AUTHORS
Dunia Hatabah, MD, Satheesh Chonat, MD2, Evan Orenstein, MD3, Marianne Elaine McPherson Yee, MD, MSc4, Ross M. Fasano I, MD5
AFFILIATIONS
1 Department of Pediatrics, Emory University School of Medicine, Atlanta, Georgia, United States
2 Aflac Cancer and Blood Disorders Center, Children's Healthcare of Atlanta, Atlanta, Georgia, United States, Department of Pediatrics, Emory University School of Medicine, Atlanta, Georgia, United States
3 Children's Healthcare of Atlanta, Atlanta, Georgia, United States, Department of Pediatrics, Emory University School of Medicine, Atlanta, Georgia, United States
4 Aflac Cancer and Blood Disorders Center, Children's Healthcare of Atlanta, Atlanta, Georgia, United States, Department of Pediatrics, Emory University School of Medicine, Atlanta, Georgia, United States
5 Aflac Cancer and Blood Disorders Center, Children's Healthcare of Atlanta, Atlanta, Georgia, United States, Emory University School of Medicine, Atlanta, Georgia, United States
ABSTRACT
Background: Red blood cell (RBC) transfusions play a central role in the management of sickle cell disease (SCD). Delayed hemolytic transfusion reaction (DHTR) a life threatening complication of RBC transfusions is more common in patients with SCD compared to the general population. Diagnosis is challenging because symptoms overlap with vaso-occlusive episodes, and serologic testing may not only be complicated or negative, but time consuming leading to delay in correct diagnosis and treatment.
Objective: Evaluate the implementation of an electronic medical record-integrated clinical decision support tool (EMR-CDST) for DHTR, and utilitizing generated data to assess diagnostic performance of an established nomogram for the diagnosis of DHTR in children with SCD.
Methods: A 2-phase EMR-CDST was implemented across Children’s Healthcare of Atlanta for children with SCD who received inpatient RBC transfusion and presented to the emergency department (ED) within 21 days of transfusion. In phase 1, a best practice alert (BPA) prompted inpatient providers to obtain post-transfusion hemoglobin electrophoresis (HbEP) and hemoglobin (Hb) testing. In phase 2, a BPA alerted ED providers to evaluate for possible DHTR if presenting within 21 days of transfusion and obtain: HbEP, Hb, absolute reticulocyte count (ARC), urinalysis, lactate dehydrogenase (LDH), direct antiglobulin test. The primary process measure was the number of encounters with hemoglobin A (HbA) values at 2 post-transfusion time points, enabling estimation of DHTR likelihood using a previously published nomogram of HbA decline over time (Dessap et al. 2016). Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated to assess the diagnostic accuracy of the nomogram. DHTR diagnostic criteria from our institutional clinical practice guideline included a diagnosis of SCD, a significant drop of Hb (>2 g/dL) within 28 days of transfusion and ³2 of the following: new RBC alloantibody, hemoglobinuria, reticulocytopenia (ARC< 200 x109/L or ½ of baseline) or reticulocytosis >2x baseline and increased LDH (>1000 U/L or >2x baseline).
Results: Between November 2022-January 2026, 1,258 transfusions triggered the phase 1 BPA, and HbA values were collected in 933 (74%) episodes. Within 21 days, 165/1258 (13%) transfused patients presented to the ED, and 67/165 (41%) had HbEP drawn after the phase 2 BPA. Paired phase 1 and 2 HbA values were obtained in 50 patients (4% of all transfused; 30% of those who re-presented to ED). Of these, 36/50 (72%) were hospitalized. Paired HbA values in 2/50 (4%) met high likelihood for DHTR, one of whom was diagnosed and treated for DHTR. Seven (14%) patients had met intermediate DHTR likelihood, 5 of whom met institutional DHTR criteria and 2 of whom were diagnosed and treated for DHTR. All 3 cases of clinically diagnosed DHTR were negative for new RBC antibodies. Forty-one out of 50 (82%) patients plotted in the low-likelihood zone, and no patients met DHTR criteria,. The nomogram yielded a sensitivity of 100%, specificity of 93%, PPV of 66%, and NPV of 100%.
Conclusion: Collectively, these results indicate that the nomogram is a robust diagnostic tool with excellent discriminatory performance in the pediatric SCD population. Its particularly high negative predictive value suggests that it may be most valuable for confidently excluding DHTR in patients classified as low risk while prompting further evaluation in higher-risk patients. A strength of this tool is the ability to prompt clinicians to consider DHTR, a diagnosis that is frequently challenging and overlooked. Importantly, clinically recognized DHTR occurred despite negative antibody testing, underscoring the limitations of serology alone..Wider implementation with improved adherence to HbA testing has the potential to facilitate earlier diagnosis and treatment of DHTR in patients with SCD.
DISCLOSURES
Dunia Hatabah, MD: No relevant conflicts of interest to declare.