The coronoid process (PCU) is the most important osseous stabiliser of the elbow joint (EJ) [1]. Biomechanical studies have extensively shown that an isolated fracture involving over 50% of the PCU leads to an instability involving dorsal luxation of the joint, despite intact capsule ligaments [2, 3]. In combination with the loss of the radial head, a PCU-deficit of 25% is sufficient to allow for a dorsal reluxation [2]. The stabilisation of the joint against varus-valgus stress is primarily effected by the ulnar and radial collateral ligaments [1, 4]. Thus, a greater valgus instability can be expected in the case of an isolated transection of the ulnar ligament apparatus when compared to a sole radial head resection [5]. Hyperextension trauma [6] and hypersupination trauma during a twisting fall [7] will be discussed later as examples of an accident causing an elbow luxation.

Overall, fractures of the coronoid process are rare and appear as concomitant injuries in 2 to 11% of elbow luxations [8, 9]. This represents an incidence of 5.21 per 100,000 people a year [10]. Approximately 23 to 61% of coronoid fractures are treated with surgery [11, 12]. The typical injury combination of a PCU and radial head fracture along with an elbow luxation, which consequently causes joint instability, is referred to as the “terrible triad” injury of the elbow joint [13]. Clinical studies on therapies of various injury combinations and their different degrees of severity are difficult to carry out as the number of cases is low and therapies are sparsely standardised across hospitals. In two cooperating trans-regional trauma centres using the same therapy strategy, we performed a retrospective follow-up examination to identify medium-term treatment outcomes and isolate prognostically relevant factors.

Seventy-seven patients with a fracture of the coronoid process were treated in an 8-year period (2001 to 2009). Within the scope of a retrospective study (ethics commission of the University of Lübeck, Ref No: 10-024), the treatment outcome of 58 patients was acquired. Ten of the 58 (17%) patients filled in the questionnaire only, while for 14 patients (24%), medical reports existed for submission to insurances and 34 patients (59%) took part in the actual follow-up examination. After an average of 48 months (SD 31; 9 to 120 months) and following initial treatment, the treatment outcome of 58 patients (75%) was obtained.

In accordance with the Mayo-Elbow-Performance-Score (MEPS) [8] and the Disabilities of the Arm, Shoulder and Hand Score (DASH) [14], subjective details regarding pain, joint stability and function in everyday life were recorded. During the clinical follow-up examination, circumference measurements and motion ranges of the arm’s joints were documented using a standardised measurement log (form F4222 [15]). The stability of the elbow joint was measured with a protractor in steps of 5 degrees via varus-valgus stress in the extended position and as a deviation of the longitudinal axis. Afterwards, force was measured in a 90° flexed position of the elbow joint, with the subject seated, using a Genius force measuring device (0 - 5 kN, FREI AG, Kirchzarten, Germany). The measurements were carried out three times per patient and the means recorded. Based on indication, an X-ray of the elbow joint was carried out (20 out of 48 patients).

The recorded data were saved in an Excel spreadsheet (Microsoft) and anonymously analysed in cooperation with a statistician employing the SPSS statistics software (Version 19, IBM, Ehningen, Germany). The data underwent descriptive statistical analysis, including analysis of frequencies and distributions as well as an analysis of variance (ANOVA). The non-parametric group comparisons were conducted utilising the Mann-Whitney U-test. Pearson’s correlation analysis was used on data subjected to normal distribution. Mean numbers of groups were rounded to integers and significances to the third decimal place. The analysis was based on MEPS and DASH. According to Kendall, both rating scores strongly correlate (τ_b =-0,867, p=0,000); therefore, only one score (MEPS) was used for statistical analysis.

Significant differences or correlations between a specific pattern of injury or a specific treatment and the follow-up examination result were explored. Thus, individual injuries as well as recurring injury combinations were analysed (Table 1).

On the whole, there are a small number of retrospective studies dealing with the clinical course of coronoid fractures [11, 20-25]. Only one study features over 40 cases [11]. They all report on injury combinations and show that therapeutic strategies have become more surgical recently than 20 years ago [12]. The reason, besides the rather unsatisfactory results after predominantly conservative therapy [12], is the sophisticated knowledge of the biomechanical implication of the coronoid process [2, 3, 5]. The poorly treatable instability of injury combinations involving PCU with RH fracture and collateral ligament ruptures (terrible triad) continues to be a therapeutic challenge [25].

The proportion of concomitant elbow joint luxations in PCU fractures has not been determined to date. In 44 out of 58 patients (76%), we noticed a luxation. Adams et al. [11] registered a luxation in only 49 out of 103 (47%) coronoid fractures in their patient population and Regan in just 12 out of 35 patients (34%) [12]. We could not verify a significant difference in the outcome regarding patients with or without luxation. The follow-up results showed preference for neither surgical nor conservative therapy of ligament injuries. The most recent publications [21, 22, 24, 26] reported on surgical reconstructions of the collateral ligament apparatus in luxation fractures, but not using any prospective studies. The only available prospective study on isolated ligament lacerations without osseous instability is [27], which shows that the follow-up results of surgical procedures are not better than those of conservative therapies.

During follow-up, the coronoid process fractures treated by us revealed - in relation to the MEPS - no significant differences between the types of fracture. Significant differences in the functional result were described by Regan and Morrey [12], where the therapy was, to a certain extent, stage-independent and predominantly conservative. Adams et al. also described an increasing restriction of ulnohumeral mobility in relation to the coronoid fracture type, where the evaluation was complicated by a rising number of additional fractures [11]. A primarily stable osteosynthesis of the PCU fracture appears to be important for a successful surgical procedure [21]. However, this has its limits in cases of reduced bone quality caused by osteoporosis (9 of the 58 patients (15%) were aged over 65 at the time of surgery) and in multi-fragment fractures with osteosynthesis plates already in place. That is why the fractured coronoid process could not be preserved in 4 cases, but had to be replaced with an autologous bone block. In fact, one of those patients needed a complete elbow prosthesis due to a persistent reluxation. In the examined population, patients with combined variations of osteosyntheses (plate and screws) showed significantly poorer results than those treated with only one type of osteosynthesis (plate or screws). A total of 15% of the coronoid fractures could not be sufficiently treated and stabilised with the implants in place. Future studies should evaluate whether a more stable treatment can be achieved by using anatomically preformed fixed-angle plates (e.g., tifix-coronoid-plate, Litos, Ahrensburg, Germany). Moreover, in selected cases, stabilisation can be achieved by a coronoid substitute prosthesis [28].

The presence of an additional RH fracture is described as the biggest negative prognostic factor by Adams et al. [11]. In isolated occurrences of RH fractures, positive long-term results are mostly expected despite complete resections [29], even if the missing radial bracing on the Capitulum humeri leads to an additional load of the medial collateral ligament apparatus and the humeroulnar articular surface. Data regarding impacts on wrists are inconsistent [30, 31]. In luxation fractures of the elbow joint, there is a risk of persistent instability and reluxation after RH resection [22]. From a therapeutic perspective, maintenance or replacement therapies of the RH are being increasingly mentioned in more recent studies [11, 24, 32], even if some studies do report positive long-term results after complete RH resections in spite of a previous luxation [33, 34]. In the examined patient population, MEPS results for patients whose RH were preserved were only tendentially better than those with a complete resection. Therefore, we cannot statistically back up a therapy recommendation. However, having interpreted our data, we think it is biomechanically sensible to either preserve or replace the RH in order to achieve a better stability of the joint. Functional results also seem to support that in the long-term course [22, 24, 35].

In both trauma centres, follow-up care aimed to attain early mobilisation of the elbow joint at all levels, given that it has been demonstrated that a remaining flexion deficit prognostically correlates with the duration of immobilisation [36]. Thus, an immobilisation period of less than 3 weeks allowed for predominantly good to excellent results [22]. In the population we examined, significant differences in the outcome were also found between patients with a cast immobilisation period of less than 3 weeks and those of more, although both patient groups showed no differences in the type or number of injuries. Therefore, we regard immobilisation as an independent influencing factor. In contrast, a comparison of patients with or without application of an external fixator showed significant differences in the injury rate per elbow joint and in the rate of required revision surgeries. Hence, the significant differences can only partially be assigned to the immobilisation by the fixator. A prospective randomised study should initially clarify [37] whether a fixator with motion capacity can yield an advantage.

A total of 69% of the interviewed patients subjectively stated that they had a stable joint; 67% of the re-examined patients were, in fact, objectively stable. With that said, the proportion of stably-healed elbow joints was smaller than those in other follow-up examinations [21, 33, 38, 39]. However, the comparability is limited by different patient groups and missing data on clinical stability checks. Statistically, we were unable to determine a correlation between specific osseous or ligament injury combinations and therapeutic strategies. The remaining instabilities after completed treatment did not lead to recurrent luxations in our patient population, but based on the significant correlation between the pain statements and the associated load reduction, they have to be rated as substantial limited functionalities. In the long term, a limited joint stability also seems to exacerbate arthrosis [25]. Hall et al. reported on the link between pain and instability [40] and Adams et al. [11] described the connection between previous ligament injury and chronic pain. Therefore, achievement of a stable joint should not only be assessed by the reluxation rate, but also via the regained load-carrying ability.

Limitations of this study include the retrospective study design and in homogenous patient population, CT data was not available in all cases. Strength is the large follow up period monitored and the number of cases overseen by this study.

At present, fractures of the coronoid process can only be limitedly reconstructed by osteosynthesis and often heal with permanent functional deficits, their prognoses largely contingent on concomitant injuries. Remaining instabilities of the elbow joint appear relatively often after complex luxation fractures; instabilities are therapeutically difficult to manage and significantly correlate with pain. Surgical therapy of ligament injuries cannot be justified statistically. A stable osseous reconstruction appears to make more sense.

Predictive statements regarding individual injuries are, for fracture-specific therapy and simultaneous injury combinations, only just possible. The strongest negative prognostic parameters in our patient population were: therapy with an external fixator, immobilisation for more than 21 days, the occurrence of complications and unstable osteosyntheses on the coronoid process.

No financial assistance has been accepted for the conduction of this clinical study and its analysis. The aforementioned tifix-coronoid-plate from Litos was developed in cooperation with the lead author.