Validation of the Plasmic Score and Development of a Novel Predictive Model for Thrombotic Thrombocytopenic Purpura: Exploring Socioeconomic Disparities and Outcomes

Background:

Thrombotic thrombocytopenic purpura (TTP) is a rare and life-threatening thrombotic microangiopathy with a mortality rate up to 90% if left untreated. The PLASMIC score is a clinical prediction tool used to estimate the pretest probability of severe ADAMTS13 deficiency. Female sex, black ethnicity, and obesity have been identified as potential risk factors for TTP, however, definitive evidence for this is lacking. Our goal was to externally validate the PLASMIC score at our institution, explore the role of socioeconomic factors on TTP outcomes, and potentially identify a more accurate predictive model for TTP.

Methods:

We conducted a retrospective analysis of patients admitted to our institution in Cleveland, Ohio with suspected TTP between January 2000 and January 2023. Any patient who had an ADAMTS13 test ordered was included in the study. The PLASMIC score was retrospectively calculated based on admission labs and correlated with ADAMTS13 levels. Logistic regression analysis was performed to identify potential predictors of TTP with effect measured by odds ratio (OR) and its 95% confidence interval (CI). The performance of 5 predictive models was summarized using the area under the curve (AUC) and compared using Delong's test (Figures 1.1&1.2). We also examined the relationship between age, race, average income, and underlying chronic diseases with TTP outcomes including length of hospital stay, number of blood transfusions, treatment options, lowest hemoglobin level during admission, and lowest platelet count during admission. Overall survival (OS) was calculated from admission to death or last follow-up for surviving patients and was evaluated using Kaplan-Meier method and Cox model.

Results:

We identified a cohort of 47 patients who underwent ADAMTS13 testing. The average age was 50 years old with 62% females and 44.7% identifying as African American. A total of 47% of patients were diagnosed with TTP and had ADAMTS13 levels below 10%. Our analysis revealed that the PLASMIC score was predictive of TTP with OR = 3.05 (P=0.003). When the PLASMIC score was dichotomized into a low-intermediate range (0-5) and a high range (6-7), the model predicted severe ADAMTS13 deficiency (at cutoff point -0.1088028) with a positive predictive value of 75%, negative predictive value of 69.6%, sensitivity of 72%, and specificity of 72.7%. Among the five predictive models compared (Figures 1.1 & 1.2), the final multivariable model that included sex and four components of the PLASMIC score (Active Cancer, Mean Corpuscular Volume, International Normalized Ratio, and Creatinine) was found to be the most accurate in predicting TTP, with an AUC 0.88. In patients with TTP, 94% received plasma exchange and 100% received systemic steroids as treatment. Only 6 TTP patients (31.5%) received rituximab, and 2 patients (10.5%) received intravenous immunoglobulin. The use of multiple treatment options was significantly higher in older patients (P=0.017), obesity (P=0.002), and those with higher average income (per 1% increase, P=0.006). The length of hospital stay was associated with older age (P=0.015), hyperlipidemia (P=0.037), and the presence of underlying autoimmune disease (P=0.009). There was no statistically significant correlation between race and any of the adverse outcomes. Patients diagnosed with TTP had a significantly better OS compared to patients without TTP (P=0.011). The OS of patients who underwent plasmapheresis was slightly better, although the difference was not statistically significant (P=0.067).

Conclusion:

Our study validates the original PLASMIC score as a reliable predictor of severe ADAMTS13 deficiency. However, we propose a new multivariable model that incorporates sex and four components of the original PLASMIC score (Active Cancer, Mean Corpuscular Volume, International Normalized Ratio, and Creatinine) as a better predictor for TTP. Further multicenter studies with larger sample sizes are necessary to validate this new model. We also identified socioeconomic disparities, such as patients with higher income who received multiple lines of treatment compared to lower income patients. Notably, patients diagnosed with TTP had lower mortality possibly due to prompt diagnosis and effective management at our institution and/or the presence of more severe alternative diagnoses in the non-TTP group.