Background: Peripheral blood lymphocyte apoptosis is a recognized feature of serious infection and sepsis and can be easily quantified by flow cytometric measurement of annexin V binding to the cell surface. Use of apoptosis as a biomarker in emergency department (ED) studies of sepsis is potentially difficult because of sample processing requirements and limited availability of a research cytometer with which to measure patient samples. Objectives: To assess, in vitro and in simulation, the relationship between sample stability, timing of patient enrollment, and diagnostic performance of a flow cytometric assay for sepsis in patients evaluated in EDs. Methods: Assuming any clinical trial would require daily sample batching, the authors measured the stability of lymphocyte samples over time, noting the rate at which annexin V-negative cells became positive as ED processing delays increased. With these data, they then optimized a study design that could evaluate lymphocyte apoptosis as a sepsis biomarker by using a series of Monte Carlo-based simulated clinical trials. Results: The authors found that annexin V-negative lymphocytes become positive during storage delays that would be encountered in an ED sepsis trial. The extent of this deterioration was least among cells left as whole blood at room temperature until just before analysis or when lymphocytes were isolated early and stored in culture media at 4°C until analysis. When the expected rate of sample deterioration was considered in simulated clinical trials, an inverse relationship was found between the rate at which patients are enrolled and the best achievable receiver operating characteristic curve a study could produce. Conclusions: Peripheral blood samples being analyzed for lymphocyte apoptosis degrade at a rate relevant to the design of ED trials of sepsis. Because of sample processing delays inherent in studying unscheduled septic patients, the performance of annexin V binding as a biomarker for sepsis can approach, but not be expected to exceed, its performance in a comparable intensive care unit-based study. © 2007 Society for Academic Emergency Medicine.