Paroxysmal cold hemoglobinuria (PCH) is a rare autoimmune hemolytic anemia often overlooked as a potential etiology of hemolysis, and is challenging to diagnose due to the complicated testing methods required. We performed a systematic review of all reported cases to better assess the clinical, immunohematologic, and therapeutic characteristics of PCH. We systematically analyzed PubMed, Medline, and EMBASE to identify all cases of PCH confirmed by Donath-Landsteiner (DL) testing. Three authors independently screened articles for inclusion, and systematically extracted epidemiologic, clinical, laboratory, treatment, and outcomes data. Discrepancies were adjudicated by a fourth author. We identified 230 cases, with median presentation hemoglobin of 6.5 g/dL (IQR 4.8-9.0 g/dL) and nadir of 5.5 g/dL (IQR 4.4-7.2 g/dL). The most common direct antiglobulin test (DAT) result was the presence of complement and absence of IgG bound to red blood cells, though other findings were observed in one-third of cases. Seventy-one patients had DL antibody class and specificity reported, of which 83.1% were IgG anti-P. The use of corticosteroids is common, though we found no significant difference in the length of hospitalization for patients with and without steroid therapy. Recent reports have highlighted the use of complement inhibitors. Amongst patients with follow-up, 99% (213/216) were alive at the time of reporting. To our knowledge, this represents the largest compilation of PCH cases to date. We discovered contemporary PCH most commonly occurs in children with a preceding viral infection, corticosteroid use is frequent (but potentially ineffective), and DAT results are more disparate than traditionally reported.

Thrombosis in extracorporeal membrane oxygenation (ECMO) circuits remains a frequent complication. We characterize the location, extent, structure, and clinical implications of thrombi in 53 ECMO circuits from 46 pediatric patients. The tubing, pump, and oxygenator were examined for visible thrombi. Representative samples of thrombi were collected for histologic, immunofluorescence, and immunohistochemical analysis. Thrombi were found in 81% of ECMO circuits. The most clinically significant were inflow oxygenator membrane surface thrombi (11% of circuits), arterial tubing thrombi (30%), and venous tubing (26%) or connector thrombi (26%). Oxygenator membrane surface thrombi resulted in rapidly increasing delta pressure across the oxygenator over 1-2 days, oxygenator failure, and circuit replacement. Oxygenator membrane surface thrombi were associated with intravascular venous thrombosis and bacterial infection before starting ECMO. Arterial cannula/tubing thrombi led in one case to aortic and mesenteric artery thrombosis followed by bowel infarction. In 11% of cases, venous tubing thrombi grew large enough to break off and embolize to the pump, resulting in increased hemolysis. Antifibrinolytic therapy during ECMO was associated with an increased risk of pump thromboembolism. Other less clinically significant thrombi included pump axle thrombi with thrombus fragments trapped in the oxygenator (45%), and deep oxygenator membrane thrombi (15%). Examination of ECMO circuits after removal is a useful quality improvement tool that can elucidate the cause of circuit problems, indicate patients at increased risk of thrombosis, and suggest areas for possible improvements.

Objectives: Anticoagulation with unfractionated heparin remains the most common therapy used to prevent circuit thrombosis during extracorporeal membrane oxygenation, but no consensus exists on the optimal method or targets for heparin monitoring. From 2015 to 2018, we switched from monitoring heparin during extracorporeal membrane oxygenation using activated clotting times to anti-Xa heparin activity assays. This study describes the transition from activated clotting time to anti-Xa heparin activity assay monitoring and the associated clinical changes. Design: Retrospective analysis at single institution. Setting: Referral Children’s Hospital. Patients: A total of 145 pediatric patients over 152 extracorporeal membrane oxygenation runs using 206 extracorporeal membrane oxygenation circuits. Interventions: Anticoagulation protocol quality improvement. Measurements and Main Results: From 2015 to 2018, heparin monitoring during extracorporeal membrane oxygenation changed from hourly activated clotting time to anti-Xa heparin activity assay every 6 hours with an associated 75% reduction in the circuit changes per extracorporeal membrane oxygenation day. Over the 4 years, patients with an average anti-Xa heparin activity assay of at least 0.25 U/mL showed a 59% reduction in circuit changes per extracorporeal membrane oxygenation day compared with less than 0.15 U/mL. In addition to its association with reduced circuit changes, anti-Xa heparin activity assay monitoring was also associated with reduced heparin dose changes per day from 11 ± 4 to 2 ± 1 (p < 0.001), smaller heparin dose changes (less variation in dose), and reduced diagnostic phlebotomy volumes from 41 ± 6 to 25 ± 11 mL/day (p < 0.001). The number of patients with reported bleeding decreased from 69% using activated clotting time to 51% (p = 0.03). Transfusion rates did not change. Conclusions: Over 4 years, we replaced the activated clotting time assay with the anti-Xa heparin activity assay for heparin monitoring during extracorporeal membrane oxygenation. Minimum anti-Xa heparin activity assay levels of 0.25 U/mL were associated with reduced circuit changes. Further studies are needed to determine the optimum anti-Xa heparin activity assay therapeutic range during extracorporeal membrane oxygenation.

Background Red blood cell (RBC) alloimmunization can occur secondary to transfusion or pregnancy. It is observed most frequently among patients with hemoglobinopathies and myeloid neoplasms. Although previous antigen exposure is generally required for alloimmunization, some alloantibodies may develop naturally without prior exposure. Other alloantibodies may become evanescent, only to reemerge at a detectable titer following a stimulatory event. In a minute fraction of cases, ‘non‐naturally occurring’ alloantibodies may appear without a known antigenic stimulus. Methods and Materials All testing (antibody detection tests and identification, antigen phenotyping, and crossmatching) was performed using the same method and reagents, but occurred at two hospitals within the Yale New Haven Hospital delivery network, and was performed by technologists utilizing different instruments and reagent lots. Results We present two cases of seemingly de novo alloimmunization (anti‐E and anti‐K), and one case of re‐emergence of a known, previously evanescent alloantibody (anti‐K) following transfusion of RBCs that were antigen‐negative for the corresponding antibodies. Conclusion While the exact mechanism underlying the development and/or re‐emergence of RBC alloantibodies in the absence of antigenic stimulation remains unclear, these cases highlight this unusual phenomenon, underscoring the general immunogenicity, as well as the potential consequences, of RBC transfusion and reiterates the importance of concluding an alloantibody specificity, even in the absence of known transfusion of RBCs with a particular antigen.