The biomechanical effect of radial shortening on the radiocapitellar articulation.

Repeated trauma to the radial head may be one of the causative factors in the genesis of osteochondritis dissecans of the capitellum. We measured the force, contact area and pressure across the radiocapitellar articulation of the elbow before and after radial shortening osteotomy in five fresh-frozen cadaver upper limbs with loads of 45, 90 and 135 N, respectively. Measurements were made on pressure-sensitive film placed in the radiocapitellar articulation with the forearm in the supinated, neutral and pronated positions before and after radial shortening. Radial shortening significantly reduced the mean force and contact area across the radiocapitellar articulation in all positions of the forearm.

A comparison of biomechanical mechanisms of whiplash injury from rear impacts.

Several hypotheses have been proposed to explain the mechanism of injury in whiplash including, pressure on nerve root ganglia, stretching of facet capsules, or damage to facet articular cartilage. These injury mechanisms have not been directly compared in the same study. A comparison could provide insight into the most likely mechanism of whiplash injury. Twenty eight volunteers underwent rear impacts with head and chest acceleration data collected. The same apparatus was used to test 11 cervico-thoracic human cadaveric spines with an instrumented headform attached. Head acceleration, individual vertebral kinematics from high speed video, local nerve root pressure, and facet joint contact pressures were collected during impacts. Each specimen was tested first at an impact acceleration similar to that of volunteers, who reported minimal or no symptoms after the test, then at double the acceleration. Head X (forward) and Z (upward) accelerations of cadaveric specimens were very similar in time sequence and magnitude to those of unprepared volunteers. Pressure around the lower cervical nerve roots ranged from 2.7kPa to 10kPa, and occurred generally after chest but before peak head acceleration. Facets at C4-5 and C5-6 had the highest probability (64% and 71% respectively) of pinching. Neither pressure rise nor pinching changed significantly with increased acceleration. Vertebral intersegmental extension rotations (4 ( o ) -9.5 ( o ) ) and posterior translations (3.7-8.9 mm) peaked near maximum head excursion into the head restraint, at the time of peak head acceleration. Vertebral shear translations showed the largest (and only significant) increases with increased impact acceleration. This data implies that facet shearing was most sensitive to the increased acceleration in this experiment and may be a primary mechanism of cervical spine injury in rear impacts.

Are drivers more likely to injure their right or left foot in a frontal car crash: a crash and biomechanical investigation.

In this study we hypothesized that for frontal crashes, the driver was more likely to sustain foot injury than the front seated passenger and the right foot was more likely to be injured than the left because the driver's right foot was positioned in dorsiflexion and eversion during the crash and less able to tolerate crash forces. Seventeen CIREN frontal crashes were studied to document the circumstances and resulting foot injuries. NASS data was sampled to determine frequencies of injury to foot side and occupant position. Biomechanically, matched pairs of cadaveric feet were tested in compression with one in dorsiflexion, and the other, dorsiflexion, and eversion. Results showed that drivers were more likely than passengers to sustain a foot or ankle injury and this was almost exclusively due to a higher frequency of right foot injuries in drivers. Braking, with significant toepan intrusion, was associated with an increase in both right and left foot injuries in drivers. Biomechanically, 9 of 15 pairs tested provided useable data. There was a significant decrease in load to failure between specimens forced into dorsiflexion and eversion (mean, 4107 N, sd = 1630 N) compared with dorsiflexion alone (mean, 6468 N, sd = 2435 N, p = 0.001). The majority (16 of 17) of foot injuries in the CIREN cases were to drivers of which 13 injured their right foot and ankle. An implication of this work is that if thresholds based on data from dorsiflexion and compression loading are used to predict foot and ankle injury in frontal crashes they may not represent that population of drivers who are braking at the time of the crash and could suffer injuries at forces lower than current threshold values.

The response of human volunteers to rear-end impacts: the effect of head restraint properties.

STUDY DESIGN: Human volunteers were subjected to a rear-end impact while sitting on a standard automobile seat, and sagittal plane kinematic responses were quantified. The effect of changing head restraint properties was determined by use of a repeated measures design. OBJECTIVE: To determine the forces acting, and relative motions resulting, on volunteers in a rear-end impact and the effect of head restraint properties. SUMMARY OF BACKGROUND DATA: In several recent studies of the kinematics of the cervical spine during rear-end impact, a forward thrust to the lower cervical spine was produced, and a transient S shape of the spine resulted while the head remained upright during the initial phase of the impact. This may result in nonphysiologic intervertebral motions and tissue strains. METHODS: Nineteen automobile seats were first tested, and a modified head restraint was designed. Each volunteer sitting on a standard vehicle seat was subjected to an impact pulse of 3g with a 4-kph speed change. Testing was performed first with the modified head restraint, then again after replacement by the head restraint that came with the seat. Kinematic responses were compared for both head restraints by use of a repeated measures analysis of variance. RESULTS: There was a measurable time difference between peak chest and peak head accelerations, which resulted in the chest being thrust forward by the seat back before the head was thrust forward by the head restraint. The modified head restraint significantly reduced the contact time difference and therefore decreased the relative chest-to-head forward motion. CONCLUSIONS: Volunteers seated on a standard automobile seat demonstrated differential sagittal plane motion between the chest and head. It is possible to significantly decrease the relative chest-to-head motion by altering the characteristics of the head restraint.

Biomechanical factors and failure of transcervical hip fracture repair.

OBJECTIVE: To assess the association between biomechanical measurements (bone quality of the femoral neck, comminution, fracture angle, and fracture level) and the likelihood of fixation failure among patients who have a multiple screw stabilisation of an intracapsular hip fracture. METHODS: A cohort study of 139 Washington State residents greater than 60 years of age who sustained a fall-related transcervical hip fracture treated from 1990 to 1996 inclusive. Measurements of bone quality, fracture angle, fracture level, and comminution were taken from perioperative X-rays. The outcome measure was clinical failure of the internal fixation procedure within 12 months of hospital discharge, as measured by readmission for further surgery to that hip. RESULTS: Of the four biomechanical aspects examined, only bone quality, as measured by presence of an ICD code for osteoporosis, was significantly associated with risk of subsequent hospitalisation for revision surgery (adjusted hazard ratio 7.7, 95% CI 1.8-32.8). CONCLUSION: A diagnosis code for osteoporosis was related to the outcome of intracapsular fractures repaired with multiple pins. Other biomechanical measurements from diagnostic X-rays were not related to the need for further surgery.

Do "whiplash" victims with neck pain differ from those with neck pain and other symptoms?

Studying 432 patients with "whiplash" neck pain after a low speed rear-end crash, we determined whether those who reported associated symptoms (arm or low back pain) differed from those with neck pain only. Exposure variables were: age, gender, height, weight, surprised by impact, rotated position of the head, and pre-existing cervical or lumbar spinal degeneration. Immediate symptoms were considered. Odds ratios (OR) were derived from logistic regression. Only pre-existing lumbar spinal degeneration was associated with having arm symptoms (OR = 9.6, CI = 1.1, 83.6) or low back pain (OR = 23.3, CI = 2.6, 206.7) along with neck pain.

Canal geometry changes associated with axial compressive cervical spine fracture.

STUDY DESIGN: A laboratory study using isolated ligamentous human cadaveric cervical spines to investigate canal occlusion during (transient) and after (steady-state) axial compressive fracture. OBJECTIVES: To determine whether differences exist between transient and postinjury canal occlusion under axial compressive loading, and to examine the effect of loading rate on canal occlusion. SUMMARY OF BACKGROUND DATA: Prior studies have shown no correlation between neurologic deficit and canal occlusion measurements made on radiographs and computed tomography scans. The authors hypothesized that postinjury radiographic assessment does not provide an appreciation for the transient occlusion that occurs during the traumatic fracture event, which may significantly affect the neurologic outcome. METHODS: Twelve human cervical spines were instrumented with a specially designed canal occlusion transducer, which dynamically monitored canal occlusion during axial compressive impact. Six specimens were subjected to a fast-loading rate (time to peak load, approximately 20 msec), and the other six were subjected to a slow-loading rate (time to peak load, approximately 250 msec). After impact, two different postinjury canal occlusion measurements were performed. RESULTS: Each of the six specimens subjected to the fast-loading rate incurred burst fractures, whereas the slow-loading rate produced six wedge-compression fractures. For the fast-rate group, the postinjury occlusion-measurements were significantly smaller than the transient occlusion. In contrast, transient occlusion was not found to be significantly different from postinjury occlusion in the slow-rate group. All of the comparisons between loading rate groups showed significant differences, with the fast-rate fractures producing larger amounts of canal occlusion in every category. CONCLUSIONS: The findings indicate that even if canal occlusion could be measured immediately after axial compressive trauma, the measurement would underestimate the maximal amount of transient canal occlusion. Therefore, postinjury measurement of canal occlusion may indicate a smaller degree of neurologic deficit than what might be expected if the transient occlusion could be measured.

Autograft replacements for the scapholunate ligament: a biomechanical comparison of hand-based autografts.

An ideal replacement for the scapholunate ligament (SLL) has not been found. The carpometacarpal bone-ligament-bone complex at the base of the second and third metacarpal is proposed as a replacement for the SLL in this study. A cadaveric study of matched SLL, second metacarpal-trapezoid ligament, third metacarpal-capitate ligament, and dorsal periosteal retinaculum was performed. Stiffness and strength were obtained from fresh-frozen specimens tested to failure with a hydraulic distractor. The second metacarpal-trapezoid ligament and the third metacarpal-capitate ligament most closely approximated the stiffness and strength of the SLL. The dorsal periosteal retinaculum was significantly less stiff and was significantly weaker than the SLL. This study indicates that the second or third carpometacarpal ligaments are grafts that approximate the mechanical properties of the SLL.

The effect of thumb metacarpophalangeal ulnar collateral ligament attachment site on joint range of motion: an in vitro study.

Although reconstruction of the torn thumb metacarpophalangeal (MCP) joint ulnar collateral ligament (UCL) has been shown to reduce symptoms, final joint motion may be different from that of the uninjured state. It was hypothesized that nonanatomic repositioning of the UCL might affect joint motion; therefore, the effect of UCL attachment site on MCP range of motion was investigated. The UCL and MCP joint capsule were visualized in each of 8 fresh cadaveric hands without otherwise disrupting the joint. The centers of the ligament attachments were marked with pins and each specimen was mounted on a testing frame capable of applying loads through the flexor and extensor tendons. After measuring the flexion, extension, and radial and ulnar deviation ranges of motion of the intact specimen, the origin of the ligament (on the metacarpal) and an attached bone block were elevated and repositioned 2 mm proximal and 2 mm palmar and range of motion was tested. The origin was reattached in its anatomic location and the insertion of the ligament was similarly elevated and displaced 2 mm in distal, dorsal, and palmar directions. Compared with the intact joint, palmar placement of the UCL origin on the metacarpal increased radial deviation (from a mean of 18 degrees to a mean of 27 degrees); proximal placement of the origin decreased it (from 18 degrees to 10 degrees). Similarly, dorsal placement of the UCL insertion on the phalanx increased radial deviation (from 18 degrees to 25 degrees) and distal positioning of the insertion decreased it (from 18 degrees to 11 degrees). Relative to intact joint flexion range of motion (mean, 57 degrees), distal placement of the UCL phalangeal insertion restricted motion (mean, 47 degrees), as did palmar placement (mean, 48 degrees). Extension and ulnar deviation motions were unaffected by ligament attachment position. This study demonstrates that nonanatomic reconstruction of the UCL alters normal MCP joint range of motion.

Support of the talus: a biomechanical investigation of the contributions of the talonavicular and talocalcaneal joints, and the superomedial calcaneonavicular ligament.

The goal of this study was to determine the magnitude of force transmission to the talus by its inferior articulations to provide insight into mechanisms involving acquired deformities of the hindfoot. Cadaver feet were mounted in a loading apparatus that applied axial force through the tibia and fibula as well as tensile loading of the tendons of extrinsic musculature. This also permitted positioning of the tibia in the sagittal plane. Eighteen specimens were tested in three selected positions of the gait cycle. In one series, pressure-sensitive film was inserted into the posterior and anteromedial facets of the talocalcaneal joint as well as into the talonavicular joint. In a second series, film was inserted between the talar head and the superomedial calcaneonavicular ligament. In stance position, the specimens were also tested without posterior tibial tendon (PTT) tension. Contact areas and force transmitted across the articulations were greatest in near toe-off position, in the posterior facet of the talocalcaneal joint. The talonavicular joint, the anteromedial facet of the talocalcaneal joint, and the calcaneonavicular ligament articulation showed sequentially decreasing amounts of contact area and force transmission. Mean pressures were similar across all articulations, except in the posterior facet in near toe-off position. From heel-strike to stance, to near toe-off, a trend to increasing contact area and force was noted. No difference in contact characteristics was found in the calcaneonavicular ligament articulation after PTT release. The contact force of the calcaneonavicular ligament against the talus was found to be much smaller than those of other talar articulations; however, its medially oriented direction must contribute to stabilization of the head of the talus against medial displacement. Loss of PTT tension was not found to alter the contact forces acting at the talar head in this model, which might indicate that it shares its talar stabilizing function with other structures.

Two reconstructive techniques for flatfoot deformity comparing contact characteristics of the hindfoot joints.

The effect of two different methods of reconstruction of flatfoot deformity and the role of the posterior tibial tendon on the contact characteristics of the hindfoot joints were quantified using pressure-sensitive film. Each of 10 cadaver feet was loaded quasi-statically by an axial compressive force to simulate varying loads. First, a specimen was tested intact, then it was tested after sectioning the spring ligament and loading the specimen cyclically to create one type of flatfoot deformity. It was then tested again after reconstructing the deformity. Reconstructions used were the Dillwyn-Evans procedure (bone graft in osteotomy of the calcaneus) or the calcaneocuboid distraction arthrodesis (CCDA). We found that surgically produced flatfoot deformity altered mainly the talonavicular joint, by decreasing its contact area. The Dillwyn-Evans method had less effect on the talonavicular joint (altering 2 of 6 measured parameters) than the CCDA (3 of 6) and more effect on the anteriomedial facet (altering 3 of 6 parameters) than the CCDA (1 of 6). The Dillwyn-Evans method had more effect on the posterior facet (altering 2 of 6 measured parameters) than the CCDA (1 of 6). Function of the posterior tibial tendon had no effect on contact characteristics of the hindfoot joints after either type of reconstruction. These findings are based on measurements using a quasi-statically-loaded foot model at three selected positions, and results may be different with dynamic loading.

Biomechanical analysis of capitate shortening with capitate hamate fusion in the treatment of Kienböck's disease.

The biomechanical effects of surgical treatment options for Kienböck's disease have been compared. However, no study has included a direct analysis of capitate shortening along with capitate-hamate fusion (CSCHF). To investigate the biomechanical effects of CSCHF, a cadaver model of the upper extremity was used to determine radiocarpal articular pressure changes resulting from this procedure using pressure-sensitive film. Ten specimens were tested by placing each in an apparatus that applied load across the radiocarpal joint through the wrist flexor and extensor tendons. Testing was performed in 3 wrist positions (ulnar deviation, radial deviation, and neutral) combined with 3 forearm positions (pronation, supination, and neutral) and neutral flexion/extension. Radioscaphoid, radiolunate, and mean contact pressures in the entire radiocarpal joint were determined for each of the 9 wrist positions, both intact and after surgery. The radioscaphoid mean pressure increased in 6 of 9 positions and was unchanged in 3 positions. The radiolunate mean pressure decreased in 9 of 9 positions. The radiocarpal mean pressure increased in 2 of 9 positions and was unchanged in 7 positions. These data suggests that CSCHF increases radioscaphoid mean pressure, decreases radiolunate mean pressure, and has little effect on radiocarpal mean pressure.

Two-, four-, and six-strand zone II flexor tendon repairs: an in situ biomechanical comparison using a cadaver model.

A dynamic in vitro model of zone II flexor tendon repair was used to compare gliding resistance, gap formation, and ultimate strength of the 2-, 4-, and 6-strand repair techniques. Each of 12 hands was mounted to a loading frame with 3 flexor tendons attached to individual pneumatic cylinders. A spring attached to a pin through the distal end of each digit provided a 1.25-kg resistance force. The force required to flex each proximal interphalangeal joint to 90 degrees was determined. Following this, the tendons were sectioned and each was repaired using a different technique so that each specimen acted as its own control. The 2- and 4-strand core sutures were placed using a suture interlock technique with radial and ulnar grasping purchase of the tendon on each side of the transverse part of the repair. Each repair was accomplished using a single core stitch with the knot buried between the tendon ends. The 4-strand repair involved an additional horizontal mattress suture with the knot buried. Repair of the dorsal side of the tendon was performed followed by core suture placement. The palmar portion of the peripheral locking suture was completed after core suture placement. Following repair, each hand was remounted on the frame and cycled 1,000 times. After cyclic loading, the resulting gap between the repaired ends of each tendon was measured, the tendons were removed from the hand, and each was loaded to failure in tension. All tendon repairs showed a small, but not statistically significant, increase in gliding resistance after reconstruction. The 2-strand repair had significantly greater gap formation after cyclic loading (mean gap, 2.75 mm) than either the 4-strand (0.30 mm) or 6-strand (0.31 mm) repair. The tensile strength of the 6-strand repair (mean, 78.7 N) was significantly greater than either the 4-strand (means, 43.0 N) or 2-strand (mean, 33.9 N) repair.

Norian SRS cement augmentation in hip fracture treatment. Laboratory and initial clinical results.

Bone quality, initial fracture displacement, severity of fracture comminution, accuracy of fracture reduction, and the placement of the internal fixation device are important factors that affect fixation stability. New high strength cements that are susceptible to remodeling and replacement for fracture fixation may lead to improved clinical outcome in the treatment of hip fractures. Norian SRS is an injectable, fast setting cement that cures in vivo to form an osteoconductive carbonated apatite of high compressive strength (55 MPa) with chemical and physical characteristics similar to the mineral phase of bone. It can be used as a space filling internal fixation device to facilitate the geometric reconstruction, load transfer, and healing of bone with defects and/or fractures in regions of cancellous bone. Furthermore, this cement can improve the mechanical holding strength of conventional fixation devices. Use of this material potentially could improve fracture stability, retain anatomy during fracture healing and improve hip function, thus achieving better clinical outcomes. In vivo animal studies have shown the material's biocompatibility, and cadaveric studies have shown the biomechanical advantage of its use in hip fractures. Initial clinical experience (in 52 femoral neck fractures and 39 intertrochanteric fractures) showed the potential clinical use of this innovative cement in the treatment of hip fractures.

Predicting the mechanical outcome of femoral neck fractures fixed with cancellous screws: an in vivo study.

OBJECTIVE: The goal of this project was to determine, by using a retrospective cohort analysis, the relative importance of fracture geometry, bone density, and quality of fracture reduction in predicting redisplacement of femoral neck fractures fixed with multiple cancellous screws in a patient population. DESIGN: A retrospective analysis of quantitative geometric and bone density data from forty-seven patients with femoral neck fractures fixed with cancellous screws was performed. SETTING: Radiographs retrieved after analysis of a statewide hospital database were digitized and analyzed quantitatively. INTERVENTION: Quantitative data from radiographs included estimates of femoral neck bone density, fracture surface orientation and location, degree of comminution, prefixation displacement, initial reduction position of the fractured component, and final displacement of the fracture after fixation. MAIN OUTCOME MEASUREMENTS: Angular rotation and inferior displacement of the fracture component after fixation. RESULTS: Femoral neck bone density can be correlated to femoral cortical thickness and can be used as a measure of bone density from plain x-rays. Significant relative risk of redisplacement of a femoral neck fracture is correlated with initial inferior offset of the fracture component and varus angulation. Relative risks of other variables, including valgus reduction, Garden Stage 3 and 4 position of the femoral head, low bone density, presence of inferior comminution, a more vertical fracture surface angle, and fracture position, were not significant. CONCLUSIONS: When compared with other geometric and mechanical variables, nonanatomic reduction of a femoral neck fracture, with either inferior offset or varus angulation, is the strongest predictor of postfixation redisplacement of the fracture.

Effect of variations in calcaneocuboid fusion technique on kinematics of the normal hindfoot.

Calcaneocuboid fusion with lengthening of the lateral column of the foot has been advocated as a method of treating flatfoot deformity. This study was designed to determine how the length of the lateral column chosen or the position of the foot selected when performing this fusion affect hindfoot kinematics in normal cadaver feet. An electromagnetic tracking system was used to monitor the positions of the talus, calcaneus, navicular, and cuboid while the intact cadaver feet were moved passively and then under reproducible loads. Calcaneocuboid fusion was then performed on these feet first with the feet in neutral position and the lateral column of normal length, then lengthened 10 mm or shortened 5 mm, and then with the lateral column lengthened 10 mm and the feet positioned in plantar flexion and eversion or dorsiflexion and inversion. Kinematic measurements were made at each stage using the same loads. Fusing the calcaneocuboid joint with lengthening or shortening the lateral column and the feet in neutral position did not affect hindfoot joint motion compared with intact. Changing the position of the foot for fusion, however, resulted in significant decreases in motion in the talocalcaneal and talonavicular joints. Tibiotalar joint motion was unaffected. This study, therefore, demonstrates that when fusing the calcaneocuboid joint, attention should be paid to maintaining a neutral position of the foot.

The effect of post-injury spinal position on canal occlusion in a cervical spine burst fracture model.

STUDY DESIGN: The canal space of burst-fractured, human cervical spine specimens was monitored to determine the extent to which spinal position affected post-injury occlusion. OBJECTIVE: To test the null hypothesis that there is no difference in spinal canal occlusion as a function of spinal positioning for a burst-fractured cervical spine model. SUMMARY OF BACKGROUND DATA: Although previous studies have documented the effect of spinal positioning on canal geometry in intact cadaver spines, to the authors' knowledge, none has examined this relationship specifically in a burst fracture model. METHODS: Eight human cervical spine specimens (levels C1 to T3) were fractured by axial impact, and the resulting burst injuries were documented using post-injury radiographs and computed tomography scans. Canal occlusion was measured using a custom transducer in which water was circulated through a section of flexible tygon tubing that was passed through the spinal canal. Any impingement on the tubing produced a rise in fluid pressure that was monitored with a pressure transducer. Each spine was positioned in flexion, extension, lateral (and off-axis) bending, axial rotation, traction, and compression, while canal occlusion and angular position were monitored. Occlusion values for each position were compared with measurements taken with the spine in neutral position. RESULTS: Compared with neutral position, compression, extension, and extension combined with lateral bending significantly increased canal occlusion, whereas flexion decreased the extent of occlusion. In extension, the observed mechanism of occlusion was ligamentum flavum bulge caused by ligament laxity resulting from reduced vertebral body height. CONCLUSIONS: Increased compression of the spinal cord after injury may lead to more extensive neurologic loss. This study demonstrated that placing a burst-fractured cervical spine into either extension or compression significantly increased canal occlusion as compared with occlusion in a neutral position.

Biomechanical comparison of internal fixation techniques for the treatment of unstable basicervical femoral neck fractures.

OBJECTIVES: The optimal method of internal fixation of basicervical femoral neck fractures is controversial. This area represents a transition zone between the intracapsular femoral neck fracture, usually fixed with multiple cancellous screws, and the extracapsular interochanteric fracture, fixed with a sliding screw device [sliding hip screw (SHS)] and derotation screw (DRS) device. The authors' specific aim was to biomechanically compare these two methods of fixation in a cadaveric model of a basicervical femoral neck fracture with posteromedial comminution. DESIGN: The authors compared the average peak force during cyclic loading and the maximum axial force sustained by matched pairs of specimens stabilized with either fixation and subjected to axial and torsional loading while flexing and extending the hip. The average peak force was defined as the mean of the peak force values measured in each loading cycle with the maximum displacement of the materials tester actuator the same for each cycle (displacement control) as opposed to the maximum force being held constant (load control). RESULTS: The cancellous screw group maintained a significantly lower average peak force, 470 Newtons (SD = 145 Newtons), compared with 868 Newtons (SD = 186 Newtons) for the SHS and DRS composite group (p < 0.01). Similarly, the cancellous screw group demonstrated a lower ultimate load to failure, 1,863 Newtons (SD = 475 Newtons) compared with 3,557 Newtons (SD = 215 Newtons) for the SHS and DRS composite group (p < 0.01). CONCLUSION: The results support the use of an SHS and DHS composite compared with three cancellous screws in the treatment of unstable basicervical femoral neck fractures.

Effects of cyclical mechanical stress on the controlled release of proteins from a biodegradable polymer implant.

The availability of osteogenic proteins for orthopedic applications has led to great interest in developing delivery systems for these substances. Standard release rate models are applicable in most biological settings, but orthopedic implants usually bear mechanical loads. To determine whether a release rate model for load bearing applications must consider mechanical stress, the effects of dynamic mechanical stress on the in vitro release kinetics of two model proteins, bovine albumin (BA) and trypsin inhibitor (TI), from a biodegradable film were evaluated. Biodegradable poly(lacticco-glycolic acid) cylindrical implants with embedded proteins were subjected to cyclic three point bending loading of 720 cycles/day at 0.4 Hz for 2 weeks. Protein release into solution, swelling and mass loss changes, molecular weight degradation, and the presence of microstructural stress cracks and pores in the polymer carrier were evaluated. Cumulative BA and TI releases with time were significantly higher when a cyclic bending load was applied and increased with the magnitude of the load. Mass loss was not significantly greater, nor was swelling or molecular weight change of the polymer carrier in this 2-week interval. Pores on the surface of the polymer in the highest stress region were elongated into cracks, compared with pores in the low-stress region of the same implant, which were roughly circular. This implies that the pores probably act as stress risers to initiate cracks, which then expose more surface area, increasing protein release.