Background: Severe hip contractures in arthrogrypsosis are multiplanar, which can preclude or can greatly complicate sitting and ambulation. The reorientational osteotomy at the intertrochanteric level preserves preoperative hip motion but moves it to a more functional domain. We retrospectively compared preoperative and postoperative hip motion and evaluated the ambulatory abilities of patients who underwent the procedure.
Methods: Since 2008, 65 patients with arthrogryposis had 119 reorientational proximal femoral osteotomies with a minimum follow-up of 2 years. The mean patient age at the time of the surgical procedure was 48 months. An intertrochanteric wedge osteotomy aligned the femoral shaft with the body axis, leaving the hip joint in its preexisting position. A cannulated hip blade plate was used for fixation. Hip motions were recorded preoperatively, at implant removal, and at the time of the latest follow-up, as was ambulatory ability.
Results: Eighty-one hips had a mean flexion contracture of 52° preoperatively, which improved by 35°; 84 hips with a mean preoperative adduction of −20° improved by 42°; 101 hips with a mean preoperative internal rotation of −16° improved by 35° (p < 0.0001 for all). The flexion-extension total arc of motion for the 119 hips improved by 13° (p < 0.0001). Only 11 of 94 hips that preoperatively flexed ≥90° did not do so postoperatively, but none of the patients reported seating difficulties and one of the patients had already regained hip flexion of >90° by a soft-tissue release. At a mean follow-up of 40 months, 36 patients were independently ambulatory and 20 patients were walker-dependent.
Conclusions: Children with arthrogryposis often have the potential for ambulation if the limb positioning can be optimized. The reorientational hip osteotomy corrects the hip contractures by altering the range of motion but not the total arc of motion.
Level of Evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.
Arthrogryposis multiplex congenita is a group of conditions that have in common congenital, nonprogressive contractures in ≥2 joints in multiple body areas1. The characteristic that these conditions share is fetal akinesia, the lack of intrauterine movement that leads to fibrosis and stiffness of involved joints2. There are approximately 400 diagnoses that lead to a child being born with arthrogryposis3, with an overall estimated incidence of 1 of 3,000 to 1 of 5,000 live births4-6.
Hip involvement in arthrogryposis multiplex congenita, both congenital contractures and dislocations, occurs in 56% to 90% of patients7-12; hip contractures alone affect 18% to 51% of all patients9-13. Severe hip contractures are multiplanar, typically flexion, abduction, and external rotational, which can preclude or can greatly complicate sitting and/or ambulation8,10-15. Animal models of congenital joint fibrosis have shown collagen proliferation, fibrotic replacement of muscle, and a markedly thickened joint capsule accompanying contractures2,16-18. Severe hip contractures are recalcitrant to nonoperative treatment because of adherence of the capsule to underlying bone, short and fibrotic ligaments, and the metaplasia of periacetabular and subcutaneous soft tissues into firm, inelastic fibrotic tissue. Limited motion prenatally and postnatally and a lack of weight-bearing lead to hypoplastic epiphyses, resulting in secondary deformity and a lack of sphericity of the femoral head. These hip contractures cannot be resolved by repositioning the femoral head anatomically within the acetabulum because of severe postoperative scarring and/or joint incongruence.
Our philosophy, from 17 years of experience in treating children with arthrogryposis multiplex congenita, is that if a hip deformity can be improved, it should be improved, either by open reduction of the congenitally dislocated hip before 3 years of age or by contracture treatment. The proximal femoral reorientational osteotomy preserves preoperative hip motion but positions the lower limb in the best functional position for that patient for sitting and ambulation. This retrospective study compares preoperative and postoperative hip ranges and total arcs of motion and evaluates the ambulatory abilities of patients who underwent reorientational osteotomies. The effects of the procedure on specific hip contractures were assessed. We reviewed the children’s success achieving ambulatory goals, which other procedures that they underwent to achieve that status, and what predictive factors were related to the hip contractures.
Materials and Methods
For this institutional review board-approved retrospective study, the medical records of all patients with arthrogryposis multiplex congenita who underwent a surgical procedure for hip contractures from December 2008 through December 2013 were reviewed. Hips that were congenitally dislocated and underwent open reduction were excluded (33 hips in 22 patients). Indications for the procedure were multiaxial hip contractures that precluded stable seating and/or efficient crawling or walking, regardless of perceived sitting or ambulatory potential. Contraindications were uniplanar contractures, for which other procedures were performed instead. For isolated severe hip flexion or flexion-abduction contractures, a percutaneous anterior hip release was performed, as described in the Surgical Techniques section. For hips only lacking flexion, a posterior hip release was performed, as described in the Surgical Techniques section. Hips with a severely narrow flexion-extension total arc of motion, not suitably aligned for sitting, and recalcitrant to soft-tissue releases were eligible for the reorientational osteotomy, specifically to improve seating. No muscle strength assessment was made as most of the patients were too young to cooperate with formal strength testing. Ninety-nine consecutive patients (175 hips) with contractures were identified. Thirty-one patients (51 hips) were excluded as their flexion or flexion-abduction contractures were satisfactorily treated with percutaneous anterior hip releases. The remaining 68 consecutive patients all underwent reorientational osteotomies for severe multiaxial hip contractures, forming the basis for this study. Three patients (5 hips) did not return for a 2-year follow-up. Of the remaining 65 patients (119 hips), 28 were female and 37 were male, and 54 patients had bilateral hip contractures. With regard to the 11 patients who underwent a unilateral reorientational osteotomy, the contralateral hip had a simultaneous open reduction for congenital dislocation in 6 patients, had a mild hip contracture requiring only soft-tissue (percutaneous anterior hip) releases in 3 patients, and did not have a clinically important abnormality in 2 patients. The mean patient age at the time of the procedure was 48 months (range, 13 months to 12 years and 10 months). The mean patient age at the time of implant removal was 5 years and 6 months (range, 25 months to 14 years and 11 months), for a mean total implant time (and standard deviation) of 18 ± 8 months. The minimum follow-up time after the index procedure was 24 months, and the mean follow-up time was 40 months.
Hip motion ranges were recorded under general anesthesia, just prior to the index procedure and at implant removal. The latest follow-up measurements were performed in the clinic. Whenever range-of-motion data were collected, special care was taken to maintain the pelvis in a neutral position, thereby preventing errors in measurement. All measurements were made with a plastic goniometer and were rounded to the nearest 5° increment. Table I lists the hip motion ranges measured. For notation purposes, the inability to bring the hip to neutral was indicated as a negative value (e.g., a 30° flexion contracture would be −30° of extension). The change in hip motion range due to the surgical procedure was evaluated for the entire group. Smaller groups with specific hip contractures were also analyzed. These specific contracture groups are also listed in Table I, along with the total arcs of motion measured.
Ambulation ability was described as non-ambulatory, walker-dependent ambulation, or a community ambulator without walker support.
Other pertinent lower-extremity procedures performed by us were also recorded; prior procedures that had been performed elsewhere were not captured. Clubfoot deformities were treated with serial Ponseti casting with or without Achilles tenotomies; congenital vertical tali were corrected by the Dobbs method19,20. Other hip procedures performed concurrently with the reorientational osteotomy, and/or at subsequent surgical settings, included percutaneous anterior and posterior hip releases. When the anterior and/or posterior hip releases were performed concurrently with the reorientational osteotomy, the goal was to increase the flexion-extension total arc of motion, limiting the extent of correction needed through the osteotomy. After 4 years of age, knee flexion contractures were addressed by posterior releases through medial and lateral incisions, followed by either anterior distal femoral hemiepiphysiodeses or knee-spanning Ilizarov fixators for gradual joint distraction21,22.
Statistical analysis was performed with SAS software (version 9.4; SAS Institute). Changes in hip motion ranges at the various time points were analyzed using paired t tests.
Patients were first evaluated for the characteristic arthrogryposis multiplex congenita soft-tissue hip flexion contractures, where a percutaneous anterior hip release (described below) might decrease the flexion component of the contracture initially, allowing for a less extreme osseous correction. Then a standard lateral approach to the proximal part of the femur was made. A marking wire was laid on the skin between the two anterior superior iliac spines, denoting the pelvic axis when viewing the hip under fluoroscopy. The hip and lower limb were positioned in their natural resting position, usually some combination of flexion, abduction, and external rotation. A guidewire was advanced into the proximal part of the femur, parallel to the pelvic axis wire both visually and on fluoroscopy (Figs. 1 and 2).
A seating chisel for the cannulated hip blade plate set (Smith & Nephew) was impacted over the guidewire, orienting the chisel blade parallel to the transverse plane of the pelvis, regardless of the position of the femur (Fig. 3). In the extreme case of a 90° hip flexion contracture, the seating chisel blade was actually positioned perpendicular rather than parallel to the vastus ridge. Often, the seating chisel had bicortical fixation, breaching the posteromedial cortex of the femoral neck.
Two osteotomy cuts were made, one proximal cut, 5 to 8 mm inferior and parallel to the flat surface of the seating chisel, and one distal cut, perpendicular to the shaft of the femur, removing a wedge of bone (Fig. 4). When joining the two osteotomy surfaces together and derotating the distal fragment, the lower extremity is positioned in neutral with respect to both internal and external rotation and abduction-adduction (Fig. 5). A 90° angled cannulated blade plate with offset was impacted over the guidewire and was then secured to the femoral shaft with bone screws.
A Petrie cast (bilateral long leg casts connected by a bar) was applied for both unilateral and bilateral procedures; a bilateral hip spica cast was applied if a contralateral open hip reduction was performed. The patients were allowed to increase hip motion by sitting upright and lying supine and even lying prone for extra extension stretching. Standing in the cast was allowed at 3 weeks if radiographs showed good healing progress. At 6 weeks, the cast was removed, and a rigorous standing and ambulation program was begun with knee-ankle-foot orthoses.
Percutaneous Anterior Hip Release
Patients with soft-tissue hip flexion contractures often benefitted from this minimally invasive procedure, either as an independent procedure, or in conjunction with the proximal femoral reorientational osteotomy. The cordlike conjoined tendon of the sartorius and tensor fasciae latae was readily palpable as emanating from the anterior superior iliac spine when the contralateral hip was maximally flexed and the ipsilateral hip was extended. The cord was transected through a percutaneous incision just below the anterior superior iliac spine, followed by incising the fascia lata to the level of the greater trochanter; the rectus femoris tendon and the fascia overlying the iliopsoas were incised as needed if they restricted extension in more severe contractures.
Posterior Hip Release
For patients with a limitation of hip flexion, this procedure was performed through the same lateral incision as the proximal femoral reorientational osteotomy, either at the time of osteotomy or at plate removal. The gluteus maximus insertion was elevated subperiosteally off the posterior part of the femur as a sleeve, along with any hip adductor insertions that limited hip flexion. Their deep fascia was incised transversely to perform an intrasubstance lengthening.
Nearly 70% of the patients presented with the characteristic combination of hip flexion, abduction, and external rotation contractures (Figs. 6 and 7). Eighty-one hips had flexion contractures of >20°, with a mean contracture (and standard deviation) of −52° ± 19° that improved by 35° to −17° ± 15° at the time of implant removal (p < 0.0001) (Tables I and II). The mean preoperative hip adduction angle for 84 hips with <15° adduction was −20° ± 24° and it improved by 42° postoperatively to 22° ± 13° (p < 0.0001). One hundred and one hips had <30° of internal rotation in flexion, which improved by 36°, from −11° ± 22° to 24° ± 20° (p < 0.0001). Similarly, 101 hips had <30° of internal rotation in extension that improved by 35°, from −16° ± 21° to 18° ± 21° (p < 0.0001). At the time of the latest follow-up, the improvements in hip extension, adduction, and internal rotation were maintained.
Hips that lacked either frog-leg abduction (n = 21) or regular abduction (n = 10) showed no change from preoperative measurements to either postoperative or follow-up measurements. Of the 119 hips, 25 had <90° flexion, which improved by 12°, from 63° ± 17° to 75° ± 21° (p = 0.018), but in many cases, this change was not enough to be clinically important for improved seating. The other 94 hips flexed ≥90° preoperatively, of which 11 hips in 9 patients did not flex to 90° postoperatively (mean, 79° [range, 65° to 85°]), a mean loss of 32° per hip; at the time of the latest follow-up, none of those patients had reported seating difficulties, and all but 1 were walking independently. The oldest of that group was 11 years old at the time of the surgical procedure and had the greatest loss of extension, from a mean of 120° to 58°. She underwent posterior hip releases 18 months later, at implant removal, and at 18 months thereafter, her hip flexions were 85° and 95°. At the time of the latest follow-up, she reported no recurrence of her preoperative back pain with ambulation, which had been caused preoperatively by her wide-based gait and 30° hip flexion contractures. For the whole group, the flexion-extension total arc of motion increased by 13° ± 23° (p < 0.0001) (Table II).
Preoperatively, 6 patients were ambulatory at a mean age of 7 years and 1 month. At the time of the latest follow-up, 36 of the 65 patients were independent ambulators, most with braces, and another 20 were walker-dependent. Nine patients were non-ambulatory, of whom 2 had the procedure performed specifically to improve seating.
Other hip procedures performed included 56 percutaneous anterior hip releases in 20 patients to improve hip extension, 26 posterior hip releases in 13 patients to improve hip flexion, and 7 adductor tenotomies in 4 patients (Table III). One hip underwent a greater trochanteric resection at the time of the osteotomy to increase hip motion. Table IV lists the other lower-extremity procedures.
Ten complications were observed. There were 4 fractures; 1 occurred in the femoral shaft 3 months after the osteotomy and was treated in a hip spica cast, and 3 occurred within 2 months of plate removal, treated with operative fixation. All healed uneventfully, without apparent loss of motion. There were 3 surgical site infections, 1 in a patient with previous methicillin-resistant Staphylococcus aureus-infected spine instrumentation, and all resolved with implant removal after the osteotomy healed. One patient had partial pullout of the plate fixation screws, seen 3 weeks postoperatively. As the patient was comfortable, and the osteotomy position was healing and maintained by the Petrie cast, the osteotomies were observed until full healing, with resultant satisfactory hip motion. In another patient, a guidewire sheared off during the osteotomy, requiring a hip arthrotomy for removal, without sequelae. One patient had recurrence of hip contracture and required a repeat osteotomy 3 years following the initial osteotomy.
Children with arthrogryposis multiplex congenita can learn to walk if they have severe foot deformities or knee contractures, but severe hip contractures limit lower-extremity positioning, greatly affecting upright and sitting positions8,14,15. In a study of 36 patients severely affected with arthrogryposis, Hoffer et al. concluded that hip flexion contractures needed to be <30° for functional ambulation15. We developed the proximal femoral reorientational osteotomy to take advantage of the preexisting hip motion sphere by translating it to a more functional domain of motion. The unexpected finding of the increased flexion-extension total arc of motion was probably exercise-induced, once the children began using the hips more functionally and routinely. Furthermore, the often dramatic remodeling of the proximal part of the femur may help to explain why hip function improves over time (Figs. 8 and 9).
The limitations of this study were its retrospective nature, in which a full complement of hip ranges was not always measured. A number of hips had implant removal around or after 24 months, and later follow-up measurements had not yet been recorded, lessening the number of follow-up measurements. It is important to note that each patient was compared with himself or herself in analyzing the data, so a patient’s data for a specific motion could not be included unless both before and after measurements had been recorded. We also acknowledge the potential for detection bias in the measurement of hip total arc of motion both preoperatively and postoperatively. This study is a preliminary report of early findings, and we expect more patients to become ambulatory over time, but realize that there may be degradation of results as the children mature. Arthrogryposis multiplex congenita is a heterogeneous grouping of conditions. Although analyzing the outcomes of the procedure by the specific condition that caused each child’s congenital contractures could be revealing, current diagnosis methods (blood tests and muscle biopsies) are difficult to access, expensive, and uncomfortable, and often the findings are inconclusive or, when they are positive, are unlikely to lead to any changes in the treatment course. Our procedure treats the hip contractures that are the result of the final common pathway of all of these arthrogryposis multiplex congenita conditions, fetal akinesia, and is unlikely to be affected by the specific cause of the fetal akinesia.
To our knowledge, the literature addressing treatment of severe arthrogrypotic hip contractures is sparse. Some authors have suggested that the contractures can readily be treated conservatively with physical therapy, traction, and casting, but provide few supportive clinical data9,13,23-25. Some surgical procedures for arthrogrypotic hip contractures have been recommended. Yau et al. treated contractures by soft-tissue releases (either iliopsoas, adductor, rectus femoris, iliotibial band, or posterior release) or by femoral extension osteotomies. Data on hip motion improvement were not given, but those authors noted that earlier intervention led to better long-term motion26. Other authors have also mentioned soft-tissue releases, but without describing the initial contracture severity, the structures released, and/or the outcomes of the procedures7,10,27,28. Intertrochanteric osteotomies have been mentioned, but specific details were not provided as to contracture severity or outcomes11,15. The English abstract of a Polish-language article described intertrochanteric femoral osteotomies improving hip flexion contractures to a mean of 53° in 12 arthrogrypotic hip contractures. The mobility status was improved for most of the patients, and recurrences were associated with more severe preoperative contractures and younger age at the time of the surgical procedure29.
The majority of patients with severe arthrogrypotic hip contractures have the potential to become ambulatory if the contractures can be sufficiently corrected. Unfortunately, most of our patients and their families were told by their previous orthopaedists that ambulation was unlikely and their treatment plan was limited to comfortable seating. In our study, 6 patients were ambulatory before the reorientational osteotomy. It is possible that some of the other younger patients would have become ambulatory by themselves anyway, had they not undergone the procedure. However, given the presenting hip positioning, even if ambulatory status had been achieved, the stance would have been wide and the body would have been hunched over, making the gait very energy-inefficient. In this preliminary report, we documented that 55% of patients were independent ambulators, and another 31% were walking with a walker. Most of the walker-dependent children had follow-up of <3 years, and many were expected to become independent over time as well (Fig. 8). Our practice is to begin aggressive standing and walking programs 1 or 2 months after the removal of the Petrie casts. Many children with arthrogryposis multiplex congenita have poor arm function and are unable to protect themselves from falls in typical ways. For these children, we institute fall therapy, teaching them to reflexively fall onto their sides and shoulders, to protect their head and arms from injury or fracture. The contractures that most affect ambulation are the flexion and abduction contractures. The procedure, often with the addition of a simple soft-tissue release, was able to decrease flexion contractures from a mean of 52° to <20°, and adduction from 20° short of neutral to 20° past neutral. In so doing, the knee can be positioned directly under the hip. As rotation can be acutely corrected by the procedure, the knee was positioned in the same sagittal plane as the hip. The range of motion was altered but not the total arc of motion. This meaningfully increases the patients’ chances of achieving their ambulatory potential. Although 11 hips lost flexion to ≥90°, none of the patients reported seating difficulties. Our plan is to address those hips with posterior hip releases as they near maturity.
Our current treatment algorithm for children with arthrogrypotic deformities is to address the foot deformities (clubfoot or congenital vertical talus) as a baby; to reduce the dislocated hips or correct the hip contractures when the patient is between 12 and 18 months of age; and to address the knee flexion contractures after the patient is 4 years of age.
Note: The authors acknowledge the important contribution of John Gaughan, MS, PhD, MBA, of Temple University, Philadelphia, Pennsylvania, in providing the statistical support needed for this research, and Carolyn Hendrix of the Shriners Hospital for Children, Philadelphia, Pennsylvania, for her proofreading and editorial assistance.
Investigation performed at the Shriners Hospital for Children, Philadelphia, Pennsylvania
Disclosure: There were no funding sources for this research. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of the article.
- Copyright © 2017 by The Journal of Bone and Joint Surgery, Incorporated