Abstract
Background
The role of hemiarthroplasty (HA) in the management of proximal humerus fractures (PHFs) and their sequalae has evolved with the development of contemporary internal fixation techniques and the widespread use of the reverse total shoulder arthroplasty. However, HA may still have a role in certain acute PHFs as well as select fracture sequalae. The aim of this investigation was to evaluate the outcomes of HA when used in acute fractures and fracture sequelae.
Methods
Over a 16-year period (2000 - 2016), 122 primary HA performed for either acute PHFs or fracture sequelae were identified. Of these, 70 (57.4%) HA were performed within 4 weeks of the injury, whereas 52 (42.6%) underwent HA for fracture nonunion, malunion, or avascular necrosis. The minimum follow-up period was 2 years. Outcomes included the visual analog scale for pain, range of motion, American Shoulder and Elbow Surgeons (ASES) score, complications, and reoperations inclusive of revision surgery. Cumulative incidence analysis was used to report implant survivorship with death as a competing risk.
Results
The mean follow-up time after HA was 4.8 years (range, 2-15 years) with no differences between groups (P = .102). Cohort comparisons demonstrated an older age (67.8 vs. 60.1; P = .004), lower rate of previous procedure (4.3% vs. 51.9%; P < .001), lower bone graft use (28.6% vs. 59.6%; P < .001), and a longer length of stay (5.9 vs. 3.0 days; P < .001) in the acute HA group. Additionally, no differences were observed between the acute and sequalae cohort in pain (2.0 vs. 2.5; P = .523), forward elevation (98° vs. 93°; P = .627), external rotation (30° vs. 23°; P = .215), internal rotation score (4.0 vs. 4.5; P = .589), satisfaction (P = .592), ASES scores (64.4 vs. 57.1; P = .168), complications (27.1% vs. 28.8%; P = .836), or reoperations (11.4% vs. 19.2%; P = .229). When comparing acute fractures and sequalae, the 15-year complication rates were 32.4% and 43.3%, respectively (P = .172), with 15-year reoperation rates of 13.7% and 24%, respectively (P = .098).
Conclusions
HA, whether performed acutely for a PHF or in a delayed fashion for fracture sequalae, demonstrated no statistically significant differences in outcomes for all examined parameters. HA in this setting may provide reasonable pain relief. However, limited motion, marginal ASES scores, and elevated rates of complications and reoperations can be expected up to 15 years postoperatively.
Keywords: Hemiarthroplasty, Proximal humerus fracture, Avascular necrosis, Implant survivorship
Adult proximal humerus fractures (PHFs) remain a common injury pattern, accounting for approximately 6% of all fractures.9,25 Fractures often vary in location and complexity, with possible involvement of the surgical and anatomic necks of the humerus, as well as the greater and lesser tuberosities. Over the past decade, management of PHFs has continued to evolve, especially as a result of modern internal fixation techniques and the widespread use of reverse shoulder arthroplasty (RSA).22 When alignment is reasonably well maintained, nonoperative management is often successful.4,15,21,29 However, substantial controversy remains in more complex fracture patterns.16,28
In patients < 65 years old without evidence of osteopenia, operative intervention is often indicated for displaced fractures; glenohumeral fracture-dislocations; displaced intra-articular fractures; and greater tuberosity displacement of > 5 mm.22,30 Conventionally, these patterns are managed with open reduction and internal fixation (ORIF). However, when the humeral head cannot be reliably reconstructed or there is concern about subsequent humeral head viability, hemiarthroplasty (HA) and RSA are reasonable options.6,12,18,19 In geriatric patients, ORIF has demonstrated considerably high failure rates leading to increased interest in both nonoperative management and shoulder arthroplasty.2,3,7,22,34 Currently, RSA has been identified as more effective, leading to a reciprocal rise in RSA use and decline in HA utilization.1,26,32,38
Despite these advances, HA may still serve a role for acute fractures with unreconstructable patterns in young patients, compromised glenoid bone stock not amenable to glenoid component implantation, or sequelae where the humeral head is not viable, but tuberosity healing is not needed (post-traumatic avascular necrosis [AVN] and malunion).20,37 In RSA, early surgery for acute PHF demonstrated quicker recovery and primary RSA may outperform salvage RSA.26,36,38 Additionally, patient-specific, evidence-based algorithms have also been implemented to help mitigate complications and improve outcomes with very limited roles for HA.35 As such, the purpose of this study was to evaluate the outcomes of HA for management of PHFs and to compare the results of HA for acute fracture and fracture sequalae.
Materials and methods
After institutional review board approval, a prospectively collected Institutional Total Joint Registry was queried to identify all HA performed between 2000 and 2016 for PHF or fracture sequelae. A total of 134 shoulders were initially identified. Exclusions consisted of patients with less than 2 years of clinical follow-up (n = 8), oncologic reconstructions (n = 3), and previous native shoulder infection (n = 1). The final cohort consisted of 122 HA (83 women and 39 men), of which 70 (52 women and 18 men) had been performed for acute PHFs (< 4 weeks from the injury) and 52 (21 women and 31 men) underwent HA for fracture sequalae (Table I). Fracture sequelae included malunion, nonunion, or AVN with or without prior ORIF. Of these 52 shoulders, 27 (51.9%) were previously treated nonoperatively and 25 (48.1%) had previously undergone ORIF. In this series, there were no “nondisplaced” or minimally displaced (Neer 1-part) fractures. Of note, 59 (48.4%) HA were performed prior to January 2005 and 63 (51.6%) HA were performed between January 2005 and December 2016. Throughout the study period, 464 RSA were performed for acute PHF and their sequelae, with the first case being performed in August 2004.
Table I.
Baseline demographic and clinical characteristics.
| Characteristics | Acute fracture (n = 70) | Sequels of trauma (n = 52) | Total (n = 122) | P value |
|---|---|---|---|---|
| Age (y) | 67.8 (13.3) | 60.1 (15.1) | 64.6 (14.5) | .004 |
| Sex | .083 | |||
| Male | 17 (24.3%) | 20 (38.4%) | 37 (30.3%) | |
| Female | 53 (75.7%) | 32 (61.6%) | 85 (69.7%) | |
| Body mass index (kg/m2) | 31.0 (8.5) | 30.1 (6.6) | 30.6 (7.7) | .964 |
| Laterality | .537 | |||
| Right | 44 (62.9%) | 29 (55.8%) | 73 (59.8%) | |
| Left | 26 (37.1%) | 23 (44.2%) | 49 (40.2%) | |
| Dominant | .213 | |||
| No | 26 (37.1%) | 27 (51.9%) | 53 (43.4%) | |
| Yes | 44 (62.9%) | 25 (48.1%) | 69 (56.6%) | |
| Previous procedure | <.001 | |||
| No | 67 (95.7%) | 25 (48.1%) | 92 (75.4%) | |
| Yes | 3 (4.3%) | 27 (51.9%) | 30 (24.6%) | |
| Time from fracture to HA (weeks) | 0.6 (0.1-3) | 31 (7-109) | 14 (0.1-109) | <.001 |
| Final follow-up (y, range) | 4.7 (2.0, 15.2) | 5.1 (2.0, 15.5) | 4.8 (2.0, 15.5) | .102 |
HA, hemiarthroplasty.
All values are reported as the mean (standard deviation) or number (percentage) except when noted.
Operative intervention was completed by one of 7 upper extremity surgeons. A standard deltopectoral approach was uniformly used. In all cases, special attention was placed on the tuberosities to achieve adequate tuberosity reduction and promote optimal conditions for tuberosity healing. Sutures around the tuberosities were secured to the stem and to the shaft. The use of bone cement for humeral component fixation was at the discretion of the surgeon, and more often used for acute fractures. Implant selection varied by surgeon with the implantation of the Cofield2 shoulder prosthesis (Smith & Nephew, Memphis, TN, USA) in 46 shoulders, Aequalis shoulder prosthesis (Tornier/Viking, Montbonnot, France) in 26 shoulders, Global advantage shoulder prosthesis (DePuy Synthes, West Chester, PA, USA) in 22 shoulders, Bio-Modular shoulder prosthesis (Zimmer Biomet, Warsaw, IN, USA) in 20 shoulders, Bigliani/Flatow shoulder prosthesis (Zimmer Biomet, Warsaw, IN, USA) in 5 shoulders, and the ReUnion shoulder prosthesis (Stryker, Kalamazoo, MI, USA) in 3 shoulders. Postoperatively, the affected shoulder was immobilized for 6 weeks, with progressive introduction of passive motion, active assisted motion, active motion, and strengthening with elastic bands.
A concurrent review of data extracted from our Institutional Total Joint Registry and medical record was reviewed to obtain preoperative and postoperative characteristics. Variables included 10-point visual analog pain scale scores (VAS), active range of motion (ROM) measurements assessed in 3 planes (forward elevation (FE) in the scapular plane (in degrees), external rotation (ER) with the arm at the side (in degrees), and internal rotation (IR) as measured by the level reached by the thumb.11 Additional outcome scores included patient satisfaction, the American Shoulder and Elbow Surgeons (ASES) score, postoperative complications, and implant survivorship free from reoperation or revision surgery (revision being defined as subsequent exchange and/or removal of any components, humeral head of stem).
Preoperative radiographs and computed tomography scans, if available, were evaluated for fracture classification. Acute fractures were categorized using Neer’s classification.8,24 Sequalae of the original fractures were evaluated for fracture union, malunion, and nonunion. A malunion of the greater or lesser tuberosity, articular surface, or articular segment were documented according to the definitions by Beredjiklian et al.5 Fractures were considered a nonunion if the fracture was not clinically or radiographically united after 6 months from definitive fixation or sooner if there was subsequent displacement of the fracture due to implant failure. AVN was assessed according to the Cruess classification.10 Postoperative radiographs included a true anteroposterior, scapular Y view, and an axillary view. These radiographs were taken at approximate intervals of 3 weeks, 6 weeks, 3 months, 6 months, 1 year, 2 years, 5 years, and every 5 years thereafter. Evaluated parameters included vertical and horizontal tuberosity reduction, tuberosity resorption, and glenoid erosion. Vertical tuberosity reduction was categorized as high (<5 mm), anatomic (5-10 mm), or low (>10 mm). Tuberosity resorption was categorized as none, partial, or complete. Glenoid erosion after HA was graded according to the method proposed by Sperling et al.33
Statistical analysis
Statistical analysis was performed using JMP 14.1.0 software (SAS Institute Inc., Cary, NC, USA). Descriptive statistics (means, ranges, and standard deviations) were calculated with comparisons performed utilizing the Student t tests, χ2, or Fisher exact test as indicated. The McNemar test was performed for paired categorical data and the Wilcoxon signed rank analysis was used for continuous variables. Univariate Cox models were utilized to evaluate for potential risk factors for complications and reoperations. Kaplan-Meier survival curves were generated to determine survivorship free from reoperation and revision. Additionally, due to the extended follow-up of the patients, a cumulative incidence analysis with all-cause mortality as a competing risk was performed. In all cases, P < .05 was considered statistically significant.
Results
A total of 122 HA met the final inclusion criteria, of which 70 (57.4%) underwent had been performed for acute fracture (<4 weeks post-injury) and 52 (42.6%) had been performed for fracture sequalae. The mean follow-up time after HA was 4.8 years (range, 2-15 years). When compared with HA for fracture sequelae, HA for acute fracture demonstrated an older age (67.8 vs. 60.1; P = .004), lower rate of a previous procedure (4.3% vs. 51.9%; P < .001), less common bone graft use (28.6% vs. 59.6%; P < .001), more common use of cement for humeral component fixation (80% vs. 34.6%; P < .001), and a longer length of stay (5.9 vs. 3.0 days; P < .001). Other notable demographic, surgical, and radiographic characteristics are outlined in Table I, Table II, Table III, respectively.
Table II.
Perisurgical characteristics categorized by timing of HA.
| Characteristic | Acute fracture (n = 70) | Sequels of trauma (n = 52) | Total (n = 122) | P value |
|---|---|---|---|---|
| Anesthesia time (min) | 313.8 (89.0) | 305.6 (133.8) | 310.3 (109.8) | .167 |
| OR time (min) | 230.6 (92.5) | 229.9 (133.9) | 230.3 (111.4) | .503 |
| Bone graft use | <.001 | |||
| None | 20 (28.6%) | 31 (59.6%) | 51 (41.8%) | |
| Autograft | 47 (67.1%) | 15 (28.8%) | 62 (50.8%) | |
| Allograft | 3 (4.3%) | 6 (11.5%) | 9 (7.4%) | |
| Humeral stem cement | <.001 | |||
| No | 14 (20.0%) | 34 (65.4%) | 48 (39.3%) | |
| Yes | 56 (80.0%) | 18 (34.6%) | 74 (60.7%) | |
| Subscapularis management | <.001 | |||
| Tenotomy | 1 (1.4%) | 24 (46.2%) | 25 (20.5%) | |
| Peel | 3 (4.3%) | 15 (28.8%) | 18 (14.8%) | |
| Through fracture | 66 (94.3%) | 12 (23.1%) | 78 (63.9%) | |
| Absent | 0 (0.0%) | 1 (1.9%) | 1 (0.8%) | |
| Length of stay (d) | 5.9 (6.7) | 3.0 (2.2) | 4.7 (5.5) | <.001 |
HA, hemiarthroplasty; OR, operating room.
All values are reported as the mean (standard deviation) or number (percentage) except when noted.
Table III.
Preoperative and postoperative radiographic characteristics.
| Acute fracture (n = 70) | Sequels of trauma (n = 52) | Total (n = 122) | |
|---|---|---|---|
| Preoperative | |||
| Neer fracture classification | |||
| 2 | 2 (2.9%) | 2 (1.6%) | |
| 3 | 27 (38.6%) | 27 (22.1%) | |
| 4 | 41 (58.6%) | 41 (33.6%) | |
| Malunion | 23 (44.2%) | 31 (18.9%) | |
| Nonunion | 20 (48.5%) | 21 (16.4%) | |
| Avascular necrosis | 9 (17.3%) | 9 (7.4%) | |
| Postoperative ∗ | |||
| Vertical tuberosity reduction | |||
| High | 31 (44.9%) | 29 (61.7%) | 60 (51.7%) |
| Anatomic | 21 (30.4%) | 15 (31.9%) | 36 (31.0%) |
| Low | 17 (24.6%) | 3 (6.4%) | 20 (17.2%) |
| Horizontal tuberosity reduction | |||
| No | 5 (11.1%) | 4 (14.8%) | 9 (12.5%) |
| Yes | 40 (88.9%) | 23 (85.2%) | 63 (87.5%) |
| Tuberosity resorption | |||
| None | 30 (50.8%) | 23 (65.7%) | 53 (56.4%) |
| Partial | 14 (23.7%) | 4 (11.4%) | 18 (19.1%) |
| Complete | 15 (25.4%) | 8 (22.9%) | 23 (24.5%) |
| Tuberosity healed | |||
| No | 17 (28.8%) | 9 (25.0%) | 26 (27.4%) |
| Yes | 42 (71.2%) | 27 (75.0%) | 69 (72.6%) |
| Glenoid erosion | |||
| None | 18 (42.9%) | 6 (19.4%) | 24 (32.9%) |
| Mild | 20 (47.6%) | 11 (35.5%) | 31 (42.5%) |
| Moderate | 4 (9.5%) | 12 (38.7%) | 16 (21.9%) |
| Severe | 0 (0%) | 2 (6.5%) | 2 (2.7%) |
All values are reported as the number (percentage) except when noted.
Missing values present (variables were not available for all 122 patients, as such cohort percentages were calculated based on categorical totals).
At the final follow-up, the entire cohort demonstrated a mean VAS of 2.2 ± 2.1, FE of 96° ± 43°, ER of 28° ± 23°, IR score of 4.2 ± 1.7, and ASES of 61.1 ± 17.1 (Table III). Half of the patients rated satisfaction as much better (n = 59; 48.4%), with the rest as better (n = 23; 18.9%), same (n = 22; 18.0%), and worse (n = 18; 14.8%). No significant differences were noted between acute and delayed HA regarding pain VAS (2.0 vs. 2.5; P = .523), FE (98° vs. 93°; P = .627), ER (30° vs. 23°; P = .215), IR score (4.0 vs. 4.5; P = .589), satisfaction (P = .592), or ASES scores (64.4 vs. 57.1; P = .168). Radiographically, vertical tuberosity reduction was anatomic in 31% of shoulders and adequate horizontal tuberosity reduction was observed in 87.5% of cases. The tuberosity healing rate was 72.6% of shoulders, but tuberosity resorption was also observed in 43.6% of shoulders. Among cemented stems, the tuberosity healing rate was 75.8%, compared to 70.2% for uncemented stems (P = .569).
Postoperative complications were reported in 34 shoulders (27.9%) with rotator cuff failure (n = 6; 4.9%) and superficial infection (n = 6; 4.9%) as the most common complications (Table IV). Comparisons of acute and delayed HA demonstrated similar rates of overall complications (27.1% vs. 28.8%; P = .836). Reoperation occurred in 18 shoulders (14.8%). A larger percentage of HA performed for sequels of fracture underwent reoperation (19.2% vs 11.4%), although with the numbers available this difference did not reach statistical significance (P = .229). In the acute HA cohort, reoperations included revision to a reverse arthroplasty (n = 5; 7.1%), irrigation and débridement for superficial wound breakdown (n = 2; 2.9%), and two stage reimplantation for deep infection (n = 1; 1.4%). In the delayed HA group, reoperations included revision to a reverse arthroplasty (n = 3; 5.8%), irrigation and débridement (n = 3; 5.8%), resection (n = 2; 3.8%), latissimus dorsi/teres major transfer to improve ER (n = 1; 1.9%), and internal fixation of a periprosthetic fracture (n = 1; 1.9%).
Table IV.
Postoperative clinical outcomes categorized by HA timing.
| Acute fracture (n = 70) | Sequels of trauma (n = 52) | Total (n = 122) | P value | |
|---|---|---|---|---|
| VAS daily pain | 2.0 (1.9) | 2.5 (2.4) | 2.2 (2.1) | .523 |
| FE, ° | 97.3 (40.9) | 92.8 (47.1) | 95.5 (43.2) | .627 |
| ER, ° | 30.4 (24.3) | 23.3 (21.3) | 27.6 (23.3) | .215 |
| IR score | 4.0 (1.8) | 4.5 (1.5) | 4.2 (1.7) | .589 |
| Satisfaction | .592 | |||
| Much better | 31 (44.3%) | 28 (53.8%) | 59 (48.4%) | |
| Better | 11 (15.7%) | 12 (23.1%) | 23 (18.9%) | |
| Same | 17 (24.3%) | 5 (9.6%) | 22 (18.0%) | |
| Worse | 11 (15.7%) | 7 (13.5%) | 18 (14.8%) | |
| ASES | 64.4 (17.0) | 57.1 (16.7) | 61.1 (17.1) | .168 |
All values are reported as the mean (standard deviation) or number (percentage) except when noted.
HA, hemiarthroplasty; VAS, visual analog scale; FE, forward elevation; ER, external rotation; IR, internal rotation; ASES, American Shoulder and Elbow Surgeons.
Accounting for death as a competing risk, the cumulative incidences of complications at 1, 5, 10, and 15 years were 19.6% (95% confidence interval [CI],11.9, 26.7), 27.9% (95% CI, 18.8, 35.9), 29.0% (95% CI, 19.7, 37.2), and 36.6% (95% CI, 24.4, 46.9), respectively (Fig. 1). When compared across the acute and delayed cohorts observed rates were 18.1% and 21.9% at 1 year, 24.7% and 33.0% at 5 years, 24.7% and 36.1% at 10 years, and 32.4% and 43.3% at 15 years, respectively. Again, the complication rate was higher at any time point when HA had been performed for sequels of fracture, although these differences did not reach statistical significance (P = .172) (Table V). With respect to reoperation/revision surgery, the cumulative incidence rates with death as a competing risk at 1, 5, 10, and 15 years were 8.2% (95% CI, 2.9, 13.2), 14.5% (95% CI, 7.4, 21.1), 15.7% (95% CI, 8.3, 22.5), and 18.0% (95% CI, 9.4, 25.8), respectively (Fig. 2). When compared across the acute and delayed cohorts observed rates were 3.1% and 15.5% at 1 year, 10.1% and 21.0% at 5 years, 10.1% and 24.0% at 10 years, and 13.7% and 24.0% at 15 years, respectively, again higher rates for HA for sequels of fracture, with not statistically significant differences (P = .098).
Figure 1.
Cumulative incidence of complications with a competing risk of death.
Table V.
Detailed breakdown of complications by group.
| Acute fracture (n = 70) | Sequels of trauma (n = 52) | Total (n = 122) | P value | |
|---|---|---|---|---|
| Rotator cuff failure | 5 (7.1%) | 1 (1.9%) | 6 (4.9%) | |
| Superficial infection | 2 (2.9%) | 4 (7.7%) | 6 (4.9%) | |
| Symptomatic glenoid arthrosis | 1 (1.4%) | 3 (5.8%) | 4 (3.3%) | |
| Greater tuberosity nonunion | 4 (5.7%) | 0 (0%) | 4 (3.3%) | |
| Instability | 3 (4.2%) | 1 (1.9%) | 4 (3.3%) | |
| Deep infection | 1 (1.4%) | 2 (3.8%) | 3 (2.5%) | |
| Periprosthetic humeral fracture | 1 (1.4%) | 2 (3.8%) | 3 (2.5%) | |
| Greater tuberosity resorption | 0 (0%) | 2 (3.8%) | 2 (1.6%) | |
| Neural palsy/neuropathy | 1 (1.4%) | 0 (0%) | 1 (0.8%) | |
| Deep vein thrombosis/Pulmonary embolus | 1 (1.4%) | 0 (0%) | 1 (0.8%) | |
| Total complications | 19 (27.1%) | 15 (28.8%) | 34 (27.9%) | .833 |
All values are reported as the mean (standard deviation) or number (percentage) except when noted.
Figure 2.
Cumulative incidence of reoperation with a competing risk of death.
With the numbers available, a subanalysis on the association of fracture characteristics and clinical outcomes demonstrated no significant correlation between fracture classification, vertical tuberosity reduction, horizontal tuberosity reduction, tuberosity resorption, or glenoid erosion and postoperative pain scores, ASES scores, or need for revision surgery (Table VI). Additionally, patient characteristics (including age, body mass index, cohort, and previous procedure) were not found to be significant risk factors for complications or reoperations inclusive of revision surgery (Table VII). An age-based subanalysis was performed evaluating the complications and reoperation rates by specific age cutoffs. No statistical differences were identified with respect to complications and reoperations when categorizing the cohort according to ages of less than 40, less than 50, or less than 60 years compared to greater than or equal to 40, 50, and 60 years of age (Table VIII).
Table VI.
Fracture characteristics and associations with clinical outcomes.
| Variable | Post op pain |
Forward elevation |
ASES |
Implant revision |
||||
|---|---|---|---|---|---|---|---|---|
| Estimate (95% CI) | P value | Estimate (95% CI) | P value | Estimate (95% CI) | P value | Estimate (95% CI) | P value | |
| Fracture classification | ||||||||
| Neer 2 | Reference | Reference | ||||||
| Neer 3 | 2.41 (−2.08, 6.90) | .291 | Reference | 0.27 (0.01, 12.8) | .513 | |||
| Neer 4 | 1.96 (−2.48, 6.41) | .384 | −0.67 (−15.8, 14.5) | .938 | 0.09 (0.01, 5.65) | .251 | ||
| Malunion | 3.00 (−1.60, 7.60) | .209 | −11.2 (−29.7, 7.25) | .233 | 0.60 (0.02, 30.0) | .795 | ||
| Nonunion | 2.67 (−2.05, 7.38) | .264 | 0.33 (−18.9, 19.6) | .971 | 0.07 (0.00, 12.3) | .326 | ||
| AVN | 1.50 (−3.85, 6.85) | .583 | −1.50 (−39.3, 36.3) | .942 | 0.50 (0.01, 78.2) | .798 | ||
| Previous ORIF | 2.14 (−2.52, 6.81) | .368 | 0.50 (−23.7, 24.7) | .975 | 0.44 (0.01, 27.1) | .694 | ||
| Vertical tuberosity reduction | ||||||||
| High | Reference | Reference | Reference | Reference | ||||
| Anatomic | 0.71 (−0.52, 1.93) | .259 | 10.4 (−12.3, 33.1) | .361 | 1.55 (−11.3, 14.4) | .816 | 3.57 (0.59, 21.6) | .172 |
| Low | −0.17 (−1.56, 1.22) | .817 | 16.2 (−8.5, 41.0) | .204 | 1.08 (−15.0, 17.2) | .892 | 2.59 (0.36, 18.6) | .348 |
| Horizontal tuberosity reduction | ||||||||
| No | Reference | Reference | Reference | Reference | ||||
| Yes | −0.62 (−2.07, 0.83) | .406 | −9.3 (35.6, 17.0) | .487 | 5.86 (−7.63, 19.35) | .395 | 0.94 (0.11, 7.82) | .959 |
| Tuberosity resorption | ||||||||
| None | Reference | Reference | Reference | Reference | ||||
| Partial | −0.66 (−2.01, 0.69) | .341 | −16.3 (−41.5, 8.8) | .203 | 10.2 (−4.21, 24.6) | .162 | 3.09 (0.44, 22.0) | .268 |
| Complete | 0.45 (−0.78, 1.68) | .463 | −14.2 (−36.1, 7.6) | .209 | 5.15 (−6.80, 17.1) | .398 | 2.11 (0.30, 15.1) | .463 |
| Tuberosity healed | ||||||||
| No | Reference | Reference | Reference | Reference | ||||
| Yes | −0.67 (−1.80, 0.45) | .248 | 13.9 (−6.6, 34.5) | .181 | −2.96 (−14.6, 8.73) | .611 | 0.38 (0.08, 1.90) | .242 |
| Glenoid erosion | ||||||||
| 0 | Reference | Reference | Reference | Reference | ||||
| 1 | −0.27 (−1.81, 1.26) | .724 | 6.2 (−22.6, 34.9) | .672 | 9.0 (−7.8, 25.8) | .284 | 0.21 (0.02, 2.42) | .215 |
| 2 | 0.58 (−0.94, 2.09) | .452 | 4.0 (−24.9, 33.0) | .788 | −0.44 (−17.0, 16.1) | .962 | 0.29 (0.04, 2.18) | .238 |
| 3 | 0.41 (−1.26, 2.08) | .629 | 12.6 (−18.2, 43.3) | .425 | 4.58 (−12.9, 22.1) | .601 | 0.47 (0.06, 3.60) | .477 |
| 4 | 0.08 (−4.24, 4.40) | .971 | ||||||
ASES, American Shoulder and Elbow Surgeons; AVN, avascular necrosis; CI, confidence interval; ORIF, open reduction internal fixation.
All values are reported as the variable estimate (95% confidence interval).
Table VII.
Risk factors for complications and reoperations/revisions.
| Variable | Complications |
Reoperations/Revisions |
||
|---|---|---|---|---|
| Hazard ratio (95% CI) | P value | Hazard ratio (95% CI) | P value | |
| Age (per year) | 1.0 (0.97, 1.02) | .856 | 0.97 (0.93, 1.0) | .055 |
| BMI (per unit) | 0.99 (0.94, 1.03) | .535 | 1.01 (0.96, 1.07) | .688 |
| Length of stay (per day) | 1.02 (0.96, 1.07) | .581 | 0.83 (0.66, 1.05) | .124 |
| Gender | .443 | .695 | ||
| Male | Reference | Reference | ||
| Female | 1.37 (0.62, 3.02) | 1.26 (0.41, 3.87) | ||
| Laterality | .222 | .337 | ||
| Right | Reference | Reference | ||
| Left | 1.52 (0.77, 2.98) | 1.61 (0.62, 4.17) | ||
| Dominant | .254 | .682 | ||
| No | Reference | Reference | ||
| Yes | 0.61 (0.27, 1.42) | 0.79 (0.25, 2.44) | ||
| Previous procedure | .108 | .061 | ||
| No | Reference | Reference | ||
| Yes | 1.83 (0.88, 3.80) | 2.52 (0.96, 6.62) | ||
| Bone graft use | .617 | .265 | ||
| None | Reference | Reference | ||
| Yes | 1.21 (0.58, 2.51) | 0.56 (0.21, 1.52) | ||
| Humeral stem cement | .619 | .389 | ||
| No | Reference | Reference | ||
| Yes | 1.22 (0.58, 2.56) | 0.64 (0.24, 1.72) | ||
| Group | .282 | .103 | ||
| Delayed | Reference | Reference | ||
| Acute | 0.69 (0.35, 1.35) | 0.44 (0.17, 1.16) | ||
BMI, body mass index; CI, confidence interval.
All values are reported as the hazard ratio (95% confidence interval).
Table VIII.
Age-based comparisons of complications and reoperations/revisions.
| Variable | Complications |
Reoperations/Revisions |
||
|---|---|---|---|---|
| N (%) | P value | N (%) | P value | |
| Age < 40 y comparison | .409 | .964 | ||
| < 40 (n = 7) | 1 (14.3%) | 1 (14.3%) | ||
| ≥ 40 (n = 115) | 33 (28.7%) | 17 (14.8%) | ||
| Age < 50 y comparison | .552 | .253 | ||
| < 50 (n = 22) | 5 (22.7%) | 5 (22.7%) | ||
| ≥ 50 (n = 100) | 29 (29%) | 13 (13%) | ||
| Age < 60 y comparison | .143 | .627 | ||
| < 60 (n = 41) | 8 (19.5%) | 7 (17.1%) | ||
| ≥ 60 (n = 81) | 26 (32.1%) | 11 (13.6%) | ||
All values are reported as the N and percentage among the group total for each age.
Discussion
Operative management of PHFs continues to present challenges, with an assortment of management strategies and no unified standard for selection of various surgical techniques. HA may play a role in select unreconstructable fracture patterns and their sequels, and the comparative performance of this procedure in the acute vs delayed setting is not well-documented. In the present cohort, early HA performed after acute PHFs and late HA performed for fracture sequalae demonstrated no statistically significant differences with respect to pain, FE, ER, IR, satisfaction, ASES scores, complications, or reoperations. There was a trend for more complications and reoperations when HA had been performed for sequels of trauma, even when adjusted for death as a competing factor, but with the numbers available these differences did not reach statistical significance.
Discussion on timing of proximal head replacement after fracture has been previously debated in the past with conflicting results. In the 1980s, Tanner et al reported outcomes of 43 Neer prostheses demonstrating higher incidences of complications when surgery was delayed for more than three weeks. Similarly, Mighell et al23 observed significantly better function in patients treated within two weeks of the initial injury. By contrast, Prakash et al27 demonstrated no significant differences in ROM in patients who received HA within 30 days of injury compared to more than 30 days. Shortly after, Kontakis et al20 performed a systematic review of available studies prior to 2008 concluding that there is no strong evidence to support the effectiveness of early HA in PHFs. The present investigation reports similar findings that correspond well with the more recent literature on timing of HA after PHF. In general, the results of delayed HA performed for fracture sequalae did not demonstrate strong statistical differences from the results of acute HA.
In the past decade, dissemination of the well-recognized PROPHER (Proximal Fracture of the Humerus: Evaluation by Randomization) trial by Rangan et al28 provided additional information on the surgical management of PHFs, but also potentially misled clinicians to treat severely displaced PHFs conservatively, allowing for harmful fracture sequalae and worse outcomes in select cases.17,26 As such, there has been a recent emergence of studies evaluating the relationship of surgical timing on management of PHFs, with a focus on RSA. Seidl et al31 analyzed 47 RSA (15 acute and 32 delayed) performed after PHFs demonstrating a higher tuberosity healing rate and improved ER in the acute RSA group. These authors further concluded that their findings may be due to the improved ability to reduce and repair the greater tuberosity in the acute setting. In a meta-analysis by Torchia et al36 no differences with respect to forward flexion, clinical outcome scores, or all-cause reoperation were identified between acute or delayed RSA. Most recently, Panagopoulos et al26 assessed 92 RSA (36 early and 56 late) also demonstrating no statistical differences in outcomes between early or late RSA for PHFs. However, they observed a shorter time from injury to regaining pain-free function in the early group, suggestive of a quicker recovery with less pain and discomfort. These recent findings on timing of RSA in PHF are akin to previous investigations of HA timing and continue to support the present study suggesting no notable differences. However, they also provide additional insight that patients undergoing early surgery may benefit from a quicker return to pain-free use of their extremity and activities of daily living sooner than those who undergo late surgery performed for fracture sequalae. In addition, the differences reported in our study regarding complications and reoperations, although not statistically significant, may be viewed by some as clinically relevant.
With respect to the various surgical options for PHFs, studies have reported mixed results. Gupta et al14 performed a systematic review on the surgical management for 3- and 4-part PHFs. After analysis of 92 studies, they demonstrated higher clinical outcome scores and ROM with ORIF compared to HA. However, ORIF also demonstrated a higher reoperation rate than HA and RSA (12.7% vs. 4.9% vs. 5.0%; P < .001). Interestingly, comparison of HA and RSA demonstrated no difference in any outcome measures. Subsequent investigations comparing RSA and HA report a more reproducible improvement of function, better motion, and lower reoperations with the utilization of the RSA.13,32,38 As such, utilization of HA continues to decline and when HA is selected, it is often recommended to use a convertible stem system to allow for revision to RSA.
The present study supports the known literature of modest clinical outcomes after HA.12 At the final follow-up, the majority of patients in this cohort continued to have some pain (VAS of 2), with limited FE to 96°, an ASES of 61.1, a 27.9% complication rate, and a 14.8% reoperation rate. Similar findings were observed in a similar recent investigation, with a 30% complication rate and 15.7% reoperation rate in HA after PHFs.38 However, when death was taken into account as a competing risk, the 15-year cumulative incidences rose to 36.6% and 18.0%, respectively. These elevated rates likely highlight the complexity of fractures treated with HA in this cohort in addition to the difficulty to obtain reliable tuberosity healing despite dedicated efforts at restoration of humeral height and anatomic tuberosity reduction and fixation. Moreover, an age-based analysis for complications and reoperations was performed demonstrating no statistical differences in groups when categorized by age cutoffs of 40, 50, and 60 years of age. As such, use of HA should be selectively utilized, and within our practice we have implemented RSA as the primary implant of choice with these fractures.
This investigation did present specific limitations, most notably those inherent to a retrospective review, including selection bias, a relatively small sample size, and a lack of power analysis due to the non-randomized nature of the study. This was evident in subgroup analyses where larger differences were not statistically significant, likely due to the underpowered nature of the study. Second, the early years of the study period included patients without reliable preoperative documentation of certain clinical values (ROM). This was due to both the longevity of the study, and through transition of the hospital electronic medical record limiting access. As a result, functional outcomes were reported as final postoperative values with limited ability to calculate improvements from preoperative baseline. Third, the delayed cohort consisted of a heterogenous patient population including malunion, nonunion, AVN, and previous ORIF. This was performed because of the small numbers and difficulty making statistical comparisons between the individual patient populations. Regardless of these differences, the analysis of the early and delayed subgroups resulted in the same conclusions. Fourth, this investigation represents the experience of a tertiary center spanning multiple years and including several surgeons, this may introduce bias and confounding variables to our methodology. This includes the evolving indication for HA over the study period as other implants became available.
Conclusion
HA performed acutely for a PHF or in delayed fashion for fracture sequalae demonstrated no statistically significant differences in outcomes of all examined parameters. HA in this setting may provide reasonable pain relief. However, limited motion and relatively poor ASES scores can be expected. Additionally, there was no clear correlation of outcomes based on fracture classification or the healing of the humeral tuberosities.
Disclaimers
Funding: No funding was disclosed by the authors.
Conflicts of interest: The authors, their immediate families, and any research foundation 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.
Footnotes
Mayo Institutional Review Board approved this study (IRB # 12-007498).
References
- 1.Acevedo D.C., VanBeek C., Lazarus M.D., Williams G.R., Abboud J.A. Reverse shoulder arthroplasty for proximal humeral fractures: update on indications, technique, and results. J Shoulder Elbow Surg. 2014;23:279–289. doi: 10.1016/j.jse.2013.10.003. [DOI] [PubMed] [Google Scholar]
- 2.Barlow J.D. Complications after ORIF of proximal humerus fractures—how to deal with them. Ann Joint. 2020;6 doi: 10.21037/aoj-20-77. [DOI] [Google Scholar]
- 3.Barlow J.D., Logli A.L., Steinmann S.P., Sems S.A., Cross W.W., Yuan B.J., et al. Locking plate fixation of proximal humerus fractures in patients older than 60 years continues to be associated with a high complication rate. J Shoulder Elbow Surg. 2020;29:1689–1694. doi: 10.1016/j.jse.2019.11.026. [DOI] [PubMed] [Google Scholar]
- 4.Bell J.-E., Leung B.C., Spratt K.F., Koval K.J., Weinstein J.D., Goodman D.C., et al. Trends and variation in incidence, surgical treatment, and repeat surgery of proximal humeral fractures in the elderly. J Bone Joint Surg Am. 2011;93:121–131. doi: 10.2106/jbjs.I.01505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Beredjiklian P.K., Iannotti J.P., Norris T.R., Williams G.R. Operative treatment of malunion of a fracture of the proximal aspect of the humerus. J Bone Joint Surg Am. 1998;80:1484–1497. doi: 10.2106/00004623-199810000-00010. [DOI] [PubMed] [Google Scholar]
- 6.Boesmueller S., Wech M., Gregori M., Domaszewski F., Bukaty A., Fialka C., et al. Risk factors for humeral head necrosis and non-union after plating in proximal humeral fractures. Injury. 2016;47:350–355. doi: 10.1016/j.injury.2015.10.001. [DOI] [PubMed] [Google Scholar]
- 7.Carbone S., Papalia M. The amount of impaction and loss of reduction in osteoporotic proximal humeral fractures after surgical fixation. Osteoporos Int. 2016;27:627–633. doi: 10.1007/s00198-015-3304-x. [DOI] [PubMed] [Google Scholar]
- 8.Carofino B.C., Leopold S.S. Classifications in brief: the neer classification for proximal humerus fractures. Clin Orthop Relat Res. 2013;471:39–43. doi: 10.1007/s11999-012-2454-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Court-Brown C.M., Caesar B. Epidemiology of adult fractures: a review. Injury. 2006;37:691–697. doi: 10.1016/j.injury.2006.04.130. [DOI] [PubMed] [Google Scholar]
- 10.Cruess R.L. Experience with steroid-induced avascular necrosis of the shoulder and etiologic considerations regarding osteonecrosis of the hip. Clin Orthop Relat Res. 1978:86–93. [PubMed] [Google Scholar]
- 11.Flurin P.-H., Marczuk Y., Janout M., Wright T.W., Zuckerman J., Roche C.P. Comparison of outcomes using anatomic and reverse total shoulder arthroplasty. Bull Hosp Jt Dis (2013) 2013;71:101–107. [PubMed] [Google Scholar]
- 12.Freeman T.R., Dunn R.H., Ko K.J.W., Seidl A.J. Hemiarthroplasty for proximal humerus fracture—a dying art. Ann Jt. 2020;6:15. doi: 10.21037/aoj-2019-mfas-08. [DOI] [Google Scholar]
- 13.Gallinet D., Ohl X., Decroocq L., Dib C., Valenti P., Boileau P. Is reverse total shoulder arthroplasty more effective than hemiarthroplasty for treating displaced proximal humerus fractures in older adults? A systematic review and meta-analysis. Orthop Traumatol. 2018;104:759–766. doi: 10.1016/j.otsr.2018.04.025. [DOI] [PubMed] [Google Scholar]
- 14.Gupta A.K., Harris J.D., Erickson B.J., Abrams G.D., Bruce B., McCormick F., et al. Surgical management of complex proximal humerus fractures—a systematic review of 92 studies including 4500 patients. J Orthop Trauma. 2015;29:54–59. doi: 10.1097/bot.0000000000000229. [DOI] [PubMed] [Google Scholar]
- 15.Han R.J., Sing D.C., Feeley B.T., Ma C.B., Zhang A.L. Proximal humerus fragility fractures: recent trends in nonoperative and operative treatment in the Medicare population. J Shoulder Elbow Surg. 2016;25:256–261. doi: 10.1016/j.jse.2015.07.015. [DOI] [PubMed] [Google Scholar]
- 16.Handoll H.H., Brorson S. Interventions for treating proximal humeral fractures in adults. Cochrane Database Syst Rev. 2015:CD000434. doi: 10.1002/14651858.CD000434.pub4. [DOI] [PubMed] [Google Scholar]
- 17.Jefferson L., Brealey S., Handoll H., Keding A., Kottam L., Sbizzera I., et al. Impact of the PROFHER trial findings on surgeons' clinical practice: an online questionnaire survey. Bone Joint Res. 2017;6:590–599. doi: 10.1302/2046-3758.610.Bjr-2017-0170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jung S.W., Shim S.B., Kim H.M., Lee J.H., Lim H.S. Factors that influence reduction loss in proximal humerus fracture surgery. J Orthop Trauma. 2015;29:276–282. doi: 10.1097/bot.0000000000000252. [DOI] [PubMed] [Google Scholar]
- 19.Klug A., Wincheringer D., Harth J., Schmidt-Horlohé K., Hoffmann R., Gramlich Y. Complications after surgical treatment of proximal humerus fractures in the elderly-an analysis of complication patterns and risk factors for reverse shoulder arthroplasty and angular-stable plating. J Shoulder Elbow Surg. 2019;28:1674–1684. doi: 10.1016/j.jse.2019.02.017. [DOI] [PubMed] [Google Scholar]
- 20.Kontakis G., Koutras C., Tosounidis T., Giannoudis P. Early management of proximal humeral fractures with hemiarthroplasty: a systematic review. J Bone Joint Surg Br. 2008;90:1407–1413. doi: 10.1302/0301-620x.90b11.21070. [DOI] [PubMed] [Google Scholar]
- 21.McLean A.S., Price N., Graves S., Hatton A., Taylor F.J. Nationwide trends in management of proximal humeral fractures: an analysis of 77,966 cases from 2008 to 2017. J Shoulder Elbow Surg. 2019;28:2072–2078. doi: 10.1016/j.jse.2019.03.034. [DOI] [PubMed] [Google Scholar]
- 22.Mease S.J., Kraeutler M.J., Gonzales-Luna D.C., Gregory J.M., Gardner M.J., Choo A.M. Current controversies in the treatment of geriatric proximal humeral fractures. J Bone Joint Surg Am. 2021;103:829–836. doi: 10.2106/jbjs.20.00665. [DOI] [PubMed] [Google Scholar]
- 23.Mighell M.A., Kolm G.P., Collinge C.A., Frankle M.A. Outcomes of hemiarthroplasty for fractures of the proximal humerus. J Shoulder Elbow Surg. 2003;12:569–577. doi: 10.1016/S1058-2746(03)00213-1. [DOI] [PubMed] [Google Scholar]
- 24.Neer Cn. Displaced proximal humeral fractures. Orthop Trauma Dir. 2007;5:25–29. [Google Scholar]
- 25.Palvanen M., Kannus P., Niemi S., Parkkari J. Update in the epidemiology of proximal humeral fractures. Clin Orthop Relat Res. 2006;442:87–92. doi: 10.1097/01.blo.0000194672.79634.78. [DOI] [PubMed] [Google Scholar]
- 26.Panagopoulos G.N., Pugliese M., Leonidou A., Butt F., Jaibaji M., Megaloikonomos P.D., et al. Acute versus delayed reverse total shoulder arthroplasty for proximal humeral fractures: a consecutive cohort study. J Shoulder Elbow Surg. 2022;31:276–285. doi: 10.1016/j.jse.2021.07.003. [DOI] [PubMed] [Google Scholar]
- 27.Prakash U., McGurty D.W., Dent J.A. Hemiarthroplasty for severe fractures of the proximal humerus. J Shoulder Elbow Surg. 2002;11:428–430. doi: 10.1067/mse.2002.126615. [DOI] [PubMed] [Google Scholar]
- 28.Rangan A., Handoll H., Brealey S., Jefferson L., Keding A., Martin B.C., et al. Surgical vs nonsurgical treatment of adults with displaced fractures of the proximal humerus: the PROFHER randomized clinical trial. JAMA. 2015;313:1037–1047. doi: 10.1001/jama.2015.1629. [DOI] [PubMed] [Google Scholar]
- 29.Rosas S., Law T.Y., Kurowicki J., Formaini N., Kalandiak S.P., Levy J.C. Trends in surgical management of proximal humeral fractures in the Medicare population: a nationwide study of records from 2009 to 2012. J Shoulder Elbow Surg. 2016;25:608–613. doi: 10.1016/j.jse.2015.08.011. [DOI] [PubMed] [Google Scholar]
- 30.Schultz B.J., Lowe D.T., Egol K.A., Zuckerman J.D. Shoulder hemiarthroplasty for proximal humerus fracture. J Orthop Trauma. 2021;35:S3–S4. doi: 10.1097/bot.0000000000002158. [DOI] [PubMed] [Google Scholar]
- 31.Seidl A., Sholder D., Warrender W., Livesey M., Williams G., Abboud J., et al. Early versus late reverse shoulder arthroplasty for proximal humerus fractures: does it matter? Arch Bone Jt Surg. 2017;5:213–220. [PMC free article] [PubMed] [Google Scholar]
- 32.Shukla D.R., McAnany S., Kim J., Overley S., Parsons B.O. Hemiarthroplasty versus reverse shoulder arthroplasty for treatment of proximal humeral fractures: a meta-analysis. J Shoulder Elbow Surg. 2016;25:330–340. doi: 10.1016/j.jse.2015.08.030. [DOI] [PubMed] [Google Scholar]
- 33.Sperling J.W., Cofield R.H., Schleck C.D., Harmsen W.S. Total shoulder arthroplasty versus hemiarthroplasty for rheumatoid arthritis of the shoulder: results of 303 consecutive cases. J Shoulder Elbow Surg. 2007;16:683–690. doi: 10.1016/j.jse.2007.02.135. [DOI] [PubMed] [Google Scholar]
- 34.Spross C., Zeledon R., Zdravkovic V., Jost B. How bone quality may influence intraoperative and early postoperative problems after angular stable open reduction-internal fixation of proximal humeral fractures. J Shoulder Elbow Surg. 2017;26:1566–1572. doi: 10.1016/j.jse.2017.02.026. [DOI] [PubMed] [Google Scholar]
- 35.Spross C., Zdravkovic V., Manser M., Farei-Campagna J.M., Jacxsens M., Jost B. Outcomes of management of proximal humeral fractures with patient-specific, evidence-based treatment algorithms. J Bone Joint Surg Am. 2021;103:1906–1916. doi: 10.2106/jbjs.20.01309. [DOI] [PubMed] [Google Scholar]
- 36.Torchia M.T., Austin D.C., Cozzolino N., Jacobowitz L., Bell J.-E. Acute versus delayed reverse total shoulder arthroplasty for the treatment of proximal humeral fractures in the elderly population: a systematic review and meta-analysis. J Shoulder Elbow Surg. 2019;28:765–773. doi: 10.1016/j.jse.2018.10.004. [DOI] [PubMed] [Google Scholar]
- 37.Valenti P., Aliani D., Maroun C., Werthel J.D., Elkolti K. Shoulder hemiarthroplasty for proximal humeral fractures: analysis of clinical and radiographic outcomes at midterm follow-up: a series of 51 patients. Eur J Orthop Surg Traumatol. 2017;27:309–315. doi: 10.1007/s00590-017-1927-7. [DOI] [PubMed] [Google Scholar]
- 38.Yahuaca B.I., Simon P., Christmas K.N., Patel S., Gorman R.A., Mighell M.A., et al. Acute surgical management of proximal humerus fractures: ORIF vs. hemiarthroplasty vs. reverse shoulder arthroplasty. J Shoulder Elbow Surg. 2020;29:S32–S40. doi: 10.1016/j.jse.2019.10.012. [DOI] [PubMed] [Google Scholar]