Eligibility simulator for IV alteplase (tPA) thrombolysis in acute ischemic stroke — "time is brain"
Every thrombolysis decision in acute ischemic stroke is built on a single, deceptively simple data point: the Last-Known-Well (LKW) time — the last moment the patient was witnessed to be at their neurological baseline. Unlike a heart attack, where pain onset is usually unambiguous, many strokes are discovered on waking or found by a bystander, making LKW an estimate rather than a certainty. Getting this timestamp right — or wrong — determines whether a patient is eligible for a therapy proven to reduce disability.
Acute ischemic stroke care is organized around a brutal arithmetic: the longer a brain artery stays occluded, the more irreversibly infarcted tissue accumulates, and the smaller the salvageable "penumbra" becomes. Alteplase (recombinant tissue plasminogen activator, rt-PA) was approved by the FDA in 1996 based on the NINDS trial, which showed a clear time-dependent benefit — patients treated within 3 hours had significantly better 90-day outcomes than placebo, and later the ECASS III trial (2008) extended the proven benefit window to 4.5 hours for a more selected population.
Because benefit shrinks and hemorrhage risk climbs the longer treatment is delayed, the entire prehospital and emergency-department pathway (EMS pre-notification, stroke code activation, parallel CT and lab draws) exists purely to compress the interval between LKW and needle time. This is why stroke systems obsess over minutes: NNT (number needed to treat for one additional patient with improved outcome) is roughly 3 at 90 minutes, 4-5 at 3 hours, and 14 near the 4.5-hour boundary.
The NINDS rt-PA trial found that patients treated within 90 minutes of onset were nearly three times more likely to have minimal or no disability at 3 months compared with placebo — a benefit that erodes steadily with every additional 15 minutes of delay.
For roughly a quarter of ischemic strokes, the patient wakes up with a deficit already present ("wake-up stroke") or is found down with no witness. In these cases LKW defaults to the last time the patient was confirmed normal — often bedtime the night before — which may place the case far outside the standard window on paper.
Modern practice uses advanced imaging to recover eligibility for these patients: MRI diffusion-weighted imaging (DWI) versus FLAIR mismatch identifies infarcts still imaging-young (DWI positive, FLAIR negative typically means onset under ~4.5h), and the WAKE-UP trial (2018) showed this MRI-based selection safely extends thrombolysis to appropriately screened wake-up strokes. Similarly, CT perfusion mismatch (small ischemic core, large at-risk penumbra) underpins extended-window protocols (EXTEND, up to 9h from onset or midpoint of sleep) for carefully selected patients — shifting stroke triage from a pure clock-based rule to a tissue-based one.
Once a stroke is suspected, two parallel processes compress into minutes: the NIH Stroke Scale (NIHSS) quantifies deficit severity across 15 items, and an emergent non-contrast head CT scans for the single finding that reverses the entire treatment plan — intracranial hemorrhage. Alteplase dissolves clots indiscriminately; giving it to a patient already bleeding into the brain converts a bad stroke into a potentially fatal hemorrhage.
The NIHSS assigns points across level of consciousness, gaze, visual fields, facial palsy, motor arm/leg, limb ataxia, sensory loss, language, dysarthria, and extinction/inattention — producing a score from 0 (no deficit) to 42 (maximal deficit). It takes an experienced examiner roughly 5-8 minutes and is the common language used to communicate severity between EMS, the emergency department, neurology, and stroke registries.
NIHSS also informs eligibility nuance: very low scores (under ~4, minor/non-disabling deficits) historically raised debate about whether the bleeding risk of tPA outweighs a small anticipated deficit, though PRISMS (2018) supports treating even minor, disabling strokes. Very high scores (over ~25) correlate with large territory infarcts and, in some protocols, prompt extra caution or additional imaging because of a higher hemorrhagic transformation risk — though high NIHSS alone is not an absolute exclusion when other criteria are satisfied.
Non-contrast head CT remains the frontline imaging study because it is fast, nearly universally available, and highly sensitive for acute hemorrhage, which appears as hyperdense (bright white) blood within minutes of onset. Ischemic changes, by contrast, may be subtle or entirely absent on CT in the first few hours — early signs include loss of grey-white differentiation, sulcal effacement, and the hyperdense MCA sign (a bright thrombus visible directly inside the vessel).
The purpose of this CT is not to confirm the stroke diagnosis (which remains clinical) but to exclude the one finding — hemorrhage — that makes thrombolysis dangerous. Many centers now pair the non-contrast CT with CT angiography (to identify a large-vessel occlusion amenable to mechanical thrombectomy) and CT perfusion (to estimate core infarct versus salvageable penumbra), all acquired in a single trip to the scanner to avoid losing minutes.
A hyperdense clot visible on plain CT, hemorrhage of any volume, or a large established infarct already occupying more than one-third of the middle cerebral artery territory are each independently sufficient to halt the thrombolysis pathway — imaging can override an otherwise "eligible" clinical picture.
Alteplase eligibility is decided by a structured checklist, not a single number. Inclusion requires a clinical diagnosis of ischemic stroke with a measurable deficit within the approved time window; exclusion criteria layer on top, screening for bleeding risk, recent trauma or surgery, uncontrolled hypertension, and interacting anticoagulation. Any single hard exclusion closes the gate regardless of how favorable the other criteria look.
• Clinical diagnosis of ischemic stroke causing a measurable neurological deficit • Time from last-known-well to expected treatment start within 4.5 hours (or an approved extended-window protocol using perfusion/DWI-FLAIR mismatch) • Age generally 18 years or older (pediatric thrombolysis is handled on a case-by-case protocol basis) • Symptoms present for a meaningful, non-resolving duration — a rapidly improving deficit does not automatically exclude treatment if a disabling deficit remains
Hemorrhagic risk: • Any evidence of intracranial hemorrhage on CT • History of prior intracranial hemorrhage • Known structural CNS lesion with bleeding tendency (aneurysm, AVM, neoplasm) • Suspected or confirmed subarachnoid hemorrhage presentation
Blood pressure: • Systolic ≥185 mmHg or diastolic ≥110 mmHg that cannot be safely lowered before infusion
Coagulation and recent anticoagulant use: • Current use of a direct oral anticoagulant within the last 48 hours (unless coagulation studies are normal) • INR >1.7 on warfarin, or elevated PTT on heparin • Platelet count <100,000/mm³
Recent trauma or surgery: • Major surgery or serious trauma within the prior 14 days • Head trauma or prior stroke within the previous 3 months • Gastrointestinal or urinary tract hemorrhage within 21 days
Other: • Blood glucose <50 mg/dL (hypoglycemia can mimic stroke) • Infective endocarditis or suspected aortic dissection • CT evidence of a large established infarct (>1/3 MCA territory)
The gate is intentionally asymmetric: dozens of favorable findings cannot overcome a single hard exclusion such as active hemorrhage. This conservative design reflects that symptomatic intracranial hemorrhage after tPA — while rare (~6%) — carries roughly 50% mortality.
Once the eligibility gate is passed, alteplase is dosed strictly by weight: 0.9 mg/kg to a hard maximum of 90 mg. Ten percent of the total dose is pushed as an immediate IV bolus over one minute, and the remaining 90% is infused continuously over the following 60 minutes. This split delivers an early pharmacologic strike against the clot while sustaining plasminogen activation throughout the critical first hour.
Alteplase is a recombinant form of tissue plasminogen activator (rt-PA), a serine protease that binds fibrin within a thrombus and catalyzes the conversion of trapped plasminogen into plasmin. Plasmin then degrades the fibrin mesh holding the clot together, restoring blood flow through the occluded artery. Because alteplase preferentially activates plasminogen that is already fibrin-bound, its action is meant to concentrate at the clot surface rather than causing widespread systemic fibrinolysis — though some systemic fibrinogen depletion still occurs, which underlies the bleeding risk.
Recanalization is not guaranteed: alteplase alone achieves early recanalization in roughly a third to half of proximal large-vessel occlusions, which is why patients with a confirmed large-vessel occlusion are simultaneously routed toward mechanical thrombectomy — a physical clot retrieval procedure that can be combined with, not substituted for, IV thrombolysis (the "bridging therapy" model).
For a 70 kg patient, the total dose is 63 mg (0.9 × 70): a 6.3 mg bolus followed by 56.7 mg infused over 60 minutes. Because the absolute maximum is 90 mg, any patient over 100 kg is capped at that ceiling regardless of actual weight.
During and after infusion, blood pressure is monitored aggressively (every 15 minutes for 2 hours, then every 30 minutes) to keep it under 180/105 mmHg, since hypertension sharply raises hemorrhagic transformation risk. A repeat neurological exam and often a follow-up CT at 24 hours screen for hemorrhagic conversion before starting antiplatelet or anticoagulant therapy. No other antithrombotic agents are given for 24 hours after alteplase to avoid compounding bleeding risk.
The 10%/90% bolus-then-infusion design mirrors the original NINDS protocol: the bolus achieves a fast rise in active drug concentration at the clot surface, while the hour-long infusion sustains plasmin generation long enough to meaningfully degrade a fibrin-rich thrombus before the drug is cleared (plasma half-life ≈5 minutes).
The clinical mantra "time is brain" is not a slogan but an approximation grounded in stroke pathophysiology: in a typical large-vessel occlusion, each untreated minute is estimated to cost about 1.9 million neurons, 14 billion synapses, and 12 kilometers of myelinated fibers. Earlier treatment does not merely improve statistics on a curve — it directly preserves the ischemic penumbra, the band of at-risk-but-still-viable tissue surrounding an irreversibly infarcted core.
When a cerebral artery occludes, blood flow does not drop uniformly across the downstream territory. A central "core" region, typically fed by no collateral circulation, infarcts within minutes and is not salvageable by any current therapy. Surrounding it, the "penumbra" receives just enough collateral flow to remain metabolically viable but electrically silent — tissue that is functionally impaired but structurally alive, and the entire target of thrombolysis and thrombectomy.
The penumbra shrinks over time as collateral flow gradually fails and core expands to consume it — the rate of this conversion varies substantially between patients depending on collateral robustness, blood pressure, and glucose control, which is why perfusion imaging (rather than the clock alone) increasingly guides treatment decisions in the extended window. But on average, restoring flow earlier converts more penumbra back to functioning brain rather than letting it become permanent infarct.
The clinical payoff of earlier treatment is measured with the modified Rankin Scale (mRS) at 90 days, where mRS 0-1 represents essentially full functional independence. Pooled trial data show a clear, continuous relationship: patients treated within 90 minutes of onset have roughly triple the odds of an excellent outcome compared with placebo; that odds ratio steadily declines through the 3-hour mark and further through 4.5 hours, where benefit is smaller but still statistically significant.
This is why stroke systems of care are built around minimizing every interval — prehospital recognition and EMS pre-notification, "code stroke" activation before hospital arrival, parallel (not sequential) lab draws and CT imaging, and point-of-care mixing of alteplase — collectively aiming to compress door-to-needle time to under 60 minutes, with top-performing centers now achieving under 30.
A useful mental model: each 15-minute delay in treatment is associated with measurably worse average 90-day disability outcomes across large pooled analyses — the eligibility window is a hard legal and safety boundary, but within it, earlier is unambiguously better.