Determining Real Change in Conditioned Pain Modulation: A Repeated Measures Study in Healthy Volunteers

Abstract

Conditioned pain modulation (CPM) is a potentially useful biomarker in pain populations; however, a statistically robust interpretation of change scores is required. Currently, reporting of CPM does not consider measurement error. Hence, the magnitude of change representing a “true” CPM effect is unknown. This study determined the standard error of measurement (SEM) and proportion of healthy participants showing a “true” CPM effect with a standard CPM paradigm. Fifty healthy volunteers participated in an intersession reliability study using pressure pain threshold (PPT) test stimulus and contact heat, cold water, and sham conditioning stimuli. Baseline PPTs were used to calculate SEM and >±2 × SEM to determine CPM effect. SEM for PPT was .21 kg/cm2. An inhibitory CPM effect (>+2 SEM) was elicited in 59% of subjects in response to cold stimulus; in 44% to heat. Intrasession and intersession reliability of within-subject CPM response was poor (kappa coefficient <.36). Measurement error is important in determining CPM effect and change over time. Even when using reliable test stimuli, and incorporating measures to limit bias and error, CPM intersession reliability was fair and demonstrated a large degree of within-subject variation. Determining “true” change in CPM will underpin future interrogations of intraindividual differences in CPM. PerspectiveThis study used a distribution-based statistical approach to identify real change in CPM, based on the SEM for the test stimulus. Healthy volunteers demonstrate substantial within-subject variation; CPM effect was paradigm dependent at intrasession testing and unstable to the same paradigm at intersession testing.