Transfusion medicine underpins multidisciplinary clinical care in perioperative care, critical care, hematology, and trauma emergencies. Artificial intelligence (AI) has shown broad potential in this field. This review examines the current domestic and international research, covering technological evolution, data infrastructure, clinical applications, management implementations, and ethical and regulatory considerations. First, it outlines the evolution of AI from statistical analysis to deep learning and describes key technologies, including the standardization of transfusion data, feature selection, and knowledge graph construction. Second, it summarizes AI applications in pre-transfusion blood compatibility screening, intraoperative risk monitoring, and post-transfusion prognosis prediction, as well as its value in blood supply-demand coordination, adverse reaction early warning, and clinical decision support systems. Finally, it discusses ethical and legal issues such as data privacy, the "black box" nature of algorithms, and regulatory adaptation; analyzes the limitations in the practical implementation of existing models; and provides a theoretical reference for the translation of AI into transfusion departments and blood collection and supply institutions in China. This review is expected to inform the clinical translation of AI in transfusion medicine and to promote safer, more efficient, and more precise blood utilization in China.
Artificial intelligence (AI) is profoundly transforming the concepts and paradigms of medical education. As a discipline characterized by strong practicality, rapid knowledge updates, and high clinical risks, transfusion medicine faces challenges in graduate education such as monotonous teaching modes, uneven distribution of practical resources, and insufficient personalized training. Focusing on AI-enabled graduate education in transfusion medicine, this review systematically examines the theoretical foundations, current applications, and implementation outcomes of relevant technologies from four dimensions: the construction of personalized adaptive learning systems, virtual reality/augmented reality-based practical skills training, cultivation of research competence and clinical decision-making thinking, and intelligent management and quality assessment of educational processes. It further analyzes practical challenges including technical costs, data security, algorithmic ethics, and limited digital literacy among faculty, and discusses future directions of collaborative educational paradigm, aiming to provide references for building an intelligent and precise system for cultivating high-level talents in transfusion medicine.
Objective To investigate the feasibility and biosafety of mesoporous silica (MS) loaded with platelet-rich plasma (PRP), as well as its modulatory effect on the wound-healing capacity of PRP and the underlying molecular mechanisms. Methods Venous whole blood was harvested from rats. After anticoagulation, PRP was extracted via a two-step centrifugation protocol. The microstructure of mesoporous silica (MS) was characterized via scanning electron microscopy. PRP was loaded onto MS by freeze-drying method to fabricate PRP@MS composites. Subsequent in vitro experiments were performed to evaluate sustained-release behaviour and biocompatibility. Thirty SD rats with full-thickness dorsal skin wounds were randomly assigned into three groups (n=10 per group) using a random number table ∶ routine dressing group (treated with normal saline), PRP group (treated with PRP), and PRP@MS group (treated with PRP@MS). Dressings were performed every four days. Wound closure rates were quantified by pixel calculation on day 4 (n=10), day 8 (n=10) and day 12 (n=5). On day 8, five rats from each group were euthanised randomly. Wound tissues were harvested for histological detection, including immunohistochemical staining targeting CD31, high-mobility group box 1 (HMGB1) and Caspase 3, alongside immunofluorescence staining of silent information regulator 3 (SIRT3) and inducible nitric oxide synthase (iNOS). Results Scanning electron microscopy revealed abundant mesoporous structures in MS, endowing it with high specific surface area. Following 1∶1 loading ratio, PRP@MS group a sustained VEGF release profile lasting over 96 hours. exhibited sustained vascular endothelial growth factor (VEGF) release profile lasting over 96 hours. The PRP@MS group showed no intradermal irritation, hypersensitivity, pyrogenic reaction or acute toxicity, demonstrating favourable biocompatibility. Compared with the routine dressing group, hematoxylin-eosin (HE), Masson's trichrome and Sirius red-picric acid staining revealed both the PRP and PRP@MS groups had more abundant granulation tissue. higher collagen content, and a more uniform and organized collagen distribution . In cellular experiments, CCK-8 assay and ELISA results illustrated that compared with the PRP group, human skin fibroblast (HSF) cultured with PRP@MS exhibited markedly elevated proliferation activity and intracellular SIRT3 expression, accompanied by decreased iNOS levels (P<0.05). In animal experiments, compared with the routine dressing group and PRP group, the PRP@MS group displayed significantly elevated wound healing rates, counts of CD31 immunohistochemistry positive cells at day 4, 8 and 12 (P<0.05). The number of HMGB1 positive cells were lower in the PRP@MS group (P<0.05). Compared with the routine dressing group, Caspase 3 positive staining was substantially lower in the PRP@MS group (P<0.05), with no statistical difference observed between the PRP and PRP@MS groups. Immunofluorescence results showed that the mean optical density of SIRT3 was significantly increased, while that of iNOS was significantly decreased in the PRP@MS group compared with the other two groups (P<0.05). Conclusion PRP@MS composites can preserve PRP bioactivity, enhance its therapeutic efficacy and achieve sustained factor release. Compared with standalone PRP, PRP@MS exhibits superior biological activity and favourable biocompatibility, thereby facilitating wound healing comprehensively. And the mechanism of action is related to the upregulation of SIRT3 and downregulation of iNOS.
Objective To verify the antiplatelet effects of quercetin (Qu), the main component of Guizhi Fuling Pill, using blood samples from healthy individuals. Methods Thromboelastography was used to measure the inhibition rates of platelets in response to arachidonic acid (AA) and adenosine diphosphate (ADP) after treatment with different concentrations of Qu. Using flow cytometry, we quantified the CD62P and PAC-1 expression levels in platelets exposed to varying concentrations of Qu. Platelet thromboxane B2 (TXB2 ) levels were also assessed by ELISA after incubation with different concentrations of Qu. Results Thromboelastography showed ADP and AA inhibition rates of the control group matched those in normal unactivated samples. Low-, medium-, and high-concentration quercetin (Qu, 10, 20, 40 μmol/L) groups exhibited significant differences in ADP and AA inhibition rates compared with blank controls (all P<0.01). Pairwise comparisons among the Qu groups showed a significant difference in the AA inhibition rate between Qu 10 μmol/L and Qu 40 μmol/L (P<0.05). Flow cytometry revealed CD62P and PAC-1 expression rates were markedly higher in positive controls than in blank controls (all P<0.01). Under thrombin stimulation, CD62P and PAC-1 expression levels at all three Qu concentrations differed significantly from those in positive controls (all P<0.01). Under ADP stimulation, only PAC-1 expression in the Qu 10 μmol/L group showed a statistically significant difference versus that in positive controls (P<0.05), whereas other Qu groups had significant differences (all P<0.01); no intergroup differences were detected among the three Qu concentrations (all P>0.05). ELISA showed TXB2 levels in agonist-activated positive controls were significantly different from those of blank controls (P<0.01). In the thrombin-activated system, TXB2 levels differed significantly in the Qu 10 μmol/L group (P<0.05) and showed a highly significant difference in the Qu 40 μmol/L group (P<0.01). In the ADP-activated system, Qu 10 μmol/L showed no significant change in TXB2 levels; Qu 20 μmol/L had a statistically significant difference (P<0.05), and Qu 40 μmol/L displayed a highly significant difference (P<0.01). No significant differences were observed among the Qu treatment groups (all P>0.05). Conclusion Qu exerts an inhibitory effect on platelet activation, specifically suppressing the activation triggered by ADP and thrombin.
Objective To investigate the dynamic changes in sphingolipid metabolism during red blood cell (RBC) storage and its association with oxidative stress and cellular maturity. Methods Three public metabolomics datasets from the Metabolomics Workbench were integrated: a longitudinal human RBC cold-storage metabolome (ST004148; sampled on days 1, 7, 14, 21, 28, 35, and 42 from 6 donors), a human reticulocyte-versus-mature RBC metabolome (ST000403), and an in-vivo RBC-aging metabolome of wild-type versus peroxiredoxin-2 knockout (Prdx2-/-) mice (ST004096). Linear mixed-effects (LME) models were used to analyze temporal trends during storage, with Benjamini-Hochberg (BH) correction for multiple testing. An elastic net regression model was established to predict storage age, with performance evaluated using leave-one-out cross-validation (LOOCV). Welch's t-test was used to compare differences between reticulocytes and mature RBCs. Results Among 165 metabolites, 124 (75.2%) showed significant changes during storage (FDR<0.05). Sphingosine accumulated significantly over time (slope 0.017 6 log10/d, P=1.6×10-19; log2FC=2.56 on day 42 vs day 1), whereas sphingosine-1-phosphate (S1P) declined (P=7.6×10-5) significantly, with no significant change in dihydrosphingosine-1-phosphate. The elastic-net model predicted storage age with a cross-validated R2 of 0.965 and a mean absolute error of 2.01 days and a mean absolute error of 2.01 d (Spearman ρ=0.985, P=3.3×10-32). In reticulocytes, de novo sphingolipid intermediates, such as 3-dehydrosphinganine (log2FC=3.99, FDR=0.001), were significantly higher than in mature RBCs, a finding independently replicated in 13C-labeled block (ρ=0.78). During in-vivo aging of RBCs in Prdx2-/- mice, stearoyl sphingomyelin SM(d18:1/18:0) showed genotype-specific loss (log2FC=-1.41, P=0.008). Conclusion Sphingolipid metabolic axis imbalance occurs during RBC storage, characterized by sphingosine accumulation and S1P depletion, which aligns with the direction of sphingomyelin loss driven by oxidative stress. The loss of de-novo sphingolipid synthesis capacity in mature RBCs may render this imbalance irreversible. The sphingolipid axis may serve as a candidate marker for assessing the RBC storage lesion, warranting further functional validation.
Objective To analyze the transfusion outcomes of two platelet transfusion strategies, minimal incompatibility and random transfusion, and to identify the optimal transfusion strategy for patients with serological cross-match incompatibility. Methods A retrospective analysis was conducted on the 24-hour Corrected Count of Increment (CCI) in 8 586 platelet transfusions at our hospital. Univariate analysis of variance was used to evaluate the efficacy of platelet transfusion and identify the primary factors; bivariate logistic regression analysis was used to predict the optimal transfusion strategy based on cutoff values stratified by the receiver operating characteristic (ROC) curve. Results When platelet antibody screening was positive and agglutination intensity was ≥2+, the transfusion efficacy in patients was significantly lower than in the platelet antibody-negative group (P<0.05). Among patients with positive platelet antibody screening, the transfusion outcomes in the matched transfusion group and the random transfusion group were superior to those in the minimal mismatch transfusion group (P<0.05). There was no statistically significant difference in transfusion failure rates between the minimal mismatch transfusion group and the agglutination intensity of the mismatched group (P>0.05). Results of binary logistic regression analysis showed no statistically significant difference between the platelet antibody ≤2+ group, the random transfusion group, and the reference group (P>0.05). Conclusion For patients with platelet antibody screening results of ≤2+, when matched platelet transfusions are unavailable, a random transfusion strategy is superior to the minimally mismatched transfusion strategy; when platelet antibody screening results are >2+, neither the minimally mismatched nor the random transfusion strategies should be selected in non-emergency situations.
Objective To evaluate the safety profile of transfusing frozen-thawed deglycerolized red blood cells (FTDRBCs) in patients with solid tumors, and to explore its application value in different clinical scenarios and significance for perioperative blood support. Methods A retrospective analysis was conducted on transfusion data from 100 patients with solid tumors treated between June 2025 and June 2026. Patients were divided into the intraoperative bleeding group (n=75) and the anemia correction group (n=25) based on the timing of the first FTDRBC transfusion.Using the transfusion episode as the analytical unit, we collected data on Hb before and after transfusion, unit transfusion efficiency (ΔHb/U), 72-hour subsequent red blood cell requirement rate, and total red blood cell exposure. Safety evaluation included dynamic changes in liver and renal function, coagulation parameters, and transfusion-related adverse reactions. The paired Wilcoxon test was used to analyze Hb changes before and after transfusion, the Mann-Whitney U test for between-group comparisons, and the chi-square test for comparing subsequent requirement rates between groups. Results In the intraoperative bleeding group, the baseline Hb was 71 (61, 83) g/L, and the median ΔHb at 24 h post-transfusion was 7 (2, 14) g/L (Z=-5.29, P<0.001); in the anemia correction group, the baseline Hb was 68 (55, 80.3) g/L, and the 24 h ΔHb was 7 (3.5, 10) g/L (Z=-3.99, P<0.001). In the transfusion-free subgroup without additional red blood cell transfusions within 72 h, the 48 h ΔHb in the two groups were 8 (-0.75, 18.75) and 8 (0.5, 19.5) g/L, respectively (Z=-0.079, P=0.937), and the 72 h ΔHb were 8.5 (-5, 15.75) and 12 (3.5, 26.5) g/L, respectively (Z=-1.099, P=0.272). The 72-hour subsequent red blood cell requirement rates in the two groups were 47% (35/75) and 32% (8/25), respectively, with no statistically significant difference (χ2=1.646, P=0.200). Multivariate linear regression showed that there was no significant correlation between storage days and 24 h ΔHb/u (P=0.520). Regarding safety, no statistically significant differences were observed in TBil, Cr, AST, ALT, APTT, TT, FIB, PT or D-dimer before and after transfusion; no transfusion-related adverse reactions were observed among the 100 patients. Conclusion Patients with solid tumors demonstrated a distinct Hb increase following the transfusion of FTDRBC. No statistically significant difference in ΔHb was observed between the two scenarios, and the need for subsequent transfusions remained controllable within 72 hours post-transfusion. Prolonged storage (with a median duration of 928 days) did not compromise transfusion efficacy. No evidence of liver or kidney function impairment, nor coagulation dysfunction, was detected, and no transfusion-related adverse reactions were reported. In contexts of limited blood resources or scarce rare blood type reserves, FTDRBC exhibits favorable safety and efficacy profiles in perioperative transfusion management for patients with solid tumors.
Objective This study aims to investigate the impact of two distinct preparation methods on the efficacy and safety of recombinant apheresis platelet transfusion, thereby providing empirical data to support its clinical application. Methods A prospective study was conducted involving 104 patients diagnosed with hematological disorders who had clear indications for platelet transfusion. Recombinant apheresis platelets were administered in cases where ABO-compatible platelets were unavailable. Based on the preparation method, the recombinant apheresis platelets were categorized into two groups: the centrifugation group and the apheresis group. The efficacy and safety of transfusions were subsequently compared between two groups. Results A total of 266 recombinant apheresis platelet transfusions were administered to the 104 patients. The centrifugation group enrolled 157 transfusions, while the apheresis group enrolled 109 transfusions. The comparison between the centrifugation and apheresis groups revealed the following: the average volume per single infusion was 0.98±0.08 for the centrifugation group and 0.96±0.09 for the apheresis group. The 12-hour post-transfusion platelet count increment was 9.5 (2, 18)×109/L for the centrifugation group and 9 (3, 17)×109/L for the apheresis group. In a comparative analysis of the two groups, the 12-hour Corrected Count Increment (CCI) was 6.08 (95%CI: 1.45, 12.19) versus 6.02 (95%CI: 1.19, 11.81), with infusion efficiencies of 57.32% and 64.22%, respectively. The subsequent platelet infusion intervals were 50.75 hours (95%CI: 24.88, 95.53) compared to 49.3 hours (95%CI: 26.58, 101.35). The rate of infusion-related adverse reactions was 0% in one group versus 0.92% in the other, with no statistically significant difference observed between the groups (P>0.05). However, the incidence of adverse events was higher in the centrifugation group at 4.46% compared to 0% in the apheresis group. The interval between platelet collection and transfusion was 2 days (range: 1-3 days) longer in the centrifugation group compared to 0 days (range: 0-1 day) in the apheresis group. The waiting time for platelet transfusion was reduced by 6.07 hours (95%CI: 3.55, 9.45) in the centrifugation group compared to 8.35 hours (95%CI: 4.99, 10.14) in the apheresis group. The number of recombinant apheresis platelet transfusions during hospitalization was 1 (range: 1-3) in the centrifugation group, which was higher than 1 (range: 1-2) in the apheresis group. Additionally, the total volume of platelet transfusion treatment during hospitalization was 7.80 units (95%CI: 4.00, 10.60) in the centrifugation group, exceeding the 6.00 units (95%CI: 3.6, 8.8) in the apheresis group.The regression analysis and mediation analysis indicated that there was no statistically significant difference in infusion efficacy [Odds Ratio (OR) 1.03—1.34, P>0.05] and safety (OR 1.99—4.35, P>0.05) between the two preparation methods. Conclusion Recombinant apheresis plateletsprepared through either centrifugation or apheresis are suitable for clinical infusion in patients with hematological disorders when ABO identical platelet is deficient. In practical applications, considerations such as current platelet inventory, infusion time constraints, patient blood type, donor blood type and quantity should be taken into account.
Objective To analyze the major causes of positive unexpected antibody screening with negative or nonspecific identification, and to develop a workflow that improves specific alloantibody detection while reducing unnecessary referrals. Methods We examined antibody screening-positive specimens referred to a reference laboratory from November 25, 2024 to February 5, 2026. Of 562 specimens completing antibody identification, 119 (21.17%) yielded negative or nonspecific results, and 50 were selected for detailed etiological investigation. Four specimens with known antibody specificity were sent to four hospital laboratories for multicenter comparison with screening cells and anti-human globulin cards unified to control variables. Seven sera, including one weak anti-E positive control, were tested weekly to monitor agglutination intensity of the same batch of screening cells stored from weeks 5 to 15; data were analyzed using a mixed-effects model (REML) and Kaplan-Meier method. Based on these findings, a three-step identification protocol was piloted in two hospitals for 2 months. Seventeen referred specimens (7 compliant) formed the intervention group, and 12 specimens from the same period of the previous year served as historical controls to compare the detection rate of specific antibodies. Results Among the 50 specimens, the main causes were false-positive screening cells, saline-reactive antibodies, weak alloantibodies and antibodies of undetermined specificity (AUS); HLA antibodies were detected in some cases. In the multicenter comparison, cold-reactive anti-M was detected only in laboratory 1, and after unified use of the anti-human globulin card, neither cold agglutinin nor anti-M was detected. Agglutination intensity of the 7 monitored specimens increased significantly with prolonged storage [F (8,42)=11.35, P<0.000 1], and was significantly higher at weeks 12 (P=0.000 5), 13 (P=0.030 9), 14 (P<0.000 1) and 15 (P=0.025 7) than at week 5. All 7 specimens developed false-positive events within 15 weeks (median, 12 weeks). During the pilot, specific antibody detection was 100% (7/7) in the compliant group versus 70% (7/10) in the non-compliant group(absolute risk difference 30%, 95%CI: 2%—58%, P=0.237). Weak-positive specimens were more frequent in the intervention group than in historical controls (64.7%) vs 16.7%, P=0.021). Conclusion sStorage-related false positivity of screening cells is an important modifiable cause of nonspecific antibody identification, especially after ≥12 weeks of storage. The three-step protocol can improve specific antibody detection. Incorporating screening cell functional status into routine quality control may reduce false-positive interference.
Objective To quantify the cumulative probability of novel antigen exposure and model-based immunological risk among individuals with distinct RhCE phenotypes under random transfusion conditions, thereby providing a quantitative rationale for proposed precision transfusion strategies. Methods This was a secondary analysis. We pooled RhD, C, c, E, and e antigen typing data from 458 542 Chinese individuals across eight studies reported by Zhao Tongmao, and confirmed Hardy-Weinberg equilibrium. Maximum-likelihood estimation was used to derive Rh haplotype frequencies and RHCE allele frequencies. On this basis, an independent-event cumulative probability model was applied to calculate the cumulative novel antigen exposure probability for nine RhD-positive phenotypes across one to five random transfusions, and a weighted immunological Risk Index (RI) was constructed to comprehensively assess the alloimmunization risk of each phenotype. The RI is a composite model-derived score and does not equate to the observed incidence of alloimmunization. Results The most common haplotype in the Chinese population was DCe (65.80%), followed by DcE (24.77%). The DCCee phenotype (46.00% of the population) represented the highest-risk group, with a single-transfusion novel antigen (E, c) exposure probability of 53.77%, rising to 97.89% after five transfusions, and a model-derived RI of approximately 43.95, classified as high risk by the model definition. The DCcEe phenotype (33.58% of the population), owing to its simultaneous carriage of C, c, E, and e antigens, had a transfusion-related RhCE novel antigen exposure probability of 0% under the conditions of this model; however, this applies only to the RhCE system and does not render such individuals universal recipients. For e antigen-negative homozygous phenotypes (e.g., DCcEE), random transfusion makes e antigen exposure virtually inevitable, with cumulative probability approaching 100% after multiple transfusions; nevertheless, because of the extremely low immunogenicity of the e antigen (weight 0.071), the model-derived RI remained limited. Conclusion The distribution of RhCE phenotypes in the Chinese population is highly uneven, and the DCCee phenotype constitutes the primary target population for proposed precision transfusion strategies. Model results indicate that both L1-A targeted antigen avoidance and L1-B (C/c/E/e) complete phenotype matching can reduce the probability of target RhCE novel antigen exposure to 0%; however, the two approaches differ in donor pool size and clinical feasibility, and neither should be equated with complete elimination of clinical alloimmunization risk. The DCcEe phenotype can be regarded as a "pan-antigen carrier" only within the RhCE system and is not a clinically universal recipient; alloimmunization against other blood group systems remains possible. Based on these model projections, we recommend that RhCE typing be progressively incorporated into pre-transfusion testing, and that RhCE-matched proposed strategies be initiated as early as possible for high-risk phenotypes and patients requiring long-term transfusion, so as to minimize the risk of alloimmunization; these specific strategies warrant further validation in multicenter prospective studies.
Objective To investigate the relationship of direct antiglobulin test (DAT) patterns and Rh phenotypes with undetermined specificity in antibody identification in patients with positive unexpected antibody screening results, thereby providing a basis for early recognition of specimens requiring complex antibody identification procedures. Methods Clinical data were retrospectively collected from 445 patients with positive unexpected antibody screening results at our hospital from January 2023 to June 2026. All patients underwent antibody identification, DAT, and Rh typing, and only the first eligible record was retained for each patient. DAT results were classified as negative, IgG-only, or C3d-containing. The primary outcome was undetermined specificity in antibody identification. Univariate and multivariable logistic regression analyses were performed, and clinical, clinical+Rh, and clinical+Rh+DAT models were developed. Results Among the 445 patients, antibody specificity could not be determined in 148 (33.3%). The rates of undetermined antibody specificity in patients with negative, IgG-only, and C3d-containing DAT results were 21.9% (63/288), 39.4% (28/71), and 66.3% (57/86), respectively (P<0.001). After adjustment for clinical disease category and Rh phenotype, the likelihood of undetermined antibody specificity remained higher in patients with IgG-only and C3d-containing DAT results than in DAT-negative patients, with adjusted odds ratios of 2.020 (95%CI: 1.106—3.686) and 6.440 (95%CI: 3.603—11.512), respectively. The CcDEe phenotype was also associated with the outcome (aOR=3.251, 95%CI: 1.909—5.537). Among 161 patients with Rh system antibodies, 148 (91.9%) showed phenotype-antibody concordance and 13 (8.1%) showed discordance; the DAT positivity rate was higher in the latter than in the former (61.5% vs 26.4%, P=0.021). The area under the receiver operating characteristic curve (AUROC) was 0.651, 0.703, and 0.770 for the three models, respectively, with P<0.001 for both adjacent-model comparisons. Conclusion Among patients with positive unexpected antibody screening results, both IgG-only and C3d-containing DAT results were independently associated with undetermined antibody specificity, with a stronger association for C3d-containing DAT results. After sequentially adding Rh phenotype and DAT results to the clinical data-based model, the model's ability to distinguish antibody specificity that was not clearly defined gradually improved, enabling early identification of specimens requiring complex identification and thereby improving the timeliness of clinical blood transfusion.
Objective To explore the association between the transfusion characteristics, transfusion volume, transfusion efficacy and prognosis of adult patients with hemophagocytic lymphohistiocytosis (HLH). Methods This study retrospectively enrolled 50 adult patients with HLH in the First People's Hospital of Guangzhou from January 2018 to December 2024. All patients met the diagnostic criteria of HLH-2004. Demographic data, laboratory indicators, infusion records of various blood products, and follow-up survival data of the patients were collected. We compared the differences in the use of blood products between survivors and non-survivors, and analyze the correlation between blood transfusion volume and the therapeutic effect of platelet transfusion, inflammatory markers, and survival outcomes. Results Among the 50 patients, a total of 439 U of platelets and 572 U of suspended red blood cell (RBC) were infused. 60% of the patients had coagulation dysfunction. The non-survivor group had significantly higher transfusion volumes of platelets, plasma, as well as daily mean RBC and platelet transfusions compared to the survivor group (all P<0.05). The positive rate of platelet antibody screening was 28%, which was positively correlated with the platelet transfusion volume, while the adjusted platelet recovery rate was negatively correlated with the total transfusion volume. With the increase of the RBCs transfusion threshold, the corrected hemoglobin increment decreased, and the 24-hour platelet post-transfusion recovery (PPR) is the lowest when the PLT<10×109/L transfusion threshold. High RBC transfusion volume (≥8 U) was associated with a synchronous increase in post-transfusion ferritin after blood transfusion, and high platelet transfusion volume (≥7 U) was associated with a synchronous increase in post-transfusion interleukin-10. Multivariate analysis identified coagulation dysfunction as a factor related to poor prognosis. Conclusion Patients with HLH have a high demand for blood transfusion. An increase in blood transfusion volume does not bring greater survival benefits. Coagulation dysfunction is a risk factor for poor prognosis of HLH.
Objective To provide a basis for improving the professional title evaluation and assessment system for transfusion physicians, we explored the practice scope, current status of professional title promotion and career development bottlenecks of transfusion physicians in China. Methods This study conducted an online questionnaire survey among transfusion physicians from transfusion departments/blood banks and blood collection and supply institutions in China in December 2025, and collected a total of 1 188 valid questionnaires. The survey covered multiple dimensions, including respondents' basic demographic information, practice registration status, residency standardized training experience, professional title distribution, difficulties in career promotion, professional identity and career development expectations. Descriptive statistical methods were employed to analyze the obtained data. Results Among the 1 188 individuals participating in this survey, 1 144 were from healthcare institutions and 44 were from blood collection and supply agencies. Their main professional backgrounds were clinical medicine (52.6%) and clinical laboratory medicine (45.6%). In terms of practice registration, 55.5% were registered as practitioners under the specialty of "Clinical Laboratory Medicine or Pathology", while 34.2% were registered under the specialty of Internal Medicine. Regarding the professional title distribution, 60.6% of participants held intermediate or senior titles. Among those with intermediate titles, 50.4% had held the title for more than five years, 22.7% of senior-title holders transferred to the Department of Transfusion Medicine after obtaining their current title, and 29.6% of transfusion physicians also held technician-level professional titles. Participants identified "incompatibility of professional specialty" as the top barrier to career advancement, accounting for 69.9% of responses, 67.1% believed that the current professional title system cannot accurately evaluate the professional competence of physicians working in the Department of Transfusion Medicine. Only 21.5% of the surveyed physicians reported a strong sense of professional belonging, while 78.2% considered their career development path to be "unclear". The most urgent demand, put forward by 55.4% of respondents, was to establish an independent professional title for "Transfusion Medicine". Conclusion At present, transfusion physicians generally face problems including mismatches between professional registration and educational background, unclear career development paths, and limited opportunities for professional promotion. These issues severely hinder the development of transfusion medicine talents and the disciplinary progress of the field.
Objective To investigate the association between the FUT2 gene variation and the Helicobacter pylori (Hp) infection in Han population from Dalian, China. Methods Peripheral blood samples were collected from 302 Dalian Han blood donors at Dalian Blood Center between October 2022 and February 2023. The coding region of the FUT2 gene was amplified by polymerase chain reaction (PCR), and genetic variants were identified by Sanger sequencing. Plasma anti-Hp antibodies were detected using a latex-enhanced immunoturbidimetric assay. We compared the associations of different FUT2 phenotypes, FUT2 genotypes, and phenotypes with Hp seropositivity, and analyzed the effects of age, sex, and other factors on antibody positivity using logistic regression. Results A total of 302 Han blood donors were enrolled, among whom 87 (28.8%) were positive for anti-Hp antibodies. Seven FUT2 variant sites were identified, and the genotype distributions of all variants were in accordance with Hardy-Weinberg equilibrium (all P>0.05). Univariate analysis showed that age, sex, alcohol consumption, and body mass index (BMI) were associated with Hp seropositivity (all P<0.05), whereas no significant associations were observed between Hp seropositivity and FUT2 variant genotypes, FUT2 genotypes, or FUT2 phenotypes (all P>0.05). Multivariate logistic regression analysis demonstrated that age was an independent risk factor for Hp seropositivity (OR=1.041, 95%CI: 1.017—1.068, P=0.001), and males had a higher risk of Hp seropositivity than females (OR=1.810, 95%CI: 1.007—3.308, P<0.05). Alcohol consumption, BMI, and FUT2 genetic variations were not significantly associated with Hp seropositivity (all P>0.05). Conclusion No significant association was found between FUT2 genetic variations and susceptibility to Hp seropositivity among Han blood donors in Dalian. Age and sex may be important factors influencing Hp seropositivity, with increasing age and male sex being associated with a higher risk of Hp seropositivity.
Objective This study explored the influencing factors and action pathways of group voluntary blood donation behavior based on group dynamics, so as to provide evidence for improving the recruitment strategies for group blood donation. Methods Research units were randomly selected from enterprises and institutions scheduled to hold group blood donation campaigns between June and July 2026. Employees were recruited at the campaign sites for questionnaire surveys and blood donation behavior follow-up. A validated Group Voluntary Blood Donation Cognitive and Intention Scale developed based on group dynamics theory was used to measure six cognitive factors including task cohesion and conformity. Bivariate correlation analysis was performed for preliminary variable screening, and structural equation modeling (SEM) was adopted to identify influencing factors and their pathways. Results Valid questionnaires were collected from 695 participants, of whom 295 completed blood donation on the same day. Participants in the donation group had significantly more previous blood donations, and scored significantly higher in all cognitive factors than the non-donation group. Structural equation modeling revealed that a moderate significant positive association between group voluntary blood donation intention and behavior (β=0.293, P<0.001). Social cohesion, task cohesion, self-monitoring, conformity, and self-efficacy influenced group voluntary blood donation behavior via moderate-to-strong positive associations with donation intention (βranging from 0.587 to 0.732, P<0.001). The model explained 19% of the variance in group blood donation behavior. Conclusion Cognitive factors from group environment and individual evaluation jointly influence group blood donation intention and behavior. Blood centers could enhance employee-oriented donation promotion and interaction through long term collaboration, empowering core volunteer, and cross-unit networking activities. By leveraging the exemplary role of donors, stratified recruitment strategies may be established to help enterprises cultivate a supportive environment and foster employees' individual identity.
Objective To investigate the association between HLA-Ⅱ gene polymorphisms and red blood cell alloantibody formation in a northern Chinese population. Methods A total of 299 clinical transfusion-related samples were enrolled, including 129 red blood cell alloantibody responders and 170 non-responder controls with multiple transfusions (≥10 U) but negative antibody screening. High-resolution genotyping of HLA-DRB1, DQA1, DQB1, DPA1, and DPB1 was done using PacBio long-read sequencing. The distribution of HLA-Ⅱ alleles was compared between responders and non-responders, followed by stratified analyses according to common antibody specificities. Overall associations were evaluated using logistic regression adjusted for age and sex, with additional adjustment for the number of previous pregnancies in sensitivity analyses; multiple comparisons were corrected for the false discovery rate (FDR) by the Benjamini-Hochberg (BH) method. Results After adjustment for age and sex, the odds ratios (ORs) were>1 for DRB1*09:01, DQA1*03:02, DQB1*03:03, and DPB1*02:02 and<1 for DRB1*13:02, DQB1*06:04, DPA1*01:03, and DPB1*04:01, all with P<0.05. However, none of the 69 HLA-Ⅱ alleles reached q<0.05 after FDR correction. After further adjustment for the number of previous pregnancies, the directions of the ORs for these alleles were generally consistent. In antibody-specific analyses, DQA1*02:01 and DQB1*02:02 in the anti-C group remained significant after within-group FDR correction (q=0.039), whereas no other antibody-specific results reached q<0.05 after correction. ConclusionHLA-Ⅱ gene polymorphisms may be associated with red blood cell alloantibody responses, with some HLA-Ⅱ alleles potentially showing different association patterns across antibody-specific subgroups. This study adds to the genetic evidence on HLA-Ⅱ in relation to red blood cell alloimmunization in the Dalian region; these findings require validation in larger samples and independent cohorts.
We report a case of therapeutic whole blood exchange (TBE) for propanil-induced methemoglobinemia (MetHb) after the cumulative dose of methylene blue (MB) exceeded the recommended limit. The patient was a 50-year-old woman who received a cumulative MB dose of approximately 8 mg/kg, exceeding the recommended cumulative limit of 7 mg/kg. Although MetHb decreased from 17.8% to 5.5%, she remained comatose and cyanotic. Given the risk of paradoxical worsening of methemoglobinemia and hemolysis with further MB, additional dosing was withheld following multidisciplinary consultation, and TBE was performed. In this case, TBE consisted of therapeutic plasma exchange (TPE) followed by red blood cell exchange (RBCX) within one continuous treatment session to remove and replace both plasma and red blood cells. MetHb was 3.6% during TBE, 4.4% at 1 and 12 hours after TBE, 2.0% at 21 hours, and 0.5% at 36 hours. Intraoperative SpO2 increased to 97%~98%. The patient regained consciousness and was extubated at 24 hours after TBE. Vital signs remained stable during the procedure, and no serious transfusion-related adverse reactions occurred. This case suggests that TBE may be considered as an individualized rescue regimen when further MB administration is limited by safety concerns and severe clinical manifestations persist. Because MB had already substantially reduced MetHb and the patient concurrently received mechanical ventilation and other supportive treatments, the independent efficacy of TBE cannot be established from this case. Further cases are needed to clarify its indications, exchange dose, and safety.
Dysregulated immune cell lipid metabolism is a critical link between metabolic homeostasis and the development and progression of immune-mediated inflammation. Lipid metabolism comprises fatty acid, cholesterol, sphingolipid, and phospholipid metabolism. Dysregulation of these metabolic processes can remodel immune cell membrane architecture and signaling platforms, thereby amplifying inflammatory signaling, promoting reactive oxygen species (ROS) accumulation, impairing phagocytosis and migratory functions, and disrupting immune cell functional states. Collectively, these alterations contribute to the development and progression of autoimmune diseases and other chronic inflammatory disorders. Emerging evidence further indicates that metabolic pathways and immune homeostasis are interconnected through complex spatiotemporal regulatory networks, and that distinct immune cell types and subsets exhibit pronounced metabolic heterogeneity. These features complicate the identification of key metabolic targets and the prediction of therapeutic responses. Meanwhile, therapeutic strategies targeting key lipid metabolic pathways have shown therapeutic potential, including statins and PCSK9 inhibitors, modulation of the LXR, AMPK, and PPARγ signaling, and inhibition of fatty acid synthesis-related pathways. Such interventions may alleviate metabolic dysregulation, reshape the immune microenvironment, and therefore mitigate inflammatory responses. In this review, we summarize the mechanisms by which immune cell lipid metabolism dysregulation contributes to inflammatory diseases, review current advances and translational applications of lipid metabolism-targeted interventions, and discuss key scientific questions and future research directions. These insights may inform precision interventions and individualized treatment of inflammatory diseases.
With the acceleration of population aging, cognitive dysfunction has become a major public health issue that seriously affects the health of the elderly. Recent studies have demonstrated that erythrocyte dysfunction is closely associated with the onset and progression of cognitive dysfunction. This article systematically reviews the associations between erythrocyte abnormalities and distinct phenotypes of cognitive impairment and explores their underlying mechanisms, aiming to provide new insights for the early identification and intervention of cognitive dysfunction. Existing evidence indicates that erythrocyte abnormalities are not merely peripheral manifestations of cognitive dysfunction, but may actively participate in disease progression via chronic cerebral hypoxia, microcirculatory disturbance, oxidative stress, and dysregulated peripheral-central signaling transmission. Among these markers, conventional hematological indices such as hemoglobin and red cell distribution width (RDW) are non-invasive, convenient and available for dynamic monitoring, and are clinically accessible. In contrast, cutting-edge research indicators, including erythrocyte membrane lipidomics and microRNA (miRNA) profiles of red blood cell-derived extracellular vesicles (RBC-EVs), remain investigational owing to high cost and technical barriers. Both marker categories hold promise for risk stratification and auxiliary diagnosis of cognitive impairment. However, their causal associations and clinical translational value remain to be validated by large-scale prospective studies and targeted interventional trials.
Red blood cell (RBC) transfusion is widely used in clinical practice. However, during storage, RBCs are susceptible to oxidative stress induced by external environmental factors and internal metabolic changes, leading to biochemical, morphological, and functional alterations that ultimately result in storage damage and cellular senescence. Oxidative stress constitutes a pivotal mechanism underlying RBC storage injury, involving free radical accumulation, depletion of endogenous antioxidant systems, and membrane lipid peroxidation. Recent studies have demonstrated that traditional Chinese medicines and their active components exhibit significant advantages in mitigating oxidative stress; consequently, research on their intervention in RBC oxidative stress and modulation of antioxidant defense systems has garnered increasing attention. This review summarizes recent domestic and international advancements and efficacy findings regarding the use of TCM active components as antioxidants to alleviate RBC oxidative stress, aiming to provide theoretical foundations and references for the clinical translation and application of these components in RBC storage.
Red blood cell (RBC) transfusion is one of the most widely used supportive therapies in clinical practice. However, single hemoglobin (Hb) threshold poorly reflects the complex status of oxygen delivery and tissue oxygen utilization in individual patients, and evidence from special populations continues to challenge the universal applicability of RBC restrictive thresholds. Based on the physiology of oxygen supply-demand balance at the tissue level, this review summarizes recent advances in physiological indicators for individualized transfusion: 1) tissue oxygenation monitoring provides noninvasive and continuous information on organ oxygenation, but heterogeneous responses to transfusion make it difficult to establish universal thresholds; 2) lactate levels and its clearance rates are better suited as tools for stratification and trend assessment than transfusion triggers. Individualized thresholds for special populations and emerging directions such as artificial intelligence are also reviewed. Physiological indicator-guided transfusion represents the core direction in the evolution of individualized transfusion, but most indicators still lack prospective validation and standardized trigger values. Future research should focus on multimodal integration, closed-loop decision-making, and prospective validation with hard endpoints, so as to optimize the benefit-risk balance of transfusion.
Patients with chronic kidney disease (CKD) have a significantly higher transfusion rate than the general population owing to renal anemia, coagulopathy and perioperative requirements, and their transfusion-related risks are characterized by multiple and overlapping risks. Current evidence indicates that CKD is an important independent risk factor for transfusion-associated circulatory overload (TACO) (OR=27, 95%CI: 5.4—143); that the incidence of transfusion-related acute lung injury (TRALI) under active surveillance varies with blood component, being 0.17, 0.31 and 3.19 per 10 000 units transfused for red blood cells, platelets and plasma, respectively; and that only 44.2% of patients on maintenance hemodialysis in China achieve the target hemoglobin level of 110~130 g/L. This review systematically elaborates the risk profiles and underlying mechanisms of TACO, TRALI, red blood cell alloimmunization, iron overload, transfusion-transmitted infections (TTIs) and transfusion-associated graft-versus-host disease (TA-GVHD), and proposes prevention and control strategies including restrictive transfusion, appropriate blood component selection, risk-stratified extended antigen matching, irradiation and leukoreduction. The roles of erythropoiesis-stimulating agents (ESAs), hypoxia-inducible factor prolyl hydroxylase inhibitors (HIF-PHIs) and patient blood management (PBM) in reducing the need for transfusion are also discussed. This review aims to provide an evidence-based reference for transfusion medicine-led decision-making and quality control of transfusion in patients with CKD.
The classic pathogenesis of hemolytic disease of the fetus and newborn (HDFN) is that maternal IgG isoantibody is transported into fetal circulation through placental FcRn, which causes immune damage of fetal red blood cells. This mechanism can explain the intrauterine and early postnatal manifestations of most children, but there are limitations in explaining that some children still have persistent or delayed hemolysis several weeks after birth. Case reports in recent years suggest that red blood cell alloantibodies in breast milk may be absorbed into the circulation through the neonatal intestine and participate in the maintenance of postpartum hemolysis. Based on the summary of existing cases, this paper reviews the source of breast milk antibody, the physiological basis of neonatal intestinal absorption, the clinical characteristics, detection methods and causal judgment of related cases. Detection of breast milk alloantibodies currently relies on serological methods, and sample processing, titer reporting, and pathogenic thresholds all lack standardization. From the mechanism point of view, breast milk antibody is more likely to play the role of "maintaining attack" and prolong the course of hemolysis during the window period when placental antibody has not been cleared, rather than starting HDFN independently. The existing evidence is not enough to support the classification of antibody-mediated hemolysis in breast milk as an independent fifth pathogenic mechanism. Prospective cohort study is necessary in the future to clarify its clinical incidence and establish standardized detection and intervention strategies.
Perioperative coagulation management is a crucial aspect for ensuring the safety of surgical procedures and the prognosis of patients. It involves multiple stages, such as preoperative risk assessment, intraoperative dynamic intervention, and postoperative continuous monitoring. With the increasing complexity of surgical procedures and the widespread use of anticoagulant drugs, the balance management of perioperative bleeding and thrombotic factors has become a major challenge in clinical practice. Although current research has established a basic framework based on coagulation function monitoring, individualized adjustment of anticoagulant drugs, and blood transfusion strategies, there are still many unresolved issues in the reversal of new oral anticoagulants, the real-time techniques of coagulation monitoring, and the application of new hemostatic materials. Moreover, there is a lack of unified high-level evidence. This article systematically reviews the latest research progress in perioperative coagulation management, focusing on bleeding risk assessment, perioperative management plans for anticoagulant drugs, blood transfusion treatment strategies, and the clinical application of new local and systemic hemostatic techniques. By integrating recent clinical studies, guideline updates, and expert consensus, the aim is to provide more scientific and individualized treatment plans for clinicians, in order to effectively reduce the incidence of perioperative bleeding and thromboembolic events and improve patient prognosis.
Platelet ABH antigens are terminal glycan structures carried by membrane glycoproteins and glycosphingolipids. Their expression reflects endogenous glycosylation along the megakaryocyte-platelet axis and adsorption of ABH-active substances from plasma. The GPⅠb-Ⅸ-Ⅴ complex and integrins can carry ABH determinants, with substantial variation among the A1 and A2 subgroups and the high-expresser phenotype (HXP) and low-expresser phenotype (LXP). HXP and LXP are assay-dependent expression phenotypes rather than new ABO subgroups; their classification varies with reagents, flow-cytometry platforms, analytical methods, and thresholds, and no cross-center standard is currently available. Available functional studies suggest that ABO-related glycosylation may affect the binding kinetics between GPⅠb-Ⅸ-Ⅴ and von Willebrand factor (VWF), Furthermore, anti-A/B antibodies have been shown to alter platelet aggregation, spreading, thrombin generation, and mediator release in vitro. However, these findings are currently insufficient to establish a causal relationship with clinical bleeding or thrombosis events. In transfusion practice, major ABO incompatibility concerns post-transfusion platelet recovery and count increments, minor incompatibility can cause rare but severe hemolysis through passively transfused antibodies, and bidirectional incompatibility combines both risks. When inventory permits, ABO-identical platelets may be considered preferentially. Platelets with A2 or LXP phenotype may be valuable in selected major ABO-incompatible settings, but standardized screening and prospective clinical-outcome evidence remain insufficient for routine recommendation.
β-Thalassemia is an autosomal recessive hemoglobin disorder caused by pathogenic variants in the β-globin gene. Red blood cell transfusion with iron chelation improves patient survival but cannot correct the underlying genetic defect. Autologous hematopoietic stem and progenitor cell gene therapy offers a novel strategy to achieve sustained transfusion independence. This review summarizes the technical advances, clinical evidence, and translational challenges of gene therapy for β-thalassemia. Lentiviral gene addition delivers an exogenous β-globin expression cassette for genomic integration, whereas CRISPR nuclease editing targeting the BCL11A erythroid enhancer reactivates fetal hemoglobin. Representative products of these strategies have obtained regulatory approvals. Nuclease-mediated editing can also target the HBG1/HBG2 promoters. Base editing enables modification of fetal hemoglobin regulatory elements without generating programmed DNA double-strand breaks, and it can repair selected eligible pathogenic variants within the β-globin gene. Base editing for fetal hemoglobin regulation has entered early-stage clinical development, yet the scope of its direct gene repair is constrained by permissible base conversion types and other limitations. Prime editing can theoretically mediate all 12 classes of base substitutions and correct a subset of small insertions and deletions. Nevertheless, direct repair of the β-globin gene via base editing and prime editing is still confined to preclinical investigation. These technologies are not a simple "new-for-old replacement" of platforms. Their clinical utility is determined by editing accuracy, long-term HSPC fitness, conditioning regimen toxicity, product consistency, manufacturing costs and patient accessibility.
Programmed death of red blood cells is an active cellular death process that occurs in mature erythrocytes under specific physiological and pathological stimuli. Unlike apoptosis in nucleated cells, mature erythrocytes lack the classic apoptotic execution structures such as a nucleus and mitochondria, and their death modes possess unique molecular features and regulatory networks. The currently known major pathways include the eryptosis pathway characterized by phosphatidylserine externalization and cell shrinkage, the newly discovered lytic programmed death Spectosis, as well as related pathways such as ferroptosis and erythronecroptosis. These pathways form a complex interactive regulatory network through key mechanisms including calcium signaling, oxidative stress, sphingolipid metabolism, and iron metabolism. Abnormal regulation of programmed erythrocyte death is widely involved in the pathological processes of various diseases, including hemolytic anemia, sickle cell disease, sepsis, malaria, chronic kidney disease, and diabetes. Its excessive activation can lead to premature clearance of red blood cells and tissue damage, while insufficient regulation may affect erythrocyte homeostasis. This article systematically reviews the main types of programmed erythrocyte death, molecular regulatory mechanisms, and cross-talk between signaling pathways, and explores their pathophysiological significance in various diseases, which could provide a theoretical basis and new ideas for the development of diagnostic markers and therapeutic intervention strategies targeting programmed erythrocyte death.
Heparin-induced thrombocytopenia (HIT) is a severe immune-mediated adverse reaction caused by heparin drugs, primarily mediated by anti-platelet factor 4 (PF4)-heparin complex IgG antibodies, and characterized by progressive thrombocytopenia and increased risk of thrombosis. Clinically, HIT is mainly classified into two types: type Ⅰ and type Ⅱ, with type Ⅱ being more harmful. If not diagnosed and treated promptly, it can lead to patient disability or even death. Currently, clinical diagnosis primarily relies on the 4Ts score combined with laboratory tests, among which immunological screening assays have high sensitivity but relatively low specificity. The serotonin release assay (SRA) is currently recognized as the functional gold standard for HIT diagnosis (with a sensitivity of 90%~95% and specificity greater than 98%). In recent years, with the continuous development of laboratory technology, the progressive application of new technologies such as automated chemiluminescence immunoassay, novel rapid functional assays, and flow cytometry, has optimized the diagnostic workflow for HIT. This article systematically reviews the pathogenic features, clinical classification, and typical symptoms of HIT, and compares the advantages and disadvantages of various HIT laboratory detection methods, as well as the latest research advances in these fields, aiming to provide a reference for rapid clinical identification of HIT and precise laboratory testing.