No significant associations between patient characteristics (e.g. age, BMI, ISS) or treatment-specific factors (e.g. time to surgery, post-operative weight-bearing protocol) and the occurrence of implant failure was observed. Factors, such as fracture severity, additional posterior stabilization, and the specific type of posterior injury did not influence implant failure rates. The Majeed Pelvic Score was higher in the implant failure group after adjusting it to the previous work of patients. The implant failure rate of 37% is comparable to previous reports [12, 15, 18, 19]. Furthermore, the median time to implant failure of approximately 10 weeks is comparable to Rojas et al. (seven weeks), Eastman et al. (13 weeks) and Avilucea (16 weeks) et al. but earlier than reported by Morris et al. (one year) [5,6,7, 12].

The inability to predict implant failure was previously reported in a more heterogeneous group of pelvic ring injuries [21]. As in the study of Frietman et al. no predictors of implant failure in patients’ demographics could be detected [15]. Tseng et al. reported, that males suffer more often from implant failure [19]. Due to the gender inhomogeneity of the cohort presented here, with 93% male patients, this finding could either be proven or disproven.

Conflicting reports exist, regarding the effect of fracture severity according to the AO classification, and it is poorly documented for the Young and Burgess classification [5, 6, 15, 20]. The advantage of our study is the use of both the Young and Burgess and AO classification, particularly because of the conflicting recommendations for comparable injuries associated with the use of different classification systems. Performing a global survey yielded a predominant use of stand-alone anterior plating especially in Europe for AO type B1.1 injuries [22]. In contrast, a survey in the UK revealed a favored treatment using an anterior plate with an additional SI screw for APC II injuries [10]. Different recommendations may result from to a more heterogeneous injury pattern and displacement within similar classified injuries as known from lateral compression fractures [23]. This hypothesis is supported by the recommendation of Gill et al. performing an individual assessment of stability and required stabilization even in similarly classified injuries [10]. The fracture classification was not predictive of implant failure in the present study.

While the choice of a two- vs. a four-hole plate affects the occurrence of implant failure [17], the choice of longer plates or double plating does not affect implant failure [6, 15, 18, 19].

Besides fracture classification, the type of posterior injury may affect implant failure. Eastman et al. determined implant failure predominantly in patients suffering from sacroiliac joint injuries [7]. This may be due to the underestimation of instability or micro-instability caused by these injuries, or the lack of ability to detect them on static imaging [7, 18]. Such instabilities could be addressed with an additional posterior fixation resulting in a reduction of implant failure [12]. However, the present study as well as previous studies were unable to support these finding [5, 6, 15, 18, 19].

In addition to different classifications, different weight bearing recommendations for the same injury pattern can affect implant failure [10]. The present study could not support this thesis, which can be explained by a possible incompliance of the patients with partial weight bearing which could not be excluded [7].

The impact of implant failure on functional outcome is still a matter of debate [15, 17]. Frietman et al. supported the view, that implant failure could be the result of healing and the return of mobility within the pelvic ring and therefore should not be considered as a complication [15]. Pain levels did not differ in this study comparable to previous reports [17].

Compared to previous studies, the Majeed Pelvic Score was higher in the present study [15, 18, 19, 24, 25]. However, there are differing opinions on the impact of implant failure on the functional outcome as followed: no impact [19, 26], a tendency for better outcome without significance for intact implants [5, 17] or implant failure [15]. In the present study, the implant failure group showed a significantly better outcome adjusting the Majeed Pelvic Score to the work category.

The present study was limited by the retrospective design, the predominance of male patients, and the small number of patients, which reduced the power. Functional outcome could be estimated in only 60% (25/42) of the cohort.

In conclusion, implant failure is a common radiologic phenomenon with little or no relevance to revision indication or functional outcome [20]. In particular, screw loosening should not be overemphasized and, as previously suggested, radiologic analysis may not necessarily predict functional outcome [15, 27]. Therefore, plate removal in asymptomatic patients is not recommended and the addition of a sacroiliac screw should be critically discussed.