Proximally displaced medial humeral epicondyle fracture in pediatric patients: A clinical case series

: 27 Background: Medial epicondyle fracture comprises a considerable proportion of pediatric elbow 28 injury. The fracture fragment is typically pulled distally by the muscle and the ligament. This 29 study aims to suggest proper recognition of a subset of the fracture that differs from its usual 30 presentation. 31 Methods: A retrospective case study was conducted during 2011-2016. Of those cases, a subset 32 was identified as proximally displaced (atypical) ones. Distinctive radiologic images, as well as 33 the injury causes, demographic data, clinical signs, treatment ways, and final follow-ups regarding 34 these atypical ones, were presented and discussed. The fracture mechanism was carefully inferred 35 from former theories and the operative findings, and a tentative management strategy was 36 suggested. 37 Results: Seven out of 112 cases were distinguished as the atypical, which represents 6.25% of the 38 whole sample. Injury causes were all direct or combined direct/indirect force injuries instead of 39 indirect force mostly seen in the typical. Five were operated while two nonoperatively treated. 40 Operated cases revealed stripping of medial epicondyle from its surrounding periosteum/muscle 41 origin or even cartilage. The fracture fragment was either pulled by proximal periosteum or even 42 proximally dissociated. The outcomes of those atypical were mostly acceptable despite some 43 minor defects. 44 Conclusion: The proximally displaced cases do constitute a portion of medial humeral epicondyle 45 fracture in children. As well as its skeletal manifestation, awareness of its injury mechanism and 46 soft tissue damage is required. Precise restoration of its anatomical structure might be vital for its 47 treatment. Further scientific work is needed regarding its mechanism and management. 48 Level of evidence: Level 4. Atypical

Four were single fracture, among them one with hemophilia; the other three were respectively 85 combined with ulnar fracture, radial head subluxation, and multiple elbow lesion. All of the seven 86 presented certain degree nerve irritation while no elbow dislocation or fragment incarceration was 87 discovered. Valgus stress tests were performed and positive on all the operated five cases. Cases 88 information was given in Table 1. 89 Though featured uneventfully in demographic data, the atypical ones did show some differences. For the operative findings, both the typical and atypical revealed a tear of periosteum in the medial 106 part of the distal humeral. In the typical, the epicondyle fragment together with the attached 107 periosteum was pulled distally. While for the seven atypical, the fragment retracted proximally 108 with the attached proximal periosteum, yet some detached periosteum was pulled distally by the 109 flexor-pronator mass [ Fig.2a-f]. On some extreme occasion, the epicondyle ossification nucleus 110 was stripped bare from the surrounding cartilage and dissociated proximally, and the remaining 111 structure was pulled distally (Case #1). 112 The follow-up duration ranged from two to three years. The last follow-up results were excellent 113 in four, good in one and fair in two cases (according to the Mayo Elbow Performance Score [8]).
Elbow was stable and pain-free for all seven. Two presented a fibrous union with no symptom. No 115 cubitus valgus or ulnar nerve palsy resided. 116

Illustrative case: 117
A seven-year-old girl fell from parallel bars with her dominant elbow hit the ground directly. Arm 118 plywood was put on immediately in a local clinic. The next day, she presented to our department, 119 with pain and swelling in the medial elbow, restriction of the elbow joint and slightly numb in the 120 little finger. Three-dimensional computed tomography (3D-CT) showed medial humeral 121 epicondyle fracture, with fragment proximally displaced and rotated and a piece of epicondyle 122 attached to fracture bed [ Fig.1a]. Open reduction, fixation with two K-wires was carried out two 123 days after injury. Intraoperative findings showed swelling of the ulnar nerve, detached 124 flexor-pronator origin which was reattached to the proximal humeral periosteum but the UCL was 125 intact [ Fig.2b,d,f]. Longarm splint was applied, pins and splint were removed four weeks after 126 surgery while active functional recovery started. Bony union achieved seven weeks 127 postoperatively. At the final follow-up two years postoperatively, though X-ray revealed slightly 128 heterotopic ossification near epicondyle, the patient regained a stable and pain-free elbow, with 129 merely seven-degree extension loss [ Fig.3a,b]. The final results were rated excellent. 130

Discussion: 131
Medial humeral epicondyle fracture accounts for 11-20% of pediatric elbow fractures [9], A 132 relatively small amount to draw enough attention from pediatric orthopedists. However, over the 133 past few decades, as diagnostic technology advanced and attitudes toward pediatric injury 134 improved, more and more concern has been given on it. Its classification has since evolved. 135 Multiple classification systems existed in English literature. In 1950, Smith [10] described five 136 types of medial epicondyle injuries based on the amount of fracture displacement and entrapment 137 of the fragment in the elbow joint. In 1982, Papavasiliou [11] modified the classification into four 138 types, making it the most succinct and widely used criterion among the orthopedic practitioners 139 since then. The Wilkins's [7] classification system, in which medial epicondyle injuries were 140 divided into acute injury patterns and chronic tension stress patterns, was a more comprehensive 141 one. The Papavasiliou four types and those with fracture lines through the epicondyle were 142 included in the acute injuries in this classification. However, all of the above categorizations 143 postulated that the separated fracture fragment was displaced towards distal and medial direction for the strong pull from the flexor-pronator mass. The proximally displaced ones were not even 145 included. To date, only one study on a single case of this type in the English literature was 146 identified, in which the authors presented a case that our Case #1 resembled [12]. Regretfully, the 147 authors did not infer the mechanism in their case study and a single case revealed little of the 148 commonality of this subset. Thus, the mechanism and the treatment algorism for this type need to 149 be further clarified. 150 Currently, three existing theories regarding the mechanism for medial epicondyle fracture have 151 been described: direct trauma [8], an avulsion mechanism involving an indirect muscular pull 152 [9,13], and a combined association with elbow dislocation [3,14]. We managed to apply these 153 theories to the cases in our series and described them as follows. As to the typical ones, direct 154 force or indirect muscular traction cause avulsion of the epicondyle and periosteum surround it. 155 Subsequently, the periosteum, cartilage, and the UCL sticking to the epicondyle as a whole, and 156 altogether they are pulled distally by the musculus flexor [ Fig.2e]. While for the atypical ones, 157 from intraoperative findings of the subset, we speculated that though the avulsion part is similar, 158 the epicondyle is dissociated from the distal periosteum or even the epicondyle ossification is 159 stripped off from the cartilage surrounding it, and due to proximally-directed force, displaced 160 proximally with or without rest attached periosteum [ Fig.2f]. 161 Management strategy somehow remains controversial even for typical medial humeral epicondyle 162 fracture. It is already consensual to perform cast immobilization on those with low-energy 163 mechanisms, stable elbows, and minimal displacement and to operate on those with 164 traumatic/high-energy injury, significant displacement, elbow instability/dislocation, incarcerated 165 fracture fragment, open fractures and ulnar neuropathy [15]. However, for those with moderate 166 displacement, debates are still going on, mostly around the exact displacement amount to justify 167 the surgical intervention. 168 Traditional treatment algorism for moderate displaced ones suggests cast immobilization when the 169 displacement is less than 5mm, and operation when the displacement is more than 5mm [1,5]. A 170 research by Edmonds [16] et al did show the relationship between displacement amount and 171 outcomes like wrist flexion strength (approximate 2% decrease for every 1mm of anterior 172 displacement due to muscle shortening). Yet the deemed displacement was usually measured on 173 AP or lateral plain X-rays, which was with great deviation and did not represent the true displacement of the fracture, making it hard to justify the treatment strategy [17]. Not until 175 recently have the researchers introduced more sophisticated ways like 45-degree oblique [5] and 176 distal humerus axial view [6] to improve plain X-ray measurement. Yet, these established 177 measurement ways were typically for the distal displacement cases, while for more complexed 178 occasions in the proximal displaced, simple and reliable ways to measure were still in need, which 179 is why this study resorted to more sophisticated ways like 3D-CT scans to get a precise 180 measurement. In a recent review, Beck [18] and her colleagues acknowledged that besides the 181 displacement measurement, the decision for surgery should be made on specific factors such as 182 arm dominance and sport type the patients were to take. 183 Taking into account these specific factors, though, non-operative treatment still represents the 184 mainstream and has been historically adopted [19]. Josefsson [20] et al. carried out a long-term 185 follow-up retrospective study of 56 non-operatively treated fractures, which showed good results 186 with minimal presence of ulnar nerve symptoms. To date, there are growing numbers of 187 comparative studies supporting similar outcomes between operative and non-operative treatment 188 [21,22]. Also, fracture displacement may improve over the conservative treatment period, a study 189 result presented by Lim [23] and his colleagues, in which an average improvement of 1.55mm 190 from 5.34mm at initial radiographs obtained. Yet high nonunion rates with up to 90% (17/19)

in 191
Farsetti [4] et al.'s and 50% (28/56) in Josefsson [18] et al.'s cohorts respectively occur in 192 nonoperatively treated patients, though always asymptomatic. Yet, symptoms relating to nonunion 193 like pain, elbow instability, and wrist flexion weakness do exist in some conservatively treated 194 patients, especially in those adolescent athletes and the deciding factors and true incidence are still 195 unknown, which explains the favor of some orthopedists for operative intervention. They believed 196 that anatomical reduction and proper fixation allows earlier return to sports and recovery to a 197 preinjury level of function [24]. However, general anesthesia, surgery-induced trauma, and extra 198 medical expenses raised further concern about the indication for surgery. 199 Although precise displacement amount was able to acquire with 3D morphological images in 200 proximally displaced cases, rigorous treatment algorithm is still lacking, as in the typical ones. 201 Considering the complex mechanism, the diversified injury extent, and the multiple combined 202 elbow injuries, the algorithm should therefore be even more complex. Deemed indications for 203 surgery roughly include fragment displacement, combined injuries, patient needs, orthopedist 204 experience. The history of direct injury or high energy injury and the indication of elbow 205 instability like valgus test positive should be seriously taken into account for treatment decisions. 206 The authors suggest open reduction when the displacement is apparent, soft tissue damage is 207 assumed severe, combined elbow injuries are present and complex, or the patient is an adolescent 208 athlete, requiring sooner return to activity and better performance on sports. It might be suggested 209 to be very cautious with non-operative treatment for the atypical, as our study revealed the 210 disruption of the anatomical structure tended to be more severe, which justified proper fixation. 211 For the operation technique, the preliminary experience gained was to try best to restore normal 212 anatomical (skeletal and ligamental) structure around the epicondyle in children. Whenever the 213 surgery is decided, keeping the separated epicondyle and periosteum/cartilage fit together closely, 214 firmly and durably is highly recommended, no matter what kind of hardware is used. 215 This case series provided a new subset of medial humeral epicondyle fracture which few 216 predecessors had mentioned in the English literature. Though a small proportion of all pediatric 217 medial humeral epicondyle fracture in this study, the proximally displaced one may update the 218 current view on this topic. Dissociation between the epicondyle and distal periosteum/cartilage 219 might be the vital pathological change. More complex and higher energy injury lead to more 220 severe soft tissue damage and more often combined elbow injures, compared with those typical 221 ones. Not only the fracture fragment but also the detached soft tissue is recommended to be 222 anatomically reduced and fixated. The separation between epicondyle and periosteum/cartilage 223 might stimulate subperiosteal ossification or entochondrostosis, which probably compromises the 224 outcome. Although this work provided preliminary discussion and tentative treatment strategy, it 225 did offer proper recognition of this unique subset of pediatric medial humeral epicondyle fracture 226 for pediatric orthopedists. Considering its distinct appearance, mechanism, and intriguing 227 treatment strategy, we cautiously recommend to add it into an even more extensive classification 228 system to facilitate future clinical practice. 229

Declarations: 233
Ethics approval and consent to participate: