PURPOSE: Oldest elderly trochanteric fractures treatments are common but long term follow up studies are indistinct. Previous data about all outcomes are not reported for oldest one. The purpose was to update all the results for these oldest old patients.

METHODS: A prospective, longitudinal study was undertaken of 455 “31 A1, A2, A3” (OTA) peritrochanteric femur fractures through 2002. There were 84% female patients. The age was 88 (80–105) ±5 years. Associated hip arthritis rate was 10%. 31 A1 and 31 A2 really trochanteric fractures were 90% (31 A1: 52% and 31 A2: 38%), 31 A3 subtrochanteric were 10%. Surgical treatment was achieved for 97%: [extramedullary internal fixation (dynamic hip screw): 36%, intramedullary (trochanteric nail): 56%, hip arthroplasty: 5%]. Functional treatment (not displaced or contra-indicated) was achieved for 3%. Katz, mental, Parker, walking scores, live place before and at long term after fracture, radiological healing and complications were registered.

RESULTS: The radiological healing rate after 3 months was 85% and after 6 was 97%. Complications rate: the general complications rate during 1st month was 12% and between 1st and 3rd month was 8%. The local complications rate during the 1st month was 4% (local infection: 1%) and between 1st and 3rd month was < 1%. The new surgical procedures rate during the 1st month was < 0.5%, between 1st and 3rd month was 3%, between 3rd and 6th month was 3%. Clinical results: Post-operative weight bearing was uncertain. 6th month after fracture 72% of the patients were alive (dead: 28% and non-surgically treated: twice more). All scores and live place demonstrated dependence increase. > 6th month after fracture mortality was common.

CONCLUSION AND SIGNIFICANCE: Intra/extra capsular hip fractures ratio increases after 80. These are older, more dependant and help demanding. Hip arthritis inflates hip fracture risk. Success points technical procedures had been demonstrated by randomised studies: dynamic hip screws for stable fractures and trochanteric nails for unstable; hip prosthesis for arthritis or poor bone. Practice surgery for all trochanteric fractures. Don’t separate trochanteric/subtrochanteric. The prognostic is rather poor (mortality rate, functional outcomes), depending on initial functional score and dependence. Modern internal fixation is reliable. Arthroplasty should be considered.

Purpose of the study: Treatment of supra- and inter-condylar fractures of the femur remains a difficult challenge, irrespective of the method used, because of the high risk of infection, disassembly, nonunion, joint stiffness, osteoarthritis, and multiple operations. Use of a supracondylar retrograde nail, accepted for C1 and C2 fractures, can be used for some C3 fractures depending on the stability of the epidphyseal assembly.

Material and methods: This series included 19 C3 fractures (AO classification) operated on in 1993–2000. Mean patient age was 54 years (range 30–81), 11 females and 8 males. This consecutive series of patients had: high energy trauma (n=14), low-energy trauma (n=5), multiple fractures (n=16), open fractures (n=10). Osteo-synthesis was performed on an ordinary table in the dorsal supine position with arthrotomy and epiphyseal screw and pin fixation followed by static supracondylar retrograde nailing (Smith and Nephez GHS), completed in two cases with an autologous corticocancellous graft. Kinetec was used for mobilization and weigh bearing delayed until bone healing.

Results: Twelve secondary operations were performed: cover with muscle flap (n=1), early revision for rotation misalignement (n=1), autologous graft (n=4), surgical arthrolysis (n=6), revision for nonunion (n=4). There were no infections. Among the four cases of nonunion, three involved epiphyseal screw failure, two cases having involved grafts. All four cases were treated by decortication, graft and plate fixation; healing was achieved. Mean time to bone healing per primam was 23 weeks on average. The 19 patients were examined at mean 44 months follow-up (range 16–78 months). Pain was noted: absent (n=8), mild (n=3), moderate (n=6), severe (n=2). Gait was noted: normal (n=7), slight limp (n=9), important limp (n=3). Mean flexion was 114° (range 85–150°). Five cass had permanent flexion < 10°. Radiologically, misalignment of +5° in the frontal plan was observed in six cases.

Discussion: Retrograde nailing of C3 fractures is difficult, but possible and requires first epiphyseal fixation then diaphyseal solidarization. The assembly is reliable, allowing immediate mobilization. Weight bearing must however be delayed to bone healing. Complementary surgery to graft bone stock or for relative arthrolysis has to be integrated into the operative plane for more than half of these difficult cases.

Purpose of the study: The place for total ankle arthroplasty versus fusion remains a subject of debate for the treatment of painful stiff ankles.

Material and methods: This series included 58 total ankle arthroplasies performed in 56 patients between 1991 and 2003. Mean paient age was 52 years (range 27–84). The underlying cause was multiple trauma (n=27), rheumatoid arthritis (n=17), chronic instability (n=11), hemophilia (n=2), primary disease (n=1). Four implants were used: New Jersey (n=22), Albatros (n=4), Star (n=10), AES (n=22). Associated procedures were: lengthening of the Achilles tendon or vastus (n=28), lateral ligamentoplasty (n=6), fibular osteotomy (n=2), medial ligamentoplasty (n=1), calcaneal osteotomy (n=3), double arthrodesis (n=1). The patients wore a plaster cast for 21 to 40 days.

Results: Eight patients were removed from the analysis: death (n=4), foreign residence (n=1, 3 follow-up shorter than one year (n=3). The analysis retained 50 total ankle arthroplasties in 48 patients reviewed with a mean 49 months follow-up (range 1–12 years). Reasons for surgery were: trauma (n=25), rheumatoid disease (n=12), instability (n=10), hemophilia (n=2), primary (n=1). Implants were: New Jersey (n=17), AES (n=19), Star (n=9), Albatros (n=3). Complications were: intra-operative medial malleolar fractures which were pinned (n=8), immediately revised radiological instability (n=2), wound dehiscence treated with a flap (n=1), secondary fusion (n=4). There were no cases of infection. There were six failures (12%) leading to implant removal for loosening (n=3), pain (n=2), instability (n=1) and revision arthrodesis (n=4) or new arthroplasty (n=2). The 44 remaining cases were analyzed: AOFAS score improved from 40/100 to 73/100 at last follow-up. Joint motion was 24° preoperatively and 20.5° postoperatively (dorsal flexion −1° to +6°, plantar flexion 25° to 14.5°).

Radiographically five prostheses were unstable with potential loosening (3 tibial and 2 talar components), one presented varus misalignment, and the others were considered correct. Moderate to severe intra-articular osteophytes were noted in 11 ankles. Three presented an undetermined defect image in the tibia.

Discussion: Indications for total ankle arthroplasty are exceptional and different from total prostheses for the knee or hip joints. Patients are young subjects with stiff, misaligned, unstable ankles, generally resulting from traumatic injury. Surgery is a challenge and requires several complementary procedures. Failure rate is higher than for the knee or the hip but mid-term results are encouraging. Further follow-up is needed for long-term confirmation.

Purpose: The purpose of this work was to validate the most reliable technique for obtaining ligament isometry of the knee in flexion.

Material and methods: This prospective series of non-cemented non-posterior stabilised Interax knee prostheses (Howmedica) implanted by the same surgeon included 57 genu varum knees with degenerative disease. Mean preoperative femorotibial varus was 8.23°. Landmarks used intraoperatively included the posterior condylar line (PCL), the biepicondylar line (BECL) and the Whiteside line (WL). Ligament balance was measured with the Derby tensor.

Results: The first part of the assessment concerned ligament balance in extension. Mean initial medial retraction, measured with the tensor, was 3.6°. Release of the concavity was performed in 62% of the knees with mean residual retraction of 1°. The second time was to evaluate balance at 90°. Using the anatomic landmarks, the PCL was parallel to the BECL in 22% of the knees and perpendicular to the WL in 26%. There was a weak angulation in 28% and 30% of the knees and in 50% and 44% respectively. There was thus a strong correlation intra-operatively between these two landmarks and the initial radiographic varus. Evaluation with the tensor showed mean 2.96° medial retraction. The correlation between the anatomic measures and the tensor ligament measures was very significant. When the BECL was parallel to the PCL, medial retraction with the tensor was 1.12°. The angulation was small, 2.25°. When the angle was wide, the mean measure was 4.4°. We found the same results with the WL. External rotation of the anteroposterior femoral cut was then guided by these different measures. It was 2.6° on average (0–6°). Residual medial retraction, measured with the tensor, was thus significantly improved, only 0.4° on average (−2 to +2).

Discussion: After ligament rebalancing in extension, there persisted frequently an imbalance in flexion (62%). This was independent of the preoperative varus. It was corrected by external rotation of the femoral implant, the value assess approximately from the landmarks. It was measured in our hands reliably with the Derby ligament tensor.

Purpose: Osteolysis is one of the important issues during the life of noncemented total hip arthroplasty (THA). The purpose of this study was to evaluate a series of 228 THA using an ABG-1 implant to determine the incidence of osteolysis and contributive factors.

Material and methods: This series of 228 THA using noncemented ABG-1 anatomic implants coated with hydroxyapatite was implanted in 210 patients, 116 women and 112 men, mean age 62.2 years. The indication was primary osteoarthritis (53.6%), primary necrosis (21.5%), posttraumatic osteoarthritis (11.8%), rheumatoid disease (8.3%) and hip dysplasia (4.8%). The bearings were, 200 zirconium-polyethylene (87.7%), 28 metal-polyethylene (12.3%). At mean 88.6 months follow-up, a minimum 60 months postoperatively, we reviewed 163 patients (37 patients died without revision, 28 (12.3%) were lost to follow-up, and 17 had undergone a revision procedure). We analyzed osteolysis on the digitalized radiograms using the Delee-Charnley classification to which we added a fourth retroacetabular zone. Polyethylene wear was measured with the Imagika(r) software using the method described by Martell.

Results: The 17 cup revisions (8.5%) were required for instability (4 pt), loosening (4 pt), osteolysis (4 pt), infection (3 pt), and limping (1 pt). Overall implant survival was 92.1%. Implant survival, considering loosening and osteolysis as failure, was 96.1%. Mean wear at last follow-up was 1.26 mm, for an annual average of 0.17 mm (0.04 – 0.69 mm/yr). Acetabular osteolysis was observed in 41.6% of cases, mainly in the Delee-Charnley zone 1, but was also found in all the other zones. The average surface area was an estimated 223 mm2. Possible factors favoring osteolysis were: aetiology, age, gender, activity level, body mass index, Charnley ABC classification, presence of preoperative acetabular defects, cup size, polyethylene insert thickness, position of the insert rim, cup inclination, complementary fixation, bearing type, polyethylene offset and wear. There was a statistically significant relationship between osteolysis and: the Charnley classification (p=0.012), presence of preoperative acetabular defects (p=0.0034), cup inclination angle (p=0.035), cup size (p=0.042), polyethylene thickness (p< 0.01), use of complementary fixation (p=0.048), and polyethylene wear (p=0.0011). Paradoxically, we did not find any relationship with gender, age, body mass index, and the other factors.

Discussion: This analysis demonstrated the determining causes of osteolysis: time, polyethylene thickness, polyethylene wear, the Zicronium-polyethylene bearing. Wear was probably not only the consequence of the Zicronium-polyethylene bearing but was also related to the quality of the polyethylene, and the instability of the insert. Diffuse osteolysis is favored by orifices in the first-generation cups which have been eliminated in the ABG-2 cups.

Purpose: Cephalo-tuberosity (CT) fractures are complex fractures with a serious prognosis. Appropriate treatment is highly debated. We report long-term clinical and radiological results in a series of 34 patients treated by osteosynthesis.

Material and methods: The series included 34 patients (21 women and 13 men) who underwent surgery between 1987 and 1997. Mean age was 61 years. The dominant side was involved in 18 cases. There were 14 traffic accident victims and 20 fall victims. Fracture types (Duparc classification) were: CT2 (n10), CT3 (n=18), CT4 (n=6). Nervous complications were associated in five cases. Treatment consisted in closed osteosynthesis for 16 cases and open plate fixation or pinning for 18 cases. Constant score was used to assess functional outcome using the pain, activity, motion, and force scales. Radiological results were assessed on the AP and lateral axillary views. Statview was used for statistical analysis. Immobilisation was maintained for 28 days.

Results: All patients were seen for physical examination and radiographs at a mean follow-up of 40 months. Only four patients were fully satisfied with the outcome. The Constant pain score was rated 9 points. Thirteen patients contralateralised since their trauma. Active antepulsion was 97 and external rotation 30 without a statistical correlation between mobility and type of fracture. The overall mean Constant score was 60 points. Reduction was considered anatomic in 12 cases with a deformed callus in 14 cases. There were eight cases of secondary cephalic necrosis (four among the CT4). The most common complication was pin migration (n=15) and disassembly (n=7) with a significant correlation between complications and age.

Conclusion: The objective and subjective results of this series demonstrated that complications increase with age. There is a significant correlation between external rotation and reduction of the greater tubercle. We did not find any radio-clinical correlation. Although the overall results were not satisfactory, they must be compared with those of arthroplasty.

Purpose: The purpose of this work was to evaluate the mid-term behaviour of 228 ABG-1 femoral stems implanted without cement in patients with at least five years follow-up.

Material and methods: The series included 228 ABG-1 (cup and femur) non-cemented hydroxyapatite-coated anatomic total hip prostheses implanted in 210 patients. Mean age was 62.2 years, 116 women, 112 men. Indications were: primary degenerative disease (53.6%), primary necrosis (21.5%), posttraumatic osteoarthritis (11.8%), rheumatoid disease (8.3%) and dysplasia (4.8%). A zincronium-polyethylene bearing was used for 200 hips (87.7%) and metal-polyethylene for 28. We retained for analysis 163 patients with mean 88.6 months follow-up [37 died without revision, 28 lost to follow-up (12.3%), 17 surgical revision]. The Postel Merle d’Aubigné (PMA) and Harris Hip Score (HHS) were noted. Radiographically, the AGORA Engh-Massin and ARA femur scores were determined; osteolysis was evaluated with the Gruen classification.

Results: The seventeen femoral revisions were related to aseptic loosening (n=4), fracture of the femur (n=4), sepsis (n=3), instability (n=3), pain (n=2), and limping (n=1). Overall survival was 92.5% at 88.6 months, relative survival (loosening) was 98.2%. At last follow-up, the PMA score was 17.3 and 92.7% of the patients had an excellent or good outcome (PMA > 14). The mean HHS was 96.4. Ninety-six percent of the patients were pain free. Radiologically, 112 stems were evaluated at last follow-up. The Engh-Massin classification showed that 102 stems presented certain bone ingrowth (85.2%), 15 suspected (12.3%), and two showed fibrous encapsulation with one case of implant instability. The ARA femur score was good or excellent in 70.5%, fair in 13.1%, poor in 16.4%. Hypertrophic cortical reaction was noted in Gruen zones 2–3 or 5–6 in 27.9% and was not correlated with pain. These reactions were however associated in more than half of the cases with a position anomaly in the frontal plan or an oversized stem. There was on incomplete pedestal and no extensive proximal stress shielding. Femoral osteolysis was very limited (mean surface area 145 mm2) in zones 1 and 7 in 21.3% of the cases. There were no cases of distal osteolysis.

Discussion: These data confirm the very good mid-term behaviour of the anatomic stems coated with hydroxyapatite. This metaphyseal coating acts like a barrier preventing diffusion of wear particles along the shaft. Stem migration was minimal, to the order of one millimetre, and did not continue beyond the first year. Over time, a constant trabecular metaphyseal endocortical bone reaction was observed, reflecting stress transmission to the proximal portion of the hydroxyapatite-coated femoral implant.

Purpose: We report the results at more than five years of a consecutive series of total knee arthroplasties (TKAs) implanted without cement and with preservation of the posterior cruciate ligament (PCL).

Material and methods: This series included 98 patients who underwent 109 primary TKAs between 1994 and 1998. Mean patient age was 67.7 years. The press-fit Interax prosthesis has a macroporous hydroxyapatite coating since 1996. The patella was resurfaced in 92% of the cases with a cemented polyethylene button. Ligament balance and joint space were controlled by progressive release of the ligaments with a tensor derby. Clinical assessment was based on the KSS. Radiological assessment included the femorotibial axis, the position of the implants, and the bone-implant interfaces.

Results: Seven (6.4%) of the patients were lost to follow-up. Fifteen (13.4%) died or were demented and two (1.8%) developed late infection and were not included in the primary analysis. Thus, 85 patients (77.9%) were retained for analysis at a mean 5.2 years follow-up. The KSS improved from 33.4 preoperatively to 79.4 at last follow-up. The KSS function score improved from 55.1 to 82.4. Knee flexion was 124.5° preoperatively and 113.1° at last follow-up. Anteroposterior laxity greater than 5 mm was observed in 4.8% at last follow-up. Radiologically, the mechanical femorotibial axis changed from 184.4° to 180.6°. Alpha 95.6°, geta 89.1°, omega 4.77°, gamma 3.8°, sigma 89.4°. Patellar height was 0.84 preoperatively and 0.65 at last follow-up. Lucent lines were observed 3.5 times less frequently with hydroxyapatite coated implants. Patellar loosening and femorotibial loosening were observed in one patient each. Revision procedures were necessary for three prostheses for bipolar loosening, painful stiff knee, and anteroposterior instability. Non-infectious survival rate was 94.1% at 5.2 years.

Discussion: These mid-term clinical results are comparable with those reported in the literature for cemented or non-cemented implants with or without preservation of the posterior cruciate ligament. Longer follow-up is however necessary to assess the anteroposterior stability. The radiographic measurements provide a good demonstration of the operative precision and proper ligament balance. Hydroxyapatite coating improves bony fixation of the implants and provides a fixation comparable with cemented implants.

Conclusion: This series confirms the mid-term reliability of the non-cemented Interax THA with preservation of the posterior cruciate ligament.

Purpose: Fractures involving both the shaft of the femur and the proximal portion of the femur are uncommon. In a meta-analysis of cases reported between 1951 and 1985, Alho recorded 659 cases where a wide range of management strategies were used. We report here a homogeneous series of 17 patients treated with ascending locked anterograde nailing.

Material: This series of 17 patients were young (mean age 36 years). These eleven men and six women were all victims of high-energy trauma; 12 had multiple fractures. The shaft fracture involved the middle third in 15 patients, the lower third in two; the shaft fracture was open in six cases. The proximal fracture was transcervical in nine patients (7 B21, 1 B22, 1 B23) and trochanteric in six (A A32, 1 A31, 2 A12, 2 A33).

Methods: The fracture was reduced under fluorescent guidance on the orthopaedic table in the supine position followed by anterograde nailing with ascending proximal locking in the axis of the neck using a Russel and Taylor reconstruction nail. The osteosynthesis was performed on day 0 in eleven patients, during the first week in three and later in three.

Results: Results are reported for 17 patients. There was one early superficial suppuration which healed favourable after local care. Two shaft fractures exhibited nonunion and were revised to decorticalise the graft. The cervical fracture exhibited early displacement in one patient who underwent revision on day 15; bone healing did not ensue and a total hip arthroplasty was implanted at ten months. All the other fractures healed within three to five months after the first-intention treatment. The long-term follow-up has revealed one case of cephalic necrosis at five years which has required a total hip arthroplasty.

Discussion: These double fractures involving the proximal femur and the shaft of the femur account for 1 to 5% of the femur fractures reported in the literature. They are observed in young victims of high-energy trauma, often associated with other multiple injuries. Diagnosis is not always easy to establish since there may be little or no displacement of the proximal fracture, which may be recognised secondarily after standard nailing (2 out of 17 cases).

The trochanteric fractures are generally easier to diagnose and reduce, and usually heal well. The shaft fractures are more often displaced and readily comminutive, sometimes open, having absorbed the greater part of the trauma energy. These fractures heal like ordinary shaft fractures. Neck fractures are often seen in the lower portion with a vertical fracture line, with or without displacement.

Using a single centromedullary nail for the osteosynthesis of both fractures is an attractive solution. The proximal fracture must however be carefully reduced with percutaneous pins before attempting nail insertion. The postoperative period is generally uneventful. Problems may be encountered if the cervical fracture cannot be perfectly reduced, in which case two separate fixations would be preferable.

Purpose: The purpose of this work was to assess the quality of the bone reconstruction in contact with the long hydroxyapatite-coated locked femoral stem used in a consecutive series of patients undergoing revision total hip arthroplasty (RTHA).

Material and methods: This series of 20 patients underwent RTHA for aseptic loosening (n=15) or septic loosening (n=5) of an AURA prosthesis. Mean age was 70.5 years. Mean time between insertion of the first stem and revision was 11 years for the aseptic patients and 2.6 years for the septic patients. Bony lesions of the femur were assessed with the SOFCOT classification: grade I=5, grade II=4, grade III=4, and grade IV=1. A transfemoral approach with a floating femoral segment was used in 14 patients and an endofemoral approach in six. The septic loosenings were reconstructed in two operative times in four patients and in one operation in one patient. An AURA reconstruction stem was used in 15 cases and a revision stem in five. A complementary cancellous bone graft with the endofemoral approach was used in six patients.

Results: Two patients died. All others were reviewed at a mean follow-up of 26 months (range 12–46 months) for clinical and radiographic assessment (five patients also had a supplementary scan at more than three years follow-up). We had three dislocations at 15 days with no recurrence and one case of sepsis at three months which cured after wash out and adapted antibiotics. All the femoral segments healed starting on the tenth week. The PMA score improved from 9.1 to 15.66 and the Harris score from 43.5 to 85.5. At last follow-up, all femoral lesions had moved to a lower SOFCOT score. There was a tight contact between the AURA stem and the femur on the last follow-up scan which showed an increase in the cortical index from 1 to 8 cm from the metaphyseal spine of the stem. None of the patients experienced secondary unlocking or required revision for a shorter stem.

Discussion: Revision after femoral failure with bone destruction using a non-cemented hydroxyapatite-coated stem allows immediate prosthetic mechanical stability and intimate bone reconstruction in the metaphyseal diaphyseal region. This reconstruction is real even if a graft is not used and appears to be favoured by the femorotomy. Implantation of the long stem is not particularly difficult and can even make the operation easier. Femorotomy has a real advantage, particularly for the revision of septic stems or in the event of difficult explantation.

Purpose: This prospective study was conducted to evaluate correction of post-traumatic deformity of the thoracolumbar and lumbar spine at consolidation after posterior instrumentation associated with transpedicular graft.

Material and method: The analysis began in 1998 and included 11 patients (mean age 37 years, age range 19–62 years) with lumbar (two L2) or thoracolumbar (four T12 and five L1) fractures. The Magerl classification was A1=2, A2=3, mixed=3. Inclusion criteria were: age < 65 years, regional kyphosis > 10°), intact pedicles, relative gain on local and/or regional kyphosis peroperatively > 70% after correction for reduction-distraction. Posterior instrumentation used the V-V assembly in two cases and 2VV-1VC assembly in nine. Evaluation criteria were, postoperatively and at consolidation: local kyphosis, corrected regional kyphosis, Cobb angle in the frontal plane. Autologous graft tissue harvested from the iliac crest was introduced via the right and left pedicle into the damaged vertebra using an adapted funnel after raising the vertebral plate with a spatula.

Results: Mean follow-up was 11 months (range 8–18). Associated procedures were laminectomy in three cases and posterior graft in two. Preoperatively local kyphosis was 19° (12° to 37°) and regional kyphosis was 17° (10° to 35°). Postoperatively, local and regional kyphosis were 4.5° (0–11°) and −2° (−10 to 5°) giving a mean relative gain of 80% (SD = 0.1). Mean relative gain at consolidation was 75% (SD = 0.2 for local kyphosis and 86% for regional kyphosis. The correction was statistically significant. There was no significant difference between the correction postoperatively and at consolidation.

Discussion: Transpedicular grafting associated with posterior instrumentation can avoid anterior access in a certain number of cases. Indications are fractures in patients under 65 years of age with a regional kyphosis > 10° and an important anterior defect after reduction (> 40% reduction in height). All patients achieved bone healing without significant loss of correction. The limitations of this technique include the requirement for intact pedicles, a posterior wall displacement of less than 60% and a peroperative relative gain greater than 70% for the local and/or regional kyphosis. An anterior approach must be associated in other cases.

Conclusion: Transpedicular grafting is a simple technique allowing true reconstruction of the vertebral body. It can avoid a certain number of supplementary anterior approaches.

Purpose: We report our experience with 10 cases of osteotomy performed at the same time as total hip arthroplasty.

Material and methods: This series included ten knees in ten patients, mean age 63 years. The knees were divided into two groups: five with tibial deviations in one or two planes (three callus deformities, one congenital varus, and one coxalgia sequela) and six mono- or biplanar femoral deviations (three callus deformities, two congenital valgus, and one coxalgia sequela). The principal extra-articular deviation was greater than 10° in all cases. A total knee arthroplasty preserved the posterior cruciate ligament in nine cases with a non-cemented femoral stem implant and cemented tibial implant with or with out a stem. The tibial osteotomies were all in the proximal metaphysis. The femoral osteotomy was in the distal metaphysis in four cases, subtro-chanteric with derotation in one and in the diaphysis with derotation in one. The correction osteotomy was performed before the prosthesis bone cut. Complementary osteosynthesis was used in all cases.

Results: Minimal follow-up was 12 months with a mean of 19.6 months. Bone healing was achieved in all cases. Complete weight bearing was achieved at a mean 2.5 months. The mean postoperative HSS score was 76. Results were excellent in four knees, good in four, fair in one and poor in one. Mean joint amplitude was 105°. Radiographically, complete tibial correction was obtained for three knees (two biplanar corrections); for two knees the biplanar correction was incomplete in one plane. Among the four biplanar femoral deviations, complete correction was achieved in three and incomplete frontal correction in one, with two varus overcorrections in the frontal plane on the single plane deviations.

Discussion: We used this method to avoid intra-articular correction of extra-articular deviations, a source of complications.

Conclusion: Combining osteotomy with total knee arthroplasty during the same operation for patients with major axial deviation and degenerative joint disease provides satisfactory results in 80% of the cases.