Comparison of baseline outcomes between surgical and nonoperative management in youth with lower extremity torsional abnormalities

1 Introduction

Lower extremity torsional abnormalities (LETA) commonly affect children and include excessive torsion of the femur and/or tibia. While most cases resolve with growth, some persist and can lead to symptoms that affect quality of life. Symptoms may include hip, knee or ankle pain, patellar instability, and functional limitations.^1–3 When symptomatic, management options include nonoperative approaches such as bracing and physical therapy or surgical correction via derotational osteotomy.

Most studies have focused on surgical outcomes or on the overall effects of torsion in mixed cohorts that include patients managed both surgically and nonoperatively, with little attention to differences between these groups.^1,4 Some studies suggest that foot orthoses may help improve abnormal foot progression angles.^2,5–8 Evidence for physical therapy is sparse. One case report described minimal improvement with physical therapy, with subsequent need for surgery.⁹ Several studies noted failed courses of physical therapy before surgery.^4,10,11 Conversely, derotational osteotomy has been shown to improve physical function, gait patterns and pain in patients with LETA.^2,5–8 However, few studies have examined baseline differences between patients who undergo surgery and those managed with nonoperative methods.

In clinical practice, many orthopedic surgeons recommend physical therapy as first-line treatment, reserving surgery for patients who do not respond to physical therapy and whose quality of life remains affected.^10,12,13 Despite this general approach, there are no clear clinical guidelines delineating which patients should undergo surgery. Identifying baseline differences between patients managed surgically and those treated with nonoperative methods may clarify factors that influence surgical decision-making and support the development of more standardized treatment criteria for young patients with LETA.

Therefore, the objective of this study was to determine whether patients with LETA who were referred to nonoperative management differed from those referred to surgery. It was hypothesized that the nonoperative group would be younger and would report less pain and fewer physical limitations than the surgical group.

2 Methods

3 Results

A total of 21 patients (female = 17, male = 4) in the surgical group, 13 patients (female = 11, male = 2) in the nonoperative group and 21 in the control group (female: 16, male = 5) were included. There was no statistically significant group difference in age (nonoperative: 14.7 ± 2 years, surgery: 16.3 ± 2.3 years, control: 15.8 ± 2.7; p = 0.06), body mass (nonoperative: 54.4 [95 % CI: 48.0, 60.8]kg, surgery: 52.2 [95 % CI: 48.3, 58.7]kg, control: 57.3 [95 % CI: 52.9, 59.8], p=0.83), height (nonoperative: 1.6 ± 0.1 m, surgery: 1.7 ± 0.1 m, control: 1.6 ± 0.1m, p = 0.32), and body mass index (nonoperative: 19.3 [95 % CI: 19.1, 21.5] kg/m², surgery: 19.1 [95 %CI: 17.6, 20.8]kg/m², control: 20.412 [95 % CI: 19.599, 22.370] kg/m², p = 0.48). Pain duration was also similar (nonoperative: 72.0 [95 % CI: 15.0, 84.0] months, surgery: 48.0 [95 % CI: 31.0, 60.0] months, p = 0.74).

3.3

3.3 Quantitative gait analysis

Kinematic patterns were similar between treatment groups across the pelvis, hip, knee and ankle (Fig. 1). Significant differences occurred at the knee during terminal stance and swing (37–47 % of the gait cycle), with the nonoperative group showing greater knee flexion than the control group. In the transverse plane, when compared to the control group, the surgical group demonstrated slightly greater external knee rotation (30–40 % of the gait cycle) and internal foot rotation (0–6 %, 33–36 %, 88–100 % of the gait cycle). Overall, kinetic data were also comparable (Fig. 2). Small, short-duration significant clusters were identified for the hip rotation moment (48–52 % and 66–68 % of the gait cycle) with the nonoperative group having lower moment than the control group. However, the magnitude of these differences was modest.

Close

Fig. 1 Lower-limb kinematics across the gait cycle. Between-group differences were confined mainly to the transverse plane (knee and foot rotation; brief clusters).

Export to PPT

Close

Fig. 2 Lower-limb kinetics across the gait cycle. Patterns were similar between groups; brief significant clusters were observed for hip rotation moments.

Export to PPT

4 Discussion

The study investigated whether patients referred to nonoperative management differed from those referred for surgery. It was hypothesized that the nonoperative group would be younger and report less pain and fewer physical limitations. While not statistically significant, the surgical group was older by 1.6 years (nonoperative: 14.7 ± 2 years, surgery: 16.3 ± 2.3 years, p = 0.055), consistent with the common clinical practice of deferring derotational osteotomy until skeletal maturity. Although this age difference did not reach statistical significance, in mid-adolescence, a 1–2-year gap can be clinically meaningful,²⁸ and the study was likely underpowered to detect this magnitude of difference. Contrary to the initial hypothesis, pain measures did not differ between treatment groups, and PROs showed a mixed pattern. However, compared to the control group, both treatment groups reported more pain and lower physical function.

Across PROs, the two treatment groups were broadly similar when analyzed alongside controls. An apparent pairwise difference on the PEDI-IKDC (surgical lower than nonoperative) was not significant after the three-group ANOVA with post-hoc testing. Nonetheless, the between-group difference fell within published minimally clinically important difference ranges for PEDI-IKDC (5.6–20.5), suggesting potential clinical relevance despite the lack of statistical significance.²⁹ In contrast, the nonoperative group scored lower than both the surgical and control groups on HAGOS Activities of Daily Living and Function in Sport and Recreation. Objective physical examination findings (ROM, relative muscle strength, and generalized joint mobility) did not differ between treatment groups, although ROM differed from controls. The clearest treatment-group differences were in patellofemoral-specific clinical tests. The surgical group had higher rates of MPFL laxity, patellar apprehension, and patellar maltracking. Taken together, these findings suggest that, overall, surgical patients exhibited more severe knee symptoms and knee-specific functional limitations than the nonoperative group, even though between-group differences were modest.

Given that elective derotational osteotomy is often undertaken for pain-related disability, higher pain was expected in the surgical group. Instead, both treatment groups had higher pain than controls, but were similar to each other. They had similarly low PODCI Comfort & Pain scores (≈58 %) and moderate APPT pain intensity (≈3–4/10), and moderate sensory, affective and evaluative pain qualities. PROMIS Pain Interference scores were higher than controls (53–54 vs 35), indicating an impact on daily activities. Notably, the surgical group had more patellofemoral symptoms, a mechanical indication that can drive surgery independent of constant pain. Patients with patellofemoral instability may experience severe pain during/after dislocation episodes. Because the PROs captured the prior week, recent dislocations may not have occurred at the time of completion, attenuating between-group differences. Consequently, only patients with more constant torsion-related pain would influence mean pain scores.

On imaging and physical examination, both treatment groups differed from normative data and from controls, but they were similar to each other in femoral anteversion, external tibial torsion, passive ROM, relative muscle strength and generalized joint laxity. Each treatment group included patients with isolated femoral anteversion, isolated external tibial torsion, and combined deformities. This case mix likely influenced group averages for femoral version, tibial torsion and ROM depending on the proportion of each pattern. In contrast, patellofemoral-specific tests (MPFL laxity, apprehension test and J-sign) were significantly more frequent in the surgical group, indicating greater patellar instability. This pattern aligns with current surgical decision-making, in which patellar instability in the context of LETA is a key indication for surgery¹³ and is consistent with reports of improved outcomes when torsional deformities are corrected at the same time with osteotomy compared with ligament surgery alone.⁶

Kinematic and kinetic differences among the three groups were limited to small regions of the gait cycle. Limited contrasts likely reflect heterogeneity in torsional deformities pattern (isolated femoral version, isolated external tibial torsion, combined torsion) and compensatory strategies. For example, some patients with external tibial torsion exhibit increased internal hip rotation during gait, whereas others do not, diluting group-level effects when averaged.³⁰ One notable kinematic difference was greater knee flexion at terminal stance in the nonoperative group, which may reflect a protective strategy to reduce patellofemoral demand or apprehension. This compensation has been described for patients with LETA.^1,30–32 Conversely, increased knee extension during stance has also been described to reduce joint loading and quadriceps demand.^33,34 Such opposing adaptations attenuate mean differences. Visual inspection of individual curves confirmed the presence of both strategies among patients. Accordingly, despite muted group differences, gait analysis remains clinically useful to identify compensatory mechanisms prior to nonoperative or surgical management, as outcomes may differ depending on whether such mechanisms are present.^2,30

PROs suggested that different joints were most impacted in each treatment pathway. The surgical group tended to show more knee-focused limitations (PEDI-IKDC), while the nonoperative group reported more hip/groin limitations (HAGOS). Practically, PEDI-IKDC is routinely collected at gait visits and readily available to surgeons, whereas HAGOS was study-specific. Therefore, knee-specific information may be more visible in routine reports. However, in clinical practice, PROs document symptom burden and track response to care. They complement, rather than determine, surgical indications, which are ultimately guided by the combination of physical examination, gait analysis, imaging findings, and patient subjective information.

5 Conclusions

In this cohort of young patients with LETA, both the surgical and nonoperative groups showed meaningful impairment on function, physical examination, and imaging, supporting the need for treatment in either pathway. However, treatment group differences were modest overall. Surgical referrals in practice appeared to align most with patellofemoral instability in the context of torsional deformity, whereas pain severity, passive ROM, and relative muscle strength did not clearly separate pathways. PRO patterns suggested a tendency towards greater knee-specific limitation in the surgical cohort and more hip/groin limitation in the nonoperative cohort, but these differences were small. Rather than relying on any single metric, surgeons use a patient-specific, integrated synthesis of imaging, physical examination, gait findings, and PROs. The interaction among these factors guides decision-making and is not fully reflected in group means. The small differences reported between both pathways truly highlight the need for clear, consensus-based guidelines to determine treatment pathways. The findings of this study also support a systematic assessment framework that includes patellofemoral testing, joint-specific PROs, imaging, and instrumented gait analysis, and they underscore the importance of screening for concomitant torsional deformity in patients presenting with patellofemoral instability.

Guardian/patient's consent

Written informed consent and assent (when applicable) were obtained from all patients and/or their legal guardians.

Credit author statement

-Funding Acquisition: Marianne Gagnon (MG), Karen M Kruger (KMK), Mitchell Bernstein (MB), Louis-Nicolas Veilleux (LNV)-Conceptualization: MG, MB, LNV-Data curation: Sena Tavukcu (ST), MG-Formal analysis: MG-Supervision: MB, LNV-Methodology: MG, MB, LNV-Project Administration: ST, KMK, LNV-Investigation: MG, KMK, MB, LNV-Writing - original draft: MG-Writing – review & editing: MG, Jeremy Bauer (JB), KMK, ST, Haluk Altiok (HA), Reggie Hamdy (RH), MB, LNV

Disclosure

During the preparation of this work, the authors used ChatGPT (OpenAI, San Francisco, CA, USA) to assist with improving readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Ethical statement

This study is approved by the McGill Institutional Review Board (A02-M04-22B/20222336). Written informed consent and assent (when applicable) were obtained from all patients and/or their legal guardians.

Funding statement

This study was funded by a developmental grant from Shriners’ s Children (#79155, 2022–2024). MG was supported by doctoral scholarships from the Fonds de recherche du Québec-Santé [#298324; 2021–2022]; and the Canadian Institutes of Health Research [#476791; 2022–2025].