To reveal trauma severity, the Injury Severity Score (ISS) 2021 and the Abbreviated Injury Scale (AIS) 22 were used. The presence or absence of a head injury was classified by initial GCS and simultaneous CT-Scan abnormalities such as fractures of the skull or intra-cranial injuries. Patients with an almost normal to normal GCS and combined anatomical lesions on the CT-scan were classified as head injured patients.

Patients with symptomatic HO at routine follow-up in our clinic were included in the present study. A great part of heterotopic ossifications cause swelling, pain or limited function to total ankylosis. Since these patients confront the clinician during every day work and utilize clinical resources, we focussed on these patients. We asked and examined the patients towards one ore more of these symptoms and took x-rays of the affected body region in standardized antero-posterior and lateral views from the follow-up appointment three to six months after the initial injury for radiologic confirmation of suspected HO (figure 1, figure 2).

Today, Brooker's classification is widely accepted for classification of the HO around the hip joints, classifying HO into 4 grades ranging from just visible (grade 1) to total ankylosis (grade 4) in standardized x-rays in two planes 23. Unfortunately, a general and comparable classification system of all joints to date does not exist. We therefore adapted and modified Brooker's classification in a similar way to the other joints and defined the extent of the heterotopic ossification accordingly (grade 1-grade 4, in the following "radiologic ossification class"). Additionally, the maximum expansion on both films was measured in mm. Furthermore, the location at the fracture site (fractured long bone between the adjacent joints) or at a site remote to the fracture site (any non-adjacent part of an extremity) was noted. All the x-rays were analysed and classified by two independent trauma surgeons (J. S. and C. P.).

Patients were defined to receive prophylactic medications, if corticoids, non-steroidal anti-inflammatory drugs (NSAIDs), muscle relaxants, diphosphonates or hyaluronidases were administered in a prophylactic regimen.

We defined surgical fracture if initially intramedullary nailing, plate osteosynthesis or external fixateurs with secondary conversion to intramedullary nailing (damage control orthopaedics, DCO) were used. No other methods of fracture care such as extension treatment or casting were used in our population.

Ventilation time and duration of intensive care unit stay were analysed.

Results are shown as mean ± standard error of the mean (SEM). For the analysis of nominal-scaled variables the Chi-squared test (Chi²) was used, for continuous data we used the student t-test. In addition, analysis of variances (ANOVA) was performed followed by post-hoc Tukey test to determine differences between groups. Level of significance was set at p < 0.05.

The study population consisted of 79 men (78.7%) and 22 women (21.3%). Average age between groups showed no significant difference (IMN: 27.1 ± 3.1 vs. PLATE 29.1 ± 2.6 years, p = 0.25). The GCS mean value was also statistically comparable between groups (IMN: 10.7 ± 0.8; PLATE 11.0 ± 1.0; p = 0.93) as was the incidence of head injuries (IMN: 33% vs. PLATE: 24%; p = 0.36).

Additionally, PLATE group showed a higher rate of articular fractures (63% vs. 28% in IMN; p = 0.003) while IMN demonstrated a higher rate of diaphyseal fractures (72% vs. 37% in PLATE; p = 0.003).

Comparing the mean ISS, and AIS max there was no statistical difference between our groups (table 1).

A significantly higher incidence of radiologic classes 3 or 4 was found for the PLATE group in comparison to the IMN group (p = 0.04; figure 3).

For the largest extension of the HO in mm in two views of plane x-rays the p-values show no significant difference but a tendency towards larger HO-formations in group PLATE (table 2).

HO occurred significantly more frequently remote to the fracture site in the IMN group in comparison to the PLATE group (p = 0.03; figure 4).

No differences of ventilation time (IMN: 12.2 ± 3.1 days vs. PLATE: 11.0 ± 2.7 days; p = 0.48), duration of the ICU-stay (IMN: 14.6 ± 3.9 days vs. PLATE: 13.2 ± 3.6 days; p = 0.76) and indomethacine prescribed (IMN: 22% vs. PLATE: 30%; p = 0.47) was demonstrated.

The role of head injuries in the formation of HO still is lively debated about in the literature. Some authors reported a stimulation of fracture healing in patients with head injuries 293031. Furthermore, a positive correlation of the severity of the head injury and the HO rate was observed 24. Other studies could not confirm a relationship between severe head trauma and HO formation. Lehmann et al. demonstrated constant expressions of the HO in multiply injured patients without head trauma in comparison to multiply injured patients with severe head trauma 4. We could confirm the findings of Lehmann et al., the present report could demonstrate comparable GCS and constant incidence of head trauma in both groups.

Interestingly, a recent study demonstrated differences in the location of the HO between polytrauma patients with and without severe head trauma. In polytrauma patients with associated head trauma, the HO was located adjacent to the fracture region. In polytrauma patients without head injury, the HO formation more frequently occurred at sites remote to the actual fracture sites 7. In our study, the incidence and severity of head injuries was distributed equally between both groups.

Nonetheless, we found a higher incidence of remote HO in the IMN group, leading to the idea of systemic factors liberated during nailing that affect HO formation such as prostaglandin E2 1332.

Surgical treatment such as osteosynthesis, manipulation at joints or traumatic haematoma is known to be a risk factor for the development of the HO 63334. In the present study, we could demonstrate a positive association of plate osteosynthesis and the development of the HO in the PLATE group.

A more invasive approach required for plate osteosynthesis is well described as one of the risk factors 6. Local fracture and soft tissue manipulation is believed to hold a substantial role in the development of the HO, possibly by the liberation of bone morphogenetic protein (BMP) or other tissue factors 3536. Home et al reported on extended HO after intramedullary nailing in combination with severe head trauma 37. However, these results could not be shown in our study potentially due to a relatively low patient number.

In the present study, there were no significant differences in ventilation time (IMN: 12.2 ± 3.1 days vs. PLATE: 11.0 ± 2.7 days; p = 0.48). Long term ventilation is widely accepted as a factor associated with HO formation 2: One study showed HO in patients after pulmonary transplantation with prolonged ventilation times at healthy joints 38. Mechanical ventilation may lead to changes in the acid-base metabolism which results in mineral accumulation in the soft tissues and therefore may lead to HO formation 5 which was also demonstrated in an experimental study 34. Other authors speculate that HO formation in shock trauma patients and mechanically ventilated patients occurs due to critical hypoxia in consequence to local tissue compression. It could be revealed that osteogenesis is induced by low oxygen concentrations 33.

Prophylactic medications to prevent or to decrease HO are widely discussed in hip and acetabular surgery. Moreover, several studies revealed the effectiveness of prophylactic treatment after knee arthroplasty 181939. Prophylactic strategies may lead to decrease the development and the resulting size of the HO; these strategies include treatment with NSAID or postoperative radiotherapy. Best evidence for prophylactic medication is shown for indomethacine for at least 7 days, other NSAIDs are also well documented 19. To our knowledge, there are no reports on the effect of prophylactic medication on HO formation in multiple trauma patients. In our study, up to 30% (group PLATE) of the patients received prophylactic medications, there were no differences of NSAIDs prescribed (IMN: 22% vs. PLATE: 30%; p = 0.47).

The missing effect of the prophylactic treatment in our study may be the result of the low fraction of patients who received prophylactic treatment. On the other hand, HO formation in multiply injured patients may result out of interactions of multiple systemic and local factors, thereby limiting the effect of a single intervention or substance.