In an animal model we determined the strength of anterior cruciate ligaments (ACL) after section and repair by four different methods and compared it with that of the intact ligament. The standard suturing technique of multiple loops through the ligament stumps was used. Stronger suture material did not give a stronger repair. Wrapping a fine polyester mesh around the ligament or placing it between the bundles before suture increased the strength of the repair. This modification, allied to protective rehabilitation, may reduce the failure rate of acute ACL repairs.

Many methods of reconstruction for ACL deficiency have been described, but little is known about their biomechanical properties. We examined extra-articular (EA), intra-articular (IA) and combined (EA+IA) reconstructions in ten cadaver knees after the ACL had been ruptured by the performance of a rapid anterior drawer movement. Stability at each stage before and after rupture and reconstruction was tested by anterior drawer, Lachman, varus-valgus and tibial rotation tests. Both IA and IA+EA reconstructions restored normal stability, while EA reconstructions improved stability but did not restore it to normal. The addition of an EA procedure to an IA procedure made no difference to knee stability. We conclude that in cases of isolated ACL deficiency there is no biomechanical basis for EA reconstruction, either alone or in addition to an IA reconstruction.

Compression testing of cadaver specimens showed that excision of the radial head allowed proximal radial displacement. The insertion of a metallic radial head restored normal mechanics, while a silicone rubber implant did not. We reviewed 31 of 36 comminuted fractures of the radial head, 21 associated with dislocation or ulnar fracture, which had been treated by primary replacement with a Vitallium prosthesis. At a mean follow-up of 4.5 years, there was reliable restoration of stability and prevention of proximal radial migration. There had been no dislocations or prosthetic failures, but two implants had been removed for loosening. The prosthesis is recommended for use as a spacer to stabilise the elbow after severe injuries while the soft tissues heal.

Using 26 cadaver shoulders, we produced a standard defect in the supraspinatus tendon and performed one of three types of repair. Their strength was found by testing in tension the force required to produce a gap of 3 mm, then 6 mm, and finally total disruption of the repair. The use of a polyethylene patch to spread the forces over the lateral bone surface and of extra sutures to grasp the tendon end raised by 2.6 times the load at which a 3 mm gap in the repair occurred and by 1.7 times the load to failure.

We excised the anterior cruciate ligament from the left stifle of 24 sheep and replaced it by a polyester fibre implant routed 'over the top' of the femoral condyle and fixed, using grommets and screws. All the joints were sound, and the animals moved normally until they were killed at six, 12 and 24 months after operation. We found that the implants were always covered by host tissue, which matured into bundles with a histological appearance similar to the natural ligament. The implants were joined to the bones by organised fibrous tissue and there was no anchorage loosening. There was no synovitis, but the operated joints showed progressive cartilage degeneration. The reconstructed joints became less stable immediately after operation, but regained normal stability as the neoligaments developed. The neoligaments lost strength with time, despite tissue ingrowth. The good functional, biomechanical, and histological results justify clinical trials of this type of implant.

This work studied the fibre bundle anatomy of the anterior cruciate ligament. Three functional bundles--anteromedial, intermediate, and posterolateral--were identified in cadaver knees. Their contributions to resisting anterior subluxation in flexion and extension were found by repeated tests after sequential bundle transection. Changes of length in flexion and extension and in tibial rotation were measured. None of the fibres were isometric. The posterolateral bundle was stretched in extension and the anteromedial in flexion, which correlated with increased contributions to knee stability and the likelihood of partial ruptures in these positions. Tibial rotation had no significant effect. The fibre length changes suggested that the 'isometric point' aimed at by some ligament replacements lay anterior and superior to the femoral origin of the intermediate fibre bundle and towards the roof of the intercondylar notch.

We report the experimental use of three different biological implants to restore articular surface defects: glutaraldehyde-fixed bovine meniscal xenograft, glutaraldehyde-fixed bovine costal cartilage xenograft, and viable osteochondral allografts. The grafts were implanted in the knees of 19 goats who were allowed free-field activity and were studied for up to one year. The natural articular surfaces of meniscal fibrocartilage provided excellent articular surfaces at all times. Equally good articular surfaces were restored by host tissue growth covering costal cartilage grafts at six months, but by 12 months this surface had degenerated. The majority of the allografts survived and integrated with the host at six months, but many showed signs of failure at 12 months. Only three out of seven ungrafted defects healed completely at six months and the healed surfaces were degenerating at 12 months.

We have assessed the biomechanical properties of a 'double-bundle' prosthetic ligament replacing the anterior cruciate in cadaver knees. We compared the results with those of single bundle 'over-the-top' and 'through-the-condyle' techniques, performing anterior drawer tests at 20 degrees and 90 degrees knee flexion. The over-the-top reconstruction gave better anteroposterior stability at 20 degrees, while the through-the-condyle repair was more stable at 90 degrees. The double-bundle reconstruction gave practically normal anterior stability at both 20 degrees and 90 degrees.

The anteroposterior stability of cadaveric knees was investigated. There was a wide range of normal laxity; knees were more stable at 90 degrees than at 20 degrees flexion. Anterior cruciate ligament implants with different stiffnesses were inserted; normal stability could always be restored, and the stiffness or extensibility of implants did not affect knee behaviour significantly. The tightness of implants was critical--small tensioning errors caused subluxation, inhibited knee extension and allowed damagingly high implant tensions. It is concluded that the tension of ligament implants could not be adjusted simply with a pre-set instrument; the procedure will remain critically dependent on the judgment of the operating surgeon.

Using dried bones which could be tilted and rotated, we assessed the accuracy of published radiographic methods for measuring the migration of prosthetic acetabular components and compared the results with a new method. The new line linking acetabular margins was significantly more accurate for proximal migration than the teardrop, the sacroiliac line or the sacroiliac-symphysis line. For medial migration, a new line tangential to the brim and through the horizontal mid-point of the obturator foramen was more accurate than Kohler's line, the ilio-ischial line or the iliopubic line. In combination, the two new lines can give a more accurate assessment of acetabular erosion than previous methods, since they are less affected by the differences in rotation commonly found in a series of radiographs.

The anterior cruciate ligament was replaced in rabbits, using implants of carbon or polyester filaments with known mechanical properties. The biocompatibility of the implants was assessed in detail using light microscopy, and scanning and transmission electron microscopy. Mechanical tests were made of stability, in comparison with normal joints and controls after excision of the ligament. Some carbon fibre implants broke down in vivo, allowing instability; the fragments caused chronic inflammation. Intact carbon implants did not induce the formation of neoligaments; they were covered by tissue, but there was no ingrowth. Polyester did not degrade mechanically and supported early collagenous ingrowth within the implant, even in the mid-joint space. It was concluded that there was no justification for the use of carbon fibres as anterior cruciate replacements; polyester appeared to be suitable.

We examined the structure of the digital flexor sheath by dissection and histology. The inner aspect of the sheath was found not to be a continuous smooth surface, as depicted in anatomical and surgical texts. The thin parts of the sheath often overlapped the pulleys before attaching to their superficial aspects, so that the pulleys possessed free edges within the sheath. The frequency of occurrence and sizes of these overlaps were studied in 48 cadaveric fingers; the largest and most frequent overlap was at the distal end of the A2 pulley. Functional studies showed an intricate mechanism of pulley approximation and sheath bulging during flexion. Sutured or partly cut tendons triggered on the free edges; this could be a major contributor to the failures of tendon repairs in "no man's land".

This study aimed to compare the holding strength of various commercially-available anchorage devices for artificial ligaments, so that surgeons might make a reasoned choice. Tensile tests to failure were performed on screws, bollards, toggles and staples which had been implanted into cadaveric bones. The holding strength of all devices correlated significantly with the local thickness of cortical bone, so it is recommended that anchorages should be placed away from the joint line, into diaphyseal bone if possible. A new trans-cortical grommet was developed which, when used around an AO screw, had significantly greater holding strength than the other devices.

The calcaneal tendons of rabbits were excised and either replaced with a carbon or polyester fibre implant, or left as controls. The strength of the neotendons and their mode of failure under tension were examined at intervals up to six months after operation. Return to near normal strength took six months to develop, suggesting that patients having ligament or tendon reconstructions should not resume normal activity for several months. Carbon fibre-based neotendons showed progressive elongation which, unless avoided by a sufficient period of immobilisation, would affect the functional result.

Carbon-fibre and polyester-fibre implants of comparable dimensions were used to replace the calcaneal tendon in 30 sheep. The neotendon produced in proximity to the polyester fibres was denser, more collagenous and more closely adherent than that in the carbon-based neotendon. Fragmentation of the carbon caused continuing cellular reaction which was associated with a poor collagen response.

The effects of synovectomy and excision of the radial head in 40 elbows affected by rheumatoid arthritis have been assessed. In contrast with many reports indicating minimal side-effects of this procedure, a common pattern of deterioration from what was often a satisfactory initial result has been demonstrated. A biomechanical theory of one of the factors responsible for failure has been put forward and the importance of conserving or replacing the radial head emphasised.