Purpose of the study: The aim of this study was to analyze rotation of the normal and prosthetic distal femur as well as the spaces from 90 to 130 degrees flexion.

Material and methods: Torsion scans were obtained preoperatively and postoperatively for 44 total knee prostheses. The difference in femoral torsion between the pre- and postoperative image was used to assess the rotation in which the femoral component was implanted. The prostheses were divided into two groups: group I when the femoral implant was implanted with external rotation of more than 5°; group II when the femoral implant was implanted with external rotation less than 5°. A preoperative stress scan was obtained in 20 patients then repeated during the year following implantation. Stress images with knee flexion at angles from 90° to 130° were obtained. The patient was installed in the ventral supine position. 8mm scan slices were centered on the lower end of the femur, ten 50ms images were acquired during flexion movement from 90° to 130°. This enabled determination of the knee flexion axis preoperatively and postoperatively, to measure the variation in the epicondylar axis compared with the mechanical axis of the tibia between 90° and 130° flexion and finally to deduct change in the femorotibial space in flexion from 90° to 130°.

Results: The 18 total knee prostheses with a femoral component implanted with external rotation greater than 5° (group I) showed significantly greater range of flexion (p< 0.05) (mean 120°, range 110°–130°) than the 26 prostheses in group II with a femoral component implanted in external rotation less than 5° (mean 100°, range 80°–115°. For the 20 knees with stress scans, the preoperative images showed an epicondylar axis about 5° fro the mechanical axis of the tibia when the knee flexed in the 90°–130° range. After surgery, the stress scans showed that this epicondylar axis of rotation of the prosthesis-bearing knees occurred especially for knees with a wide range of flexion. The 20 knees with flexion limited to 100° did not present an epicondylar rotation axis compared with the mechanical axis of the tibia. The 15 knees with 125° flexion or more had an epicondylar axis of rotation after 90° flexion. Rotation of the epicondylar axis in relation to the mechanical axis of the tibia between 90° and 130° flexion was the consequence of a femorotibial space which changed in the medial and laeral femorotibial compartments between 90° and 130° flexion: after 90° flexion, the medial femorotibial space decreased and the lateral femorotibial space increased. This explains why movement from 90° flexion to 130° flexion was facilitated by placing the femoral piece in external rotation.

Discussion: Search for ligament balance for knee flexion above 90° is logical only if the goal is to obtain knee stability in extension and flexion at 90°. It is probably no rational if the goal is to allow the knee to reach flexion in the 120°–130° range. Ligament balance in flexion above 90° is important and should be maintained up through 130° flexion. The other solution is to empirically increase external rotation of the femoral component a few degrees in order to allow greater range of flexion.

Purpose of the study: The humeral head is the second most frequent localization of non-traumatic osteonecrosis. For certain etiologies, for example sickle-cell anemia, the frequency is similar to that observed for the femoral head. There have nevertheless been very few publications on this pathology and its treatment. The purpose of this study was to assess outcome in a series of 771 cases of humeral head osteonecrosis in order to establish the natural history of the disease, criteria predictive of outcome, and therapeutic options.

Material and methods: The diagnosis of osteonecrosis of the humeral head was established for 771 humeri in 424 patients between 1981 to 2000. Minimum follow-up was five years (maximum 23 and mean 13 years). Outcome was assessed in terms of the clinical course, specifically the need for surgery due to pain or functional impotency. The radiological assessment was made on serial AP and lateral views taken every year or two years. The extent of the osteonecrosis was assessed on the basis of the magnetic resonance imaging (MRI) findings when available (after 1985). The ARlet and Ficat classification established for the femoral head was adapted to the shoulder: grade I: osteonecrosis of the humeral head visualized solely with MRI; grade II: radiologically detectable osteonecrosis; grade III: subchondral dissection without loss of spherical shape: grade IV: loss of spherical shape without visible osteoarthritis: grade V: osteoarthritis.

Results: Bilateral osteonecrosis was observed in 82% of the 424 patients. This gave 771 cases of humeral head necrosis. There was no gender predominance. Mean age at diagnosis was 32 years (range 18–57 years). The most frequent etiology was sickle-cell anemia (307 patients), followed by corticosteroid therapy (80 patients). Other etiologies were much less frequent: alcohol abuse, Gaucher’s disease, hyperlipidemia. Osteonecrosis of the humeral head was generally associated with another localization, particularly involving the hip and the knee joints. Multifocal osteonecrosis was also a common finding. Among the patients whose dignosis of osteonecrosis was established before symptom onset (scintigraphy or MRI performed in patients with multifocal osteonecrosis), the natural history was on average three years between MRI diagnosis and onset of pain. For 46% of the cases, pain appeared at grade I, before the development of radiographic signs. In 54% of the cases, grade II occurred before pain. It took six years before all of the cases with osteonecrosis diagnosed in a non-symptomatic phase produced pain. Factors affecting the rapidity of the radiological course were: etiology, size of the necrotic focus, presence and rapidity of osteonecrosis in other localizations (hip and knee). The humeral head lost its spherical shape on average four to five years after the diagnosis of osteoarthritic degradation of the joint, at about seven to eight years of evolution. Among the 256 patients followed for more than ten years, 51% required surgery. These 131 operations were for: drilling with bone marrow grafting (grade I or II) (n=62), cimentoplasty after loss of spherical shape but before glenohumeral osteoarthritis (n=15), resection of sequestered necrosis after loss of spherical shape (n=12), shoulder arthroplasty (n=42).

Discussion and conclusion: This study demonstrated that the natural history of osteonecrosis of the humeral head has a poor long-term outcome. Shoulder arthroplasty is rarely required during the first decade of the disease. Other therapeutic alternatives can help avoid or retard the need for shoulder arthroplasty in these very young patients.

Purpose of the study: It is currently accepted that ligament balance should be one of the goals for total knee arthroplasty (TKA) and that this balance should be obtained by correct bone cuts or appropriate ligament procedures. There is however no standard way of assessing this balance. The purpose of this study was to define limit values for knee laxity observed in a series of normal knees and in a series of 54 TKA reviewed at more than ten years.

Material and methods: Laxity in extension of normal knees was measured on forced varus and valgus films using the contralateral knees of patients who had undergone knee surgery for osteotomy or prosthesis implantation. Laxity in extension of TKA knees was measured the first postoperative year and at last follow-up by measuring the decoaptation between the tibial and femoral pieces on single-leg stance films. The change in decoaptation over time was compared with the postoperative and last follow-up goniometry figures, the IKS knee score, the number of loosenings and the number of lucent lines. Multifactorial analysis was considered significant at p< 0.05.

Results: For the normal knees in extension, the medial compartment gap was 2 mm on average (range 1.5–3.5 mm) on the forced valgus images and the lateral compartment gap was 3 mm on average (range 2–4 mm) on the forced varus images. The corresponding angular value was 1° decoaptation on the forced valgus images and 1.5° on the forced varus images. Among the 54 knees with a TKA, the first postperative single-leg stance image revealed a lateral decoaptation _ 3° for 12 knees considered to present laxity, and was _ 2° for 42 knees considered not to present laxity. At last follow-up (13 years on average, range 11–14 years) the 42 knees without laxity remained unchanged without decoaptation, including the 34 normocorrected knees (±3°) and the eight undercorrected knees presenting more than 3° varus (mean undercorection 5°, range 3–7°). The 12 knees presenting postoperative radiographic decoaptation _ 3° showed at last follow-up a significant increase in laxity (p< 0.05) and 2.5° further increase in decoaptation. The increase in decoaptation occurred on normocorrected (n=7) or undercorrected (n=5) knees. This increase in decoaptation was greater with greater residual genu varum. Four groups of knees could be distinguished: normocorrected and stable; normocorrected and unstable; undercorrected and stable; overcorrected and unstable. The number of loosenings requiring revision and the number of progressive lucent lines were significantly greater among unstable knees (two loosenings, and five progressive lucent lines) than among stable knees (no loosening or lucent lines). They were also greater in the group of normocorrected and unstable knees (one loosening and two lucent lines) than in the group of undercorrected and stable knees (no loosening or lucent line). The IKS knee score of stable knees was higher than that of unstable knees irrespective of the correction (p< 0.05).

Discussion: Postoperative laxity in varus with angular decoaptation greater than 3° corresponds to a lateral compartment gap and should be avoided even if the knee is properly aligned postoperatively. If the knee is stable, moderate undercorrection (3–5° varus) does not appear to have an unfavorable long-term effect on knee laxity or on the femoral and tibial pieces.

Conclusion: For knees with constitutional genu varum, moderate undercorrection with a stable knee is preferable to normocorrection at the cost of lost stability.