We studied the complications after open-wedge osteotomy by hemicallotasis in 308 consecutive patients, most of whom had osteoarthritis of the knee. The participating surgeons, who worked at 17 hospitals, used their discretion in selecting patients, operating techniques and external fixators. The general complications included 11 cases of deep-vein thrombosis (4%), six of nonunion (2%) and one of septic arthritis of the knee. There were technical complications in 13 patients (4%). In 157 patients (51%) pin-site infections were recorded; of these, 96% were minor and responded to wound toilet and antibiotic treatment. A total of 18 revision procedures was carried out.

We describe the results of 50 operations carried out on 46 patients with medial osteoarthritis of the knee of Ahlbäck grade 1 to 3. Patients were randomised either to a closed-wedge high tibial osteotomy (HTO) or an open-wedge procedure based on the hemicallotasis technique (HCO). Their median age was 55 years (38 to 68). The preoperative median hip-knee-ankle (HKA) angle was 171° (164 to 176) in the HTO group and 173° (165 to 179) in the HCO group. After six weeks, the median HKA angle was 185° (176 to 194) in the HTO group and 184° (181 to 188) in the HCO group. In the HTO group, seven patients were within the range of 182° to 186° compared with 21 in the HCO group (p < 0.001). One year later, ten HTO patients were within this range while the HKA angulation in the HCO group was unchanged. At two years the numbers were 11 and 18, respectively.

We evaluated the clinical results on the Hospital for Special Surgery, Lysholm and Wallgren-Tegner activity scores, and patients completed part of the Nottingham Health Profile questionnaire. An impartial observer at the two-year follow-up concluded that all scores had improved, but found no clinical differences between the groups.

To assess migration of the tibial component we used roentgen stereophotogrammetric analysis in 40 patients who had had a total knee arthroplasty after failure of a closing wedge osteotomy and compared them with 40 matched patients after primary total knee arthroplasty.

We found no difference in migration over time or in the tendency for continuous migration between the two groups. There were no differences in alignment or position of the knee prosthesis or in the clinical outcome.

Our findings show that revision of a failed high tibial osteotomy to a total knee arthroplasty is effective.

We measured the insertion and extraction torque forces in a randomised study of 76 external fixation screws in 19 patients treated by hemicallotasis for osteoarthritis of the medial side of the knee. The patients were randomised to have either standard tapered screws (Orthofix 6/5 mm) or the same screws with hydroxyapatite (HA) coating. One patient had two standard and two HA-coated screws. All patients had an anterior external fixator (Orthofix T-garche), with two screws in the proximal tibial metaphysis parallel to and about 2 cm below the joint surface and two in the tibial diaphysis.

The mean torque forces for insertion of the standard screws were 260 Ncm for the proximal to medial screw, 208 for the proximal to lateral screw and 498 and 546 Ncm for the diaphyseal pins. The corresponding forces for the HA-coated pins were not significantly different. The torque forces for the extraction of the standard pins were 2 Ncm for the proximal pins, 277 and 249 Ncm for the distal pins and 482, 478, 585 and 620 Ncm, respectively (p < 0.005) for the HA-coated pins.

All 18 of the metaphyseal standard screws were loose at extraction (extraction force < 20 Ncm), but only one of the HA screws in the metaphysis was loose. In the diaphysis the standard screws lost about 40% of their fixation in contrast to the HA-coated screws which retained full fixation strength.