Innovations in the Management of Displaced Proximal Humerus Fractures

Shane J. Nho, MD Abstract Robert H. Brophy, MD The management of displaced proximal humerus fractures has Joseph U. Barker, MD evolved toward humeral head preservation, with treatment Charles N. Cornell, MD decisions based on careful assessment of vascular status, bone quality, fracture pattern, degree of displacement, and patient age John D. MacGillivray, MD and activity level. The AO/ASIF fracture classification is helpful in guiding treatment and in stratifying the risk for associated disruption of the humeral head blood supply. Nonsurgical treatment consists of sling immobilization. For patients requiring Dr. Nho is Resident, Department of surgery, options include closed reduction and percutaneous Orthopaedic Surgery, Hospital for fixation; transosseous suture fixation; open reduction and internal Special Surgery, New York, NY. Dr. Brophy is Resident, Department of fixation, with either conventional or locking plate fixation; bone Orthopaedic Surgery, Hospital for graft; and hemiarthroplasty. Proximal humerus fractures must be Special Surgery. Dr. Barker is Resident, evaluated on an individual basis, with treatment tailored according Department of Orthopaedic Surgery, to patient and fracture characteristics. Hospital for Special Surgery. Dr. Cornell is Attending Orthopaedic Surgeon, Department of Orthopaedic Surgery, Hospital for Special Surgery. Dr. roximal humerus fractures are ticular fractures5-11 and in locking MacGillivray is Assistant Attending Pincreasingly common in socie- plate technology,8,12 which are par- Orthopaedic Surgeon, Department of ties with maturing populations.1 ticularly relevant to the care of these Orthopaedic Surgery, Hospital for These fractures are not simple to fractures.7,13-15 Locking plate tech- Special Surgery. treat. A variety of options exists; nology and the use of osteobiologics None of the following authors or the however, outcomes are less than ide- may become increasingly important departments with which they are al in many patients.2,3 Most proxi- in the management of displaced affiliated has received anything of value mal humerus fractures are either proximal humerus fractures, facili- from or owns stock in a commercial nondisplaced or minimally displaced tating humeral head preservation in company or institution related directly or and can be treated nonsurgically.4 appropriately selected patients. indirectly to the subject of this article: Nonsurgical options focus on early Dr. Nho, Dr. Brophy, Dr. Barker, Dr. functional exercises with the goal of Epidemiology Cornell, and Dr. MacGillivray. achieving a functionally acceptable Reprint requests: Dr. MacGillivray, range of motion (ROM). For the 15% Proximal humerus fracture is the Department of Orthopaedic Surgery, to 20% of displaced proximal hu- second most common fracture of the Hospital for Special Surgery, 535 East merus fractures that may benefit upper extremity, following distal 1 70th Street, New York, NY 10021. from surgery, no single approach is forearm fracture. In people older considered to be the standard of care. than age 65 years, fracture of the J Am Acad Orthop Surg 2007;15:12-26 Surgeons should be familiar with the proximal humerus is the third most Copyright 2007 by the American different treatment options avail- common fracture, after hip fracture Academy of Orthopaedic Surgeons. able, including recent advances in and Colles’ fracture.1 Proximal hu- the management of complex periar- merus fracture is associated with

12 Journal of the American Academy of Orthopaedic Surgeons Shane J. Nho, MD, et al significant morbidity, leading to Figure 1 functional impairment lasting at least 3 months.16 Displaced proximal humerus fractures generally re- sult in long-term functional disabil- ity. This type of injury usually is sustained after a moderate-energy fall in an individual with low bone density.16

Anatomy The proximal humerus is divided into the shaft, surgical neck, anatomic neck, lesser tuberosity, and greater tuberosity. Classification of proximal humerus fractures based on these anatomic divisions predi- cates treatment and predicts outcome. The humeral head has a spherically shaped articular surface; the anatomic neck demarcates the junction of the head and metaphy- sis. The neck-shaft angle is 130°, and the head is retroverted 19° to 22° relative to the shaft.17 The greater and lesser tuberosities are located adja- cent to the articular margin of the head; they are separated by the bicipital groove. The surgical neck serves as the junction between the metaph- ysis and the humeral diaphysis. Hertel radiographic criteria. A and B, Metaphyseal head extension is a radiographic Tendinous insertions contribute measurement of the articular fragment from the head-neck junction to the inferior to the pattern of fracture displace- extent of the medial cortex. A, Metaphyseal head extension >8 mm. B, Metaphyseal head extension <8 mm. C and D, Medial hinge integrity is determined by the ment around the proximal humerus continuity of the medial calcar. C, Undisplaced medial hinge. D, Medial hinge by transmitting deforming muscular displaced >2 mm. (Adapted with permission from Hertel R, Hempfing A, Stiehler forces to the bony fragments. The in- M, Leunig M: Predictors of the humeral head ischemia after intracapsular fracture of sertions of the supraspinatus, in- the proximal humerus. J Shoulder Elbow Surg 2004;13:427-433.) fraspinatus, and teres minor tendons onto the greater tuberosity contribute to the typical posterior and supe- tends to be very dense and strong, disrupted, implying, in the fractured rior retraction of this fragment. The thus providing a potential site for proximal humerus that perfusion rotator interval functions as a check- fracture fixation, it is important from the posterior circumflex hu- rein on the humeral head fragment when using suture fixation to re- meral artery alone may be sufficient and limits displacement of two-part member that the tendons are even for humeral head survival. Hertel et fractures and most three-part frac- stronger than the bone.18 al22 conducted perfusion studies in tures. Functionally significant tears The anterior circumflex humeral proximal humerus fractures to deter- of the rotator interval are uncom- artery and the arcuate artery are con- mine factors predictive of humeral mon. The pull of the subscapularis sidered to be the major sources of head ischemia. They demonstrated muscle tends to retract lesser tuber- humeral head perfusion,19 but stud- the significance of metaphyseal head osity fragments medially. When the ies characterizing the vascular anat- extension (ie, the amount of me- lesser tuberosity remains attached to omy were performed in the intact taphysis, typically posteromedial) the head fragment, the head frag- proximal humerus.20,21 Even in sim- that remains attached to the humer- ment is rotated internally. Although ple fracture patterns, the anterior al head determines the amount of the bone at the tendinous insertion circumflex humeral artery is often head extension22 (Figure 1, A). Meta-

Volume 15, Number 1, January 2007 13 Innovations in the Management of Displaced Proximal Humerus Fractures

Figure 2 angulation >45°), direction of dislocation, and involvement of the articular surface.4 The AO/ASIF classification system for proximal humerus fractures broadly groups fractures based on the degree of articular involvement and likelihood of vascular injury23 (Figure 2). Observer reli- ability and reproducibility for both the Neer and the AO/ASIF classification is fair to poor; it is unlikely that one orthopaedic surgeon will assign the same classification to a proximal humerus fracture at two separate times, just as it is unlikely that two orthopaedic surgeons will agree on a classification. Given these limita- tions, the AO/ASIF classification for the three basic types of humerus fracture is more user-friendly than the Neer classification,24 although still complex.

Treatment Algorithm Although the management of displaced proximal humerus fractures has evolved toward humeral head preservation, treatment should be guided by careful assessment of vascular status, bone quality, fracture pattern, and degree of comminution, as well as patient factors, such as age and activity level. Patients who are either medically unstable or inactive AO/ASIF classification for proximal humerus fractures. Top, A, Unifocal, extra- are poor candidates for surgery and articular two-part fracture with intact vascular supply. Middle, B, Bifocal extra- instead may be treated with sling articular fracture with possible injury to the vascular supply. Bottom, C, Articular immobilization until the fracture fracture involving the anatomic neck, with high likelihood of necrosis. (Adapted with permission from Müller ME: Appendix A: The comprehensive classification of heals. The ultimate goal is maxi- fractures of long bones, in Müller ME, Allgöwer M, Schneider R, Willenegger H mum shoulder function and mini- [eds]: Manual of Internal Fixation: Techniques Recommended by the AO-ASIF mal shoulder pain. A proposed treat- Group. Berlin, Germany: Springer-Verlag, 1991, pp 118-125.) ment algorithm (Figure 3) provides a working guideline for the management of proximal humerus fractures, physeal head extension <8 mm was humeral head ischemia.22 but treatment, whether nonsurgical found to be a good predictor of is- or surgical, must be individualized chemia (accuracy, 0.84).22 Another according to patient and fracture Classification predictor of ischemia was medial characteristics. hinge disruption >2 mm (accuracy, The Neer classification4 and the AO/ The likelihood of humeral head 0.79)22 (Figure 1, B). In this series, the ASIF classification23 are the most osteonecrosis is implicit in the combination of metaphyseal head widely used systems to evaluate and AO/ASIF classification; thus, deter- extension <8 mm with medial hinge determine treatment of proximal mining the AO/ASIF fracture type is disruption >2 mm as well as an ana- humeral fractures. The Neer classi- the initial step in determining the tomic neck fracture pattern had a fication is based on the number of probability of humeral head preser- positive predictive value of 97% for fracture parts (displacement >1 cm, vation (Figure 3). Type A is a unifo-

14 Journal of the American Academy of Orthopaedic Surgeons Shane J. Nho, MD, et al

Figure 3

Displaced proximal humerus fracture

AO/ASIF AO/ASIF AO/ASIF type A type B type C (unifocal) (bifocal) (anatomic neck)

<4 mm Cortex >4 mm Cortex <4 mm Cortex >4 mm Cortex <4 mm Cortex >4 mm Cortex

2-SN ORIF ORIF ORIF 2-SN 2-SN 2-SN (>66%) with with with (<66%) (>66%) (<66%) CRPF Sling Hemi LCP LCP Hemi LCP Sling TOSF Sling or ± BG ± BG ± BG ORIF

Treatment algorithm for proximal humerus fracture. The systematic approach includes assessing vascular status, assigning fracture type using the AO/ASIF classification, assessing bone quality (ie, measure cortex), and applying the Hertel criteria (ie, medial hinge [>2 mm], metaphyseal extension [<8 mm]). 2-SN (<66%) = two-part surgical neck fracture with <66% translation, 2-SN (>66%) = two-part surgical neck fracture with >66% translation, BG = bone graft or bone graft substitute, CRPF = closed reduction and percutaneous fixation, Hemi = hemiarthroplasty, LCP = locking compression plate, ORIF = open reduction and internal fixation, TOSF = transosseous suture fixation cal, extra-articular fracture with an first level. The combined cortical in the management of displaced intact vascular supply. Type B is a bi- thickness is the average of the medi- proximal humerus fractures.3 Court- focal, extra-articular fracture with al and lateral cortical thickness at Brown and colleagues26,27 recom- possible injury to the vascular sup- the two levels25 (Figure 4). Nonsurgi- mend 2 weeks of sling immobiliza- ply. Type C is an articular fracture cal treatment, suture fixation, and tion followed by physical therapy for involving the anatomic neck with a hemiarthroplasty may be the best patients with two-part surgical neck high likelihood of osteonecrosis. options for the patient with com- fractures27 and valgus-impacted frac- Cortical thickness of the humer- bined cortical thickness <4 mm. tures.26 Two-part proximal humerus al diaphysis is a more reliable and re- Cortical thickness ≥4 mm is consid- fractures with >66% translation producible predictor of both bone ered necessary for adequate screw were treated with either a sling or mineral density and the potential for purchase with standard internal fix- with internal fixation with flexible success of internal fixation than is ation.25 intramedullary nailing and tension- age.25 After adjusting for magnifica- band wires.26,27 No statistical differ- tion, medial and lateral cortical ence was reported between the Nonsurgical thickness is measured from the an- groups in terms of Neer score,4,28 re- Management teroposterior view of the proximal turn to activities of daily living, and humerus.25 The first level, the most Nonsurgical management tradition- fracture union.26,27 The data demon- proximal aspect of the humeral di- ally has been recommended for non- strate that the raw Constant score aphysis, occurs at the level in which displaced and minimally displaced (Table 1) deteriorates with advancing the endosteal borders of the medial proximal humerus fractures. Sling age and degree of displacement. and lateral cortices are parallel. The immobilization with or without However, when calculated based on second level is 20 mm distal to the closed reduction also has a role age-adjusted Constant score, the

Volume 15, Number 1, January 2007 15 Innovations in the Management of Displaced Proximal Humerus Fractures

FIgure 4 no. 2 or no. 5 nonabsorbable sutures are inserted into the supraspinatus and subscapularis tendons to obtain control of the head fragment. The head is reduced over the humeral shaft, and drill holes are made on either side of the bicipital groove to re- approximate the head to the shaft (Figure 5). With a three-part fracture, in order to stabilize the greater tuberosity fragment, the sutures in the fragment are passed through transosseous humeral shaft drill holes after the head is reduced to the shaft and the rotator interval is closed. Sever- al studies report similarly good clinical results in two-part, three-part, and valgus-impacted proximal humerus fractures, as well as a low rate of complications, including osteonecrosis.19,32,34

Closed Reduction and Percutaneous Fixation Anteroposterior radiographs of the proximal humerus demonstrating the two levels Closed reduction and percutane- used to measure the combined medial and lateral cortical thickness of the humeral ous fixation with either threaded diaphysis. Level one, the most proximal aspect of the humeral diaphysis, occurs pins or screws is a reasonable option at the level in which the endosteal borders of the medial and lateral cortices are parallel. Level two is 20 mm distal to the first level. Examples of patients with low in AO/ASIF type A (unifocal) frac- bone mineral density (A) and high bone mineral density (B). (Reproduced with tures with translation >66% and permission from Tingart MS, Apprelexa M, von Stechow D, et al: The cortical combined cortical thickness >4 mm. thickness of the proximal humeral diaphysis predicts bone mineral density of the The advantage of closed reduction proximal humerus. J Bone Joint Surg Br 2003;85:611-617.) with percutaneous fixation is that it requires minimal surgical dissection with less disruption of the remain- older patients actually had better stronger fixation than does the soft ing vascular supply. Adequate corti- scores than did the younger pa- bone of the humeral head.18 Tension cal purchase is required; thus, os- tients.26,27,29,30 Therefore, sling immo- band, figure-of-8 constructs incorpo- teoporotic bone with extensive bilization is an appropriate treat- rating the rotator cuff tendons have comminution is a relative contrain- ment option for patients older than proved to be an excellent method of dication. age 60 years with valgus-impacted, fixation for tuberosity fractures.18,31 Using fluoroscopy, an indirect two-part surgical neck or two-part For isolated greater tuberosity frac- closed reduction is performed. If tuberosity fractures. tures with >5 mm of displacement, necessary, a pointed hook retractor Flatow et al32 describe a transosseous can be introduced into the subacro- suture fixation technique using a lat- mial space to assist in the reduction. Surgical Management eral approach to the shoulder. Four Two to three threaded Kirschner Transosseous Suture or five no. 2 nonabsorbable sutures pins (0.045 to 0.0625 in) engage the Fixation are passed through the supraspinatus lateral cortex distal to the deltoid in- AO/ASIF type A fractures with tendon, and drill holes are created in sertion and then are advanced into >66% translation and combined cor- the humerus to secure anatomic re- the subchondral bone of the hu- tical thickness <4 mm are appropri- duction of the greater tuberosity meral head without penetrating the ate candidates for suture fixation fragment. articular surface. For greater tuber- with tension band constructs. Ten- Transosseous suture fixation also osity fractures in isolation or in con- sion band fixation takes advantage of can be done on two-part surgical junction with surgical neck frac- the soft-tissue anchorage of the rota- neck fractures and three-part proxi- ture, two pins should purchase the tor cuff tendon, which provides mal humerus fractures.32-34 Three medial cortex >20 mm distal to the

16 Journal of the American Academy of Orthopaedic Surgeons Shane J. Nho, MD, et al

Table 1 Figure 5 Constant Score29 Subjective Shoulder Assessment (35 total points) Criteria Points Pain (15 points) None 15 Mild 10 Moderate 5 Severe 0 Activities of daily living (10 points) Ability to work 0-4 Ability to engage in recreational activities 0-4 Ability to sleep 0-2 Ability to work at a speciﬁc level (10 points) Waist 2 Chest 4 Neck 6 Transosseous suture fixation for Head 8 AO/ASIF type A extra-articular, proximal Above head 10 humerus fracture with <4 mm combined cortical thickness. The tension-band construct with Objective Shoulder Assessment (65 points) transosseous suture fixation can be used for type A fractures involving the Criteria Points surgical neck, greater tuberosity, and Flexion and abduction (scored separately) lesser tuberosity. (Reproduced with >150° 10 permission from Flatow EL, Cuomo F, 121°-150° 8 Macky MG, et al: Open reduction and 91°-120° 6 internal fixation of two-part displaced 61°-90° 4 fractures of the greater tuberosity of 31°-60° 2 the proximal part of the humerus. 0°-30° 0 J Bone Joint Surg Am 1991;73: 1213-1218.) Combined active external rotation (10 points) Hand behind head, elbow forward 2 Hand behind head, elbow back 2 Hand on top of head, elbow forward 2 achieved with closed reduction and Hand on top of head, elbow back 2 percutaneous fixation in three- and Full elevation from top of head 2 four-part fractures, even though Combined active internal rotation (10 points) these fractures are technically Interscapular region 10 more demanding than two-part frac- Inferior tip of scapula 8 tures11 (Table 2). Twelfth rib 6 Lumbosacral junction 4 Open Reduction and Buttock 2 Internal Fixation Lateral thigh 0 Conventional Plate Fixation Strength (25 points) 1/lb Proximal humerus fractures typically occur in elderly patients with osteoporotic bone; thus, successful inferior border of the head. How- graphs should be taken weekly to implant fixation in these fractures ever, if anatomic reduction cannot ensure that the pins have not migrat- must account for the mechanical be obtained with indirect reduction ed. The pins may be removed 6 to 8 and biologic aspects of osteoporotic maneuvers, the surgeon should have weeks later once fracture union is bone. Poor bone mineral density cor- a low threshold for converting to present, at which time active and relates with reduced ability to se- open reduction techniques. active-assisted ROM exercises are curely hold conventional plates and Pendulum exercises may be initi- begun. Good clinical results and low screws,31 making it difficult to ated after surgery, and serial radio- rates of osteonecrosis have been achieve sufficient screw torque to

Volume 15, Number 1, January 2007 17 Innovations in the Management of Displaced Proximal Humerus Fractures

Table 2 Recent Studies on the Treatment of Proximal Humerus Fractures No. of Fracture Pattern Complications Study Shoulders (No.) Treatment Outcome Score (No.) Court- 125 Minimal (57), greater Sling Mean Constant, 71.8 NR Brown tuberosity (19), surgical Mean Neer, 80.6% good to et al26 neck (18), greater excellent tuberosity and surgical neck (31) Chen 19 Two-part (13), CRPF Neer, 84% good to Painful et al33 three-part (6) excellent hardware (1 [6%]), osteonecrosis (0) Resch 27 Three-part (9), four-part CRPF Average Constant, Osteonecrosis in et al11 (18) three-part: 85.4 (91%) four-part Average Constant, (2 [11%]) four-part: 82.5 (87%) Park 28 Greater tuberosity (13), TOSF Average ASES, 87.1 Second surgery et al34 surgical neck (9), greater Average Neer, 78% (2 [7%]: 1 tuberosity and surgical excellent, 11% satisfactory, irrigation and neck (6) 11% unsatisfactory débridement, 1 capsulotomy), osteonecrosis (0) Gerber 34 Two-part (2), three-part Internal Mean Constant for all Osteonecrosis et al7 (16), four-part (16) ﬁxation shoulders, 78 (89%) (12 [35%]), (CRPF, Constant without malunion (4 TOSF, osteonecrosis, 83 (96%) [12%]), ORIF) ± Constant with arthroplasty (2 bone graft osteonecrosis, 66 (76%) [6%], (P < 0.0005) second surgery (6 [18%]: 4 SAD, 1 hematoma evacuation, 1 capsulotomy)

ASES = American Shoulder and Elbow Surgeons, CRPF = closed reduction and percutaneous fixation, NR = not reported, ORIF = open reduction and internal fixation, SAD = subacromial decompression, TOSF = transosseous suture fixation prevent plate and fracture motion.37 and internal fixation with buttress teoporotic patients, reported rates of Screw loosening and pull-out, not plate fixation is appropriate for mul- osteonecrosis may be as high as 35% implant breakage, is the reason for tifragmentary fractures (ie, AO/ASIF with open reduction and internal fix- construct failure and related poor type B). The goal of buttress plate fixation with plate fixation.7,35 Al- clinical outcomes.5,33 Furthermore, ation is anatomic reduction, which though patients treated with conven- with extensive comminution, par- typically can be accomplished in tional plate fixation may have ticularly in valgus-impacted frac- fractures with an adequate diaphy- satisfactory clinical results even tures, using conventional screws to seal cortex (>3.5 mm) and no meta- when osteonecrosis develops,7,35 pa- lag the fracture fragments to the physeal comminution.7 Disruption tients without humeral head necro- plate typically results in malreduc- of the medial hinge is a relative con- sis have significantly (P < 0.0005) tion. Conventional plate-and-screw traindication to the use of buttress better clinical outcomes than do pa- fixation of displaced proximal hu- plate fixation.22 Buttress plate appli- tients in whom necrosis develops.7 merus fractures has historically been cation requires extensive soft-tissue Plates should not be used to associated with osteonecrosis, non- dissection, resulting in an increased bridge gaps in osteoporotic bone; union, malunion, screw cutout, and risk of disruption of the intraosseous however, double-plating on the op- screw loosening.7 blood supply, which predisposes the posite cortex may be a reasonable al- In young patients, open reduction patient to osteonecrosis. In older, os- ternative for stabilizing such gaps.5

18 Journal of the American Academy of Orthopaedic Surgeons Shane J. Nho, MD, et al

Table 2 (continued)

No. of Fracture Pattern Complications Study Shoulders (No.) Treatment Outcome Score (No.) Wijgman 60 Three- and ORIF with Constant, 87% good or Osteonecrosis (22 [37%]: et al35 four-part cerclage wires excellent, 13% poor 17 [77%] had good or or T-plate excellent Constant score) Hessmann 98 Two-part (50), ORIF with Constant, 69% good to Malunion (20 [20%]), et al36 three-part (37), T-plate ± bone excellent osteonecrosis (3 [4%]), four-part (6) cement or Neer, 59% good to nonunion (1 [1%]) bone graft excellent Wanner 71 (60 Two-part (10), ORIF with Mean Constant, 61 (78%) Complications (7 [12%]): et al10 available three-part (41), double fracture displacement, for four-part (20) one-third osteonecrosis, frozen follow-up) tubular plate shoulder, SAD, loosening Age and pattern had no effect on outcome Robinson 25 Valgus Screws with Constant (median), 80 Impingement (3 [12%]), and impacted Norian Dash (median), 22 wound infection Page9 Skeletal (2 [8%]), adhesive Repair capsulitis (1 [4%]), System, or osteonecrosis (0) buttress plate with Norian Skeletal Repair System Fankhauser 29 AO type A (4), ORIF with Mean Constant (all Redislocation (4 [14%]), et al6 type B (15), locking plate shoulders), 74.6 varus malunion type C (9) Type A, 82.6 (1 screw cutout; 2 Type B, 78.3 impingement) (3 [10%]), Type C, 64.6 partial osteonecrosis (2 [7%]), plate breakage (1 [3%]), deep infection (1 [3%])

Wanner et al10 managed 60 shoulders fractures in osteoporotic bone, par- small blade plates,8 several long with one-third tubular plates fixed ticularly with metaphyseal commi- metaphyseal screws locked to the orthogonally on the anterior and lat- nution. The key to this technology is plate perpendicular to the limb axis eral cortices (Figure 6). Sixty-three the fixed-angle relationship between are better able to resist cantilever percent of patients had good or very the screws and plate. The threaded bending forces and torsional forces.8 good results. Seven patients (12%) screw heads are locked into the Precontoured, fixed-angle plate de- experienced complications, includ- threaded plate holes to prevent screw signs likely provide a mechanical ad- ing fracture displacement, osteone- toggle, slide, and pull-out, thus di- vantage in fractures with metaphy- crosis, adhesive capsulitis, subacro- minishing the possibility of primary seal comminution that lack bony mial impingement, and hardware or secondary loss of reduction.38 The contact opposite the plate.8 loosening.10 load is transmitted from the bone to For AO/ASIF type B (bifocal) and the plate via the screw-plate type C (anatomic neck) proximal hu- Locking Plate Fixation threaded connection. Because the merus fractures, humeral head pres- Locking plate technology has plate and screws form a single stable ervation may be accomplished with been developed as a potential solu- system, the stability of the fracture the contoured proximal humerus tion to the difficulties encountered depends on the stiffness of the entire locking compression plate6,12 and, using conventional plating to treat construct.38 Analogous to multiple, when necessary, bone graft or bone

Volume 15, Number 1, January 2007 19 Innovations in the Management of Displaced Proximal Humerus Fractures

Figure 6

Open reduction and internal fixation of four-part (AO/ASIF type B) proximal humerus fracture with double, one-third tubular plates. A, Anteroposterior (left) and scapular-Y (right) injury radiographs demonstrating four-part proximal humerus fracture. B, Postoperative anteroposterior (left) and scapular-Y (right) radiographs following placement of one-third tubular plates on the lateral and anterior cortices of the proximal humerus. (Reproduced with permission from Wanner GA, Wanner-Schmid E, Romero J, et al: Internal fixation of displaced proximal humeral fractures with two one-third tubular plates. J Trauma 2003;54:536-544.) graft substitute.7,9,39 Open reduction disimpact the head fragment. The al head using the insertion guide and and locked plate fixation is contrain- fracture must be reduced and provi- sleeve assembly. At least three distal dicated in some fracture-dislocation sionally fixed into position before shaft screws are inserted. The final patterns, head-splitting fractures, and permanent fixation with the con- fluoroscopic images should demon- impression fractures involving >40% toured proximal humerus locking strate anatomic reduction of the to 50% of the articular surface.15,40,41 plate. proximal humerus fracture (Figure 7, The tagged sutures are passed C and D). Surgical Technique through the proximal humerus The patient may be positioned in plate, and the plate is positioned di- Results the lateral decubitus position, su- rectly in the middle of the lateral Fankhauser et al6 reported the pine on a radiolucent table, or in the cortex (Figure 7, A). On the antero- first series of patients treated with beach chair position. We prefer to posterior view, the plate should be proximal humerus locking plates. place the patient in the beach chair positioned roughly 8 mm distal to Twenty-nine shoulders with proxi- position and to use a standard del- the superior tip of the greater tuber- mal humerus fractures (5 AO/ASIF topectoral approach. The anterior osity; from the lateral view, the plate type A, 15 type B, 9 type C) were eval- one third of the deltoid is dissected should be centered against the later- uated at minimum 1-year follow-up off its insertion into the deltoid tu- al aspect of the greater tuberosity using Constant scores and plain ra- bercle. When necessary, the greater (Figure 7, B). An adequate gap should diographs. Mean Constant score for and lesser tuberosity fragments are be left between the plate and the bi- all patients at 1 year was 74.6 (type tagged with 5-0 braided, nonabsorb- ceps tendon to prevent disruption of A, 82.6; type B, 78.3; type C, 64.6). able suture. For three- and four-part the anterior humeral circumflex ar- Four patients had redislocation; one fractures, the tuberosity fragments tery. The initial screw is placed in fracture was associated with a deep are reduced to the lateral cortex of the elongated hole in the humeral infection. Three patients had varus the shaft. Reduction of the tuberos- shaft so that the height of the plate malunion in which the screws cut ities may indirectly reduce the head can be adjusted. Once appropriate through the humeral head; in two fragment; alternatively, to restore fracture reduction and plate position of these patients, subacromial im- the medial calcar of the proximal hu- have been achieved, the locked pingement developed because the merus, an elevator can be inserted to screws are inserted into the humer- plate was positioned too proximal.

20 Journal of the American Academy of Orthopaedic Surgeons Shane J. Nho, MD, et al

Only 2 of the 29 patients in the Figure 7 study had partial osteonecrosis of the humeral head. In one patient, the plate fractured before fracture healing, necessitating revision surgery.6 Strohm et al12 reported 64% good and very good results in 35 patients treated with the proximal humerus locking plate. Partial humeral head necrosis was noted in 16% of patients. The authors reported that some patients have a high risk of proximal screw perforation of the humeral head. Although prelimi- nary results are promising with locking plates, additional studies with larger cohorts and longer follow-up are necessary to better define the appropriate indications for and expected outcomes with this technology.

Osteobiologics Displaced proximal humerus fractures may have significant metaphyseal comminution, which often is more apparent after anatomic reduction of fracture fragments. Bone graft or bone graft substitute may be used as a local adjuvant to fracture healing. The indications have not yet been clearly defined, but improved rates of fracture healing and decreased rates of malunion have been reported.7,36 Autogenous bone graft, which has been shown to have osteoinductive, osteoconductive, and osteogenic31 properties, can stimulate new bone Treatment of AO/ASIF type B proximal humerus fracture with locked plate fixation. formation in metaphyseal defects A, Tagged sutures are passed through the proximal humerus locking plate. B, associated with complex proximal Locked plate on the lateral cortex with suture fixation to the plate. Anteroposterior humerus fractures. Elderly, os- (C) and internal rotation (D) radiographs demonstrating anatomic reduction after teoporotic patients have limited open reduction and internal fixation with a proximal humerus locking plate. quantity and quality of available il- iac crest graft, necessitating larger exposure, thus placing them at risk Calcium phosphate cements (eg, No- seous union without evidence of os- of increased complications associ- rian Skeletal Repair System, Syn- teonecrosis in all fractures, and the ated with a second surgical site.5 thes, West Chester, PA; Alpha BSM, average head-neck inclination re- Bone graft substitutes are an at- Etex, Cambridge, MA) can be in- mained within previously published tractive alternative, particularly in jected or molded into small bone de- normal values9 (Figure 8). osteoporotic patients. Substitutes in- fects; they provide structural support The disadvantages of calcium clude synthetic osteoconductive ma- with good compressive strengths.42 phosphate cements are poor shear terials, allograft bone, and deminer- Using the Skeletal Repair System, strength, slow resorption, and com- alized allograft bone matrix.5,7,36 Robinson and Page9,39 achieved os- plete lack of osteoinductivity.42 For

Volume 15, Number 1, January 2007 21 Innovations in the Management of Displaced Proximal Humerus Fractures

Figure 8

Use of the Norian Skeletal Repair System (Synthes) to repair metaphyseal defects in the humeral head. A, After provisional fixation of the fracture and tagged tuberosity sutures, the metaphyseal defect is evident between the articular fragment and the tuberosity fragments. B, Fluoroscopic image demonstrating the area of radiolucency after anatomic reduction. C, The metaphyseal cavity is closed by suturing together the tuberosity fragments. Anteroposterior (D) and modified axial (E) radiographs demonstrating anatomic reduction after open reduction and internal fixation and use of the Norian Skeletal Repair System (the radiodense area behind the humeral head). GT = greater tuberosity, LT = lesser tuberosity. (Reproduced with permission from Robinson CM, Page RS: Severely impacted valgus proximal humeral fractures. J Bone Joint Surg Am 2004; 86[suppl 1 pt 2]:143-155.)

larger metaphyseal defects (>20 mL as with other allografts, there are union and malunion is unknown. volume), structural allograft, morcel- risks, including unpredictable qual- lized allograft, or autograft may be ity and the possibility (however Hemiarthroplasty used to fill the void.39 Demineralized slight) of disease transmission.42 Bone Traditionally, the indications for human bone matrix can enhance graft substitute in conjunction with hemiarthroplasty in the manage- structural integrity and has variable locked plate fixation has not yet been ment of proximal humerus fractures osteoinductive properties; however, studied, and its effect on fracture are four-part fractures, three-part

22 Journal of the American Academy of Orthopaedic Surgeons Shane J. Nho, MD, et al fractures in older patients with secure positioning of the tuberosities noid wear after shoulder hemiarthro- osteoporotic bone, fracture-disloca- through a variety of suturing tech- plasty found a decrease in average tions, head-splitting fractures, and niques. Bone graft may be necessary. joint space from 2.9 to 1 mm over a impression fractures involving The shoulder should be taken mean 43-month period in eight pa- >40% to 50% of the articular sur- through a ROM before closing. tients.50 The data suggest that func- face.15,40,41 The tenuous fixation of tional outcome, as measured by fracture fragments in osteoporotic Results Constant score, negatively correlat- bone15 and the high rates of osteone- Recent studies have focused on ed with joint space narrowing.50 crosis in the humeral head after maximizing tuberosity healing in an Prosthetic loosening occurs less three- or four-part fracture suggest anatomic position and optimizing often, with reported rates of 3% to that hemiarthroplasty is a better the degree of component retrover- 6%,43,45 although higher rates of ra- treatment alternative for these frac- sion.43,49 Most authors recommend diolucencies (30%) have been report- ture patterns. Although hemiarthro- 30° to 40° of retroversion, typically ed.47 Infection usually occurs at an plasty has been shown to provide using the bicipital groove as the even lower rate, typically 1% to good pain relief, achieving excellent landmark for prosthesis orienta- 2%,41,48 although one study reported ROM has been less predictable. tion.49 However, an individualized a 7.7% rate of infection.43 Although Rehabilitation, particularly early approach has been proposed, in up to 15% of patients may present passive ROM43,44 and long-term ac- which the contralateral humerus is preoperatively with nerve injury tive ROM and strengthening, is con- used to estimate the proper retrover- from the original trauma,41,43,48 the sidered essential to achieve optimal sion for each patient.43 A recent reported rate of intra- or postopera- outcome after shoulder hemiarthro- study found no difference in terms of tive nerve injury typically is close to plasty.43,45 In recent studies, the mean Constant score when compar- zero.41,43,45,48 However, one study re- results of hemiarthroplasty for ing patients treated with these dif- ported a rate of 4.5%.47 proximal humerus fractures have ferent techniques for achieving prop- Both surgeon and patient must been reasonable, although not quite er version, although the sample size have realistic postoperative expecta- as good as the results in earlier (26 patients) and resulting statistical tions for proximal humerus fracture studies.15,41,43-48 significance were limited.43 after hemiarthroplasty. Typical end Complications after hemiarthro- ROM achieves shoulder-level func- Surgical Technique plasty include tuberosity malunion tion. Although the mean reported Shoulder hemiarthroplasty tech- or nonunion, heterotopic ossifica- ROM for forward flexion is often nique is well-described in the lit- tion, degenerative changes in the above 100°,41,43,47 only 42% of pa- erature. Typically, a standard del- glenoid, prosthetic loosening, and tients could flex above 90° in the se- topectoral approach is used. Deep nerve injury. Malunion or nonunion ries reported by Kralinger et al.44 soft-tissue dissection may be mini- of tuberosity osteosynthesis occurs Most studies suggest mean internal mal, depending on the fracture pat- in 4% to 50% of shoulders and is the rotation to approximately L2/ tern and soft-tissue injury. It is essen- most significant complication after L3,41,44,45 with mean external rota- tial to identify and tag the tuberosity hemiarthroplasty for displaced prox- tion typically varying between 20° fragments. The humeral head and imal humerus fractures.40,45,47 In one and 40°.41,43,44,46,47 Functional results shaft fragments are exposed; the gle- study, the factors that led to tuberos- are usually satisfactory, with recent noid should be carefully inspected to ity malunion were poor intraopera- studies showing a mean Constant determine whether a glenoid compo- tive positioning of the prosthesis (ex- score between 55 and 70.43,44,47 nent is warranted. After adequately cessive height and/or retroversion), The importance of anatomic res- exposing and cleaning the joint, prep- age and sex (women older than age toration of the tuberosities with aration of the humeral shaft begins. 75 years), and initial malposition of secure fixation and restoration of It is critical to place the humeral the greater tuberosity.47 humeral length and retroversion component in the correct retrover- Boileau et al47 reported an 11% cannot be overemphasized (Figure 9). sion (typically 30° to 40°) and height. rate of heterotopic ossification, sim- One recent cadaveric study evaluat- Different modular heads can be tri- ilar to the rate reported by Prakash et ed a design in which the bone- aled to identify the optimal configu- al45 (9%), but lower than that report- tendon interface of the rotator cuff is ration. With most fractures, the stem ed by Mighell et al41 (27%) and Beck- fixed to holes in the articular rim of should be cemented to ensure rota- er et al46 (56%). Mighell et al41 re- the prosthesis.51 The fixation was tional control of the prosthesis. Once ported an 8% rate of degenerative strong enough to resist fracture frag- the humeral component is secure, changes in the glenoid at 36-month ment displacement under the forces the proximal anatomy is restored, average follow-up. A recent quanti- associated with activities of daily with particular emphasis on correct, tative radiographic analysis of gle- living, suggesting that this design

Volume 15, Number 1, January 2007 23 Innovations in the Management of Displaced Proximal Humerus Fractures

Figure 9

Anteroposterior (A), scapular-Y (B), and axillary (C) radiographs following shoulder hemiarthroplasty. might allow earlier rehabilitation In their retrospective multicenter reports tend to show better results even in the presence of poor bone review, Kralinger et al44 (P = 0.0001) after early treatment,41,46 although quality. Biomechanical studies have found that patients with healed, one study found no difference be- shown that incorporation of medial nondisplaced tuberosities had signif- tween early (≤30 days postinjury) circumferential cerclage around the icantly higher Constant scores and and late (>30 days postinjury) treat- tuberosities decreases interfragmen- subjective patient satisfaction scores ment.45 A series of 27 displaced four- tary motion and strain, maximizes (P = 0.031) than did patients whose fragment fractures of the proximal fracture stability, and facilitates tuberosities did not heal or healed in humerus treated with hemiarthro- postoperative rehabilitation.52 displacement >0.5 cm. However, plasty showed better shoulder mo- Postoperative outcomes depend pain did not correlate with displace- tion (measured by video motion on several variables, chief among ment of the tuberosity. Although analysis) and better clinical out- them displacement of the tuberosi- Coleman and Craig53 also reported comes (Neer and Constant scores) ties. Robinson et al48 retrospectively no difference in overall pain score when treated within 2 weeks of inju- reviewed the results of shoulder between patients with a healed ry, compared with those treated lat- hemiarthroplasty for proximal hu- greater tuberosity and those with- er.46 Mighell et al41 found significant- merus fractures at a single center. out, their retrospective review did ly (P < 0.01) greater ASES scores in They found consistent improvement find greater forward flexion and ex- patients treated within 2 weeks of in the Constant score from 6 weeks ternal rotation in patients with a initial injury. Prakash et al45 did not to 6 months postoperatively but lit- healed greater tuberosity. In a retro- find any difference in final ROM be- tle change thereafter. At 1 year, pa- spective review of the results of a tween patients who underwent sur- tients reported reasonable pain relief single surgeon, Mighell et al41 found gery within 30 days of initial injury but poorer scores for function, ROM, significantly lower ASES (P < 0.01) and patients who underwent surgery and strength. The factors assessed 6 and Simple Shoulder Test (P < 0.001) more than 30 days after injury, even weeks postoperatively that predicted scores as well as decreased forward though the mean time from injury to 1-year Constant score included pa- flexion in patients with superior mi- surgery among the latter group was tient age, persistent neurologic defi- gration of the greater tuberosity. 13 months. cit, need for early revision, as well as The optimal timing of hemiar- the degree of displacement of the throplasty is important—specifical- Summary prosthetic head from the central axis ly, acute versus late reconstruction. of the glenoid and the degree of dis- Because it is technically easier, acute Proximal humerus fracture manage- placement of the tuberosities (seen treatment is generally preferable to ment is constantly evolving, partic- radiographically). later hemiarthroplasty. Most recent ularly in light of improved under-

24 Journal of the American Academy of Orthopaedic Surgeons Shane J. Nho, MD, et al standing of proximal humerus References 13. Resch H, Beck E, Bayley I: Recon- fracture characteristics and innova- struction of the valgus-impacted hu- Evidence-based Medicine: There are tions in surgical technique and tech- meral head fracture. J Shoulder no level I or II randomized prospec- Elbow Surg 1995;4:73-80. nology. The orthopaedic surgeon tive studies referenced. The majori- 14. Zyto K, Ahrengart L, Sperber A, should approach proximal humerus Tornkvist H: Treatment of displaced ty of references are observational co- fractures on a case-by-case basis and proximal humeral fractures in elderly hort or case control studies (level III/ tailor treatment according to pa- patients. J Bone Joint Surg Br 1997; IV) or expert opinion (level V). tient and fracture characteristics. 79:412-417. 15. Zyto K, Wallace WA, Frostick SP, Pre- Treatment decision-making should Citation numbers printed in bold ston BJ: Outcome after hemiarthro- include assessing the vascular sta- type indicate references published plasty for three- and four-part frac- tus of the humeral head, determin- within the past 5 years. tures of the proximal humerus. ing the optimal fixation constructs, J Shoulder Elbow Surg 1998;7:85-89. and implementing local adjuvants 1. Baron JA, Barrett JA, Karagas MR: The 16. Lee SH, Dargent-Molina P, Breart G, epidemiology of peripheral fractures. EPIDOS Group. Epidemiologie de as needed to enhance anatomic frac- Bone 1996;18(3 suppl):209S-213S. l’Osteoporose Study: Risk factors for ture healing. The likelihood of hu- 2. Naranja RJ Jr, Iannotti JP: Displaced fractures of the proximal humerus: meral head ischemia is established three- and four-part proximal humer- Results from the EPIDOS prospective using the AO/ASIF classification us fractures: Evaluation and manage- study. J Bone Miner Res 2002;17:817- and Hertel’s radiographic criteria. ment. J Am Acad Orthop Surg 2000; 825. 8:373-382. 17. Robertson DD, Yuan J, Bigliani LU, Fractures with AO/ASIF type C pat- 3. Schlegel TF, Hawkins RJ: Displaced Flatow EL, Yamaguchi K: Three- tern, metaphyseal extension <8 proximal humerus fractures: Evalua- dimensional analysis of the proximal mm, or medial hinge displacement tion and treatment. J Am Acad part of the humerus: Relevance to ar- >2 mm are associated with high Orthop Surg 1994;2:54-66. throplasty. J Bone Joint Surg Am 4. Neer CS II: Displaced proximal hu- 2000;82:1594-1602. probability of osteonecrosis and meral fractures: I. Classification and 18. Hawkins RJ, Bell RH, Gurr K: The probably are best treated with hemi- evaluation. J Bone Joint Surg Am three-part fracture of the proximal arthroplasty. 1970;52:1077-1089. part of the humerus: Operative treat- When fixation is required, frac- 5. Cornell CN: Internal fracture fixation ment. J Bone Joint Surg Am 1986;68: in patients with osteoporosis. J Am 1410-1414. ture pattern and cortical thickness Acad Orthop Surg 2003;11:109-119. 19. Panagopoulos AM, Dimakopoulos P, should guide the treatment approach 6. Fankhauser F, Boldin C, Schippinger Tyllianakis M, et al: Valgus impacted and fixation technique. AO/ASIF G, Haunschmid C, Szyszkowitz R: A proximal humeral fractures and their type A fractures are typically treated new locking plate for unstable frac- blood supply after transosseous sutur- tures of the proximal humerus. Clin ing. Int Orthop 2004;28:333-337. with sling immobilization. Proximal Orthop Relat Res 2005;430:176-181. 20. Gerber C, Schneeberger AG, Vinh TS: humerus fractures with surgical 7. Gerber C, Werner CM, Vienne P: In- The arterial vascularization of the hu- neck translation >66% or tuberosity ternal fixation of complex fractures of meral head: An anatomical study. displacement >5 mm may be treated the proximal humerus. J Bone Joint J Bone Joint Surg Am 1990;72:1486- Surg Br 2004;86:848-855. 1494. with transosseous suture fixation 8. Haidukewych GJ: Innovations in 21. Brooks CH, Revell WJ, Heatley FW: (cortex <4 mm) or closed reduction locking plate technology. J Am Acad Vascularity of the humeral head after and percutaneous fixation (cortex >4 Orthop Surg 2004;12:205-212. proximal humeral fractures: An ana- mm). For multifragment fractures, 9. Robinson CM, Page RS: Severely im- tomical cadaver study. J Bone Joint pacted valgus proximal humeral frac- Surg Br 1993;75:132-136. locked compression plating creates a tures: Results of operative treatment. 22. Hertel R, Hempfing A, Stiehler M, stable construct with preservation of J Bone Joint Surg Am 2003;85:1647- Leunig M: Predictors of humeral head the periosteal blood supply. Calcium 1655. ischemia after intracapsular fracture phosphate cement, demineralized 10. Wanner GA, Wanner-Schmid E, of the proximal humerus. J Shoulder Romero J, et al: Internal fixation of Elbow Surg 2004;13:427-433. bone matrix, and allografts are im- displaced proximal humeral fractures 23. Müller ME: Appendix A: The compre- portant local adjuvants that may im- with two one-third tubular plates. hensive classification of fractures of prove rates of osseous union and J Trauma 2003;54:536-544. long bones, in Müller ME, Allgöwer minimize malunion. Given recent 11. Resch H, Povacz P, Frohlich R, M, Schneider R, Willenegger H (eds): Wambacher M: Percutaneous fixation Manual of Internal Fixation: Tech- developments in locking plate and of three- and four-part fractures of the niques Recommended by the AO- osteobiologics technology, the most proximal humerus. J Bone Joint Surg ASIF Group. Berlin, Germany: relevant question may be whether Br 1997;79:295-300. Springer-Verlag, 1991, pp 118-125. open reduction and internal fixation 12. Strohm PC, Kostler W, Sudkamp NP: 24. Siebenrock KA, Gerber C: The repro- Locking plate fixation of proximal ducibility of classification of fractures will in future produce better results humerus fractures. Techniques in of the proximal end of the humerus. than hemiarthroplasty and, if so, in Shoulder & Elbow Surgery 2005;6:8- J Bone Joint Surg Am 1993;75:1751- which patients. 13. 1755.

Volume 15, Number 1, January 2007 25 Innovations in the Management of Displaced Proximal Humerus Fractures

25. Tingart MJ, Apreleva M, von Stechow 35. Wijgman AJ, Roolker W, Patt TW, Hemiarthroplasty for severe fractures D, Zurakowski D, Warner JJ: The cor- Raaymakers EL, Marti RK: Open re- of the proximal humerus. J Shoulder tical thickness of the proximal hu- duction and internal fixation of three Elbow Surg 2002;11:428-430. meral diaphysis predicts bone mineral and four-part fractures of the proximal 46. Becker R, Pap G, Machner A, Neu- density of the proximal humerus. part of the humerus. J Bone Joint mann WH: Strength and motion after J Bone Joint Surg Br 2003;85:611-617. Surg Am 2002;84:1919-1925. hemiarthroplasty in displaced four- 26. Court-Brown CM, Cattermole H, Mc- 36. Hessmann M, Baumgaertel F, Gehling fragment fracture of the proximal hu- Queen MM: Impacted valgus frac- H, Klingelhoeffer I, Gotzen L: Plate merus: 27 patients followed for 1-6 tures (B1.1) of the proximal humerus: fixation of proximal humeral frac- years. Acta Orthop Scand 2002;73: The results of non-operative treat- tures with indirect reduction: Surgi- 44-49. ment. J Bone Joint Surg Br 2002;84: cal technique and results utilizing 47. Boileau P, Krishnan SG, Tinsi L, 504-508. three shoulder scores. Injury 1999; Walch G, Coste JS, Mole D: Tuberos- 27. Court-Brown CM, Garg A, McQueen 30:453-462. MM: The translated two-part fracture 37. Egol KA, Kubiak EN, Fulkerson E, ity malposition and migration: Rea- of the proximal humerus: Epidemiol- Kummer FJ, Koval KJ: Biomechanics sons for poor outcomes after hemiar- ogy and outcome in the older patient. of locked plates and screws. J Orthop throplasty for displaced fractures of J Bone Joint Surg Br 2001;83:799-804. Trauma 2004;18:488-493. the proximal humerus. J Shoulder 28. Neer CS II, Watson KC, Stanton FJ: 38. Wagner M: General principles for the Elbow Surg 2002;11:401-412. Recent experience in total shoulder clinical use of the LCP. Injury 48. Robinson CM, Page RS, Hill RM, replacement. J Bone Joint Surg Am 2003;34(suppl 2):B31-42. Sanders DL, Court-Brown CM, Wake- 1982;64:319-337. 39. Robinson CM, Page RS: Severely im- field AE: Primary hemiarthroplasty 29. Constant CR, Murley AH: A clinical pacted valgus proximal humeral frac- for treatment of proximal humeral method of functional assessment of tures: Surgical techniques. J Bone fractures. J Bone Joint Surg Am 2003; the shoulder. Clin Orthop Relat Res Joint Surg Am 2004;86(suppl 1 pt 2): 85:1215-1223. 1987;214:160-164. 143-155. 49. Kontakis GM, Damilakis J, Christo- 30. Katolik LI, Romeo AA, Cole BJ, 40. Bosch U, Skutek M, Fremerey RW, forakis J, Papadakis A, Katonis P, Pras- Verma NN, Hayden JK, Bach BR: Nor- Tscherne H: Outcome after primary sopoulos P: The bicipital groove as a malization of the Constant score. and secondary hemiarthroplasty in landmark for orientation of the hu- J Shoulder Elbow Surg 2005;14:279- elderly patients with fractures of the meral prosthesis in cases of fracture. 285. proximal humerus. J Shoulder Elbow J Shoulder Elbow Surg 2001;10:136- 31. Cornell CN, Levine D, Pagnani MJ: Surg 1998;7:479-484. 139. Internal fixation of proximal humerus 41. Mighell MA, Kolm GP, Collinge CA, 50. Parsons IM IV, Millet PJ, Warner JJP: fractures using the screw-tension Frankle MA: Outcomes of hemiar- Glenoid wear after shoulder hemiar- band technique. J Orthop Trauma throplasty for fractures of the proxi- throplasty: Quantitative radiographic 1994;8:23-27. mal humerus. J Shoulder Elbow Surg analysis. Clin Orthop Relat Res 32. Flatow EL, Cuomo F, Maday MG, 2003;12:569-577. 2004;421:120-125. Miller SR, McIlveen SJ, Bigliani LU: 42. Ladd AL, Pliam NB: Use of bone-graft 51. De Wilde LF, Berghs BM, Beutler T, Open reduction and internal fixation substitutes in distal radius fractures. Ferguson SJ, Verdonk RC: A new pros- of two-part displaced fractures of the J Am Acad Orthop Surg 1999;7:279- thetic design for proximal humeral greater tuberosity of the proximal part 290. fractures: Reconstructing the gleno- of the humerus. J Bone Joint Surg Am 43. Christoforakis JJ, Kontakis GM, Kato- humeral unit. J Shoulder Elbow Surg 1991;73:1213-1218. nis PG, Stergiopoulos K, Hadjipavlou 2004;13:373-380. 33. Chen CY, Chao EK, Tu YK, Ueng SW, AG: Shoulder hemiarthroplasty in the 52. Frankle MA, Ondrovic LE, Markee BA, Shih CH: Closed management and per- management of humeral head frac- Harris ML, Lee WE III: Stability of tu- cutaneous fixation of unstable proxi- tures. Acta Orthop Belg 2004;70:214- berosity reattachment in proximal hu- mal humerus fractures. J Trauma 218. meral hemiarthroplasty. J Shoulder 1998;45:1039-1045. 44. Kralinger F, Schwaiger R, Wambacher Elbow Surg 2002;11:413-420. 34. Park MC, Murthi AM, Roth NS, M, et al: Outcome after primary hemi- 53. Coleman SH, Craig EV: Hemiarthro- Blaine TA, Levine WN, Bigliani LU: arthroplasty for fracture of the head of plasty for complex fractures of the Two-part and three-part fractures of the humerus: A retrospective multi- proximal humerus: Surgical tech- the proximal humerus treated with centre study of 167 patients. J Bone nique and results with the Atlas tri- suture fixation. J Orthop Trauma Joint Surg Br 2004;86:217-219. modular prosthesis. Am J Orthop 2003;17:319-325. 45. Prakash U, McGurty DW, Dent JA: 2002;31(1 suppl):11-17.

26 Journal of the American Academy of Orthopaedic Surgeons