Tranexamic Acid: A Clock That Started Before He Reached the Hospital
The trauma team wants TXA reflexively, the way it's given in every major hemorrhage protocol. The actual question isn't whether to give it — it's whether the three-hour clock the evidence is built on already ran out somewhere on the highway.
A.N., a 41-year-old man, is brought in after a high-speed rollover with an open femur fracture and a grade IV splenic laceration, bleeding into both his thigh and his abdomen. The trauma team's reflexive instinct is to add tranexamic acid to the massive hemorrhage protocol the moment blood products are ordered — but the actual time of injury turns out to be genuinely uncertain: he was found by a passing driver an unknown interval after the crash, EMS response time was itself delayed by the rural location, and the best reconstruction anyone in the room can put together places injury-to-hospital-arrival somewhere between two and three and a half hours, straddling exactly the threshold that matters.
CRASH-2, the trial that established TXA's mortality benefit in traumatic hemorrhage, randomized over 20,000 patients and found a real reduction in all-cause and bleeding-related death — but its own prespecified subgroup analysis by time-to-treatment found that benefit concentrated almost entirely in patients treated within three hours of injury, with a signal toward possible harm, not neutral effect, in those treated later. That single finding is the reason this isn't a reflexive add-on the way it can feel in a busy massive-transfusion activation: giving TXA to a patient genuinely outside the window doesn't just fail to help, it plausibly costs him something, which makes reconstructing his actual timeline — imprecise as it inevitably is — the load-bearing step before the drug goes in, not an afterthought to it.
The passing driver who found him remembers checking her phone before calling 911 and estimates roughly ten minutes had passed since she first noticed the overturned pickup from the road, though she also admits she isn't certain how long it had been there before she saw it. EMS's own dispatch log gives a firmer anchor — the 911 call itself is timestamped — but the true injury time, the actual moment of the crash, still has to be inferred backward from a witness's uncertain memory rather than read directly off any record, which is exactly the kind of imprecise reconstruction the team has to work with honestly rather than pretend away.
In the trauma bay, reconstructing the timeline
Give the TXA now — it's already part of our massive hemorrhage protocol, the drug's own safety profile at trauma dosing is favorable, and we're never going to be fully certain of his exact injury time anyway. Waiting for timeline certainty that isn't coming just delays a low-risk intervention.
I'd slow down on "low-risk" specifically for the timing question. CRASH-2's own subgroup analysis found the mortality benefit was concentrated within three hours of injury and reversed toward a harm signal after that — this isn't a theoretical caveat, it's the actual randomized finding the trauma indication rests on. Our best reconstruction puts him possibly past that window, not comfortably inside it.
"We're never fully certain of injury time" is true in general, but here we have a specific estimate that straddles the exact threshold that matters — that's different from genuine unknowability.
Worth being explicit that this is a different question from the elective cardiac-surgery TXA pivot, where the tradeoff is dose-dependent seizure risk in a controlled setting — that safety concern doesn't transfer here. And I won't manufacture a number we don't have. The 911 call is timestamped; the crash is not, and nothing in this chart converts one into the other — the dispatch log anchors when a stranger noticed him, not when he was hurt. What we actually hold is a range, two to three and a half hours, straddling the threshold. CRASH-2's late stratum isn't neutral ground to default onto: it is a relative risk of 1.44 for bleeding death. An interval we cannot place on one side of a line where the two sides point in opposite directions is not an argument for giving the drug anyway. Give it only if someone can put the injury inside three hours on something firmer than a witness's recollection — and if that can't be done, this is one of the uncommon times the honest answer is to withhold it and write down why.
Agreed: balanced-component massive transfusion and operative source control of both the splenic laceration and femur fracture. He stabilized intraoperatively and was extubated on post-operative day one.
NOT agreed, and deliberately left unresolved in the record: whether the TXA should have been given at all. The trauma attending gave it, reasoning the 911 timestamp made an interval under three hours more likely than not. The anesthesiologist and the clinical pharmacologist both documented that the timestamp dates his discovery and not his crash, and that CRASH-2's late stratum carries a relative risk of 1.44 rather than a neutral result. What went into the chart was the disagreement itself, on the shared view that a reconstructed interval written down as a finding would have been the more dangerous entry.