Potential modifications of the plasmic scoring system for predicting thrombotic thrombocytopenic purpura: Sometimes, less is more

Abstract

IntroductionThrombotic thrombocytopenic purpura (TTP) is a life-threatening occlusive disease of the microcirculation characterized by systemic platelet plugs, organ ischemia, deep thrombocytopenia, and fragmentation of erythrocytes. One of the widely used scoring system to determine the clinical probability of TTP is the PLASMIC scoring system. This study aimed to evaluate the contribution of PLASMIC score modifications to sensitivity and specificity in patients with microangiopathic hemolytic anemia (MAHA) undergoing plasma exchange with a prediagnosis of TTP at our center.Materials and MethodsThe data of patients who were hospitalized with a previous diagnosis of MAHA and TTP and underwent plasma exchange at Bursa Uludag University, Faculty of Medicine, Department of Hematology between January 2000 and January 2022 were retrospectively analyzed.ResultsOverall, 33 patients (including 15 and 18 with and without TTP, respectively) were included in this study. Receiver operating characteristic (ROC) analysis revealed that the area under the curve (AUC) for the original PLASMIC score was 0.985 (95% confidence interval [95% CI]: 0.955-1.000), and AUC for the PLASMIC score without mean corpuscular volume (MCV) was 0.967 (95% CI: 0.910-1.000), which is close to the original AUC. With the removal of MCV from the scoring system, the sensitivity decreased from 100% to 93%, whereas the specificity increased from 33% to 78%.ConclusionsBased on the results of this validation study, removing MCV from the PLASMIC score led to the categorization of eight non-TTP cases in the low-risk category, and this could avoid unnecessary plasma exchange. However, in our study increasing the specificity was at the expense of the sensitivity by missing one patient with this new scoring system without MCV. Further multicenter studies with large sample sizes are required owing to the fact that different parameters may be effective in TTP prediction among different populations.