Key words: Throwing, shoulder, maximum external rotation, elbow valgus stress, prevention.

Twenty high school baseball players who had experienced medial elbow pain induced only by throwing within the previous month prior to the experiment but were not experiencing the pain on the day of the experiment were recruited as elbow-injured group. Careful consideration was required in this type of selection of the subject. If the subject displayed the elbow pain while being tested, it would be difficult to decide whether the shoulder kinematic data was the cause or the result of the medial elbow pain. In addition, if the pain subsided well before the day of the experiment, we might not be able to identify the pathological throwing mechanics in the result. Therefore, we recruited baseball players who had a recent history of medial elbow pain, but did not have any elbow pain on the day of the experiment as the elbow injured group.

A clinical examination was conducted carefully according to the predetermined procedure. After the history of the elbow injury was asked by an experienced physical therapist, the presence of tenderness and motion pain on the elbow were examined. Joint laxity and any sign of joint abnormality were then evaluated by orthopedic tests. Those who displayed any tenderness, subjective pain during motion (active, passive, or both), and / or abnormal joint laxity on elbow joint in the clinical examination were excluded from the experiment. Three subjects had missed several sessions of baseball practice due to the medial elbow pain, although no structural abnormality was found in a roentgen examination. The average age, height, weight, years of baseball experience were 17.1 ± 0. 6years, 1.70 ± 0.05cm, 63.8 ± 12.7kg, and 7.7 ± 1.6years, respectively. The elbow-injured group consisted of 3 pitchers, 1 catcher, 12 infielders, and 4 outfielders. In addition, another group of twenty high school baseball players who had never experienced any medial elbow pain was recruited as the control group. The average age, height, weight, and years of baseball experience in the control group were 17.0 ± 0.7years, 1.70 ± 0.07cm, 62.2 ± 7.4kg, and 7.7 ± 2.4years, respectively. The control group consisted of 4 pitchers, 1 catcher, 8 infielders, and 7 outfielders. Prior to the experiment, all subjects signed their names on the informed consent form approved by the Hiroshima University Ethics Committee.

ER was measured bilaterally by a 2D-3D motion analyzer (Frame- DIAS II, DKH, Tokyo, Japan). Subjects were placed in a sitting position with shoulder abduction and elbow flexion angle at 90º. The standard manual muscle testing manoeuvre for ER is used in a supine position. However, since the posture in throwing was an upright position, the sitting position must have simulated actual throwing motions more accurately. After the shoulder was passively moved until the maximum position, the measurement angle was determined by one experienced physical therapist. Excellent intra-tester repeatability was confirmed with intraclass correlate coefficient (ICC) (ICC = 0.97). The shoulder position was captured by a digital video camera and was transferred to a personal computer for the further analysis. During the ER measurement, the passive force of 20 Newton (N) was applied to the distal end of forearm. The magnitude of the passive force was monitored in a hand-held dynamometer (Micro FETII, Hoggan Health Ind. Inc., USA). Lumbar extension was not observed during the measurement.

Data collection for the throwing was performed on an outdoor baseball field. After a warm-up period, the subject executed five throwing trials at his maximum effort. The throwing distance, throwing target, outfit, and equipment used in this experiment were selected to simulate real practices or games. A throwing distance of 27.43m, the regulation distance between bases, was used. The throwing target was a baseball glove held in front of the catcher's chest. In addition, the subject held a glove on his non-throwing hand and wore a pair of baseball uniform pants, spiked shoes, and a sleeveless shirt with a hole on the back for marker placement during the testing.

Reflective markers were placed at the bony landmarks of subjects' upper body: acromion process, the dorsal side of the distal end of the humerus and forearms, and the spinous process of the 8^th thoracic vertebrae (Figure 1).

Direct Linear Transformation (DLT) method (Abdel-Aziz and Karara, 1971) was performed to establish three-dimensional (3D) coordinates of the shoulder complex in throwing with a 2D-3D motion analyzer (Frame-DIAS II, DKH, Tokyo, Japan). The throwing attempt in which the ball was the most accurately controlled to the target was used for the analysis. Two high-speed video cameras (HSV-400, NAC Image Technology, Inc., Tokyo, Japan) were used to collect the throwing motion data sampled at 200Hz. These cameras were located to the right and left rear of the subject. Throwing motion data for the three throws was transferred to a personal computer for marker digitizing. The reflective makers were automatically tracked by the motion analyzer.

The MER was determined by a kinematic model with two segments. One was described by lines between the markers on dorsal side of distal forearm (wrist marker), elbow joint (humerus marker) and acromion process of the throwing shoulder (shoulder marker), and the other was described by lines between markers on the humerus marker, the shoulder marker, and the spinous process of the 8th thoracic vertebrae (Th 8 marker) (Figure 1a). First, two normal unit vectors projected from the two segments were computed. Inner product between the normal unit vectors and their cosine angle was then defined as the external rotation angle of the shoulder (Figure 1c). Our MER computation method allows us to obtain a net external rotation angle of shoulder complex without an influence of thoracic and lumber movements. Shoulder external rotation angle where the planes intersected at an angle of 90º was defined as zero degree. The direction of external rotation was expressed as a positive value in this study (Figure 1c).

The ratio of MER to ER was calculated, which was expressed as MER divided by the ER. This proposed index was designed to assess the magnitude of the valgus stress imposed on the elbow.

A Mann-Whitney test was performed to compare ER on throwing shoulder, MER in throwing, and the ratio of MER to ER between the two groups. In order to determine whether the degree of ER represented inherent or acquired characteristics of the subjects, ER on non-throwing shoulder was also compared between the two groups. We used commercial statistical ad-in software (Statcel, OMS, Japan) for the statistical analysis. Statistical significance was defined as an alpha level of 0 .05 in this study.