15283-B-Blokker

87 Agreement between MIA and CA: a cross-sectional study 5 and radiological abnormalities, such as acute cellular lung rejection and disseminated intravascular coagulation. The methods here applied for MIA appeared adequate, and are feasible in any larger general hospital. Post-mortem angiography, an important technical advancement of post-mortem imaging, 80,108,133 appears not necessary for establishing COD in most of the cases if the cardiac coil and specialized cardiac protocols are used during PMMR and coronary calcium scores are examined at PMCT. For subclinical arterial stenoses and localizing the origin of bleeding the technique is indispensable. However, the logistics of post- mortem angiography is more demanding than the methods applied in our MIA. Moreover, it is more time-consuming, in particular the placement of the catheters, and requires expertise not normally available in general hospitals. There are a number of potential sources of bias in this study. The teams performing MIA and CA were blinded, however, for the pathologist performing CA the biopsy sites could potentially lead to increased suspicion of pathology in the biopsied areas. A further limitation of this study is the use of consensus COD as gold standard, because errors are not uncommon with CA. 56,57,90 This appeared also true for our study where CA was performed in the daily routine of an academic hospital, whereas MIA was carried out by a dedicated research team. Another argument against using CA as gold standard, is the fact that some diagnoses are more readily made on imaging than with CA. The use of consensus COD as the gold standard prohibited calculation of sensitivity and specificity of the two compared methods. Therefore, only agreement between the MIA and CA for COD, and agreement of the two methods with the consensus COD was calculated. Also, for diagnoses, we did not calculate sensitivity and specificity. Specificity cannot be calculated in a meaningful way, since it directly depends on the total number of diagnoses in the entire study population, due to the fact that the number of true negative diagnoses per case increases when more diagnoses are encountered in the population. We decided not to calculate sensitivity, because without a related specificity it is not relevant. The list of ICDcodeswas composedby theMIA-researcher, basedon the autopsy reports of both methods without input from a pathologist involved in CA. As a precaution, this researcher was not involved in scoring diagnoses. The ICD code list appeared accurate because when scoring the diagnoses for MIA and CA all diagnoses could be coded. This process of objective scoring of all and major diagnoses, using the ICD-list, resulted in a large number of diagnoses, often involved in the same pathological process. It required grouping of the individual major diagnoses to GMD, to adequately describe