Biomechanics of Windmill Softball Pitching With Implications About Injury Mechanisms at the Shoulder and Elbow

Biomechanics of Windmill Softball Pitching With Implications About Injury Mechanisms at the Shoulder and Elbow Steven W. Barrentine, MS ' Glenn S. Fleisig, Ph D * james A. Whiteside, M D Rafael F. Escamilla, PhD james R. Andrews, MD 0 veruse injuries caused by the overhand throwing motion in men's baseball pitching have been welldocumented (3,13,16,17,25). To address this problem, several studies have investigated the kinematics and kinetics during overhand pitching (5,6,8,10,19,20,24). Forces produced at the shoulder and elbow during baseball pitching are equivalent to 80-120% of body weight (BW) for compressive forces and 30 to 45% BW for shear forces (8,lO-12,24). Torques exerted at the shoulder and elbow to generate the high velocities during throwing are equivalent to 3-7% of body weight multiplied by height (BW X HT). Critical instances of the overhand pitching motion have been identified and related to injury mechanisms for rotator cuff tensile failure, subacromial impingement, and anterosuperior labrum tear at the shoulder (10). Underhand pitching has received little attention in the sports medicine literature. It has been noted that the limited attention may be due to the perception that the underhand motion creates less stress on the arm and, thus, fewer injuries occur (14). Using a survey of eight collegiate softball teams participating in the 1989 Underhand pitching has received minimal attention in the sports medicine literature. This may be due to the perception that, compared with overhead pitching, the underhand motion creates less stress on the arm, which results in fewer injuries. The purpose of this study was to calculate kinematic and kinetic parameters for the pitching motion used in fast pitch sofiball. Eight female fast pitch soflball pitchers were recorded with a four-camera system (200 Hz). The results indicated that high forces and torques were experienced at the shoulder and elbow during the delivery phase. Peak compressive forces at the elbow and shoulder equal to 70-98% of body weight were produced. Shoulder extension and abduction torques equal to 940% of body weight x height were calculated. Elbow flexion torque was exerted to control elbow extension and initiate elbow flexion. The demand on the biceps labrum complex to simultaneously resist glenohumeral distraction and produce elbow flexion makes this structure susceptible to overuse injury. Key Words: biomechanics, underhand pitching, softball, shoulder, elbow ' Biomechanist, American Sports Medicine Institute, Birmingham, AL Smith & Nephew Chair of Research, American Sports Medicine Institute, 1313 13th Street South, Birmingham, A1 35205 ' Eminent Scholar of Sports Medicine; Professor, College of Health and Human Sciences, Troy State University, Troy, A1 Assistant Professor, Orthopaedic Surgery, Duke University Medical Center, Durham, NC Medical Director, American Sports Medicine Institute, Birmingham, AL NCAA softball tournament, Loosli et a1 (14) attempted to quantify the frequency of injury to underhand pitchers. Twenty of the 24 pitchers that participated in the survey reported 26 injuries, with 17 of these injuries involving the upper extremity. Eightytwo percent of all time-loss injuries (grade I1 and 111) involved the upper extremity. Almost one-half of all timeloss injuries (five of 11) involved injuries to the shoulder and elbow, including bicipital and rotator cuff tendinitis and strain. It was con- cluded that softball pitchers experience significant time-loss injuries as a direct result of the underhand pitching motion. Injuries to softball pitchers have prompted investigations of the underhand pitching motion. Based on three reported cases of fatigue fractures of the ulna, Tanabe et a1 (21) used high-speed cinematography and computed tomography scans of six collegiate pitchers (three males, three females) to analyze the underhand pitching motion and the result- JOSFT - Volume 28 * Number 6 - December 1998

RESEARCH STUDY -... - -.......... -.. -. -...... - - FIGURE 1. A view of laboratory set-up with a softball pitcher during testing. ing effects on the ulna. The authors attributed the fractures to torsional stress exerted on the forearm as the arm pronated after release of the ball (21). The occurrence of injury to softball pitchers has also prompted the investigation of muscle firing patterns during the windmill fast pitch motion (15). The electromyographic activity of eight shoulder muscles for 10 collegiate pitchers were analyzed, and contributions to joint stabilization and arm velocity were determined. The dearth of research studies investigating the underhand pitching motion has limited the understanding of the mechanics and the stresses that result on the shoulder and elbow. Based on the results reported by Loosli et al, it is hypothesized that the stresses on the shoulder and elbow generated during underhand pitching are similar to the stresses experienced during overhand throwing. The purpose of this study was to calculate kinematic and kinetic parameters at the shoulder and elbow that occur during underhand windmill pitching in women's fast pitch softball. These results will be compared to values calculated for over- hand throwing, with emphasis on the forces and torques experienced during the acceleration and deceleration phases. Critical instances of the pitching motion will be determined and used to investigate proposed mechanisms of overuse injuries. METHODS Eight healthy female pitchers were used as subjects. A pitcher was considered healthy if she met three criteria: I) she was not currently injured or recovering from an injury at the time of testing; 2) she had not undergone surgery for at least 12 months prior; and 3) she felt that she was able to pitch with the same intensity as she would in a game environment. Six of the subjects were collegiate pitchers, and two subjects were former collegiate pitchers (competitive on semi-professional teams at the time of data collection). The average age was 21 (SD = 4) years, the average weight was 65 (SD = 5) kg, and the average height was 1.73 (SD = 0.08) m. A brief questionnaire was completed before the testing session concerning the medical history, pitching background, and the cur- rent physical fitness level. Written consent was obtained from each sub ject prior to testing. Height, weight, and length of the radius and humerus of the throwing arm were measured. The length of the humerus was measured from the lateral tip of the acromion to the lateral humeral epicondyle. The length of the radius was measured from the lateral humeral epicondyle to the radial styloid process. Each subject was then instructed to perform her normal warm-up routine that included stretching, throwing, and additional nonthrowing drills. Data collection and analyses consisted of a procedure similar to the method previously described (10,ll). Spherical (3.8 cm in diameter) reflective markers were used to identify anatomical landmarks for digitization. Markers were placed bilaterally on the distal end of the midtoe, lateral malleolus, lateral femoral epicondyle, greater trochanter, lateral tip of the acromion, and lateral humeral epicondyle. A reflective band was wrapped around the wrist on the throwing arm, and a reflective marker was attached to the ulnar styloid of the nonthrowing arm. A reflective band was also attached to the softball to determine the location of the ball. The testing set-up is shown in Figure 1. Ten trials were collected for each pitcher when throwing fastball pitches. All subjects threw to a strike zone net located behind a home plate placed 12.19 m from the pitching rubber. A threedimensional, automatic digitizing system (Motion Analysis Corporation, Santa Rosa, CA) was used to digitize the location of the reflective markers. Four electronically synchronized 200 Hz charged couple device cameras transmitted pixel images of the reflective markers directly into a video processor without being recorded onto video. Threedimensional marker locations were calculated with Motion Analysis Expertvision three-dimensional software that utilized the direct linear transformation method (1). Volume 28 Number 6 December 1998 JOSPT

RESEARCH STUDY FIGURE 2. Sequence of motion in windmill pitching: a-c) wind up, d-0 stride, g-j) delivery, k-i) follow through. Parameter Underhand Overhand 2 SD Range Stride Linear velocity of hips (dsec) 3.2 0.4 - Delivery Shoulder flexiodaddudion (0-5090 of delivery) Shoulder flexion velocity ("/set) 5260 2390 - Pelvidu~wr torso rotation (50-75% of delivery) Pelvis rotation velocity (O/sec) 430 140 640-660~~~ Upper torso rotation velocity (O/sec) 650 120 1170-1220"eb Internal rotation (75-100% of delived Elbow extension velocity (O/sec) 570 310 2200-2440a99R Internal rotation velocity ("/set) 4650 1200 6073-7550~~,~#~ Instant at ball release Ball speed (rn/sec) 25 2 34-38- Follow through Elbow flexion velocity (O/sec).-. -.... - " Uata trom tscamrlla et a1 (71. 880 360 - Data from Fleisig et a1 (1 I). Data from Fleisig et a1 (10). Data from Feltner and Dapena (8). Data from Werner et a/ (24). ' Data from Dillman et a1 (5). Data from Vaughn (22). TABLE 1. Maximum velocities during underhand and overhand pitching. The root mean square error in calculation of threedimensional marker location was determined to be less than 1.0 cm (1 1). A global reference frame was established using X, Y, and Z directions. The global X direction was defined as direction from the pitching rubber toward home plate. Global Y was a direction perpendicular to global X directed toward first base. Global Z was defined as a direction perpendicular to both global X and Y directed vertically. Local reference frames were established at the trunk, shoulder, and elbow (11). Anterior, superior, and lateral axes defined the trunk reference frame. The shoulder reference frame were defined by anterior, superior, and distal axes of the upper arm, while the elbow reference frame was defined by anterior, medial, and distal axes of the forearm. The locations of the midhip, midshoulder, elbow joint center, and shoulder joint center were calculated in each frame as described by Dillman et al (5). Midhip was determined to be the midpoint of a line segment between the two hip markers, and the midshoulder was established at the midpoint of a line segment between the two shoulder markers. Digitized locations for the throwing shoulder and elbow markers were translated to the estimated joint center location (1 1). Ball speed was recorded as it left a pitcher's hand with a Jugs Tribar Sport radar gun (Jugs Pitching Machines Company, Tualatin, OR). Kinematic variables (angular displacement and velocity) at the shoulder and elbow joints were calculated as previously described (5,8,11,22). Rotation of the forearm about the upper arm's long axis was used to calculate shoulder external rotation. Shoulder flexion was calculated as the angle between the upper arm and the trunk in the sagittal plane. Elbow flexion was defined as the angle between the distal directions of the upper arm and forearm. Pelvis orientation angle was defined as the JOSPT Volume 28 Number 6 December 1998 407

Parameter h!kry Shoulder flexiodadduction (&SO% of delivewl Shoulder adductionlhorizontal adduction toque (%BW X HT) Shoulder internal rotation toque (%BW x Hn Shoulder medial force (%BW Pelvidu~wr torso rotation (50-75% of delivery) Shoulder anterior force (%BW Elbow compressive force (%BW Elbow flexion torque (%BW X HT) Internal rotation (75-1 00% of delivew) Shoulder superior/compressive force (%BW Shoulder abduction toque (%BW X HT) Shoulder lateral force (%BW Elbow valgudvarus toque (%BW X HT) Elbow anterior force (%BW Elbow lateral/medial force (YoBW Shoulder extension toque (%BW x HT) Am deceleration/follow through Shoulder extensionlhorizontal abduction torque (%BW X HT) Shoulder posterior force (%BW Shoulder compressive force (%BW Elbow compressive force (%BW Elbow extension toque (%BW " Data irorn Fleisig et a1 (1 11. Data from Escamilla et a1 (12). Data from Fleisig et a1 (101. Data from Feltner and Dapena (8). Data from Werner et a/ (24). BW = Body weight. HT = Height. Underhand R SD TABLE 2. Magnitudes for kinetic data during underhand and overhand pitching. angle between the pelvis line segment (lead hip to throwing hip) and the global Y direction in the global XY plane. Upper torso orientation angle was defined as the angle between the upper torso line segment (leading shoulder to throwing shoulder) and the global Y direction in the global XY plane. Linear and angular velocities and accelerations were determined with finite differences utilizing the fivepoint central difference method (18). Angular velocities of the pelvis and upper torso were calculated with a method published by Feltner and Dapena (9). Kinetic values (joint force and torque) were calculated with a previously described procedure that used kinematic data, docu- mented cadaveric segment parameters, and inverse dynamics equations (9-11). The mass and the center of mass locations of the forearm and upper arm were determined using previously published cadaveric data (4). Kinetic values were reported as the force and torque applied to the forearm at the elbow and applied by the trunk to the upper arm at the shoulder (10). The force at the shoulder was separated into three components: anterior-posterior, superior-inferior, and medial-lateral. Torque at the shoulder joint was sep arated into adductionabduction, external-internal rotation, and flexionextension. Force applied to the forearm at the elbow was separated into three components: medial-lateral, anterior-posterior, and compressive. Elbow torque was separated into two components: flexion-extension and varus-valgus. Supination-pronation torque at the elbow could not be calculated with the methods available. Individual forces were divided by body weight (BW) and torques were divided by body weight and height (BW X HT) to normalize individual results and eliminate any effects due to the size of the subject. Temporal data were normalized by aligning the instant of foot contact and ball release for each subject. The timing of kinematic and kinetic data were reported as a percentage of the delivery phase completed, where 0% corresponded to the instant the front foot contacted the ground and 100% at the instant of ball release. Data for the three fastest pitches thrown by each pitcher into the strike zone were averaged. To simplify the interpretation of data, the pitching motion was separated into four phases: wind up, stride, delivery, and follow through (Figure 2). The wind-up phase was defined as the time from initial movement from the ready position until lead foot toe-off (Figure 2a-c). The stride phase was defined as the time from lead foot toe-off to lead foot contact (foot flat) with the ground (Figure 2d-0. The delivery phase was defined as the time from foot contact to release of the ball (Figure 2g-j). The final phase was follow through, which occurred from the instant of ball release until the forward motion of the throwing arm had stopped (Figure 2k-1). During the follow through, the forearm flexed at the elbow and continued to flex until the arm and forearm were decelerated. Means and standard deviations for all subjects were determined for 14 kinematic and kinetic parameters. The results were compared with overhand pitching data for qualitative analysis. The force and torque parameters were normalized 408 Volume 28 Number 6 December 1998 JOSFT

-14 I 1 0 25 50 7s 100 125 150 PC Time (% pitch) FIGURE 3. Toques (% body weight x height) applied to the arm at the shoulder for A) adduction (+)/abduction (-); B) internal (+)/external (-) rotation; and C) flexion (+)/extension (-) VS. time (% pitch). Graphs represent mean and standard deviation data for all subjects. The instances of foot contact (FC) and ball release (REL) are shown. to body weight and height to allow for comparison. RESULTS During the wind-up and stride phases of the motion, the majority of kinematic and kinetic parameters had low magnitudes. During the wind-up JOSPT * Volume 28 * Number 6 - December 1998 REL phase, the arm was hyperextended at the shoulder as the pitcher pushed off the pitching rubber with the pivot foot to initiate forward translation of the body. Emphasis on forward translation during the stride phase was illustrated by the linear velocity of the hips (Table 1). As the pitcher reached foot contact, the trunk (pelvis, upper torso) was re tated toward third base and the humerus was flexed beyond 180" to an extended position. During the delivery phase, the ball was accelerated forward with a combination of trunk (pelvis and upper torso) rotation, arm (flexion and internal) rotation at the shoulder, and flexion at the elbow. The highest magnitudes for kinematic and kinetic parameters occurred during the delivery phase as the arm was accelerated (Tables 1 and 2). These occurred during an average time interval of 0.102 2 0.014 seconds for all subjects. From 0 to 50% of the delivery phase, a maximum adduction torque was exerted at the shoulder (Table 2, Figure 3). A maximum internal rotation torque was also exerted at the shoulder as forward flexion of the arm reached a maximum velocity greater than 5,000 /sec (Table 1, Figure 4). At approximately 45% of the delivery, a maximum medial (74% BW) force was produced at the shoulder. This was followed by a maximum anterior (38% BW) force at the shoulder which occurred at approximately 55% of the delivery. These forces were produced as the arm was adducted and flexed forward (Table 2, Figure 5). From 50 to 75% of the delivery, maximum pelvis rotation velocity was reached which was then followed by maximum upper torso rotation velocity (Table 1). As the humerus was flexed forward, the forearm extended at the elbow creating a maximum extension velocity of 570 /sec (Table 1, Figure 4). A flexion torque was initiated at the elbow and reached a maximum at the end of the delivery phase (Table 2, Figure 6). During this time, a maximum compressive force (70% BW) was experienced at the elbow (Figure 7). This was followed by a maximum superior force (98% BW) at the shoulder that occurred at approximately 77% of the delivery phase (Figure 5). The last 25% of the delivery phase was characterized by internal rotation of the humerus, which reached a maxi-

RESEARCH STUDY velocity (880 /sec) was reached as the forearm continued to decelerate (Figure 4). 0 25 50 75 100 125 150 FC Time (% pitch) REL FIGURE 4. Angular velocity ("/second) for A) shoulder flexion; Bj elbow extension (+)flexion (-); and C) shoulder internal (+)/external (-) rotation vs. time (% pitch). Graphs represent mean and standard deviation data for all subjects. The instances of foot contact (FC) and ball release (REL) are shown. mum velocity (4,600 /sec) just prior to ball release (Table 1, Figure 4). During this time, a maximum abduction torque (9% BW X HT) and a maximum extension torque (10% BW X HT) were generated at the shoulder (Table 2, Figure 3). Maximum lateral force (47% BW) and valgus torque (4% BW X HT) were generated at the elbow. Just prior to ball release, elbow extension was terminated and elbow flexion was initiated. After ball release, the arm and forearm were decelerated. A second peak extension torque (9% BW X HT) was exerted and a maximum posterior force (59% BW) was experienced at the shoulder (Table 2, Figure 5). During the follow through phase, a second peak elbow compressive force occurred (56% BW) as a maximum extension torque (2% BW X HT) was exerted (Figures 6 and 7). A maximum elbow flexion DISCUSSION The results of this study are compared with results from previous studies of overhand pitching in Tables 1 and 2. The purpose of the comparison is for qualitative analysis of the loads experienced during underhand pitching. Since baseball pitching studies used male subjects and the current study used female subjects, musculoskeletal and social differences between genders must be recognized. Typically, a female's upper torso and arms possess less muscle mass and strength than the male. At the elbow, the carrying angle is larg er, and there is often more ligamentous laxity in the female. Grip strength and hand sizes are usually less for women. While a starting baseball pitcher seldom pitches in a game without 3-4 days of rest, a female windmill pitcher may pitch 2 days in a row or twice in one day during a tournament. During overhand pitching, the stability of the glenohumeral joint is compromised when the humerus rotates internally and adducts horizontally while maintaining a position of 90" of abduction. his potential for instability is magnified as the forces to resist glenohumeral distraction and anterior translation reach a peak after ball release during deceleration. Although the humerus is not held in an abducted position during underhand pitching, the motion does require resistance to distraction while also controlling internal rotation and elbow extension. The total circumduction of the arm about the glenohumeral joint from the initiation of stride phase to completion of the movement is about 485". Of significance is that this windmill motion is performed rapidly with a softball that weighs 6% to 7 oz by design compared with the baseball weight of 5 Volume 28 Number 6 December 1998 JOSPT

RESEARCH STUDY -100 J! I 0 25 50 75 I00 125 150 FC Time (% Pitch) FIGURE 5. Forces (% body weight) applied to the an at the shoulder for A) anterior (+)/posterior (-I; B) superior (+)/inferior(-); and C) lateral (+)/medial(-) vs. time (% pitch). Graphs represent mean and standard deviation data for all subjects. The instances of foot contact (FC) and ball release (REL) are shown. oz. Most of the circumduction motion is performed with the elbow in full extension, which accentuates the centrifugal distraction force on the glenohumeral joint. In the first three-dimensional analyses of the windmill pitching motion, Werner (23) observed a steady increase in the resultant force component directed from the throwing elbow to the shoulder during the REL pitching motion until just prior to ball release. This force corresponds to the superior force calculated in this study that reached a peak value during the middle of the delivery phase. Using the average weight of 635 N for all pitchers, the peak superior force of 98% BW equates to a 625 N force acting on the shoulder. As noted by Werner (23), this force had the effect of maintaining joint stability by resisting the distraction of the humerus from the shoulder joint caused by the windmilling motion (23). Similar resultant forces resisting distraction at the shoulder have been calculated for overhand pitching (8, 10-12). A difference between the methods of pitching is the timing of the peak values for the resultant forces. The greatest resistance to distraction (compressive, superior forces) occurs during the delivery or acceleration phase for underhand pitching, while the greatest resistance during overhand pitching is produced in the deceleration phase (Table 2). The magnitude of the forces were similar, as normalized forces during underhand pitching were 80-95% of the normalized values determined for overhand pitching. Forces acting to resist distraction at the elbow were also similar with the magnitudes during underhand pitching equaling 67-79% of the values calculated for overhand pitching. Although the forces were slightly less during underhand pitching, many of the torques were equivalent or slightly greater. This is especially true for the extension (10% BW X HT) and abduction (9% BW X HT) torques exerted at the shoulder. Werner (23) concluded that a major contributor to ball velocity was internal rotation of the humerus produced by shoulder internal rotation torque. In this study, internal rotation torque generated early in the delivery phase was similar in magnitude to the torques calculated during overhand pitching (Table 2). The resulting internal rotation velocity was over 4,000 /sec which, although smaller than overhand pitching, is a high magnitude (Table 1). Maffett et a1 (15) observed high levels of activity for the pectoralis major and subscap ularis, contributing to adduction and internal rotation of the humerus during windmill pitching. Although it is difficult to determine how these forces and torques relate to the incidence of injury, the JOSPT Volume 28 Number 6 December 1998 41 1

RESEARCH STUDY - 0 25 50 75 100 125 150 FC REL Time (% pitch) FIGURE 6. Toques (% body weight x heightj applied to the foream at the elbow for A) extension (+)fixion (-), and B) varus (t )/valgus (-) vs. time (% pitch). Graphs represent mean and standard deviation data for all subjects. The instances of foot contad (FC) and ball release (REL) are shown. types of injuries reported by Loosli et a1 appear to be related to overuse and the accumulative stress at the shoulder and elbow (14). Tendinitis, rotator cuff and tendon strain, and ulnar nerve damage were the majority of injuries reported for all grades of injuries. Similar to overhand throwing, the causes of these injuries may be related to the mechanisms of maintaining joint stability. A common complaint of softball pitchers is anterior shoulder discomfort near the origin of the long head of the biceps tendon. The diagnosis of bicipital tendinitis is often made with little consideration to the possibility of subscapularis or pectoralis strain, which may occur as increased shoulder extension is followed by large forces and torques produced during the delivery motion (Figures 3 and 5). Maffett et al (15) concluded that the high levels of activity for the pectoralis major and subscapularis observed during early delivery also contribute to the stabilization of the anterior capsule of the shoulder. Despite the possibility of injury to these structures, pain in the anterior shoulder is often treated by injection of a steroid/analgesic mixture into the bicipital tendon area. Unfortunately, repeated injections in the area of the anterior shoulder, with continued pitching and without sufficient rehabilitation, may actually lead to weakness, degradation, and even disrup tion of the biceps tendon. The long head of the biceps tendon, by reason of its insertion into the superior glenoid labrum, functions as a humeral head depressor normally. In the lax glenohumeral joint, it is speculated that the biceps tendon undergoes stretching and surface fibrillation. As discussed by Fleisig et al (10). during overhand pitching, the biceps brachii functions to provide elbow flexion torque and aids in resisting humeral distraction. With the attachment of the long head of the biceps brachii to the glenoid labrum, the contraction of the biceps brachii to control elbow extension creates tension on the biceps tendon labrum complex. At a similar time in the motion, compressive force is needed to resist distraction, which further increases the demand on the biceps labrum complex. This mechanism can be applied to underhand pitching during the delivery phase. Forces to resist distraction reach a peak at a time during delivery that elbow flexion torque is exerted to control elbow extension and initiate elbow flexion (Figures 5 and 6). The demand on the biceps labrum complex to both resist glenohumeral distraction and produce elbow flexion torque makes this structure susceptible to overuse injury. Internal ro~tion of the humerus and pronation of the forearm further complicate the mechanism. Although not measured in the current study, the torsional stress that occurs as the forearm is pronated through ball release has been related to stress fracture injuries to underhand pitchers (21). Softball pitchers often are required to pitch multiple games in one day and or pitch consecutive days throughout the season. It would seem reasonable to speculate that, even with perfect pitching mechanics, overuse type injuries will occur. Loosli et al (14) reported that pitchers reporting grade I or I1 type injuries (did not result in missed games or practices) on average pitched more innings per season than uninjured pitchers. In the softball pitcher, forceful deceleration of the upper arm at or just prior to ball release, similar to baseball pitching, places a Volume 28 Number 6 December 1998 0JOSPT

RESEARCH STUDY 0 25 50 75 100 12s 150 REL FC Time (% pitch) FIGURE 7. Forces (% body weight) applied to the forearm at the elbow for A) medial (+)/lateral (-1; 8) anterior (+)/posterior (-); and C) compressive vs. time (% pitch). Graphs represent mean and standard deviation data for all subjects. The instances of foot contact (FC) and ball release (REL) are shown. substantial burden on the posterior rotator cuff muscles (Figures 3 and 5). Maffett et al (15) have shown that, similar to baseball pitching, the teres minor is very active in decelerating the humerus. However, it was concluded that the stress on the teres minor is less during softball pitching because acceleration forces seem to be dissipated through contact of the arm against the lateral thigh. This conclusion does not appear to agree with the findings of this study. Using the average weight of 635 N and height of 1.73 m for all pitchers, torques applied at the shoulder to control adduction and flexion were equivalent to 100 Nm. These torques are required to decelerate the high rotational velocities achieved during delivery and to transfer momentum from proximal to distal segments down the throwing arm to the ball (2). Alexander and Haddow (2) analyzed the angular kinematics of the upper arm, lower arm, and hand of four skilled windmill pitchers. They observed a specific sequence of motions and concluded that the larger, more proximal segments attained peak acceleration before the more distal segments. After reaching peak velocity, the proximal segment was decelerated in order to transfer momentum to the distal segment. In the current study, this sequence was illustrated by the occurrence of peak flexion velocity of the humerus at the shoulder early in the delivery phase, which was followed by the generation of an extension torque at the shoulder to decelerate the humerus (Figures 3 and 4). A peak shoulder extension torque was reached as elbow flexion was initiated, enabling the momentum from the upper arm to be transferred to the lower arm. Werner (23) also observed a similar extension or negative "windmilling torque" just prior to ball release and believed the purpose was to control the windmilling motion. Posterior shoulder symptoms, which are usually present after pitching, appear to be the second most common site of pain and soreness in the female windmill pitcher. Eccentric loading and stretching of the posterior muscle girdle with overuse could significantly contribute to dynamic anterior instability of the humeral head. Ultimately, failure or insufficiency of the posterior support structures to keep the humeral head properly seated in the glenoid could propagate symptoms of posterior shoulder pathology. In baseball pitching, the production of varus torque to resist valgus motion often leads to elbow injury (10). Only occasionally does ulnar collateral ligament injury become apparent in the softball pitcher because only a small amount of varus torque is produced (Figure 6). Ulnar nerve neuritis does occur in windmill JOSm Volume 28 Number 6 December 1998

RESEARCH STUDY pitching, although the cause is not the same as in overhand pitching. The injury in windmill pitching is attributed to poor mechanics and occurs as the medial elbow contacts the hip just before ball release. CONCLUSIONS This study presented kinematics and kinetics for underhand pitching motion of eight collegiate fast pitch softball pitchers throwing the fastball pitch. Comparison of underhand and overhand pitching illustrated similar joint speeds and loads for each m e tion. One of the critical instances for underhand pitching wa. during the delivery or acceleration, where the forces to resist distraction at the shoulder and elbow were the greatest. This differed from overhand pitching, where the peak forces needed to resist distraction occurred during deceleration. Obvious differences, including gender, size and weight of the ball, and pitching environment (height of mound), prevent a direct comparison to overhand pitching; however, these results question the assumption that underhand pitching does not create significant stress on the shoulder and elbow. Further investigation is needed to fully determine the influence of underhand pitching on overuse injuries. Attention should be directed toward prevention of injury by teaching proper pitching mechanics, strengthening the shoulder and rotator cuff musculature, and regulating the number of pitches and mound a p pearances for the female fast pitch softball pitcher. JOSPT ACKNOWLEDGMENTS The authors would like to thank Anthony DeMonia and Phillip Sutton for their assistance with data analysis and Dan Nichols for his assistance with illustrations. The authors would also like to thank Jennifer Hogan, Cheri Kempf, and pitchers from Samford University and Troy State University for their assistance with this study. REFERENCES 1. Abdel-Aziz Yl, Karara HM: Direct linear transformation: From comparator coordinates into objects coordinated in close-range photogrammetry. In: American Society of Photogrammetry, Symposium on Close-Range Photogrammetry, pp 1-19. Falls Church, VA: American Society of Photogrammetry, 1971 2. Alexander M, Haddow J: A kinematic analysis of an upper extremity ballistic skill: The windmill pitch. Can J Appl Sport Sci 7(3):209-2 17, 1982 3. Atwater AE: Biomechanics of overarm throwing movements and of throwing injuries. Exerc Sport Sci Rev 7:43-85, 1979 4. Clauser CE, McConville JT, Young JW: Weight, volume, and center of mass of segments of the human body. AMRL- TR-69-70, Wright-Patterson Air Force Base, 1969 5. Dillman CJ, Fleisig GS, Andrews JR: Biomechanics of pitching with emphasis upon shoulder kinematics. J Orthop Sports Phys Ther 18(2):402-408, 1993 6. Elliott B, Grove JR, Gibson B, Thurston B: Threedimensional cinematographic analysis of the fastball and curveball pitches in baseball. Int J Sport Biomech 2(1):20-28, 1 986 7. Escamilla RF, Fleisig GS, Barrentine SW, Zheng N, Andrews JR: Kinematic comparisons of throwing different types of baseball pitches. J Appl Biomech l4(l): 1-23, 1998 8. Feltner M, Dapena J: Dynamics of the shoulder and elbow joints of the throwing arm during a baseball pitch. Int J Sport Biomech 2(4):235-259, 1986 9. Feltner ME, Dapena J: Three-dimensional interactions in a two-segment kinetic chain. Part I: General model. Int J Sport Biomech 5(4):403-4 19, 1989 10. Fleisig GS, Andrews JR, Dillman CJ, Escamilla RF: Kinetics of baseball pitching with implications about injury mechanisms. Am J Sports Med 23(2): 233-239, 1995 11. Fleisig GS, Escamilla RF, Andrews JR, Matsuo T, Satterwhite Y, Barrentine SW: Kinematic and kinetic comparison between baseball pitching and football passing. J Appl Biomech 12(2):207-224, 1996 12. Escamilla RF, Fleisig GS, Alexander E, Andrews JR: A kinematic and kinetic comparison while throwing different types of baseball pitches. Med Sci Sports Exerc 26(5):Sl75, 1994 (abstract) 13. lobe F, Kvitne R: Shoulder pain in the overhand or throwing athlete: The relationship of anterior instability and rotator cuff impingement. Orthop Rev 18(9):963-975, 1989 14. Loosli AR, Requa RK, Garrick JG, Hanley E: Injuries to pitchers in women's collegiate fast-pitch softball. Am 1 Sports Med 20(1):35-37, 1992 15. Maffett M W, lobe FW, Pink MM, Brault J, Mathiyakom W: Shoulder muscle firing patterns during the windmill softball pitch. Am J Sports Med 25(3):369-374, 1997 16. McLeod WD: The pitching mechanism. In: Zarins B, Andrews JR, Carson WG (eds), Injuries to the Throwing Arm, pp 22-29. Philadelphia: W.B. Saunders Company, 1985 17. McLeod WD, Andrews JR: Mechanisms of shoulder injuries. Phys Ther 66(12): 1901-1904, 1986 18. Miller Dl, Nelson RC: Biomechanics of Sport-A Research Approach, Philadelphia: Lee & Febiger, 1973 19. Pappas AM, Morgan WJ, Schulz LA, Diana R: Wrist kinematics during pitching: A preliminary report. Am J Sports Med 23(3):312-3 15, 1995 20. Sakurai S, lkegami Y, Okamoto A, Yabe K, Toyoshima S: A three-dimensional cinematographic analysis of upper limb movement during fastball and curveball baseball pitches. J Appl Biol 9(1): 47-65, 1993 2 1. Tanabe S, Nakahira 1, Bando E, Yamaguchi ti, Miyamoto H, Yamamoto A: Fatigue fracture of the ulna occurring in pitcherr of fast-pitch softball. Am J Sports Med 19(3):317-32 1, 199 1 22. Vaughn RE: Threedimensional kinematics of the baseball pitch. In: Terauds J, Barham JN (eds), Biomechanics in Sports 11, pp 72-78. Del Mar, CA: International Symposium on Biomechanics in Sports, 1985 23. Werner SL: Selected kinematic and kinetic parameters of the windmill softball pitch. Unpublished master's thesis, Indiana University, Bloomington, IN, 1987 24. Werner SL, Fleisig GS, Dillman CJ, Andrews JR: Biomechanics of the elbow during baseball pitching. J Orthop Sports Ph ys Ther 17(6):274-278, 1993 25. Wilson FD: Valgus extension overload in the pitching elbow. Am J Sports Med 1 1 (2):83-88, 1983 414 Volume 28 Number 6 December 1998 JOSPT