Almost all authors have included ‘patellofemoral arthritis’ in the list of contraindications to unicompartmental arthroplasty and it may strike the reader as strange that we have not yet mentioned it in the discussion of AMOA, because, intuitively, associated patellofemoral joint OA must influence the outcome of UKA.
In anteromedial OA, the patellofemoral compartment very commonly exhibits chondromalacia, fibrillation, and cartilage erosions that sometimes expose bone. These lesions are mainly on the medial longitudinal (or ‘odd’) and medial facets of the patella and the equivalent surfaces of the femoral trochlea (see Fig. 4.2(g)), but they are also seen astride the median ridge of the patella and in the groove of the trochlea. They are much less common on the lateral facets. Marginal osteophytes are often seen on the preoperative radiographs and even more commonly when the joint is open to inspection.
The presence of any of these lesions has frequently been taken to contraindicate unicompartmental replacement. However, there are some evidence-based arguments for believing that this is unnecessary.
We first reached this opinion in 1986 (Goodfellow & O’Connor 1986) based on a study of 125 bicompartmental Oxford arthroplasties performed for OA (n = 74) or rheumatoid arthritis (n = 51). In these procedures only the tibiofemoral articular surfaces were replaced; the patella and the trochlea were retained. The state of the patella’s articular surface was recorded intra-operatively. At postoperative review (mean follow-up 49 months), no correlation was found between the intraoperative state of the patellofemoral joint and the patients’ postoperative complaints of pain (Table 5.1).
Table 5.1 Patellofemoral state versus postoperative pain (a) at rest and (b) during activity. Each entry in a table gives the number of knees with PFJ state at surgery as defined by the column and the report of pain at last follow-up as defined by the row. Each table is also a graph on which is plotted the linear regression line (shown dashed) that fits the data with minimum squared error. Both lines are nearly horizontal showing that the outcome was independent of the preoperative state.
Despite the mixture of diagnoses in this study (and its questionable relevance to unicompartmental replacement), at that time it provided the only scientific evidence on which to base our practice. Accordingly, ever since that publication, we have continued to ignore the state of the patellofemoral joint, whether assessed clinically, radiographically or intraoperatively, when deciding between OUKA and TKA. Subsequently, Carr et al. (1995) found no correlation between the perioperative state of the PFJ and the patients’ postoperative complaints of pain at a mean 44 months after surgery in a series of 121 knees treated for anteromedial OA by OUKA. Despite the degeneration of the PFJ in these patients (and in those of surgeons to whom we have given similar recommendations), patellofemoral problems have rarely been the cause of failure after OUKA (Price & Svard, 2011). In all the published series of OUKA with 10-year or longer results that we are aware of, including nearly 10,000 patients, there was not one revised for PFJ problems. The 2004 report from the Swedish Knee Arthroplasty Register gave causes for the 50 revisions from a total of 699 OUKAs, only one of which was for PFJ problems (Lidgren et al. 2004).
In 28 knees, the state of the PFJ was assessed on radiographs taken 1–2 years after OUKA and was compared with films taken 10+ years later. No significant difference between them was found (Weale et al., 1999). (This study was based on anteroposterior and lateral radiographs of the patellofemoral joint because ‘skyline’ views were not available.)
Beard et al. (2007b) have reported on 100 consecutive OUKA cases for anteromedial OA. In all these cases, the location of preoperative pain (anterior, medial, lateral, generalised) was independently determined and the radiological status of the patellofemoral joint was defined using Altman systems (Altman et al., 1987). There was no relationship between the presence of the pre-operative anterior knee pain (AKP) and the state of the PFJ. Also pre-operative AKP settled in every case and did not compromise the outcome. We therefore do not consider AKP to be a contraindication.
In this study (Beard et al., 2007b), arthritis, however severe, seen on the medial side of the PFJ on the skyline view did not compromise the outcome, so we ignore medial PFJ OA. However, although the numbers were small, there was some evidence to suggest that severe lateral PFJ OA does compromise the outcome. Therefore in the rare cases (less than 1% of patients) when there is severe lateral PFJ OA with bone loss, grooving and subluxation, we would now recommend a TKA.
In a further study of 824 knees in 793 patients, the state of the PFJ peri-operatively was correlated with the clinical outcome (Beard et al., 2007a.). There was exposed bone in the trochlea in 15%, on the medial side of the patella in 9%, and on the lateral side in 4%. Exposed bone did not compromise the outcome. Therefore we do not consider exposed bone seen in the PFJ to be a contraindication.
Berend et al. (2011) correlated the pre-operative state of the PFJ assessed radiographically with the six year survival in 626 OUKA. 61% of the cases had a normal PFJ and a survival of 94%; 39% had an abnormal PFJ with a 98% survival; and 15% had significant PFJ damage and a 97% survival. There was no statistically significant difference in survival between the groups.