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.
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.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.
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.
When to Transfuse โ Component by Component
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.
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.
Massive Transfusion โ Balanced Resuscitation
Before surgical/radiological control of bleeding, target a lower BP (e.g. SBP ~80โ90 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.
Coagulopathy of the Critically Ill
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: โplatelets, โPT/aPTT, โfibrinogen, โโD-dimer; use the ISTH DIC score (โฅ5 = overt DIC).
Reversing the Bleeding Anticoagulated Patient
| Agent | Reversal | Dose / Notes |
|---|---|---|
| Warfarin (major bleed) | 4-factor PCC + Vitamin K | PCC 25โ50 IU/kg (INR-guided) IV โ faster & better than FFP; + Vitamin K 5โ10 mg IV (sustains reversal). FFP only if PCC unavailable. |
| Warfarin (high INR, no bleed) | Vitamin K ยฑ hold | INR >9: oral vit K 2.5โ5 mg + hold. Minor bleed: vit K IV. Don't over-reverse if mechanical valve. |
| Dabigatran (direct thrombin inh.) | Idarucizumab | 5 g IV (two 2.5 g vials) โ specific antidote, rapid. Dialysis also removes dabigatran. |
| Apixaban / Rivaroxaban (anti-Xa) | Andexanet alfa or PCC | Andexanet alfa (where available/affordable) OR 4-factor PCC 50 IU/kg as practical alternative. |
| Unfractionated heparin | Protamine sulphate | 1 mg per 100 U heparin given in last 2โ3 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) bleed | Cryoprecipitate ยฑ TXA | Stop infusion; cryoprecipitate (fibrinogen), FFP, platelets; TXA; reverse the fibrinolytic state. |
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 <2โ4 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.
Recognising the Dangerous Reactions
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.
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.
Common Mistakes in Transfusion & Coagulopathy
Liberal transfusion to Hb 90โ100 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.
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.
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).
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).
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.
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.
References
- Carson JL, Stanworth SJ, Guyatt G et al. Red Blood Cell Transfusion: 2023 AABB International Guidelines. JAMA 2023;330:1892โ1902.
- 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;340:409โ417.
- Holst LB, Haase N, Wetterslev J et al. (TRISS). Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med 2014;371:1381โ1391.
- 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;313:471โ482.
- CRASH-2 Collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients. Lancet 2010;376:23โ32.
- Marino PL. The ICU Book, 5th Ed. Anaemia, Transfusion & Haemostatic Disorders. Wolters Kluwer; 2025.
- Levi M, Toh CH, Thachil J, Watson HG. Guidelines for the diagnosis and management of disseminated intravascular coagulation. Br J Haematol 2009;145:24โ33 (ISTH/BSH).
- Washington Manual of Critical Care, 4th Ed. Chapters: Transfusion Medicine; Coagulopathy. Wolters Kluwer 2023.