For most of the twentieth century, transfusion practice was governed by the so-called 10/30 rule, under which clinicians transfused red blood cells whenever hemoglobin fell below 10 grams per deciliter or hematocrit dropped below 30 percent. That convention rested largely on expert judgment rather than controlled evidence. The landscape shifted decisively in 1999 with the Transfusion Requirements in Critical Care (TRICC) trial, in which 838 critically ill adults in Canadian intensive care units were randomized to a restrictive strategy, transfusing at a hemoglobin threshold of 7 grams per deciliter, or a liberal strategy with a threshold of 10. The restrictive arm produced no increase in 30-day mortality, and in younger and less severely ill patients it was actually associated with lower mortality, alongside a substantial reduction in blood product use.

Subsequent landmark trials extended that message across specialties. The FOCUS trial, published in 2011, enrolled more than 2,000 older patients with cardiovascular risk factors after hip fracture surgery and found no differences in death, inability to walk unaided at 60 days, cardiovascular events, or hospital stay between a liberal threshold of 10 grams per deciliter and a restrictive threshold of 8. The TRACS trial in cardiac surgery compared a restrictive hematocrit target of roughly 24 percent with a liberal target of 30 percent and showed non-inferiority for mortality and major complications. The TITRe2 trial, spanning more than 2,000 cardiac surgery patients in the United Kingdom, confirmed fewer transfusions with a restrictive threshold of 7.5 grams per deciliter, although a secondary analysis hinted at a modest mortality increase that was not the primary endpoint, leaving interpretation contested.

Some of the most striking results came from gastroenterology. In a 2013 study, Villanueva and colleagues randomized patients with acute upper gastrointestinal bleeding to restrictive or liberal transfusion, using thresholds of 7 and 9 grams per deciliter respectively. The restrictive approach was associated with significantly lower mortality, less rebleeding, shorter hospital stays, and fewer complications, an effect particularly pronounced in patients with portal hypertension and cirrhosis. The proposed mechanism is physiological: excessive transfusion raises portal venous pressure, which can worsen variceal bleeding and impede hemostasis. More recently, the TRISS trial in septic shock randomized over 1,000 patients to hemoglobin targets of 7 or 9 grams per deciliter and found no significant differences in mortality or ischemic events, while the restrictive arm cut red cell use by roughly half.

The physiological rationale for these findings lies in the body’s remarkable capacity to compensate for moderate anemia. Oxygen delivery to tissues depends not only on hemoglobin concentration but also on cardiac output, arterial oxygen saturation, and microcirculatory function. As hemoglobin falls, the heart increases rate and stroke volume, peripheral vasodilation redistributes flow toward vital organs, and tissues extract a larger fraction of the oxygen carried in blood, rising from a normal utilization of about 25 percent to substantially higher levels. Red cells themselves adapt, increasing production of 2,3-bisphosphoglycerate, which shifts the oxygen-hemoglobin dissociation curve to the right and promotes oxygen unloading at the tissue level. These mechanisms explain why many stable patients tolerate hemoglobin levels well below traditional transfusion triggers.

Compensation, however, has limits. Below a critical threshold of oxygen delivery, oxygen consumption becomes supply-dependent, and tissue hypoxia, anaerobic metabolism, lactate accumulation, and organ dysfunction follow. This is why overly restrictive transfusion carries its own dangers, particularly in patients with limited physiological reserve: the elderly, those with coronary or severe lung disease, actively bleeding patients, neonates, and the critically ill with multiple comorbidities. The review emphasizes that the same hemoglobin value can mean very different things in different bodies, depending on cardiovascular reserve, metabolic demand, inflammatory burden, and ongoing blood loss.

Transfusion itself is not a benign intervention. Allogeneic blood exposure carries risks including transfusion-associated circulatory overload, transfusion-related acute lung injury, hemolytic reactions, alloimmunization, iron overload, transfusion-transmitted infection, and transfusion-related immunomodulation. Prolonged storage of red cell units produces the so-called storage lesion, a set of biochemical and structural changes that may impair erythrocyte deformability, reduce nitric oxide availability, and affect oxygen transport and microvascular flow, although its clinical significance remains debated. These hazards, together with global blood shortages and economic constraints, have driven the adoption of Patient Blood Management, an integrated approach endorsed by the World Health Organization, the AABB, the European Society of Anaesthesiology and Intensive Care, and the UK’s National Institute for Health and Care Excellence, which focuses on diagnosing and treating anemia before surgery, minimizing perioperative blood loss, and boosting the patient’s own tolerance of anemia.

Current international guidelines now converge on restrictive thresholds for most hemodynamically stable adults. The AABB recommends considering transfusion below 7 grams per deciliter in stable hospitalized adults, with a slightly higher threshold of about 8 for orthopedic and cardiac surgery patients and those with pre-existing cardiovascular disease. NICE recommends a threshold of 7 with a post-transfusion target of 7 to 9 grams per deciliter, and 8 to 10 for patients with acute coronary syndromes, along with single-unit transfusions followed by clinical reassessment in non-bleeding adults. The Surviving Sepsis Campaign and the European Society of Intensive Care Medicine similarly endorse a 7 gram per deciliter threshold for stable critically ill patients, while the American Society of Anesthesiologists favors individualized perioperative decisions informed by signs of inadequate oxygen delivery such as tachycardia, hypotension, metabolic acidosis, and elevated lactate.

Yet major uncertainties persist in precisely the populations where transfusion decisions matter most. Myocardial tissue extracts oxygen at near-maximal levels even under normal conditions, leaving little compensatory capacity, and trials in acute coronary syndromes have yielded conflicting results: the REALITY trial found restrictive transfusion non-inferior for major cardiovascular events in anemic myocardial infarction patients, while the MINT trial showed only suggestive trends favoring liberal transfusion that did not reach consistent statistical significance. In traumatic brain injury, cerebral tissue is exquisitely vulnerable to hypoxia, but liberal transfusion has not consistently improved neurological outcomes. Evidence also remains thin for patients on extracorporeal membrane oxygenation, those with hematologic malignancies, and frail older adults, who are simultaneously more susceptible to anemia and to transfusion-associated circulatory overload. In massive hemorrhage and trauma, fixed thresholds are largely abandoned in favor of goal-directed, balanced blood component therapy guided by clinical status and viscoelastic testing.

The review argues that the next frontier is precision hemotherapy: replacing the single-number trigger with multivariable algorithms that integrate hemoglobin with serum lactate, central or mixed venous oxygen saturation, near-infrared spectroscopy-derived tissue oxygenation, echocardiographic assessment, and emerging biomarkers of hypoxia, endothelial dysfunction, and iron metabolism such as hepcidin and erythroferrone. Artificial intelligence and machine learning promise to synthesize laboratory data, vital signs, comorbidities, and longitudinal electronic health records to estimate individual transfusion benefit and predict adverse reactions, though the author cautions that prospective validation, algorithmic transparency, and bias mitigation are prerequisites for safe clinical deployment. Digital health tools, continuous physiological monitoring, and wearable biosensors could allow transfusion decisions to evolve dynamically with the patient’s condition. Combined with Patient Blood Management, particularly in resource-constrained settings where blood shortages and weak hemovigilance remain acute challenges, and with greater attention to patient-reported outcomes such as fatigue, functional recovery, and quality of life, these innovations could transform hemotherapy from a threshold-driven reflex into a personalized, physiology-based therapy, conserving a finite biological resource while ensuring that every unit reaches the patient who truly needs it.

Subject of Research: Restrictive versus liberal blood transfusion strategies and the shift toward precision hemotherapy

Article Title: Towards safer and more effective hemotherapy: addressing critical unmet needs in restrictive and liberal blood transfusion strategies

Article References: Obeagu, E. I. (2026). Towards safer and more effective hemotherapy: addressing critical unmet needs in restrictive and liberal blood transfusion strategies. Frontiers in Hematology, 5, Article 1913853. https://doi.org/10.3389/frhem.2026.1913853

Image Credits: AI Generated

DOI: 10.3389/frhem.2026.1913853

Keywords: blood transfusion, restrictive transfusion strategy, liberal transfusion strategy, hemoglobin threshold, patient blood management, TRICC trial, transfusion-related complications, precision hemotherapy, acute coronary syndrome, artificial intelligence, tissue oxygenation, transfusion medicine