Risk factors for avascular necrosis after closed reduction for developmental dysplasia of the hip

Purpose The purpose of this study was to identify and evaluate risk factors of avascular necrosis (AVN) after closed treatment for developmental dysplasia of the hip (DDH). Methods A retrospective review of children diagnosed with DDH at a tertiary-care children’s hospital between 1986 and 2009 was performed. The presence of AVN was assessed according to Salter’s classification system. Results Eighty-two affected hips in 70 children with an average age of 10 months at closed reduction (range 1–31 months) and 5 years (range 2–19 years) of follow-up met the inclusion criteria. Twenty-nine (of 82, 35 %) affected hips developed AVN. The use of pre-reduction traction (p = 0.019) increased the risk of AVN, while preoperative Pavlik harness or brace trial (p = 0.28), presence of ossific nucleus at the time of closed reduction (p = 0.16), and adductor tenotomy (p = 0.37) were not significant factors. Laterality (right vs. left) was also not a significant risk factor (p = 0.75), but patients who underwent closed reduction for bilateral DDH were less likely to develop AVN (p = 0.027). Overall, the degree of abduction did not affect the rate of AVN (p = 0.87). However, in patients treated with closed reduction younger than 6 months of age, the rate of AVN was increased with abduction ≥50° (9/15, 60 %) compared to abduction <50° (0/8, 0 %) (p = 0.007). Patients who developed AVN were more likely to require subsequent surgery (p = 0.034) and more likely to report a fair/poor clinical outcome (p = 0.049). Conclusions The risk of AVN (35 %) following closed reduction and spica casting for DDH is high. The degree of abduction in spica casts appears to be a risk factor in patients ≤6 months old. The authors recommend that abduction in spica casts should be limited to <50° in children younger than 6 months of age. Level of evidence IV.

The exact etiology of AVN is unknown but likely multifactorial, with vascular and iatrogenic components [18,[26][27][28]. Prior literature identifies both a number of risk & Paul D. Choi pchoi@chla.usc.edu factors, such as hip abduction angle, and protective factors, such as early age at reduction, the presence of an ossific nucleus, and the use of an adductor tenotomy in the development of AVN [18-20, 22-25, 29-31]. Although some of these factors have been previously analyzed, the significance of each in the development of AVN has remained unclear. The purpose of this study was to identify and evaluate the risk factors and outcomes associated with AVN after closed reduction and spica cast application for children with DDH.

Materials and methods
After institutional review board approval, initial screening using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes for the diagnosis of DDH and Current Procedural Terminology (CPT) codes for closed reduction between January 1986 and January 2009 identified 287 consecutive patients at a single tertiary pediatric center. Inclusion criteria were a diagnosis of a subluxated or dislocated hip requiring closed reduction under general anesthesia in the operating room plus spica cast application, computed tomography (CT) scan following reduction, and a minimum of 18 months of clinical and radiographic follow-up. Exclusion criteria were a diagnosis of neuromuscular disease or teratologic dislocations, history of open reduction as initial management, incomplete radiographic data, history of multiple closed reductions, and onset of AVN after subsequent surgery. Retrospective electronic and physical chart reviews were performed to identify demographic information, clinical data, and length of follow-up. The need for subsequent procedures, whether open reduction of the hip or a reconstructive procedure (acetabular augmentation and/or femoral osteotomy), was also recorded. Images were reviewed by a single observer not involved in patient treatment. The initial preoperative plain radiographs were reviewed for the presence of a proximal femoral ossific nucleus. CT scans following closed reduction and spica casting were reviewed to measure the hip abduction angle by the method described by Browning et al. [32]. Plain radiographs at 1 year and final follow-up were reviewed to look for the development of AVN and to measure the acetabular index. The presence of AVN was evaluated using Salter's criteria [18]. Temporary irregular ossification, as defined by Salter et al., was not considered AVN [18]. Clinical outcomes were graded as excellent, good, fair, or poor based on the criteria devised by Brougham et al. [22]. Radiographic outcomes were graded based on the acetabular index, which was measured manually. Acetabular dysplasia was defined by an acetabular index C30°in the affected hip at 2 years of age [33].
Statistical analysis was performed using Stata 12 (Sta-taCorp LP, College Station, TX) and Microsoft Excel (2010). Univariate statistical analysis was performed by two-tailed Student's t-test and the Chi-squared test or Fisher' exact test in the case that values were too small for Chi-squared. A general estimating equation (GEE) was used to analyze the association of AVN with multiple variables while accounting for the cases of bilateral affected hips. Logistic multivariate regression was used for subgroup (closed reduction within six months of age) analysis, as the population was too small to use a GEE. Variables were included in either multivariate model if the univariate p-value \0.15 [34]. p-Value were considered significant at p \ 0.05. Univariate p-values are given for variables that did not meet the multivariate inclusion threshold (p \ 0.15). Effect sizes are reported as odds ratio (OR) or risk difference (RD) when unable to calculate the OR due to a zero being included in the calculation.

Results
Seventy patients with 82 hips underwent closed reduction under general anesthesia and spica cast application for a subluxated or dislocated hip and met the inclusion criteria. Sixty-two of these patients (89 %) were female and eight (11 %) were male. The left hip was involved in 40 patients, the right in 18 patients, and both hips in 12 patients. The mean age at the time of diagnosis was 8.0 months (range 0-31 months) and the mean age at the time of closed reduction was 10 months (range 1-31 months). The mean length of follow-up was 5 years (range 2-19 years).
Twenty-two (of 70, 31 %) had undergone Pavlik harness or hip abduction orthosis treatment prior to the closed reduction procedure under general anesthesia (duration 1-12 weeks) ( Table 1). In general, Pavlik harness or hip abduction orthosis was used as the initial treatment for a subluxated or dislocated hip in younger patients (younger than 6-9 months of age). Traction was used in 17 patients Twenty-one of these patients (75 %) were female and seven (25 %) were male. The left hip was involved in 18 patients, the right in nine, and both hips in one patient. Male patients were significantly more likely to develop AVN compared to female patients (p = 0.027). Patients with bilateral DDH were less likely to develop AVN (p = 0.027), while laterality (right vs. left) did not appear to affect the risk of AVN (p = 0.85).
No significant difference in the development of AVN was seen whether patients had undergone prior Pavlik harness or hip abduction orthosis treatment (p = 0.28) or whether adductor tenotomy was performed during the closed reduction procedure (p = 0.37). The use of pre-reduction traction appeared to increase the risk of AVN, with AVN developing in 65 % (11/17) of patients who underwent traction versus 32 % (17/53) who did not (p = 0.019). There was no significant difference in the rates of AVN in patients who underwent longitudinal versus overhead (vertical) traction (p = 0.71). Four of the 17 patients (five affected hips, one bilateral case) that underwent pre-reduction traction had undergone and failed brace treatment (Pavlik harness or hip abduction orthosis).
A prior history of brace treatment did not affect the risk of AVN in patients who underwent pre-reduction traction; the risk of AVN remained increased. In pre-reduction patients with history of brace treatment, three (of five) hips (in two of four patients) developed AVN, while in pre-reduction patients without history of brace treatment, nine (of 16) hips (in 9 of 13 patients) developed AVN (p [ 0.99).
No significant difference in the risk of AVN was seen whether proximal femoral ossific nucleus was present or not at the time of closed reduction (p = 0.16). The hip abduction angle also did not appear to affect the risk of AVN (p = 0.87) in the overall study population. However, in patients B6 months of age at the time of the closed reduction procedure, hip abduction angle C50°did significantly increase the risk of AVN [p = 0.007, RD 0.6, 95 % confidence interval (CI) 0.19-0.80] (Fig. 1). The proximal femoral ossific nucleus was present in six (of 23) hips that underwent closed reduction younger than 6 months of age; 42 (of 59) hips that underwent closed reduction were older than 6 months. The risk of AVN remained increased in this group of patients B6 months of age at the time of closed reduction and hip abduction angle C50°whether proximal femoral ossific nucleus was present or not. In patients younger than 6 months old and hip abduction angle C50°, eight (of 12) hips (in seven of 11 patients) developed AVN when the ossific nucleus was absent, compared to one (of three) hips (in one of three patients) when the ossific nucleus was present (p = 0.53). Furthermore, in patients without ossific nuclei (29 of 70 patients), hip abduction angle C50°did not significantly increase the risk of AVN (p = 0.29). The results of univariate and multivariate analyses to determine the significance of risk factors of AVN (including effect sizes and confidence intervals) can be found in Table 2.   = 0.59). The results of the univariate and multivariate analyses to determine the relationship between AVN and clinical and radiographic outcomes can be found in Table 5.

Discussion
AVN of the femoral head is a known complication following closed reduction and spica cast application for DDH (a subluxated or dislocated hip), with rates reported as high as 47 % [22]. Earlier studies have cited various possible risk factors of AVN, including age, gender, laterality, absence of proximal femoral ossific nucleus, use of pre-reduction traction, preliminary/history of hip abduction   and identified open reduction combined with concomitant osteotomies, redislocation, and need for secondary procedures after initial reduction as risk factors of AVN. Age, preliminary traction, and spica cast immobilization were not found to increase the risk of AVN [24]. In this series of 82 hips, the incidence of AVN was 35 % (29 of 82 affected hips). Similar to previous studies, a large number of risk factors for AVN after closed reduction and spica casting for a dislocated or subluxated hip were comprehensively reviewed. One specific focus of the study was the effect of hip abduction angle on the risk of AVN, in particular, whether the effects of hip abduction angle occur in an age-dependent manner.
In this series, gender was found to be a significant risk factor of AVN, with seven (of eight, 88 %) affected hips in males compared to 21 (of 62, 34 %) hips in females developing AVN (p = 0.027). A trend toward higher risk of AVN in male patients has been previously reported; however, it is unclear why male patients would be at higher risk [22,37]. Similar to previous studies, laterality (right vs. left) in this series was not a significant risk factor of AVN (p = 0.85) [22,35]. But unlike previous studies, this study revealed a decreased likelihood of developing AVN in patients undergoing treatment for bilateral DDH (p = 0.027). Morbi et al. found an increased risk of AVN in patients treated for bilateral DDH (33.3 %, 24 of 72 hips) compared to unilateral cases (11.2 %, 25 of 224 hips) [38]. Previous studies (like that of Morbi et al.) reporting an increased risk of AVN for bilateral DDH, however, have included patients treated with (closed and) open reduction [38,39]. The true effect of bilaterality on the risk of AVN is not clear from this study, as it was limited to patients undergoing closed reduction alone and excluded patients treated with open reduction.
The use of pre-reduction traction was found to be significantly associated with the development of AVN, with an increased rate of AVN in patients who had undergone pre-reduction traction (65 vs. 32 %) (p = 0.019). The higher risk of AVN in patients who underwent pre-reduction traction conflicts with traditional thinking that traction can decrease the risk of AVN, as reported in multiple studies [12,28,40,41]. These multiple studies explain that pre-reduction traction can decrease the need for open reduction and, thereby, reduce the risk of AVN. The exact relationship between pre-reduction traction and AVN, however, is unclear, as other studies have not found a decreased rate of open reduction or AVN with pre-reduction traction [8,22,24,35,42,43]. Subgroup analysis was performed to determine whether history of prior brace treatment (Pavlik harness or hip abduction orthosis) influenced the role of pre-reduction traction and risk of AVN. Four of the 17 patients (five affected hips, one bilateral case) who underwent pre-reduction traction had history of failed brace treatment. The risk of AVN remained increased in patients who had undergone pre-reduction traction-whether they had history of prior brace treatment or not (p [ 0.99). The risk of AVN in pre-reduction traction patients was also increased regardless of the type of traction (longitudinal vs. overhead-vertical) (p = 0.71).
Similar to other studies, the risk of AVN in this series was not affected by: presence of proximal femoral ossific nucleus, preliminary/history of hip abduction bracing, and adductor tenotomy [22-24, 35, 37, 44]. Upon initial analysis, hip abduction in the spica cast did not appear to affect the risk of AVN (p = 0.87): mean abduction angle of 57.3°  in hips without AVN compared to 57.7°  in hips that developed AVN. However, upon closer inspection of younger patients (less than or equal to 6 months of age at the time of the closed reduction procedure), abduction in the spica cast greater than or equal to 50°did significantly increase the risk of AVN (p = 0.007). Further analysis was performed to explore whether it was age and/or the presence of ossific nucleus that influenced this higher risk of AVN with high hip abduction angle. The risk of AVN remained increased in younger patients with hip abduction angle greater than 50°whether proximal femoral ossific nucleus was present or not (p = 0.53). Also, in all patients without proximal femoral ossific nuclei, hip abduction angle C50°did not increase the risk of AVN (p = 0.29). Previous literature considering excessive hip abduction as a potential risk factor of AVN lacks a consistent conclusion. Older reports by Salter et al. [18] and Gage and Winter [19] suggest extreme positions of abduction as a potential contributor to the development of AVN. In the years following, this has been both corroborated, in reports from Smith et al. [45] and Gregosiewicz and Wósko [29], and contradicted, as reported by Stanton and Capecci [21]. While our initial analysis of all 70 patients found no significant association between hip abduction and AVN, a more in-depth look identified a positive association in a subset of patients who underwent closed reduction at B6 months of age. Also notable was that, within this group, AVN only developed in patients 3-6 months of age at the time of closed reduction. This age-dependent association is consistent with prior clinical, experimental, and anatomical investigations which describe a changing vascular structure mid-way through the first year of life, with significant vulnerabilities to extreme abduction in its transition from bilateral to unilateral supply [18,23,27,[46][47][48][49][50]. In the future, a gadolinium-enhanced perfusion magnetic resonance imaging (MRI) after closed reduction/spica casting may help to decrease the risk of AVN by allowing for adjustments in the treatment protocol, including abduction angle in the spica cast, while maintaining perfusion/vascularity of the femoral head [51].
Outcome when AVN develops after closed reduction procedure for DDH can be compromised. The development of AVN increased the likelihood of subsequent surgery (p = 0.034), most commonly an acetabular augmentation procedure (13 cases), such as a Salter or Dega pelvic osteotomy. Most commonly, subsequent surgeries were needed because of persistent acetabular dysplasia and/or incomplete reduction. The development of AVN led to poorer clinical grades, with eight (of 29, 28 %) AVN-affected hips reported as fair to poor clinical outcome compared to three (of 53, 6 %) non-AVN hips.
There were several limitations in our study. First, because of its retrospective nature, the study was limited by available patient data. Second, the study period of 23 years was long, resulting in heterogeneity in treatment protocols. Third, with multiple treating surgeons, the role of surgeon skill is unclear and unknown. Finally, radiographic outcome data is subject to rater reliability and image accuracy. Despite the limitations, our study comprehensively examined risk factors for the development of AVN and outcomes following closed reduction and spica casting for DDH in a large patient pool treated at a single tertiary pediatric center. It is also the first to report on the relationship of hip abduction angle and young age on the risk of AVN.
In conclusion, although closed reduction and spica cast application remain an effective and successful treatment option for many patients with DDH, the development of AVN following a closed reduction procedure can be problematic, with our study reporting an incidence of 35 %. On the basis of our study, male gender and the use of pre-reduction traction may increase the risk of AVN. The degree of hip abduction appears to play a significant role in the development of AVN in younger patients (less than or equal to 6 months of age). Based on our results, the authors recommend that abduction in spica casts should be limited to less than 50°in children less than or equal to 6 months of age.
Informed consent For this type of study, formal consent is not required. A waiver of informed consent was obtained from our institutional review board for all individual participants included in the study.

Conflict of interest None in relation to the work submitted.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea tivecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.