Differences between skeletally immature and mature patients in recurrent patellar dislocation: a radiological study Authors: Yi Hua, Hua Wang, Xu Yang Published: 2026 ABSTRACT AbstractObjectiveTo compare radiological characteristics between skeletally immature and mature patients with recurrent patellar dislocation (RPD).MethodsA retrospective study was conducted involving 85 patients (106 knees) with RPD treated between January 2016 and September 2024. Patients were divided into skeletally immature and skeletally mature groups based on epiphysis closure. Patellar height was assessed using the Caton-Deschamps Index (CDI). Trochlear and patellar morphology were classified according to the Dejour and Wiberg classifications, respectively on computed tomography (CT). Additional CT measurements included tibial tubercle–trochlear groove (TT-TG) distance and patellar tilt angle (PTA). Lower-limb rotational alignment was evaluated in a subset of patients using tibial external rotation angle (TERA) and femoral anteversion angle (FAA) on full-length lower-limb CT.ResultsThe immature group comprised 24 patients (33 knees, mean age 13.5 ± 1.4 years), and the mature group 61 patients (73 knees, mean age 23.3 ± 6.0). Wiberg type III patellar dysplasia was significantly more prevalent in the mature group (71.2% vs.18.2%, P < 0.001). Type B trochlear dysplasia predominated in the immature group (51.5%), whereas type A was most common in the mature group (46.6%). Mean CDI was significantly higher in the immature group (1.5 ± 0.2 vs. 1.3 ± 0.2, P < 0.001), with patella alta being more frequent (84.8% vs. 60.3%, P = 0.012). PTA (31.3° vs. 27.1°, P = 0.043) and FAA (28.3 ± 9.8° vs. 21.1 ± 9.6°, P = 0.017) were significantly greater in the immature group.ConclusionsIn patients with RPD, patella alta is more severe and prevalent in skeletally immature individuals. Immature patients also exhibit greater patellar tilt and increased femoral anteversion but less severe patellar dysplasia compared with skeletally mature adults.Level of evidenceIII.Keywords: Recurrent patellar dislocation, Skeletally immature, Patella alta, Trochlear dysplasia, Patellar dysplasia, Femoral anteversion Introduction Patellar dislocation accounts for approximately 3% of all knee injuries and most commonly affects adolescents and young adults [1–3]. Recurrent patellar dislocation (RPD) is the predominant form of patellofemoral instability in this population [4–6]. The reported incidence in children and adolescents reaches 50 per 100,000 [7]. RPD frequently causes persistent pain, swelling, cartilage damage, and early patellofemoral osteoarthritis, significantly impairing quality of life and sports participation [8–10]. RPD is typically associated with underlying anatomical risk factors, including increased tibial tubercle-trochlear groove (TT-TG) distance, patella alta, trochlear dysplasia, excessive femoral anteversion, and increased tibial external torsion [11–14]. However, current studies have mostly focused on the clinical outcomes of RPD, with less attention paid to fundamental aspects such as radiological parameters [15–17]. The interaction between skeletal maturity and anatomical risk factors remains poorly understood. During growth, some osseous deformities (such as trochlear dysplasia and TT-TG distance) may persist or even worsen, while others (such as patella alta) may partially improve as the skeleton matures [18–21]. Furthermore, skeletal maturity may mediate biomechanical changes in the patellofemoral joint, thereby influencing the development of its anatomical structures [22]. In particular, whether these deformities improve with skeletal maturation remains unclear. The aim of this study was to compare radiographic characteristics of RPD between skeletally immature and mature patients. We hypothesize that skeletally immature patients present more severe anatomical risk factors, particularly trochlear dysplasia and lower limb torsional deformities. Methods This retrospective study was approved by the Institutional Ethics Committee of our hospital (approval no. 2025238 K). Informed consent requirements were fulfilled. Patients Patients diagnosed with RPD between January 2016 and September 2024 were screened. Inclusion criteria were: (1) clinically confirmed RPD (two or more episodes of patellar dislocation), (2) no previous knee surgery, and (3) complete imaging data (lateral knee radiographs and knee CT), (4) no history or imaging evidence of patellar fracture. Patients were classified as skeletally immature or mature based on epiphyseal closure on radiographs. Basic information of the patients, such as sex (male/female), age, body mass index (BMI), history of knee injury, age at onset, affected side (left/right), and unilateral/bilateral, was extracted from medical records. Physeal closure was assessed on anteroposterior knee radiographs by a single trained observer (Y.H.) using predefined criteria: an open physis was defined as a visible radiolucent line between the epiphysis and metaphysis, whereas a closed physis was defined as complete obliteration of this line with trabecular bone continuity. During the measurement of Caton-Deschamps index (CDI) on lateral knee radiographs, all assessments were subsequently reviewed and confirmed by a second observer (X.Y.) to ensure consistency. Radiology assessment All patients underwent lateral knee radiographs (30°flexion) and knee CT. Full-length lower- limb CT was performed when rotational deformity was suspected clinically. Radiation exposure was minimized in immature patients using lead shielding. Patellar height was measured using the CDI [23]. Patella alta was defined as CDI ≥ 1.3 in immature patients and ≥ 1.2 in mature patients [24]. TT-TG distance and patellar tilt angle (PTA) were measured on axial CT according to Dejour et al. [25]. Trochlear morphology was classified using Dejour’s criteria [26], and patellar shape using Wiberg classification [27]. Femoral anteversion (FAA) and tibial external rotation angle (TERA) were measured on full-length CT Fig. 1 [28, 29]. All patients were supine with legs in full extension. The slice thickness was 4 mm. Measurement of lower extremity torsion; A: the line of posterior femoral condyle (red dotted line); B:FAA: the angle formed between a line passing through the centers of the femoral head and neck(red solid line) and the posterior femoral condylar line (red dotted line); C: the line to the posterior condyles of the proximal tibia(yellow dotted line); D:TERA: the angle formed between the line to the posterior condyles of the proximal tibia(yellow dotted line) and the malleolar line (yellow solid line); TERA, tibial external rotation angle; FAA, femoral anteversion angle All measurements were performed independently by two experienced observers (Y.H. and X.Y.). Intraclass correlation coefficients (ICC) were calculated, reliability was considered excellent if ICC > 0.75, fair to good if 0.4–0.75, and poor if < 0.40. Discrepancies in classifications were resolved by consensus. Statistical analysis Statistical analyses were performed using SPSS Version: 27.0 (IBM Corp., Armonk, NY, USA). Normal distribution was assessed with the Kolmogorov–Smirnov test. Continuous variables were compared using independent-samples t-tests (normally distributed data) or Mann-Whitney U tests (non-normal data). Categorical variables were compared using Pearson’s χ2 or Fisher’s exact test. Effect sizes were reported as Cohen’s d. A p-value < 0.05 was considered statistically significant. Results Of 122 individuals (146 knees) diagnosed with RPD in our department, 85 patients (106 knees) met the inclusion criteria Fig. 2. A total of 37 patients (40 knees) were excluded from this study: history of knee surgery (2 knees) and incomplete clinical or radiological data (38 knees). The immature group included 24 patients (33 knees; mean age 13.5 ± 1.4 years, range, 11–16), and the mature group 61 patients (73 knees; mean age 23.3 ± 6.0 years, range, 14–42) Table 1. Patient screening flowchart Table 1Patient characteristics aImmature group(n = 24, 33 knees)Mature group(n = 61, 73 knees)P-valueAge, y, range13.5 ± 1.4 (11–16)23.3 ± 6.0 (14–42)< 0.001Female sex, n (%)19 (79.2)44 (72.1)0.51BMI, kg/m221.2 ± 4.223.3 ± 4.80.05History of knee injury22 (66.7)51 (69.9)0.74Affected side (Left/Right)17/1635/380.73Bilateral affected, n (%)9 (37.5)12 (19.7)0.20Age at onset, y12.5 ± 1.819.7 ± 6.4< 0.001aData are presented as mean ± SD unless otherwise noted; Immature, skeletally immature; Mature, skeletally mature; BMI, body mass index. Patient characteristics a aData are presented as mean ± SD unless otherwise noted; Immature, skeletally immature; Mature, skeletally mature; BMI, body mass index. Trochlear and patellar dysplasia Wiberg type III patellar dysplasia was significantly more common in the mature group (71.2% vs. 18.2%, p < 0.001). Type B trochlear dysplasia predominated in the immature group (51.5%), whereas type A was most frequent in the mature group (46.6%) Table 2. Table 2Dejour and Wiberg classification aImmature group(n = 33 knees)Mature group(n = 73 knees)P-valueICC95%CIDejour classification––0.160.560.42 to 0.68Normal2 (6.1)2 (2.7)–––Type A11 (33.3)34 (46.6)–––Type B17 (51.5)22 (30.1)– ––Type C1 (3.0)9 (12.3)–––Type D2 (6.1)6 (8.2)–––Wiberg classification–– < 0.001 0.820.75 to 0.87Type I4 (12.1)5 (6.8)–––Type II23 (69.7)16 (22.0)–––Type III6 (18.2)52 (71.2)–––aData are presented as n (%); Immature, skeletally immature; Mature, skeletally mature; ICC, intraclass correlation coefficient; CI, confidence interval. Bolded values indicate statistical significance (p < 0.05). Dejour and Wiberg classification a aData are presented as n (%); Immature, skeletally immature; Mature, skeletally mature; ICC, intraclass correlation coefficient; CI, confidence interval. Bolded values indicate statistical significance (p < 0.05). Patellar height Mean CDI was significantly higher in the immature group (1.5 ± 0.2 vs.1.3 ± 0.2, p < 0.001; Cohen d = 0.79). Patella alta was present in 84.8% of immature knees versus 60.3% of mature knees (p = 0.012) Table 3. Table 3Radiological measurements outcome and interrater reliabilities aImmature group(33 knees)Mature group(73 knees)P-valueCohen dICC95% CICDI1.5 ± 0.21.3 ± 0.2 < 0.001 0.790.850.79–0.90Patella alta (≥ 1.3/1.2), n (%)28 (84.8)44 (60.3) 0.01 –NA–TT-TG (mm)19.0 ± 4.119.6 ± 4.30.480.150.890.85–0.93≥ 20 (mm), n (%)13 (39.4)32 (43.8)0.67–NA–PTA (°)31.3 ± 1127.1 ± 8.6 0.04 0.440.950.93–0.97Bold values indicate statistical significance (p < 0.05)aData are presented as mean ± SD unless otherwise noted; Immature, skeletally immature; Mature, skeletally mature; ICC, intraclass correlation coefficient; CDI, Caton−Deschamps index; TT-TG, tibial tuberosity‐trochlear groove; PTA, patellar tilt angle; ICC, intraclass correlation coefficient; CI, confidence interval.Bolded values indicate statistical significance (p < 0.05). Radiological measurements outcome and interrater reliabilities a Bold values indicate statistical significance (p < 0.05) aData are presented as mean ± SD unless otherwise noted; Immature, skeletally immature; Mature, skeletally mature; ICC, intraclass correlation coefficient; CDI, Caton−Deschamps index; TT-TG, tibial tuberosity‐trochlear groove; PTA, patellar tilt angle; ICC, intraclass correlation coefficient; CI, confidence interval.Bolded values indicate statistical significance (p < 0.05). Other measurements TT-TG distance did not differ significantly between groups (p = 0.484). PTA was significantly greater in the immature group (31.3°±11.0° vs. 27.1°±8.6°, p = 0.043, Cohen d = 0.44) Table 3. Lower limb rotational deformity Full-length CT was performed in 39 patients (54 knees) with suspected rotational deformity. FAA was significantly higher in the immature group (28.3 ± 9.8° vs. 21.1 ± 9.6°, p = 0.017, Cohen d = 0.75). No significant difference was found in TERA (35.0 ± 9.4° vs. 32.5 ± 5.6°, p = 0.34, Cohen d = 0.29) Table 4. Table 4Patients with lower limb rotational deformity aImmature group(n = 11,15 knees)Mature group(n = 28, 39 knees)P-valueCohen dICC95% CIAge, y13.6 ± 1.525.7 ± 6.8 < 0.001 NANANABMI, kg/m222.3 ± 5.223.4 ± 4.40.487NANANACDI1.5 ± 0.21.3 ± 0.2 0.021 0.72NANATT-TG19.0 ± 3.619.3 ± 5.00.8290.06NANAFAA (°)28.3 ± 9.821.1 ± 9.6 0.017 0.750.900.83–0.94TERA (°)32.5 ± 5.635.0 ± 9.40.3400.290.910.80–0.95aData are presented as mean ± SD unless otherwise noted; Immature, skeletally immature; Mature, skeletally mature; BMI, body mass index; CDI, Caton−Deschamps index; FAA, femoral anteversion angle; TERA, tibial external rotation angle; ICC, intraclass correlation coefficient; CI, confidence interval. Bolded values indicate statistical significance (p < 0.05). Patients with lower limb rotational deformity a aData are presented as mean ± SD unless otherwise noted; Immature, skeletally immature; Mature, skeletally mature; BMI, body mass index; CDI, Caton−Deschamps index; FAA, femoral anteversion angle; TERA, tibial external rotation angle; ICC, intraclass correlation coefficient; CI, confidence interval. Bolded values indicate statistical significance (p < 0.05). Inter-rater reliability was excellent for CDI, PTA, FAA, TERA, and Wiberg classification (ICC 0.82–0.95), but only fair for Dejour classification (ICC 0.56). Discussion The principal findings of this study are that skeletally immature patients with RPD exhibit more severe and prevalent patella alta, greater patellar tilt, and increased femoral anteversion, whereas skeletally mature patients demonstrate more severe (Wiberg type III) patellar dysplasia. However, because full-length lower limb CT was performed only in patients with suspected rotational deformities, conclusions regarding lower extremity torsional deformities should be interpreted with caution. Patella alta is a well-established risk factor for patellar instability [30, 31]. Higher normal CDI values are observed in children, and different thresholds are therefore applied [12, 24]. In this study, CDI ≥ 1.3 was defined as patella alta for skeletally immature patients [32]. Megremis et al. [33] reported 60% of knees (51/85) with patella alta in children with RPD. Additionally, Huang et al. [18] observed significant patella alta in children with RPD, with a mean CDI of 1.3. However, this finding is limited by the small sample size (18 patients, 22 knees). In this study, patella alta was present in 60.3% of knees in the skeletally mature group. Furthermore, a study by Zhang et al. [21] reported that 47% of knees in adults with RPD had patella alta. This condition appears to be more prevalent among skeletally immature patients. Furthermore, Patel et al. [34] reported that patella alta originates in early childhood rather than during adolescence. In this study, patient grouping was based on skeletal maturity (physeal closure status) rather than chronological age, because this distinction directly guides clinical decision-making (whether bony procedures are contraindicated). Although it is known that the CDI correlates with age [24], subgroup analysis by age could not be performed due to the limited sample size. Moreover, to the best of available knowledge, no published study has definitively determined whether chronological age or skeletal maturity exerts a stronger influence on CDI. Nevertheless, grouping by skeletal maturity is more clinically relevant in this population, as it directly reflects the patient’s growth potential and surgical options. Our findings of a mean CDI of 1.3 and patella alta in 84.8% of immature knees are consistent with previous paediatric series and confirm that patella alta is more pronounced in younger patients with RPD. In contrast, severe (Wiberg type III) patellar dysplasia was markedly more common (71.2%) in mature patients. In the immature group, Wiberg type II patellar dysplasia was the most prevalent form, accounting for 69.7% of cases. This supports the hypothesis that repetitive lateral stress and altered patellofemoral loading over time contribute to progressive flattening of the lateral facet, transforming type II into type III morphology. Huang et al. [18] found that 90.9% of knees (20/22) had Wiberg type II patellar dysplasia in children with RPD and Zhang et al. [35] reported that 89.1% of knees (49/55) had severe patellar dysplasia in skeletally mature patients with RPD. These previously reported rates closely align with our findings. Fucentese et al. [36] concluded that the decreased traction of medial patellofemoral may be related to patellar morphology. Panni et al. [37] found that increased lateral stresses may contribute to the development of Wiberg type III patellar dysplasia. Meissburger et al. [38] demonstrated a linear correlation in the severity of patellar and trochlear dysplasia; however, a possible genetic component for patellar morphology remains an open question. The relationship between patellar dysplasia and trochlear dysplasia in the knee has been underreported in the literature, yet the morphology of the patella may serve as a predisposing factor for patellofemoral instability. Although the distribution of trochlear dysplasia types differed between groups, the relatively low interobserver reliability of the Dejour classification (ICC 0.56) and its limitations in paediatric patients must be acknowledged. The Dejour classification is based on two-dimensional imaging data and cannot fully and accurately assess the geometric morphology of the femoral trochlea [39]. Furthermore, Martinez-Cano et al. [40] found that the Dejour classification for trochlear dysplasia has slight interobserver reliability and substantial intraobserver reliability. Yang et al. [41] developed a three-dimensional classification system for femoral trochlear dysplasia based on the 3D trochlear morphology. Sieberer et al. [39] proposed the concept of the entry point to trochlear groove angle (EPTG) to quantify the severity of femoral trochlear dysplasia in 3D models. Future studies will require larger sample sizes and more scientific assessment methods for trochlear dysplasia to validate the conclusions of this study. Increased patellar tilt in immature patients likely reflects the combined effect of more severe patella alta and rotational deformity rather than being an independent risk factor [26, 42, 43]. Factors such as tight lateral retinaculum, medial laxity, patella alta, and trochlear dysplasia are likely responsible for the increased patellar tilt [12, 44]. Kang et al. [45]found significant correlation between an increased patellar tilt and an increased femoral torsion in patients with RPD. According to the current literature, patellar tilt is more frequently a consequence resulting from a combination of multiple anatomical risk factors. The significantly higher femoral anteversion observed in the immature group aligns with previous reports linking excessive anteversion to patellar instability and a flatter trochlea. The relationship between lower limb rotational deformities and RPD is a significant topic in contemporary research. Excessive femoral rotation and tibial rotation have been identified as critical anatomical risk factors for recurrent patellar instability [46–49]. In the current study, the immature group exhibited lower tibial external rotation and significantly greater femoral anteversion compared to the mature group. Previous studies have linked excessive femoral anteversion to a flatter, more dysplastic trochlea [50, 51]. However, Qiao et al. [52]reported that lower extremity torsion was unrelated to trochlear dysplasia; their findings suggest that excessive femoral torsion originates primarily from the femoral shaft and distal femur. Regarding tibial external torsion, both groups demonstrated a high degree of torsion (> 30°) [53], although they did not differ statistically. Prakash et al. [54] concluded that tibial rotation in relation to the femur might be associated with the TT-TG distance for RPD patients. However, in the current study, the TT-TG distance was less than 20 mm in both groups, with no statistical difference between them. Additionally, Qiao et al. [52] demonstrated that both the proximal and distal tibia segments contribute to excessive torsion of the tibial. The correlation and causation between lower limb rotational deformities and patellar instability remain to be established through further research. Limitations This study has several limitations. Its retrospective design and exclusion of patients with incomplete imaging may introduce selection bias. Different CDI thresholds were used for immature and mature patients, reflecting current literature. However, owing to the limited sample size and retrospective design, multivariate analysis and age-subgroup comparisons were not performed. Age may influence the Caton-Deschamps index, which should be investigated in future larger studies. Full-length CT was performed only in patients with clinically suspected rotational deformity, potentially introducing selection bias. Moreover, MRI was not used to evaluate the imaging characteristics of the patients. In addition, physeal closure status was evaluated by a single observer, which may introduce bias despite the use of clear criteria. A few patients contributed both knees, which may have violated statistical independence. Advanced methods were not applied, but this potential issue does not undermine the main findings. Future studies should address this correlation. Finally, the cohort consisted exclusively of Han Chinese individuals and lacked a control group. Conclusion Skeletally immature patients with recurrent patellar dislocation present more severe patella alta, greater patellar tilt, and increased femoral anteversion compared with skeletally mature adults. Conversely, severe patellar dysplasia (Wiberg type III) is more common in mature patients. These age-related differences should be considered when planning surgical management of RPD. Acknowledgements Not applicable. Abbreviations Recurrent patellar dislocation Caton-Deschamps index Body mass index Tibial tubercle–trochlear groove Patellar tilt angle Tibial external rotation angle Femoral anteversion angle Author contributions Y.H. and X.Y. were responsible for data collection, analysis, and manuscript drafting. H. W. was responsible for study design and revision. All the authors read and approved the final manuscript. Funding None. Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval and consent to participate Ethical approval was obtained from the Ethics Committee of Zhongnan Hospital of Wuhan University (No. 2025238 K). And this study was conducted in accordance with the Declaration of Helsinki of the World Medical Association. Informed consent was obtained from all individual participants included in the study. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Clinical trial number Not applicable. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.