Thank you for your time and attention. Our disclosures are available on the MyAcademy website. Patellofemoral anatomy is such that allows for substantial freedom of motion. In part due to this, patellar dislocation accounts for 2 to 3% of knee presentations in clinic. Now, the risk factors for instability have been previously well established, and these include young age, trochlear dysplasia, and high demand activities. However, to this date, accurate prognostication remains elusive, especially after first time dislocation. This is likely in part due to the fact that there are multifactorial risk factors contributing to this, and also that there's a long history of natural recurrence taking multiple years to recur. Therefore, the purpose of our study was to describe the clinical history of a series of primary lateral patellar dislocations, and to determine long-term predictors of recurrent instability. Our goal in doing so was to create a prognostic stratification tool for predicting recurrent instability. We use the Rochester Epidemiology Project, or REP for this, a point of particular pride at Mayo Clinic. It consists of the complete medical record for Olmstead County residents and also neighboring areas, including parts of western Wisconsin, and having a catchment area of over half a million patients. It allows access to clinical and surgical notes, as well as imaging, and allows patients to be longitudinally followed, regardless of the treating institution. Therefore, it's ideal for studies such as this. The REP database was reviewed from 1990 to 2010, and inclusion criteria consisted of first-time lateral patellar dislocations. Exclusion criteria consisted of chronic patellar subluxations, no MRI available at the time of injury, and less than four years of follow-up. Records were then manually verified to ensure that the patients actually did have a lateral patellar dislocation, and rates of recurrence, as well as surgery, were determined. In terms of creating our recurrent instability of the patellar RIP score, we used a competing risks analysis methodology. Points were assigned using methods analogous to the now-famous Framingham Heart Study, and in short, fine and gray proportional hazards were employed, and sub-distribution hazards of the individual risk factors were used as a basis for point standardization. This allowed for the creation of a multivariable predictive model, as compared to previous univariate models. Model checking was then performed with cumulative incidence and ROC curves. Our study population consisted of 81 patients, followed for a mean of 10 years, 43 females, 38 males, with a mean age just under 20 years. Recurrent instability was present in 38 patients, or less than half, but almost half of the patients, and mean time to recurrence was three years, again highlighting that we need long-term follow-up for these patients. Surgical management was employed in 30 patients, and seven patients underwent surgical management before recurrent dislocation, mostly driven by intra-articular loose bodies. When looking at univariate recurrent instability predictors, young age, low body mass index, skeletal immaturity, any type of de jure dysplasia, patella alta as measured by Catan-Deschamps indices greater than or equal to 1.3, and tibial tubercle to trochlear groove distances standardized over patellar length, or TTTG over PLs, which are increased, were all univariate predictors of recurrence. Now if we put them all together and find which factors are additive compared to which ones are collinear, we built the RIP score. Age of less than 25 years was given two points, skeletal immaturity, one point, de jure dysplasia of any type, one point, and TTTG over PL greater than or equal to .5, one point, for a total of zero to five points. If we look at the RIP score performance, patients with a RIP score of zero to one points were deemed to be low risk. They had a 0% cumulative incidence of recurrence at 10 years. Patients with a RIP score of two to three points were intermediate risk. They had a 31% cumulative incidence of recurrent instability at 10 years. And finally, patients with a RIP score of four to five points were high risk, with a cumulative incidence of 79%. If we looked at the RIP score prognostication ability, the receiver operating curve characteristics were quite satisfactory. Area under the curve was .875, and the p-value was less than .001. In discussion, historically, 30 to 50% of patients will experience recurrent dislocation, and clinically, this is substantial, given that recurrence is associated with increasing rates of osteochondral injury with each recurrence. Now while individual risk factors have been described in the literature, there currently exists no good multivariable system for predicting long-term recurrent instability. Previous systems have been described, but they have limited discriminatory value, they're often based on short-term follow-up, and furthermore, they publish odds ratios, which are difficult to combine in a multivariable model, plus they're hard to use clinically. In contrast, we believe the RIP score can be used to predict recurrence of primary lateral patellar dislocations at over 10 years of follow-up. Furthermore, the RIP score is now available on MDCalc, it has at-the-fingertips clinical access online, there's expert input and commentary by Dr. Farr, who was kind enough to help us with this, and it's also free. In conclusion, patients can be readily classified into low, intermediate, and high-risk categories using the RIP score based on age, skeletal maturity, trochlear dysplasia, and TTTG over PL ratio. Finally, long-term risk gratification holds significant clinical utility for determination of patients who are at high risk for recurrent instability following primary patellar dislocation. I thank you all for your time and attention.

patellar dislocation

recurrent instability

prognostic tool

Rochester Epidemiology Project

Rochester Instability Prognostication Score