We reviewed the clinical data of 274 patients with UCMFs and identified age at injury and DD-RH as risk factors for the development of HO. The current results partially confirmed our study hypothesis, which tested whether HO in such patients is caused by the direction and distance of the dislocated RH from the anatomic position and the interval from injury to diagnosis.

In our study, we found that the rate (14.5%) of HO in group A patients was higher than that (0%) in group B. We hypothesized that such a result may be the result of mechanical traction of the biceps brachii on the proximal radius, especially in group A patients. In fact, the biceps brachii tendon terminates at the radial tuberosity, thus traction may result in RH hypermobility, which would mechanically stimulate the surrounding soft tissues and subsequent formation of HO [23,24,25,26,27,28]. Several studies have shown that the development of HO is intrinsically related to mechanical stimulation on soft tissues [25,26,27,28]. On the contrary, especially in patients with lateral or posterior RH dislocation, the movement of the dislocated RH is more limited due to the lack of a strong muscle terminating at the lateral or posterior proximal radius [23, 24]. In addition, there was sufficient soft tissue in the anterior parts of the elbow [23, 24]. Therefore, continuous mechanical stimulation on the soft tissues surrounding the dislocated RH can be considered one of the main causes of HO formation in group A patients, as several reports have indicated that mechanical stimulation can induce soft tissue ossification [29, 30].

Our study concluded that an age at injury > 6.78 years was associated with an increased rate of HO in patients with UCMFs (Fig. 6). In addition, we found that the bone density of HO tends to increase with age at injury (Fig. 6), as the osteogenic capacity is greater in older children [31, 32].

Radiographs of two patients with UCMFs of different ages at the time of injury, including a 6-year-old girl (A and B) and an 8-year-old boy (C and D), with HO (black arrow) and lower (A and B) and higher (C and D) bone densities, respectively, than the surrounding normal bone

Our current study also identified DD-RH as a risk factor for the development of HO, especially in those with a DD-RH < 1.59 (Fig. 7). This result may be mainly due to the position of the annular ligament in relation to the dislocated RH. The annular ligament, which is the primary stabilizer of the proximal radius, tears when the RH is dislocated and may be displaced into the radiocapitellar joint space in patients with a greater DD-RH [33,34,35]. At this level, RH instability may also stimulate annular ligament ossification; in this regard, previous studies have shown that HO in patients with UCMFs may be secondary to annular ligament calcification and ossification [7,8,9]. In addition, our current results also showed that 80.6% of HO cases were circular (type I) and similar in shape to the annular ligament. Notably, our study revealed that HO did not occur in patients with PAMFs who underwent surgery < 48 h after injury, even though demographics, including age at injury and DD-RH, were similar between patients with PAMFs and those with UCMFs (Fig. 8).

Radiographs of two patients with UCMFs, with and without HO (black arrow), in one patient with a DD-RH > 1.59 (A and B) and in the other patient with a DD-RH < 1.59 (C and D)

Preoperative (A and B), 7-week (C and D) and 31-week (E and F) postoperative radiographs of 6-year-old girl with surgically treated PAMFs without HO

Radiographs of two patients with UCMFs with an interval from injury to diagnosis of 1 month (A and B) and 48 months (C and D), respectively, showing HO (black arrow) with lower (A and B) and higher (C and D) bone densities than those of the surrounding normal bone tissue

Previous reports have shown that osteogenic factors and bone cell progenitors with osteogenic potential from the post-injury hematoma and a suitable mechanical microenvironment are critical for the development of HO [25, 26, 36,37,38,39,40]. In patients with UCMFs, the instability of the unreduced RH creates a favourable osteogenic environment [25, 26, 36,37,38,39,40]. However, when the RH is reduced and stabilized within 48 h of trauma, the osteogenic potential decreases due to the lack of mechanical stimulus [41, 42]. In fact, several experimental studies have shown that the lack of mechanical stimulation limits osteogenic potential even when osteogenic factors are added to the culture [41, 42]. In addition, we also found that the bone density of HO increased during the interval from injury to diagnosis, which corresponds to the period of mechanical stimulation due to unreduced RH instability (Fig. 9). Therefore, the higher rate of HO in patients with UCMFs than in those with PAMFs could be explained by the interaction between osteogenic factors, bone cell progenitors with osteogenic potential, and direct mechanical stimulation.

Interestingly, our study did not reveal an association between a longer interval from injury to diagnosis and an increased rate and size of HO in patients with UCMFs. This finding may be explained by the fact that the development and maturation of HO can occur within a short time after injury (Fig. 10). Several reports have shown that HO can be detected 2–4 weeks after injury [7,8,9]. Our study included only patients with UCMFs diagnosed > 4 weeks after injury, which was sufficient for HO development (Fig. 10). In addition, the distance between the RH and the radiocapitellar joint did not change substantially after injury due to surrounding scar tissue, and consequently, the size of the HO did not change significantly during the time from injury to diagnosis (Fig. 10).

Radiographs of a 7-year-old girl with UCMFs showing no HO (black arrow) in radiographs taken 6 h (A and B) and 12 days (C and D) after injury, except for radiographs taken 24 days (E and F) and 13 months (G and H) after injury

Our analysis showed that the associated radial or median nerve injury in patients with UCMFs did not lead to the development of HO, although previous studies have revealed that central nerve injury is a risk factor for HO [43,44,45]. This discrepancy could be attributed to central nerve injury disrupting the blood‒brain barrier through the release of osteogenic factors into the blood and subsequent increases in osteogenesis, whereas this does not occur in UCMFs with peripheral nerve injury [43,44,45].

In particular, our study had several limitations. First, this was a retrospective study, and the presence and characteristics of HO were evaluated only via radiographs rather than computed tomography. However, according to several studies, radiographs have been reported to help in the diagnosis of HO, and the minimum interval from injury to diagnosis is 4 weeks, which is sufficient for the development and maturation of HO [7,8,9, 46, 47]. Second, the correlation between the severity of initial displacement of the fractured ulna and the incidence of HO was not evaluated due to incomplete radiographic data. Nevertheless, we evaluated the direction of RH dislocation and DD-RH, which may reflect the severity of UCMFs. Third, the molecular mechanism underlying the development of HO was not elucidated in this study. Fourth, the correlation between the presence of HO and clinical outcomes, including pain, limited elbow range of motion and decreased quality of life, was not evaluated in the current study because the study was designed to evaluate the risk factors for HO formation on the basis of radiographic analysis. Additional studies are needed to further investigate the relationship between HO formation and clinical outcomes in patients with UCMFs. Despite these limitations, to the best of our knowledge, the current study is the first to evaluate the risk factors for the development of HO in patients with UCMFs on the basis of an adequate sample size.

In conclusion, the overall rate of HO was 13.1%, with UCMFs with anterior, anterior-lateral or anterior-medial displacement of the RH being more prone to developing HO than those with lateral or posterior-lateral displacement, whereas timely reduction of the RH in patients with acute injury prevents HO. The bone density of HO increases with age at injury and interval from injury to diagnosis. Age at injury greater than 6.78 years and DD-RH < 1.59 are risk factors for HO in patients with UCMFs. Despite these findings, additional studies are needed to confirm our results and to elucidate the pathogenic mechanism of HO in patients with UCMFs.