Introduction

Pediatric supracondylar fractures are the second most common fracture in all pediatric fractures [1, 2]. At childhood, they are the most common fractures in elbow region [3]. These fractures are most often seen between ages of four and seven [1, 2]. Accounting for 97–99 % of these fractures is extension type [4]. Reduction is difficult due to anatomy of the region and swelling.

Stable fractures require only simple external immobilization, whereas unstable displaced fractures require surgical intervention [4–6]. The aim of treatment of these fractures is to achieve anatomic reduction and stable fixation, and good functional and cosmetic results. Surgical treatment consists of open or closed reduction with K-wire fixation. Treatment of these fractures with closed reduction and percutaneous fixation with cross K-wires is widely accepted and has successful results [6–10].

In this study, we compared operation duration, radiological, and functional results of the open reduction with either posterior or lateral approach and closed reduction with joystick method in unsuccessful closed reduction of displaced (Gartland type III) supracondylar humeral fractures.

Materials and methods

Between February 2010 and August 2011, 239 pediatric patients with acute supracondylar humeral fractures were admitted in our emergency department. Of these fractures, 52 (21.7 %) were classified as Gartland type I, 33 (13.8 %) were classified as type II, and 154 (64.5 %) were classified as type III. Patients who had <6 months follow-up period, patients undergoing initial treatment at another center, patients treated with primary open reduction, revision cases, open fractures, patients with neurovascular injuries in admission, divergent or parallel K-wire fixed cases were excluded from study. Thirty-seven (40.6 %) of 91 patients were not obtained satisfactory reduction with classic closed reduction attempts for three times in operating room. These 37 patients had Gartland type III fractures and were included study. Patients were treated with three different surgical methods and were divided into three groups. All fractures were fixed with cross K-wire. Surgeon’s preferences determined methods of operation. The study was performed retrospectively by reviewing the hospital records.

Group I have 13 patients who had joystick and lateral K-wire-assisted closed reduction, group II have 12 patients who had open reduction by lateral approach, and group III have 12 patients who had open reduction by posterior approach. The demographic distributions of patients were given in Table 1. In all three groups were similar respect to gender, side, and follow-up period (Table 1). Group II was younger than the other groups (p < 0.001), but group I and III were similar respect to age (p = 0.422).

Table 1 Demographics of patients
Full size table

Surgical technique

In group I, after unsuccessful attempts of closed reduction with fluoroscopic control under general anesthesia, a 5 mm incision was made just distal to the deltoid muscle attachment. In accordance with the patient’s age, 2.5 or 3 mm K-wire was placed as double cortex, perpendicular to the long axis of the humerus. After classic closed reduction manipulation, distal fragment of fracture was usually internal rotation. Proximal fragment using joystick K-wire was fitted to distal fragment. When accurate reduction in AP view and partial reduction in lateral view achieved, a 1.2 mm soft K-wire was placed from lateral as double cortex. While the distal assisted K-wire was maintaining to reduction in coronal plane, accurate reduction was achieved using joystick K-wire in lateral view. Then, second 1.6 or 2 mm K-wire was placed from lateral to the far cortex. Another K-wire was placed from medial with 1 cm medial mini-incision in elbow flexion more than 90°, while reduction was maintaining with lateral K-wires and joystick K-wire. Joystick K-wire was removed after fixation, the other K-wires bent outside to skin. Operative stages are shown in Fig. 1.

Fig. 1
figure 1

Respectively stages of fluoroscopic views of joystick and lateral K-wire assisted closed reduction and fixation. Placing joystick K-wire (a) and assisted K-wire (b, c). Bending of assisted K-wire after full reduction in lateral view (d). Placing of lateral and medial K-wires after accurate reduction in both views (e–h)

Full size image

In group II and III, after attempts of closed reduction with classical method for three times, open reduction with lateral approach lateral approach between ekstensör carpi radialis, triceps and brachialis, and cross K-wire fixation done for group II. Open reduction with posterior triceps split approach and cross K-wire fixation done for group III.

In all cases, the quality of the reduction was controlled by fluoroscopy. Long arm splint was applied with 90° elbow flexion and neutral forearm position. All patients monitored 1 day in hospital concerning edema, circulation problems and neurological deficit. First, second, and fourth weeks of follow-up, clinical and radiological examinations were performed. Active motion was allowed by removing splint and K-wires in 4th week. AP and lateral radiographs of both elbows were taken at last control.

In last control, bilateral Baumann angles and lateral humerocapitellar angles were measured. Carrying angles and elbow range of motion were measured by goniometer. These angles and operation times compared between the groups. Cosmetic and functional assessment of patients in all groups was evaluated by criteria of Flynn et al. [11].

Statistical analysis

IBM SPSS Statistic Version 20.0 software was used for statistical analysis. In evaluation of the data descriptive statistical methods (mean, standard deviation) and in analyzing of independent groups of quantitative data showing no normal distribution, Mann–Whitney U test and Kruskal–Wallis test were used. For crude analysis of independent groups of qualitative data, Fisher’s exact test was used. About, 95 % confidence interval significance at p < 0.05 was accepted.

Results

Difference between Baumann angles, lateral humerocapitellar angles and carrying angles of fractured and uninjured sides are listed in Table 2. Between group I and group II and between group I and group III, Baumann angles differences are p = 0.473, p = 0.639, respectively, lateral humerocapitellar angles differences are p = 0.175, p = 0.978, respectively, carrying angles differences are p = 0.736, p = 0.416, respectively. There was no significant difference.

Table 2 Difference between angles of fractured and uninjured sides
Full size table

Operation time between three groups were no significant difference (p = 0,092) but between group I and II and between group I and III have meaningfully differences (respectively, p = 0.0076, p = 0.0051).

Table 3 describes the cases according to criteria of Flynn et al. [11]. Excellent, good and fair results considered satisfactory. Functional scoring in our cases showed that in group I and group II all patients have satisfactory results; however, in group III, three patients (25 %) had poor results. In the perspective of cosmetic results, all three groups have satisfactory results. There was no neurovascular injury in our cases. Only one patient in group II had pin infection, one patient in group III had superficial wound infection. Both patients had fully recovered after antibiotic treatment and removal of K-wires.

Table 3 Distribution of cases according to criteria of Flynn et al. [11]
Full size table

Discussion

Supracondylar fractures are common fractures accounting for 17.9 % of all pediatric fractures [2]. They are the most commonly surgically treated fractures during childhood [4]. Anatomic reduction, stable fixation, providing full range of motion, and good cosmetic appearance are surgical goals of these fractures. As a fast and easy method closed reduction and percutaneous pining is accepted by many authors [4–13]. However, in soft tissue interposition, severely displaced fractures, very edematous extremities closed reduction may not be achieved [14–16].

Usually, open reduction is the chosen method for open fractures, accompanying neurovascular injuries and irreducible fractures with closed methods [4, 7, 17]. Posterior triceps split approach have the advantage of visualizing both colons and easier anatomical reduction [9]. But it has the disadvantage of muscle fibrosis and range of motion deficit because of muscle weakness [4, 9]. Lateral and medial approaches visualize one colon and anterior portion of distal humerus, reduction of other colon is checked by palpation when using these approaches [8, 18].

Advantages of closed reduction and percutaneous fixation are shorter operation time, less infection risk and no opening fracture site; however, disadvantages of this technique are; needs more experience, fluoroscopy usage, risk of iatrogenic nerve injuries and not able to visualizing quality of reduction directly [11, 19–21]. Also aggressive and repeated reduction attempts may lead to neuropraxia, elbow stiffness, and myositis ossificans [13, 15]. By using joystick and lateral K-wire-assisted closed reduction and cross K-wire fixation method, we think that the complications related with repeated manipulation and open approaches can be avoided. Main advantage of this technique is single reduction manipulation. While joystick K-wire supplies necessary rotation to proximal fragment, lateral K-wire keeps fracture reducted position in the coronal plane, also in lateral view accurate reduction can be achieved by using these two wires.

Kazımoglu et al. [13] and Kaewpornsawan [8] in their studies compared primary open lateral approach and closed reduction. They found no differences in Flynn cosmetic and functional results. Aktekin et al. [9] found better cosmetic and functional scores with closed reduction group comparing posterior approach group. Ozkoc et al. [10] mentioned better functional results with closed reduction than primary open posterior approach. They stated no differences in cosmetic scores. All groups in our study had the similar results for Flynn cosmetic criteria. There were not any poor result in group I and group II according to functional results. In group III, three patients had more than 15° range of motion deficit and had poor result. In one of these patients, there was triceps fibrosis and the other two patients had elbow stiffness.

Method of K-wire fixation is controversial. Biomechanical studies imply the most stable fixation as medial and lateral cross K-wire fixation [22]. Some authors state the importance of additional medial K-wire after fixation with two lateral K-wire [7, 12, 23]. But there is the risk of ulnar nerve injury with additional medial K-wire fixation. Ulnar nerve injuries are the most common iatrogenic nerve injuries in supracondylar fractures [7, 20]. To reduce the risk of ulnar nerve injuries we used a mini-incision on medial side and bend the elbow <90° during medial pin placement as stated in previous studies [6, 7, 12, 23]. This technique allows placement of the medial pin safely and reduction of fractures does not deteriorate during medial pin placement.

During closed reduction it is difficult to maintain reduction before fixation especially in severely displaced and swollen elbows. Besides these situations this new technique may be useful patients whom can not be treated with open reduction due to soft tissue problems (Fig. 2). Also during surgery rotation of elbow for fluoroscopic lateral view control a cause displacement of the fracture side. Joystick and lateral K-wire assisted closed reduction is an effective way to maintain reduction during fluoroscopic visualization. There is a risk of iatrogenic radial nerve deficit, but with careful application and proper placement (distal to the lateral deltoid insertion) this risk can be reduced.

Fig. 2
figure 2

Preoperative photograph of the supracondylar humeral fracture with severe edema and soft tissue problem (a). Intraoperative fluoroscopic images of after joystick and assisted lateral K-wires reduction and fixation (b)

Full size image

Lee and Kim [24] described pin leverage technique after three unsuccessful closed reduction attempts in patients with Gartland type III fractures and edema. In this technique, they used a Steinmann pin applied from posterior to the fracture line for reduction and they had good results, but this technique was not useful for multifragmented fractures. Parmaksizoglu et al. [25] described joystick pin reduction technique in failed closed reduction cases. After fixation of proximal fragment below deltoid muscle insertion, distal fragment can easily manipulated to correct coronal tilt, translation and rotation. They had good results according to Flynn criteria at the ratio of 95.6 %. Novais et al. [26] in their retrospective study had used a joystick K-wire through capitellum to establish reduction in multi-directional unstable fractures. Li et al. [27] compared closed reduction with minimally invasive mosquito forceps assisted reduction. They found similar functional results. In none of cases treated with joystick technique, open reduction was not applied as similar with the literature [25–27]. In addition, we used joystick technique in two patients with multifragmented fractures (Fig. 3). Both of them had successful results as radiologic and functionally. Another advantage of this technique allows better cosmesis due to its no incision scar compared to open reduction.

Fig. 3
figure 3

Third week direct radiographies of multifragmanted supracondylar humerus fracture maintaining reduction and showing callus formation (a). In 1 year follow-up radiographies, fracture totally remodeled (b)

Full size image

In our study, we calculated the time of operation as starting and ending of duration of anesthesia as the study of Lee and Kim [24]. We found shorter operation time in joystick and lateral K-wire assisted closed reduction group than open reduction groups. Weaknesses of our study are short follow-up period, small number of cases and its retrospective nature. However, patients of group II were younger than group I and group III. We believe that this finding was due to lateral approach tendency of surgeons in younger children in the present study.

Conclusion

As a conclusion, when classical closed reduction fail, lateral joystick and K-wire-assisted reduction is a useful way to make and maintain the reduction. Functional and radiological results are as good as lateral and posterior open approaches. Short operation time is an advantage. This technique is also useful in multifragmented fractures. This method reduces the risk of complications due to repeated closed reduction attempts such as neuropraxia, elbow stiffness and myositis ossificans and risks of open reduction such as triceps fibrosis, joint stiffness and unsightly scarring in unsuccessful closed reduction in pediatric supracondylar humeral fractures.