A prospective, randomised, controlled trial compared two different techniques of high tibial osteotomy with a lateral closing wedge or a medial opening wedge, stabilised by a Puddu plate. The clinical outcome and radiological results were examined at one year.

The primary outcome measure was the achievement of an overcorrection of valgus of 4°. Secondary outcome measures were the severity of pain (visual analogue scale), knee function (Hospital for Special Surgery score), and walking distance.

Between January 2001 and April 2004, 92 patients were randomised to one or other of the techniques. At follow-up at one year the post-operative hip-knee-ankle angle was 3.4° (± 3.6° sd) valgus after a closing wedge and 1.3° (± 4.7° sd) of valgus after an opening wedge. The adjusted mean difference of 2.1° was significant (p = 0.02). The deviation from 4° of valgus alignment was 2.7° (± 2.4° sd) in the closing wedge and 4.0° (± 3.6° sd) in the opening-wedge groups. The adjusted mean difference of 1.67° was also significant (p = 0.01).

The severity of pain, knee score and walking ability improved in both groups, but the difference was not significant.

Because of pain, the staples required removal in 11 (23%) patients in the closing-wedge group and a Puddu plate was removed in 27 (60%) patients in the opening-wedge group. This difference was significant (p < 0.001).

We conclude that closing-wedge osteotomy achieves a more accurate correction with less morbidity, although both techniques had improved the function of the knee at one year after the procedure.

Deformity increases the risk of progression of osteoarthritis of the knee.1 Patients with osteoarthritis of the medial compartment of the knee which is refractory to conservative treatment can be treated by a valgus high tibial osteotomy, and various techniques are available. These include a closing-wedge osteotomy, an acute opening-wedge osteotomy, a dome osteotomy and a hemicallotasis progressive opening-wedge osteotomy controlled by an external fixator.26 Each option has individual advantages and complications.79 The most important message which has emerged from retrospective studies is that the appropriate selection of patients and the achievement and maintenance of an adequate operative correction are required for a successful outcome.2,4,6,10 Loss of correction correlates with the type of fixation, the degree of correction and the time to bony union.11,12

A randomised, controlled trial comparing the hemicallotasis opening-wedge with a closing-wedge technique showed that there was a significantly greater loss of correction after closing-wedge osteotomy at follow-up at one year.3

We therefore undertook a prospective, randomised study to compare lateral closing-wedge and medial opening-wedge techniques, in regard to achievement and maintenance of adequate operative correction.

Patients and Methods

This randomised controlled trial was carried out on 92 patients between January 2001 and April 2004, after approval of the Ethics Committee.

The criteria for inclusion in the study were radiological evidence of medial compartment osteoarthritis (OA) with medial joint pain and varus malalignment. The grade of OA was scored according to Ahlbäck13 and measured on standing short posteroanterior (PA) radiographs.

Malalignment was assessed using the hip-knee-ankle angle and the mechanical axis was obtained on a full-length standing radiograph from the mechanical axis of the femur (from the centre of the femoral head, using Mose circles to the central point between the tibial spines), and the mechanical axis of the tibia (from the centre of the tibial spines to the centre of the ankle). We used lateral fluoroscopic control by superimposing the dorsal aspect of the femoral condyles to ensure a perfect anteroposterior full-length exposure.14

The criteria for exclusion were symptomatic OA of the lateral compartment, rheumatoid arthritis, range of movement of < 100°, grade-315 collateral ligament laxity, history of fracture or previous open operation of the lower limb and a flexion contracture of > 10°. Patients with a contralateral high tibial osteotomy were excluded if the first knee had been included in this trial; thus, if both knees were symptomatic, only the first knee was included.

After obtaining informed consent and baseline measurements, the patients were randomised by a computer-generated procedure in blocks of 16. Sealed envelopes contained the group assignment. An independent assistant not directly involved in the study opened these sealed envelopes after enrolment of the patients by one of the four orthopaedic surgeons (RWB, JANV and two who are not authors), who performed the operations. All surgeons were experienced with both techniques.

Treatment groups.

Randomisation was to one of the following procedures: a) closing-wedge high tibial osteotomy and a plaster cylinder cast for six weeks post-operatively (Fig. 1) or b) opening-wedge osteotomy fixed using a Puddu plate (Arthrex; Naples, Florida) (Fig. 2). For the closing-wedge technique we used the Allopro (Zimmer; Winterthur, Switzerland) calibrated osteotomy guide to obtain accurate resection of bone. The common peroneal nerve was exposed and retracted. Subsequently, the anterior part of the proximal fibular head representing the anterior part of the proximal tibiofibular syndesmosis, was resected. The osteotomy was fixed with two staples. At the end of the procedure a fasciotomy of the anterior compartment was performed to prevent compartment syndrome.

Fig. 1

Radiograph of a closing-wedge high tibial osteotomy stabilised by two staples.

Fig. 2

Radiograph of an opening-wedge high tibial osteotomy secured by a Puddu plate and ipsilateral iliac-crest bone graft.

The extent of the opening-wedge depended on the length of the osteotomy and the diameter of the proximal tibia and was calculated pre-operatively. The Arthrex instruction manual provides a goniometric formula table, which gives the extent of the opening-wedge for a specific correction. Additionally, the degree of correction performed during the procedure was controlled by fluoroscopic assistance in which the centre of the hip and the centre of the ankle were marked. If the opening-wedge was > 7.5 mm, the open gap was filled with bone harvested from the ipsilateral iliac crest. There was a second randomisation (in blocks of eight) for treatment after the osteotomy, namely, with or without a plaster cast to determine whether a cast influenced the outcome.

In both techniques the goal was to achieve a correction of 4° physiological valgus.

The patients were mobilised on the first post-operative day and partial weight-bearing was allowed in all groups. After six weeks any plaster cast used was removed.

Initial evaluation.

Age, gender, body mass index (BMI), the severity of medial and lateral OA (Ahlbäck score, 0 to 3), the aetiology of the OA, primary versus secondary (post meniscectomy or previous cruciate ligament injury), varus malalignment (hip-knee-ankle angle), severity of pain measured by a visual analogue scale (VAS),16 knee function using the Hospital for Special Surgery score (HSS score),15 and walking distance (km) were scored pre-operatively.

Outcome assessments at one year.

The primary outcome measure was achievement of an overcorrection of the hip-knee-ankle angle by 4° of valgus. The differences between the achieved valgus correction and the objective of 4° of overcorrection was determined. In addition a dichotomous outcome was an achievement of a valgus alignment between 0° and 6°.

RWB, together with the surgeon who performed the operations, made the pre-operative radiological measurements; post-operative measurements were made by RWB and TMR.

Secondary outcome measures were severity of pain (VAS), walking distance and the knee function score (HSS). The HSS score is divided into six categories (pain, function, range of movement, muscle strength, flexion deformity, and instability) and consists of a questionnaire and a physical examination. The examination was undertaken by one non-blinded assessor (RWB). Complications, further surgery including removal of hardware and morbidity from the donor site in the iliac crest were noted.

Sample size.

The sample size was calculated based on an expected increase in the success rate from 60% for the closing-wedge to 85% for the opening-wedge osteotomy. A successful operative result was defined as achievement of approximately 4° of valgus alignment. To detect such a difference with one-sided testing (α = 0.05 and a power of 80%) 46 patients were required in each group.

Statistical analysis.

A multivariate linear regression method was used to analyse the impact of closing- versus opening-wedge osteotomy on post-operative alignment, the VAS and HSS knee scores, walking distance and complications at follow-up at one year. A multivariate logistic regression method was used for the dichotomous outcome measures.

All data were analysed according to an intention-to-treat principle, implying that all patients who were randomised were included in the analyses, and that they were analysed according to the group to which they had been allocated.

Gender, age, BMI and baseline values for the hip-knee-ankle angle, pain, HSS score, walking distance, medial OA of more than loss of joint space alone, concurrent OA of the lateral compartment, contralateral arthritis of the knee and ipsilateral arthritis of the hip were considered as possible confounders and were included in the regression models only if they changed the relationship between the dependent variable and the type of osteotomy by at least 10%. The same was done for the relationship between these described dependent variables and the type of post-operative treatment (plaster versus no plaster) in the group with the opening-wedge technique. For patients who were lost to follow-up or had further surgery during follow-up, the last available measurement was used.

The SPSS program (SPSS Inc., Chicago, Illinois) was used for the statistical analysis and a p-value of 0.05 was considered to be statistically significant.


One patient (closing-wedge group) was lost to follow-up. Another (opening-wedge group) had a one year post-operative pain and HSS score, a standing standard short PA radiograph, but no full-length radiograph because of emergency treatment of an unrelated condition, precluding his full radiological evaluation. Both were included in the analysis.

The mean age of the patients was 50.2 years (21 to 67). The mean pre-operative hip-knee-ankle angle was 6.3° (0° to 14°) of varus and differed significantly (p < 0.05) between the two groups. In the closing-wedge group the mean was 6.8° (2° to 14°) of varus and in the opening-wedge group it was 5.7° (0° to 12°) of varus (Table I).

View this table:
Table I.

Baseline characteristics for the total series and separately for both groups

A total of 47 patients had a closing-wedge osteotomy and 45 an opening-wedge osteotomy. In the opening-wedge group 22 patients had a plaster cast after operation and 23 did not.

Primary outcome measures (Table II).

View this table:
Table II.

Continuous outcomes (mean, sd) for closing-wedge osteotomy versus opening-wedge osteotomy at follow-up at one year

The power calculation was based on one-sided testing, because we expected a higher success rate in the opening-wedge group. However, because the raw data showed better results for the closing-wedge group, it was decided to test two-sided with a signifiance level of 0.05.

At follow-up of one year the mean post-operative hip-knee-ankle angle was 3.4° (± 3.6° sd) of valgus in the closing-wedge group and 1.3° (± 4.7° sd) of valgus in the opening-wedge high tibial osteotomy group. The adjusted mean difference of 2.12° (95% confidence interval (CI) 0.38 to 3.86) was significant (p = 0.019).

The mean deviation from valgus alignment of 4° was 2.7° (± 2.4° sd) in the closing-wedge group and 4° (± 3.6° sd) in the opening-wedge group. The adjusted mean difference of 1.67° (95% CI 0.42 to 2.92) was also significant (p = 0.011).

The dichotomous outcome measure (achievement of a valgus alignment within 0° to 6°) was achieved in 37 (79%) in the closing-wedge group and in 25 patients (56%) in the opening-wedge, resulting in an odds ratio for successful overcorrection of 3.44 (95% CI 1.29 to 9.16) in the closing-wedge compared with the opening-wedge group. This difference was also significant (p = 0.01).

Secondary outcome measures (Table II).

The mean pain (VAS) score decreased in both groups, 2.3 points after closing-wedge and 2.7 points after opening-wedge osteotomy which was not significant (p = 0.93).

The mean HSS knee score and the mean walking distance increased in both groups, 8.5 points and 1.7 km after closing-wedge and 9.4 points and 2.2 km after opening-wedge osteotomy. These differences were also not significant (p = 0.78, p = 0.65, respectively).

Plaster versus non plaster in the opening-wedge group.

The baseline characteristics in the subgroup were similar for the plaster and the non-plaster groups, without any significant differences. There were no significant differences in primary or secondary outcome measures between the plaster and non-plaster subgroups (Table III).

View this table:
Table III.

Continuous outcomes (mean, sd) for plaster versus no plaster in the opening-wedge osteotomy after follow-up at one year

Complications during follow-up.

One patient in the closing-wedge group required a corrective varus osteotomy because of overcorrection of the initial osteotomy and another underwent total knee replacement because of progression of symptoms. Three patients in the opening-wedge group had a further valgus osteotomy because of recurrent varus alignment.

Because of pain, the staples or plate was removed in 11 patients (23%) in the closing-wedge and in 27 patients (60%) in the opening-wedge group. This difference was significant (p < 0.001; odds ratio 0.15; 95% CI 0.06 to 0.41).

In the opening-wedge group 33 patients required bone grafting. Two patients in this group developed nonunion, one with an opening-wedge of 7.5 mm without bone graft and the other with an opening-wedge of 12.5 mm who had bone graft. Persistent pain at the iliac crest was reported in eight patients. One of these had further surgery because of a symptomatic exostosis at the donor site. Another sustained an injury to the lateral femoral cutaneous nerve. The remaining complications are shown in Table IV.

View this table:
Table IV.

Complications after closing-wedge and opening-wedge osteotomy


Our results showed that the closing-wedge osteotomy achieved significantly more accurate correction with less deviation from our objective of 4° valgus overcorrection at follow-up at one year.

We expected a better outcome in the opening-wedge group, because of the precise positioning possible with the Puddu plate and because it is performed under fluoroscopic control. Additionally, it is easier to create an opening wedge than to remove a wedge of bone from the proximal tibia. In our experience, with the closing-wedge technique it can be difficult to remove the wedge totally, especially at its apex at the medial side. An important reason for inadequate correction noted at one year is that the Puddu plate is not strong enough to maintain the per-operative correction.17 A new design, in which the screw-head locks into a more rigid plate, may provide more stability and give better results. Lobenhoffer and Agneskirchner5 used a medial plate in 92 opening-wedge osteotomies without loss of correction.

There was no statistical difference in the improvements in both groups. However, a follow-up of one year may be too short to demonstrate whether the benefit of the slight overcorrection approximating to 4° greater than physiological valgus remains true as previously suggested.3 To determine whether one type of osteotomy delays the requirement for total knee replacement more than the other requires a much longer follow-up.

Application of a plaster cast after an opening-wedge osteotomy did not appear to prevent loss of correction, with no difference in the mean hip-knee-ankle angles between the plaster and non-plaster groups at one year. However, the sample size in our study was calculated to detect a difference between the closing-wedge and opening-wedge techniques. This study lacks sufficient power to make a confident statement about the advantages or otherwise of using a plaster cast in conjunction with an opening-wedge osteotomy.

There were some limitations to our study. First, we were unable to obtain the standing full-length radiographs on the first day after operation because the patients could not stand on their operated leg.

Secondly, four orthopaedic surgeons were involved. Although all had experience of both techniques, one single surgeon would have been preferable to reduce possible operator-dependent variability but this was not clinically feasible. Finally, the outcome assessor was not blinded for physical examination of the HSS score and post-operative hip-knee-ankle angle.

Based on our study we conclude that closing-wedge high tibial osteotomy achieves a more accurate correction and that both techniques reduce pain and improve function at follow-up of one year. The opening-wedge osteotomy with a Puddu plate is associated with a higher number of complications.


  • We thank R. M. D. Bernsen for the statistical analysis.

  • No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

  • Received February 7, 2006.
  • Accepted June 1, 2006.


View Abstract