ORIGINAL ARTICLE
Proximal Humeral Fractures Treated Conservatively Settle During Fracture Healing Antonio M. Foruria, MD, PhD,* Milagros Martí, MD, PhD,† and Joaquin Sanchez-Sotelo, MD, PhD‡
Objectives: Determine the relative position of the main fractured fragments of proximal humerus fractures treated conservatively to compare displacement at 2 time points: injury (0), and 1 year later (1).
Design: Prospective, comparative cohort study. Setting: Level I trauma center. Patients: Eighty-nine prospectively enrolled adults. Intervention: Six weeks of sling immobilization and a homebased program rehabilitation protocol started 2 weeks after injury.
Main Outcome Measurements: Standardized radiographs of injured shoulders were obtained in all patients at times 0 and 1. Computed tomography scans were also obtained at these times in 73 cases. Forty-two computer-assisted measurements of displacement were performed at times 0 and 1 and then compared. Factors related to progression of displacement were analyzed. Results: Ninety percent of fractures were classified into 1 of 4 patterns: posteromedial (varus) impaction (46), lateral (valgus) impaction (13), isolated greater tuberosity (15), and anteromedial impaction (6). Head– shaft displacement increased over time. In posteromedial impaction fractures, average fracture settling included 9 degrees in varus, 7 degrees in retroversion, and 3.2 mm in posterior shortening. In valgus-impacted fractures, a decrease in valgus tilt and a tendency toward a more anterior orientation of the articular surface was observed. Greater tuberosity displacement increased more than 5 mm in less than 20% of cases. Age and initial displacement were related to progression of displacement. Conclusions: Proximal humerus fractures treated conservatively settle at the head–shaft junction during healing. Substantial additional displacement of tuberosities was seldom observed.
INTRODUCTION Currently, early mobilization is the mainstay of conservative treatment of proximal humeral fractures.1 Several articles have been published showing the benefits of early versus delayed mobilization,2 encouraging the orthopaedic community to narrow the time that patients should be completely immobilized before starting a range of motion exercises. Although such advocated early mobilization protocols often start even before soft callus formation occurs,3 especially in impacted surgical neck fractures,4 nonunion rates remained low. We demonstrated that fracture pattern and initial fragment displacement is a main factor determining the final outcome after 1 year of standard conservative treatment of proximal humerus fractures5; initial fragment displacement and proximal humeral relationships with surrounding structures, measured with dedicated software on image test, could explain most but not all interpatient outcome variability. Other factors may influence final outcome, so predictions based on initial fracture morphology may not be completely accurate. We hypothesized that fragments displacement and anatomic proximal humeral relationships may change during the fracture healing process in the setting of standard conservative treatment protocols including early mobilization. Our primary aim was to analyze morphologic changes that occur during the healing process of proximal humerus fractures treated with standard conservative protocols, focusing on identifying possible fragment displacement progression and changes in anatomic relationships of the proximal humerus. Our secondary aim was to identify factors associated with progression of displacement during the fracture consolidation process.
Key Words: proximal humeral fractures, fracture healing, conservative treatment
Level of Evidence: Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence. (J Orthop Trauma 2015;29:e24–e30) Accepted for publication June 26, 2014. From the *Department of Orthopedic Surgery, Fundación Jiménez Díaz, Madrid, Spain; †Department of Radiology, La Paz University Hospital, Madrid, Spain; and ‡Department of Orthopedic Surgery, The Mayo Clinic, Rochester MN. The authors report no conflict of interest. Reprints: Antonio M. Foruria, MD, PhD, Department of Orthopaedic Surgery, Shoulder and Elbow Surgery Unit, Fundación Jiménez Díaz, Avda Reyes Católicos, 2, Spain, Madrid 28040 (e-mail: [email protected]). Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
e24
| www.jorthotrauma.com
PATIENTS AND METHODS Patients The study proceeded after approval from the Institutional Review Board of La Paz University Hospital (identification code: PI-382) as part of a large prospective research project focusing on prognostic factors of conservatively treated proximal humerus fractures.5 All patients with a proximal humerus fracture evaluated at the Emergency Department of a single institution between March and September 2005 were considered for inclusion into our study. The treating orthopaedic surgeon on call recommended nonoperative or surgical treatment based on morphology of the fracture, age of the patient, functional demands, and comorbidities. J Orthop Trauma Volume 29, Number 2, February 2015
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Patients selected for conservative treatment were asked to participate in the study if they were aged 18 years or older, had a unilateral fracture, had no other associated injuries or previous fracture involving the shoulder, and had a preserved cognitive function as determined by the Pfeiffer test—6 patients excluded.6 During this period, 132 patients were evaluated, in whom surgical treatment was proposed and accepted in 10 patients. Of the remaining 122, 111 fulfilled the inclusion criteria and were included in the study. All gave informed consent. Within the first 12 months after their fracture, 7 died as a result of unrelated causes and 11 were lost to follow-up. The remaining 93 patients were followed up prospectively for 1 year after their fracture and form the basis of this study. There were 70 females (79%) and 19 males, with an average age of 70 6 13 years (range, 26–93 years). The average age was 73 years for females and 59 years for males. The mechanism of injury was a fall from a standing height in 75 shoulders (84%), high-energy injury (motor-vehicle accident of fall from a height) in 8 shoulders (9%), and a sportrelated injury in 6 shoulders (7%). Comorbidities included heart disease (15 cases, 17%), diabetes mellitus (10 cases, 11%), gastrointestinal disease (9 cases, 10%), and pulmonary disease (7 cases, 8%).
Treatment Protocol Treatment was identical in all patients. They were instructed and supervised by the same orthopaedic surgeon (A.M.F.). The affected upper extremity was protected in a commercially available shoulder immobilizer for the first 2 weeks after the fracture. A home program of passive and active assisted range of movement exercises was then initiated within the range of pain tolerance for 4 more weeks, protecting the arm in the sling between the exercise periods. The immobilizer was removed and stretching exercises were added at 6 weeks, and shoulder elevation and external and internal rotation elastic rubber band strengthening exercises7 were started at 12 weeks and continued for 3 months.
Radiologic Evaluation All radiographs were standardized according to an established protocol and carried out under direct supervision by the same physician (A.M.F.). The projections included an anteroposterior (AP) radiograph in the plane of the scapula with the arm in 20 degree of external rotation, a lateral “y view” with the forearm parallel to the image receptor, and an axillary view taken with the patient standing with the x-ray beam coming from the floor toward the axilla, and the image receptor over the patient’s shoulder, the arm being separated from the body passively as in a pendulum exercise.7 Radiographs were obtained for both shoulders at the time of the fracture, and for the affected shoulder at 3, 6, and 12 months. Additional AP and lateral radiographs were taken at 2 and 6 weeks to assess possible gross progression of displacement. A computed tomography (CT) scan, with 3D and 3-planar reconstruction referenced to the scapular plane, of the affected shoulders were also obtained in all patients immediately after the injury, and in 73 of the 89 shoulders (82%) at 12 months. Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Proximal Humeral Fractures Settle
All the digital radiographs and CT scans were corrected for magnification. Morphology of each individual fracture was analyzed based on these, including spatial rotation of 3D reconstructions. Commercially available software (AutoCAD 2004; Autodesk Inc, San Rafael, CA) was used to calculate the measurements listed in Table 1 for both the radiographs and CT reconstructions (Figs. 1A–F). The diaphyseal humeral axis was used to define inferior, superior, medial, and lateral on the AP and lateral radiographs and on the coronal and sagittal CT reconstructions; the measurements were referenced to this axis. The perpendicular to the glenoid surface was used to define anterior, posterior, medial, and lateral on the axillary radiographs and on the axial CT reconstructions, and the measurements were also referenced to this axis. For each individual fracture, the CT scan that showed that the maximum displacement was selected for each individually analyzed fragment. Measurements obtained from image test obtained at the moment of the injury were compared with those performed at 1 year of follow-up. The same researcher (M.M.) formatted all the image tests and another (A.M.F.) performed all the measurements. Fractures were classified according to our morphologic classification system described elsewhere5 that included 4 main fracture patterns: posteromedial (varus) impacted fractures, lateral (valgus) impacted fractures, isolated greater tuberosity fractures, and anteromedial impacted fractures; injuries were also classified according to Neer8 and OTA9 criteria.
Statistical Analysis Normal distribution of quantitative variables was analyzed with the Kolmogorov–Smirnov test. Differences in measurements performed at baseline and at 1 year of followup were compared with the paired 2-sampled Student t or the Wilcoxon signed-rank tests accordingly. No data was available before the study to calculate the sample size; post hoc sample size calculation was performed after data were available.10 Effect size was also calculated (general linear modeling, partial eta squared: h2p). Correlations between quantitative variables (age and progression of displacement, initial displacement and progression of displacement) were tested with the Pearson product-moment and the Spearman’s rank correlation coefficients.
RESULTS Radiographs and CT scans at 1 year showed fracture union in 87 cases (98%). Two patients developed a nonunion at the surgical neck level. There was radiographic evidence of avascular necrosis in 6 patients, being partial in 5 patients with minimal symptoms, and most of the humeral head in the remainder patient who complained of severe pain.
Fracture Classification There were 46 (52%) posteromedial (varus) neck impaction fractures; of these, 25 (54%) had a fractured greater tuberosity, and 3 (6%) had both tuberosities fractured. There were 13 (14%) lateral (valgus) impaction fractures, all with an associated greater tuberosity fracture and 5 (38%) with both www.jorthotrauma.com |
e25
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Foruria et al
TABLE 1. Measurements Performed Both in CT Reconstructions and X-Rays for Each Patient Measurements Performed on CT Reconstructions and X-Rays Measurements Axial CT reconstruction-axillary x-ray GT-AS overlapping angle and % LT-AS overlapping angle and % GT-glenoid rim distance LT-glenoid rim distance GT-LT tuberosity distance and angle GT medial displacement LT medial displacement Cephalic axis–glenoid axis angle Cephalic axis–diaphyseal axis angle Coronal CT reconstruction-AP x-ray Cephalodiaphyseal angle GT-AS Medial–lateral translation Medial impaction Lateral impaction GT acromion distance AS acromion distance Sagittal CT reconstruction lateral x-ray Cephalodiaphyseal angle Translation in AP plane Flexion—extension of surgical neck Anterior impaction Posterior impaction Cephalic axis–scapular axis angle Diaphyseal axis–scapular axis angle
Definition Arc of overlapping of the GT over the AS referenced to the center of the AS, and percentage of the AS covered by the GT Arc of overlapping of the LT over the AS referenced to the center of the AS, and percentage of the AS covered by the LT Lesser distance between the GT and the posterior glenoid rim Lesser distance between the LT and the anterior glenoid rim Lesser distance between the GT and the LT and arc of displacement between both tuberosities referenced to the center of the AS Medial displacement of the GT Medial displacement of the LT Angle between the axis of the humeral AS and the perpendicular to the glenoid AS Angle between the axis of the AS and the diaphyseal axis. Only in x-rays Angle between the diaphyseal axis and the axis of the AS Distance between the highest points of the AS and the GT perpendicularly projected on the diaphyseal axis Medial or lateral displacement of the head related to the diaphysis measured at the medial cortical level Magnitude of medial surgical neck cortex impaction/distraction Magnitude of lateral surgical neck cortex impaction/distraction Lesser distance between the GT and the acromion Lesser distance between the AS and the acromion Angle between the diaphyseal axis and the axis of the AS Anterior or posterior displacement of the head related to the diaphysis measured at the posterior cortical level Angle between head and diaphyseal longitudinal axis Magnitude of anterior surgical neck cortex impaction/distraction Magnitude of posterior surgical neck cortex impaction/distraction Angle between the axis of the AS and the longitudinal axis of the scapula. Only performed in x-rays Angle between the diaphyseal axis and the longitudinal axis of the scapula. Only performed in x-rays
The numbers correspond to those between brackets in Figure 1. The diaphyseal axis was used to define what is inferior or superior and medial or lateral in AP and lateral x-rays, and coronal and sagittal CT reconstructions, being measurements referenced to this axis. The perpendicular to the glenoid surface was used to define anterior, posterior, medial, and lateral on axillary x-rays and axial CT reconstructions, being measurements referenced to this axis. AS, articular surface; GT, greater tuberosity; LT, lesser tuberosity. Reproduced with permission from Foruria et al.5 Copyright @2011, the British Editorial Society of Bone and Joint Surgery.
tuberosities fractured. Fifteen (17%) patients had an isolated greater tuberosity fracture; 3 of these presented at the emergency department with an associated dislocation that was conservatively reduced by the on-call team. Six (7%) shoulders had an anteromedial surgical neck impaction fracture. The remaining 9 (10%) fractures did not fit into any of the 4 common patterns described above. After analyzing x-rays and CT scans, 36 (40.4%) fractures were considered to be displaced according to Neer8 criteria. Table 2 shows the injury distribution according to the OTA classification.9 When initial and final simple x-rays and CT scans were compared, we observed that fractures settled at the surgical neck level during healing, increasing their deformity as a result of the progression of their initial displacement. In posteromedial surgical neck impaction fractures (46 cases, characterized by varus and posterior impaction at the surgical neck of the humerus, and increased retroversion
of the humeral head), x-rays and CT scans showed an increase of the tridimensional deformity and a change in the relationships of the proximal humerus with the acromion (Table 3): 1. Increase in varus deformity: with a mean decrease in the cephalodiaphyseal angle on AP x-rays of 8.6 6 8 degrees (median, 7; range, 29 to 27; P , 0.001), and a mean increase in the medial impaction of the surgical neck on the same projection of 3.5 6 5.7 mm (median, 2.2; range, 27 to 24; P , 0.001). 2. Increase in humeral head retroversion: with a mean increase in the cephalic axis–glenoid axis angle on axial CT scans of 7 6 17 degrees (median, 10; range, 242 to 33; P = 0.014), a decrease in the distance between the lesser tuberosity and the anterior glenoid rim of 3.8 6 6 mm (median, 4.5; range, 29 to 14; P , 0.001), and a mean increase in the cephalodiaphyseal angle in the sagittal CT of 13 6 39 degrees (median, 7; range, 271 to 117; P = 0.049), toward a more posterior–inferior orientation of the cephalic axis.
e26
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Progression of Deformity
| www.jorthotrauma.com
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Proximal Humeral Fractures Settle
FIGURE 1. Examples of computed-assisted measurements performed after correction for magnification. Between brackets is written the number corresponding to measurements represented in Table 1. A, AP x-ray. B, Lateral x-ray. C, Axillary x-ray. D, Axial CT reconstruction. E, Coronal CT reconstruction. F, Sagittal CT reconstruction. Reproduced with permission from Foruria et al.5 Copyright @2011, the British Editorial Society of Bone and Joint Surgery.
3. Increase in the mean posterior surgical neck impaction on the sagittal CT of 3.2 6 6 mm (median, 2.1; range, 29 to 19; P = 0.003). 4. A decrease in the distance between the acromion and the humerus on the AP x-rays: with a mean decrease of TABLE 2. Distribution of Fractures According to the OTA Classification System Fracture Type
N
%
A11 A12 A13 A21 A22 A23 A32 A33 B11 B13 B21 C21 C23 Nonclassifiable Total
9 1 4 6 19 4 2 1 28 2 1 3 1 8 89
10.1 1.1 4.5 6.7 21.3 4.5 2.2 1.1 31.5 2.2 1.1 3.4 1.1 9.0 100.0
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
the distance between the humeral articular surface and the acromion of 3.9 6 5 mm (median, 2.8; range, 25 to 21; P , 0.001), and a mean decrease of the distance between the greater tuberosity and the acromion of 4.5 6 5 mm (median, 4.6; range, 24 to 15; P , 0.001). Lateral (valgus) impaction fractures (13 cases) showed a mean reduction of valgus tilt of the articular surface– cephalodiaphyseal angle of 3.3 6 6 degrees (median, 4; range, 215 to 12; P = 0.023) in AP x-rays. On lateral x-rays, articular surfaces showed a tendency to a more anterior orientation cephalodiaphyseal angle after consolidation (mean, 5.2 6 8 degrees; median range, 25 to 24; P = 0.061). There were no significant differences on AP or mediolateral translations of the articular surface with respect to the diaphysis, or on impaction of the diaphysis inside the head segment. Fractured greater tuberosities (63 cases) remained stable in the majority of the cases; mean differences in displacement between initial and 1-year image tests, measured as the medial displacement of the greater tuberosity, the distance from the greater tuberosity to the glenoid rim or the distance between tuberosities, were lesser than 2 mm for the 3 measurements and not statistically significant (P = 0.34, P = 0.9, and P = 0.15, respectively). Six patients (9%) showed an increase of medial displacement of the greater tuberosity greater than 5 mm (range, 5.5–16.2 mm). Eleven cases (17%) showed www.jorthotrauma.com |
e27
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Foruria et al
TABLE 3. Comparison of Measurements Performed on Image Tests of Patients With Posteromedial Impaction Fractures (n = 46) at 2 Time Points: When the Fracture Occurred (0) and at 1 Year of Follow-up (1) Measurement AP x-rays Cephalodiaphyseal angle, degrees GT-AS distance, mm AS-AC distance, mm Medial impaction, mm Lateral x-rays Cephalodiaphyseal angle, degrees GT-AC distance, mm AS-AC distance, mm Posterior impaction, mm Axial CT LT-glenoid distance, mm Cephalic axis–glenoid axis angle, degrees Coronal CT GT-AC distance, mm AS-AC distance, mm Sagittal CT Cephalodiaphyseal angle, degrees Posterior impaction, mm
Mean 6 SD
Means Difference
P
h2p
(0) 117 6 19, (1) 109 6 18 (0) 0.1 6 5, (1) 21 6 5 (0) 18 6 6, (1) 14 6 5 (0) 12.6 6 8, (1) 16.1 6 10
8.6 1 24 3.5
5.9 0.01 22.5 1.7
to to to to
11.2 2.2 24.5 5.2
,0.001 0.048 ,0.001 ,0.001
0.5 0.1 0.4 0.3
(0) 74 6 25, (1) 85.4 6 26 (0) 14.3 6 5, (1) 9.7 6 3 (0) 15.9 6 5, (1) 12.2 6 5 (0) 13.2 6 11, (1) 16.4 6 11
11.4 24.5 23.7 3.2
5.7 22.7 21.9 0
to to to to
17 26.3 25.5 6.5
,0.001 ,0.001 ,0.001 0.05
0.3 0.4 0.3 0.1
(0) 19.3 6 5, (1) 15.5 6 5 (0) 31.5 6 20, (1) 38.6 6 18
23.8 7
21.9 to 25.7 1.5 to 12.6
,0.001 0.014
0.3 0.2
(0) 8.6 6 4, (1) 6.1 6 4 (0) 9.3 6 4, (1) 8.1 6 4
22.4 21.2
20.7 to 24.1 0 to 22.4
0.007 0.05
0.2 0.1
(0) 54.7 6 38, (1) 67.8 6 33 (0) 11.6 6 9, (1) 14.8 6 1
13 3.1
0.6 to 26 1.2 to 5.1
0.049 0.003
0.1 0.2
95% CI
AC, acromion; AS, articular surface; CI, confidence interval; GT, greater tuberosity; h2p, partial eta squared representing effect size.
a decrease in the distance between the greater tuberosity and the posterior glenoid rim greater than 5 mm (range, 5.6–30.5). Eight (13%) patients showed a more superior position of the greater tuberosity in relation to the articular surface greater than 5 mm (range, 5.1–25), as a consequence of increased varus tilt of the humeral head, leaving the greater tuberosity more superiorly located as opposed to a superior migration of the isolated tuberosity fragment. The majority of fractured lesser tuberosities (9 cases) remained also stable, without significant progression of displacement. Only 1 case showed progression of medial displacement of the lesser tuberosities of 5 mm.
1 year of follow-up. On the contrary, for superposition of the greater tuberosity over the articular surface (posterior displacement), the lesser the initial displacement, the greater the progression of displacement, but this last result could be influenced by the increment of head retroversion described above (Table 4).
DISCUSSION
Age was associated with progression of surgical neck displacement in posteromedial impaction fractures: the older the patient, the greater the increase of medial impaction (P = 0.04; correlation coefficient, 0.31), the greater the increase of posterior impaction (P , 0.001; correlation coefficient, 0.57), and the greater the increase of retroversion (the lesser tuberosity–glenoid distance decreased more with age, P = 0.023, correlation coefficient 0.37; and the angle formed by the cephalic axis and the perpendicular to the glenoid surface on axial CT showed a tendency toward greater increase with age, P = 0.059; correlation coefficient, 0.31). Gender did not show correlation with progression of displacement in the context of a lack of statistical power. On posteromedial impaction fractures, initial displacement at the surgical neck level correlated with progression of deformity; for a given measurement, the greater the initial displacement, the greater the further displacement observed at
We verified our main hypothesis: fracture deformity increases during standard conservative treatment of proximal humerus fractures. Classic conservative treatment protocols underline the necessity of physical examination before starting the exercise program to test the stability of the fracture and avoid possible fragment migration and eventually the development of a nonunion.11 However, fracture instability may not be observed during the examination unless gross surgical neck instability is present. The recent description of proximal humerus fractures as impacted in the majority of the cases,5,12 and the existence of articles supporting even immediate mobilization of impacted injuries,4 could make the orthopaedic community think that there is no risk of further displacement when conservative treatment is implemented. Surgical neck impaction could be responsible for inability to identify fragment motion during physical examination; but the increase of deformity might be the consequence of a single event (like a fall) or due to the continuous cyclic stress forces crossing the fracture during mobilization protocols, in conjunction with the pull of periarticular muscles on fracture fragments. This cyclic stress could wear off the resistance of humeral head trabecular bone, allowing the diaphysis to penetrate progressively
e28
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Factors Related to Progression of Deformity
| www.jorthotrauma.com
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
TABLE 4. For a Given Measurement, Correlations Between Initial Value and the Difference Obtained From Subtracting the Final Value From the Initial One in Posteromedial Impaction Fractures (n = 46) Measurement AP x-rays Cephalodiaphyseal angle, degrees GT-AS distance, mm AS-AC distance, mm Lateral x-rays GT-AC distance, mm AS-AC distance, mm Axial CT LT-glenoid distance, mm Cephalic axis–glenoid axis angle, degrees GT-AS superposition angle, degrees GT-glenoid distance Coronal CT GT-AC distance, mm AS-AC distance, mm Sagittal CT Cephalodiaphyseal angle, degrees
Pearson Correlation Coefficient
P
0.31
0.034
0.35 0.62
0.017 ,0.001
0.82 0.58
,0.001 ,0.001
0.56 0.52
,0.001 0.001
20.38
0.019
0.42
0.009
0.54 0.37
,0.001 0.023
0.62
,0.001
AC, acromion; AS, articular surface; GT, greater tuberosity; LT, lesser tuberosity.
deeper on it in the same direction the initial deformity was produced, as it is seen on our results. Early mobilization protocols continue to be the mainstay of conservative treatment. However, we know from our previous study5 that the greater the initial deformity, the greater the pain and the poorer the outcome after conservative treatment of proximal humerus fractures. In this context, we believe that the increase in deformity occurring during the healing process could play a relevant role on final outcome in some patients. For this reason, we think it is important to identify the subset of patients with higher risk of important progression of displacement and isolate them from the majority of patients in which such a risk would not be clinically relevant. We identified elderly patients and those with greater initial displacement as the ones in which the progression of the deformity could be worse, both factors probably related to poor bone quality. Age has been also described as a risk factor for sustaining a more displaced fracture12; moreover, age is an independent factor on treatment decision making, and the older the patient, surgery is offered less likely for treatment.1 All these factors in combination could increase the incidence of malunion resulting in a poorer outcome in older patients. This mentioned “at risk” group could hypothetically benefit of either surgery if they are physically fit or a delayed treatment protocol (eg, until soft callus formation at 3–4 weeks after injury), as other authors have described after proximal humerus open reduction and internal fixation to minimize the chances of fixation failure with varus collapse and screw penetration.13 Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Proximal Humeral Fractures Settle
Our results showed that tuberosities usually remained stable during fracture healing; however, in some cases, an increase of displacement was observed. We failed to identify factors associated with significant tuberosity displacement, but we found that this occurred only in 17% of the cases. Our study has several weaknesses. A wide range of ages and different mechanisms of injuries were included, which could be a source of bias; however, we believe that including all patients best represents the full spectrum of injury. Nonoperative treatment was recommended for all patients included in this study at the discretion of the on-call orthopaedic team; some of these patients may had benefited from surgery, but at the time of completion of this study, most surgeons in our institution were more inclined to recommend nonoperative treatment for proximal humerus fractures, because available clinical evidence still failed in supporting surgical treatment for these injuries.14 Not all patients had a CT scan at last follow-up. A reproducibility analysis of measurements has not been performed. In addition, the sample size of the study is not large enough to analyze progression of displacement in every fracture pattern, and our findings may not be applicable to the entire spectrum of injuries. Finally, post hoc sample size calculation showed that our sample was not big enough to demonstrate statistically significant differences in all measurements performed; however, P values remained significant in the most clinically relevant ones. Our study also presents several strengths: its prospective nature; the fact that all the patients underwent the same treatment protocol and were evaluated by a single investigator using the same standardized and corrected for magnification imaging studies; and comprehensive image analysis was carried out using specific computer-assisted measurements and dedicated software. To the best of our knowledge, this is the first study analyzing and demonstrating humeral morphologic changes during conservative treatment of proximal humerus fractures. Proximal humerus fractures treated with standard conservative treatment protocols settle at the head–shaft junction during the consolidation process. These morphologic changes could be a source of interindividual outcome variability, and potentially be harmful in a subset of patients with older age and greater initial displacement. Substantial additional displacement of the tuberosities was seldom observed.
ACKNOWLEDGMENTS The authors wish to acknowledge Professor L. Munuera Martínez for his contribution to this scientific project; and to Dr. JJ Domínguez Reboiras for his expert support during the clinical phase of this study. REFERENCES 1. Murray IR, Amin AK, White TO, et al. Proximal humeral fractures: current concepts in classification, treatment and outcomes. J Bone Joint Surg Br. 2011;93:1–11. 2. Handoll HH, Ollivere BJ, Rollins KE. Interventions for treating proximal humeral fractures in adults. Cochrane Database Syst Rev. 2010;12:CD000434. 3. Hodgson SA, Mawson SJ, Stanley D. Rehabilitation after two-part fractures of the neck of the humerus. J Bone Joint Surg Br. 2003;85:419–422. 4. Lefevre-Colau MM, Babinet A, Fayad F, et al. Immediate mobilization compared with conventional immobilization for the impacted nonoper-
www.jorthotrauma.com |
e29
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Foruria et al
5. 6.
7. 8. 9.
atively treated proximal humeral fracture. A randomized controlled trial. J Bone Joint Surg Am. 2007;89:2582–2590. Foruria AM, de Gracia MM, Larson DR, et al. The pattern of the fracture and displacement of the fragments predict the outcome in proximal humeral fractures. J Bone Joint Surg Br. 2011;93:378–386. Martinez de la Iglesia J, Duenas Herrero R, Onis Vilches MC, et al. Spanish language adaptation and validation of the Pfeiffer’s questionnaire (SPMSQ) to detect cognitive deterioration in people over 65 years of age [in Spanish]. Med Clin (Barc). 2001;117:129–134. Neer CS II. Shoulder rehabilitation. In: Reines L, ed. Shoulder Reconstruction. Philadelphia, PA: W.B. Saunders Company; 1990:487–533. Neer CS II. Displaced proximal humeral fractures. I. Classification and evaluation. J Bone Joint Surg Am. 1970;52:1077–1089. Marsh JL, Slongo TF, Agel J, et al. Fracture and dislocation Classification compendium - 2007: Orthopaedic Trauma Association Classification,
e30
| www.jorthotrauma.com
10. 11. 12. 13. 14.
Database and Outcomes Committee. J Orthop Trauma. 2007;21(suppl 10): S1–S163. Snedecor GW, Cochran WG. Statistical Methods. Ames, IA, Iowa: Iowa State University; 1980. Neer CS II. Four-segment classification of proximal humeral fractures: purpose and reliable use. J Shoulder Elbow Surg. 2002;11: 389–400. Court-Brown CM, Garg A, McQueen MM. The epidemiology of proximal humeral fractures. Acta Orthop Scand. 2001;72:365–371. Barlow JD, Sanchez-Sotelo J, Torchia M. Proximal humerus fractures in the elderly can be reliably fixed with a “hybrid” locked-plating technique. Clin Orthop Relat Res. 2011;469:3281–3291. Foruria AM, Fox TJ, Sperling JW, et al. Clinical meaning of unexpected positive cultures (UPC) in revision shoulder arthroplasty. J Shoulder Elbow Surg. 2013;22:620–627.
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Proximal Humeral Fractures Treated Conservatively Settle During Fracture Healing Antonio M. Foruria, MD, PhD,* Milagros Martí, MD, PhD,† and Joaquin Sanchez-Sotelo, MD, PhD‡
Objectives: Determine the relative position of the main fractured fragments of proximal humerus fractures treated conservatively to compare displacement at 2 time points: injury (0), and 1 year later (1).
Design: Prospective, comparative cohort study. Setting: Level I trauma center. Patients: Eighty-nine prospectively enrolled adults. Intervention: Six weeks of sling immobilization and a homebased program rehabilitation protocol started 2 weeks after injury.
Main Outcome Measurements: Standardized radiographs of injured shoulders were obtained in all patients at times 0 and 1. Computed tomography scans were also obtained at these times in 73 cases. Forty-two computer-assisted measurements of displacement were performed at times 0 and 1 and then compared. Factors related to progression of displacement were analyzed. Results: Ninety percent of fractures were classified into 1 of 4 patterns: posteromedial (varus) impaction (46), lateral (valgus) impaction (13), isolated greater tuberosity (15), and anteromedial impaction (6). Head– shaft displacement increased over time. In posteromedial impaction fractures, average fracture settling included 9 degrees in varus, 7 degrees in retroversion, and 3.2 mm in posterior shortening. In valgus-impacted fractures, a decrease in valgus tilt and a tendency toward a more anterior orientation of the articular surface was observed. Greater tuberosity displacement increased more than 5 mm in less than 20% of cases. Age and initial displacement were related to progression of displacement. Conclusions: Proximal humerus fractures treated conservatively settle at the head–shaft junction during healing. Substantial additional displacement of tuberosities was seldom observed.
INTRODUCTION Currently, early mobilization is the mainstay of conservative treatment of proximal humeral fractures.1 Several articles have been published showing the benefits of early versus delayed mobilization,2 encouraging the orthopaedic community to narrow the time that patients should be completely immobilized before starting a range of motion exercises. Although such advocated early mobilization protocols often start even before soft callus formation occurs,3 especially in impacted surgical neck fractures,4 nonunion rates remained low. We demonstrated that fracture pattern and initial fragment displacement is a main factor determining the final outcome after 1 year of standard conservative treatment of proximal humerus fractures5; initial fragment displacement and proximal humeral relationships with surrounding structures, measured with dedicated software on image test, could explain most but not all interpatient outcome variability. Other factors may influence final outcome, so predictions based on initial fracture morphology may not be completely accurate. We hypothesized that fragments displacement and anatomic proximal humeral relationships may change during the fracture healing process in the setting of standard conservative treatment protocols including early mobilization. Our primary aim was to analyze morphologic changes that occur during the healing process of proximal humerus fractures treated with standard conservative protocols, focusing on identifying possible fragment displacement progression and changes in anatomic relationships of the proximal humerus. Our secondary aim was to identify factors associated with progression of displacement during the fracture consolidation process.
Key Words: proximal humeral fractures, fracture healing, conservative treatment
Level of Evidence: Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence. (J Orthop Trauma 2015;29:e24–e30) Accepted for publication June 26, 2014. From the *Department of Orthopedic Surgery, Fundación Jiménez Díaz, Madrid, Spain; †Department of Radiology, La Paz University Hospital, Madrid, Spain; and ‡Department of Orthopedic Surgery, The Mayo Clinic, Rochester MN. The authors report no conflict of interest. Reprints: Antonio M. Foruria, MD, PhD, Department of Orthopaedic Surgery, Shoulder and Elbow Surgery Unit, Fundación Jiménez Díaz, Avda Reyes Católicos, 2, Spain, Madrid 28040 (e-mail: [email protected]). Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
e24
| www.jorthotrauma.com
PATIENTS AND METHODS Patients The study proceeded after approval from the Institutional Review Board of La Paz University Hospital (identification code: PI-382) as part of a large prospective research project focusing on prognostic factors of conservatively treated proximal humerus fractures.5 All patients with a proximal humerus fracture evaluated at the Emergency Department of a single institution between March and September 2005 were considered for inclusion into our study. The treating orthopaedic surgeon on call recommended nonoperative or surgical treatment based on morphology of the fracture, age of the patient, functional demands, and comorbidities. J Orthop Trauma Volume 29, Number 2, February 2015
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Patients selected for conservative treatment were asked to participate in the study if they were aged 18 years or older, had a unilateral fracture, had no other associated injuries or previous fracture involving the shoulder, and had a preserved cognitive function as determined by the Pfeiffer test—6 patients excluded.6 During this period, 132 patients were evaluated, in whom surgical treatment was proposed and accepted in 10 patients. Of the remaining 122, 111 fulfilled the inclusion criteria and were included in the study. All gave informed consent. Within the first 12 months after their fracture, 7 died as a result of unrelated causes and 11 were lost to follow-up. The remaining 93 patients were followed up prospectively for 1 year after their fracture and form the basis of this study. There were 70 females (79%) and 19 males, with an average age of 70 6 13 years (range, 26–93 years). The average age was 73 years for females and 59 years for males. The mechanism of injury was a fall from a standing height in 75 shoulders (84%), high-energy injury (motor-vehicle accident of fall from a height) in 8 shoulders (9%), and a sportrelated injury in 6 shoulders (7%). Comorbidities included heart disease (15 cases, 17%), diabetes mellitus (10 cases, 11%), gastrointestinal disease (9 cases, 10%), and pulmonary disease (7 cases, 8%).
Treatment Protocol Treatment was identical in all patients. They were instructed and supervised by the same orthopaedic surgeon (A.M.F.). The affected upper extremity was protected in a commercially available shoulder immobilizer for the first 2 weeks after the fracture. A home program of passive and active assisted range of movement exercises was then initiated within the range of pain tolerance for 4 more weeks, protecting the arm in the sling between the exercise periods. The immobilizer was removed and stretching exercises were added at 6 weeks, and shoulder elevation and external and internal rotation elastic rubber band strengthening exercises7 were started at 12 weeks and continued for 3 months.
Radiologic Evaluation All radiographs were standardized according to an established protocol and carried out under direct supervision by the same physician (A.M.F.). The projections included an anteroposterior (AP) radiograph in the plane of the scapula with the arm in 20 degree of external rotation, a lateral “y view” with the forearm parallel to the image receptor, and an axillary view taken with the patient standing with the x-ray beam coming from the floor toward the axilla, and the image receptor over the patient’s shoulder, the arm being separated from the body passively as in a pendulum exercise.7 Radiographs were obtained for both shoulders at the time of the fracture, and for the affected shoulder at 3, 6, and 12 months. Additional AP and lateral radiographs were taken at 2 and 6 weeks to assess possible gross progression of displacement. A computed tomography (CT) scan, with 3D and 3-planar reconstruction referenced to the scapular plane, of the affected shoulders were also obtained in all patients immediately after the injury, and in 73 of the 89 shoulders (82%) at 12 months. Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Proximal Humeral Fractures Settle
All the digital radiographs and CT scans were corrected for magnification. Morphology of each individual fracture was analyzed based on these, including spatial rotation of 3D reconstructions. Commercially available software (AutoCAD 2004; Autodesk Inc, San Rafael, CA) was used to calculate the measurements listed in Table 1 for both the radiographs and CT reconstructions (Figs. 1A–F). The diaphyseal humeral axis was used to define inferior, superior, medial, and lateral on the AP and lateral radiographs and on the coronal and sagittal CT reconstructions; the measurements were referenced to this axis. The perpendicular to the glenoid surface was used to define anterior, posterior, medial, and lateral on the axillary radiographs and on the axial CT reconstructions, and the measurements were also referenced to this axis. For each individual fracture, the CT scan that showed that the maximum displacement was selected for each individually analyzed fragment. Measurements obtained from image test obtained at the moment of the injury were compared with those performed at 1 year of follow-up. The same researcher (M.M.) formatted all the image tests and another (A.M.F.) performed all the measurements. Fractures were classified according to our morphologic classification system described elsewhere5 that included 4 main fracture patterns: posteromedial (varus) impacted fractures, lateral (valgus) impacted fractures, isolated greater tuberosity fractures, and anteromedial impacted fractures; injuries were also classified according to Neer8 and OTA9 criteria.
Statistical Analysis Normal distribution of quantitative variables was analyzed with the Kolmogorov–Smirnov test. Differences in measurements performed at baseline and at 1 year of followup were compared with the paired 2-sampled Student t or the Wilcoxon signed-rank tests accordingly. No data was available before the study to calculate the sample size; post hoc sample size calculation was performed after data were available.10 Effect size was also calculated (general linear modeling, partial eta squared: h2p). Correlations between quantitative variables (age and progression of displacement, initial displacement and progression of displacement) were tested with the Pearson product-moment and the Spearman’s rank correlation coefficients.
RESULTS Radiographs and CT scans at 1 year showed fracture union in 87 cases (98%). Two patients developed a nonunion at the surgical neck level. There was radiographic evidence of avascular necrosis in 6 patients, being partial in 5 patients with minimal symptoms, and most of the humeral head in the remainder patient who complained of severe pain.
Fracture Classification There were 46 (52%) posteromedial (varus) neck impaction fractures; of these, 25 (54%) had a fractured greater tuberosity, and 3 (6%) had both tuberosities fractured. There were 13 (14%) lateral (valgus) impaction fractures, all with an associated greater tuberosity fracture and 5 (38%) with both www.jorthotrauma.com |
e25
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Foruria et al
TABLE 1. Measurements Performed Both in CT Reconstructions and X-Rays for Each Patient Measurements Performed on CT Reconstructions and X-Rays Measurements Axial CT reconstruction-axillary x-ray GT-AS overlapping angle and % LT-AS overlapping angle and % GT-glenoid rim distance LT-glenoid rim distance GT-LT tuberosity distance and angle GT medial displacement LT medial displacement Cephalic axis–glenoid axis angle Cephalic axis–diaphyseal axis angle Coronal CT reconstruction-AP x-ray Cephalodiaphyseal angle GT-AS Medial–lateral translation Medial impaction Lateral impaction GT acromion distance AS acromion distance Sagittal CT reconstruction lateral x-ray Cephalodiaphyseal angle Translation in AP plane Flexion—extension of surgical neck Anterior impaction Posterior impaction Cephalic axis–scapular axis angle Diaphyseal axis–scapular axis angle
Definition Arc of overlapping of the GT over the AS referenced to the center of the AS, and percentage of the AS covered by the GT Arc of overlapping of the LT over the AS referenced to the center of the AS, and percentage of the AS covered by the LT Lesser distance between the GT and the posterior glenoid rim Lesser distance between the LT and the anterior glenoid rim Lesser distance between the GT and the LT and arc of displacement between both tuberosities referenced to the center of the AS Medial displacement of the GT Medial displacement of the LT Angle between the axis of the humeral AS and the perpendicular to the glenoid AS Angle between the axis of the AS and the diaphyseal axis. Only in x-rays Angle between the diaphyseal axis and the axis of the AS Distance between the highest points of the AS and the GT perpendicularly projected on the diaphyseal axis Medial or lateral displacement of the head related to the diaphysis measured at the medial cortical level Magnitude of medial surgical neck cortex impaction/distraction Magnitude of lateral surgical neck cortex impaction/distraction Lesser distance between the GT and the acromion Lesser distance between the AS and the acromion Angle between the diaphyseal axis and the axis of the AS Anterior or posterior displacement of the head related to the diaphysis measured at the posterior cortical level Angle between head and diaphyseal longitudinal axis Magnitude of anterior surgical neck cortex impaction/distraction Magnitude of posterior surgical neck cortex impaction/distraction Angle between the axis of the AS and the longitudinal axis of the scapula. Only performed in x-rays Angle between the diaphyseal axis and the longitudinal axis of the scapula. Only performed in x-rays
The numbers correspond to those between brackets in Figure 1. The diaphyseal axis was used to define what is inferior or superior and medial or lateral in AP and lateral x-rays, and coronal and sagittal CT reconstructions, being measurements referenced to this axis. The perpendicular to the glenoid surface was used to define anterior, posterior, medial, and lateral on axillary x-rays and axial CT reconstructions, being measurements referenced to this axis. AS, articular surface; GT, greater tuberosity; LT, lesser tuberosity. Reproduced with permission from Foruria et al.5 Copyright @2011, the British Editorial Society of Bone and Joint Surgery.
tuberosities fractured. Fifteen (17%) patients had an isolated greater tuberosity fracture; 3 of these presented at the emergency department with an associated dislocation that was conservatively reduced by the on-call team. Six (7%) shoulders had an anteromedial surgical neck impaction fracture. The remaining 9 (10%) fractures did not fit into any of the 4 common patterns described above. After analyzing x-rays and CT scans, 36 (40.4%) fractures were considered to be displaced according to Neer8 criteria. Table 2 shows the injury distribution according to the OTA classification.9 When initial and final simple x-rays and CT scans were compared, we observed that fractures settled at the surgical neck level during healing, increasing their deformity as a result of the progression of their initial displacement. In posteromedial surgical neck impaction fractures (46 cases, characterized by varus and posterior impaction at the surgical neck of the humerus, and increased retroversion
of the humeral head), x-rays and CT scans showed an increase of the tridimensional deformity and a change in the relationships of the proximal humerus with the acromion (Table 3): 1. Increase in varus deformity: with a mean decrease in the cephalodiaphyseal angle on AP x-rays of 8.6 6 8 degrees (median, 7; range, 29 to 27; P , 0.001), and a mean increase in the medial impaction of the surgical neck on the same projection of 3.5 6 5.7 mm (median, 2.2; range, 27 to 24; P , 0.001). 2. Increase in humeral head retroversion: with a mean increase in the cephalic axis–glenoid axis angle on axial CT scans of 7 6 17 degrees (median, 10; range, 242 to 33; P = 0.014), a decrease in the distance between the lesser tuberosity and the anterior glenoid rim of 3.8 6 6 mm (median, 4.5; range, 29 to 14; P , 0.001), and a mean increase in the cephalodiaphyseal angle in the sagittal CT of 13 6 39 degrees (median, 7; range, 271 to 117; P = 0.049), toward a more posterior–inferior orientation of the cephalic axis.
e26
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Progression of Deformity
| www.jorthotrauma.com
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Proximal Humeral Fractures Settle
FIGURE 1. Examples of computed-assisted measurements performed after correction for magnification. Between brackets is written the number corresponding to measurements represented in Table 1. A, AP x-ray. B, Lateral x-ray. C, Axillary x-ray. D, Axial CT reconstruction. E, Coronal CT reconstruction. F, Sagittal CT reconstruction. Reproduced with permission from Foruria et al.5 Copyright @2011, the British Editorial Society of Bone and Joint Surgery.
3. Increase in the mean posterior surgical neck impaction on the sagittal CT of 3.2 6 6 mm (median, 2.1; range, 29 to 19; P = 0.003). 4. A decrease in the distance between the acromion and the humerus on the AP x-rays: with a mean decrease of TABLE 2. Distribution of Fractures According to the OTA Classification System Fracture Type
N
%
A11 A12 A13 A21 A22 A23 A32 A33 B11 B13 B21 C21 C23 Nonclassifiable Total
9 1 4 6 19 4 2 1 28 2 1 3 1 8 89
10.1 1.1 4.5 6.7 21.3 4.5 2.2 1.1 31.5 2.2 1.1 3.4 1.1 9.0 100.0
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
the distance between the humeral articular surface and the acromion of 3.9 6 5 mm (median, 2.8; range, 25 to 21; P , 0.001), and a mean decrease of the distance between the greater tuberosity and the acromion of 4.5 6 5 mm (median, 4.6; range, 24 to 15; P , 0.001). Lateral (valgus) impaction fractures (13 cases) showed a mean reduction of valgus tilt of the articular surface– cephalodiaphyseal angle of 3.3 6 6 degrees (median, 4; range, 215 to 12; P = 0.023) in AP x-rays. On lateral x-rays, articular surfaces showed a tendency to a more anterior orientation cephalodiaphyseal angle after consolidation (mean, 5.2 6 8 degrees; median range, 25 to 24; P = 0.061). There were no significant differences on AP or mediolateral translations of the articular surface with respect to the diaphysis, or on impaction of the diaphysis inside the head segment. Fractured greater tuberosities (63 cases) remained stable in the majority of the cases; mean differences in displacement between initial and 1-year image tests, measured as the medial displacement of the greater tuberosity, the distance from the greater tuberosity to the glenoid rim or the distance between tuberosities, were lesser than 2 mm for the 3 measurements and not statistically significant (P = 0.34, P = 0.9, and P = 0.15, respectively). Six patients (9%) showed an increase of medial displacement of the greater tuberosity greater than 5 mm (range, 5.5–16.2 mm). Eleven cases (17%) showed www.jorthotrauma.com |
e27
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Foruria et al
TABLE 3. Comparison of Measurements Performed on Image Tests of Patients With Posteromedial Impaction Fractures (n = 46) at 2 Time Points: When the Fracture Occurred (0) and at 1 Year of Follow-up (1) Measurement AP x-rays Cephalodiaphyseal angle, degrees GT-AS distance, mm AS-AC distance, mm Medial impaction, mm Lateral x-rays Cephalodiaphyseal angle, degrees GT-AC distance, mm AS-AC distance, mm Posterior impaction, mm Axial CT LT-glenoid distance, mm Cephalic axis–glenoid axis angle, degrees Coronal CT GT-AC distance, mm AS-AC distance, mm Sagittal CT Cephalodiaphyseal angle, degrees Posterior impaction, mm
Mean 6 SD
Means Difference
P
h2p
(0) 117 6 19, (1) 109 6 18 (0) 0.1 6 5, (1) 21 6 5 (0) 18 6 6, (1) 14 6 5 (0) 12.6 6 8, (1) 16.1 6 10
8.6 1 24 3.5
5.9 0.01 22.5 1.7
to to to to
11.2 2.2 24.5 5.2
,0.001 0.048 ,0.001 ,0.001
0.5 0.1 0.4 0.3
(0) 74 6 25, (1) 85.4 6 26 (0) 14.3 6 5, (1) 9.7 6 3 (0) 15.9 6 5, (1) 12.2 6 5 (0) 13.2 6 11, (1) 16.4 6 11
11.4 24.5 23.7 3.2
5.7 22.7 21.9 0
to to to to
17 26.3 25.5 6.5
,0.001 ,0.001 ,0.001 0.05
0.3 0.4 0.3 0.1
(0) 19.3 6 5, (1) 15.5 6 5 (0) 31.5 6 20, (1) 38.6 6 18
23.8 7
21.9 to 25.7 1.5 to 12.6
,0.001 0.014
0.3 0.2
(0) 8.6 6 4, (1) 6.1 6 4 (0) 9.3 6 4, (1) 8.1 6 4
22.4 21.2
20.7 to 24.1 0 to 22.4
0.007 0.05
0.2 0.1
(0) 54.7 6 38, (1) 67.8 6 33 (0) 11.6 6 9, (1) 14.8 6 1
13 3.1
0.6 to 26 1.2 to 5.1
0.049 0.003
0.1 0.2
95% CI
AC, acromion; AS, articular surface; CI, confidence interval; GT, greater tuberosity; h2p, partial eta squared representing effect size.
a decrease in the distance between the greater tuberosity and the posterior glenoid rim greater than 5 mm (range, 5.6–30.5). Eight (13%) patients showed a more superior position of the greater tuberosity in relation to the articular surface greater than 5 mm (range, 5.1–25), as a consequence of increased varus tilt of the humeral head, leaving the greater tuberosity more superiorly located as opposed to a superior migration of the isolated tuberosity fragment. The majority of fractured lesser tuberosities (9 cases) remained also stable, without significant progression of displacement. Only 1 case showed progression of medial displacement of the lesser tuberosities of 5 mm.
1 year of follow-up. On the contrary, for superposition of the greater tuberosity over the articular surface (posterior displacement), the lesser the initial displacement, the greater the progression of displacement, but this last result could be influenced by the increment of head retroversion described above (Table 4).
DISCUSSION
Age was associated with progression of surgical neck displacement in posteromedial impaction fractures: the older the patient, the greater the increase of medial impaction (P = 0.04; correlation coefficient, 0.31), the greater the increase of posterior impaction (P , 0.001; correlation coefficient, 0.57), and the greater the increase of retroversion (the lesser tuberosity–glenoid distance decreased more with age, P = 0.023, correlation coefficient 0.37; and the angle formed by the cephalic axis and the perpendicular to the glenoid surface on axial CT showed a tendency toward greater increase with age, P = 0.059; correlation coefficient, 0.31). Gender did not show correlation with progression of displacement in the context of a lack of statistical power. On posteromedial impaction fractures, initial displacement at the surgical neck level correlated with progression of deformity; for a given measurement, the greater the initial displacement, the greater the further displacement observed at
We verified our main hypothesis: fracture deformity increases during standard conservative treatment of proximal humerus fractures. Classic conservative treatment protocols underline the necessity of physical examination before starting the exercise program to test the stability of the fracture and avoid possible fragment migration and eventually the development of a nonunion.11 However, fracture instability may not be observed during the examination unless gross surgical neck instability is present. The recent description of proximal humerus fractures as impacted in the majority of the cases,5,12 and the existence of articles supporting even immediate mobilization of impacted injuries,4 could make the orthopaedic community think that there is no risk of further displacement when conservative treatment is implemented. Surgical neck impaction could be responsible for inability to identify fragment motion during physical examination; but the increase of deformity might be the consequence of a single event (like a fall) or due to the continuous cyclic stress forces crossing the fracture during mobilization protocols, in conjunction with the pull of periarticular muscles on fracture fragments. This cyclic stress could wear off the resistance of humeral head trabecular bone, allowing the diaphysis to penetrate progressively
e28
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Factors Related to Progression of Deformity
| www.jorthotrauma.com
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
TABLE 4. For a Given Measurement, Correlations Between Initial Value and the Difference Obtained From Subtracting the Final Value From the Initial One in Posteromedial Impaction Fractures (n = 46) Measurement AP x-rays Cephalodiaphyseal angle, degrees GT-AS distance, mm AS-AC distance, mm Lateral x-rays GT-AC distance, mm AS-AC distance, mm Axial CT LT-glenoid distance, mm Cephalic axis–glenoid axis angle, degrees GT-AS superposition angle, degrees GT-glenoid distance Coronal CT GT-AC distance, mm AS-AC distance, mm Sagittal CT Cephalodiaphyseal angle, degrees
Pearson Correlation Coefficient
P
0.31
0.034
0.35 0.62
0.017 ,0.001
0.82 0.58
,0.001 ,0.001
0.56 0.52
,0.001 0.001
20.38
0.019
0.42
0.009
0.54 0.37
,0.001 0.023
0.62
,0.001
AC, acromion; AS, articular surface; GT, greater tuberosity; LT, lesser tuberosity.
deeper on it in the same direction the initial deformity was produced, as it is seen on our results. Early mobilization protocols continue to be the mainstay of conservative treatment. However, we know from our previous study5 that the greater the initial deformity, the greater the pain and the poorer the outcome after conservative treatment of proximal humerus fractures. In this context, we believe that the increase in deformity occurring during the healing process could play a relevant role on final outcome in some patients. For this reason, we think it is important to identify the subset of patients with higher risk of important progression of displacement and isolate them from the majority of patients in which such a risk would not be clinically relevant. We identified elderly patients and those with greater initial displacement as the ones in which the progression of the deformity could be worse, both factors probably related to poor bone quality. Age has been also described as a risk factor for sustaining a more displaced fracture12; moreover, age is an independent factor on treatment decision making, and the older the patient, surgery is offered less likely for treatment.1 All these factors in combination could increase the incidence of malunion resulting in a poorer outcome in older patients. This mentioned “at risk” group could hypothetically benefit of either surgery if they are physically fit or a delayed treatment protocol (eg, until soft callus formation at 3–4 weeks after injury), as other authors have described after proximal humerus open reduction and internal fixation to minimize the chances of fixation failure with varus collapse and screw penetration.13 Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Proximal Humeral Fractures Settle
Our results showed that tuberosities usually remained stable during fracture healing; however, in some cases, an increase of displacement was observed. We failed to identify factors associated with significant tuberosity displacement, but we found that this occurred only in 17% of the cases. Our study has several weaknesses. A wide range of ages and different mechanisms of injuries were included, which could be a source of bias; however, we believe that including all patients best represents the full spectrum of injury. Nonoperative treatment was recommended for all patients included in this study at the discretion of the on-call orthopaedic team; some of these patients may had benefited from surgery, but at the time of completion of this study, most surgeons in our institution were more inclined to recommend nonoperative treatment for proximal humerus fractures, because available clinical evidence still failed in supporting surgical treatment for these injuries.14 Not all patients had a CT scan at last follow-up. A reproducibility analysis of measurements has not been performed. In addition, the sample size of the study is not large enough to analyze progression of displacement in every fracture pattern, and our findings may not be applicable to the entire spectrum of injuries. Finally, post hoc sample size calculation showed that our sample was not big enough to demonstrate statistically significant differences in all measurements performed; however, P values remained significant in the most clinically relevant ones. Our study also presents several strengths: its prospective nature; the fact that all the patients underwent the same treatment protocol and were evaluated by a single investigator using the same standardized and corrected for magnification imaging studies; and comprehensive image analysis was carried out using specific computer-assisted measurements and dedicated software. To the best of our knowledge, this is the first study analyzing and demonstrating humeral morphologic changes during conservative treatment of proximal humerus fractures. Proximal humerus fractures treated with standard conservative treatment protocols settle at the head–shaft junction during the consolidation process. These morphologic changes could be a source of interindividual outcome variability, and potentially be harmful in a subset of patients with older age and greater initial displacement. Substantial additional displacement of the tuberosities was seldom observed.
ACKNOWLEDGMENTS The authors wish to acknowledge Professor L. Munuera Martínez for his contribution to this scientific project; and to Dr. JJ Domínguez Reboiras for his expert support during the clinical phase of this study. REFERENCES 1. Murray IR, Amin AK, White TO, et al. Proximal humeral fractures: current concepts in classification, treatment and outcomes. J Bone Joint Surg Br. 2011;93:1–11. 2. Handoll HH, Ollivere BJ, Rollins KE. Interventions for treating proximal humeral fractures in adults. Cochrane Database Syst Rev. 2010;12:CD000434. 3. Hodgson SA, Mawson SJ, Stanley D. Rehabilitation after two-part fractures of the neck of the humerus. J Bone Joint Surg Br. 2003;85:419–422. 4. Lefevre-Colau MM, Babinet A, Fayad F, et al. Immediate mobilization compared with conventional immobilization for the impacted nonoper-
www.jorthotrauma.com |
e29
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 2, February 2015
Foruria et al
5. 6.
7. 8. 9.
atively treated proximal humeral fracture. A randomized controlled trial. J Bone Joint Surg Am. 2007;89:2582–2590. Foruria AM, de Gracia MM, Larson DR, et al. The pattern of the fracture and displacement of the fragments predict the outcome in proximal humeral fractures. J Bone Joint Surg Br. 2011;93:378–386. Martinez de la Iglesia J, Duenas Herrero R, Onis Vilches MC, et al. Spanish language adaptation and validation of the Pfeiffer’s questionnaire (SPMSQ) to detect cognitive deterioration in people over 65 years of age [in Spanish]. Med Clin (Barc). 2001;117:129–134. Neer CS II. Shoulder rehabilitation. In: Reines L, ed. Shoulder Reconstruction. Philadelphia, PA: W.B. Saunders Company; 1990:487–533. Neer CS II. Displaced proximal humeral fractures. I. Classification and evaluation. J Bone Joint Surg Am. 1970;52:1077–1089. Marsh JL, Slongo TF, Agel J, et al. Fracture and dislocation Classification compendium - 2007: Orthopaedic Trauma Association Classification,
e30
| www.jorthotrauma.com
10. 11. 12. 13. 14.
Database and Outcomes Committee. J Orthop Trauma. 2007;21(suppl 10): S1–S163. Snedecor GW, Cochran WG. Statistical Methods. Ames, IA, Iowa: Iowa State University; 1980. Neer CS II. Four-segment classification of proximal humeral fractures: purpose and reliable use. J Shoulder Elbow Surg. 2002;11: 389–400. Court-Brown CM, Garg A, McQueen MM. The epidemiology of proximal humeral fractures. Acta Orthop Scand. 2001;72:365–371. Barlow JD, Sanchez-Sotelo J, Torchia M. Proximal humerus fractures in the elderly can be reliably fixed with a “hybrid” locked-plating technique. Clin Orthop Relat Res. 2011;469:3281–3291. Foruria AM, Fox TJ, Sperling JW, et al. Clinical meaning of unexpected positive cultures (UPC) in revision shoulder arthroplasty. J Shoulder Elbow Surg. 2013;22:620–627.
Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2014 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
