A predictive algorithm involving lactate dehydrogenase to serum creatinine ratio may assist in identifying patients with thrombotic thrombocytopenic purpura.

Thrombotic thrombocytopenic purpura (TTP) is a rare but fatal disease requiring urgent diagnosis. The PLASMIC scoring model has been reported to be a valuable model for identifying TTP. This study aimed to investigate the diagnostic accuracy of this model and the diagnostic utility of lactate dehydrogenase (LDH)-to-serum creatinine (sCr) ratio in identifying TTP in Chinese patients. The records of 61 patients with suspected TTP tested for ADAMTS13 activity in our hospital between June 2016 and April 2024 were retrospectively analyzed. Receiver operating characteristic (ROC) analysis was performed to evaluate the diagnostic accuracy of LDH, sCr, and the LDH-to-sCr (LDH/sCr) in predicting TTP. Multivariate logistic regression analysis was used to screen for independent risk factors of TTP, and a combined predictive algorithm was established. The AUC for the PLASMIC scoring model distinguishing TTP from non-TTP was 0.850 (optimal cutoff: 5), with 92.8% sensitivity, 76.5% specificity, 75.8% positive predictive value (PPV), and 92.9% negative predictive value (NPV). The AUC derived from the LDH/sCr ratio was higher than that derived from LDH or sCr (P < 0.05). A combined predictive algorithm, termed the LCRP algorithm (including the LDH/sCr ratio, reticulocyte percentage, and platelet count), was developed for TTP. The AUC (0.955) derived from the LCRP algorithm, with 96.3% sensitivity, 88.2% specificity, 86.7% PPV, and 96.8% NPV, was higher than that of the PLASMIC scoring model (P < 0.05). The LDH/sCr ratio is more effective than isolated LDH or sCr measurements for predicting TTP. The LCRP algorithm involving the LDH/sCr ratio showed superior predictive performance and may hold significant potential for early identification of Chinese patients with TTP.