In this meta-analysis, we summarized findings in the clinical literature on the outcomes of operative versus conservative treatment for patellar dislocation. On the basis of the available evidence, this study demonstrates a higher rate of recurrent patellar dislocation events post-treatment, higher Tegner score and higher Hughston VAS score in patients managed non-operatively compared to patients managed operatively.
All RCTs [2, 7–9, 28–30] suggested that operative treatment of patellar dislocation results in a lower risk of recurrent patellar dislocation compared to non-operative treatment. There were six RCTs in which patients were managed operatively compared to non-operatively following the first patellar dislocation among the seven included RCTs. Therefore, the finding should be interpreted with great caution with reference to recurrent patellar dislocation. In 2011, Smith et al. showed equivalent results in their meta-analysis . They considered that the result should be interpreted with great caution due to statistically significant funnel plot asymmetry. In a study by Palmu et al. , although the rate of recurrent patellar dislocation was higher in patients following non-operative treatment, there was no significant difference in functional assessment. This suggested that these patients were able to perform all their activities of daily living, irrespective of recurrent patellar instability and dislocation events. Therefore, assessment of functional-based outcomes should be paid more attention in future studies.
This study reported that there was a difference in Hughston VAS score when comparing pain between operative and non-operative management strategies. The degree of pain was lower in the non-operative group. However, the finding should be interpreted with great caution because the data involved in this result was extracted from only two RCTs and the Kujala score [2, 7–9, 28–30] correlated better with the subjective result and the recurrence of patellar dislocation than the Hughston VAS score . There was no significant difference in Kujala score between the two treatment groups (MD = 6.38, 95% CI: -5.32 to 18.08, P = 0.29), with a high degree of heterogeneity across the studies (P < 0.00001, I
2 = 98%) in this analysis. The reasons for this result may include the following factors. First, the Kujala score is a subjective evaluation method so that there may have been differences between the included RCTs. Second, the inclusion criteria were different for each study included. Third, there was some variability in treatment methods, especially in operative treatment. The seven RCTs reported using a number of different operative interventions including lateral release, medial retinaculum or MPFL repair, or Roux–Goldthwaite procedures.
Only one study was identified which solely assessed the incidence of patellofemoral osteoarthritis between operative and non-operative groups.  This reported that there was no statistically significant difference between patients treated operatively and those treated non-operatively in respect to articular cartilage lesions within the patellofemoral joint. Repeated chondral injury may predispose patients to osteoarthritis [31–33]. Due to the relative scarcity of RCTs assessing the incidence of patellofemoral osteoarthritis in patients managed following patellar dislocation, it will be necessary to observe whether there is a difference in clinical outcomes between patients managed non-operatively and operat'ively.
The results of this review should be interpreted and generalized with caution due to the limited number of the studies and the high risk of bias inherent in the studies. First, it included only a limited number of studies and of subjects. After a careful search, only 7 RCTs were included in the final analysis, giving a total population of 402 subjects. The publication bias was not tested in our analysis in consideration of the low power due to the small number of studies included.
Second, of the seven RCTs, only two used an appropriate concealed allocation method for randomization, and none reported adequate intention-to treat analysis. This might have introduced selection bias. In addition, blinding of outcome assessors was not used in any of the RCTs included, and thus detection bias might have been introduced.
Third, most of the studies did not screen participants for stress and anxiety levels, which might have weakened the evidence of the study. Four, while operative and non-operative interventions were compared in those RCTs included, the majority of studies poorly described the specific management procedures in detail, therefore limiting the ability to replicate these clinical trials. In particular, the non-operative management strategies were poorly described in all RCTs.
Finally, as regards the functional outcome of patients following patellar dislocation, there was no difference in Kujala score between operative and non-operative management strategies with significant heterogeneity. This subjective result was not effectively evaluated.
In future, in order to better evaluate the outcomes of the two treatment strategies, it is suggested to define the population, standardize the interventions prescribed to those patients, and evaluate this area of therapy through a well-designed randomised controlled trial.