Reevaluating Patellar Height Indices in Pediatric Patellofemoral Instability: Comparative Accuracy of Caton Deschamps
Giulia Beltrame1,2, Marco Crippa1,3, Samuel A Beber1
1Hospital for Special Surgery, New York, New York, USA.
Background:
Patella alta is a key anatomic risk factor for patellofemoral instability (PFI), especially among young athletes performing pivoting sports, as it reduces patellofemoral engagement during early knee flexion. The Caton-Deschamps index (CDI) is a commonly used radiographic measure of patella alta, but it does not provide information on patellofemoral engagement, which is relevant in PFI. Magnetic resonance imaging (MRI)-based indices, such as the patellotrochlear index (PTI) and sagittal patellofemoral engagement (SPE), have been proposed as more appropriate assessments of functional patella alta.
Purpose:
To (1) compare the performance of PTI and SPE in evaluating functional patella alta in pediatric PFI as determined by CDI; and (2) explore the relationship among indices.
Study Design:
Case-control study; Level of evidence, 4.
Methods:
We analyzed 252 participants: 126 patients with PFI and 126 age- and sex-matched controls. Radiographic and MRI-based measurements of CDI, PTI, and SPE were compared between groups. Correlations between indices and receiver operating characteristic (ROC) analyses were conducted, using the CDI as a reference standard.
Results:
The CDI was significantly higher in patients with PFI than controls (1.317 ± 0.215 vs 1.094 ± 0.171; P < .001), whereas PTI and SPE did not differ significantly. SPE values were lower in patients with patella alta (0.403 ± 0.156 vs 0.476 ± 0.169; P = .013). PTI and SPE were strongly correlated (r = 0.898; P < .001), both demonstrating weak inverse correlations with CDI. When stratifying by patella alta, 40.5% of alta cases had an SPE of >0.45, and 97.3% had a PTI of >0.125. Univariable logistic regression demonstrated that SPE predicted patella alta (odds ratio = 0.263 [95% CI, 0.024-2.847]; P = .031). ROC analysis demonstrated poor discriminative ability of PTI and SPE for patella alta (area under the ROC curve [AUC], 0.534-0.594) and PFI (AUC, 0.461-0.488), while CDI strongly predicted PFI with sensitivity and specificity exceeding 73%.
Conclusion:
PTI and SPE demonstrated poorer discriminative ability for patella alta than CDI. However, MRI-based indices frequently indicated preserved patellofemoral engagement despite CDI-defined patella alta, underscoring their complementary role.
