One-to-One-to-One: A Massive Transfusion Ratio Against the Clock
A motorcycle collision brings a patient into the trauma bay in real hemorrhagic shock before any laboratory value can return — and the massive transfusion protocol's own internal ratio becomes the actual, unresolved question.
Desmond K., a 44-year-old man, moved into his own apartment for the first time in three years two weeks ago, finally out of his brother's spare room after a stretch of steady work paid off — a detail his brother gave paramedics at the scene, since Desmond was unconscious by the time they arrived. He was riding his motorcycle home from a late shift when a car turned into his lane; EMS found him hypotensive with an obvious pelvic deformity and brought him in on a backboard with two large-bore IVs already running. In the trauma bay his systolic pressure sat at 78 despite two liters of crystalloid, his heart rate at 132, and a bedside ultrasound showed free fluid in the abdomen alongside portable imaging confirming an unstable pelvic fracture — a lactate of 6.2 and a base deficit of 9 confirming what the vitals already suggested: real, ongoing hemorrhagic shock, not a transient response to pain or fear.
The massive transfusion protocol activates on vitals like these before any laboratory value returns, which is exactly the problem the debate over its own internal ratio is trying to solve. PROPPR, the largest randomized comparison of transfusion ratios in trauma, found no significant difference in 24-hour or 30-day mortality between a 1:1:1 ratio of plasma, platelets, and red cells and a plasma-sparing 1:1:2 ratio — but its own secondary endpoints favored 1:1:1 for achieving hemostasis and for fewer deaths specifically from exsanguination in the first 24 hours, the outcome that matters most in the exact window Desmond is now in. The more recent iTACTIC trial tested a different axis entirely — viscoelastic-testing-guided component ratios against conventional coagulation-panel-guided ratios — and was flatly neutral: 67% of the viscoelastic arm were alive and free of massive transfusion at 24 hours against 64% of the conventional arm, with no mortality difference and, by the authors' own statement, no difference in any pre-specified subgroup on the primary outcome. Its one suggestive signal sat in 74 patients with severe traumatic brain injury, and even there the odds ratio of 2.12 carried a confidence interval from 0.84 to 5.34 — hypothesis-generating, not a demonstrated benefit, and in any case a population Desmond's normal neurologic exam and head CT put him outside of. What iTACTIC did show, and what cuts against the intuition that testing spares product, is that the viscoelastic arm received roughly twice as many hemostatic interventions as the conventional arm. Neither trial answers the actual question in front of the team: not which ratio is correct in principle, but how long to run an empiric ratio before real-time viscoelastic data can meaningfully change it.
In the trauma bay, before the OR
Start 1:1:1 right now, in the trauma bay, before he leaves for interventional radiology. PROPPR is the largest trial we have on this exact question, and while its primary mortality endpoint was neutral, its own secondary results favored 1:1:1 for achieving hemostasis and for fewer deaths specifically from exsanguination in the first twenty-four hours — that's the outcome that matters for a patient with his numbers right now, not a null result on a broader endpoint measured weeks out.
I'm not disputing PROPPR's secondary result, but a fixed ratio applied uniformly means some patients get more plasma and platelets than their actual clotting picture calls for, and that's not a free choice — dilutional coagulopathy, TRALI, and TACO are real costs of over-transfusion, not theoretical ones.
And I'll concede the trial I'd most like to lean on doesn't give me what I want. iTACTIC randomized viscoelastic-guided against conventional-test-guided protocols and came back neutral — and its viscoelastic arm ended up receiving about twice the hemostatic interventions, not fewer. So I can't tell you assay-guided titration spares product; the one trial that tested it found the opposite. What I can tell you is that it titrates to his clotting rather than to an average, and that is still the argument. Get a viscoelastic assay running the moment he hits the bay, rather than committing to a fixed ratio for the whole resuscitation before we have any real data on his own clotting.
You're right that a fixed ratio can over-transfuse some patients relative to their real clotting picture — but Desmond doesn't have the several minutes even the fastest viscoelastic assay needs to run and result, not at a lactate of 6.2 and a pressure that's still 78 on two liters.
I'd start 1:1:1 immediately, exactly as proposed, and commit right now, out loud, to pulling a viscoelastic sample the moment he's stable enough for a dedicated line, with an explicit plan to retitrate off it as soon as it results — the empiric ratio is the plan for the first several minutes, not the whole resuscitation.
Agreed within minutes, before he left the trauma bay: start the massive transfusion protocol at a 1:1:1 ratio immediately, draw a viscoelastic sample as soon as a line could be dedicated to it, and retitrate component ratios off that result the moment it returned rather than running the empiric ratio for the length of the whole resuscitation. He went to interventional radiology for pelvic angioembolization, then to the OR for splenic repair; six units of red cells, six of plasma, and one apheresis platelet dose later, his lactate had cleared to 2.1 and his pressure held without further escalation.
Not agreed: how many units of the empiric ratio to give before switching over if the viscoelastic result were delayed by lab logistics rather than by his own instability. The anesthesiologist's own comfort ceiling was roughly four units of each component before insisting on a result regardless of how busy the lab was; the hematologist would have pushed harder for the sample the moment vascular access allowed, arguing that a five- or six-unit empiric run is exactly the scenario where over-transfusion risk stops being theoretical. Desmond's own resuscitation didn't test that disagreement — the sample resulted before either ceiling was reached — so it stayed exactly where it started, unresolved.