Prediction of intraoperative red blood cell transfusion in valve replacement surgery: machine learning algorithm development based on non-anemic cohort.

Background: Our study aimed to develop machine learning algorithms capable of predicting red blood cell (RBC) transfusion during valve replacement surgery based on a preoperative dataset of the non-anemic cohort. Methods: A total of 423 patients who underwent valvular replacement surgery from January 2015 to December 2020 were enrolled. A comprehensive database that incorporated demographic characteristics, clinical conditions, and results of preoperative biochemistry tests was used for establishing the models. A range of machine learning algorithms were employed, including decision tree, random forest, extreme gradient boosting (XGBoost), categorical boosting (CatBoost), support vector classifier and logistic regression (LR). Subsequently, the area under the receiver operating characteristic curve (AUC), accuracy, recall, precision, and F1 score were used to determine the predictive capability of the algorithms. Furthermore, we utilized SHapley Additive exPlanation (SHAP) values to explain the optimal prediction model. Results: The enrolled patients were randomly divided into training set and testing set according to the 8:2 ratio. There were 16 important features identified by Sequential Backward Selection for model establishment. The top 5 most influential features in the RF importance matrix plot were hematocrit, hemoglobin, ALT, fibrinogen, and ferritin. The optimal prediction model was CatBoost algorithm, exhibiting the highest AUC (0.752, 95% CI: 0.662-0.780), which also got relatively high F1 score (0.695). The CatBoost algorithm also showed superior performance over the LR model with the AUC (0.666, 95% CI: 0.534-0.697). The SHAP summary plot and the SHAP dependence plot were used to visually illustrate the positive or negative effects of the selected features attributed to the CatBoost model. Conclusions: This study established a series of prediction models to enhance risk assessment of intraoperative RBC transfusion during valve replacement in no-anemic patients. The identified important predictors may provide effective preoperative interventions.

Understanding Intraoperative Transfusion Decision-Making Variability: A Qualitative Study.

There is evidence of significant intraoperative red blood cell (RBC) transfusion variability that cannot be explained by case-mix, and may reflect unwarranted transfusions. The objective was to explore the source of intraoperative RBC transfusion variability by eliciting the beliefs of anesthesiologists and surgeons that underlie transfusion decisions. Interviews based on the Theoretical Domains Framework were conducted to identify beliefs about intraoperative transfusion. Content analysis was performed to group statements into domains. Relevant domains were selected based on frequency of beliefs, perceived influence on transfusion, and the presence of conflicting beliefs within domains. Of the 28 transfusion experts recruited internationally (16 anesthesiologists, 12 surgeons), 24 (86%) were Canadian or American and 11 (39%) identified as female. Eight relevant domains were identified: (1) Knowledge (insufficient evidence to guide intraoperative transfusion), (2) Social/professional role and identity (surgeons/anesthesiologists share responsibility for transfusions), (3) Beliefs about consequences (concerns about morbidity of transfusion/anemia), (4) Environmental context/resources (transfusions influenced by type of surgery, local blood supply, cost of transfusion), (5) Social influences (institutional culture, judgment by peers, surgeon-anesthesiologist relationship, patient preference influencing transfusion decisions), (6) Behavioral regulation (need for intraoperative transfusion guidelines, usefulness of audits and educational sessions to guide transfusion), (7) Nature of the behaviors (overtransfusion remains commonplace, transfusion practice becoming more restrictive over time), and (8) Memory, attention, and decision processes (various patient and operative characteristics are incorporated into transfusion decisions). This study identified a range of factors underlying intraoperative transfusion decision-making and partly explain the variability in transfusion behavior. Targeted theory-informed behavior-change interventions derived from this work could help reduce intraoperative transfusion variability.

Timing of Blood Transfusions and 30-Day Patient Outcomes After Coronary Artery Bypass Graft Surgery.

OBJECTIVE: Packed red blood cell transfusion during coronary artery bypass graft surgery is known to be associated with adverse outcomes. However, the association of the timing between transfusions in relation to discharge and 30-day postoperative outcomes has not been studied. The study authors investigated the impact of transfusion timing on 30-day surgical outcomes. DESIGN: A retrospective review. SETTING: At a single tertiary-care academic hospital. PARTICIPANTS: A total of 2,481 adult patients underwent primary coronary artery bypass graft surgery between January 2014 and December 2020. MEASUREMENTS AND MAIN RESULTS: The relationship between the timing of packed red blood cell transfusion (intraoperative, postoperative, or both) and 30-day postoperative outcome variables was calculated as an odds ratio. The influence of timing of transfusion on adjusted probability of postoperative complications was plotted against the lowest intraoperative hematocrit. The median age of the population was 67 years (60.0-74.0), body mass index was 28.5 (25.6-32.3) kg/m2, and 497 (20.0%) were female. A total of 1,588 (36%) patients received packed red blood cell transfusions; 182 (7.3%) received intraoperative transfusions, 489 (19.7%) received postoperative transfusions, and 222 (9.0%) received both (intraoperative and postoperative transfusions). Postoperative transfusion was associated with significantly higher odds of readmission (1.83 [1.32-2.54], p = 0.002) and heart failure (1.64 [1.2-2.23], p = 0.008) compared to patients with no transfusions; whereas intraoperative transfusions were not. CONCLUSION: The authors' data suggested that the postoperative timing of transfusion in patients undergoing coronary artery bypass graft surgery may be associated with an increased incidence of 30-day heart failure and readmission. Prospective research is needed to conclusively confirm these findings.

Volume of packed red blood cells and fresh frozen plasma is associated with intraoperative hypocalcaemia during large volume intraoperative transfusion.

BACKGROUND: Severe hypocalcaemia is associated with increased transfusion in the trauma population. Furthermore, trauma patients developing severe hypocalcaemia have higher mortality and coagulopathy. Electrolyte abnormalities associated with massive transfusion have been less studied in the surgical population. Here, we tested the primary hypothesis that volume of packed red blood cells and fresh frozen plasma transfused intraoperatively is associated with lower nadir ionised calcium in the surgical population receiving massive resuscitation. METHODS: We performed a retrospective observational study at an academic quaternary care centre to characterise hypocalcaemia following large volume (4 or more units packed red blood cells) intraoperative transfusion. We used multivariable linear regression to assess if volume of transfusion with packed red blood cells and fresh frozen plasma were independently associated with a lower ionised calcium. We then used multivariable logistic regressions to assess the association between ionised calcium and transfusion with: (i) mortality, (ii) acute kidney injury, and (iii) postoperative coagulopathy. RESULTS: Hypocalcaemia following large volume resuscitation in the operating room is a very frequent occurrence (70% of cases). After controlling for demographic variables and intraoperative variables, the volume transfused intraoperative was independently associated with hypocalcaemia on multivariable linear regression. Hypocalcaemia, intraoperative transfusion of packed red blood cells, and intraoperative transfusion of fresh frozen plasma were not shown to be associated with clinical outcomes. CONCLUSIONS: Hypocalcaemia was associated with increased transfusion volume in this single-centre study. Unlike the trauma population, hypocalcaemia was not associated with increased mortality during surgical care. Our findings suggest that despite improved practice patterns of calcium supplementation, intraoperative hypocalcaemia occurs with relatively high frequency following large volume intraoperative transfusion.

Association between preoperative use of antithrombotic medications and intraoperative transfusion in older patients undergoing cancer surgery.

BACKGROUND: Management of antiplatelet agents and other chronic anticoagulation medications in patients scheduled for surgery can reduce intraoperative bleeding complications. However, few studies on the association of antithrombotics, relative to their duration of action, with intraoperative transfusion have been conducted. We aimed to determine the association of recent use of antithrombotics, relative to their duration of action, with intraoperative transfusion in elderly people undergoing cancer surgery. METHODS: The study subjects were patients aged 65 years or older who were scheduled for cancer surgery and presented for comprehensive geriatric assessment. We reviewed the baseline patient characteristics obtained from electronic medical records and the patients' preoperative medication history, including anticoagulants, antiplatelet agents, and streptokinase/streptodornase. RESULTS: A total of 475 cancer patients were included. Multivariate analysis showed that long-acting anticoagulant therapy before surgery was a significant risk factor for intraoperative transfusion. Long-acting anticoagulants increased the risk of transfusion approximately 15.9-fold (95% CI 1.9-136.2). The attributable risk of long-acting anticoagulants to transfusion was approximately 93.7%. Also, low body mass index (BMI) and hepato-pancreato-biliary (HPB) surgery were significantly associated with intraoperative transfusion. The adjusted odds ratios for low BMI (<18.5 kg/m2) and HPB surgery (reference: lower gastrointestinal surgery) were 5.3 (95% CI 1.8-15.4) and 4.9 (95% CI 1.9-12.5), respectively. CONCLUSIONS: It was found that the perioperative use of long-acting anticoagulants was associated with an increased risk of intraoperative transfusion, further highlighting the importance of medication optimization for elderly patients with cancer surgery.

Impact of Intra- and Postoperative Blood Transfusion on the Incidence, Timing, and Pattern of Disease Recurrence After Radical Cystectomy.

BACKGROUND: The administration of blood transfusion (BT) has been associated with a decrease in survival expectancies in patients treated with radical cystectomy (RC), as a consequence of the immunosuppressive effect mediated by BT. We sought therefore to evaluate if the usage of BT may influence the risk and pattern location of distant recurrences after RC, which may be influenced by this effect. METHODS: Data from 2 independent cohorts of consecutive patients with bladder cancer treated with RC were analyzed. Distant recurrence included all recurrence locations outside of the true pelvis, such as lung, liver, bone, extra pelvic lymph nodes, peritoneal, or brain recurrences. Cox regression analyses evaluating the risk of developing distant recurrence after RC were built. RESULTS: In the testing cohort, composed of 1081 patients, 41.2% received a perioperative BT. Within a median follow-up of 52 months (interquartile range, 44-61 months), 277 (25.6%) patients experienced a distant recurrence. In the validation cohort, composed of 433 patients, 42.3% received perioperative BT within a median follow-up of 83 months, and 127 (28.3%) patients experienced distant recurrence. On multivariable analyses predicting distant recurrences, BT was not associated with the risk of distant recurrence stratified by location and time (within first year or later after RC; all P ≥ .2) in both cohorts. CONCLUSIONS: BT administration was not associated with a different pattern, timing, or rate of distant recurrences in patients when compared with those who did not receive BT. New data are needed to investigate the mechanisms behind the association between BT and survival in RC patients.

Timing of blood transfusion and not ABO blood type is associated with survival in patients treated with radical cystectomy for nonmetastatic bladder cancer: Results from a single high-volume institution.

INTRODUCTION AND OBJECTIVES: Perioperative transfusions have been recently associated to poor outcomes as an indirect consequence of immune-hematological changes related to transfusion itself and blood type. We tested the role of blood transfusion on cancer-specific mortality (CSM) and overall mortality (OM), considering the effect of ABO system, Rh factor, and timing of transfusions. MATERIALS AND METHODS: The study focused on 728 patients with bladder cancer treated with radical cystectomy at a single tertiary care referral center between January 1995 and August 2013 with complete ABO blood type information. Kaplan-Meier analysis was used to assess the effect of transfusions, stratified according to ABO type and Rh factor, on CSM and OM. The same endpoints were tested in Cox regression models, after adjusting for all available confounders. RESULTS: A total of 341 (46.8%), 277 (38.0%), 83 (11.4%), and 27 (3.7%) patients had blood type O, A, B and AB, respectively. Overall, 630 (86.5%) and 98 (13.5%) patients were Rh-and Rh+, respectively. At a median follow-up time of 65 months, 225 (30.9%) and 282 (38.7%) patients recorded CSM and OM, respectively. At univariable analyses, ABO blood type and Rh status were not associated to either CSM or OM (all P>0.2). Similar results were observed when ABO blood type and Rh factor were tested in multivariable models (all P>0.3). Conversely, Charlson score, preoperative hemoglobin, number of nodes removed, pathological T stage, and number of positive nodes were associated to both CSM and OM (all P<0.05). Interestingly, intraoperative transfusion (all P<0.03) but not the administration of blood units in the postoperative period (P>0.05) was associated with an increase of CSM and OM. CONCLUSIONS: Although ABO type or Rh factor or both were associated with several adverse outcomes in many cancers, we were not able to confirm this association in bladder cancer. Based on our results, the effect of transfusion on survival is independent by ABO type but is associated to the timing of blood supply administration.

Effect of Allogeneic Intraoperative Blood Transfusion on Survival in Patients Treated With Radical Cystectomy for Nonmetastatic Bladder Cancer: Results From a Single High-Volume Institution.

BACKGROUND: Previous studies have demonstrated that perioperative blood transfusion (BT) is associated with a significantly increased risk of cancer recurrence and mortality after radical cystectomy (RC). Recently, it was shown for the first time that intraoperative transfusion has a detrimental effect on cancer survival. The aim of the current study was to validate this finding in a single European institution. PATIENTS AND METHODS: The study focused on 1490 consecutive nonmetastatic bladder cancer patients treated with RC at a single tertiary care referral center between January 1990 and August 2013. Kaplan-Meier analyses and Cox regression analyses were used to assess the effect of timing of BT administration (no transfusion vs. intraoperative transfusion vs. postoperative transfusion vs. intraoperative and postoperative transfusion) on cancer-specific mortality (CSM), overall mortality (OM), and disease recurrence. RESULTS: Mean age at the time of RC was 67 years. Overall, 322 (21.6%) patients received intraoperative BT and 97 (6.5%) received postoperative BT. At a mean follow-up time of 125 months (median, 110 months), the 5- and 10-year CSM rate was 846 (58%) and 715 (48%), respectively. In multivariable analyses patients who received intraoperative BT had greater risk of disease recurrence (hazard ratio [HR], 1.24; P < .04), CSM (HR, 1.60; P < .02), and OM (HR, 1.45; P < .03). Conversely, this effect disappears with postoperative BT (all P > .2). CONCLUSION: Our study confirms that intraoperative, but not postoperative BT, are related to a detrimental effect on survival after RC. These results should be take into account by physicians to administer BT using the correct timing.

Blood loss, predictors of bleeding, transfusion practice and strategies of blood cell salvaging during liver transplantation.

Blood loss during liver transplantation (OLTx) is a common consequence of pre-existing abnormalities of the hemostatic system, portal hypertension with multiple collateral vessels, portal vein thrombosis, previous abdominal surgery, splenomegaly, and poor "functional" recovery of the new liver. The intrinsic coagulopathic features of end stage cirrhosis along with surgical technical difficulties make transfusion-free liver transplantation a major challenge, and, despite the improvements in understanding of intraoperative coagulation profiles and strategies to control blood loss, the requirements for blood or blood products remains high. The impact of blood transfusion has been shown to be significant and independent of other well-known predictors of posttransplant-outcome. Negative effects on immunomodulation and an increased risk of postoperative complications and mortality have been repeatedly demonstrated. Isovolemic hemodilution, the extensive utilization of thromboelastogram and the use of autotransfusion devices are among the commonly adopted procedures to limit the amount of blood transfusion. The use of intraoperative blood salvage and autologous blood transfusion should still be considered an important method to reduce the need for allogenic blood and the associated complications. In this article we report on the common preoperative and intraoperative factors contributing to blood loss, intraoperative transfusion practices, anesthesiologic and surgical strategies to prevent blood loss, and on intraoperative blood salvaging techniques and autologous blood transfusion. Even though the advances in surgical technique and anesthetic management, as well as a better understanding of the risk factors, have resulted in a steady decrease in intraoperative bleeding, most patients still bleed extensively. Blood transfusion therapy is still a critical feature during OLTx and various studies have shown a large variability in the use of blood products among different centers and even among individual anesthesiologists within the same center. Unfortunately, despite the large number of OLTx performed each year, there is still paucity of large randomized, multicentre, and controlled studies which indicate how to prevent bleeding, the transfusion needs and thresholds, and the "evidence based" perioperative strategies to reduce the amount of transfusion.

BMI Is an Independent Preoperative Predictor of Intraoperative Transfusion and Postoperative Chest-Tube Output.

Background An increasing obese population in the United States focuses attention on perioperative management of obese and overweight patients. Objective We sought to determine if obesity, determined by body mass index (BMI), was a preoperative indicator of bleeding in coronary artery bypass graft (CABG) surgery as measured by intraoperative packed red blood cell transfusion frequency and 24-hour chest-tube output amount. Methods A retrospective chart review examined 290 consecutive patients undergoing single-surgeon off-pump or on-pump CABG surgery between November 2003 and April 2009. Preoperative variables of age, gender, hematocrit, platelet count, and BMI, chest tube output during the immediate 24-hour postoperative period, and the type of procedure (on-pump vs. off-pump) were analyzed. Logistic regression analysis was used to evaluate the likelihood of intraoperative transfusion. Linear regression analysis was used to evaluate 24-hour chest-tube output. Results Preoperative variables that significantly increased the likelihood of intraoperative transfusions were older age and low hematocrit; a significant decrease in likelihood was found with male gender, overweight BMI, and off-pump procedures. Preoperative variables that significantly increased 24-hour chest-tube output were low hematocrit, high hematocrit, and low platelets while a significant decrease in output was seen with overweight BMI and obese BMI. Conclusion Overweight and obese BMI are significant independent predictors of decreased intraoperative transfusion and decreased postoperative blood loss.

Is close monitoring in the intensive care unit necessary after elective liver resection?

PURPOSE: Many surgical patients are admitted to the intensive care unit (ICU), resulting in an increased demand, and possible waste, of resources. Patients who undergo liver resection are also transferred postoperatively to the ICU. However, this may not be necessary in all cases. This study was designed to assess the necessity of ICU admission. METHODS: The medical records of 313 patients who underwent liver resections, as performed by a single surgeon from March 2000 to December 2010 were retrospectively reviewed. RESULTS: Among 313 patients, 168 patients (53.7%) were treated in the ICU. 148 patients (88.1%) received only observation during the ICU care. The ICU re-admission and intensive medical treatment significantly correlated with major liver resection (odds ratio [OR], 6.481; P = 0.011), and intraoperative transfusions (OR, 7.108; P = 0.016). Patients who underwent major liver resection and intraoperative transfusion were significantly associated with need for mechanical ventilator care, longer postoperative stays in the ICU and the hospital, and hospital mortality. CONCLUSION: Most patients admitted to the ICU after major liver resection just received close monitoring. Even though patients underwent major liver resection, patients without receipt of intraoperative transfusion could be sent to the general ward. Duration of ICU/hospital stay, ventilator care and mortality significantly correlated with major liver resection and intraoperative transfusion. Major liver resection and receipt of intraoperative transfusions should be considered indicators for ICU admission.

Is close monitoring in the intensive care unit necessary after elective liver resection? / 대한외과학회지

PURPOSE: Many surgical patients are admitted to the intensive care unit (ICU), resulting in an increased demand, and possible waste, of resources. Patients who undergo liver resection are also transferred postoperatively to the ICU. However, this may not be necessary in all cases. This study was designed to assess the necessity of ICU admission. METHODS: The medical records of 313 patients who underwent liver resections, as performed by a single surgeon from March 2000 to December 2010 were retrospectively reviewed. RESULTS: Among 313 patients, 168 patients (53.7%) were treated in the ICU. 148 patients (88.1%) received only observation during the ICU care. The ICU re-admission and intensive medical treatment significantly correlated with major liver resection (odds ratio [OR], 6.481; P = 0.011), and intraoperative transfusions (OR, 7.108; P = 0.016). Patients who underwent major liver resection and intraoperative transfusion were significantly associated with need for mechanical ventilator care, longer postoperative stays in the ICU and the hospital, and hospital mortality. CONCLUSION: Most patients admitted to the ICU after major liver resection just received close monitoring. Even though patients underwent major liver resection, patients without receipt of intraoperative transfusion could be sent to the general ward. Duration of ICU/hospital stay, ventilator care and mortality significantly correlated with major liver resection and intraoperative transfusion. Major liver resection and receipt of intraoperative transfusions should be considered indicators for ICU admission.