Transfusion

Transfusion in Critical Care: 2026 Evidence & Management

Current standard of care for blood transfusion in the critically ill — red-cell thresholds, massive hemorrhage, platelets and plasma, and reaction management — organized by decision, with the trials behind each recommendation.

Reviewed by the DosePilot Medical Team · Last reviewed: Jun 2026 · Independent — no industry funding
How to read this page. Each decision opens with the current approach, then the trials behind it (design, key results, limitations), and a short practical note (In practice). All thresholds, doses, and figures are taken from the cited primary sources (DOI + PMID). The default position across modern critical care is restrictive — give the fewest units that keep the patient safe — with defined exceptions (active major hemorrhage, acute coronary syndrome). The AABB 2023 red-cell guideline and current society guidance on massive hemorrhage and platelets anchor the recommendations. In practice notes are general guidance, not a substitute for clinical judgment.
How we verify evidence & editorial policy
Written and maintained by the DosePilot Medical Team. Recommendations are based on randomized controlled trials and current society guidelines (GRADE). Every citation is verified directly against PubMed (PMID) and the publisher DOI — no memory-based or AI-generated references. Each topic is peer-reviewed by the clinical team before publication and re-reviewed at least every 12 months or when major new evidence appears.

Funding & conflicts of interest. Produced independently; no industry funding was received for this content and contributors report no relevant financial conflicts of interest.
On this page
  1. Evidence timeline
  2. When to transfuse — physiology & thresholds
  3. Red-cell threshold — the restrictive standard
  4. Red-cell threshold by population
  5. Massive hemorrhage & transfusion protocols
  6. Tranexamic acid & antifibrinolytics
  7. Platelet transfusion
  8. Plasma, fibrinogen & reversal
  9. Transfusion reactions & complications
  10. Patient blood management & special products
  11. Where it’s heading
  12. Bottom line
  13. FAQ
  14. Related DosePilot tools
  15. Guidelines & evidence basis
At a glance — current standard of care
SettingCurrent approachKey trial(s)
General ICU / most patientsRestrictive — transfuse at Hb <7 g/dL, target 7–9TRICC; AABB 2023
Septic shockRestrictive 7 g/dL (no benefit above)TRISS
Cardiac surgeryRestrictive <7.5 g/dL non-inferiorTRICS-III
Acute upper GI bleedingRestrictive 7 g/dL (stable) — better survivalVillanueva
Acute MI + anemiaA higher threshold (8–10) is reasonableMINT; REALITY
DecisionCurrent approachKey trial(s)
Massive hemorrhageFixed ratio ~1:1:1, MTP, replace ionized calciumPROPPR
Tranexamic acidYes in trauma & PPH <3 h; no for GI bleedingCRASH-2; WOMAN; HALT-IT
Platelets (prophylaxis)Transfuse at ≤10×10⁹/L (hypoproliferative)TOPPS; AABB
PlasmaFor active bleeding/coagulopathy — not a mild high INRAABB; guidelines
ReactionsTACO is the leading cause of transfusion deathHemovigilance

1Evidence timeline (1999–2026)

1999TRICC — a restrictive red-cell threshold (7 g/dL) is at least as safe as a liberal one (10 g/dL) in the ICU; the foundational transfusion trial
2010CRASH-2 — tranexamic acid reduces all-cause and bleeding death in trauma hemorrhage, but only when given early (<3 h)
2011FOCUS — after hip surgery in patients with cardiovascular disease or risk factors, a liberal threshold (10 g/dL) gives no benefit over 8 g/dL
2013Villanueva — in acute upper GI bleeding, a restrictive strategy (7 g/dL) improves 6-week survival and reduces rebleeding; TOPPS — prophylactic platelets reduce bleeding in hematology patients
2014TRISS — in septic shock, transfusing above 7 g/dL confers no survival benefit
2015PROPPR — 1:1:1 vs 1:1:2 plasma:platelets:red cells gives equal mortality, but 1:1:1 achieves hemostasis faster and reduces death from exsanguination
2017TRICS-III — a restrictive threshold (<7.5 g/dL) is non-inferior in cardiac surgery; WOMAN — early tranexamic acid reduces death from bleeding in postpartum hemorrhage
2019A consensus redefinition of TRALI (Type I / Type II) sharpens the diagnosis of transfusion-related lung injury and separates it from circulatory overload (TACO)
2020HALT-IT — high-dose tranexamic acid does not reduce death in GI bleeding and increases venous thromboembolism and seizures
2021REALITY — a restrictive strategy is non-inferior to liberal for 30-day cardiovascular events in acute MI with anemia (smaller trial, open-label)
2023The AABB international red-cell guideline consolidates a restrictive default (7 g/dL; 8 g/dL for orthopedic surgery and pre-existing cardiovascular disease, 7.5 g/dL for cardiac surgery); MINT — in acute MI with anemia a liberal strategy showed a borderline signal toward fewer events; CRYOSTAT-2 — empiric high-dose cryoprecipitate adds no survival benefit in trauma
2025–26Whole blood and prehospital transfusion expand, cold-stored platelets and lyophilized products advance, and patient blood management is standard — but no new evidence overturns the restrictive default outside acute coronary syndrome

2When to transfuse — physiology & thresholds

Current approach (2026)
Restrictive defaultTrigger ≠ a single number
A red-cell transfusion exists to restore oxygen delivery, not to normalize a laboratory value. Oxygen delivery (DO₂) is the product of cardiac output and arterial oxygen content, and content is dominated by hemoglobin × SaO₂; healthy tissue extracts only ~25% of delivered oxygen, so most patients tolerate moderate anemia by raising cardiac output and extraction. Modern practice therefore uses a restrictive hemoglobin trigger (transfuse when Hb falls below ~7 g/dL) combined with physiologic signs of impaired delivery — ongoing bleeding, lactate or central-venous-oxygen evidence of a supply–demand mismatch, refractory tachycardia, or active myocardial ischemia. Transfuse one unit at a time and reassess; a single unit raises hemoglobin by roughly 1 g/dL in a non-bleeding adult. The hemoglobin number guides the decision but never replaces the clinical assessment of whether oxygen delivery is actually inadequate.

Why restrictive. Across more than 45 randomized trials, giving fewer red cells has not worsened mortality, cardiac events, or recovery for most hospitalized patients, while avoiding the volume, immune, and logistical costs of unnecessary units. The burden of proof now sits with the decision to transfuse above 7 g/dL, not below it.

In practice

Anchor on Hb 7 g/dL for the stable, non-bleeding critically ill patient, then ask the physiologic question: is oxygen delivery inadequate right now? Bleeding, ischemia, or a clear supply–demand mismatch — not the number alone — justify going higher. Order single units and recheck rather than reflexively transfusing two.

Sources (2)
  • Carson JL, Stanworth SJ, Guyatt G, et al. Red blood cell transfusion: 2023 AABB international guidelines. JAMA. 2023;330(19):1892–1902. DOI 10.1001/jama.2023.12914. PMID 37824153.
  • Vincent JL, Jaschinski U, Wittebole X, et al; ICON Investigators. Worldwide audit of blood transfusion practice in critically ill patients. Crit Care. 2018;22(1):102. DOI 10.1186/s13054-018-2018-9. PMID 29673409.

3Red-cell threshold — the restrictive standard

Current approach (2026)
AABB 2023: restrictiveRCTs: no mortality difference
For the hemodynamically stable, non-bleeding critically ill adult, use a restrictive threshold of 7 g/dL (AABB 2023). This holds for the general ICU population (TRICC) and for septic shock (TRISS), where transfusing to keep hemoglobin above 7 confers no survival benefit. Target a post-transfusion hemoglobin of roughly 7–9 g/dL. The 8 g/dL threshold is reserved for specific groups in Section 4 (orthopedic surgery, pre-existing cardiovascular disease). Restrictive strategies roughly halve red-cell exposure without harming survival, length of stay, or organ recovery.
Evidence — old → new
1999
TRICC. 838 euvolemic ICU patients randomized to a restrictive (transfuse <7, maintain 7–9 g/dL) versus liberal (<10, maintain 10–12) threshold. 30-day mortality 18.7% vs 23.3% (not significant overall); in-hospital mortality was lower with restrictive (22.2% vs 28.1%), and the benefit was clearest in less-ill (APACHE II ≤20) and younger (<55 y) patients. Established that withholding red cells is safe — often safer.
2014
TRISS. 998 patients with septic shock and Hb ≤9 g/dL randomized to a 7 vs 9 g/dL threshold. 90-day mortality 43.0% vs 45.0% (relative risk 0.94; P=0.44) with no difference in ischemic events or life support — but ~50% fewer transfusions. Extended the restrictive default into the sickest, vasopressor-dependent patients.
2023
AABB international guideline. A synthesis of 45 RCTs (≈20,600 patients) recommending a restrictive threshold of 7 g/dL for most hospitalized adults (8 g/dL for orthopedic surgery or pre-existing cardiovascular disease, 7.5 g/dL for cardiac surgery), with moderate-certainty evidence that restrictive strategies do not worsen patient-important outcomes.

What changed. The field moved from a reflexive “10/30 rule” to a restrictive default grounded in randomized evidence — fewer units, equal outcomes — with the open questions now confined to acute coronary syndrome and a handful of special populations.

In practice

For the stable ICU patient — including septic shock — hold transfusion until Hb <7 g/dL, then give one unit and reassess. Document a physiologic trigger if you transfuse earlier. “Normal” hemoglobin is not the goal; adequate oxygen delivery is.

Sources (3)
  • Hébert PC, Wells G, Blajchman MA, et al; Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care (TRICC). N Engl J Med. 1999;340(6):409–417. DOI 10.1056/NEJM199902113400601. PMID 9971864.
  • Holst LB, Haase N, Wetterslev J, et al; TRISS Trial Group, Scandinavian Critical Care Trials Group. Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med. 2014;371(15):1381–1391. DOI 10.1056/NEJMoa1406617. PMID 25270275.
  • Carson JL, Stanworth SJ, Guyatt G, et al. Red blood cell transfusion: 2023 AABB international guidelines. JAMA. 2023;330(19):1892–1902. DOI 10.1001/jama.2023.12914. PMID 37824153.

4Red-cell threshold by population

Current approach (2026)
Cardiac surgery: restrictive OKACS: lean higher
Most populations follow the restrictive default; the practical exceptions are surgery with cardiovascular risk (8 g/dL) and acute coronary syndrome (lean toward a higher threshold). Acute upper GI bleeding is firmly restrictive when the patient is hemodynamically stable.
PopulationThresholdKey trial
General ICU / septic shock7 g/dLTRICC; TRISS
Cardiac surgery (on-pump)<7.5 g/dL (restrictive non-inferior)TRICS-III
Orthopedic surgery + CV risk8 g/dLFOCUS
Acute upper GI bleeding (stable)7 g/dLVillanueva
Acute MI + anemia8–10 g/dL reasonableMINT; REALITY
Evidence — by population
2011
FOCUS. 2016 patients with cardiovascular disease or risk factors undergoing hip-fracture surgery, randomized to an 8 vs 10 g/dL threshold. No difference in death or independent walking at 60 days — anchoring 8 g/dL for this surgical group.
2013
Villanueva. 921 patients with acute upper GI bleeding randomized to a 7 vs 9 g/dL threshold. The restrictive group had better 6-week survival (mortality 5% vs 9%; HR 0.55) and less rebleeding (10% vs 16%), partly by avoiding the portal-pressure rise that liberal transfusion causes in cirrhosis. Excludes exsanguinating hemorrhage, where resuscitation takes priority.
2017
TRICS-III. 5243 cardiac-surgery patients (EuroSCORE I ≥6) randomized to restrictive (<7.5 g/dL) vs liberal (<9.5 intraop/ICU, <8.5 ward). The composite of death, MI, stroke, or new dialysis was non-inferior (11.4% vs 12.5%), and held at 6 months — restrictive is safe even on bypass.
2021–23
REALITY & MINT — acute MI. REALITY (n=668) found a restrictive threshold (≤8) non-inferior to liberal (≤10) for 30-day MACE (11% vs 14%). The larger MINT (n=3504; restrictive 7–8 vs liberal ≥10) showed 30-day death-or-MI of 16.9% vs 14.5% (RR 1.15; 95% CI 0.99–1.34; P=0.07) — a non-significant trend on the primary outcome, but with a significantly higher cardiac-death rate in the restrictive arm (5.5% vs 3.2%; RR 1.74, 95% CI 1.26–2.40). On balance a higher threshold (≈10 g/dL) is reasonable in active myocardial ischemia.

What changed. The single exception to “restrictive everywhere” has crystallized around the ischemic heart: where the myocardium is actively starved, the balance of evidence now tips toward a more liberal target, while surgery and GI bleeding stay restrictive.

In practice

Default to 7; use 8 g/dL after orthopedic surgery or with stable cardiovascular disease, and a higher target (≈8–10) during active acute coronary syndrome. In upper GI bleeding, stay restrictive once the patient is resuscitated — but treat torrential, hemodynamically unstable bleeding as massive hemorrhage (Section 5).

Sources (5)
  • Carson JL, Terrin ML, Noveck H, et al; FOCUS Investigators. Liberal or restrictive transfusion in high-risk patients after hip surgery. N Engl J Med. 2011;365(26):2453–2462. DOI 10.1056/NEJMoa1012452. PMID 22168590.
  • Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368(1):11–21. DOI 10.1056/NEJMoa1211801. PMID 23281973.
  • Mazer CD, Whitlock RP, Fergusson DA, et al; TRICS Investigators. Restrictive or liberal red-cell transfusion for cardiac surgery (TRICS-III). N Engl J Med. 2017;377(22):2133–2144. DOI 10.1056/NEJMoa1711818. PMID 29130845.
  • Carson JL, Brooks MM, Hébert PC, et al; MINT Investigators. Restrictive or liberal transfusion strategy in myocardial infarction and anemia (MINT). N Engl J Med. 2023;389(26):2446–2456. DOI 10.1056/NEJMoa2307983. PMID 37952133.
  • Ducrocq G, Gonzalez-Juanatey JR, Puymirat E, et al; REALITY Investigators. Effect of a restrictive vs liberal blood transfusion strategy on major cardiovascular events among patients with acute myocardial infarction and anemia (REALITY). JAMA. 2021;325(6):552–560. DOI 10.1001/jama.2021.0135. PMID 33560322.

5Massive hemorrhage & transfusion protocols

Current approach (2026)
Balanced ratio ~1:1:1Activate an MTP
In active major hemorrhage, thresholds no longer apply — the goal is to replace whole-blood loss and stop bleeding. Activate a massive transfusion protocol (MTP) that delivers plasma, platelets, and red cells in a balanced ratio (~1:1:1), or use low-titer group O whole blood where available. Give tranexamic acid early (Section 6), control the source (surgery, angioembolization, endoscopy), and treat the “lethal triad”: keep the patient warm, correct acidosis, and replace ionized calcium — citrate in stored products chelates calcium, and hypocalcemia worsens coagulopathy and cardiac function. Use viscoelastic testing (TEG/ROTEM) to guide targeted component therapy once the initial empiric phase is underway, and de-escalate to lab-guided replacement as bleeding slows.
ElementTargetNote
Component ratio~1:1:1 plasma:platelets:RBCor low-titer O whole blood
Tranexamic acid1 g IV ≤3 h; trauma +1 g/8 htrauma & PPH
Ionized calcium>1.1 mmol/Lreplace aggressively
Fibrinogen>1.5–2 g/Lcryo / concentrate if low
Temperature / pH>36 °C / correct acidosisthe lethal triad
Evidence — old → new
2015
PROPPR. 680 severely injured patients with major bleeding randomized to 1:1:1 vs 1:1:2 plasma:platelets:red cells. No difference in mortality at 24 h (12.7% vs 17.0%) or 30 days (22.4% vs 26.1%), but the 1:1:1 group achieved hemostasis more often and had fewer deaths from exsanguination at 24 h (9.2% vs 14.6%). Cemented the balanced-ratio MTP as the empiric standard.
2023
CRYOSTAT-2. 1604 trauma patients on a major-hemorrhage protocol randomized to early empiric high-dose cryoprecipitate (3 pools ≈ 6 g fibrinogen, within 90 min of admission and 3 h of injury) vs standard care. No improvement in 28-day mortality — supporting targeted fibrinogen replacement (guided by level/viscoelastic testing) over blanket empiric cryoprecipitate.

What changed. Damage-control resuscitation reframed massive transfusion around reconstituting whole blood and limiting crystalloid; the refinements since — calcium, fibrinogen, and viscoelastic guidance — are about precision, not more product.

In practice

Activate the MTP early and don’t wait for labs to start balanced product; give TXA within 3 hours; check ionized calcium with the first cooler and replace it; keep the patient warm. Transition from fixed-ratio to viscoelastic/lab-guided replacement as the bleeding is controlled, and stop the protocol deliberately to avoid overtransfusion.

Sources (3)
  • Holcomb JB, Tilley BC, Baraniuk S, et al; PROPPR Study Group. Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma (PROPPR). JAMA. 2015;313(5):471–482. DOI 10.1001/jama.2015.12. PMID 25647203.
  • Davenport R, Curry N, Fox EE, et al; CRYOSTAT-2 Principal Investigators. Early and empirical high-dose cryoprecipitate for hemorrhage after traumatic injury (CRYOSTAT-2). JAMA. 2023;330(19):1882–1891. DOI 10.1001/jama.2023.21019. PMID 37824155.
  • Rossaint R, Afshari A, Bouillon B, et al. The European guideline on management of major bleeding and coagulopathy following trauma: sixth edition. Crit Care. 2023;27(1):80. DOI 10.1186/s13054-023-04327-7. PMID 36859355.

6Tranexamic acid & antifibrinolytics

Current approach (2026)
Trauma & PPH: benefit if earlyGI bleeding: no benefit, more VTE
Tranexamic acid (TXA) stabilizes clot by inhibiting fibrinolysis. Give it early in traumatic hemorrhage and postpartum hemorrhage — 1 g IV within 3 hours of onset, then a second dose — where it reduces death from bleeding. The timing rule is critical: benefit is greatest within the first hour and is lost or reversed after 3 hours. By contrast, TXA does not reduce death in gastrointestinal bleeding and increases venous thromboembolism and seizures, so it is not recommended there. There is no role for routine high-dose prolonged infusions outside a trial.
Evidence — old → new
2010
CRASH-2. 20,211 trauma patients with or at risk of significant bleeding randomized to TXA (1 g over 10 min, then 1 g over 8 h) vs placebo. All-cause mortality 14.5% vs 16.0% (RR 0.91; P=0.0035) and death due to bleeding 4.9% vs 5.7%. An exploratory analysis showed benefit only when given ≤3 h; later treatment increased death from bleeding.
2017
WOMAN. 20,060 women with postpartum hemorrhage randomized to TXA (1 g, repeated once if bleeding continued) vs placebo. Death due to bleeding fell (1.5% vs 1.9%; RR 0.81), with the effect concentrated in those treated within 3 hours; no increase in thromboembolic events. Established early TXA as standard in obstetric hemorrhage.
2020
HALT-IT. 12,009 patients with acute GI bleeding randomized to high-dose TXA (1 g then 3 g/24 h) vs placebo. No reduction in death from bleeding, and more venous thromboembolic events and seizures. Closed the door on TXA for GI hemorrhage.

What changed. TXA went from a presumed universal hemostatic to a context-and-time-specific drug: clearly useful early in trauma and obstetric bleeding, clearly not useful (and potentially harmful) in GI bleeding.

In practice

In trauma, give 1 g TXA as soon as possible and within 3 hours, then 1 g over 8 hours; in postpartum hemorrhage give 1 g and repeat 1 g if bleeding continues after 30 minutes or restarts within 24 hours (WOMAN/WHO) — either way, don’t start it late. Do not use TXA for GI bleeding. Weigh the small thrombotic risk in patients with a high baseline VTE risk.

Sources (3)
  • CRASH-2 trial collaborators; Shakur H, Roberts I, Bautista R, et al. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2). Lancet. 2010;376(9734):23–32. DOI 10.1016/S0140-6736(10)60835-5. PMID 20554319.
  • WOMAN Trial Collaborators. Effect of early tranexamic acid administration on mortality, hysterectomy, and other morbidities in women with post-partum haemorrhage (WOMAN). Lancet. 2017;389(10084):2105–2116. DOI 10.1016/S0140-6736(17)30638-4. PMID 28456509.
  • HALT-IT Trial Collaborators. Effects of a high-dose 24-h infusion of tranexamic acid on death and thromboembolic events in patients with acute gastrointestinal bleeding (HALT-IT). Lancet. 2020;395(10241):1927–1936. DOI 10.1016/S0140-6736(20)30848-5. PMID 32563378.

7Platelet transfusion

Current approach (2026)
Prophylaxis ≤10×10⁹/LAvoid in TTP/HIT
For hypoproliferative thrombocytopenia (chemotherapy, marrow failure), give prophylactic platelets at ≤10×10⁹/L to prevent spontaneous bleeding (AABB). Use higher thresholds for procedures: ~20×10⁹/L for central-line placement, ~50×10⁹/L for most surgery, lumbar puncture, or major bleeding, and ~100×10⁹/L for neurosurgery or CNS bleeding. One apheresis unit (or pool) typically raises the count by ~30–50×10⁹/L (less with bleeding, sepsis, DIC, or splenomegaly). Platelets are best avoided in thrombotic thrombocytopenic purpura (TTP) and heparin-induced thrombocytopenia (HIT) — they can fuel thrombosis, so reserve them for active bleeding or an urgent invasive procedure rather than prophylaxis — and they are usually unnecessary in immune thrombocytopenia (ITP) without major bleeding.
SituationPlatelet targetBasis
Prophylaxis (hypoproliferative)≤10×10⁹/L (trigger)TOPPS; AABB
Central venous catheter≥20×10⁹/LAABB
Major surgery / LP / bleeding≥50×10⁹/Lguideline consensus
Neurosurgery / CNS bleed≥100×10⁹/Lguideline consensus
TTP / HITAvoid prophylaxis; give for bleeding/procedurethrombosis risk
Evidence — old → new
2010
PLADO. Randomized low- vs medium- vs high-dose prophylactic platelets in 1272 patients. Bleeding rates were similar across doses — low-dose uses less product for the same protection — so the trigger, not the dose, drives practice.
2013
TOPPS. 600 hematology patients randomized to prophylactic platelets (≤10×10⁹/L) vs no prophylaxis. Prophylaxis reduced WHO grade ≥2 bleeding (43% vs 50%), validating the 10×10⁹/L trigger — though autologous-transplant patients benefited less.

What changed. Prophylactic platelet practice settled on a low trigger and a low dose, and on recognizing the settings — TTP, HIT, and ITP — where transfused platelets do nothing useful or actively harm.

In practice

Transfuse prophylactically at 10×10⁹/L; raise the bar to 50 for procedures and 100 for the brain. Before giving platelets for thrombocytopenia, ask why the count is low — in TTP and HIT, platelets can be dangerous, and the right move is to treat the underlying disorder.

Sources (3)
  • Kaufman RM, Djulbegovic B, Gernsheimer T, et al. Platelet transfusion: a clinical practice guideline from the AABB. Ann Intern Med. 2015;162(3):205–213. DOI 10.7326/M14-1589. PMID 25383671.
  • Stanworth SJ, Estcourt LJ, Powter G, et al; TOPPS Investigators. A no-prophylaxis platelet-transfusion strategy for hematologic cancers (TOPPS). N Engl J Med. 2013;368(19):1771–1780. DOI 10.1056/NEJMoa1212772. PMID 23656642.
  • Slichter SJ, Kaufman RM, Assmann SF, et al. Dose of prophylactic platelet transfusions and prevention of hemorrhage (PLADO). N Engl J Med. 2010;362(7):600–613. DOI 10.1056/NEJMoa0904084. PMID 20164484.

8Plasma, fibrinogen & reversal

Current approach (2026)
Plasma: only for bleeding/coagulopathyWarfarin: 4-factor PCC
Transfuse plasma (FFP) only for active bleeding or a planned invasive procedure with documented coagulopathy (typically INR > ~1.5–2 from multiple factor deficiencies), or as part of a massive-transfusion ratio. A mildly elevated INR does not predict procedural bleeding, and prophylactic plasma for minor abnormalities does not help and risks TACO/TRALI. Effective dosing is ~10–15 mL/kg. Replace fibrinogen (cryoprecipitate or fibrinogen concentrate) when it falls below ~1.5 g/L in bleeding (≤2 g/L in obstetric hemorrhage). For warfarin-associated major bleeding, use 4-factor prothrombin complex concentrate (PCC, ~25–50 U/kg by INR) plus IV vitamin K (5–10 mg) — faster and lower-volume than plasma; for direct oral anticoagulants, give idarucizumab for dabigatran; for factor-Xa inhibitors, 4-factor PCC is the practical option — andexanet alfa was withdrawn from the US market in December 2025 after ANNEXA-I showed excess thrombosis (14.6% vs 6.9% vs PCC), and where still available it must be weighed against that thrombotic risk.
Evidence & rationale
—
Plasma for high INR. Observational and randomized data show that mild INR elevations (≤1.8) correlate poorly with bleeding and that prophylactic plasma fails to correct them meaningfully while adding volume — the basis for “treat the patient, not the number.”
2015
4-factor PCC vs plasma (warfarin). Randomized trials in warfarin-associated major bleeding showed 4F-PCC achieves faster INR correction and effective hemostasis with less volume and fewer overload events than plasma — now first-line for urgent reversal.

What changed. Plasma transfusion narrowed sharply: away from “topping up” a borderline INR and toward defined roles — active coagulopathic bleeding, massive transfusion, and (for warfarin) replacement by PCC.

In practice

Don’t transfuse plasma to fix a number before a low-risk procedure. Reserve it for coagulopathic bleeding and MTP ratios; replace fibrinogen when low; and reverse warfarin with 4-factor PCC plus vitamin K, not plasma, when bleeding is severe.

Sources (4)
  • Sarode R, Milling TJ, Refaai MA, et al. Efficacy and safety of a 4-factor prothrombin complex concentrate in patients on vitamin K antagonists presenting with major bleeding: a randomized, plasma-controlled, phase IIIb study. Circulation. 2013;128(11):1234–1243. DOI 10.1161/CIRCULATIONAHA.113.002283. PMID 23935011.
  • Müller MC, Arbous MS, Spoelstra-de Man AM, et al. Transfusion of fresh-frozen plasma in critically ill patients with a coagulopathy before invasive procedures: a randomized clinical trial. Transfusion. 2015;55(1):26–35. DOI 10.1111/trf.12750. PMID 24912653.
  • Connolly SJ, Sharma M, Cohen AT, et al; ANNEXA-I Investigators. Andexanet for factor Xa inhibitor-associated acute intracerebral hemorrhage (ANNEXA-I). N Engl J Med. 2024;390(19):1745–1755. DOI 10.1056/NEJMoa2313040. PMID 38749032.
  • Carson JL, Stanworth SJ, Guyatt G, et al. Red blood cell transfusion: 2023 AABB international guidelines. JAMA. 2023;330(19):1892–1902. DOI 10.1001/jama.2023.12914. PMID 37824153.

9Transfusion reactions & complications

Current approach (2026)
TACO: leading transfusion deathStop & assess first
For any acute reaction, stop the transfusion, maintain IV access, and reassess (vitals, airway, the unit and patient identity). The two leading causes of transfusion-related death are pulmonary: TACO (transfusion-associated circulatory overload — hydrostatic pulmonary edema from volume/rate, now the commonest cause of transfusion mortality) and TRALI (transfusion-related acute lung injury — non-cardiogenic edema within 6 h, redefined in 2019 into Type I/II). Distinguish them by volume status and response to diuresis. Acute hemolytic reactions (usually ABO mismatch — a clerical error) cause fever, flank pain, and hemoglobinuria and are an emergency. Milder events — febrile non-hemolytic and allergic/urticarial — are common and managed symptomatically. Infectious transmission is now very rare; bacterial contamination (chiefly of platelets) is the main residual microbial risk.
ReactionClueFirst step
TACOHypertension, dyspnea, ↑BNP; fluid overloadStop; diurese; slow future rates
TRALIHypoxemia & bilateral infiltrates ≤6 h, normal fillingStop; supportive/lung-protective care
Acute hemolytic (ABO)Fever, flank pain, hemoglobinuriaStop; fluids; recheck ID; lab work
Febrile non-hemolyticIsolated fever/chillsStop; antipyretic; exclude hemolysis
Allergic / anaphylacticUrticaria → airway/shockAntihistamine; epinephrine if severe

What changed. As infectious risk collapsed, attention shifted to the non-infectious hazards — TACO above all. Slower rates, single units, and restrictive thresholds are themselves the most effective prevention; mitigation (leukoreduction, male-predominant plasma, pathogen reduction) addresses the rest.

In practice

Stop first, then diagnose. In a dyspneic patient mid-transfusion, weigh TACO (overloaded, hypertensive, diuresis helps) against TRALI (normal filling, diuresis doesn’t). Prevent TACO with single units, slower rates, and pre-emptive diuretics in at-risk patients. Treat suspected ABO hemolysis as an emergency and report every serious reaction to the transfusion service.

Sources (3)
  • Vlaar APJ, Toy P, Fung M, et al. A consensus redefinition of transfusion-related acute lung injury. Transfusion. 2019;59(7):2465–2476. DOI 10.1111/trf.15311. PMID 30993745.
  • Wiersum-Osselton JC, Whitaker B, Grey S, et al. Revised international surveillance case definition of transfusion-associated circulatory overload: a classification agreement validation study. Lancet Haematol. 2019;6(7):e350–e358. DOI 10.1016/S2352-3026(19)30080-8. PMID 31080132.
  • Goel R, Tobian AAR, Shaz BH. Noninfectious transfusion-associated adverse events and their mitigation strategies. Blood. 2019;133(17):1831–1839. DOI 10.1182/blood-2018-10-833988. PMID 30808635.

10Patient blood management & special products

Current approach (2026)
Treat anemia, don’t just transfuse itIrradiate for at-risk hosts
Patient blood management (PBM) treats the patient’s own blood as the resource to protect: diagnose and treat anemia (iron deficiency is the leading reversible cause — give IV iron when oral is inadequate or time is short), minimize iatrogenic blood loss (small-volume/closed sampling), and optimize hemostasis before procedures. Reserve transfusion for genuine need. Special product modifications target specific risks: irradiated cellular components prevent transfusion-associated graft-versus-host disease — indicated for components from first- or second-degree relatives or HLA-matched donors, and for at-risk hosts (Hodgkin lymphoma, purine-analogue/anti-thymocyte therapy, allogeneic stem-cell transplant, congenital T-cell immunodeficiency, and intrauterine or neonatal exchange transfusion); CMV-safe components (leukoreduced ± CMV-seronegative) for CMV-negative pregnancy, transplant, and severe immunosuppression; washed components for recurrent severe allergic reactions or IgA deficiency. Most countries now provide universal leukoreduction, which lowers febrile reactions, CMV transmission, and alloimmunization.

Why it matters. The single biggest lever on transfusion outcomes is not transfusing when the anemia can be treated at its cause. Preoperative iron-deficiency correction, restrictive triggers, and blood-loss reduction together cut exposure — and exposure drives both cost and the non-infectious hazards in Section 9.

In practice

Work up anemia before reaching for a unit — check iron studies and give IV iron when indicated rather than transfusing iron-deficiency anemia. Order the right product modification for the host: irradiated for the severely immunocompromised and neonates, CMV-safe for at-risk seronegative patients, washed for recurrent severe allergy. Lean on leukoreduction as the baseline.

Sources (3)
  • Mueller MM, Van Remoortel H, Meybohm P, et al; ICC PBM Frankfurt 2018 Group. Patient blood management: recommendations from the 2018 Frankfurt consensus conference. JAMA. 2019;321(10):983–997. DOI 10.1001/jama.2019.0554. PMID 30860564.
  • Foukaneli T, Kerr P, Bolton-Maggs PHB, et al; British Society for Haematology. Guidelines on the use of irradiated blood components. Br J Haematol. 2020;191(5):704–724. DOI 10.1111/bjh.17015. PMID 32808674.
  • Carson JL, Stanworth SJ, Guyatt G, et al. Red blood cell transfusion: 2023 AABB international guidelines. JAMA. 2023;330(19):1892–1902. DOI 10.1001/jama.2023.12914. PMID 37824153.

11Where it’s heading (2026)

Whole blood returns. Low-titer group O whole blood — one balanced product instead of three components — is expanding in both military and civilian trauma and prehospital systems; randomized and large observational data are maturing, with a survival signal in the most severely injured still being defined.

Prehospital & earlier resuscitation. Trials of prehospital plasma and red cells (e.g., PAMPer, RePHILL) gave mixed results, but the direction of travel is toward delivering balanced resuscitation and TXA before hospital arrival, where the time-dependent benefit is greatest.

Better-stored, longer-lasting products. Cold-stored platelets (longer shelf life, more hemostatically active for bleeding), pathogen-reduced components, and freeze-dried/lyophilized plasma aim to widen availability and shrink the residual infectious and logistical risks.

Precision over fixed ratios. Viscoelastic-guided (TEG/ROTEM) algorithms, fibrinogen concentrate, and point-of-care testing are refining massive transfusion from a single 1:1:1 recipe toward individualized, lab-guided replacement — without abandoning the empiric balanced start.

12Bottom line — the transfusion workflow

  1. Default restrictive. Transfuse red cells at Hb <7 g/dL for the stable, non-bleeding critically ill patient — including septic shock (TRICC, TRISS, AABB 2023).
  2. Know the exceptions. Use 8 g/dL after orthopedic surgery or with cardiovascular disease (FOCUS); a higher target (≈8–10) in active acute coronary syndrome (MINT, REALITY). Cardiac surgery and GI bleeding stay restrictive.
  3. Transfuse to physiology, one unit at a time. The hemoglobin number guides; bleeding, ischemia, or a supply–demand mismatch decide. Reassess after each unit.
  4. Major hemorrhage = MTP. Balanced ~1:1:1 (or whole blood), source control, and the lethal triad — warm, correct acidosis, replace ionized calcium (PROPPR).
  5. Early TXA where it works. 1 g ≤3 h in trauma and postpartum hemorrhage; not in GI bleeding (CRASH-2, WOMAN, HALT-IT).
  6. Platelets by setting. Prophylaxis at ≤10×10⁹/L; 50 for surgery/LP, 100 for the brain; avoid in TTP/HIT (TOPPS, AABB).
  7. Plasma for coagulopathic bleeding, not a number. Skip prophylactic plasma for a mildly high INR; reverse warfarin with 4-factor PCC + vitamin K.
  8. Watch the lungs. TACO is the leading transfusion death — single units, slower rates, pre-emptive diuresis; distinguish from TRALI.
  9. Manage the patient’s own blood. Treat iron deficiency (IV iron), limit phlebotomy, and order the right product modification (irradiated, CMV-safe, washed).
  10. Stop deliberately. End the MTP and re-evaluate thresholds as bleeding is controlled — overtransfusion has its own harms.

13FAQ

What hemoglobin level should trigger a blood transfusion in the ICU?
For a hemodynamically stable, non-bleeding critically ill adult, transfuse red cells when hemoglobin falls below 7 g/dL (the restrictive threshold endorsed by the 2023 AABB guideline, TRICC, and TRISS). Use 8 g/dL after orthopedic surgery or with pre-existing cardiovascular disease, and consider a higher target (about 8–10) during active acute coronary syndrome. Above these triggers, base the decision on physiology — bleeding or inadequate oxygen delivery — not the number alone.
Is a restrictive transfusion strategy safe in septic shock?
Yes. The TRISS trial randomized 998 patients with septic shock to a 7 versus 9 g/dL threshold and found no difference in 90-day mortality or ischemic events, while the restrictive group used about half as much blood. A threshold of 7 g/dL is the standard in septic shock for patients who are not actively bleeding or acutely ischemic.
When should tranexamic acid be used, and when should it be avoided?
Give tranexamic acid early — 1 g IV within 3 hours — in traumatic hemorrhage (CRASH-2) and postpartum hemorrhage (WOMAN), where it reduces death from bleeding; the benefit is lost or reversed if started after 3 hours. Do not use it for gastrointestinal bleeding: the HALT-IT trial showed no reduction in death and more venous thromboembolism and seizures.
What platelet count requires a prophylactic transfusion?
For hypoproliferative thrombocytopenia (for example from chemotherapy), transfuse prophylactically at a count of 10×10⁹/L or below (AABB; TOPPS). Use higher targets before procedures: about 20×10⁹/L for a central line, 50×10⁹/L for most surgery or lumbar puncture, and 100×10⁹/L for neurosurgery. Avoid platelets in TTP and HIT unless there is life-threatening bleeding, because they can promote thrombosis.
How do you tell TACO from TRALI?
Both cause pulmonary edema during or after transfusion. TACO (circulatory overload) is hydrostatic — the patient is volume-overloaded and hypertensive, has a raised BNP, and improves with diuresis. TRALI (acute lung injury) is non-cardiogenic edema within 6 hours, with normal filling pressures that does not respond to diuresis and needs lung-protective supportive care. TACO is now the leading cause of transfusion-related death, so slower rates and single units matter.
Should plasma be given to correct a mildly elevated INR before a procedure?
No. A mildly elevated INR (up to about 1.8) predicts procedural bleeding poorly, and prophylactic plasma neither corrects it meaningfully nor reduces bleeding, while adding the risk of circulatory overload and lung injury. Reserve plasma for active coagulopathic bleeding or massive transfusion, and reverse warfarin-associated major bleeding with 4-factor prothrombin complex concentrate plus vitamin K.

15Guidelines & evidence basis

Badges summarize each recommendation’s basis: Guideline: strong Guideline: conditional RCT benefit RCT neutral RCT harm. Guideline strengths use the GRADE classes quoted from the documents below.

Guidelines referenced (3)
  • Carson JL, Stanworth SJ, Guyatt G, et al. Red blood cell transfusion: 2023 AABB international guidelines. JAMA. 2023;330(19):1892–1902. DOI 10.1001/jama.2023.12914. PMID 37824153. — restrictive threshold of 7 g/dL for most hospitalized adults (8 g/dL for orthopedic surgery and pre-existing cardiovascular disease; 7.5 g/dL for cardiac surgery).
  • Kaufman RM, Djulbegovic B, Gernsheimer T, et al. Platelet transfusion: a clinical practice guideline from the AABB. Ann Intern Med. 2015;162(3):205–213. DOI 10.7326/M14-1589. PMID 25383671. — prophylactic platelet threshold of 10×10⁹/L and procedural thresholds.
  • Rossaint R, Afshari A, Bouillon B, et al. The European guideline on management of major bleeding and coagulopathy following trauma: sixth edition. Crit Care. 2023;27(1):80. DOI 10.1186/s13054-023-04327-7. PMID 36859355. — damage-control resuscitation, balanced ratios, early TXA, fibrinogen and calcium targets.

Disclaimer. This page is an educational summary of published evidence for clinicians and is not a substitute for individual clinical judgment or institutional protocols. Thresholds, doses, ratios, and targets are taken from the cited trials and guidelines; verify against current product labeling, local transfusion-service policy, and patient-specific factors before applying. “In practice” notes are general practical guidance. Recommendation strengths are quoted from the cited guidelines (GRADE). Reviewed by the DosePilot Medical Team (Jun 2026); all citations verified against PubMed (PMID) and DOI.