Clinician-reported lesion measurements in skin infection trials: Definitions, reliability, and association with patient-reported pain

Outcome assessments as clinical trial endpoints should be well-defined, reliable, and reflect meaningful treatment benefits. For acute bacterial skin and skin structure infections (ABSSSI) trials, recent recommendations suggest a primary endpoint of reduction in skin lesion area. Objectives were: evaluate ABSSSI lesion area measurement reliability, evaluate impact of various lesion area definitions on treatment effect size, and explore relationships between lesion area and pain.Data from two randomized, double-blinded Phase 3 trials comparing tedizolid to linezolid in ABSSSI and one open-label, non-comparative Phase 2 study of tedizolid in cellulitis/erysipelas and skin abscess were analyzed. Repeated lesion area measurements were prospectively obtained in all studies. In the open-label study, lesion area was measured by two investigators, using four different definitions. Repeated pain assessments using two patient-reported outcome instruments (Visual Analog Scale [VAS] and Faces Rating Scale [FRS]) were elicited in the randomized trials.At baseline, lesion size did not correlate with pain intensity: r=0.02 for VAS and r0.01 for FRS pain scores. However, decreasing lesion size and decreasing pain were strongly associated over time, regardless of initial lesion size or pain intensity (r=0.20 for VAS and r=0.21 for FRS scores at Day 10-13). Each lesion area definition demonstrated high inter-observer reliability (intra-class correlation coefficient0.95).Decreasing lesion area (indirect clinician-reported measure of benefit) and pain (direct patient-reported measure of benefit) were strongly associated over time, and lesion area measurements were reliable, regardless of their definition. These findings support both measures as outcome assessments in ABSSSI clinical trials.Clinicaltrials.govNCT01519778, NCT01170221, and NCT01421511.