Benjamin Kopp MD1 and David Ring MD PhD1 Department of Surgery and Perioperative Care, Dell Medical School – The University of Texas at Austin, Austin, TX, USA Surgeons and patients hope that more detailed imaging will better guide management and lead to improved outcomes. Due to complex three‐dimensional (3D) anatomy of proximal humerus fractures, some can be challenging to characterize on radiographs. CT scans are commonly used to get a more detailed image of the fracture pattern, which could potentially lead to better tailored treatments and improve outcomes. Two prospective studies have evaluated the reliability and accuracy of classification using radiographs and CT scans (level II).1,2 One study assessed interobserver reliability of AO classification between 2D and 3D CT scans (level III).3 Multiple retrospective studies studied inter‐ and intraobserver reliability regarding specific fracture characteristics (level III)4 and fracture classification with AO classification (level III)5 or Neer classification (level III).6 In a prospective diagnostic study three observers classified 44 consecutive fractures and found better assessment of relevant structures (tuberosities, the glenoid, and humeral head) using a four‐grade scoring system (1 = excellent, 2 = good, 3 = fair, and 4 = inadequate) based on CT compared to radiographs (AP view, scapular Y‐views, and axillary views) independent of fracture severity (i.e. Neer two‐, three‐, or four‐part fractures) (p <0.05) (level II).1 A study comparing 2D to 3D CT scans found higher, but still only fair, interobserver reliability for 3D CT scans regarding displacement of the greater tuberosity (κ = 0.35 vs 0.30, p <0.001).3 In a retrospective study three3 observers evaluated 40 nonconsecutive fractures and found better assessment of fracture displacement, impaction, and anatomic neck involvement with CT compared to radiographs, but no influence on AO classification or decision on whether to operate (level III).5 In another retrospective study four observers evaluated 40 consecutive fractures and documented better interobserver reliability by adding CT: “moderate” for radiographs (κ = 0.42) and 2D CT (κ = 0.56) to “good” for 2D CT with 3D volume renderings (κ = 0.76) for the Neer classification system intraobserver reliability improved (p <0.001) from “moderate” for radiographs (κ = 0.48) and 2D (κ = 0.63) to “excellent” for 2D CT with 3D volume renderings (κ = 0.84) for the Neer classification (level II).2 In another study, seven observers evaluated 40 consecutive fractures and compared 2D versus 3D CT with no difference in intraobserver reliability.6 There was a difference in interobserver reliability among the junior resident observers (level III). This difference amongst more junior surgeons was also reflected in a prior study comparing 2D and 3D CT scans where less experienced surgeons had significantly higher levels of agreement with 3D (κ = 0.14) CT than 2D (κ = 0.17) (p = 0.014).3 In another study, three observers reviewed 20 fractures with no difference in diagnosis of fracture characteristics (existence of medial hinge, metaphyseal extension) on radiographs compared to 2D CT (level III).4 Immediate initiation of exercises might result in better final shoulder motion but could also theoretically interfere with healing. Controversy exists between early (within one week) or late (three weeks or greater, once healing is underway) initiation of exercises after a proximal humerus fracture. Most fractures of the proximal humerus are adequately aligned, stable, and associated with limited functional impairment after nonoperative treatment.7,8 Many of these have good results treated in a simple arm sling. Due to the impact on activities of daily living in the generally older patient population, early return of function is important for maintaining independence.9 Three prospective randomized controlled trials (RCTs) (level II) have studied the effect of mobilization within one week of injury on multiple patient‐reported outcome measures (PROMs).9–11 Two prospective RCTs (160 patients) with methodological limitations (possible bias; allocation concealment unclear, some blinding outcome assessors, blinding patients impossible, inclusion/exclusion criteria not clearly defined) showed that early mobilization within one week resulted in significantly better Constant Shoulder Scores at 12 weeks9 for impacted (stable) proximal humeral fractures (weighted mean difference [MD]: 9.9; 95% confidence interval [CI]: 2.1–17.7; p <0.05) and 16 weeks10 for non‐ and minimally displaced two‐part fractures (MD: 16.0; 95% CI: 7.1–24.9; p <0.001). There were no significant differences at six months9 (MD: 6.1; 95% CI −0.2–12.4, p = 0.06) and one year10 (MD: 7.0; 95% CI: −3.4–17.4; p = 0.19) after fracture (level II).
80
Proximal Humeral Fractures
Clinical scenario
Top three questions
Question 1: In patients with a proximal humerus fracture, does adding CT imaging improve classification of fractures or improve patient outcomes compared with radiographs alone?
Rationale
Clinical comment
Available literature and quality of the evidence
Findings
Resolution of clinical scenario
Question 2: In patients choosing nonoperative treatment of a fracture of the proximal humerus, does early initiation of exercises (before one week) improve pain or patient‐reported function compared with delayed exercise programs (after three weeks)?
Rationale
Clinical comment
Available literature and quality of the evidence
Findings
Shoulder function