Selective validation of the WHO Bleeding Scale in patients with chronic immune thrombocytopenia

Objective: To evaluate the World Health Organization's (WHO) Bleeding Scale in two studies of eltrombopag in adults with chronic immune thrombocytopenia (ITP). Research design and methods: Validated scales assessing bleeding in adults with ITP are lacking. Data from two long-term, phase 3 clinical trials (RAISE: NCT00370331; EXTEND: NCT00351468) that assessed eltrombopag in adults with chronic ITP were analyzed to evaluate the performance of the WHO Bleeding Scale. Results: In RAISE, effect size (0.71), standardized response (0.75), and responsiveness statistics (0.57) were moderate for bleeding and bruising assessments. In EXTEND, effect size (0.62) and responsiveness statistics (0.59) were moderate; the standardized response statistic was 0.487. Intraclass correlation for test-retest reliability was 0.75 in RAISE and 0.71 in EXTEND. A positive correlation was observed between the WHO Bleeding Scale and the ITP Bleeding Scale. Bleeding scores and quality-of-life measures were inversely correlated (p<0.05 for all). Minimal important differences for the WHO Bleeding Scale were 0.33-0.40 at baseline and last on-treatment assessment in both studies. Limitations: The majority of bleeding in these studies was mild to moderate, so this analysis cannot provide strong evidence of the validity of the WHO Bleeding Scale in patients with more severe bleeding. Potential limitations to the WHO Bleeding Scale itself include dependence on clinician interpretation of patient recall, inability to distinguish among bleeding events occurring at different anatomical sites, and an inherent assumption of linear increases in severity of bleeding across the response categories. Conclusions: These findings suggest potential usefulness of the WHO Bleeding Scale in adult patients with chronic ITP for standardizing grading of bleeding across research studies and in clinical practice.