Recommendations for implementing patient blood management—an updated modified Delphi consensus from a multidisciplinary expert panel in Hong Kong

Statement 1. In most surgical populations, there is no significant difference in clinical outcomes between restrictive and liberal transfusion strategies; however, restrictive strategies (i.e. lower Hb threshold and fewer transfusions of RBCs) aim to minimise unnecessary blood use and optimise patient safety by reducing transfusion risk. Statement 2. A restrictive transfusion threshold of Hb 7–8 g/dL is recommended for most surgical populations, including those undergoing non-cardiac surgery and those with critical conditions (e.g. patients admitted to the intensive care unit (ICU)). Statement 3. A transfusion threshold of Hb 8 g/dL is recommended for patients with stable cardiovascular disease. Consider a more liberal threshold (Hb 9–10 g/dL) for acute coronary syndrome (ACS) or anaemic heart failure. Statement 4. Transfusion thresholds should be individualised in patients receiving ECMO (VV-ECMO target Hb 7 g/dL (mainly for respiratory failure) VA-ECMO target Hb 8–9 g/dL (mainly for circulatory failure)), accounting for individual clinical status and context. Statement 1. In most surgical populations, there is no significant difference in clinical outcomes between restrictive and liberal transfusion strategies; however, restrictive strategies (i.e. lower Hb threshold and fewer transfusions of RBCs) aim to minimise unnecessary blood use and optimise patient safety by reducing transfusion risk. Statement 2. A restrictive transfusion threshold of Hb 7–8 g/dL is recommended for most surgical populations, including those undergoing non-cardiac surgery and those with critical conditions (e.g. patients admitted to the intensive care unit (ICU)). Statement 3. A transfusion threshold of Hb 8 g/dL is recommended for patients with stable cardiovascular disease. Consider a more liberal threshold (Hb 9–10 g/dL) for acute coronary syndrome (ACS) or anaemic heart failure. Statement 4. Transfusion thresholds should be individualised in patients receiving ECMO (VV-ECMO target Hb 7 g/dL (mainly for respiratory failure) VA-ECMO target Hb 8–9 g/dL (mainly for circulatory failure)), accounting for individual clinical status and context. A transfusion trigger is the clinical threshold, primarily a low Hb level, that indicates the need for a blood transfusion to ensure sufficient oxygen delivery to organs and tissues. Despite the lack of a universally agreed definition, restrictive transfusion strategies typically allow for a higher tolerance of anaemia, initiating transfusions only when Hb levels drop to ~7 g/dL; in contrast, liberal transfusion strategies tend to commence transfusions at a higher threshold, e.g. Hb ~10 g/dL. When clinical outcomes are similar, a restrictive approach is preferred over a liberal approach to minimise unnecessary blood transfusions and associated adverse effects. A meta-analysis included 48 randomised trials with 21,433 patients across varying clinical conditions (e.g. orthopaedic or cardiovascular surgery; critical care; acute blood loss (including gastrointestinal bleeding); ACS; cancers) [ 112 ]. Compared with liberal transfusion strategies (Hb threshold, 9.0–10.0 g/dL), restrictive transfusion strategies (Hb threshold, 7.0–8.0 g/dL) reduced the need for transfusions by 41% (RR, 0.59; 95% CI, 0.53–0.66; high-quality evidence), with no substantial impact on the risk of 30-day mortality (RR, 0.99; 95% CI, 0.86–1.15; moderate-quality evidence) or other outcomes assessed, including cardiac events (low-quality evidence), myocardial infarction, stroke, and thromboembolism (all high-quality evidence) [ 112 ]. Therefore, a restrictive transfusion threshold of Hb 7–8 g/dL should be considered for most surgical populations who are haemodynamically stable [ 106 , 113 , 114 ]. Although restrictive transfusion strategies appear generally favourable, their safety in populations with acute brain injury [ 115 ] or cardiovascular problems, or undergoing cardiovascular surgeries remains less certain [ 2 , 106 , 112 , 116 , 117 ]. A systematic review and frequentist-Bayesian meta-analysis included four randomised trials that compared liberal (Hb ≥ 9 g/dL) versus restrictive (Hb ≥ 7 g/dL) transfusion in adults with acute brain injury (i.e. traumatic brain injury, subarachnoid haemorrhage, and intracranial haemorrhage) and a Glasgow Coma Scale score ≤ 13 [ 115 ]. Pooled data from trials with a low risk of bias (with no heterogeneity, I 2 = 0%) showed that the liberal strategy was associated with improved neurological outcomes (RR, 0.74; 95% CI, 0.63–0.87) [ 115 ]. Patients with coronary artery disease often have reduced physiological reserve that may necessitate a higher transfusion threshold. The randomised MINT trial demonstrated that, among patients with acute myocardial infarction and anaemia, restrictive transfusion (Hb 7–8 g/dL) appeared to be associated with a worse composite outcome of recurrent myocardial infarction or death at 30 days (RR, 1.15; 95% CI, 0.99–1.34; P = 0.07) compared with liberal transfusion (Hb 10 g/dL) [ 116 ]. A post hoc analysis of MINT assessed restrictive and liberal transfusion strategies among patients with and without baseline heart failure [ 117 ]. The rate ratio of death or heart failure at 30 days between restrictive and liberal transfusion was numerically higher in those with baseline heart failure than in those without (RR, 1.20 [95% CI, 0.99–1.45] vs. 0.94 [95% CI, 0.70–1.26]; P interaction = 0.18) [ 117 ]. Patients with ECMO require individualised transfusion thresholds. A retrospective study in Hong Kong showed that, in a cohort of patients with ECMO (VV, 30%; VA, 50%), using transfusion thresholds of Hb ≤ 8.5 g/dL and > 8.5 g/dL yielded no significant differences in ICU mortality (adjusted odds ratio (OR), 0.86; 95% CI, 0.56–1.30; P = 0.47), hospital mortality (adjusted OR, 0.79; 95% CI, 0.52–1.21; P = 0.28), or 90-day mortality (adjusted OR, 0.84; 95% CI, 0.55–1.28; P = 0.42) [ 118 ]. In the subgroup of patients with ECMO-VV, Hb ≤ 8.5 g/dL (vs. > 8.5 g/dL) was associated with a reduced risk of ICU mortality (adjusted OR, 0.36; 95% CI, 0.17–0.73; P = 0.005) [ 118 ]. The international prospective observational PROTECMO study demonstrated that, among patients with ECMO-VV (mean pretransfusion Hb, 8.1 g/dL), Hb < 7 g/dL was associated with a higher risk of death in the ICU (HR, 2.99; 95% CI, 1.95–4.60) compared with higher Hb levels, while packed RBC transfusion was associated with a lower risk of mortality (HR, 0.15; 95% CI, 0.03–0.74) only when administered with Hb 10 g/dL) [ 119 ]. Statement 5. When indicated, single-unit transfusions with reassessment should replace multiunit orders. Statement 5. When indicated, single-unit transfusions with reassessment should replace multiunit orders. In patients who are haemodynamically stable and not actively bleeding, single-unit transfusions, followed by clinical reassessments before the administration of additional units, should be considered, because this strategy is often adequate to restore oxygen delivery and treat anaemia, while offering the potential benefits of reducing unnecessary transfusions, risks of transfusion-related complications, and healthcare costs [ 113 , 120 , 121 ]. Statement 6. Transfusion strategies should be adjusted based on patient-specific factors, including laboratory data, clinical context, symptoms, and signs (Table 2 ), in addition to Hb levels. Statement 6. Transfusion strategies should be adjusted based on patient-specific factors, including laboratory data, clinical context, symptoms, and signs (Table 2 ), in addition to Hb levels. Notably, restrictive and single-unit transfusions may not be suitable in patients who are haemodynamically unstable, hypovolaemic, actively bleeding, or severely anaemic. Rather than a routine practice, these strategies should be implemented in specific patient groups after consideration of individual patient factors and conditions. Statement 7. Routine use of physiologic transfusion triggers (e.g. systemic oxygen delivery, ST segment changes on electrocardiogram, mixed venous oxygen saturation, lactate levels, near-infrared spectroscopy) is not recommended because of a lack of high-level evidence. Statement 7. Routine use of physiologic transfusion triggers (e.g. systemic oxygen delivery, ST segment changes on electrocardiogram, mixed venous oxygen saturation, lactate levels, near-infrared spectroscopy) is not recommended because of a lack of high-level evidence. The Hb level remains the standard transfusion trigger that facilitates continuous monitoring of anaemic status and thus timely transfusions. Alternatively, there are a range of alternative blood transfusion triggers that are mainly based on the physiological parameters of oxygen delivery/consumption ratio, attempting to address the insufficiency of the isolated Hb level in reflecting oxygen delivery [ 122 ]. Although physiological triggers are theoretically optimal to determine the need for transfusions, their use in routine clinical practice should be further verified in large-scale randomised trials [ 123 ]. Statement 8. Suggested key components of a hospital-wide transfusion protocol include the following: Clear Hb thresholds tailored to patient subgroups Mandatory Hb check before transfusions when clinical context allows, and with application of POC coagulation testing as necessary Assessment of coagulation, biochemical and metabolic parameters Mandatory assessment of patient conditions before transfusion Single-unit transfusion orders as far as possible Checklists to ensure reassessments after single-unit transfusions Monitoring for adverse effects (e.g. ischaemic complications, haemodynamic instability, organ dysfunction) Clear Hb thresholds tailored to patient subgroups Mandatory Hb check before transfusions when clinical context allows, and with application of POC coagulation testing as necessary Assessment of coagulation, biochemical and metabolic parameters Mandatory assessment of patient conditions before transfusion Single-unit transfusion orders as far as possible Checklists to ensure reassessments after single-unit transfusions Monitoring for adverse effects (e.g. ischaemic complications, haemodynamic instability, organ dysfunction) The panel highlighted the importance of developing a hospital-wide transfusion protocol with the above items. The principle is to facilitate efficient and safe transfusions via mandatory assessments of Hb levels and clinical conditions (e.g. active bleeding, ischaemic symptoms), implementation of the single-unit policy (investigation of abnormalities, e.g. occult bleeding and hypoxia, may be considered when additional units are required), and monitoring for adverse effects (e.g. troponin rise, changes in electrocardiogram, tachycardia, hypotension, acute kidney injury, altered mental status). Statement 9. Suggested measures to facilitate the implementation of transfusion strategies include the following: Staff training on transfusion indications, restrictive thresholds, and the single-unit policy Electronic reminders to confirm the need for transfusion in patients with Hb 7–8 g/dL without symptoms Transfusion audits to monitor compliance, transfusion reactions, and clinical outcomes Feedback and refreshment training for clinicians based on audit data Refinement of protocols based on audit data Staff training on transfusion indications, restrictive thresholds, and the single-unit policy Electronic reminders to confirm the need for transfusion in patients with Hb 7–8 g/dL without symptoms Transfusion audits to monitor compliance, transfusion reactions, and clinical outcomes Feedback and refreshment training for clinicians based on audit data Refinement of protocols based on audit data Staff training and decision-support tools, e.g. electronic alerts for Hb thresholds and checklists to ensure reassessment following a single-unit transfusion, are crucial to facilitate the implementation of a transfusion protocol. Audits should be conducted to monitor protocol compliance and a range of metrics, including the percentage of single-unit transfusions, incidences of transfusion-related reactions, and clinical outcomes (e.g. mortality rates, length of hospital stay, complications). Protocols can be refined based on audit results. Statement 10. PBM programmes consistently deliver healthcare savings via improvements in patient outcomes, including shortening hospital stays, reducing postoperative complications, and minimising unnecessary blood transfusions. Statement 11. PBM programmes enhance resource sustainability and ethical use of blood products. Statement 10. PBM programmes consistently deliver healthcare savings via improvements in patient outcomes, including shortening hospital stays, reducing postoperative complications, and minimising unnecessary blood transfusions. Statement 11. PBM programmes enhance resource sustainability and ethical use of blood products. PBM programmes that include preoperative screening for anaemia and suboptimal iron stores, perioperative bleeding reduction measures, and restrictive transfusion strategies are generally considered to be cost-effective (reducing unnecessary transfusions and related complications) and to improve patient outcomes [ 124 – 127 ]. A retrospective study showed that, over 6 years, a jurisdiction-wide PBM programme initiated across four major hospitals in Western Australia significantly reduced the hospital mortality rate by 28%, average length of hospital stays by 15%, incidence of hospital-acquired infections by 21%, incidence of heart attack or stroke by 31%, and units of transfusions per admission by 41%, translating to substantial cost savings [ 127 ]. In addition to being cost-effective, PBM programmes support the ethical use of blood products by minimising the need for transfusions and improving patient outcomes, considering their limited availability, reliance on voluntary donation, and inherent worth [ 125 ]. Statement 12. Disease- and specialty-specific protocols are needed to consistently implement PBM programmes. Statement 13. Key challenges and barriers towards the implementation of PBM include hospital culture, inadequate staff awareness, poor interdisciplinary communication or collaboration, absence of electronic monitoring systems, and resource limitations (staff, time, and finances). Statement 12. Disease- and specialty-specific protocols are needed to consistently implement PBM programmes. Statement 13. Key challenges and barriers towards the implementation of PBM include hospital culture, inadequate staff awareness, poor interdisciplinary communication or collaboration, absence of electronic monitoring systems, and resource limitations (staff, time, and finances). Consistent implementation of a hospital-based PBM programme requires dedicated protocols that address perioperative bleeding reduction interventions, Hb thresholds for transfusions, and the application of single-unit transfusions for different surgical populations. Notably, more local research is warranted to justify enhancements in investments and resources to help resolve the challenges and barriers towards implementation of PBM programmes [ 1 ].

The president and vice-president of the HKSCBM served as co-chairs responsible for convening a consensus panel of 11 Society members who were invited on the basis of diverse special interests and extensive clinical experience (> 10 years) in medical or surgical aspects of PBM. These 13 panellists comprised six anaesthesiologists (including both co-chairs), three specialists in obstetrics and gynaecology, two specialists in internal medicine, haematology and haematological oncology, and transfusion medicine, one specialist in orthopaedics and traumatology, and one anaesthetic nurse, most of whom practised in the public healthcare sector. Compared with the previous consensus panel of 11 members predominantly from anaesthesiology, an expansion of the number and specialty of panellists was expected to address recent developments across diverse PBM practices. The panel’s composition reflected the practical application of PBM practices in Hong Kong’s clinical environment, where most patients receive care in public hospitals, with anaesthesiologists typically assuming a leadership role while other specialists participate to varying extents. Prior to consensus meetings, the co-chairs formulated discussion questions based on three pillars of PBM: (1) optimising patients’ red blood cell (RBC) mass and improving anaemia management; (2) minimising perioperative blood loss; and (3) rationalising the use of blood and blood components. With the aim of providing a relatively rapid and broad overview of the topic to contextualise panel discussions, the consensus panel commissioned a medical communications agency to assist with a narrative rather than systematic literature search. The PubMed database was searched for meta-analyses, randomised controlled trials, observational studies, reviews, and clinical guidelines that addressed PBM-related areas for the publication period between June 2020 and June 2025, using the following keywords: ‘anaemia’; ‘fibrinogen’; ‘haemodynamic monitoring’; ‘haemostatic agent’; ‘intravenous iron’; ‘iron deficiency’; ‘nurse-led clinic’; ‘oral iron’; ‘postpartum haemorrhage’; ‘restrictive transfusion’; ‘single-unit transfusion’; ‘thromboelastography’; ‘tranexamic acid’; ‘transfusion threshold’; and ‘transfusion trigger’. The titles and abstracts of the search results were screened to identify relevant papers for discussion. Using the modified Delphi method [ 6 ], the panel formulated consensus statements through a series of discussions and votes. Based on personal interest and expertise, three subgroups were formed to address the discussion areas. In three online meetings held on 8th September, 22nd September, and 15th October 2025, the panellists took a turn to present the relevant literature and their clinical experiences regarding their designated discussion questions, followed by panel discussions and comments. These meetings familiarised all panellists with all discussion areas addressed by each of them and facilitated the subsequent statement drafting and voting processes. Draft consensus statements, prepared by the medical communications agency based on the meeting proceedings, were circulated among panellists via a messaging platform. The panel suggested adding, removing, merging, and revising statements where appropriate. All panellists participated in the review, revision, and voting processes for the final consensus statements. At a physical meeting held on 21 st November 2025, an updated set of statements was discussed, revised, and finalised, followed by anonymous panel voting (including co-chairs) via an electronic system. A consensus statement was accepted only if ≥ 80% of the panel chose ‘accept completely’ or ‘accept with some reservation’, based on a 5-point Likert scale; the other options were ‘accept with major reservation’, ‘reject with reservation’, and ‘reject completely’. On 9th December 2025, the panel reviewed the voting results in an online meeting. Among a total of 118 statements (including sub-statements), 114 and four were accepted and rejected, respectively. The wording of the accepted statements remained unchanged. For the remaining four, one was rejected because its suggestion (an optimal preoperative haemoglobin (Hb) level ≥ 13.0 g/dL in both males and females) was overridden by another (≥ 13.0 g/dL in males and ≥ 12.0 g/dL in females). The other three rejected statements were modified and eventually accepted through anonymous online voting. Additional file 2 reports the final voting records.

The panel accepted 117 consensus statements, of which 29 were chosen with a vote of ‘accept completely’ by < 80% of panellists (Table  1 ). The rationale for each statement is discussed below. The updated version includes additions to a number of important areas, as follows: identification of groups at high risk for iron deficiency anaemia (IDA); preoperative screening and management of iron deficiency and/or anaemia among general and special surgical populations; recognition of bleeding-prone surgeries and conditions; recent measures to minimise intraoperative blood loss, including point-of-care (POC) coagulation testing, tranexamic acid (TXA), fibrinogen concentrate, reversal agents for anticoagulation, advanced haemodynamic monitoring, and uterotonics for postpartum haemorrhage (PPH); suggested Hb thresholds for transfusions among different surgical populations; and implementation of transfusion protocols. Although the literature search was cut off in June 2025, the panel noted that, to the best of their knowledge, no studies published between July 2025 and April 2026 substantially impacted the established statements. Table 1 All accepted consensus statements % of panellists ( N = 13) # Accepted statements Accept completely Accept with some reservation Part 1. Optimising patients’ red blood cell mass and improving anaemia management 1.1. Groups at high risk for iron deficiency and anaemia 1 Groups at high risk for iron deficiency and anaemia include (Table  2 ):  1a Women at reproductive age 100 0  1b Antenatal or postnatal women 100 0  1c Patients with cancer 100 0  1d Patients with heart disease 92 8  1e Patients with bowel disorders (e.g. peptic ulcer, inflammatory bowel disease) 100 0  1f Patients with renal disease 100 0  1g Elderly populations 92 8  1h Patients on extracorporeal membrane oxygenation (ECMO) 92 8 2 Other conditions that are prone to adverse effects of iron deficiency anaemia (IDA):  2a Acute brain injury 46* 54  2b Septic shock 54* 46 1.2. Preoperative screening for IDA 3 For surgical populations, the optimal preoperative haemoglobin (Hb) levels should be ≥ 13.0 g/dL in males and ≥ 12.0 g/dL in females 62* 31 4 All surgical populations should be routinely screened for anaemia, focusing on absolute or functional iron deficiency 85 15 5 Preoperative screening for IDA facilitates diagnosis and earlier administration of oral or intravenous iron therapy, thereby reducing hospitalisations for blood transfusions and contributing to higher rates of same-day surgery 85 15 6 It may be worthwhile to adopt preoperative protocol-driven evaluation for IDA 100 0 7 Hospital- or department-specific algorithms should be established to govern preoperative screening and management of IDA, including the roles of nurses, anaesthesiologists, physicians, haematopathologists, and surgeons 100 0 8 A nurse-led clinic can be implemented to facilitate preoperative screening for IDA, with the roles including:  8a Interpretation of protocol-driven anaemia evaluation 92 8  8b Patient education on iron deficiency treatments, including effectiveness and tolerability of oral and intravenous iron therapies 100 0  8c Administration of intravenous iron therapy when necessary, using dosing tools 69* 31  8d Alerting attending surgeons and anaesthesiologists to unexplained or severe anaemia cases, and to consider referral to physicians for further work-up 100 0 9 Clinical management systems should be enhanced to track patient outcomes and automatically notify attending surgeons and anaesthesiologists if rechecked Hb remains below the target within 7–10 days before surgery 77* 23 10 Further local research should be conducted to assess the prevalence of anaemia and their causes among surgical populations in Hong Kong 100 0 1.3. Preoperative measures to optimise Hb levels in the general population 1.3.1. Patients undergoing elective surgery 11 In patients with suboptimal preoperative Hb levels, delay of elective surgery should be considered, and an iron supplement should be administered to correct the deficiency while investigating the underlying causes 69* 23 12 Transfusions should be avoided in patients undergoing elective surgery when their IDA is clinically stable and correctable 77* 23 1.3.2. Patients undergoing urgent surgery 13 In patients who require more urgent surgery (e.g. within 2–4 weeks), preoperative IDA should be treated with intravenous iron therapy 77* 23 14 To treat preoperative anaemia in patients who are clinically unstable or require urgent surgery, transfusions (preferably with a restrictive Hb threshold and a single-unit approach [refer to Part 3]) may be considered 85 8 1.4. Preoperative measures to optimise Hb levels in special populations 1.4.1. Gynaecological patients 15 In patients with heavy menstrual bleeding and severe anaemia (i.e. Hb < 7 g/dL) due to fibroids:  15a Preoperative use of gonadotropin-releasing hormone (GnRH) agonists for 3–4 months can be considered to reduce intraoperative blood loss 54* 46  15b Intravenous iron therapy ± a single-unit transfusion can be considered to optimise preoperative Hb levels, provided the patient is haemodynamically stable 85 15  15c Early surgery should be considered in those who have already planned for the procedure 85 15 1.4.2. Obstetric patients 16 In antenatal patients, screening of serum ferritin levels can be considered for early detection of IDA 100 0 17 The following treatment measures can be considered in antenatal patients:  17a Iron-rich diet 85 15  17b Oral iron supplementation 92 8  17c Intravenous iron therapy in the 2nd or 3rd trimesters (especially when oral iron therapy is contraindicated, intolerable, or ineffective) 100 0 18 Regarding the administration of intravenous iron therapy, the need for monitoring for any potential adverse reactions under an institutional protocol should be discussed with parturients in advance 85 15 1.4.3. Renal patients 19 Higher thresholds (i.e. serum ferritin ≤ 800 ng/mL and transferrin saturation < 30%) would increase the sensitivity of identifying IDA in patients with chronic kidney disease 77* 23 20 To treat anaemia in patients with chronic kidney disease:  20a Intravenous iron therapy is preferred over oral iron therapy, especially when response to oral iron therapy is suboptimal 38* 62  20b Erythropoiesis-stimulating agents can be considered in selected cases 100 0  20c A conservative Hb target should be maintained 85 15  20d Coordination with nephrologists should be considered when necessary 100 0 1.4.4. Patients with bowel disorders 21 Screening for haematinics deficiency, including iron profile, vitamin B12 and folate, is recommended 85 8 22 Intravenous iron therapy, instead of oral iron therapy, is recommended in patients with iron deficiency due to malabsorption by the gastrointestinal tract 85 15 1.4.5. Elderly patients 23 Routine screening for medications that pose a high risk for anaemia is recommended 85 8 24 Intravenous iron therapy may be preferred over oral iron therapy in patients with polypharmacy 69* 31 1.5. Patient empowerment 25 Public education in the primary and community healthcare settings should be enhanced to raise awareness of IDA, especially among women at reproductive age 100 0 26 Patients should always be counselled regarding risks associated with anaemia and blood transfusions 100 0 27 Patient preferences, acceptance, or rejection regarding blood components and/or blood conservation modalities should be discussed preoperatively 100 0 28 Related consent forms and advanced directives should be obtained and documented preoperatively to ensure that acceptable options for optimal care are provided 92 8 Part 2. Minimising perioperative blood loss 2.1. Surgeries and conditions associated with substantial blood loss 1 Perioperative bleeding is common in diverse surgical fields, including:  1a Trauma surgery 100 0  1b Orthopaedic surgery 100 0  1c Neurosurgery 62* 38  1d Visceral and transplant surgery 92 8  1e Cardiac and vascular surgery 100 0  1f Obstetric and gynaecological surgery 100 0 2 In the preoperative phase, interventions should be initiated to identify patients who may be at higher risk for bleeding, including those with:  2a Underlying coagulation abnormalities (inherited or acquired) 100 0  2b Antithrombotic medications 100 0 2.2. Measures to minimise intraoperative blood loss in the general population 2.2.1. Surgical techniques 3 Judicious use of minimally invasive surgery, electrocautery, tourniquets, topical haemostatic agents (including mechanical and active biologic agents), and intraoperative blood salvage should be considered to reduce blood loss 85 15 2.2.2. Patient positioning 4 Correct patient positioning is a simple and effective intervention to minimise intraoperative blood loss 92 8 5 The general principles of patient positioning include elevation of surgical sites and slow transition of positions 92 8 2.2.3. Normothermia 6 Maintaining perioperative normothermia is crucial to reduce blood loss and the need for blood transfusions 100 0 2.2.4. Point-of-care (POC) coagulation testing 7 POC testing of blood coagulation using a viscoelastic haemostatic assay (e.g. rotational thromboelastometry or thromboelastography) or ultrasound-induced resonance helps reducing the requirement for blood product transfusions by guiding the haemostatic therapy 92 8 8 POC testing of blood coagulation is recommended in cases of suspected coagulopathy or surgical settings where massive haemorrhage is anticipated (e.g. trauma, cardiac, liver, obstetric and gynaecological surgery) 100 0 2.2.5. Haemostatic agents 9 Tranexamic acid can be administered perioperatively to reduce the risk of major bleeding in non-cardiac surgeries 100 0 10 Fibrinogen concentrate helps minimising intraoperative blood loss in certain surgical populations, including cardiac surgery, massive obstetric haemorrhage, or polytrauma with severe bleeding 100 0 11 In patients with expected massive haemorrhage awaiting viscoelastic or laboratory tests, administering 2-g fibrinogen concentrate based on clinical criteria at admission—such as low systolic blood pressure, metabolic acidosis, or low Hb levels—helps to provide initial coagulation support and correct hypofibrinogenemia 77* 23 2.2.6. Reversal agents for anticoagulation 12 Prothrombin complex concentrate (PCC) and vitamin K1 are indicated for the urgent reversal of anticoagulation in patients with major acute bleeding, such as intracerebral haemorrhage, or those requiring emergency surgery who are taking warfarin 85 15 13 Regarding the use of specific reversal agents:  13a Idarucizumab should be considered in patients taking dabigatran if urgent reversal of anticoagulation is indicated; for example, in major acute bleeding or before emergency surgery 85 15  13b Andexanet alfa, the reversal agent for factor Xa inhibitors (e.g. rivaroxaban, apixaban, and edoxaban), is not readily available in Hong Kong. If rapid reversal of an oral factor Xa inhibitor is indicated, PCC could be considered 85 15 14 PCC may serve as a treatment option for acute massive haemorrhage (not warfarin-induced) in patients undergoing surgery 38* 54 15 In acute massive haemorrhage, the advantages of PCC over plasma may include faster onset of action, off-the-shelf availability, no thawing requirement, and reduced risk of fluid overload 92 8 16 The limitations of PCC in the management of acute massive haemorrhage include the inability to replenish all clotting factors and the thrombotic risk associated with an overdose. Its use is preferably guided by POC coagulation testing 85 15 2.2.7. Advanced haemodynamic monitoring 17 In surgical populations at high risk for massive bleeding, advanced haemodynamic monitoring providing continuous data (e.g. cardiac output, fluid responsiveness) can be considered to facilitate targeted and timely interventions, thereby minimising intraoperative blood loss and the need for blood transfusions, and improving patient outcomes 92 8 2.3. Measures to minimise intraoperative blood loss in special populations 2.3.1. Orthopaedic patients 18 During prone spine surgery, the abdomen should be well positioned to avoid compressing the inferior vena cava 100 0 19 Closed suction drains are not recommended in hip and knee arthroplasty 69* 31 2.3.2. Obstetric patients 20 Prevention of postpartum haemorrhage (PPH; defined as a blood loss of ≥ 500 mL within 24 h after vaginal or caesarean delivery) is important in obstetric populations 100 0 21 To reduce the risk of PPH, the umbilical cord can be managed as follows:  21a For vaginal delivery, controlled cord traction can be offered routinely during the 3rd stage of labour, provided that the birth attendant has the necessary skills 100 0  21b For caesarean delivery, controlled cord traction is recommended for the removal of the placenta 100 0  21c Delayed (instead of early) cord clamping is recommended for all births unless the neonate is asphyxiated and needs to be moved immediately for resuscitation 100 0 22 The use of oxytocin or carbetocin to prevent PPH during the 3rd stage of labour is recommended for all births:  22a Carbetocin is recommended in women undergoing caesarean delivery, and for those undergoing vaginal delivery who are at increased risk for PPH 100 0  22b Oxytocin is recommended in women undergoing vaginal delivery who do not have risk factors for PPH 85 15 23 The use of prophylactic tranexamic acid is recommended for high-risk patients with PPH 100 0 24 Second-line measures, including uterine compression sutures, balloon tamponade, and uterine artery embolisation, should be implemented early and with a lower threshold to help preventing PPH 100 0 25 Recommended treatments for PPH include:  25a Uterine massage 100 0  25b Intravenous oxytocin alone as the first-line uterotonic treatment 69* 31  25c Second-line uterotonics (e.g. syntometrine, carboprost, misoprostol) if bleeding does not respond to oxytocin 92 8  25d Intravenous tranexamic acid administered as soon as possible (within 3 h) after bleeding onset, in addition to standard care 100 0  25e Uterine balloon tamponade as a non-surgical treatment approach for PPH due to uterine atony if uterotonics are ineffective or unavailable 100 0  25f Surgical intervention (e.g. compression suture, uterine and internal iliac artery ligation, hysterectomy) if PPH does not respond to uterotonics or other conservative treatments 100 0  25g Single-unit transfusions may be considered 46* 38  25h Iron repletion may be considered 77* 8 26 Placenta accreta spectrum, often related to prior caesarean sections, is a major risk factor for massive PPH 100 0 27 In patients with placenta accreta spectrum following caesarean sections, conservative management involving retention of the placenta in situ may be considered to minimise total blood loss, although the possibility of requiring hysterectomy cannot be ruled out 92 8 2.3.3. Patients on antithrombotic therapy 28 In patients on antithrombotic therapy (e.g. warfarin, direct oral anticoagulants, antiplatelets) who plan to undergo elective surgery, the decision to stop antithrombotic therapy should consider the following factors:  28a Thrombotic risk associated with anticoagulation interruption in the perioperative period 100  28b Surgery/procedure-related bleeding risk 92 8 29 Antithrombotic therapy can be continued in surgeries/procedures with minimal bleeding risk, whereas heparin bridging may be needed in patients taking warfarin and at high risk for thromboembolism during surgeries/procedures with bleeding risk (e.g. recent venous thromboembolism and mechanical heart valves) 69* 23 30 Considering the wide variation in surgical procedures and patient comorbidities, thrombotic and bleeding risk assessment should be individualised and managed through a multidisciplinary approach 100 0 Part 3. Rationalising the use of blood and blood components 3.1. Transfusion triggers 1 In most surgical populations, there is no significant difference in clinical outcomes between restrictive and liberal transfusion strategies; however, restrictive strategies (i.e. lower Hb threshold and fewer transfusions of red blood cells) aim to minimise unnecessary blood use and optimise patient safety by reducing transfusion risk 77* 23 2 A restrictive transfusion threshold of Hb 7–8 g/dL is recommended for most surgical populations, including those undergoing non-cardiac surgery and those with critical conditions (e.g. patients admitted to the intensive care unit) 62* 38 3 A transfusion threshold of Hb 8 g/dL is recommended for patients with stable cardiovascular disease. Consider a more liberal threshold (Hb 9–10 g/dL) for acute coronary syndrome or anaemic heart failure 92 0 4 Transfusion thresholds should be individualised in patients receiving ECMO (venovenous-ECMO target Hb 7 g/dL [mainly for respiratory failure]; venoarterial-ECMO target Hb 8–9 g/dL [mainly for circulatory failure]), accounting for individual clinical status and context 69* 31 5 When indicated, single-unit transfusions with reassessment should replace multiunit orders 100 0 6 Transfusion strategies should be adjusted based on patient-specific factors, including laboratory data, clinical context, symptoms, and signs (Table  2 ), in addition to Hb levels 92 8 7 Routine use of physiologic transfusion triggers (e.g. systemic oxygen delivery, ST segment changes on electrocardiogram, mixed venous oxygen saturation, lactate levels, near-infrared spectroscopy) is not recommended because of a lack of high-level evidence 54* 46 3.2. Implementation of transfusion protocols 8 Suggested key components of a hospital-wide transfusion protocol include:  8a Clear Hb thresholds tailored to patient subgroups 85 15  8b Mandatory Hb check before transfusions when clinical context allows, and with application of POC coagulation testing as necessary 62* 38  8c Assessment of coagulation, biochemical and metabolic parameters 69* 23  8d Mandatory assessment of patient conditions before transfusion 100 0  8e Single-unit transfusion orders as far as possible 85 15  8f Checklists to ensure reassessments after single-unit transfusions 85 15  8g Monitoring for adverse effects (e.g. ischaemic complications, haemodynamic instability, organ dysfunction) 92 8 9 Suggested measures to facilitate the implementation of transfusion strategies include:  9a Staff training on transfusion indications, restrictive thresholds and the single-unit policy 92 8  9b Electronic reminders to confirm the need for transfusion in patients with Hb 7–8 g/dL without symptoms 54* 38  9c Transfusion audits to monitor compliance, transfusion reactions, and clinical outcomes 92 8  9d Feedback and refreshment training for clinicians based on audit data 100 0  9e Refinement of protocols based on audit data 85 15 3.3. Advantages and challenges of patient blood management (PBM) programmes 10 PBM programmes consistently deliver healthcare savings via improvements in patient outcomes, including shortening hospital stays, reducing postoperative complications, and minimising unnecessary blood transfusions 92 8 11 PBM programmes enhance resource sustainability and ethical use of blood products 92 8 12 Disease-specific and specialty-specific protocols are needed to consistently implement PBM programmes 77* 23 13 Key challenges and barriers towards the implementation of PBM include hospital culture, inadequate staff awareness, poor interdisciplinary communication or collaboration, absence of electronic monitoring systems, and resource limitations (staff, time, and finances) 77* 23 * Statements with < 80% of panellists choosing ‘accept completely’ All accepted consensus statements * Statements with < 80% of panellists choosing ‘accept completely’

The HKSCBM updated its consensus statements to facilitate the implementation of PBM across three pillars: optimisation of RBC mass and anaemia management, minimisation of perioperative blood loss, and rational use of blood and blood components. The consensus panel’s multidisciplinary composition addressed measures to enhance PBM practices for various surgical populations, including obstetric, gynaecologic, orthopaedic, renal, gastrointestinal, cardiovascular, oncological, emergency, traumatic, and elderly patients. These statements are intended to serve as a practical framework for clinicians and healthcare providers involved in PBM. While recognising the importance of patient empowerment in PBM, the panel did not include patient, caregiver, or public representatives in the present consensus process. This consensus primarily focused on establishing clinician-directed recommendations and promoting implementation among healthcare professionals. Further panel initiatives will aim to enhance public awareness through patient education, engagement with patient advocacy groups, and incorporation of these perspectives into subsequent consensus development. The HKSCBM will continue to advance research and cooperate with relevant stakeholders, including hospital administrators, medical and surgical colleagues, and patients, to promote optimal PBM practices across divisions, departments, and hospitals. Notably, PBM practices for other highly specialised areas, such as cardiothoracic surgery and paediatrics, will be addressed in future updates by expanding the consensus panel to include related experts.

Patient blood management (PBM) represents a comprehensive approach for optimising patient outcomes and safety through a range of screening, medical, and surgical measures to manage iron deficiency with or without anaemia, coagulopathy, and at the same time preserve the usage of blood components [ 1 , 2 ]. Endorsed by the World Health Organisation (WHO), the notion of PBM has gained substantial traction globally through growing awareness of the risk of transfusion-related adverse effects, including infection, prolonged hospitalisation, and morbimortality, alongside considerable healthcare costs [ 3 ]. In 2020, the Hong Kong Society of Clinical Blood Management (HKSCBM) established a consensus statement to facilitate the implementation of PBM in the locality [ 4 ]. With the continuous development of PBM initiatives, the HKSCBM issued an updated set of consensus statements on the basis of recent evidence. Because the evidence base regarding PBM remains limited, heterogeneous, and in some areas conflicting, it was considered insufficient to support a formal systematic review or evidence-based guideline development process. Therefore, a consensus approach was deemed the most appropriate methodology to address the topic, integrating the available literature with expert insights, clinical experience, and practical considerations relevant to current practice. Although this consensus protocol was not prospectively registered, its reporting aligned with the ACCORD guidelines (Additional file 1) [ 5 ].

Additional file 1. ACCORD checklist. Description: Completion of the checklist demonstrated that this consensus-based research was reported as per the ACCORD guidelines.  Additional file 2. Full voting records for all statements. Description: The file tabulates the detailed voting results for each accepted and rejected statement. Additional file 1. ACCORD checklist. Description: Completion of the checklist demonstrated that this consensus-based research was reported as per the ACCORD guidelines. Additional file 2. Full voting records for all statements. Description: The file tabulates the detailed voting results for each accepted and rejected statement.