Introduction

If both the radial head and the coronoid are involved in elbow dislocations, these lesions are referred as ‘terrible triad injuries’ (TTI) [1]. Due to the loss of the postero-lateral stabilisation of the lateral ulnar collateral ligament (LUCL), the valgus buttress of the radial head, and the anterior buttress of the coronoid, these injuries are prone to result in chronic instability, elbow stiffness and post-traumatic arthritis, if not treated adaequately [2,3,4,5].

In this context, the coronoid process plays a significant role in stabilizing the ulnohumeral joint, as it forms an anterior buttress that in combination with the radial head prevents the elbow joint from posterior dislocation. Due to the posterolateral trauma mechanics [6], most coronoid fractures in TTI are small transverse fractures of the coronoid tip [7, 8].

Currently, there is a consensus that only selected patients with a TTI can be treated non-operatively [9,10,11], whilst the vast majority requires surgical treatment to achieve a stable elbow which permits early rehabilitation [12,13,14]. If surgery is performed, each of the individual bony and soft-tissue components of the injury should be addressed based on the most recent treatment suggestions [14,15,16], leading to satisfactory clinical outcomes [17,18,19,20].

However, the decision, if a concomitant coronoid tip fracture should be fixed, remains controversial [21]. While some authors recommend skillful neglect [10] or excision of the fragment [22, 23], others suggest that any associated coronoid fracture, regardless of fracture classification, should be fixed [15, 22, 24,25,26]. However, most of these studies lack adequate cohort sizes or comparison group as well as a proper classification system for coronoid tip fractures. While attempts to classify these fragments according to height as defined by Regan and Morrey [27] have been inconsistent and contentious [28], O’Driscoll et al. [29] suggested a classification of coronoid fractures according to fracture morphology and injury pattern. Based on their classification system, coronoid tip fractures are referred as O’Driscoll type 1, which are usually associated with elbow dislocation in the context of a TTI. Therefore, our primary objective was to evaluate clinical outcomes of patients with a O’Driscoll type 1 coronoid fracture in a large sample of terrible triad lesions at mid-term follow-up. The secondary objective was to determine, if there is a difference in functional and clinical outcome between a surgical or a non-operative approach.

Methods

Study population

This is a cohort analysis of a retrospective case series at a single level-I trauma center. After approval by the Regional Ethics Committee (FF92/2018), patients were selected by searching the clinic’s patient management system (medico® by Cerner Health Services GmbH, Idstein, Germany) from 2010 to 2018 for all surgically treated terrible triad injuries with coronoid tip fracture, using the code S53.12 of the International Statistical Classification of Diseases and Related Health Problems, Tenth Edition (ICD-10). Of those, 62 adult patients with TTI and concomitant coronoid tip fracture who underwent surgery in our hospital within 14 days after injury and without previous elbow surgery or elbow joint-specific comorbidities could be included for final assessment [Fig. 1].

Fig. 1
figure 1

Flowchart diagram of the process for patient inclusion in this study

Full size image

Fractures of the coronoid were classified by two independent investigators (AN, PH), specialized in the field of elbow surgery, according to the O’Driscoll classification system [29]. Only type 1 fractures (coronoid tip fracture) were included (subtype 1.1, ≤ 2 mm coronoid height, subtype 1.2, > 2 mm coronoid height). Fractures of the radial head were classified according to its displacement pattern equivalent to the Mason system [30]; Open fractures were graded by the Gustilo and Anderson system [31].

Operative technique

Standard surgical approach was through Kocher’s interval, while in cases where the MCL had to be addressed, we used a combined approach. Structures were generally addressed in a deep to superficial manner (coronoid, radial head, LUCL). All Mason type I fractures were treated non-operatively. In reconstructable Mason type II and type III fractures, the radial head was fixed using mini-screws or low-profile locking plates. If unreconstructable, radial head arthroplasty (RHA) was performed (n = 20) (15 × MoPyc Bioprofile, Tornier, France; 5 × SBi rHead, Stryker, USA). Coronoid fractures were treated based on surgeons’ preference either non-operatively (n = 36) or reduced by screws (n = 11), if amendable for direct fixation, or by non-absorbable sutures in transosseous or suture-anchor-based manner (n = 15) [Figs. 2 and 3], if too small or comminuted for direct fixation with screws. The lateral collateral ligaments were repaired using 3.5 mm suture anchors (Arthrex, Naples, USA) in all cases. Stability of the elbow was then tested with the hanging arm test under flouroscopy. If instability, defined as a non-concentric reduction of the ulnohumeral joint through a range of 20° to 130° of flexion–extension, persisted at this point, we proceeded to repair the medial collateral ligament (MCL) (n = 27). A hinged protective external fixator was applied in three patient due to excessive swelling after trauma, and was removed after 3 weeks.

Fig. 2
figure 2

Preoperative CT scans (A, B) of a 53-years-old female after a staircase fall, showing a terrible triad injury with a radial head fracture type Mason III equivalent and a coronoid tip fracture type O’Driscoll 1.2

Full size image
Fig. 3
figure 3

Postoperative radiographs (ap, lateral) of the same patient as in Fig. 1. Radial head has been replaced by a modular monopolar model (MoPyc Bioprofile, Tornier, France) and the coronoid tip fracture has been reattached using a transosseous suture and button fixation at the posterior cortex of the ulna

Full size image

Postoperatively, a hinged elbow brace in neutral forearm rotation was applied, from 20° of extension to 120° of flexion, for six weeks. Physiotherapy with active (without weight-bearing) and passive motion was indicated from the first postoperative day. All patients received oral nonsteroid anti-inflammatory medication for two weeks as an ossification prophylaxis.

Clinical and radiographic evaluation

After a minimum follow-up of 2 years, patients were invited for clinical evaluation by an independent investigator.

Range of motion of both elbow joints was tested using a standard goniometer and ligamentous stability was measured in maximum extension and at 30° of flexion. Elbow stiffness was rated according to the degree of the residual arc of motion (severe when the total arc was 60° or less, moderate when it was between 61 and 90°) [17]. Functional outcome was assessed using the Mayo Elbow Performance Score (MEPS) [32], Oxford Elbow Score (OES) [33], and the German Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire [34]. Posterolateral rotatory stability was evaluated using the posterolateral rotatory instability test and graded normal (0), mild (I), moderate (II), or severe (III) according to the grade of joint dislocation [35].

A Jamar dynamometer (Fabrication Enterprises Inc., White Plains, New York) was used to assess grip strength compared with the uninjured side, using a correction factor of 1.07 for the dominant over the nondominant hand [36].

Pain-level was rated using the visual analogue scale (VAS).

All available radiographs were evaluated by two independent investigators (AN, PH). Coronoid fracture height was measured based on pre-operative CT scans using the technique described by Doornberg et al. [28]. Radiographic signs of osteoarthritis (OA) were rated according to Broberg and Morrey [37] and heterotopic ossifications (HO) were graded using the system of Hastings et al. [38]. Stability was assessed at final examination as a congruent joint on radiographs and absence of dislocation or subluxation events. Complications were defined as adverse events directly related to the chosen treatment, and were graded as ‘major’, if they required revision or if elbow stiffness (defined as range of motion < 60° in extension-flexion) persisted after 6 months. Revision was defined as any subsequent surgical intervention related to the index procedure.

Sample size calculation and statistical analysis

To determine the adequate sample size for group comparisons between the surgical treatment options of the coronoid tip (operative vs. non-operative), we used the OES as our primary outcome variable, which has a minimal clinically important difference of 8.2 for the elbow [39]. Consequently, a two-sided unpaired t test with an alpha-level of 0.05, a power of 80%, and an allocation ratio of 1, requires 26 patients in each group to detect an effect size of 0.8.

Statistical analysis was conducted using IBM SPSS statistics version 25 (IBM Germany GmbH, Ehningen, Germany). Fisher’s exact test was used for determining statistical differences for categorical data such as stability, complication or reoperation rate, HO and OA. Mean values were compared using independent t tests for normally distributed variables and Mann–Whitney U tests for non-parametric variables, when applicable. If more than two groups were compared, either a one-way ANOVA or the Kruskal–Wallis test were performed. P values of < 0.05 were considered significant.

Results

Patient demographics are detailed in Table 1. The average follow-up was 4.2 years (range 24–110 months). At final follow-up, mean flexion–extension range of motion (ROM) was 121° (range 85°–150°) and mean pronation-supination ROM of the forearm was 162° (range 90°–180°), both differing significantly to the unaffected side (P < 0.001).

Table 1 Patients demographics
Full size table

Average pain-level measured using the VAS was 1.5 (SD 2.1), although five patients had to rely on pain medication due to chronic elbow pain. At the final visit, mean MEPS was 86.9 (range 35–100), mean OES was 36.0 (range 11–48) and mean DASH score was 20 (range 0–61), with about 75% of patients reporting excellent or good results (n = 46). A mean grip strength of 88.7% (range 43–130%) was measured compared with the uninjured arm, with about half of patients regaining at least 90% of the uninjured side. Grade-I posterolateral, and mild valgus instability were detected in one and four cases, with no correlation between the occurrence of a valgus instability and the treatment of the coronoid or MCL-complex.

Impact of coronoid treatment

All fractures showed no involvement of the anteromedial facet, with all fractures affecting the tip and/or the anterolateral facet of the coronoid. Average coronoid fracture height was measured 5.4 mm (range 0.5–11.1 mm), ranging between 3 and 30 percent of complete coronoid height. While the (percentage) height of the coronoid fractures had no statistical influence on the final outcome (PMEPS = 0.482; POES = 0.323), patients with a O’Driscoll subtype 1.2 fracture showed a greater impairment in function and range of motion than those with subtype 1.1 fracture [Table 2].

Table 2 Comparison of the different types of coronoid tip fractures with regard to functional and clinical outcomes (abbreviations in the main text)
Full size table

Considering the treatment applied to the coronoid, there were no statistical differences in demographics or concomitant treatment procedures between both groups, except for a higher rate of radial head arthroplasty in patients with surgically treated coronoid [Table 3].

Table 3 Patient demographics for different coronoid fracture procedures (abbreviations in the text)
Full size table

Overall, no significant benefit in outcome variables was detected when the coronoid tip fracture has been fixed [Table 4], with the non-fixation group even showing significantly more favorable results without consideration of the group heterogenity. In order to minimize selection bias in both groups, stratification for radial head treatment also revealed better movement and function for an non-operative treatment of the coronoid, but no statistical differences was reached (PMEPS = 0.112) in all outcome variables.

Table 4 Comparison of the different treatment procedures for coronoid fractures with regard to functional and clinical outcomes (abbreviations in the main text)
Full size table

Radiographic evaluation

Follow-up radiographs showed the development of HO in 16 cases (12, stage 1; 1, stage 2A; and 3, stage 2C ossification), with no significant reference to a certain treatment procedure of the coronoid (Fisher’s exact, P = 0.177). However, the occurrence of HO led to significantly worse range of motion and functional outcome parameters [Table 5].

Table 5 Outcomes of patients with heterotopic ossifications (HO) or osteoarthritis (OA) on their latest radiographs (abbreviations in the text)
Full size table

Eighteen patients showed signs of OA on their latest radiographs independent of coronoid treatment procedure (P = 0.165), including 12 with grade I and 6 grade II degenerative changes. Similar to HO, OA was also associated with significantly inferior outcome scores [Table 5].

Complications and reoperations

With all 62 patients included, the total incidence of major complications was 43.5% (n = 27), with eighteen patients (24.2%) requiring at least one second surgical procedure during follow-up. These included three cases of postoperatively persisting humeroulnar subluxation (two in the fixation and one in the no-fixation group), which were treated by revision surgery with an additional medial ligamentous repair in two and application of a hinged external fixator in one patient, who already had fixation of his coronoid during index procedure. In one patients, an early peri-implant infection was successfully treated by surgical revision. During follow-up, no secondary dislocation of the elbow occurred. The largest portion of complications was due to patients with ongoing severe joint stiffness (n = 18), of which seven patients underwent surgical arthrolysis during the first postoperative year. Additionally ulnar neurolysis due to persisting ulnar dysesthesia was necessary in three patients. Overall, no statistical difference in complication or revision rate could be detected between both treatment groups, although severe joint stiffness and subsequent arthrolysis was more frequently performed in the coronoid fixation group [Table 3].

Discussion

The treatment of TTI of the elbow is challenging and can lead to an unsatisfactory outcome. In this context, the optimal treatment of certain aspects of these injuries remain unclear. Surgical reconstruction aims to restore sufficient elbow stability to allow early mobilization within a stable arc of motion. While most authors agree that the radial head should be fixed or replaced [20, 40, 41], and the LUCL should be repaired to provide posterolateral stability [16, 42], the optimal treatment strategy for coronoid (tip) fractures remains topic of current discussion. We, therefore, present the results of the largest comparative study of these lesions yet published. Based on our findings, no benefits on clinical and functional outcome could be detected, if the coronoid tip fracture was surgically fixed. Furthermore no residual instability was evident in any patient at the latest follow-up visit. However, we detected a rather high complication (43.5%) and revision rate (24.2%) in our study cohort, independent of the coronoid tip treatment. While, these rates seem high compared to previous studies, it should be noted that our main complication (as well as main indication for revision) has been symptomatic elbow stiffness, which has not always been considered a “complication” in most other studies. However, as these conditions are usually very limiting for the patients, we think, this should also be seen as such, thus leading to a higher complication/revision rate in our cohort. As those injuries are regularly very severe and challenging, this only emphasizes the importance of an adequate treatment. Due to its eminent role for elbow stability, most coronoid fractures in TTI have historically been treated surgically [15, 43, 44].

However, evolving scientific research and biomechanical analyses over the last decade have led to a better understanding of the complexity of elbow injuries and the importance of restoring stability through the coronoid [22, 23, 45].

Based on the kinematic analyses by Doornberg et al., the majority of coronoid fractures in TTI can be classified as small transverse (anterolateral) tip fractures (type 1 according to O’Driscoll) [8], rarely exceeding 30% of the coronoid height, which can be confirmed by the data of our study cohort. Due to the attachment of the anterior capsule many authors suggest repair of these lesions (and its capsular attachment) to sufficiently restore the anterior column of the elbow joint [4, 15, 25, 42, 46]. However, cadaver studies showed no impairment in elbow stability if less than 30% of the coronoid have been resected, as long as the MCL and radial head remain intact [22, 26, 45].

However, caution has to be exercised when attempting to transfer the results of biomechanical studies to the clinical setting, because the complexity of elbow injuries, coupled with in vivo motion, and the stresses across the human elbow, cannot be completely replicated in laboratory settings.

Antoni et al. [47] reported that reattaching the anterior capsule in terrible triad injuries did not improve the final clinical and radiographic outcomes, although radiographic evidence of humero-radial osteoarthritis was significantly more common in the absence of re-attachment. They concluded that elbow stability can be achieved without coronoid fixation if a coronoid process fracture does not involve anteromedial facet or the fracture is less than 50% of the coronoid height. These findings are in line with the conclusions drawn by our study, as no functional benefit could be detected when the coronoid was surgically repaired. Similar findings have also been reported by Papatheodorou et al.[10], who found very good functional outcome scores [mean Broberg and Morrey score of 90 (range 70–100) and an average DASH score of 14 (range 0–38)] without any residual instability in a series of 14 terrible triad injuries with Reagan-Morrey type I or II coronoid fractures, that were treated without fixation. Although this study lacks a comparison group, a higher number of cases and a more specific classification system, they suggested a non-operative treatment for these lesions barring intraoperative elbow stability can be achieved after reconstruction of a radial head fracture and the LCL complex. In the most recent study, Kim et al. [48] presented the outcomes of 24 patients, that were treated by a single surgeon with a standardized surgical protocol, including no reconstruction of coronoid (Reagan-Morrey) type I and II fractures. At the final follow-up, the mean MEPS and Quick-DASH score were 91.5 and 17.3, respectively, with no recurrent instability. Although no comparison group has been established, the authors conclude, that Reagan-Morrey type I and II coronoid fractures in terrible triad injuries do not need to be fixed if the radial head and ligamentous complex are completely reconstructed. Moreover, data of our study suggest even worse outcome scores and ROM, when the coronoid tip had been reconstructed (although not statistically significant). We suppose, that fixation of the coronoid tip (and its attachment of the anterior capsule) could lead to more “stiff” elbow postoperatively, which temporarily or permanently limits elbow function in the rehabilitation process, resulting in a greater loss of ROM in this group. Likewise, a higher rate of severe joint stiffness and surgical arthrolysis procedures were performed in the fixation group. However, in this context group, heterogeneity also has to be taken into consideration, as significantly more RHA had been implanted in the (coronoid) surgical intervention group, which was previously reported to be associated with a significantly inferior outcome in the context of a TTI [20, 49]. However, when stratified for radial head treatment, results between both treatment groups remained the comparable.

Although we detected no significant differences in injuries or treatment procedures to further structures around the elbow between each treatment group, the inferior outcomes following surgical treatment of the coronoid might also be related to a more severe injury mechanism. This might be evidenced by greater number of radial head arthroplasties, indicating a more severe (soft) tissue damage. Additionally, HO itself is a severe complication of elbow trauma that shows its highest prevalence in elbow dislocations [50, 51]. It frequently leads to functional impairment of the elbow, as also seen in our study population. Similarly, OA was frequently detected in our patients, and was associated with significantly inferior outcomes. However, both treatment groups showed no statistical difference in the occurrence of this disease, which is why we could not determine the extent to which treatment to the coronoid (occult microinstability) or other structures (radial head) or even the initial injury (occult cartilage lesions) affected the outcome.

Taking these aspects into consideration, treatment of the coronoid tip fractures seems to play a rather minor role for final outcome of TTI compared to the treatment of the radial head and ligamentous structures [49, 52,53,54], as long as a stable elbow joint has intraoperatively been achieved. Nevertheless, recognition of all coronoid lesions and structured treatment is mandatory to minimize complications like OA or HO following TTI.

This study has some limitations: patients were not randomized to any particular treatment method; thus, surgeon and selection bias may have affected the results. However, not only the surgeon, but also the severity of the injury itself might have influenced the choice for a certain treatment method, with surgical treatment of the coronoid more likely being performed in cases where the radial head had to be replaced, surrogating for a higher energy injury or a better surgical access. However, the study population represents, by far, the largest comparative study on coronoid tip fractures in context of TTI in the current literature. Furthermore, no substantial group differences, such as age, sex distribution, or mean time from injury to operation could be detected, suggesting enrolment occurred without bias. Ultimately, the development of late complications, such as late onset post-traumatic arthritis could not be detected because of the mid-term design of this study. However, a review of the literature indicates that a follow-up of two years is considered sufficient for the assessment of fracture healing, stability, range of motion, and early surgical complications in fracture pattern and treatment studies.

Conclusion

Recognition of all associated injury patterns helps to provide optimal treatment strategies in terrible triad lesions of the elbow. This leads to sufficient elbow stability and good outcome scores in the majority of cases. While bias in treatment allocation and group heterogeneity cannot be completely omitted, our clinical and statistical analysis detected a beneficial effect of non-surgical treatment of O’Driscoll type 1 coronoid fractures on the midterm outcome, especially in terms of elbow function, ROM, incidence of arthrosis and postoperative complications, although it did not reach statistical significance. Therefore, future research including randomized-controlled trials is needed in order to improve treatment recommendations in these complex injuries.