Sample sizes for clinical trials using sputum eosinophils as a primary outcome

Area of Science:

• Pulmonary Medicine
• Clinical Trial Methodology
• Biostatistics

Background:

• Inconsistent reporting of sputum eosinophil data in clinical trials hinders cross-study comparisons and sample size evaluation.
• Sputum eosinophil counts are crucial biomarkers in managing airway inflammatory diseases like asthma and COPD.

Purpose of the Study:

• To systematically review the reporting of effect size and sample size calculations in randomized controlled trials (RCTs) using sputum eosinophil count as a primary outcome.
• To demonstrate sample size estimation using different data representation methods with mepolizumab treatment data.

Main Methods:

• Systematic literature search of Ovid MEDLINE for RCTs in adults with asthma or COPD over the past 10 years, focusing on sputum eosinophil count outcomes.
• Analysis of reporting practices for sample size calculations and effect size representation in identified studies.
• Sample size estimation using simulated data based on an anti-eosinophil treatment (mepolizumab).

Main Results:

• Of 20 identified RCTs, only nine reported sample size calculations.
• Effect sizes were reported inconsistently, including absolute percentage change, percentage change from baseline, and fold changes.
• An estimated 18 subjects per arm are needed for 80% power with ANCOVA to detect a 15% absolute reduction in sputum eosinophils.

Conclusions:

• Standardized reporting of sputum eosinophil data and sample size calculations is essential for robust clinical trial design and interpretation.
• ANCOVA is recommended for analyzing sputum eosinophil data, with specific sample size recommendations provided.
• Improved methodological consistency will enhance the evaluation of anti-eosinophil therapies in respiratory diseases.