Classifications in Brief: Mayo Classification of Olecranon Fractures

History

Olecranon fractures are relatively common in adults. The various fracture patterns differ in their mechanical characteristics and how they might best be treated surgically [4]. Colton [6] developed a classification system of olecranon fractures in 1973. The Colton classification was based on fracture morphology, mechanism of injury, and stability of the ulnohumeral joint. The AO classification was introduced in 1987 as a systematic way to classify long bone fractures based on location of the fracture line and degree of articular comminution [5, 16]. In this system olecranon fractures were grouped with fractures of the radial head and neck. Cabanela and Morrey [5] first introduced the Mayo classification in 1993 to provide a simplified classification of olecranon fractures based on fracture comminution, displacement, and stability of the ulnohumeral joint.

Purpose

The Mayo classification [5] provides a treatment algorithm and prognosis based on fracture type [5, 13, 15]. It was meant to simplify the classification of olecranon fractures into a practical system that is clinically useful [15]. Type I injuries typically are treated nonoperatively with a short period of immobilization followed by gradually increasing active ROM exercises [16]. Type II fractures with significant comminution can be considered for excision and triceps advancement in elderly, low-demand individuals [16]. Type IIA injuries can be treated with tension band wiring [13, 16]. Most Type IIB and Type III injuries should be treated with rigid fixation with a plate and screw construct provided the patient is healthy enough for surgery [13, 16].

Description

The Mayo classification divides olecranon fractures into Types I to III based on stability and the displacement seen on plain radiographs (Fig. 1). Type I represents 5% of all fractures and are nondisplaced [16]. Mayo Type II fractures account for 80% to 85% of olecranon fractures and are displaced with a stable ulnohumeral joint indicating intact ligamentous structures, particularly the anterior portion of the medial collateral ligament [13, 14]. Type III injuries have a displaced olecranon fracture with an unstable ulnohumeral joint and torn collateral ligaments [12, 13, 15]. Each type is further subdivided into A, noncomminuted, and B, comminuted [5, 13, 15].

Fig. 1.

in the Mayo classification of olecranon fractures, each type is further subclassified to indicate noncomminuted (A) and comminuted (B) fractures. Reproduced with permission from Morrey BF, Sanchez-Sotelo J, Morrey ME. Morrey's the Elbow and Its Disorders. 5th ed. Philadelphia, PA, USA: Elsevier Inc; 2017.

In the initial description of this classification, Types I, II, and III prognoses were excellent, good, and guarded, respectively [5, 13]. A good prognosis was expected with Type II injuries as a result of the maintained ligamentous structures and prior studies, which had shown 97% good to excellent functional outcomes in patients treated with tension band wiring for isolated, displaced olecranon fractures [23]. In addition to the substantial displacement and comminution seen with Type III injuries, they often have associated injuries including radial head fractures, coronoid fractures, and complex instability [14]. The associated injuries and instability may result in worse functional outcomes than Types I and II injuries [18]. An important aspect of this classification is the differentiation between a Type II and Type III injury. In a Type III injury, the ulnohumeral joint is dislocated indicating that the collateral ligaments are torn. In such injuries, rigid fixation of the olecranon is imperative to restore stability of the ulnohumeral joint [15].

Validation

Two studies of which we are aware have assessed the reproducibility of the Mayo classification in comparison to the Colton, Schatzker, and AO classifications [3, 21]. The referenced studies assessed agreement using a κ coefficient rather than observed agreement. This coefficient assesses the level of agreement between observers that is beyond agreement that would occur by chance alone. Kappa values range from 0 to 1 with a value of 0 denoting agreement at the level of random chance and 1 representing complete agreement [11]. Kappa values of ≤ 0.5 are considered poor, 0.51 to 0.74 good, and ≥ 0.75 an excellent level of agreement [20]. Interobserver agreement refers to the agreement between different observers and intraobserver agreement is the measure of repeated agreement of the same observer at different time points. These studies showed an interobserver agreement of the Mayo classification system of κ = 0.19 [3] and κ = 0.32 [19], which represent poor reliability, with interobserver agreement being only marginally greater than chance alone.

By contrast, there was moderate intraobserver reliability in one of these two studies, at κ = 0.64 [21], whereas the other reported poor intraobserver agreement for specialists (κ = 0.18) and moderate among nonspecialists (κ = 0.51) [3].

When comparing between classifications, these studies showed conflicting results when the Mayo classification was compared with the Colton, Schatzker, and AO classifications. Benneton et al. [3] showed the worst interrater agreement for the Mayo classification with a κ coefficient of 0.19 compared with κ = 0.67 for the Colton classification. Tamaoki et al. [21] found the Mayo classification to have the highest intrarater agreement (κ = 0.64) and second highest interrater agreement (κ = 0.32) compared with the other classification systems of olecranon fractures. Despite the relative improvement in reproducibility shown in this study, the interrater reliability remained poor.

Both existing studies showed that the Mayo classification has poor reproducibility indicating low reliability [3, 21]. The reproducibility of the Mayo classification may be worse than other classification systems for olecranon fractures [3, 21]. Although one might expect interrater reliability to be improved with a simpler classification, in this instance, that did not occur.

Despite concerns about its reliability, the Mayo classification is frequently used in outcome-based studies [1, 2, 7-10, 17, 18, 22]. However, the use of a classification system with poor reproducibility will generate confusing results when patient outcomes are assessed, and that seems to be exactly what has occurred in studies that used the Mayo classification. For example, several studies have compared Mayo Type II and Type III fracture patterns treated with plate fixation. Although one might anticipate a worse outcome with Type III injuries, in fact most of these studies found no substantial differences in outcomes between the two groups [2, 7, 10]. It seems possible that this may be at least partially attributed to misclassification. Although another study found decreased postoperative ROM in patients with Mayo Type III injuries, 95% of the patients in that series were treated with a tension band construct, which may not have provided adequate fixation for a Type III injury [18].

Limitations

Despite frequent use in both research and clinical settings, the Mayo classification has several limitations. Most importantly, as already noted, its reliability is poor. That alone should cause the reader to consider not using it in practice, and that has been the suggestion of both studies that evaluated the reliability of the Mayo classification [3, 21].

Static radiographs alone may provide insufficient information to determine the instability requisite of a Type III injury. Type III injuries generally include ligamentous instability and have a high risk of associated injuries about the elbow, including coronoid and proximal radius fractures. CT may be used to further elucidate the fracture pattern when it is necessary to determine treatment. A recent study used three-dimensional CT to identify fracture morphology, comminution, and articular surface involvement based on fracture type using the Mayo classification [12]. The authors showed that patients with Type III injuries had larger proximal fragments indicating loss of stability conferred by the ulnohumeral joint. This technology, in conjunction with traditional radiographs, may provide a means for more reproducible classification of olecranon fractures, but future studies are needed to confirm or refute this contention.

Conclusions

The Mayo classification was devised as a practical means to classify olecranon fractures based on displacement, comminution, and stability. Although seemingly intuitive, the use of this classification in research and the clinical setting may create confusion resulting from poor reproducibility. Despite frequent use in outcome-based studies, the only two studies of which we are aware that have tried to validate the Mayo classification system for olecranon fractures found it inadequate for this purpose [3, 21]. Studies evaluating treatment outcomes based on Mayo type have generally shown no differences in outcomes between fracture types [2, 7, 10], which could represent a problem for the prognostic schema its authors initially proposed [5], but could as easily be a manifestation of users experiencing problems of reliability with the classification system. The existing validation studies demonstrate that although the Mayo classification was designed to simplify categorization of olecranon fractures, it does not achieve this goal [3, 21]. It is possible that the use of CT may improve the intra- and interrater reliability. Future studies should incorporate use of CT in addition to plain radiographs to assess the reliability of the Mayo classification.