–80 mL spontaneous ICH. Mean utility-weighted mRS at 180 days was 0.458 versus 0.374 (posterior probability of superiority 0.981). The benefit was in lobar hemorrhage (between-group difference 0.127). Enrollment of anterior basal ganglia hemorrhages was stopped for futility (difference −0.013). Thirty-day death was 9.3% versus 18.0%. Mention ENRICH for selected lobar ICH in that volume window at centers that perform the procedure. Do not generalize it to routine craniotomy for deep ICH, and do not treat it as a reversal of STICH/STICH II for open surgery.

Glucose, VTE, and seizures—what is not routine

Treat hyperglycemia and avoid insulin-induced hypoglycemia; INTERACT3’s stratified glucose targets are a practical bundle rather than a license for intensive insulin. For venous thromboembolism, start intermittent pneumatic compression on admission in non-ambulatory patients (CLOTS 3 logic; 2022 AHA/ASA Class 1). Pharmacologic prophylaxis with low-molecular-weight or unfractionated heparin is reasonable 24–48 hours after onset once a stability scan shows no important expansion. Graduated compression stockings alone do not prevent DVT. Seizure prophylaxis is not routine, including for lobar ICH without a clinical or electrographic seizure. Treat witnessed seizures. Continuous EEG is reasonable when consciousness is unexplained or events might be nonconvulsive. Prophylactic corticosteroids and standing hyperosmolar therapy do not improve ICH outcome.

Exam traps: mixing the ischemic-stroke lytic ceiling (185/110 mm Hg) with ICH targets; treating ATACH-2’s 110–139 mm Hg arm as “more is better”; quoting STICH II as a positive surgical trial; transfusing platelets for aspirin-associated ICH; starting levetiracetam on every lobar bleed “because CAA seizes.”