Transfixing Kirshner wires for fixation of intertrochanteric valgus osteotomies in management of pediatric coxa vara

The normal pediatric femoral neck–shaft angle (FNSA) ranges from 135° to 145°. Coxa vara is a condition in which the femoral neck–shaft angle on one side is decreased compared to that on the unaffected side in unilateral cases, or the neck–shaft angle is 120° or less in bilateral cases. Coxa vara is a relatively uncommon condition with an estimated incidence of 1 in 25,000 live births. It is equally likely to occur in either gender [1].

Coxa vara is not a pathological diagnosis; it is a radiological term describing a deformed proximal femur that could result from different pathological conditions (Fig. 1). Coxa vara could follow healed rickets or other bone-softening diseases (Fig. 1a), could be developmental (Fig. 1b) or associated with congenital femoral deficiency or could be post-traumatic, but there is no universally accepted classification system for this condition [2].

Developmental coxa vara is a pathobiological condition characterized by an ossification defect typically affecting the medial portion of the femoral neck (Fig. 1c); however, in some rare cases, it can also affect the lateral portion of the femoral neck (Fig. 1d). Developmental coxa vara is associated with progressive pathomechanical changes affecting the proximal femur that—in addition to the clinical manifestations—can markedly affect the growing pediatric hip, leading to premature degenerative arthritis of the hip joint.

Once deformed, the hip joint is progressively subjected to bending stresses and shear forces, leading to a progressive decrease in the neck–shaft angle and a more vertically oriented proximal femoral physeal plate due to the bending stresses, as well as neck shortening and overgrowth of the greater trochanter [2, 3], thus triggering a vicious circle.

Indications for surgical correction are based on the clinical picture and the patient’s symptoms, the neck–shaft angle, and—more importantly—the Hilgenreiner epiphyseal angle (HEA) [3, 4]. The goal of surgical intervention is to correct varus malalignment of the neck, thereby changing the loading characteristics from shear to compression forces, which can stimulate ossification and healing of the defective femoral neck and subsequent normalization of the neck–shaft angle. Proximal femoral valgus osteotomy is considered the standard surgical treatment for coxa vara, and multiple surgical techniques with different fixation methods, including external fixators and internal fixation with pins, cerclage wiring, and a variety of plates, have been described [5].

The postoperative follow-up period ranged from 14 to 102 months, with an average duration of 33.3 ± 27.7 months. Clinically, there was a statistically significant (p < 0.001) improvement in the mean postoperative Iowa Hip Score at the last follow-up (95.06 ± 2.6) compared to the mean preoperative score (57.4 ± 3.6) (Table 1). Correction of LLD in unilateral cases, improved gait, and improved range of hip motion (especially abduction and easier ablution) were of paramount importance to parents as good indicators of successful surgery.

Two cases (both affected unilaterally) had LLD preoperatively; the LLD was corrected in one case and still present in the other. The latter case involved a revision surgery for a residual coxa vara after a previous valgus osteotomy fixed with a plate and screws 11 months before presentation. In that case, there was LLD of about 1 cm due to a subtracted bony wedge after the second osteotomy, resulting in residual shortening of the operated side; the patient tilted their pelvis slightly while walking to compensate (Fig. 6).

The osteotomies were performed in the metaphyseal, highly cancellous, proximal part of the femur characterized by high healing potential, and were completely united by an average of 11.7 ± 2.2 weeks (range: 9–15 weeks) without losing the correction achieved in any of the cases studied. The union time was not affected by the number of transfixing wires used, whether the coxa vara was unilateral or bilateral, or the size of the wedge removed. The wires could be safely removed with no fear of deformation or late secondary displacement once the callus was mature, but wires were not removed earlier than 6 months postoperatively in any of the cases.

The rationale for delaying the removal of wires until at least 6 months postoperatively—irrespective of whether there was full consolidation and callus maturation at the osteotomy site—was to allow the patients sufficient time to restore hip range and abductor power before they suffered a second operative trauma during wire removal. The wires were not removed in one patient (two cases) who refused early wire removal after complete union of the osteotomy, while the wires were removed in the other 14 cases at 6–10 months postoperatively (mean: 7.3 ± 1.4 months).

There were statistically significant improvements (p < 0.001) in the early postoperative HEA and FNSA compared to the preoperative angles. The mean HEA was corrected from 81.6 ± 8.2° (pre-op.) to 25.7 ± 2.5° (early post-op.), and the mean FNSA was improved from 89.7 ± 11.7° (pre-op.) to 137.3 ± 3.7° (early post-op.). The improvements in both the HEA and FNSA continued to be statistically significant (p < 0.001) at the last follow-up (24.3 ± 3.5° and 138.6 ± 3.5°, respectively) when compared to the preoperative values. Although progressive improvements of these angles—that could be attributed to the correction of the mechanical loading conditions and improved orientation of the growth plate—were reported in some cases with longer follow-up periods (Table 1), the overall early post-operative correction angles were statistically non-significantly different from the angles at the final follow-up. There was also progressive ossification of the neck defect in the 14 cases with developmental coxa vara (Fig. 7).

Although the procedure decreases soft-tissue stripping and completely avoids the femoral neck and physeal plate, premature physeal closure occurred unexpectedly in one case (Fig. 8a, corresponding to the case presented in Fig. 1d). This can be explained by the marked severity of the deformity in this case of developmental coxa vara, where there was an atypical defect affecting both the superior (lateral) and inferior (medial) parts of the femoral neck. Medialization of the distal fragment after osteotomy (Fig. 8b) occurred in three cases (two primary cases and one revision case), but there were no significant clinical implications of the medialization during the follow-up period.

Heterotopic bone formation in the gluteal region occurred in two cases. Bone formation started as a hazy shadow in the abductors, which could have been due to muscle irritation by the wires. This gradually matured into a well-formed bony mass that had no significant clinical implications (Fig. 8a, c) and gradually decreased in size over the subsequent follow-up. The wires were sunken in the trochanteric apophysis in one patient (two cases; Fig. 8d) who refused early wire removal after complete union of the osteotomy but suffered no related complications throughout 8.5 years of follow-up.

None of the cases had developed avascular necrosis, complications related to this method of fixation, recurrence of the hip deformity, or genu valgum deformity of the knees by the last follow-up; this was even true of the three cases in which medialization of the distal fragment occurred, as remodeling of the proximal femur in this age group is excellent, permitting progressive correction and restoration of alignment with no residual radiological or clinical effects.

Theoretically, surgical correction of the vertically oriented femoral physis to a relatively horizontal position with a Hilgenreiner epiphyseal angle (HEA) of less than 38° will normalize the biomechanical forces, converting the poorly tolerated shear forces to more physiological and desirable compressive forces. This leads to greater protection of the physis and stimulates proper ossification of the femoral neck defect, reducing the risk of recurrence of coxa vara whatever the etiology of the deformity and the age of the patient [3, 5].

Different techniques based on the same surgical principles but utilizing different implants for internal or external fixation of either intertrochanteric or subtrochanteric valgus osteotomies have been described for the surgical correction of coxa vara. The surgical technique most commonly used to achieve correction is the laterally based, intertrochanteric Y-shaped wedge osteotomy described by Pauwels [6], considered to be state of the art in the surgical management of infantile coxa vara [2]. Pauwels’s Y-shaped osteotomy is performed between the lesser and greater trochanters and perpendicular to the femoral shaft. It corrects the neck–shaft angle but the femoral shaft is positioned medially. However, it compromises the hip joint’s integrity, potentially endangering the nourishing vessels of the femoral head and thus increasing the risk of avascular necrosis [2]. Medial displacement in such patients should be strongly avoided as it exacerbates the loading of the distal femoral physis and the lateral compartment of the knee in patients with knee deformity (genu valgum), which is commonly also seen in cases of coxa vara [7].

Borden et al. [8] presented a subtrochanteric valgization osteotomy, which is a perpendicular osteotomy in the subtrochanteric region just distal to the lesser trochanter. They introduced the concept of end-to-side valgization osteotomy in cases of infantile coxa vara, which presents only a minimal risk of head necrosis due to its extra-articular approach and lack of medialization of the distal femoral shaft. It also requires only one osteotomy with easier insertion of the angled plate, making it less technically demanding than the standard intertrochanteric technique [2]. Percutaneous subtrochanteric osteotomy with external fixation has been reported as an alternative to internal fixation [9], but that study included only limited follow-up.

In the present study, proper correction of the FNSA and HEA was achieved in all of the studied cases. There was no medialization of the distal fragment (Fig. 9a) after osteotomy in 13 cases, while there was medialization of the distal fragment (Fig. 9b) after osteotomy in three cases (two primary cases and one revision case). In contrast to the classical technique (Pauwels’s Y-shaped osteotomy; Fig. 10-1), the osteotomy performed in this work was a V-shaped, intertrochanteric osteotomy that—like Pauwels’s osteotomy—corrected the neck–shaft angle appropriately, but with less medialization of the distal fragment after osteotomy.

It was noted—retrospectively—that the best proportion and coaptation between the proximal and distal bone fragments after wedge removal occurred with both the upper and lower arms of the osteotomy oblique (Fig. 10-2), while medialization occurred in cases where either the proximal or the distal arm of the osteotomy was horizontal (Fig. 10-3, -4).

Although remodeling can improve the alignment during the follow-up period (Fig. 11), rendering medialization a benign issue, Medialization of the distal fragment should be avoided to decrease the incidence of post-operative genu valgus which is a common finding with severe pre-operative coxa vara or after proximal femoral valgus osteotomies if medialization occurred. Prober alignment of the osteotomy ends with no—or minimal—medialization facilitates transfixing wiring with better bone hold and less metal prominence under the skin.

It is reported in the literature that there is limited choice when selecting an implant that allows secure fixation of such a small proximal fragment of bone while achieving stable fixation of the osteotomy without violating the proximal femoral growth plate. Angled blade plates have been the standard devices used for fixation of proximal femoral osteotomies in children [10]. Nevertheless, there are problems with this technique; for example, the plate is not a suitable size for some children and considerable technical skill is required for its insertion. Experimental studies have indicated that the tension-band wiring fixation of the proximal femur, as proposed by Pauwels and adapted by Weber, could overcome these limitations and provide simple and stable fixation [11, 12]. External fixators have been used as an alternative to internal fixation in percutaneous subtrochanteric osteotomy, but the reported results include only limited follow-up [9], and there are potential risks of pin tract infection and loosening.

In this study, the selected method of fixation was to apply transfixing K wires protected with a hip spica cast for an average duration of 6 weeks. Transfixing K wires are versatile, malleable tools for stable fixation, even with larger angles of correction that cannot usually be addressed using plates without contouring them to such an extent that the plate strength may be comprised. The application of transfixing K wires requires less dissection, thus allowing the soft tissues around the proximal femur above the level of the trochanteric flare to be largely preserved and decreasing the healing time, it also completely avoids violating the femoral neck affording marked protection to the precious growth plate.

In this study, the only disadvantage of transfixing wires was excessive protrusion of the wires from the medial cortex, leading to a risk of injury to the medially located vessels. Meticulous insertion and application of the transfixing wires with just few millimeters of protrusion from the medial cortex of the proximal femur is recommended and can be assured through the use of an image intensifier.

Using transfixing K wires without a tension band made the procedure easier and faster, with no fear of premature closure of the trochanteric apophysis. As they are hidden within the wound and protected by the cast, transfixing K wires can give superior results to external fixators, with no risk of pin tract infection or loosening. The hip spica used to protect against displacement at the osteotomy site was not unique or specific for transfixing wires, but it was used in some studies for several weeks after surgery, even when that involved fixation by plates and screws [3, 8, 13].

The results of this study were comparable to—and, to some extent, better than—the results of other studies concerning coxa vara, regardless of the etiological cause, osteotomy site, and mode of fixation. The mean HEA had been corrected from 81.7 ± 2.2° to 24.3 ± 3.5° by the last follow-up and the mean FNSA had improved from 86.9 ± 4.2° to 138.6 ± 3.5°. In a study carried out by Cordes et al. [5], which included 18 cases of coxa vara in 14 patients with various etiologies, the mean postoperative correction of the FNSA was 141° and that of HEA was 29°. Another study, performed by Desai and Johnson [14], included 20 cases of coxa vara in 12 patients. Their mean postoperative correction of the FNSA was 136° and that of the HEA was 30°. In a study by Hefny et al. [15] of the use of external fixators to treat coxa vara in 15 cases in 13 patients, the mean postoperative correction of the FNSA was 133° and the HEA was 34°.

In a more recent study by Elzohairy and Khairy [13], in which a T plate was employed in the fixation of intertrochanteric valgus osteotomy to treat 18 cases of developmental coxa vara in 18 patients, the mean postoperative correction of the FNSA was 129.9° and that of the HEA was 27.8.

The literature reports recurrence rates of between 30 and 70% [16]. Carroll et al. [3] said that correction of the HEA to <38° lowers the recurrence rate to about 5% of the corrected cases. Carroll et al. [3] concluded that the neck–shaft angle was not a reliable indicator of appropriate correction or outcome, and they reported that recurrence was mainly due to the loss of the correction angle resulting from weak fixation. This explains why there was no recurrence of deformity in any of the cases studied right up to the last follow-up, as the mean correction of the HEA was 24.3 ± 3.5°, which is quite close to the normal (anatomic) value of HEA: about 22° [3]. It also indicates that using only transfixing wires with a protective cast is an efficient and reliable method of fixation for these osteotomies, as it prevents any postoperative displacement and loss of the correction angle.

Intraoperatively, it was found that patients with unilateral coxa vara gain length after the corrective osteotomy, with the reported preoperative LLD eliminated. In bilateral cases, there was lengthening of the first limb to be operated by about 1–2 cm in comparison to the other side after the corrective osteotomy, although patients had no LLD preoperatively. Such deformity was corrected and led to overall lengthening of both sides. In other words, although a closing wedge (subtraction) osteotomy was performed, the overall length of each limb was increased after proper redirection of the femoral neck and correction of the deformity.

The relatively short mean follow-up period in such a small number of cases, the lack of alternatives (different types of osteotomy and different fixation methods) for use as a control group, and the absence of any biomechanical lab studies comparing different fixation methods under different loading conditions are limitations of this study. However, these limitations do not appear to undermine the results achieved in this study. Further research correlating different variables such as the age of the patient, the etiology and severity of their deformity, the fixation method used, the radiological healing time, the functional results, and complications is strongly recommended.

In conclusion, transfixing wires protected in a hip spica cast represent a simple, easy, and reliable fixation method for valgus osteotomies performed to correct different varieties of pediatic coxa vara (even revision cases) with different degrees of severity. It assures stable fixation and rapid healing of the osteotomy without loss of the achieved correction. It completely avoids the femoral neck affording marked protection to the growth plate.