# In patients with suspected thrombotic thrombocytopenic purpura, what is the optimal time to therapeutic plasma exchange?

## TL;DR

This study examines the best time to perform plasma exchange in patients with suspected thrombotic thrombocytopenic purpura to improve outcomes.

## Contribution

The study identifies 13.5 hours as the optimal time for plasma exchange, challenging the existing eight-hour guideline.

## Key findings

- The optimal time for plasma exchange was found to be 13.5 hours after symptom onset.
- Time to plasma exchange had a low predictive capacity for clinical outcomes (area under the curve: 0.62).
- The relationship between time to plasma exchange and outcomes was non-linear.

## Abstract

In patients with suspected immune thrombotic thrombocytopenic purpura, guidelines suggest that therapeutic plasma exchange should be initiated within eight hours. However, this time threshold may be difficult to attain. This study sought to identify the optimal time to plasma exchange to maximize outcomes.

Patients with international classification of disease codes for thrombotic microangiopathy were identified in a retrospective cross-sectional analysis of public use data files from the Recipient Epidemiology and Donor Evaluation Study-III (REDS-III). The assumption of linearity between time to therapeutic plasma exchange and the composite outcome of bleeding, thrombosis, and mortality were evaluated. Subsequently, the optimal time for plasma exchange was identified using a nonparametric approach with bootstrapping.

For 149 patients with a suspected diagnosis of thrombotic thrombocytopenic purpura, the association between time to plasma exchange and the primary outcome was non-linear. With regard to the primary composite outcome, this time had a low predictive capacity (area under the curve: 0.62). The optimal time that maximized outcomes was 13.5 h.

Although this study found that time to therapeutic plasma exchange did not independently predict outcome, future studies might evaluate how this time interacts with other variables to predict clinical outcomes.