In this retrospective cohort study the overall rate of conversion to THA after surgically treated acetabular fractures was 13.9%. T-shape fracture patterns were found to be associated with an increased risk of early conversion to THA. Other fracture patterns classified by Letournel were not identified as independent risk factors. However, bivariate analysis suggested that fractures involving the posterior wall and associated both column fractures were associated with a higher incidence of early conversion to THA, but multivariate analysis did not prove these fractures to be independent risk factors in this study. Furthermore, higher age at time of trauma and poor degree of reduction were associated with a higher risk of early conversion to THA.

The conversion rate to THA found in this study was in line with rates found in existing literature, 13.9% vs. 8.5–30.95% [1, 3, 10, 11]. The main reason for conversion to THA is post-traumatic osteoarthritis. In literature reported risk factors for post traumatic osteoarthritis are mainly higher age (> 40 years), quality of reduction (Matta’s criteria) and AVN which is seen in 5.6% of patients after femoral head trauma [3, 11].

Ziran et al. [11] performed a review of available literature to identify independent risk factors for conversion to THA after surgically treated acetabular fractures. They found that the presence of fractures involving the posterior wall increased the risk of conversion to THA. According to Firoozabadi et al. [10] fractures involving the posterior wall are the most frequent fractures overall and are known to have a conversion rate to THA of 16.9% [10]. However, in this study both column fractures were as common as posterior wall fractures (17.1% for both) and the conversion rate to THA for fractures involving the posterior wall was 4.7% (2/48). Moreover, we did not find a statistically significant increased risk of conversion to THA for fractures involving the posterior wall. Briffa et al. [1] found that fractures involving the posterior column are associated with poor outcomes. However, in our study bivariate analysis showed that fractures involving the posterior column are not a risk factor for conversion to THA. Surprisingly, no significant differences in conversion to THA were found in patients with femoral head fractures or cartilage defects of the femoral head. This may be biased by the fact that femoral head fractures are relatively rare, and that cartilage defects are not seen on conventional imaging or CT. Therefore, cartilage defects were only noted if they were visualized during surgery. This could have led to an under-reporting of such lesions.

In this study we found that higher age is an independent risk factor for conversion to THA. Ziran et al. [11] found that age > 40 years was an independent risk factor for early conversion to THA. Rollmann et al. [24] found that the risk of conversion to THA increased with 6% per year of higher age. In this study the mean age at time of trauma was 49.5 years (SD ± 17.2 years). This is similar when compared to the mean ages at time of trauma found by Firoozabadi et al. (43.7 years) and Rollmann et al. (51.9 years (non-THA group) & 57.1 years (THA group)) [10, 24].

Poor fracture reduction was also found to be a risk factor in previous literature according to Ziran et al. [11]. Verbeek et al. [25] further specified that residual gap (> 5 mm) was associated with an increased risk of early conversion to THA. Briffa et al. [1] found that t-shaped fracture patterns were associated with difficult reduction and poor outcomes. Trouwborst et al. [26] found that CT based fracture displacement of < 2 mm had limited risk in conversion to THA in non-operatively treated acetabular fractures. Step-off ≥ 2 mm and age > 60 years were found to be predictors for conversion to THA.

Delay from time of injury to initial surgery of more than 15 days and more than 5 days for elementary and associated fracture patterns respectively was associated with an increased risk of conversion to THA according to Ziran et al. [11]. In this study analysis showed that delay to initial surgery did not differ between the THA and non-THA group and was therefore not included in the logistic regression model.

Primary THA as treatment for acute acetabular fractures is of growing interest. However, data is limited and variable [19, 20]. De Bellis et al. [19] found that in elderly patients with acetabular fractures and poor bone quality, combined acetabular and femoral neck fractures or pathological fractures primary THA is as successful as delayed THA. Iqbal et al. [22] found that primary THA for complex acetabular fractures in selected elderly patients resulted in favourable Harris Hip Scores (HHS). Thirty-seven patients (78%) who underwent primary THA for complex acetabular fractures had excellent outcomes as per HHS. Jauregui et al. [20] describes higher overall incidence of femoral head dislocation and heterotopic ossification after primary THA for acetabular fractures compared to THA in a non-traumatic setting. There is no literature available that specifies outcomes of primary THA for T-shaped fracture patterns. Manson et al. [27] found that in T-type fracture patterns, bone structures that are essential for acetabular component stability for THA are disrupted. A combination of ORIF with THA could restore acetabular component stability. With growing knowledge about which risk factors contribute to conversion to THA, more research should be done to investigate whether primary THA is a successful treatment option for selected acetabular fractures.

Sample size and the monocentre setup are the main limitations of our study. The sample size needed for a power of 80% at a significance level of 95% for operatively treated acetabular fractures and conversion to THA can be estimated from Tannast et al. [6]. A total of 254.341 acetabular fractures would be needed to analyse the different types adjusted for the prevalence of each type. Since acetabular fractures are relatively rare our study population is limited. Collaboration between hospitals in future research to increase sample size could be done to further support existing literature.

Another limitation of this study is that due to some high-risk fracture patterns being rare, such as the T-shape fracture. Surgeons might be less experienced treating these fracture types. Larger volumes of rare fracture types would be needed to distinguish risk of conversion solely due to fracture type or due to lack of experience by the surgeon.

Despite this study had a retrospective setup, there was limited recall bias. Data was prospectively collected in our electronic database and processed immediately during treatment and further follow up. A few patients were lost to follow up and therefore not included in this study.