LAUSR

To the Editor: We read with interest the article by Granja et al (1), published in the recent issue of Critical Care Medicine, on multi-modal characterization of extracorporeal membrane oxygenation (ECMO)-induced coagulopathy (EIC). The authors’ study is one of the first to compare coagulation changes during venoarterial and venovenous ECMO. Some important changes that contribute to excess bleeding during ECMO are highlighted in the study including: reduced platelet aggregation, delayed extrinsic pathway activation represented by an increased international normalized ratio, and mildly reduced Factor XIII activity. Similar changes have been reported in prior studies (2). The authors concluded that coagulopathy should be evaluated using a combination of point-of-care and conventional coagulation tests. However, there were inconsistencies between test results in their study, which highlights the limitations of currently available tests. As an example, Factor XIII activity was lower in venovenous ECMO patients compared with venoarterial ECMO patients (Fig. 4D in [1]). This change theoretically should reduce polymerization of fibrin monomers. However, polymerized fibrin levels measured by FIBTEM amplitude (Fig. 2C in [1]) were actually greater in venovenous ECMO patients. Further, Granja et al (1) did not report on platelet adhesion abnormalities that occur during ECMO. Both the platelet glycoprotein (GP)1bα receptor and von Willebrand factor (VWF) play an important role in localizing platelets to a site of vascular injury. Exposure of blood to high shear stress during ECMO causes platelet GP1bα shedding and loss of large VWF multimers. These changes result in poor platelet adhesion to collagen (3). In one prior study, where blood from venoarterial ECMO patients was evaluated in a flow chamber model, platelet adhesion was severely reduced, even after treatment with VWF concentrate (4). Plasma glycocalicin levels were also elevated. These findings were highly suggestive of GP1bα shedding in addition to VWF multimer loss. The final point of Korevaar et al (1) is that “the article focuses on the comparison between the YEARS algorithm versus the Well’s score without d-dimer is clinically poorly relevant.” We have assessed d-dimer separately, by calculating the discriminative ability of d-dimer alone using the area under the receiver operating characteristic curve (AUROC). Additionally, we have analyzed the significant value of d-dimer (≥ 500) in combination with Wells’ score. Our findings are based on analyzing patients who have already received the gold standard test for PE. In our article (2), we have mentioned the retrospective nature of our study as a limitation and we cannot control clinicians’ decisions in the past. However, the fact that Wells’ score was found to have been applied by our clinicians (2) in only 92 patients (11.6%) contradicts the Korevaar et al’s (1) assumption that Wells’ score was commonly used as a triage tool. Additionally, the number of scans conducted in proportion to the size of our hospital is quite a high number (940 scans assessed for eligibility, Figure 1 of our article (2)), supporting the assumption that all patients with suspected PE underwent CTPA. Comparing the percentage of positive YEARS in the original YEARS study (3) versus our study (2) should include comparing multicenter (3) versus single-centre (2) study and comparing clinical follow up-based assessment (3) versus gold standard CTPA-based assessment (2), respectively. All authors participated in preparing, writing, and revising the response. The study was funded by departmental resources. The authors have disclosed that they do not have any potential conflicts of interest.