Radiological outcomes in late-onset Perthes after femoral varus osteotomy

Abstract

Purpose

Treatment of late-onset Legg-Calvé-Perthes disease (LCPD) remains challenging, with outcomes varying by age and extent of involvement. This study aimed to evaluate the radiological outcomes following femoral varus osteotomy (FVO) in children presenting with LCPD after 8 years of age, to determine whether containment surgery provides benefit even in older children.

Methods

We retrospectively reviewed patients aged > 8 years with LCPD who underwent FVO between January 2008 and April 2019 at our institution, with a minimum two-year radiographic follow-up. Disease severity was graded using the modified Elizabethtown classification, Herring’s lateral pillar classification, and epiphyseal quotient (EQ). Radiographic outcomes at the latest follow-up were assessed using Deformity Index (DI), Caput Index (CI), Epiphyseal Extrusion Index (EEI), Sphericity Deviation Score (SDS), and Stulberg classification. Correlation analysis was performed using Spearman’s rank correlation.

Results

Twenty-seven children (22 boys, 5 girls; mean age 9.6 years, range 8.3–11.9) were included, with a median follow-up of 3.1 years (range 2–11). The mean DI was 0.31 (range 0.08–0.63) and showed a moderate positive correlation with the modified Stulberg grade (ρ = 0.57, p < 0.05). Fourteen of 27 hips (52%) achieved a DI < 0.3 at final follow-up. A lower postoperative EEI (p < 0.001) and lesser degree of lateral pillar collapse (p = 0.01) were significantly associated with improved femoral head sphericity.

Conclusion

The Herring classification and postoperative EEI significantly correlated with final femoral head sphericity. Containment by FVO in children presenting after 8 years of age results in good radiological outcomes in over half (50%) of cases, indicating that age alone should not preclude surgical containment in late-onset LCPD.

Level of evidence

Level III retrospective therapeutic study.

Introduction

Legg-Calvé-Perthes disease (LCPD) is a childhood disorder of the femoral head characterized by temporary interruption of its blood supply, resulting in varying degrees of osteonecrosis and deformation. The prognosis depends on the child’s age at onset and the extent of epiphyseal involvement. Younger children (< 6 years) usually experience a milder course, whereas late-onset cases, defined by disease presentation after 8 years of age, often have a more severe progression and poorer outcomes [1]. Nonoperative management has limited influence on the natural history of severe LCPD in older children. Surgical containment procedures aim to improve femoral head sphericity and long-term congruence with the acetabulum. Among these, femoral varus osteotomy (FVO) has been widely performed and shown to improve containment and radiological outcomes compared with conservative management or natural history studies in older children [2–6]. Joseph et al. suggested that FVO should ideally be performed before the late fragmentation stage (stage IIb) for optimal results [7]. However, recent studies by Stepanovich et al. and Yavuz et al. have demonstrated that favourable outcomes can still be achieved with containment, independent of age, even in later stages of Perthes disease [8, 9].

A variety of surgical options have been described for late-onset or severe Perthes disease, including arthrodiastasis, labral shelf arthroplasty, and transtrochanteric rotational osteotomy. Arthrodiastasis techniques using the Ilizarov or Orthofix systems can improve joint motion but are technically demanding and associated with complications such as pin tract infections and poor tolerance in children [10, 11]. Labral shelf procedures and transtrochanteric osteotomy have also shown benefit but require advanced surgical skill and are often reserved for salvage or late-stage situations [12, 13]. FVO remains a time-tested, technically straightforward containment procedure with predictable results and minimal additional morbidity.

It is important to distinguish between late-onset and late-stage Perthes disease. “Late-onset” in this study refers to children who presented after the age of 8 years, not to those in advanced radiographic stages of disease. To ensure comparability of outcomes and avoid the confounding effect of irreversible femoral head collapse, patients with advanced stages (modified Elizabethtown grades III and IV) were excluded. Therefore, the study specifically evaluates outcomes of containment surgery in older children presenting with active but reconstructible disease. Based on existing literature and our institutional experience, we hypothesized that older children with late-onset LCPD, even with substantial femoral head involvement but without end-stage collapse, can benefit radiologically from FVO. This study analyses the correlation between various preoperative and postoperative radiographic parameters and final femoral head sphericity to identify factors influencing the outcome of containment surgery in this group.

Materials and methods

This retrospective study was conducted at a single tertiary paediatric orthopaedic center after obtaining institutional review board approval (IRB Min. No. 13210, dated 22 July 2020). The IRB provided a waiver for both ethical approval and the requirement for informed consent due to the retrospective design. All children diagnosed and treated for LCPD between January 2008 and April 2019 were reviewed. Children older than 8 years of age at the time of presentation who underwent femoral varus osteotomy (FVO) for LCPD, with a minimum of 2 years of radiographic follow-up and complete perioperative records, were included in the study. Patients were excluded if they had received prior treatment for LCPD, had a history of significant trauma, systemic disease (such as coagulopathy or sickle cell disease), or presented with advanced stage (modified Elizabethtown grade IV), as this stage represents irreversible collapse .

During the study period, 276 children underwent surgical management for LCPD at our institution. Of these, 27 children (22 boys and 5 girls) fulfilled the inclusion criteria. Standardized anteroposterior (AP) pelvic radiographs and Lauenstein lateral views at initial presentation and latest follow-up were reviewed for each patient. Preoperative staging and disease severity were graded according to the modified Elizabethtown classification, Herring’s lateral pillar classification, and epiphyseal quotient (EQ) [7, 14, 15]. Final radiographic outcomes were assessed using the Deformity Index (DI) [16], Caput Index (CI) [17], articular-trochanteric distance index (ATDI) [18], epiphyseal extrusion index (EEI) [19], sphericity deviation score (SDS) [20], and modified Stulberg classification [20]. For ease of comparison, Stulberg classes 1 and 2 (spherical, congruent hips) were grouped as A, class 3 (congruent but ovoid) as B, and classes 4 and 5 (aspherical/flat hips) as C.

All radiographic measurements were independently performed by a paediatric orthopaedic fellow and a consultant musculoskeletal radiologist, both blinded to preoperative classification and patient outcomes. Interobserver differences were resolved by consensus. In a subset of patients, intravenous pamidronate was administered perioperatively following informed discussion with parents.

Surgical technique

The FVO technique was performed as described by Axer et al. [21] All patients had a minimum hip abduction of 30° preoperatively. The osteotomy was performed at the intertrochanteric level and stabilized without rotation. The degree of varus correction was determined intraoperatively to achieve optimal acetabular containment. Fixation was achieved using either a paediatric angled blade plate (DePuy Synthes Cannulated Paediatric Osteotomy system) or a bent dynamic compression plate in earlier cases [4]. Older children typically required greater varus correction. Patients were gradually mobilized with crutches and toe-touch weight bearing over 3–4 months postoperatively.

Statistical analysis

Data analysis was performed using IBM SPSS Statistics, version 21.0 (IBM Corp., Armonk, NY, USA). Normality of continuous variables was assessed using the Shapiro–Wilk test. Categorical variables were analysed using Fisher’s exact test, while continuous variables were compared using the Student’s t-test or Wilcoxon rank-sum test, as appropriate. Correlations between radiographic parameters (e.g., DI, EEI, CI, SDS) and the modified Stulberg groups were analysed using Spearman’s rank correlation coefficient (ρ). Statistical significance was set at p < 0.05.

Results

During the 11-year study period (2008–2019), 276 children underwent surgical treatment for LCPD at our institution. Twenty-seven patients (22 boys and 5 girls) met the inclusion criteria and were analysed. The mean age at diagnosis was 9.6 years (range, 8.3–11.9; SD = 0.91), and the median radiographic follow-up duration was 3.1 years (range, 2–11; IQR = 2.59).

Based on the Herring’s lateral pillar classification, one patient was type A, eleven were type B, five were type B/C, and ten were type C. According to the modified Elizabethtown classification, six children were type I, two were type IIa, eighteen were type IIb, and one was type III. None were grade IV, as these were excluded per study criteria. Ten children received perioperative intravenous bisphosphonate infusion (pamidronate).

At the final follow-up, 11 hips (41%) were classified as Stulberg Group A (spherical/congruent), 14 hips (52%) as Group B (ovoid but congruent), and 2 hips (7%) as Group C (aspherical/flat) (Figs. 1 and 2). The mean Deformity Index (DI) was 0.31 (range, 0.08–0.63; SD = 0.12) and demonstrated a moderate positive correlation with the modified Stulberg group (Spearman’s ρ = 0.57, p = 0.002).

Fig. 1.

a: Preoperative radiograph with AP and Launstein views in a child with right hip Herring type B LCPD. b: Intraoperative fluoroscopy images to determine adequate varus and ensure epiphyseal containment. c: Postoperative AP view at 18 months follow-up with spherical concentric right femoral head (modified Stulberg Group A)

Fig. 2.

a: Preoperative AP radiograph in a child with Right hip Herring type C LCPD. b: Preoperative Launstein view in a child with right hip Herring type C LCPD. c: Postoperative AP view at 79 months follow-up demonstrating concentric spherical right femoral head (modified Stulberg Group A)

The mean Caput Index (CI) was 0.81 (SD = 0.16), mean Epiphyseal Quotient (EQ) was 67.6% (SD = 19.9), and mean Sphericity Deviation Score (SDS) was 37.3 (SD = 17.8). These variables showed no significant correlation with either DI or modified Stulberg classification (p > 0.05). The mean postoperative Epiphyseal Extrusion Index (EEI) was 0.11 (range, 0–0.46; SD = 0.14), and the median Articular-Trochanteric Distance Index (ATDI) was 0.37 (range, − 0.89 to 1.03; IQR = 0.56). Seven patients (25%) required revision surgery for symptomatic impingement during follow-up (one valgus osteotomy and six osteochondroplasties) [Table 1].

Table 1.

Baseline clinical and radiological characteristics of children who underwent femoral varus osteotomy for late-onset Perthes disease (n = 27)

Characteristics	Sub-category	n (%) / Value
Total patients	—	27
Age at presentation (years)	—	Mean 9.6 (SD 0.91); range 8.3–11.9
Duration of follow-up (years)	—	Median 3.1 (IQR 2.59); range 2–11
Gender	Male	22 (81%)
	Female	5 (19%)
Herring’s type	A	1
	B	11
	B/C	5
	C	10
Modified Elizabethtown classification	I	6
	IIa	2
	IIb	18
	III	1
	IV	0
Adjunct bisphosphonate treatment	Yes	10
	No	17
Deformity Index (DI)	—	Mean 0.31 (SD 0.12); range 0.08–0.63
Caput Index (CI)	—	Mean 0.81 (SD 0.16); range 0.38–0.99
Articular–Trochanteric Distance Index (ATDI)	—	Median 0.37 (IQR 0.56); range − 0.89 to 1.03
Post-operative Epiphyseal Extrusion Index (EEI)	—	Mean 0.11 (SD 0.14); range 0–0.46
Epiphyseal Quotient (EQ)	—	Mean 67.6% (SD 19.9); range 21–94
Sphericity Deviation Score (SDS)	—	Mean 37.3 (SD 17.8); range 10.8–72.5
Modified Stulberg classification	Group A	11
	Group B	14
	Group C	2
Salvage procedure	Valgus osteotomy	1
	Osteochondroplasty	6

Patients were categorized based on final femoral head sphericity using the Deformity Index (DI).Group I (DI < 0.3) included 14 hips (52%),Group II (DI ≥ 0.3) included 13 hips (48%).

The mean DI was 0.23 (SD = 0.06) in Group I and 0.41 (SD = 0.09) in Group II (p < 0.0001). Mean age, sex distribution, and follow-up duration were comparable between the two groups. Three children in Group I and four in Group II required secondary impingement surgery (p = 0.67).

Nine of the fourteen hips (64%) in Group I had < 50% lateral pillar collapse (Herring B/B-C), while eight of thirteen hips (62%) in Group II had > 50% lateral pillar collapse (Herring C). A significant association was observed between the Herring classification and final sphericity outcome (p = 0.016). The modified Elizabethtown classification did not demonstrate a significant relationship with final DI (p = 0.59).

The mean postoperative EEI was 0.02 (SD = 0.05) in Group I and 0.20 (SD = 0.15) in Group II, indicating superior containment among patients with better sphericity outcomes (p < 0.001). The Caput Index (CI), EQ, SDS, and ATDI did not differ significantly between the two groups (p > 0.05). The improved containment (lower EEI) was the only surgeon-controlled factor significantly associated with better femoral head sphericity [Table 2].

Table 2.

Comparison of factors affecting femoral head sphericity following femoral varus osteotomy in late-onset Perthes disease

Variable	Group I (DI < 0.3)	Group II (DI ≥ 0.3)	p value
Total (n)	14	13	—
Mean age (years)	9.85 (SD 1.06); range 8.7–11.9	9.41 (SD 0.69); range 8.5–10.9	0.22
Gender (M/F)	12 / 2	10 / 3	0.38
Median follow-up (years)	3.5 (IQR 2.79); range 2–11	4.0 (IQR 2.73); range 2–9	—
Herring’s type	B = 9 B/C = 3 C = 2	A = 1 B = 2 B/C = 2 C = 8	0.016
Modified Elizabethtown classification	I = 4 IIa = 0 IIb = 9 III = 1	I = 2 IIa = 2 IIb = 8 III = 1	0.59
Adjunct bisphosphonate use (Yes/No)	5 / 9	5 / 8	1.00
Mean Deformity Index (DI)	0.23 (SD 0.06); range 0.08–0.29	0.41 (SD 0.09); range 0.30–0.63	< 0.0001
Mean Caput Index (CI)	0.82 (SD 0.13); range 0.68–0.95	0.79 (SD 0.19); range 0.38–0.99	0.86
Median Articular–Trochanteric Distance Index (ATDI)	0.36 (IQR 0.63); range − 0.89 to 1.03	0.39 (IQR 0.44); range − 0.55 to 0.84	0.94
Mean Post-operative Epiphyseal Extrusion Index (EEI)	0.02 (SD 0.05); range 0–0.14	0.20 (SD 0.15); range 0–0.46	< 0.001
Mean Epiphyseal Quotient (EQ)	74.2% (SD 13.3); range 62–100	60.4% (SD 23.6); range 21–94	0.07
Mean Sphericity Deviation Score (SDS)	34.6 (SD 16.2); range 10.8–61.2	40.2 (SD 19.7); range 14.8–72.5	0.20
Modified Stulberg classification (A/B/C)	9 / 5 / 0	2 / 9 / 2	—

Discussion

The primary goal in managing Legg-Calvé-Perthes disease (LCPD) is to restore a spherical femoral head and maintain acetabular congruence at skeletal maturity. Multiple surgical containment procedures have been described, including Salter innominate osteotomy (SIO), triple innominate osteotomy (TIO), labral shelf acetabuloplasty, and femoral varus osteotomy (FVO), each with variable success rates in older children [8, 10–13, 22–24]. Our study focused on children who presented after eight years of age—the late-onset group—evaluating whether FVO could still yield satisfactory radiological containment and sphericity outcomes.

A key clarification is that late-onset refers to chronologic age at presentation and not the radiographic stage of disease. Patients in advanced fragmentation or collapse (modified Elizabethtown IV) were deliberately excluded because irreversible deformity prevents containment benefit. Our conclusions apply to older children with reconstructible disease and not to end-stage Perthes hips.

Our results are consistent with prior studies reporting beneficial outcomes of FVO in older children. Joseph et al. achieved satisfactory sphericity in patients treated before late fragmentation [7]. Terjesen et al. demonstrated improved femoral-head sphericity in severe forms of LCPD following FVO, even in children over nine years [6]. Similarly, Joshi et al. and Elzohairy et al. described short-term improvement in containment indices and radiological remodeling after varus osteotomy in late-presenting cases [25, 26]. These observations support the concept that age alone should not preclude surgical containment when the head remains mouldable.

Other authors have reported comparable or superior outcomes using pelvic procedures for late-onset LCPD. Stepanovich et al. reported 72% spherical heads (Stulberg I/II) after TIO, while Yavuz et al. achieved 55% good results with SIO in late-presenting hips [8, 9]. However, these operations demand longer surgical time, greater blood loss, and risk of over-coverage or pincer impingement. FVO, by contrast, remains technically straightforward, reproducible, and associated with fewer complications, particularly in centres where advanced pelvic reconstruction may not be routinely practiced.

In this series, 52% of hips achieved near-spherical configuration (DI < 0.3) at a mean follow-up of 3.5 years. The post-operative Epiphyseal Extrusion Index (EEI) was the only surgeon-controlled factor showing significant association with final sphericity (p < 0.001), confirming that better containment predicts improved femoral-head shape. The Herring’s lateral-pillar classification also correlated with outcome (p = 0.016), while variables such as CI, EQ, and SDS did not. The moderate Spearman correlation (ρ = 0.57, p < 0.05) between DI and modified Stulberg groups reinforces that DI is a reliable radiographic surrogate for long-term sphericity assessment.

These findings emphasize that adequate containment, rather than age alone, determines prognosis in older children. Achieving lower EEI values may require slightly greater varus correction to ensure acetabular coverage in hips with borderline containment. Concerns regarding limb-length discrepancy or knee malalignment after FVO have been mitigated in recent long-term series, where discrepancies ≤ 0.5 cm were clinically insignificant [18, 27]. Thus, FVO remains an appropriate containment option for late-onset LCPD when performed with good pre-operative planning to ensure adequate containment.

Although systemic bisphosphonates have been investigated to reduce femoral-head deformity in LCPD, evidence remains inconclusive [28]. In our cohort, bisphosphonate use did not significantly influence radiological outcome (p = 1.00), consistent with prior controlled studies showing limited efficacy in preserving femoral-head shape [29]. Larger prospective trials are required to clarify their role.

This study has several limitations. The retrospective design and modest sample size introduce inherent bias. In addition, a heterogeneous treatment cohort in which some patients received adjunct intravenous pamidronate may have influenced the rate of femoral-head remodeling, potentially confounding interpretation of the surgical effect alone. The absence of a control group (non-operative or alternative containment procedures) further restricts comparison with natural disease progression or other interventions. Functional outcomes and patient-reported scores were not evaluated, limiting direct correlation between radiographic improvement and clinical function. Despite these constraints, uniform surgical technique, consistent radiographic measurement by blinded observers, and a minimum two-year follow-up lend reasonable reliability to the study’s conclusions.

Conclusion

This study demonstrates that femoral varus osteotomy provides satisfactory radiological outcomes in more than half (50%) of children presenting after eight years of age with reconstructible Legg-Calvé-Perthes disease (LCPD). The degree of containment, reflected by a lower postoperative epiphyseal extrusion index (EEI), and the Herring’s lateral pillar classification were the strongest predictors of final femoral-head sphericity. These findings suggest that adequate containment, rather than age alone, determines prognosis in older children. While future prospective, comparative studies are warranted, our results support offering timely containment surgery to children with late-onset LCPD before irreversible collapse occurs for improved outcomes.

Abbreviations

LCPD

Legg–Calvé–Perthes disease

FVO

Femoral varus osteotomy

EQ

Epiphyseal Quotient

DI

Deformity Index

CI

Caput Index

EEI

Epiphyseal Extrusion Index

SDS

Sphericity Deviation Score

ATDI

Articular–Trochanteric Distance Index

Authors’ contributions

S.P.T., A.G., and T.P. conceived the study and designed the research.S.P.T. and J.R. prepared the figures, acquired data, and assisted with bibliography and references.T.P., A.G., and J.R. compiled the tables, verified data accuracy, and contributed to statistical analysis.J.P. reviewed and graded all radiographs and contributed to radiological measurements and manuscript review.All authors contributed to drafting and critical revision of the manuscript, approved the final version, and agree to be accountable for all aspects of the work.

Funding

No external funding was received for this study.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was reviewed and approved by the Institutional Review Board (IRB), Christian Medical College, Vellore, India, Office of Research (Ethics Committee – Silver) (IRB Min. No. 13210 [Retro], dated 22 July 2020). The Institutional Review Board granted a waiver of informed consent due to the retrospective nature of the study and the use of anonymized patient data. All procedures involving human participants were conducted in accordance with the ethical standards of the institutional research committee and with the principles of the Declaration of Helsinki (as revised in 2013).

Consent for publication

The Institutional Review Board granted a waiver for obtaining individual patient consent for publication, as no identifiable patient data were included.

Competing interests

The authors declare no competing interests.