Introduction

High valgus tibial osteotomy has been proven effective in the treatment of medial femoro-tibial osteoarthritis [6, 9, 12]. In most studies, patellar height increased after closing-wedge osteotomy and decreased after opening-wedge osteotomy [3, 5, 8, 13–15, 22]. These changes in patellar height may adversely affect functional outcomes and create technical difficulties during subsequent knee arthroplasty. However, patellar height is not usually measured before performing an osteotomy. No studies have assessed whether taking patellar height into account improves treatment decisions and patient outcomes.

The primary objective of this prospective comparative observational study in a large patient population was to compare patellar height after medial opening-wedge and lateral closing-wedge osteotomy. Our hypothesis was that patellar height decreased after OWO and increased after CWO. The secondary objective was to determine whether frontal axis correction and tibial slope modification correlated with changes in patellar height. Such correlations would allow selection of the osteotomy type depending on preoperative patellar height and required axis correction.

Materials and methods

Patients

This prospective comparative observational study sponsored by the French Society for Orthopaedic Surgeons (SOFCOT) was conducted in ten research centres* in France between January 2008 and March 2009. Opening- or closing-wedge HVTO was performed according to standard practice in each study centre: two centres performed only CWO and four only OWO, whereas the remaining four centres used both techniques with a preference for OWO in patients having more than 8° of varus malalignment. In none of the study centres was patellar height used as a criterion to choose between closing- and opening-wedge osteotomy.

Consecutive patients requiring isolated high valgus tibial osteotomy at the supra-tuberosity level were eligible for the study. A follow-up of at least 6 months was required, to permit an evaluation of anatomical results, albeit not of clinical outcomes. Exclusion criteria were complex malunion of the proximal tibia requiring an intra-articular procedure, revision osteotomy, and need for combination with a ligamentous procedure or femoral osteotomy.

Data collection

Each centre recorded the data for each patient prospectively, in an anonymised electronic file. Then, the data were entered into a single database and checked.

We collected age, gender, and body mass index (BMI); type of osteotomy, cause of osteoarthritis (primary, post-traumatic, post-ligamentoplasty, other); and pre- and postoperative mobility and the visual-analogue-scale pain score. Radiological data used for the study were medial osteoarthritis classified using Ahlbäck criteria [1], the presence of osteoarthritis in the other knee compartments, pre- and postoperative patellar heights, pre- and postoperative tibial slopes measured according to Brazier et al. [2] (Fig. 2) as detailed by Ducat et al. [7], pre- and postoperative hip-knee-ankle (HKA) angle, and degree of correction. Pre- and postoperative long-leg radiographs were used to determine the mechanical weight-bearing axis and HKA angle and pre- and postoperative knee radiographs to measure tibial slope. The HKA and tibial slope were measured to the nearest degree. A lateral knee radiograph was used to measure patellar height based on the Caton-Deschamps index [4] obtained by dividing the distance between the inferior patellar edge and antero-superior tibial corner by the length of the patellar cartilage (Fig. 1). The percentage difference between pre- and postoperative patellar heights was computed; the value was negative when patellar height decreased and positive when it increased. The sub-population of patients (n = 41) with pre-operative patella baja (defined as a Caton-Deschamps index between 0.6 and 0.8) or patella infera (defined as a Caton-Deschamps index lower than 0.6) was evaluated separately.

Fig. 1
figure 1

Brazier method for tibial slope calculation

Full size image
Fig. 2
figure 2

Caton index for the patellar height calculation

Full size image

Statistical analysis

Data were analysed using Statview® software (v 5.0, SAS Institute, Cary, NC, USA). Comparisons relied on the Student-Fisher t test for qualitative variables and on Pearson’s coefficient of correlation and Yates’ Chi-square test for quantitative variables. Values of p lower than 0.05 were considered significant.

Results

Table 1 reports the main characteristics of the 321 patients, 205 males and 116 females, with a mean age of 52 ± 9 years and a mean BMI of 28 ± 5 kg/m². OWO was performed in 224 patients and CWO in 97.

Table 1 Description of the population
Full size table

There was a significant difference between the two populations (CWO and OWO) in regard to the initial axis malalignment (HKA, 173° ± 3.2° in the opening-wedge group and 175° ± 3° in the closing-wedge group, p < 0.000001). Patellar height was not significantly modified by CWO (Caton-Deschamps index, 1.07 ± 0.2 pre-operatively and 1.06 ± 0.19 postoperatively, n.s). In contrast, OWO was followed by a significant 9 ± 22 % decrease in patellar height (0.98 ± 0.19 pre-operatively and 0.88 ± 0.21 postoperatively, p < 0.0001). After opening-wedge HVTO, 28 % (n = 49) of patients had a greater than 20 % decrease in patellar height (Tables 2, 3).

Table 2 Patellar height distribution
Full size table
Table 3 Patellar height variations depending on population
Full size table

The patellar height decrease correlated with the magnitude of axis correction (p = 0.015), the more extensive the opening was, the less was the patellar height. Modification in patellar height was not correlated with modification in tibial slope in the CWO group. In contrast, after OWO, the increase in posterior tibial slope correlated significantly with the decrease in patellar height (p < 0.005).

Of 38 patients with patella baja, 28 underwent opening- and 10 closing-wedge osteotomy. All 3 patients with patella infera underwent OWO. Patellar height in these 41 patients was unchanged after high valgus tibial osteotomy (Caton-Deschamps index, 0.75 ± 0.07 pre-operatively and 0.77 ± 0.07 postoperatively). Moreover, the presence of patella baja or infera did not significantly affect the overall postoperative axis correction (183.8° ± 3.6 vs. 183.9° ± 3.4 in the patients without patella baja or infera).

Discussion

The most important finding of the present study was the mean 9 % decrease in patellar height after OWO, contrasting with the absence of significant change after CWO. Our working hypothesis was that patellar height decreased after OWO and increased after CWO.

The patellar height decrease after OWO of 9 ± 22 % was consistent with earlier reports of decreases ranging from 9 to 16 % [3, 5, 8, 13–15, 21]. After CWO, patellar height remained unchanged in our study but decreased by 5 % in a study by Tigani et al. [19]. However, in keeping with our results, Song et al. [17] found that patellar height decreased after OWO and remained unchanged after CWO. A patellar height decrease has important clinical implications, as lower patellar height can generate technical difficulties and impair functional outcomes after subsequent knee arthroplasty [16, 18, 20, 21].

We used the Caton-Deschamps index to assess patellar height. A decrease in the Caton-Deschamps index can be due either to lowering of the patella or to a decreased distance between the inferior patellar edge and antero-superior tibial corner related to a change in tibial slope. The Insall-Salvati index [11] is not affected by tibial slope modifications. However, we felt the Caton-Deschamps index would be more reproducible in this multicentre study, as the patellar tendon may be difficult to assess, most notably on postoperative radiographs [8].

The patellar height decrease correlated significantly with the magnitude of axis correction in our study. In contrast, El Amrani et al. [8] reported that opening-wedge size was not correlated with patellar lowering as assessed using the Caton-Deschamps, Insall-Salvati or Blackburne-Peel index. This difference with our study may be ascribable to the smaller sample size of only 40 patients, compared to 321 patients in our study.

The patellar height decrease after OWO correlated significantly with the increase in tibial slope. As the importance of the axis correction is not modifiable, care should be taken to minimise changes in tibial slope. Hernigou and Ma [10] advocated positioning the bone graft wedge posteriorly to avoid tibial slope changes due to OWO. We did not assess whether graft position correlated with changes in tibial slope or patellar height. In the study by Song et al. [17] showing patellar height changes similar to those in our study, the 3-year follow-up allowed an evaluation of clinical outcomes, and no difference was found between closing- and opening-wedge HVTO in terms of anterior pain or knee flexion. The 6-month follow-up in our study was too short for an assessment of clinical outcomes.

Limitations of our study include the non-randomised design with statistically significant baseline differences between the two groups for HKA, age, and BMI. In addition, OWO was used far more often than CWO (224 vs. 97 patients). The surgical procedures were not standardised; instead, each centre followed routine local procedures regarding the material and techniques. This point explains the predominance of OWO, which is consistent with current trends in France, as well as the differences in baseline HKA and axis correction, with some centres preferring OWO only when the varus malalignment angle exceeded 8°. However, our study is the largest to date comparing OWO and CWO. Other strengths include the prospective design and inclusion of consecutive patients at centres having considerable experience with HVTO. Usually, we do not measure patellar high before performing an osteotomy: only HKA angle is considered. Now, we use to measure patellar height before osteotomy, and in case of major axis correction with an OWO, we take care to open posteriorly to avoid a decrease of patellar height.

Conclusion

This prospective multicentre study in a large sample size established that patellar height decreased after OWO, by 9 ± 22 %, and that the decrease correlated with the magnitude of axis correction and tibial slope change. Patellar height remained unchanged after CWO. These results suggest that patellar height should be measured routinely before high valgus tibial osteotomy and, when major axis correction is required, that the opening wedge should be positioned posteriorly to limit the change in tibial slope and therefore the decrease in patellar height.