Cerebral Haemorrhage Diagnosis
Cerebral Haemorrhage Diagnosis
The first CT in a suspected stroke does one job, and it does it well: it establishes whether the artery has blocked or burst. Everything after that is a second job, and it is the one that decides the patient's treatment. A scan that reports only a haemorrhage has answered the easy question. The report that changes management describes how much blood there is, where it sits, whether it has reached the ventricles, and whether the bleeding appears to be continuing — and the history adds the sixth item, which is what the patient takes.
Why the CT comes first, and what it settles
Non-contrast CT is the first investigation in every suspected stroke, and it is not a preliminary to something better — it is the decision point. Fresh blood within the brain appears as a bright, well-defined area, whereas an ischaemic stroke in the first hours usually looks normal or only subtly darker. That single difference determines whether the patient is a candidate for reopening an artery or for reversing a clotting problem, and the two paths do not overlap.
The scan is also fast and universally available, which matters more than elegance in an emergency. It is read immediately for five findings: whether there is blood at all, how much there is, where it lies, whether it has entered the ventricles, and whether the ventricles are already enlarged by obstructed cerebrospinal fluid. In a patient who is deteriorating, the scan is repeated, because the question that matters most in the first hours — is this still happening — cannot be answered from a single image.
Measuring the volume, and reading the location
Volume is estimated from the CT by a simple and deliberately approximate method — multiplying the three diameters and dividing by two — and supratentorial bleeds are then grouped as small (under about 10 mL), moderate (10–30 mL) or large (over 30 mL). The grouping exists because published series report steadily worse outcomes across it, and because it is one of the inputs to the prognostic scores used in the first hours.
Volume alone is a poor guide to what should be done, and the reason is that the same number means different things in different places. The space behind the brainstem is tight, so a cerebellar bleed of 3 cm or more that compresses the brainstem or obstructs the flow of cerebrospinal fluid is a surgical emergency, while a considerably larger bleed in a silent part of a frontal lobe may be managed without an operation. Bleeding in the pons is devastating at volumes that would be unremarkable elsewhere.
The location also points at the mechanism, which is a different kind of information. Bleeding that starts deep — in the putamen, thalamus, pons or cerebellum — is the classic pattern of long-standing high blood pressure. Bleeding that starts at the outer rim of a hemisphere, in an older patient, points instead towards cerebral amyloid angiopathy. That distinction is not settled by CT, but the CT is what raises it, and it is the reason the next investigation is chosen rather than ordered routinely.
Looking for a cause, and for continuing bleeding
CT angiography adds two things that the plain scan cannot provide.
The first is an underlying structural cause: an arteriovenous malformation, an aneurysm, a venous sinus thrombosis, or a narrowing that explains the event. This matters most in a patient who is younger, who has no history of high blood pressure, or whose bleed is in an unusual place — the situations in which a treatable malformation is most likely to be found.
The second is a sign that the bleeding may still be active: a small spot of contrast visible within the haematoma, described as the spot sign. It is an imperfect predictor — it is present in a minority of patients who go on to expand and absent in some who do — but it is the closest thing to a live answer that imaging offers, and where it is seen it reinforces the urgency of blood-pressure control and reversal of anticoagulation. Catheter angiography is reserved for when a malformation is suspected and its treatment is being planned, because it delivers the detail required to decide whether something can be embolised, removed or treated with radiation.
Separating blood pressure from amyloid angiopathy
The question that MRI answers is one CT cannot: what made this vessel fail. It is not an academic question, because the two common answers carry different risks of bleeding again and therefore different answers to the question of restarting an anticoagulant.
MRI with susceptibility-weighted or gradient-echo sequences shows small bleeds that have happened previously — cerebral microbleeds — and the haemosiderin staining left along the cortical surface. A pattern of these confined to the lobar regions, in an older patient, is what points to cerebral amyloid angiopathy: a protein fragment deposited in the walls of the small vessels on the brain's surface, making them brittle with age. Where those findings are absent and the bleed is deep, the likelier explanation is the small-vessel damage of long-standing high blood pressure — a cause that can be treated, and one whose future risk is substantially reduced by doing so.
Clotting studies: what must be reversed, and how fast
This is the part of the assessment that is most often under-weighted, and it is the one with the clearest deadline. The laboratory work is straightforward — clotting screen, platelet count, renal function, glucose — but its purpose is specific: to identify the drug or the disorder that has to be undone in the first hour.
The reversals are not interchangeable. Warfarin is reversed with vitamin K and prothrombin complex concentrate. Dabigatran has a specific antidote. The factor Xa inhibitors are reversed with a specific agent where it is available, or with concentrate. Heparin is reversed with protamine. A low platelet count or a bleeding disorder requires its own correction, and advanced liver disease or kidney failure changes both the risk of bleeding and how the drugs behave. Giving the wrong reversal, or giving none because the history was incomplete, wastes the window in which reversal still helps.
Two further points follow. A normal clotting screen does not exclude the presence of a direct oral anticoagulant, which the standard screen does not detect — the specific assay or a clear medication history is needed. And in a patient taking an antiplatelet drug who bleeds spontaneously, platelet transfusion has been studied and found not to help, so it is not given routinely. Knowing which drug is involved is therefore not bookkeeping; it is the difference between a treatment that works and one that does not.
What a complete assessment must answer
The value of the assessment is not in how many investigations were performed but in whether it answers six questions. Any one left blank makes the plan incomplete, however many images were taken.
Is it a bleed? — established by the first CT, before any treatment. How much, and where? — volume and location, reported as numbers and named structures rather than "a haematoma". Has it reached the ventricles, and is the fluid pathway blocked? — the finding that decides whether a drain is needed. Is it still growing? — answered by the direction of the examination and the repeat scan, and by the spot sign where it is present. What is the patient taking? — the anticoagulant and its timing, and the antiplatelet drugs, since these decide what is reversed and how urgently. And what caused it? — blood pressure or amyloid angiopathy in most patients, a structural malformation in a minority, and the answer that governs both the treatment now and the prevention afterwards.