Authors: Kazuya Tamai, Hiroyasu Mizuhara, Hideaki Asai, Yuji Yamaguchi, Yuichi Nagase, Sakae Tanaka
Categories: Full Length Article, Proximal humeral fracture, Three-dimensional computed tomography, Shield fracture, Bicipital groove, Greater tuberosity fracture, Supraspinatus tendon insertion
Source: JSES Reviews, Reports, and Techniques
Authors: Kazuya Tamai, Hiroyasu Mizuhara, Hideaki Asai, Yuji Yamaguchi, Yuichi Nagase, Sakae Tanaka
Although the Neer classification of proximal humeral fractures typifies fracture anatomy, clinical practice often presents cases in which distinguishing between fracture types can be challenging. Moreover, many atypical fractures, including shield fractures, have been documented. These indicate that the fracture anatomy of the tuberosities is not as simple as Neer's description, particularly in multifragmentary fractures. We hypothesized that three-dimensional computed tomography (3DCT) could provide a comprehensive view of the fractured tuberosities in multifragmentary proximal humeral fractures.
A retrospective study was conducted on 80 patients, aged 34 to 94 years, who sustained multifragmentary proximal humeral fractures. The initial diagnosis of the surgeons identified 37 three-part fractures, 16 four-part fractures, 19 four-part valgus-impacted fractures, and 8 fracture-dislocations, according to the Neer classification. Pretreatment volume rendering 3DCT scans, including a superior view, were reviewed by 2 orthopedic surgeons. Tuberosity fractures were categorized as single tuberosity pattern (displacement of either the greater or the lesser tuberosity), dual tuberosity pattern (separate displacement of both tuberosities), or shield pattern (displacement of both tuberosities as well as the bicipital groove, encircling the humeral head).
The interobserver reliability for categorizing tuberosity fracture patterns was 0.725 in terms of the Cohen's kappa. Tuberosity fractures were identified as a single tuberosity pattern in 41% of cases, a dual tuberosity pattern in 13%, and a shield pattern in 36%. Most of the three-part fractures had a single tuberosity pattern, while all four-part valgus impacted fractures had a shield pattern. Detailed observation of the single tuberosity patterns revealed that, in 68% of cases, the fracture line was located 5 to 10 mm posterior to the anterior margin of the greater tuberosity, leaving the anteriormost portion of the greater tuberosity unfractured.
3DCT clearly demonstrated 3 patterns of tuberosity single, dual, and shield patterns. A shield pattern, an anatomical neck fracture laterally, was common among multifragmentary proximal humeral fractures. In greater tuberosity fractures, the fracture line was located posterior to the supraspinatus tendon insertion in many instances, while a fracture of the entire greater tuberosity was uncommon.
The Neer classification, a four-segment system for categorizing proximal humeral fractures,^13^^,^^14^ is renowned for its simplicity, comprehensibility, and widespread clinical application. This classification serves as a conceptual framework or mental image,^14^ guiding surgeons in fracture classification based on typical pathoanatomical patterns. However, clinical practice often presents cases in which distinguishing between three-part, four-part, or four-part valgus-impacted fractures can be challenging. Moreover, some instances exhibit combined displacement of greater and lesser tuberosities without separation, or a displacement of the greater tuberosity with the lesser tuberosity staying at the upper end of the shaft, resulting in a “three-part” fracture that is clinically equivalent to a four-part fracture in terms of the head lacking soft tissue attachments.^17^^,^^19^ In addition, shield fractures are reported to be common.^1^^,^^3^ The “shield” refers to a bone segment encircling the head fragment, comprising both tuberosities held together with the bicipital groove.^1^ Taken together, these indicate that the fracture anatomy of the tuberosities is not as simple as Neer's description in multifragmentary proximal humeral fractures.
The management of the tuberosities is of clinical importance. Improper handling can lead to rotator cuff dysfunction, resulting in shoulder motion loss, muscle weakness, and humeral head instability. In proximal humeral fractures, particularly multifragmentary fractures, it is essential to accurately identify the anatomy of the fractured tuberosities to develop an effective treatment plan. We hypothesized that three-dimensional computed tomography (3DCT) could provide a comprehensive view of the fractured tuberosities in multifragmentary proximal humeral fractures.
This study retrospectively examined multifragmentary proximal humeral fractures treated at 3 medical centers between 2011 and 2024. Multifragmentary fractures were defined as those in which the proximal humerus was divided into 3 or more segments. We excluded patients with one- or two-part fracture, open fracture, associated scapula, clavicle or humeral shaft fracture, polytrauma, pathological fracture, and previous shoulder surgery. The research enrolled 80 individuals in which pretreatment 3DCT scans were available. The study population comprised 22 men and 58 women, aged 34 to 94 years (mean, 70 years) at the time of injury. There were 37 three-part, 16 four-part, 19 four-part valgus-impacted fractures, and 8 fracture-dislocations (three- or four-part), according to the Neer classification judged by the surgeons at the time. All patients had a surgical neck fracture. The injuries affected 44 right and 36 left shoulders. Treatment methods included conservative treatment in 7, osteosynthesis in 37, hemiarthroplasty in 20, and reverse shoulder arthroplasty in 16 cases.
This study was approved by the Ethics Committee of Tohto Bunkyo Hospital #2502, including anonymity for the patients and publication of the data.
CT scans were performed at the 3 institutions using 2-mm slices without gaps. Volume rendering 3D reconstructions were viewed from anterior, posterior, and superior angles on a picture archiving and communication system viewer (WeVIEW-Z edition, FUJIFILM Medical Co., Ltd., Tokyo, Japan). Two shoulder surgeons (K.T. and H.M.), with 40 and 10 years of orthopedic experience, respectively, independently assessed the 3DCT images to determine the anatomy of the fractured tuberosities. The interobserver reliability was calculated using the kappa statistic. If the 2 observers agreed, the fracture type was adopted. If they did not agree, discrepancies were resolved through multiple discussions to reach a consensus on the most appropriate fracture pattern. To measure the true distances of the fracture line from the bony landmarks, a reference ruler on the picture archiving and communication system viewer was used.
Tuberosity fractures were categorized as single tuberosity, dual tuberosity, or shield patterns. A single tuberosity pattern was identified by displacement ( ≥ 1 cm) of either the greater or the lesser tuberosity with the bicipital groove remaining intact, which corresponded to the Neer's three-part fracture. A dual tuberosity pattern was identified by separate displacement ( ≥ 1 cm) of both greater and lesser tuberosities, with the bicipital groove remaining intact, which corresponded to the Neer's four-part fracture. The diagnosis of shield pattern was determined when both tuberosities, as well as the bicipital groove, were fractured off and encircled the humeral head like a shield. The shield pattern thus differed from the dual tuberosity pattern in that the bicipital groove was fractured and emerged from its base.
The first-time image review for diagnosing tuberosity fracture patterns yielded a concordance rate of 82%. The interobserver reliability was 0.725 in terms of the Cohen's kappa.
Of a total of 80 cases, 33 (41%) were categorized as the single tuberosity pattern, with 31 as a greater tuberosity fracture (Fig. 1 A, B) and two as a lesser tuberosity fracture. Ten (13%) cases were categorized as the dual tuberosity pattern (Fig. 2). Twenty-nine (36%) cases were categorized as the shield pattern (Fig. 3 A, B). Of the 8 (10%) anterior fracture-dislocations, one was associated with the shield pattern.Figure 13DCT images of the single tuberosity pattern. (A) A 69-year-old woman. The posterior two-thirds of the greater tuberosity is fractured. The superior facet (white**arrow) remains attached to the humeral head segment, along with the bicipital groove (G) and the lesser tuberosity. (B) Another 69-year-old woman. The entire greater tuberosity, inclusive of the superior facet (black**arrow), is fractured, while the bicipital groove (G) and the lesser tuberosity remain intact. 3DCT, three-dimensional computed tomography.Figure 23DCT image of the dual tuberosity pattern. A 77-year-old woman. The greater and lesser tuberosities (arrowheads) are displaced separately, with the bicipital groove remaining intact. This is consistent with the four-part fracture described by Neer. G, bicipital groove; 3DCT, three-dimensional computed tomography.Figure 33DCT images of the shield pattern. (A) A 73-year-old woman. The greater tuberosity, the bicipital groove (G), and the lesser tuberosity are fractured, forming a protective barrier around the humeral head. Note that the superior facet of the greater tuberosity (arrowheads) is not separated from the bicipital groove segment, while a fracture line traverses the greater tuberosity posterior to the superior facet (arrow). (B) An 83-year-old woman. The injury presents with a shield pattern involving the tuberosities and the bicipital groove. The greater and lesser tuberosities are displaced separately, resembling the dual tuberosity pattern, but the bicipital groove is also fractured off in this patient. 3DCT, three-dimensional computed tomography.
Table I shows the tuberosity fracture patterns of each fracture type initially diagnosed with the Neer classification. Most of the fractures diagnosed as three-part initially had a single tuberosity fracture pattern, while some had a dual or a shield pattern. Fractures diagnosed as four-part initially had a dual pattern or a shield pattern. All the fractures diagnosed as four-part valgus impacted fractures initially had a shield pattern of tuberosity fractures.Table ITuberosity fracture patterns in relation to primary diagnoses.# of fracture types (Neer) as a primary diagnosis# of tuberosity fracture patterns identified with 3DCTThree-part, 37Single, 33Dual, 2Shield, 2Four-part, 16Dual, 7Shield, 9Four-part valgus impacted, 19Shield, 193DCT, three-dimensional computed tomography.Fracture-dislocations are excluded.
A detailed observation of the fracture line of the greater tuberosity was performed and summarized in Fig. 4. In 21 (68%) of the 31 cases of single tuberosity pattern involving the greater tuberosity (corresponding to Neer's three-part greater tuberosity fractures), the fracture line was noted to be situated 5 to 10 mm posterior to the anterior margin of the greater tuberosity (Figs. 1A and 4A). In these instances, the anterior portion of the greater tuberosity remained unfractured and attached to the humeral head segment along with the bicipital groove. On the other hand, in 7 (23%) cases, the fracture line extended up to the anterior margin of the greater tuberosity, encompassing the lateral wall of the bicipital groove, so the entire greater tuberosity was fractured (Fig. 1B). In the remaining 3 (9.7%) cases with a greater tuberosity fracture, the precise location of the fracture line was not determined due to the illegibility of the 3DCT images.Figure 4Representative fracture line locations in tuberosity fractures. Representative fracture lines (solid black lines) for each fracture pattern are illustrated in a bone model. (A) In single tuberosity patterns, the fracture line typically traverses the greater tuberosity posterior to the supraspinatus facet. (B) In dual tuberosity patterns, the fracture line of the greater tuberosity is often located posterior to the supraspinatus facet. (C) In shield patterns, the fracture line separates the lesser tuberosity, the bicipital groove, and the greater tuberosity from the humeral head. Not infrequently, the fracture line traverses the greater tuberosity posterior to the supraspinatus facet (dotted black line). G, bicipital groove.
In 4 (44%) of the 9 dual tuberosity patterns (corresponding to Neer's four-part fractures), the fracture line of the greater tuberosity existed several millimeters posterior to the lateral wall of the bicipital groove. This resulted in the anterior portion of the greater tuberosity remaining attached to the humeral head (Figs. 2 and 4B). Similarly, in 16 (53%) of the 30 shield patterns, the greater tuberosity was found to be split several millimeters posterior to the lateral wall of the bicipital groove, leaving the anterior portion of the greater tuberosity attached to the bicipital groove segment (Figs. 3A and 4C).
This study demonstrated that proximal humeral fractures initially diagnosed as three-part, four-part, or four-part valgus-impacted fractures of Neer had various patterns of tuberosity single (41% of cases), dual (13%), and shield (36%) patterns. Most of the fractures diagnosed as three-part initially had a single tuberosity fracture pattern as expected. Fractures diagnosed as four-part initially had a dual pattern or a shield pattern. All the fractures diagnosed as four-part valgus impacted initially had a shield pattern of tuberosity fractures. Furthermore, a detailed observation of the single tuberosity patterns revealed that, in 68% of cases, the fracture line was located 5 to 10 mm posterior to the anterior margin of the greater tuberosity, leaving the anteriormost portion of the greater tuberosity remained unfractured and attached to the humeral head segment.
It should be noted that the shield pattern of tuberosity fracture is common, accounting for 36% of cases. Recognition of the shield pattern was possible with 3DCT, particularly with the superior view, which clearly revealed the presence or absence of a fracture line behind the bicipital groove, thereby establishing that the lesion is unambiguously distinct from the single or the dual pattern. The high prevalence of the shield pattern may be associated with the common injury mechanism, a fall on the outstretched hand. An experiment with this parachute reflex simulation produced minimally displaced greater tuberosity fractures, head-splitting fractures, or shield fractures in a cadaveric bone model with the joint capsule left intact.^11^ Another experiment observed that the initial impaction of the head segment against the glenoid formed a fracture of the head and the tuberosities.^9^ The articular margin of the humeral head is known to have a lower cortical thickness,^10^ so head-to-glenoid impaction in a fall on the outstretched hand can result in a shield pattern tuberosity fracture.
The shield pattern is an anatomical neck fracture laterally, potentially resulting in humeral head ischemia. The ascending branches of the anterior humeral circumflex artery enter the bone at the lateral wall of the bicipital groove 12 mm distal from the superior apex of the greater tuberosity.^4^^,^^8^ Another study demonstrated that these branches enter the bone on the medial surface of the lateral wall of the bicipital groove 10 mm distal from the superior apex of the greater tuberosity.^18^ These ascending vessels along the bicipital groove are very likely to sustain damage in shield pattern tuberosity fractures. Nevertheless, osteonecrosis of the humeral head is rarely documented in conservatively managed shield fractures.^2^ This may be due to preserved blood flow from the posterior circumflex artery, which has several ascending branches^7^ that enter the bone more distally (20 mm from the superior apex of the greater tuberosity^8^) than those of the anterior circumflex artery. Or, more importantly, the blood flow of the humeral head is highly dependent on the medial calcar condition, that is, whether the calcar is attached to the humeral head or to the shaft^15^^,^^16^ and how long the remaining calcar is.^6^ Therefore, in the presence of a shield pattern of tuberosity fractures, it is imperative to pay close attention to the medial calcar for estimation of the humeral head vascularity.
Another noteworthy finding in the present study is that the anteriormost part of the greater tuberosity is preserved in the majority of the single tuberosity pattern (three-part fractures). Again, this finding was most evident in the superior view of the 3DCT image. Similar 3DCT images have been documented in the literature.^2^ This anteriormost part of the greater tuberosity that escaped fracture corresponds to the supraspinatus tendon footprint.^12^ Hasan and colleagues reported that the fracture lines involving the greater tuberosity were most often between the supraspinatus and infraspinatus tendon footprints.^5^ This may be attributed to the fact that the greater tuberosity has a lower cortical thickness than the bicipital groove or the lesser tuberosity^10^ and that the footprint protected by the tendon is less likely to fracture. Sparing of the anteriormost part of the greater tuberosity was also observed in dual tuberosity and shield patterns. Thus, it should be remembered that, in many instances, the supraspinatus tendon footprint remains connected with the bicipital groove-lesser tuberosity complex. Clinically, it is crucial to prevent accidental stripping of the supraspinatus tendon during surgical intervention for these fractures. Additionally, in fracture reduction procedures, care must be taken not to pull down the greater tuberosity fragment excessively downward due to the false assumption that it is displaced upward by the pull of the supraspinatus.
The primary strength of this study lies in its clear demonstration of the previously ambiguous anatomy of tuberosity fractures. A shield pattern proved to be common. The tuberosity fracture anatomy varied even for fractures that fall into the same category of the Neer classification. The second strength of this study is that it identified the location of the fracture line of the greater tuberosity, thus provided information regarding the possible direction of displacement of the fractured segment.
The present study has several limitations. The number of cases was rather small, and only cases with access to 3DCT were included, so the study population was potentially biased. Secondly, observations were focused on the tuberosity fractures and not directed to the medial calcar, which is likely more deeply involved in osteonecrosis of the humeral head. Furthermore, the retrospective observational design precluded the determination of an appropriate treatment method for each fracture pattern. In the future, a prospective study must be conducted in a larger number of cases, investigating the relationship between pretreatment 3DCT findings and functional prognosis.
3DCT of multifragmentary proximal humeral fractures revealed various patterns of tuberosity fracture. A shield pattern, an anatomical neck fracture laterally, accounted for over one-third of cases. In more than a half of greater tuberosity fractures, the fracture line was located posterior to the supraspinatus tendon insertion.
Funding: No funding was disclosed by the authors.
Conflicts of The authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.