Thorac Cardiovasc Surg 2013; 61(08): 744-746
DOI: 10.1055/s-0033-1353533
Letter to the Editor
Georg Thieme Verlag KG Stuttgart · New York

Multiple Electrode Aggregometry and Prediction of Bleeding and Transfusion Outcomes in Adult Cardiac Surgery Patients: Methodological Challenges and Opportunities for Future

Mate Petricevic
1  Department of Cardiac Surgery, University of Zagreb School of Medicine, Kispaticeva 12, 10000 Zagreb, Croatia
,
Bojan Biocina
2  Department of Cardiac Surgery, School of Medicine University of Zagreb, Zagreb, Croatia
,
Davor Milicic
3  Department of Cardiovascular Diseases, School of Medicine University of Zagreb, Zagreb, Croatia
,
Ante Lekic
2  Department of Cardiac Surgery, School of Medicine University of Zagreb, Zagreb, Croatia
,
Ivica Safradin
2  Department of Cardiac Surgery, School of Medicine University of Zagreb, Zagreb, Croatia
,
Hrvoje Gasparovic
2  Department of Cardiac Surgery, School of Medicine University of Zagreb, Zagreb, Croatia
› Author Affiliations
Further Information

Publication History

05 July 2013

10 July 2013

Publication Date:
21 November 2013 (online)

We read with great interest the recently published study by Schimmer et al.[1] The purpose of the retrospective study was to assess whether multiple electrode aggregometry (MEA) can predict bleeding as well as transfusion outcomes in group of patients undergoing cardiac surgery.[1]

When assessing MEA as a predictor of bleeding and transfusion outcomes, some methodological considerations should be addressed inevitably. This study was retrospective, and both preoperative and postoperative pathologic MEA results served as transfusion triggers in the platelet concentrate (PC) transfusion management.[1] Because PC transfusion inevitably influences bleeding extent as well as transfusion outcomes per se, correlations between bleeding and transfusion outcomes are distorted (biased) by inclusion of MEA in transfusion algorithm. This could explain nonsignificant differences among groups with respect to the bleeding extent as well as re-exploration for bleeding proportions. In addition, the authors used predefined cutoff values for different MEA tests.[1] Because those values are not specifically validated in setting of cardiosurgical patients, sensitivity and specificity are probably underestimated, below the optimal values. Prospective observational study with attending physicians blinded to MEA results would provide more precise and reliable results on prediction of bleeding and transfusion outcomes. Such a setting provides correlations that are not distorted by inclusion of MEA in regular clinical practice as a part of decision-making process. Using receiver operating curve (ROC) analysis, we recently delineated MEA test values that depict excessive bleeding in group of patients undergoing coronary artery surgery.[2] In addition, we also published the article that addressed the issue of excessive bleeding prediction using concomitantly MEA and rotational thromboelastometry (ROTEM).[3] In prospective observational (clinicians blinded) setting, we defined values that have the best sensitivity and specificity for prediction of excessive bleeding.[2] [3]

In this study by Schimmer et al, 16.1% of patients required emergency operations.[1] The proportion of patients with preoperative exposure to dual antiplatelet therapy (dAPT) regime remains unclear in emergency subgroup. It would be interesting to see, if patients exposed to dAPT have had pronounced platelet inhibition, especially those who were exposed to loading dose of adenosine diphosphate (ADP)-antagonists before surgery (i.e., previously considered as PCI treatment failure). Recent exposure to loading dose of ADP-antagonists could explain to some degree the fact that despite higher proportions of patients transfused with PCs, the postoperative ADP test values remained quite similar in Group 1 (preoperative ADP 306.4 vs. postoperative ADP 309.1). Furthermore, when analyzing the relation between platelets dysfunction and need for re-exploration due to excessive bleeding, patients with detected obvious surgical cause of bleeding should, in our opinion, be excluded. Inclusion of patients who were reoperated for excessive bleeding due to surgical cause lacks rationale in assessment of relatively infrequent outcome prediction using tests that detect hemostatic disorder. Finally, the heterogeneity of the observed patient population makes it somehow difficult to exclude the effect of the complexity of the cardiosurgical procedure and the duration of cardiopulmonary bypass on postoperative bleeding and transfusion outcomes because they certainly influence hemostatic blood properties in way other than solely platelet function.

We congratulate authors on their timely, elegant, and valuable contribution[1] to the current knowledge. Strategies to prevent bleeding and transfusion outcomes are essential for the successful management of patients, however, require comprehensive approach. Our recent study supports the authors' statement that it seems to be reasonable to use hemostatic properties optimization algorithms in cardiac surgery based on both MEA and ROTEM.[3] However, novel hemostatic approaches that should be evaluated through randomized trials should be targeted and triggered after cutoff values that are obtained through prospective observational studies in which we have specific cutoffs that are obtained through ROC analysis with the best sensitivity/specificity ratio as we clearly showed recently.[3] Therefore, we call for multicentric prospective randomized trial that will evaluate the role of targeted hemostatic management on the basis of cutoffs that are obtained and validated using ROC analysis in specific cohort of patients undergoing cardiac surgery.[2] [3] Such methodological approach would enable us to take advantage of MEA and ROTEM to the greatest extent providing results with minimum risks of limitations that require cautious results interpretation.