Development of a preoperative risk score predicting allogeneic red blood cell transfusion in children undergoing spinal fusion

Abstract

Background Children undergoing spinal fusion often receive blood products. The goal of this study was to develop a preoperative score to help physicians identify those who are at risk of allogeneic red blood cell (RBC) transfusion. Study design and methods This retrospective study of children undergoing spinal fusion in the ACS‐NSQIP Pediatric database (2016–2019) aimed at identifying risk factors associated with allogeneic RBC transfusion. Univariable logistic regression and multivariable logistic regression were performed using preoperative patient characteristics and aided in the creation of a simplified scoring system. Results Out of 13,929 total patients, 2990 (21.5%) were transfused. We created a risk score based on 10 independent predictors of transfusion: age, sex, race, weight < 3rd percentile, American Society of Anesthesiologists physical status classification, cardiac risk factors, hematologic disease, preoperative anemia, deformity type, and number of spinal levels to be fused. Patients in both the training and testing cohorts were assigned a score ranging from 0 (lowest risk) to 21 (highest risk). The developed transfusion risk score showed 77% accuracy in distinguishing patients who did not receive a transfusion during or soon after surgery (AUROC 0.7736 [95% CI, 0.7619–0.7852]) in the training cohort and 77% accuracy (AUROC 0.7732 [95% CI, 0.7554–0.7910]) in the testing cohort. Discussion Our score, based on routinely available preoperative data, accurately estimates the risk of allogeneic RBC transfusion in pediatric patients undergoing spinal fusion. Future studies will inform whether patient blood management interventions targeted to high‐risk patients can help reduce the need for transfusion and improve outcomes.

Allogeneic Red Blood Cell Transfusion and Infectious Complications Following Pediatric Spinal Fusion

“…Based on our prior work 44 demonstrating that anemia is an independent predictor of ARBT rate, and the observed association between ARBT and infections, we expected to find a stronger relationship between anemia and infection in our unadjusted analysis. However, we previously showed the impact of anemia on ARBT to be less pronounced in younger patients and in those with neuromuscular scoliosis 44 , groups found in the present study to be at the greatest risk of postoperative infection. Studying a heterogeneous group with disparate risks of infection dictated in large part by deformity type may have thus minimized any impact of preoperative anemia.…”

“…These findings strengthen evidence for an association between ARBT and postoperative infections in children treated with spinal fusion. To the extent that the need for ARBT may be reduced through risk stratification and targeted preventive measures, including the treatment of preoperative anemia and other blood conservation strategies, we hope that further study can contribute to a reduction in postoperative infections and associated morbidity 40 , 44 .…”

“…Covariates predictive of transfusion in previous work and/or deemed by 2 attending physicians (L.E., L.G.L.) to contribute to infectious outcomes were included in the analysis 44 , 45 . Patient age, sex, race, year of procedure, transfer status, American Society of Anesthesiologists (ASA) physical status classification, structural pulmonary abnormality, steroid use, nutritional support (enteral feeding or total parenteral nutrition), neuromuscular spinal deformity, hematological disorder, surgical duration, surgical extent (number of levels fused), cardiac risk factors, and application of antibiotics to the wound were available directly from the NSQIP-P data set.…”

“…Weight below the third percentile for the patient’s age, determined from weight and age data by sex based on a WHO algorithm (z-score), was considered a marker of poor nutritional status and has been previously identified as a predictor of transfusion 43 , 46 . Continuous variables (age, operative time) were categorized based on prior research, and complex categorical variables (e.g., ASA status, number of levels fused) were reduced to fewer categories to simplify analyses 44 , 47 , 48 . Preoperative anemia was defined based on hematocrit, using an algorithm for age- and sex-specific thresholds 47 .…”

Factors affecting need for blood transfusion in paediatric patients undergoing open surgery for hip dysplasia

“…For the DD patients, the mean (SD) length of stay was increased (4.1 [3.5] days without transfusion vs. 4.6 [6.7] days with transfusion; p = 0.49), likely owing to a more extensive surgical procedure. Similarly, multiple studies have identified a decrease in the rate of allogeneic packed red blood cells (PRBC) transfusion in children over time, likely owing to more conservative thresholds in transfusion guidelines [14–17].…”