Using kinematics to re-define the pull test as a quantitative biomarker of the postural response

Background Quantitative biomarkers are needed for the diagnosis, monitoring and therapeutic assessment of postural instability. Objective Create a practical and objective measure of postural instability using kinematic measurement of the pull test to incorporate into clinical practice. Methods Twenty patients evaluated and treated for normal pressure hydrocephalus were tested over a number of sessions. Patients were fitted with 15 inertial measurement units during each session. At each session, the patient underwent 10-20 pull tests performed by a trained clinician. The clinician purposefully induced a range of perturbations during each session to assess the patient’s step response. Kinematic data was extracted for each pull test and aggregated. Results Patients participated in 57 sessions for a total of 860 trials and were separated into groups by pull test score. The center of mass velocity profile easily distinguished between groups such that score increases correlated with decreases in peak velocity and later peak velocity onset. All patients except those scored as “3” demonstrated an increase in step length and decrease in reaction time with increasing pull intensity. Groups were distinguished by differences in overall step length or reaction time regardless of pull intensity (y-intercept). A logistic regression model including only kinematic variables identified three kinematic factors that could be used to identify trials in which patients needed to be caught (i.e. “fell”). Conclusion An instrumented, purposefully varied pull test produces kinematic metrics useful for distinguishing clinically meaningful differences between and within NPH patients. These metrics should be followed prospectively to examine fall risk.