Head impact exposure sustained by football players on days of diagnosed concussion

Abstract

Purpose: This study compares the frequency and severity of head impacts sustained by football players on days with and without diagnosed concussion and to identify the sensitivity and specificity of single-impact severity measures to diagnosed injury.

Methods: One thousand two hundred eight players from eight collegiate football teams and six high school football teams wore instrumented helmets to measure head impacts during all team sessions, of which 95 players were diagnosed with concussion. Eight players sustained two injuries and one sustained three, providing 105 injury cases. Measures of head kinematics (peak linear and rotational acceleration, Gadd severity index, head injury criteria (HIC15), and change in head velocity (Δv)) and the number of head impacts sustained by individual players were compared between days with and without diagnosed concussion. Receiver operating characteristic curves were generated to evaluate the sensitivity and specificity of each kinematic measure to diagnosed concussion using only those impacts that directly preceded diagnosis.

Results: Players sustained a higher frequency of impacts and impacts with more severe kinematic properties on days of diagnosed concussion than on days without diagnosed concussion. Forty-five injury cases were immediately diagnosed after head impact. For these cases, peak linear acceleration and HIC15 were most sensitive to immediately diagnosed concussion (area under the curve = 0.983). Peak rotational acceleration was less sensitive to diagnosed injury than all other kinematic measures (P = 0.01), which are derived from linear acceleration (peak linear, HIC15, Gadd severity index, and Δv).

Conclusions: Players sustained more impacts and impacts of higher severity on days of diagnosed concussion than on days without diagnosed concussion. In addition, of historical measures of impact severity, those associated with peak linear acceleration are the best predictors of immediately diagnosed concussion.

Figure 1

Hypotheses tested within this communication are based on a subset of biomechanical and clinical data that was collected as part of a longitudinal study to investigate the biomechanical bases of mild traumatic brain injury. Data reported in this study are derived from the samples highlighted in the above flowchart.

Figure 2

Number of impacts per day (all, > 50^th percentile, > 95^th percentile linear acceleration) for players on days with and without diagnosed concussion.

Figure 3

Receiver Operating Characteristic (ROC) curves indicating the sensitivity and specificity of historical measures of impact severity. A 50% probability line is included to indicate the level of guessing (50 – 50 chance). Peak linear acceleration and HIC₁₅ are the most sensitive single impact measures to immediately diagnosed concussion followed by GSI, change in velocity, and peak rotational acceleration. The horizontal axis is reduced to highlight only the top 90^th percentile of all impact events which is inclusive of most impacts associated with injury.