๐Ÿฉธ Transfusion & Coagulopathy in the ICU

AABB 2023 ISTH BSH TRICC/TRISS
Restrictive Hb 70 g/L MTP 1:1:1 ยท DIC Anticoagulant Reversal AABB 2023 ยท TRICC/TRISS/TRIGGER ยท CRASH-2 ยท Marino 5th Ed (2025)
๐Ÿ“… Last reviewed July 2026 ยท Next review January 2027 ยท Compiled by Dr. Anmol Srivastava Anaesthesia, Emergency Medicine & Critical Care Medicine ยท Reviewed by Dr. Tanya Chawla Anaesthesia & Critical Care
๐Ÿ“˜ 1 ยท Washington Manual of Critical Care, 4th Ed

Overview โ€” Less Blood Is Usually Better Blood

"For decades the reflex response to a low haemoglobin was to transfuse toward 'normal'. A generation of randomised trials has overturned that instinct: in most critically ill patients a restrictive strategy โ€” transfusing only when haemoglobin falls below 70 g/L โ€” is as safe as, or safer than, a liberal one. Every unit of blood carries cost and risk; the burden of proof now lies with the decision to transfuse, not to withhold."

Washington Manual of Critical Care, 4th Ed. Wolters Kluwer 2023. Chapter: Transfusion Medicine.

Why the Pendulum Swung to Restrictive

  • TRICC (1999): the landmark trial โ€” a restrictive threshold (70 g/L) was at least as good as liberal (100 g/L), with a survival signal favouring restrictive in younger and less sick patients.
  • TRISS (2014): in septic shock, 70 g/L was as safe as 90 g/L.
  • TRIGGER / upper-GI-bleed (Villanueva 2013): a restrictive threshold (70 g/L) improved survival and reduced rebleeding vs liberal (90 g/L).
  • Transfusion is not benign: immunomodulation (TRIM), volume overload (TACO), lung injury (TRALI), and โ€” in resource-limited settings โ€” infection risk and scarcity.
๐Ÿ“— 2 ยท Marino's The ICU Book, 5th Edition (2025)

Marino Physiology โ€” Oxygen Delivery, Not the Number

"The purpose of red cells is to deliver oxygen, and the relevant question at the bedside is never the haemoglobin in isolation but whether oxygen delivery is adequate for the patient's needs. A haemoglobin of 75 g/L in a young, euvolaemic, well-perfused patient is entirely adequate; the same value in a patient with ongoing ischaemia and a rising lactate may not be. Treat the patient and the physiology, not the laboratory value."

Marino PL. The ICU Book, 5th Ed. Anaemia & Red Cell Transfusion. Wolters Kluwer; 2025.
๐Ÿ“— Marino โ€” The Oxygen Delivery Equation

DOโ‚‚ = Cardiac Output ร— [(1.34 ร— Hb ร— SaOโ‚‚) + (0.003 ร— PaOโ‚‚)]

Haemoglobin is only one of three levers on oxygen delivery โ€” the others are cardiac output and arterial saturation. A patient with low DOโ‚‚ may be better served by improving cardiac output (fluids/inotropes) or oxygenation than by transfusion. The body also compensates for anaemia by increasing cardiac output and oxygen extraction, which is why chronic, euvolaemic anaemia is tolerated to surprisingly low levels.

๐Ÿ“— Marino โ€” Why Stored Blood Isn't Fresh Blood

Stored red cells undergo the "storage lesion": 2,3-DPG falls (left-shifting the Oโ‚‚ dissociation curve, so transfused cells initially release less Oโ‚‚), cells become less deformable, and potassium leaks into the supernatant. This is why transfusion does not instantly improve tissue oxygenation, and why large-volume transfusion of older units can cause hyperkalaemia (relevant in massive transfusion and in neonates/renal failure). Age of blood does not change the threshold โ€” large RCTs (ABLE, INFORM) found fresh blood gave no benefit over standard-issue blood.

๐Ÿ“‹ 3 ยท Transfusion Thresholds (AABB 2023)

When to Transfuse โ€” Component by Component

70 g/L
RBC โ€” most ICU patients
80 g/L
RBC โ€” cardiac surgery / pre-existing CAD
<10 ร—10โน
Platelets โ€” prophylactic (stable)
<50 ร—10โน
Platelets โ€” bleeding / pre-procedure
Red Cells
Hb <70 g/L โ€” restrictive threshold for most haemodynamically stable adults AABB 2023 Strong.
Hb <80 g/L โ€” for cardiac surgery and patients with pre-existing cardiovascular disease/active ACS.
Transfuse one unit at a time and reassess โ€” most non-bleeding patients need only a single unit. Do not transfuse on the number alone if asymptomatic and not bleeding.
Platelets
Prophylactic if <10 ร—10โน/L (stable, no bleeding). <20 with fever/sepsis or minor procedure. <50 for active bleeding, major surgery, or invasive procedure. <100 for neurosurgery/CNS bleeding. Avoid in TTP/HIT unless life-threatening bleeding โ€” adds fuel to thrombosis
Fresh Frozen Plasma (FFP) & Cryoprecipitate
FFP (12โ€“15 mL/kg) only for bleeding or pre-procedure with significant coagulopathy (e.g. INR >1.5โ€“2.0) โ€” NOT to "correct" a mildly abnormal INR in a non-bleeding patient. Cryoprecipitate for fibrinogen <1.5 g/L (<2.0 in obstetric haemorrhage) โ€” the most rapidly depleted factor in major bleeding. Vitamin K for warfarin/nutritional coagulopathy.
โš ๏ธ Don't Transfuse Plasma for a Number

The INR poorly predicts procedural bleeding risk, and FFP barely corrects a mildly raised INR (you would need huge volumes). Giving FFP to a non-bleeding patient with INR 1.4 before a line insertion exposes them to TACO/TRALI for no benefit. Reserve plasma for active bleeding or genuinely deranged coagulation before a high-risk procedure.

๐Ÿšจ 4 ยท Massive Haemorrhage & Transfusion Protocol

Massive Transfusion โ€” Balanced Resuscitation

Activate Early, Resuscitate Balanced
Massive transfusion โ‰ˆ โ‰ฅ10 units RBC/24h, or โ‰ฅ4 units/1h with ongoing bleeding. Give RBC:FFP:Platelets in a 1:1:1 ratio PROPPR 2015 โ€” empiric balanced resuscitation reproduces whole blood and reduces death from exsanguination versus RBC-heavy strategies. Activate the institutional Massive Haemorrhage Protocol early; don't wait for lab results to start.
Tranexamic Acid โ€” Give It Early
TXA 1 g IV over 10 min โ†’ 1 g over 8 h within 3 hours of injury CRASH-2. In trauma it reduces death from bleeding; benefit is lost (and possibly harmful) after 3 h. Also benefits postpartum haemorrhage (WOMAN trial) and is used in major surgical/GI bleeding.
Address the "Lethal Triad" & Calcium
Aggressively prevent/correct hypothermia, acidosis, and coagulopathy โ€” each worsens the others. Replace calcium: citrate in stored blood chelates ionised calcium; massive transfusion causes hypocalcaemia which impairs both clotting and cardiac contractility. Give calcium gluconate/chloride guided by ionised Caยฒโบ. Use goal-directed correction with viscoelastic testing (ROTEM/TEG) where available to target the specific deficit (fibrinogen, platelets, factors).
Permissive Hypotension (in trauma, pre-haemostasis)

Before surgical/radiological control of bleeding, target a lower BP (e.g. SBP ~80โ€“90 mmHg, or a palpable radial pulse / mentation in penetrating trauma) to avoid "popping the clot" and dilutional coagulopathy โ€” until haemorrhage is controlled. This does not apply to traumatic brain injury, where cerebral perfusion must be maintained.

๐Ÿงฌ 5 ยท DIC & ICU Thrombocytopenia

Coagulopathy of the Critically Ill

๐Ÿ“— DIC โ€” Consumption Gone Wrong

Disseminated intravascular coagulation is not a primary diagnosis โ€” it is always secondary to a trigger (sepsis is the commonest; also trauma, malignancy, obstetric catastrophe, snakebite). Systemic activation of coagulation consumes platelets and factors and generates fibrin (microvascular thrombosis + organ failure) while simultaneously causing bleeding. Labs: โ†“platelets, โ†‘PT/aPTT, โ†“fibrinogen, โ†‘โ†‘D-dimer; use the ISTH DIC score (โ‰ฅ5 = overt DIC).

DIC Management โ€” Treat the Cause
Treat the underlying trigger โ€” this is the only definitive therapy. Support with blood products only if bleeding or pre-procedure: platelets (<50 if bleeding), FFP (prolonged PT/aPTT + bleeding), cryoprecipitate (fibrinogen <1.5). In the predominantly thrombotic phenotype (or to prevent VTE), prophylactic heparin is reasonable once bleeding is controlled. Do not chase the numbers in a non-bleeding patient.
The ICU Thrombocytopenia Differential
Common and multifactorial: sepsis/DIC (most common), drug-induced, dilutional (massive transfusion), HIT, TTP/HUS (the microangiopathies โ€” do NOT give platelets), hypersplenism, immune (ITP), mechanical (IABP/ECMO/CRRT circuits).
HIT โ€” Don't Miss It (4Ts Score)
Heparin-Induced Thrombocytopenia: a prothrombotic immune reaction, typically a >50% platelet fall 5โ€“10 days after heparin exposure, with new thrombosis. Use the 4Ts score (Thrombocytopenia, Timing, Thrombosis, oTher causes). If suspected: STOP all heparin (including LMWH and flushes), start a non-heparin anticoagulant (argatroban/fondaparinux/danaparoid), send HIT antibodies โ€” and crucially do NOT transfuse platelets (fuels thrombosis) and do not start warfarin until platelets recover (venous limb gangrene risk).
๐Ÿ’Š 6 ยท Anticoagulant Reversal

Reversing the Bleeding Anticoagulated Patient

AgentReversalDose / Notes
Warfarin (major bleed)4-factor PCC + Vitamin KPCC 25โ€“50 IU/kg (INR-guided) IV โ€” faster & better than FFP; + Vitamin K 5โ€“10 mg IV (sustains reversal). FFP only if PCC unavailable.
Warfarin (high INR, no bleed)Vitamin K ยฑ holdINR >9: oral vit K 2.5โ€“5 mg + hold. Minor bleed: vit K IV. Don't over-reverse if mechanical valve.
Dabigatran (direct thrombin inh.)Idarucizumab5 g IV (two 2.5 g vials) โ€” specific antidote, rapid. Dialysis also removes dabigatran.
Apixaban / Rivaroxaban (anti-Xa)Andexanet alfa or PCCAndexanet alfa (where available/affordable) OR 4-factor PCC 50 IU/kg as practical alternative.
Unfractionated heparinProtamine sulphate1 mg per 100 U heparin given in last 2โ€“3 h (max 50 mg); slow IV (hypotension).
LMWH (enoxaparin)Protamine (partial)~60% reversal: 1 mg per 1 mg enoxaparin if within 8 h. No complete antidote.
Thrombolytic (tPA) bleedCryoprecipitate ยฑ TXAStop infusion; cryoprecipitate (fibrinogen), FFP, platelets; TXA; reverse the fibrinolytic state.
๐Ÿ‡ฎ๐Ÿ‡ณ Indian ICU Context

Idarucizumab and andexanet alfa are expensive and not universally stocked โ€” 4-factor PCC (or, where PCC is unavailable, FFP) is the pragmatic reversal for DOAC-associated life-threatening bleeding in many centres; activated charcoal helps if the last dose was <2โ€“4 h ago.

Snakebite (viper) consumption coagulopathy is a major cause of DIC-like bleeding in India โ€” the specific treatment is polyvalent anti-snake venom (ASV), not blood products alone; use the 20-minute whole-blood clotting test (20-WBCT) to guide ASV and recovery. Avoid unnecessary FFP/platelets unless actively bleeding after adequate ASV.

Blood product availability can be limiting โ€” restrictive thresholds and single-unit transfusion are doubly important stewardship in our setting.

โš ๏ธ 7 ยท Transfusion Reactions โ€” TACO vs TRALI

Recognising the Dangerous Reactions

TACO (Circulatory Overload) vs TRALI (Acute Lung Injury)
Both present as acute dyspnoea/hypoxaemia with bilateral infiltrates during or within 6 h of transfusion โ€” distinguishing them matters:
TACO: hydrostatic/volume overload โ€” hypertension, raised JVP/BNP, responds to diuretics. Commonest cause of transfusion death. Risk: elderly, cardiac/renal failure, fast/large transfusion. Prevent: transfuse slowly, single units, pre-emptive diuretic in at-risk patients.
TRALI: immune-mediated permeability lung injury โ€” often hypotension, fever, normal/low filling pressures, does not respond to diuretics; supportive/lung-protective care. Mitigated by male-predominant plasma donor policies.
Acute Haemolytic Reaction (ABO Incompatibility)
A medical emergency, usually from clerical/identification error. Fever, loin/flank pain, hypotension, haemoglobinuria, DIC, AKI โ€” even after a small volume. STOP the transfusion immediately, maintain IV access with saline, support BP and urine output, recheck identity/labels, send samples (DAT, repeat crossmatch, haemolysis screen), inform the blood bank. Prevention is rigorous bedside identity checking.
First Move for Any Acute Reaction

Stop the transfusion, keep the line open with saline, and assess ABC. Recheck the patient's identity against the unit. Mild febrile non-haemolytic or urticarial reactions may allow cautious restart after treatment; fever with hypotension, respiratory distress, or haemolysis means the unit does not go back up โ€” escalate and notify the blood bank.

โŒ 8 ยท Common Mistakes

Common Mistakes in Transfusion & Coagulopathy

โŒ Mistake 1 โ€” Transfusing to a "Normal" Haemoglobin

Liberal transfusion to Hb 90โ€“100 g/L in stable patients offers no benefit and may harm (TRICC, TRISS, TRIGGER). Use a restrictive 70 g/L threshold (80 for cardiac disease), transfuse single units, and reassess between units.

โŒ Mistake 2 โ€” Giving FFP to Correct a Number

FFP for a mildly raised INR in a non-bleeding patient barely changes the INR and exposes the patient to TACO/TRALI. Reserve plasma for active bleeding or significant coagulopathy before a high-risk procedure โ€” and use PCC (not FFP) for warfarin-related major bleeding.

โŒ Mistake 3 โ€” Transfusing Platelets in TTP or HIT

In the thrombotic microangiopathies (TTP/HUS) and in HIT, platelet transfusion can precipitate catastrophic thrombosis. Withhold platelets unless there is life-threatening bleeding; treat the underlying disorder (plasma exchange for TTP; non-heparin anticoagulation for HIT).

โŒ Mistake 4 โ€” Forgetting Calcium & Warmth in Massive Transfusion

Citrate chelates ionised calcium and stored blood is cold โ€” hypocalcaemia and hypothermia both worsen coagulopathy and cardiac function. Warm the blood and patient, monitor ionised calcium, and replace it during massive transfusion. And give TXA early (within 3 h).

โŒ Mistake 5 โ€” Treating DIC With Products Instead of the Cause

DIC is always secondary. Pouring in FFP and platelets without controlling the trigger (sepsis source control, delivery in obstetric DIC, ASV in snakebite) is futile. Support with products only if bleeding/pre-procedure; the cure is treating the cause.

โŒ Mistake 6 โ€” Blaming "TRALI" and Reaching for Diuretics, or Vice Versa

TACO and TRALI look alike but are managed oppositely โ€” diuresis helps TACO and is useless/harmful thinking in TRALI. Assess filling status (JVP, BNP, response to diuretic, blood pressure) to distinguish them, and prevent the commoner TACO by transfusing slowly in single units with pre-emptive diuretics in at-risk patients.

๐Ÿ“‘ 9 ยท References

References

  1. Carson JL, Stanworth SJ, Guyatt G et al. Red Blood Cell Transfusion: 2023 AABB International Guidelines. JAMA 2023;330:1892โ€“1902.
  2. Hรฉbert PC, Wells G, Blajchman MA et al. (TRICC). A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340:409โ€“417.
  3. Holst LB, Haase N, Wetterslev J et al. (TRISS). Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med 2014;371:1381โ€“1391.
  4. Holcomb JB, Tilley BC, Baraniuk S et al. (PROPPR). Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs 1:1:2 ratio. JAMA 2015;313:471โ€“482.
  5. CRASH-2 Collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients. Lancet 2010;376:23โ€“32.
  6. Marino PL. The ICU Book, 5th Ed. Anaemia, Transfusion & Haemostatic Disorders. Wolters Kluwer; 2025.
  7. Levi M, Toh CH, Thachil J, Watson HG. Guidelines for the diagnosis and management of disseminated intravascular coagulation. Br J Haematol 2009;145:24โ€“33 (ISTH/BSH).
  8. Washington Manual of Critical Care, 4th Ed. Chapters: Transfusion Medicine; Coagulopathy. Wolters Kluwer 2023.