Patient blood management in general intensive care patients Authors: Patrick Meybohm, David M. Baron, Dietmar Fries, Sigismond Lasocki, Alexander P. J. Vlaar, Kai Zacharowski, Suma Choorapoikayil Published: 2026 ABSTRACT PurposeCritically ill and high-risk perioperative patients requiring intensive care are often multimorbid and depend on rapid, highly specialized management. While most comorbidities are difficult to modify in the acute setting, anemia, particularly iron-defi ciency anemia, represents a potentially modifiable risk factor. Clinicians are also often confronted with complex alterations in hemostasis that require rapid assessment and targeted therapeutic interventions, including the optimal use of blood products. This narrative review summarizes the current evidence on Patient Blood Management strategies, including anemia management, the use of small-volume tubes, and the appropriate use of blood products in intensive care unit patients.ResultsIntravenous iron supplementation can safely raise hemoglobin to a clinically meaningful degree. Erythropoietin therapy may also raise hemoglobin, but its use should remain selective given uncertain thromboembolic risk. Small-volume blood collection tubes and closed blood-conservation systems should be routinely used to reduce iatrogenic anemia. Current evidence supports restrictive red blood cell transfusion strategies in most clinical settings, particularly in patients with gastrointestinal bleeding. Similarly, a restrictive platelet transfusion strategy is supported by current evidence. Prophylactic platelet transfusion should be reserved for high-risk hematologic malignancies. In other critically ill patients with severe thrombocytopenia, a therapeutic (bleeding-driven) approach is preferred. Finally, current evidence does not support prophylactic fresh frozen plasma transfusion in non-bleeding patients.ConclusionsOverall, these findings support the implementation of Patient Blood Management strategies that optimize blood health and promote safer, more individualized care in critically ill patients.Visual abstract Introduction Anemia is very common in intensive care patients and is associated with increased transfusion requirements, prolonged intensive care unit (ICU) stays, and worse outcomes [1]. Its etiology is multifactorial and key contributing factors include hemolysis, frequent diagnostic phlebotomy, bleeding, reduced erythropoiesis, and (functional) iron deficiency, all of which contribute to a reduction in red blood cell (RBC) mass. In addition, the inflammatory state accompanying critical illness and major surgery or trauma promotes anemia of inflammation, characterized by altered iron metabolism, impaired iron mobilization, and reduced erythropoietin (EPO) production, further limiting erythrocyte synthesis. In intensive care patients, the diagnostic evaluation of anemia in ICU patients requires a systematic assessment to identify potentially reversible causes. Iron homeostasis is tightly regulated to ensure sufficient availability for oxygen transport, metabolic processes, and redox balance, while preventing excess free iron that may catalyze the formation of reactive oxygen species and cause cellular injury. Both acute and chronic anemia elicit compensatory physiological responses including increased cardiac output and altered oxygen extraction, which may further burden intensive care patients with limited physiological reserve. In addition, hemostatic abnormalities are common in intensive care patients. Hemostasis represents a complex, tightly regulated equilibrium between bleeding and clot formation, which is often disrupted in intensive care patients due to various factors, including disseminated intravascular or sepsis-associated coagulopathy, extracorporeal membrane oxygenation (ECMO), renal replacement therapy, and cardiac assistance devices [2]. In some cases, these disturbances can result in the co-occurrence of both thrombotic and hemorrhagic complications [3, 4]. Platelet and fresh frozen plasma (FFPs) transfusions are often administered to critically ill patients. However, the optimal transfusion thresholds remain poorly defined, and these blood products are often used prophylactically. In addition, both blood products are associated with adverse events, such as allergic and febrile reactions, as well as transfusion-related acute lung injury. Notably, the infection risk is greater with platelet transfusion due to platelet storage conditions [5–7]. Therefore, the appropriate use of platelets and FFPs is vital. Because certain forms of anemia and hemostatic disturbance are potentially modifiable, their detection and targeted treatment should be considered important therapeutic goals and form an essential part of comprehensive Patient Blood Management (PBM) strategies aimed at optimizing patient outcomes (Fig. 1). PBM is an interdisciplinary diagnostic, evidence-based approach that focuses on optimizing blood health and thereby preserving patients’ own blood resources [8–11]. PBM measures can be categorized into three pillars: (1) comprehensive anemia management; (2) minimizing surgical, procedural, and laboratory blood losses and managing coagulopathic bleeding throughout the care episode; and (3) evidence-based transfusion strategies at any time during the ICU stay and beyond.Fig. 1Schematic representation of risk factors in intensive care unit patients targeted by Patient Blood Management interventions and their potential benefits Schematic representation of risk factors in intensive care unit patients targeted by Patient Blood Management interventions and their potential benefits Several studies have demonstrated the efficacy of a multimodal PBM approach. In an analysis of 1618 patients, Lasocki et al. reported that PBM implementation was associated with a significant reduction in hospital length of stay and transfusion rate, as well as a 2.37-fold increase in the likelihood of preoperative anemia management (95% confidence interval [CI]: 1.2–4.7) [8]. Similarly, Meybohm et al. showed that among approximately 1.2 million surgical patients, PBM implementation was associated with a 13.9% relative reduction in the mean number of RBC units transfused per 1000 patients and a 10.5% relative reduction in transfusion rate [12]. In addition, a meta-analysis of 17 studies involving 235,779 surgical patients reported that implementation of a comprehensive PBM program was associated with clinically meaningful reductions in transfusion rates (− 39%), hospital length of stay (− 0.45 days), overall complication rates (− 20%), and mortality (− 11%) [9]. In recent decades, PBM has led to a more restrictive and patient-oriented approach to RBC transfusion. In the next step, attention is increasingly shifted toward the development of evidence-based strategies for the use of FFP and platelets, aiming to improve patient outcomes and ensure the safe and appropriate use of all blood products. PBM programs comprise multiple complementary measures that can be tailored to institutional needs and resources while enabling an individualized, patient-centered approach, which has been extensively optimized for the perioperative setting. Given the heterogeneous pathophysiology of critically ill patients, ranging from postoperative recovery to severe trauma or hemodynamic instability, PBM programs provide a framework of strategies to optimize blood health in this complex population. This narrative review summarizes the current evidence on anemia management and appropriate use of blood products, with a focus on the intensive care setting, including studies involving surgical patients requiring intensive care as well as studies exclusively assessing medical ICU patients. Intravenous iron in the intensive care unit Nutritionally derived anemia most often results from inadequate erythropoiesis due to deficiencies in iron, Vitamin B12, or folic acid. Among these, iron deficiency is the most common cause of anemia [13, 14]. Several randomized controlled trials (RCTs) and meta-analyses have evaluated intravenous (IV) iron supplementation in critically ill patients, including those recovering from major surgery. Overall, IV iron administration during critical illness resulted in a moderate increase in hemoglobin (Hb) concentrations, typically ranging from 0.3 to 0.7 g/dL, compared with placebo or standard care [15–17]. Although modest, this increase can be considered clinically meaningful in intensive care patients, given the need for frequent phlebotomy and impaired erythropoiesis. In the absence of erythrocyte destruction or sequestration, transfusion of one RBC unit is generally associated with a 1 g/dL increase in Hb concentration, or 3% increase in hematocrit, within 24 h [18–20]. Although the observed increases in Hb concentration, typically 0.3–0.7 g/dL, are modest compared with the post-transfusion increment, these changes may still be clinically meaningful, particularly those close to transfusion thresholds. Notably, the IRONMAN trial showed that IV iron supplementation not only increased Hb concentrations compared with no treatment (10.7 g/dL vs. 10.0 g/dL) but also reduced transfusion requirements (97 vs. 136 RBC units) [21]. A meta-analysis of 378 intensive care patients from four RCTs revealed that iron supplementation significantly increased mean Hb concentration by 0.31 g/dL (95% CI 0.04–0.59) after 10 days, whereas no significant increase was observed before 10 days [15]. Geneen et al. analyzed the effect of IV iron supplementation among 1198 patients admitted to the ICU due to general critical illness, trauma, cardiothoracic/cardiovascular surgery, and mixed causes, finding that IV iron was associated with higher Hb concentrations at 10–30 days, but not within the first 10 days [17]. In this context, it is notable that iron-dependent erythroid maturation takes approximately 4–6 days in healthy individuals. Therefore, a measurable effect on Hb following IV iron supplementation in intensive care patients is most likely detectable after at least 7 days and typically becomes more apparent within 10–30 days. Contemporary IV iron formulations appear to be well tolerated in intensive care patients, with no consistent evidence of increased risks of infection or other major adverse events across RCTs [17, 21]. However, most studies have been insufficiently powered to reliably detect rare adverse events, underscoring the need for ongoing pharmacovigilance. In this context, it is important to note that iron is not only an integral part of oxygen-carrying proteins but also of electron transport proteins and cytochromes. This transition metal readily accepts and donates electrons through reversible cycling between the ferrous (Fe2+) and ferric (Fe3+) states, a process that can catalyze the formation of reactive oxygen species. Therefore, systemic iron balance must be tightly regulated. Poorly ligated iron, also called catalytic iron, has been associated with neurological disorders, cancer, and cardiovascular disease [22, 23] and has been implicated in the pathogenesis of acute kidney injury. In a multivariable analysis adjusted for age and preoperative estimated glomerular filtration rate, Leaf et al. demonstrated that higher catalytic iron levels were associated with increased odds of in-hospital mortality, need for renal replacement therapy, and development of acute kidney injury in critically ill patients undergoing cardiac surgery with cardiopulmonary bypass [24, 25]. Potential sources of catalytic iron include hemolysis of transfused or endogenous RBCs, rhabdomyolysis, ischemia–reperfusion injury, hepatocellular injury, and infection-related cell death [26]. Lasocki et al. compared oxidative stress, using 8α-isoprostanes, advanced oxidation protein products, myeloperoxidase, reduced glutathione (GSH), and oxidized glutathione (GSSG) in critically ill patients and healthy volunteers after IV iron supplementation. The analysis revealed no difference in iron-induced plasma oxidative stress between both groups after iron supplementation [27]. Oxidative stress has been observed to occur transiently following IV iron supplementation and may be minimized by slow iron infusion [28]. Oral iron is a widely used and generally well-tolerated treatment for iron deficiency. However, its utility is substantially reduced in the ICU setting. It often fails to increase systemic iron stores because inflammation-driven elevation of hepcidin markedly suppresses intestinal iron absorption and iron utilization, particularly in critically ill patients and patients recovering from major surgery [29]. While evidence in general ICU cohorts remains limited, insights from related populations may inform practice. Following cardiac surgery, IV iron supplementation has shown potential to aid patient recovery and reduce persistent anemia [30]. Similarly, multiple RCTs and meta-analyses have demonstrated that IV iron supplementation improves functional capacity, reduces rehospitalizations, and may confer survival benefits in patients with heart failure [31–34]. Although these findings cannot be directly extrapolated to the ICU population, they highlight the pathophysiological relevance of iron deficiency and support ongoing investigation of targeted IV iron supplementation in intensive care patients, particularly during the rehabilitation phase. Third-generation IV iron formulations are associated with fewer complications compared with first-generation formulations. However, a potential risk of hypophosphatemia has recently been discussed controversially. Several contributing factors have been identified, including the severity of iron deficiency (as reflected by lower ferritin levels), pre-existing low serum phosphate levels, vitamin D deficiency, hyperparathyroidism, and the use of osteoporosis medications [35]. Notably, hypophosphatemia can be rapidly treated, particularly in the ICU setting. A practical recommendation for iron dosing and timing cannot currently be formulated due to a lack of data and substantial variation in patient populations, interventions, and overall patient management across ICUs (Table 1). However, it should be noted that persistent iron deficiency during the recovery phase may be associated with adverse outcomes, including cognitive dysfunction, fatigue, and cardiopulmonary impairment. Therefore, treatment of iron deficiency should be considered, taking into account several important contraindications and precautions [36]: known hypersensitivity to IV iron formulations, a history of serious allergic reactions to any parenteral iron product, iron overload states (e.g., hemochromatosis), chronic inflammatory conditions with unclear iron status, severe liver disease, and a history of asthma, eczema, or multiple drug allergies. Each patient should be carefully evaluated before initiating IV iron supplementation.Table 1Selected overview of treatment regimesAuthorStudy designDrugPatientID cut offXu et al. 2019Randomized, single‑blindedIV iron (iron sucrose Venofer) (weight based: weight [kg] × [target Hb − actual Hb] [g/L] × 0.24 + iron storage [mg]). 200 mg every other day until the total iron deficiency was achievedCardiovascular surgeryFerritin 30–100 mg/L or transferrin saturation < 20%IRONMAN et al. 2016Multicentre, randomized, blinded500 mg of ferric carboxymaltose in 100 ml of 0.9% saline delivered in two consecutive 50-ml syringesMixed (general surgical, cardiothoracic, neurosurgical, medical, trauma)Ferritin < 1200 mn/mlGarrido-Martin et al. 2012Randomized, double-blind, placebo-controlledIV iron (iron sucrose Venofer) (3 × 100 mg/day)Cardiothoracicn.a.Lasocki et al. 2016Multicenter, randomized, single‑blindedFerric carboxymaltose (1 g of iron over 15 min, according to product characteristics, seeICUAbsolute ID: hepcidin level of < 20 μg/LFunctional ID: as a 20 ≤ hepcidin < 41 μg/LPierraci et al. 2014Multicenter, randomized, single-blind, placebo-controlledIV Iron sucrose (Venofer; Luitpold Pharmaceuticals) 100 mg thrice weekly for up to 6 doses or until ICU dischargeTraumaPatients were excluded: with iron ferritin ≥ 1000 ng/mL, iron > 160 μg/dL, or transferrin saturation ≥ 50%Pierraci et al. 2009Randomized, double-blind, placebo-controlledEnteral ferrous sulphate 325 mg (oral solution or capsule) (Rockwell Compounding Inc.) thrice daily until hospital dischargeGeneral surgical, burns, neurosurgicaln.a.Madi-Jebara et al. 2004RandomizedIV iron (Venofer; Luitpold Pharmaceuticals) 200 mg/day to reach total iron deficitCardiothoracicn.a.van Iperen et al. 2000RandomizedIV iron saccharate (20 mg daily Venofer; Vifor)Mixed (surgical, medical, neurological, trauma)Patients were excluded: with ferritin < 50 µg/L Table 1Selected overview of treatment regimesAuthorStudy designDrugPatientID cut offXu et al. 2019Randomized, single‑blindedIV iron (iron sucrose Venofer) (weight based: weight [kg] × [target Hb − actual Hb] [g/L] × 0.24 + iron storage [mg]). 200 mg every other day until the total iron deficiency was achievedCardiovascular surgeryFerritin 30–100 mg/L or transferrin saturation < 20%IRONMAN et al. 2016Multicentre, randomized, blinded500 mg of ferric carboxymaltose in 100 ml of 0.9% saline delivered in two consecutive 50-ml syringesMixed (general surgical, cardiothoracic, neurosurgical, medical, trauma)Ferritin < 1200 mn/mlGarrido-Martin et al. 2012Randomized, double-blind, placebo-controlledIV iron (iron sucrose Venofer) (3 × 100 mg/day)Cardiothoracicn.a.Lasocki et al. 2016Multicenter, randomized, single‑blindedFerric carboxymaltose (1 g of iron over 15 min, according to product characteristics, seeICUAbsolute ID: hepcidin level of < 20 μg/LFunctional ID: as a 20 ≤ hepcidin < 41 μg/LPierraci et al. 2014Multicenter, randomized, single-blind, placebo-controlledIV Iron sucrose (Venofer; Luitpold Pharmaceuticals) 100 mg thrice weekly for up to 6 doses or until ICU dischargeTraumaPatients were excluded: with iron ferritin ≥ 1000 ng/mL, iron > 160 μg/dL, or transferrin saturation ≥ 50%Pierraci et al. 2009Randomized, double-blind, placebo-controlledEnteral ferrous sulphate 325 mg (oral solution or capsule) (Rockwell Compounding Inc.) thrice daily until hospital dischargeGeneral surgical, burns, neurosurgicaln.a.Madi-Jebara et al. 2004RandomizedIV iron (Venofer; Luitpold Pharmaceuticals) 200 mg/day to reach total iron deficitCardiothoracicn.a.van Iperen et al. 2000RandomizedIV iron saccharate (20 mg daily Venofer; Vifor)Mixed (surgical, medical, neurological, trauma)Patients were excluded: with ferritin < 50 µg/L Selected overview of treatment regimes Absolute ID: hepcidin level of < 20 μg/L Functional ID: as a 20 ≤ hepcidin < 41 μg/L In summary: (1) Evidence suggests that IV iron supplementation can safely achieve a modest but clinically meaningful increase in Hb concentration. (2) Early treatment of iron deficiency is recommended. (3) If anemia is multifactorial, then iron alone may be insufficient to stimulate erythropoiesis. (4) Future research should aim to define homogenous patient populations and stratify analyses by relevant confounding factors when assessing the impact of IV iron supplementation on Hb. However, given the current limitations of the evidence, high-quality prospective trials in ICU populations are needed to better define the clinical efficacy and safety of IV iron supplementation. Erythropoietin therapy in the intensive care unit Erythropoiesis-stimulating agents (ESAs), including EPO, have been studied extensively in ICU cohorts over the past 2 decades, and conflicting results have been reported regarding their ability to increase Hb and hematocrit concentrations [37–41]. Their regulatory approval varies by country but generally includes anemia associated with chronic kidney disease, myelosuppressive chemotherapy, zidovudine-treated HIV infection, and anemic preterm infants. Carpentier et al. examined the impact of EPO on RBC transfusion requirements and Hb concentrations in 190 severely ill patients. The authors found that those treated with EPO required significantly fewer RBC transfusions (0 [0–2] vs. 1 [0–3]; p < 0.001), and Hb concentrations were 1.2 g/dL higher [42]. In addition, several RCTs suggest a reduction in RBC transfusion requirements [37, 43]. For example, Litton et al. reported that, among 4075 critically ill patients from 10 RCTs, EPO administration was associated with a significant 12% reduction in the risk of receiving an RBC transfusion (relative risk [RR 0.88]: 95% CI 0.78–0.98; p = 0.024) [43]. In addition, EPO has been shown to reduce mortality in ICU patients [37, 38, 42, 44]. Subgroup analyses, particularly in trauma or neurotrauma populations, have hinted at more important survival benefits in these patients [43]. The EPO-TRAUMA trial might shed further insight into whether EPO can reduce mortality and severe disability in critically ill trauma patients following trauma. However, concerns have been raised regarding potential thromboembolic complications of EPO therapy. Currently, pooled analyses have not demonstrated a consistent increase in thrombotic risk, although most studies were underpowered to detect rare but clinically important adverse events [37, 43]. Proinflammatory cytokines, such as interleukin-1 alpha (IL1A/IL1) and tumor necrosis factor (TNF/TNF-α), suppress EPO production, thereby inhibiting the proliferation of erythroid progenitors. Administration of high EPO concentrations promotes the generation of more normoblasts and, eventually, reticulocytes [45]. Importantly, the increase in circulating reticulocytes becomes significant only after 3–4 days [44]. Furthermore, the Hb concentration increases slowly in patients with inflammatory conditions, which may contribute to the observed variability in outcomes such as Hb response or RBC transfusion. Cost-effectiveness analysis of EPO therapy in the ICU setting is challenging due to the heterogeneous patient population, difficulties in defining clinically relevant outcomes, and high cost variability. MacLaren et al. performed a decision analysis incorporating costs associated with RBC production and transfusion, as well as the adverse effects of recombinant human EPO (rHuEPO) and RBC transfusion. Using rHuEPO therapy to reduce blood transfusions added approximately US$1400–$1900 per patient while providing a small health benefit, translating into a cost of roughly US$34,000–$47,000 per quality-adjusted life year (QALY) gained. Overall, the authors concluded that rHuEPO therapy was cost-effective in about half of the simulated scenarios, assuming society is willing to pay up to $50,000 for each additional year of healthy life (per QALY) [46]. However, the EPO patent expired several years ago, and cost savings may now be greater. The current prices for 40,000 UI/mL epoetin-α range between €35 in Germany, €49 in Spain, €61 in France, up to €632 in the USA. Estimated costs per life saved range from €3900 in Germany up to €70,100 in the USA [40]. Currently, the European Society of Intensive Care Medicine’s clinical practice guidelines do not recommend routine EPO therapy in unselected ICU patients. Instead, EPO therapy remains indicated in specific populations such as chronic kidney disease, chemotherapy-induced anemia, and perioperative anemia [37, 43]. In summary: (1) Current evidence suggests that EPO therapy use in critically ill or intensive care patients may reduce RBC transfusion requirements, increase Hb concentrations, and reduce mortality in selected subgroups. (2) In critical illness, selective use of EPO therapy may be considered in highly transfusion-dependent patients or those with iron deficiency, where EPO may be considered to support erythropoiesis. (3) However, the risk of thromboembolic complications associated with EPO therapy remains uncertain due to limited evidence. (4) Future research is needed to clarify the benefit–risk profile of EPO therapy in critically ill patients. Optimizing diagnostic blood sampling practices Laboratory testing is an essential component of perioperative care. However, clinicians are often unaware of the frequency of routine laboratory testing and the volume of blood involved. Diagnostic phlebotomy is believed to contribute to hospital-acquired anemia (HAA) and to the severity of anemia, both of which have been shown to adversely affect health recovery [47, 48]. The daily diagnostic blood volume varies from 22 to 377 mL [49–52]. Over the past several years, healthcare providers have recognized the potential to minimize the blood required for diagnostic testing and to preserve patients’ blood resources. However, only a few hospitals have successfully decreased the total volume of blood drawn for diagnostic purposes [49, 52–54]. Siegal et al. conducted one of the largest stepped-wedge, cluster-randomized trials, evaluating the transition from standard-volume (4.0–6.0 mL) to small-volume tubes (1.8–3.5 mL) blood collection tubes in 25 medical-surgical ICUs across Canada, involving more than 21,000 patients. Among patients with an ICU stay of ≥ 24 h, the transition from standard- to small-volume blood collection tubes was associated with an absolute difference of 9.84 (95% CI 0.24–20.76) RBC units transfused per 100 patients [55]. In a cohort of 400 major surgical patients, Neef et al. reported that switching from standard blood collection tubes to small-volume blood collection tubes (blood gas analysis: from 2 to 1 mL, EDTA: from 2.7 to 1.8 mL, SCC: from 3.0 to 1.8 mL, and serum: from 4.7 to 2.6 mL) was associated with a reduction of 82.8 mL in blood volume for patients with an ICU stay of 2 days and up to 824.0mL for those with an ICU stay of ≥ 11 days [56]. Beyond diagnostic phlebotomy, blood is also lost through catheter clearance, with discarded volumes of 2–10 mL [57]. In addition to small-volume blood collection tubes, the use of closed blood sampling systems is recommended to minimize iatrogenic blood loss and reduce the risk of anemia in intensive care patients [52, 58]. Given that bedside waste per blood draw is 3.9 mL for arterial lines, 5.5 mL for central venous catheters (CVC), and 6.3 mL for peripheral venous catheters [57], the use of blood-conservation devices is recommended. In the ENCLOSE trial, Keogh et al. showed that the use of closed blood sampling systems and clinically indicated blood sampling significantly reduced the mean daily blood sample losses by 50% [59]. The adoption of small-volume blood collection tubes has been hindered by unsubstantiated concerns about rejected samples due to insufficient volume and erroneous laboratory results. However, laboratory instrumentation has advanced, enabling testing of smaller blood volumes. Taken together, multiple studies have consistently demonstrated that small-volume blood collection tubes are an excellent alternative to standard-sized tubes without compromising patient care or the quality of laboratory results. Particularly, the Intensive Care Medicine Rapid Practice Guideline by Callum et al. strongly recommends the use of small-volume blood collection tubes in ICU patients [60]. The greatest benefit may be achieved by combining several measures that minimize blood loss and enhance erythropoiesis. For instance, Warner et al. demonstrated the efficiency and feasibility of a multifaceted anemia prevention and treatment bundle consisting of optimized phlebotomy practices, clinical decision support, and anemia treatment with IV iron supplementation. They found that Hb concentrations 1 month post-discharge were higher in the intervention cohort than in the standard care cohort (12.2 [11.8–13.0] g/dl vs. 11.5 [10.2–12.6] g/dL) [61]. In summary: (1) Current evidence suggests the use of small-volume blood collection tubes, blood conservation devices, and bundling of tests onto single collection tubes. (2) The use of point-of-care tests using minute blood volumes is also warranted. (3) Future research should further investigate the association between Hb recovery and ICU and post-ICU outcomes, including transfusion requirements and related adverse events. Red blood cell transfusion thresholds The appropriate use of allogeneic blood products has been rigorously debated for decades, driven by uncertainty about transfusion thresholds. The lack of consensus regarding the “optimal” transfusion threshold was further highlighted by the InPUT study, one of the largest observational studies to date, including 3643 critically ill patients admitted to 233 ICUs across 30 countries on six continents. Of these patients, 894 (25%) received a total of 1727 RBC transfusions, corresponding to a median of 2 transfusions per patient (interquartile range [IQR] 1–4), with 1 RBC unit per transfusion (IQR 1–2) and 0.5 RBC units per day (IQR 0.25–1) with marked regional variability in transfusion practices [62]. Several RCTs across various patient populations suggest that a restrictive transfusion threshold of Hb 7.0 g/dL (4.3 mmol/L) is safe for most critically ill patients [63–68]. However, mixed observations have been reported in neurocritical care patients and patients with acute coronary syndrome (ACS), for whom the optimal transfusion threshold may be slightly higher (Hb thresholds < 9 to ≤ 10 g/dL) [69–71]. In this context, the compensatory capacity may differ among neurocritical care patients with traumatic brain injury (TBI), ischemic stroke, subarachnoid hemorrhage (SAH), and intracerebral hemorrhage (ICH). The brain compensates for reduced oxygen delivery through several mechanisms, including increased cardiac output, redistribution of blood flow to prioritize cerebral perfusion, vasodilation to enhance cerebral blood flow, and increased cerebral oxygen extraction. However, these mechanisms may be impaired following brain lesions, leaving patients at risk of anemia-induced secondary brain injury. The SAHARA trial showed that, in patients with acute aneurysmal SAH, restrictive transfusion (optional transfusion at Hb ≤ 8 g/dL) was more often associated with an unfavorable neurological outcome in 37.7% of patients, compared with 33.5% in the liberal transfusion group (mandatory transfusion at Hb ≤ 10 g/dL) (RR: 0.88 [95% CI 0.72–1.09] p = 0.22) [71]. In patients with TBI, the HEMOTION trial demonstrated that an unfavorable outcome (Glasgow Outcome Scale < 4) occurred in 68.4% of patients in the liberal strategy group (transfusions at Hb ≤ 10 g/dL) and in 73.5% in the restrictive strategy group (transfusions at ≤ 7 g/dL), with an absolute difference of 5.4% (95% CI − 2.9–13.7) [72]. The TRAIN trial, which included patients with TBI, SAH, or ICH and a Glasgow Outcome Extended Scale ≤ 13, showed that the proportion of patients with unfavorable neurological outcome at 180 days (Glasgow Outcome Extended Scale 1–5) was significantly lower in the liberal transfusion group (transfusion at Hb < 9 g/dL) (62.6%) than in the restrictive transfusion group (transfusion Hb < 7 g/dL) (72.6%) (RR 0.86 [95% CI 0.79–0.94] p = 0.002) [70]. These studies indicate that a more liberal transfusion strategy may offer clinical benefits and individualized Hb targets of 9–10 g/dL should be considered. Kougias et al. compared liberal (Hb > 10 g/dL) and restrictive (Hb > 7 g/dL) postoperative transfusion strategies in 1424 veteran patients with cardiovascular diseases undergoing vascular surgery, finding no significant difference in ischemic events or mortality [73]. However, female patients were underrepresented, thereby limiting the generalizability of these findings. In patients with acute myocardial infarction (MI) and anemia, the compensatory capacity to increase coronary blood flow is limited, which may exacerbate myocardial ischemia. The MINT trial revealed no overall differences in outcomes, as maintaining a higher Hb concentration was not superior in reducing mortality or the incidence of new or recurrent MIs after randomization [64]. Interestingly, a secondary analysis of the MINT trial showed that patients with acute-onset anemia experienced more severe clinical courses than those with chronic anemia. However, no significant differences were observed between the restrictive and liberal transfusion groups in this context [74]. Notably, when distinguishing between type 1 and type 2 MI, a restrictive transfusion strategy was associated with higher rates of mortality in patients with type 1 MI compared with type 2 MI [75]. A physiological explanation for this finding may relate to the high baseline myocardial oxygen extraction ratio, which approaches maximal levels at rest, thereby limiting capacity to further increase oxygen extraction in response to reduced arterial oxygen content. Therefore, reductions in Hb concentrations may disproportionately impair myocardial oxygen delivery, increasing the susceptibility to ischemia compared with other organ systems. In a recently published Cochrane review, Carson et al. compared outcomes across 27,639 patients in 61 trials who received a transfusion according to either a restrictive (7.0–8.0 g/dL) or a liberal (7.0–8.0 g/dL) Hb threshold. Across most clinical settings, restrictive transfusion strategies were not associated with a difference in 30-day all-cause mortality when compared with liberal transfusion strategies (RR 1.01 [95% CI 0.90–1.14] high-certainty evidence). However, in patients with gastrointestinal bleeding, restrictive transfusion strategies were associated with a lower risk of 30-day mortality (RR 0.63 [95% CI 0.42–0.95]). Furthermore, no significant effects were observed for patients with MI (RR 1.10 [95% CI 0.96–1.27]), stroke (RR 1.10 [95% CI 0.96–1.27]), thromboembolic events (RR 0.88 [95% CI 0.67–1.15]), or infections (RR 1.02 [95% CI 0.94–1.11]), based on evidence of moderate to high certainty [76]. Furthermore, the authors highlighted that transfusion thresholds should be individualized according to specific clinical contexts while also accounting for relevant physiological parameters beyond Hb in clinical decision-making, particularly indicators of oxygen delivery to vital organs and markers of metabolic stress when determining RBC transfusion volume. Ongoing studies continue to investigate whether the 7.0 g/dL Hb threshold is beneficial for the general ICU population and healthcare providers. Especially in elderly patients, the potential benefit of a restrictive transfusion remains unclear. Because elderly patients have a higher prevalence of cardiovascular comorbidities and reduced physiological reserves, their compensatory mechanisms for anemia are often impaired. As a result, they may be more vulnerable to ischemic events and perioperative complications. The ongoing LIBERAL trial is assessing whether a liberal transfusion strategy may reduce adverse outcomes in these patients [77]. Patients with acute respiratory distress syndrome (ARDS) may require ECMO support. In this setting, bleeding and thrombotic complications are common, and the optimal transfusion threshold remains uncertain. Martucci et al. reported in a cohort of 604 patients with ARDS receiving venovenous ECMO that RBCs were administered on 31% of ECMO days and that transfusion at a Hb threshold < 7.0 g/dL was associated with improved survival. Notably, the authors found that optimal transfusion Hb thresholds during ECMO may vary over time, depending on the intensity of extracorporeal support of the patient’s fluid balance [78]. The OBLEX study assessed major hemorrhage and blood product use among 545 patients receiving venoarterial ECMO. In total, 406 major hemorrhagic events were observed in 286 patients, and 481 patients received RBC transfusions. ECMO initiated during cardio-pulmonary resuscitation (incidence rate ratios 1.50 [95% CI 1.19–1.89]) and prior use of antiplatelets (incidence rate ratios 1.43 [95% CI 1.13–1.80]) were identified as predictors for RBC transfusion [79]. Data on 534 patients in the OBLEX study were used to emulate a target trial comparing in-hospital mortality between a liberal (Hb ≥ 9 g/dL) and a restrictive (Hb ≤ 7 g/dL) transfusion strategy with respect to hospital mortality. After adjusting for confounding factors, including age, gender, renal replacement prior to ECMO, extracorporeal cardiopulmonary resuscitation, surgery, bleeding, mechanical circulatory support, and renal replacement during ECMO, the liberal transfusion strategy was associated with improved survival within the first 2 days of ECMO (12% [95% CI 3–21%]). No differences in survival were observed beyond day 3. However, the authors emphasized the need for further prospective studies to determine the optimal transfusion strategy [80]. A meta-analysis of five studies further indicated that a restrictive transfusion Hb threshold of 7 g/dL was associated with improved 28-day survival (OR 4.03 [95% CI 1.35–11.84] p = 0.011) [81]. However, based on the currently available evidence, no definitive transfusion Hb threshold can be recommended for this patient population. A one-size-fits-all approach to transfusion Hb thresholds for all patients of the ICU cannot be established, because of the heterogeneity in patients’ illness severity and clinical contexts. Consequently, balancing the risks of over-transfusion and under-transfusion is particularly challenging in intensive care patients. Higher Hb transfusion thresholds may be indicated during perioperative acute blood loss with hemodynamic instability than in hemodynamically stable, normovolemic intensive care patients, who can tolerate a lower Hb threshold [82, 83]. However, it is evident that judicious stewardship of the valuable blood resource is urgently required. In summary: (1) Current evidence supports the safety of restrictive transfusion strategies across most clinical settings, particularly in patients with gastrointestinal bleeding. (2) In contrast, liberal transfusion strategies may be superior in patients with type 1 MI and in neurocritically ill patients. (3) Future research is warranted to define optimal transfusion thresholds in specific clinical conditions and vulnerable patient populations. Such clarification may help inform the design of future RCTs in the broader ICU population. Platelet transfusion thresholds Trials comparing prophylactic platelet transfusion thresholds of 10,000/mm3 versus 20,000/mm3 in patients with hematologic malignancies have demonstrated that a lower threshold was safe and significantly reduced transfusion requirements [84–87]. Randomized trials such as PATCH [88] and MATISSE [89] have shown that platelet transfusion can be detrimental in patients with intracranial injury or those at risk of intracranial complications. Both trials demonstrated that platelet transfusion was associated with worse long-term outcomes compared with no transfusion or a restrictive transfusion strategy. These findings were subsequently applied in the ICU and other settings and incorporated into current guidelines [63, 87]. The 2020 European Society of Intensive Care Medicine guidelines recommend against routine prophylactic platelet transfusion prior to invasive procedures in non-bleeding critically ill patients with platelet counts above 100 × 109/L [63]. They also suggest a threshold of 50 × 109/L for selected procedures such as central venous catheter placement and tracheostomy (based on low to very low certainty evidence), with no recommendation for counts between 10 and 50 × 109/L. In addition, The 2025 AABB and ICTMG International Clinical Practice Guidelines on Platelet Transfusion from the American Association of Blood Banks and the International Collaboration for Transfusion Medicine Guidelines recommend a restrictive platelet transfusion strategy in the following clinical scenarios: (1) non-bleeding patients with hypoproliferative thrombocytopenia receiving chemotherapy or undergoing allogeneic hematopoietic stem cell transplantation, (2) preterm neonates without major bleeding, (3) patients undergoing lumbar puncture; and (4) patients with dengue-associated consumptive thrombocytopenia in the absence of major bleeding [87]. The PACER trial compared platelet transfusion strategies in ICU and hematologic patients with severe thrombocytopenia undergoing CVC placement [90]. Withholding prophylactic platelet transfusion in patients with platelet counts between 10,000 and 50,000/mm3 did not meet the predefined noninferiority margin and resulted in more CVC-related bleeding events. Subgroup analyses revealed that these bleeding events occurred mainly in hematologic patients, suggesting that prophylactic platelet transfusion remains appropriate in this group. Furthermore, a therapeutic approach may be reasonable in intensive care patients with severe thrombocytopenia. Furthermore, when classifying platelet disorders or quality control of platelet concentrates, platelet function may be more relevant than platelet count. Therefore, measuring platelet aggregation in response to stimulation is crucial for the diagnosis of platelet disorders and for monitoring patients receiving antiplatelet therapy [91, 92]. Recently, Coz Yataco et al. developed seven Population, Intervention, Comparator, and Outcome (PICO) questions addressing platelet and FFP transfusions in critically ill patients and conducted a comprehensive evidence review. Based on 16 studies with a low level of certainty, the authors formulated recommendations for critically ill patients, excluding those with trauma or neurocritical conditions. In stable, non-bleeding critically ill patients with thrombocytopenia and without a high risk of spontaneous bleeding, platelet transfusion is recommended when platelet counts are < 10 × 109/L. In patients with thrombocytopenia and active major bleeding, platelet transfusion is recommended when platelet counts are < 50 × 109/L. Furthermore, in critically ill patients with an increased risk of bleeding due to thrombocytopenia undergoing CVC or arterial line insertion, bedside thoracentesis or paracentesis, flexible bronchoscopy without biopsy, or gastrointestinal endoscopy, routine prophylactic platelet transfusion is not recommended. In critically ill patients with an increased risk of bleeding due to thrombocytopenia undergoing bedside lumbar puncture, platelet transfusion is recommended when platelet counts are ≤ 40–50 × 109/L [93]. Importantly, the authors emphasized the lack of well-designed prospective trials and the urgent need for further research in this area. In summary: (1) Current evidence supports a restrictive platelet transfusion strategy. (2) Prophylactic transfusion should be reserved for high-risk hematologic patients, while a therapeutic approach is reasonable for critically ill patients with severe thrombocytopenia. (3) Future research should validate whether patients with sepsis and preserved bone marrow function can safely tolerate lower platelet thresholds in the ICU setting, determine optimal transfusion strategies for moderate-to-severe bleeding and prophylaxis before surgery, and identify measurable predictors of bleeding risk. Fresh frozen plasma in the intensive care unit Fresh frozen plasma is still frequently administered to patients with coagulopathy, although the clinical benefit of FFP transfusion remains elusive. Additional reports have consistently linked FFP transfusion to complications such as nosocomial infections, ARDS, prolonged mechanical ventilation, multi-organ failure, and transfusion-associated circulatory overload [94]. To date, no RCTs or high-quality prospective studies have demonstrated a reduction in morbidity or mortality associated with FFP transfusion in intensive care patients, and current evidence is largely derived from observational studies that are inherently susceptible to confounding by indication. However, a prospective cohort study by Johnson et al. found that early FFP transfusion may increase the risk of multi-organ failure in critically injured patients [95]. Importantly, evidence-based guidelines do not support the routine use of FFP for prophylactic correction of coagulation abnormalities or for volume replacement in critically ill patients. Even in patients with acute bleeding, fibrinogen is typically the first coagulation factor to reach critically low levels, and FFP transfusion is not recommended as first-line replacement therapy [82, 83, 96, 97]. It is crucial to distinguish between prophylactic plasma transfusion in non-bleeding intensive care patients and the management of active major bleeding, as these represent fundamentally different clinical scenarios. In the absence of active bleeding, prophylactic FFP transfusion aimed solely at correcting laboratory coagulation abnormalities is not supported by evidence and is not recommended. In contrast, during major bleeding or massive transfusion, plasma transfusion may be used as part of a broader hemostatic resuscitation strategy. However, current best practice increasingly favors individualized, goal-directed coagulation management guided by viscoelastic testing and early use of factor-specific concentrates rather than empiric high-volume plasma transfusion [82, 83]. Coz Yataco et al. recommended that routine prophylactic FFP transfusion should not be performed in non-trauma, non-neurocritically ill patients at increased risk of bleeding due to coagulopathy who are undergoing CVC or arterial line insertion, bedside thoracentesis or paracentesis, or gastrointestinal endoscopy. In contrast, they recommended FFP transfusion before the procedure for patients undergoing bedside lumbar puncture or flexible bronchoscopy without biopsy. However, the authors also highlighted that the certainty of the evidence is low [93]. In summary: (1) Current evidence does not support prophylactic FFP transfusion in non-bleeding patients. (2) FFP may be indicated in the setting of major bleeding or massive transfusion. (3) Individualized hemostatic strategy, ideally guided by viscoelastic testing and factor-specific concentrates is preferred. (4) Future research should perform stratified analyses based on relevant confounders. Conclusions PBM focuses on preserving circulating blood volume and supporting blood health. In this context, small-volume blood collection tubes and closed blood-conservation systems should be routinely used to reduce anemia resulting from diagnostic blood loss. The current evidence supports restrictive RBC transfusion strategies in most clinical settings, particularly in patients with gastrointestinal bleeding, whereas liberal strategies may be preferable in patients with type 1 MI or brain injury. Strategies to support erythropoiesis include IV iron supplementation and EPO therapy. IV iron supplementation can safely result in a clinically meaningful rise in Hb. EPO therapy may reduce RBC transfusion requirements and increase Hb in selected subgroups. However, its use should be selective, as the risk of thromboembolic complications remains uncertain. Similarly, current evidence supports a restrictive platelet transfusion strategy. Prophylactic transfusion should be limited to high-risk hematologic patients, while a therapeutic approach is appropriate for critically ill patients with severe thrombocytopenia. Finally, prophylactic FFP transfusion in non-bleeding patients is not supported by current evidence. FFP transfusion should be reserved for major bleeding or massive transfusion, with a preference for an individualized hemostatic approach guided by viscoelastic testing and factor-specific concentrates (Fig. 2).Fig. 2Schematic summary of Patient Blood Management in the ICU, showing core measures and corresponding recommendations Schematic summary of Patient Blood Management in the ICU, showing core measures and corresponding recommendations