Simplifying Eligibility Criteria for Clinical Trials in Immune Thrombocytopenia Using the Platelet Count Variability Index

Introduction
Up to 15% of immune thrombocytopenia (ITP) patients are misdiagnosed because there is no unique ITP-specific biomarker and many syndromes can present with thrombocytopenia (Arnold et al, Blood Advances 2017). Errors in diagnosis can result in case mix and diluted treatment effects in clinical trials if non-ITP patients are permitted to enter. Clinical trial eligibility criteria that rely solely on the presence of severe thrombocytopenia may not adequately identify non-ITP patients. Conversely, the useful criterion of a previous response to ITP treatment is either too subjective and prone to recall bias, or too cumbersome to be feasible. The platelet count variability index (PVI) is a simple metric that quantifies platelet count fluctuations over time from 3 or more platelet count values. A high PVI can distinguish ITP patients from non-ITP patients independent of treatment. In this study, we modeled the performance of PVI as an eligibility criterion for ITP clinical trials.

Methods
Using consecutive patients with thrombocytopenia from the McMaster ITP registry, we determined the number of patients that were correctly identified as ITP based on the presence of severe thrombocytopenia plus a high PVI index, compared with patients with severe thrombocytopenia alone (worst-case scenario) or patients with severe thrombocytopenia plus a documented response to prior ITP treatment (best-case scenario). Patients were classified as ITP or non-ITP based on the clinical assessment at the last follow up in the registry. Severe thrombocytopenia was defined as an average platelet count <30 x10⁹/L from 3 platelet counts in a 3-month period and no platelet count >35 x10⁹/L. Previous response to treatment was defined as a platelet count >50 x 10⁹/L within 4 weeks of starting corticosteroids, IVIG or Rh-immune globulin with a baseline platelet count <30 x10⁹/L. We previously reported that the median (IQR) PVI for patients with definite ITP was 11.1 (9.7, 12.7) (Li et al, Blood Advances, 2021). In the context of improving clinical trial eligibility criteria, we prioritized a high positive predictive value (PPV) in our evaluation of PVI to minimize the risk of admitting non-ITP patients into an ITP trial.

Results
We identified 862 thrombocytopenic patients, of whom 501 (58.1%) had ITP and 361 (41.8%) had non-ITP. In the worst-case scenario, eligibility criteria that used severe thrombocytopenia alone identified 129 eligible patients, of whom 31 (24.0%) did not have ITP (PPV = 76.0%). In the best-case scenario, eligibility criteria that used severe thrombocytopenia plus a documented response to previous ITP treatment identified 42 eligible patients, of whom 2 (4.8%) did not have ITP (PPV = 95.2%). In comparison, our new strategy that used severe thrombocytopenia plus PVI >9.7 (25th quartile for definite ITP) identified 70 eligible patients, of whom 6 (8.6%) did not have ITP (PPV = 91.4%).

Conclusion
PVI is an objective metric that can be easily derived from a patient’s previous platelet count values. A high PVI index was comparable to a previous response to ITP treatment in its ability to identify ITP patients and exclude non-ITP patients. PVI may be a useful eligibility criterion for ITP clinical trials.