ORIGINAL ARTICLE
Predisposing Effect of Elbow Alignment on the Elbow Fracture Type in Children Seungcheol Kang, MD and Soo-Sung Park, MD, PhD
Objectives: Under the hypothesis that the elbow alignment,
Key Words: elbow fracture, carrying angle, pediatrics, supracondylar humeral fracture, lateral condylar fracture, radial neck fracture, trauma
namely the carrying angle, could predispose individuals to a specific type of pediatric elbow fracture after a fall onto an outstretched arm, we investigated the relationship between radiographic carrying angle and elbow fracture type in children.
Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
Design: Retrospective case–control study.
(J Orthop Trauma 2015;29:e253–e258)
Setting: Level I pediatric trauma center. Patients/Participants: We reviewed 374 children who were diagnosed with supracondylar fracture (SCF, n = 208), lateral condylar fracture (LCF, n = 132), and radial neck fracture (RNF, n = 34). Intervention: The association between the radiographic carrying angle and the fracture type was investigated. Main Outcome Measurements: To adjust for bias, 2 statistical methods were used: multivariate analysis using a baseline-category logistic model and a case-matching method using propensity score analysis. Results: In the multivariate analysis, with SCF patients set as the baseline category, a more valgus-deviated elbow (increased carrying angle, P = 0.011) predisposed individuals to RNF, whereas a more varus-deviated elbow (decreased carrying angle, P , 0.001) predisposed them to LCF. In the case-matched analysis, there were also significant differences in carrying angles between RNF and casematched SCF patients (14.3 vs. 11.4 degrees, P = 0.013) and between LCF and case-matched SCF patients (7.7 vs. 11.7 degrees, P , 0.001).
Conclusions: Elbow alignment, which may influence the transmission of traumatic force during a fall onto an outstretched elbow, could be a predisposing factor for specific types of pediatric elbow fracture. The results provide the additional information about the injury mechanisms of pediatric elbow fracture and may deepen our understanding of the fractures.
Accepted for publication February 23, 2015. From the Department of Orthopaedic Surgery, Asan Medical Center Children’s Hospital, University of Ulsan College of Medicine, Seoul, South Korea. Presented in part at the 58th Annual Fall Congress of the Korean Orthopaedic Association, October 18, 2014, Seoul, South Korea. The authors report no conflict of interest. Reprints: Soo-Sung Park, MD, PhD, Department of Orthopedic Surgery, Asan Medical Center Children’s Hospital, University of Ulsan College of Medicine, 88, Olympic-ro, 43-gil, Songpa-gu, Seoul 138-736, South Korea (e-mail: [email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Orthop Trauma Volume 29, Number 8, August 2015
INTRODUCTION Elbow fracture is one of the most common traumatic problems in children, comprising 7%–20% of all pediatric fractures.1,2 Of the various types of elbow fractures, the supracondylar fracture (SCF) is generally the most common, followed by the lateral condylar fracture (LCF) and radial neck fracture (RNF).3 The typical cause of these elbow fractures in children is a fall onto an outstretched elbow4,5: when a child falls onto an outstretched arm, the injury mechanism of hyperextension with vertical stress associates with SCF, hyperextension with valgus stress associates with RNF, and hyperextension with varus stress associates with LCF.4 The degree of angulation or alignment generally affects the mechanism of load transfer.6 Thus, the degree of elbow angulation might theoretically influence the transmission of traumatic force and associate with a specific elbow fracture type. However, few studies have examined the association between elbow alignment and fracture type. In a previous experimental study, possible associations of elbow alignment with SCF and LCF were suggested,7 but that study used only 2 cadavers for each fracture. Therefore, we hypothesized that the elbow alignment in the coronal plane, namely the carrying angle, might favor a specific type of elbow fracture in children after a fall onto an outstretched arm. More specifically, a more valgus-deviated elbow might predispose individuals to RNF, whereas a more varus-deviated elbow might predispose individuals to LCF. However, there are many potential bias problems when assessing the association between carrying angle and fracture type. Some different characteristics according to the fracture type have already been reported, such as patient age and sex and the injured side.3,8,9 Patient’s height and weight or their combination would affect the injury mechanism.10,11 Moreover, the carrying angle itself differs according to age and sex.12,13 Thus, adequate adjustment of confounding factors needs to be achieved to evaluate the correct relationship. Pursuant to these concerns, under the hypothesis that elbow alignment could affect the type of pediatric elbow fracture, we measured and compared the radiographic carrying angle of children who were diagnosed with isolated SCF, LCF, or RNF after a fall www.jorthotrauma.com |
e253
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Kang and Park
onto an outstretched arm, and finally evaluated the effect of radiographic carrying angle on the fracture type after bias adjustments. To adjust for bias, we used 2 statistical methods: multivariate analysis using a baseline-category logistic model and a case-matching method using propensity score analysis.
PATIENTS AND METHODS Patients Selection Patients younger than 16 years who were diagnosed with SCF, LCF, or RNF (13-A, 13-B, or 21-A, respectively, by AO/OTA classification14) in our institution from January 2005 to December 2013 were retrospectively reviewed. Of the 240 SCF patients, those with combined fracture (n = 17) or without an outstretched injury mechanism (direct injury or undocumented injury mechanism, n = 15) were excluded. Of the 142 LCF patients, those with combined fracture (n = 2) or without an outstretched injury mechanism (n = 8) were excluded. Of the 46 RNF patients, those with combined fracture (n = 12) were excluded; all remaining RNF patients had an outstretched injury mechanism. Finally, these exclusion criteria left 374 patients for the review (208 patients with SCF, 132 with LCF, and 34 with RNF). The institutional review board of our institute approved this study.
J Orthop Trauma Volume 29, Number 8, August 2015
S.K.) performed assessments in 30 randomly selected cases. The interobserver reliabilities of the measurements were evaluated using intraclass correlation coefficients. The intraclass correlation coefficient of interobserver reliabilities for the measurements was greater than 0.9, representing satisfactory agreement. Thus, measurements taken by a single investigator (S.K.) were used in the analyses. For clarity, we refer to the “carrying angle” as the “radiographic carrying angle” throughout the Patients and Methods, and Results section.
Statistical Methods For the purpose of comparing the patients according to the fracture type, the significance of differences between frequencies was calculated using the x2 test. Comparisons of mean values among the 3 fracture groups were performed using analysis of variance. When significant differences were detected in the analyses, intergroup comparisons were performed using Scheffe post hoc tests. To evaluate the independent effect of carrying angle on the fracture type, we used 2 statistical methods to eliminate bias: multivariate analysis using
Investigated Characteristics We investigated the radiographic carrying angle; age, height, weight, and body mass index (BMI) at injury; sex; and injured side (right or left). Simple radiographs of the uninjured elbow taken at the first visit to our institution were used to assess the radiographic carrying angle. In our institution, the radiographs of the contralateral uninjured elbow were routinely checked for the children with elbow fracture at their initial visit and regarded as a reference for treatment of injured elbow. All of the first visits were within 5 days of the injury. There was no significant left–right difference of carrying angle [10.3 6 5.3 degrees for right side (n = 152, measured from uninjured left side) vs. 10.7 6 4.8 degrees for left side (n = 222, measured from uninjured right side), P = 0.457]. The elbow anteroposterior radiographs were obtained with a following standardized technique: a child seats at the end of a table; the arm is fully extended with shoulder and elbow on same horizontal plane, and the forearm is supinated; the x-ray beam is angled perpendicularly to the table centering the elbow joint. Before passing the image to our image view system, the digitalized x-ray is prechecked by a disciplined radiographer. Then, the following items are checked: the rotation of epicondyle, the center of x-ray beam, and the overlap between the proximal ulna and the radial head, radial neck, and radial tuberosity.15 If there is any insufficiency or blurring, the x-ray is repeated. The radiographic carrying angle was measured as the angle between the following 2 lines (Fig. 1): a line connecting the midpoint of the metaphysis and the diaphysis of the distal humerus and a line connecting the midpoint of the ulna at the level of the bicipital tuberosity of the radius and the midpoint of the proximal end of the ulna.16–18 To test the interobserver reliability for the assessment of the carrying angle, 2 observers (S.-S.P. and
FIGURE 1. The radiographic carrying angle was measured as the angle between the following 2 lines: a line connecting the midpoint of the metaphysis and the diaphysis of the distal humerus and a line connecting the midpoint of the ulna at the level of the bicipital tuberosity of the radius and the midpoint of the proximal end of the ulna.
e254
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
| www.jorthotrauma.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
Predisposing Effect of Elbow Alignment
a baseline-category logistic model and case-matching analysis using propensity score analysis.19 For the multivariate analysis, using the baseline-category logistic model, the SCF patients were set as a baseline or reference category. Thus, odds ratios of .1 in each category meant that the variable or its increasing value favors the categorized fracture compared with SCF. Carrying angle, age, sex, injured side, and BMI were included in the analysis, but weight and height were excluded because they interfered with BMI. Statistical analyses were conducted with SPSS for Windows statistical package (version 21; IBM Co, Armonk, NY) and P , 0.05 were considered significant. For the selection of a case-matched control group, we used propensity score analysis. Among the 208 SCF patients, 34 and 132 patients were selected as control groups (1:1 ratio) for RNF and LCF patients, respectively. The propensity score was calculated for each patient based on logistic regression analysis of the probability of experiencing RNF and LCF using patient characteristics of age (in years), sex (male or female), height (in centimeters), weight (in kilograms), and side of injured arm (right or left). These characteristics were selected based on a discussion between 2 experienced and board-certified orthopaedic surgeons (S.K. and S.-S.P.) after a literature review.3,8–13 Using the propensity scores, we created a 1:1 matching with the nearest neighbor approach. Propensity score analysis was performed with R statistical software (version 3.1.0; http://cran.r-project.org/). Then, the SPSS software was also used to compare the variables between RNF or LCF patients and their respective case-matched SCF patients. The significance of differences between mean values, including age, height, weight, BMI, and carrying angle, was calculated using the independent t test. The significance of differences between frequencies, including sex and injured side, was calculated using the x2 test; Fisher exact test was used in the case that the expected number of frequency was less than 5.
RESULTS Comparisons Among the Three Fracture Groups Comparisons among the fracture groups are shown in Table 1. The carrying angle was different in each group: the mean carrying angles were 14.3 6 4.3 degrees for the RNF
group, 11.8 6 4.5 degrees for the SCF group, and 7.7 6 4.4 degrees for the LCF group (P , 0.001). Subgroup analyses using Scheffe post hoc tests showed significant difference between each group (P = 0.010 for RNF and SCF groups, P , 0.001 for SCF and LCF groups, and P , 0.001 for RNF and LCF groups). Besides the carrying angle, there were significant differences in age, height, weight, and BMI.
Results of Multivariate Analysis Using a Baseline-Category Logistic Model The results of multivariate analysis for each type of fracture are shown in Table 2. A larger carrying angle (P = 0.011), older age (P = 0.025), and larger BMI (P = 0.020) favored RNF occurrence compared with SCF, and a smaller carrying angle (P , 0.001), younger age (P = 0.023), and larger BMI (P , 0.001) favored LCF occurrence compared with SCF.
Results of Case-Matched Comparisons Using Propensity Score Analysis The results of the case-matched comparisons are shown in Table 3. There were no significant differences in matched variables reflecting well-balancing of the confounding factors, such as age, sex, height, weight, and injured side. The carrying angle was significantly larger in RNF patients than in their case-matched SCF patients and was significantly smaller in LCF patients than in their case-matched SCF patients.
DISCUSSION Traumatic forces in outstretched arm injuries are the main determinants of the type of pediatric elbow fracture. Elbow alignment can theoretically influence the transmission of traumatic force, but its association with fracture type has scarcely been studied. Therefore, we hypothesized that elbow alignment could also result in a specific fracture type by affecting the transmission of traumatic force in an outstretched arm. The results of the present study revealed the significant relationships between carrying angle and the specific fracture type. This is the first study about the predisposing effect of carrying angle on the specific type of
TABLE 1. Characteristics of the Children in This Study Diagnosed With RNF, SCF, and LCF Variable Age (years) Sex, n (%) Male Female Height (cm) Weight (kg) Injured side, n (%) Right Left BMI (kg/m2) Carrying angle (degrees)
Total (n = 374)
RNF (n = 34)
Patients With SCF (n = 208)
Patients With LCF (n = 132)
P
6.0 6 2.7
7.8 6 2.7
6.1 6 2.7
5.5 6 2.6
,0.001* 0.100
248 (66.3) 126 (33.7) 116.1 6 17.4 23.2 6 9.1
19 (55.9) 15 (44.1) 124.2 6 16.2 27.6 6 9.7
133 (63.9) 75 (36.1) 116.4 6 17.3 22.6 6 8.4
96 (72.7) 36 (27.3) 113.6 6 17.5 23.1 6 9.7
152 (40.6) 222 (59.4) 16.7 6 2.6 10.6 6 5.0
13 (38.2) 21 (61.8) 17.4 6 2.6 14.3 6 4.3
78 (37.5) 130 (62.5) 16.2 6 2.2 11.8 6 4.5
61 (46.2) 71 (53.8) 17.3 6 3.0 7.7 6 4.4
0.007* 0.010* 0.268
,0.001* ,0.001*
*P , 0.05.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jorthotrauma.com |
e255
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
Kang and Park
TABLE 2. Results of Multivariate Analysis Using a Baseline-Category Logistic Model RNF (vs. SCF) Carrying angle (degrees) Age (years) Sex Male Female Affected side Right Left BMI (kg/m2)
LCF (vs. SCF)
OR (95% CI)
P
OR (95% CI)
P
1.122 (1.027–1.225) 1.178 (1.021–1.359)
0.011* 0.025* 0.109
0.820 (0.772–0.870) 0.885 (0.797–0.983)
,0.001* 0.023* 0.345
Variable
0.519 (0.233–1.156) 1 (reference)
1.299 (0.755–2.234) 1 (reference) 0.718
1.159 (0.521–2.578) 1 (reference) 1.207 (1.031–1.414)
0.217 1.369 (0.832–2.254) 1 (reference) 1.225 (1.098–1.366)
0.020*
,0.001*
*P , 0.05. OR, odds ratio; CI, confidence interval.
pediatric elbow fracture, with bias adjustment using both multivariate analysis and case matching. There are a few considerations when interpreting the results of this study. First, we included patients with an outstretched arm injury but did not investigate the detailed traumatic force applied during the trauma, whether varus, valgus, or only vertical stress. In our anecdotal clinical experience, many children do not remember the exact injury mechanism, and some children even regard their injury situations irrelevant to the real traumatic force when compared with injury witnesses. For this reason, we were unable to investigate the detailed traumatic force and its correlation with elbow alignment. Second, the retrospective nature and relatively small number of RNF patients may cause bias problems that limit the interpretation of our results. To overcome these limitations, we used 2 statistical methods, multivariate analysis and case matching. Although these methods can provide excellent ways of minimizing bias, no statistical process is able to completely eliminate the inherent differences. Third, although there are many possible methods used to determine the carrying angle,18 the carrying angle is
traditionally measured from the shoulder to the wrist. However, we measured the radiographic carrying angle to establish the elbow alignment because the transmission of the traumatic force through the elbow is directly affected by angulation around the elbow, not the whole arm. Furthermore, bony alignment seemed to be more objective and more appropriate for assessing the transmission of the force. Radiographic carrying angle is reported to be substantially reliable,20 and radiographic measurement is thought to be more precise than clinical measurement.21 Fourth, we did not investigate the effect of sagittal alignment. There are various possible methods for evaluating sagittal alignment, such as measuring the maximal degree of elbow extension, ligamentous laxity, and/or humeroulnar angle on lateral radiograph with elbows fully extended. However, we did not routinely evaluate the degree of elbow extension or the ligamentous laxity of the patients with fractures, and the lateral radiographs were taken with the elbow flexed at a right angle. Although the fracture types were different from our study, McLauchlan et al22 already reported no difference of elbow extension, namely sagittal alignment, between the children
TABLE 3. Results of Case-Matched Comparisons Variable Age (years) Sex, n (%) Male Female Height (cm) Weight (kg) Injured side, n (%) Right Left BMI (kg/m2) Carrying angle (degrees)
Patients With RNF (n = 34)
Control Group* (n = 34)
7.8 6 2.7
8.1 6 3.1
19 (55.9) 15 (44.1) 124.2 6 16.2 27.6 6 9.7
18 (52.9) 16 (47.1) 127.0 6 20.2 28.7 6 10.1
13 (38.2) 21 (61.8) 17.4 6 2.6 14.3 6 4.3
8 (26.5) 25 (73.5) 17.4 6 2.6 11.4 6 5.0
P 0.717 0.808
0.525 0.652 0.300
0.914 0.013‡
Patients With LCF (n = 132)
Control Group† (n = 132)
5.5 6 2.6
5.3 6 2.4
96 (72.7) 36 (27.3) 113.6 6 17.5 23.1 6 9.7
133 (63.9) 75 (36.1) 111.3 6 17.0 21.5 6 8.6
61 (46.2) 71 (53.8) 17.3 6 3.0 7.7 6 4.4
78 (37.5) 130 (62.5) 16.7 6 2.1 11.7 6 4.7
P 0.501 0.100
0.266 0.156 0.268
0.069 ,0.001‡
*SCF patients who were case matched to RNF patients. †SCF patients who were case matched to LCF patients. ‡P , 0.05.
e256
| www.jorthotrauma.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
with SCF and distal radius fracture. We also suppose the sagittal alignment would not differ among the elbow fractures, but further studies are needed. Fifth, radiographic measurement is affected by positioning of the elbow, and the positioning can be difficult especially in children.15 The mean difference of carrying angle between RNF and SCF patients is only 2.5 degrees. Therefore, the possibilities of measurement error should be considered during interpretation of the results. Finally, the correlation between the elbow alignment and the transmission of traumatic force could not be proven by the results of this study. The results of this study should be supported by a future well-designed biomechanical study. In the multivariate analysis with SCF patients set as a baseline category, a more valgus-deviated elbow (increased carrying angle, P = 0.011) and older age (P = 0.025) predisposed individuals to RNF, whereas a more varus-deviated elbow (decreased carrying angle, P , 0.001) and younger age (P = 0.023) predisposed individuals to LCF. A higher BMI was significantly associated with both RNF (P = 0.020) and LCF (P , 0.001) when compared with SCF. The significant difference in carrying angle was also confirmed by case-matching analysis using propensity score analysis; when other characteristics were distributed equally, the carrying angle was significantly different between RNF and SCF patients (14.3 6 4.3 vs. 11.4 6 5.0 degrees, P = 0.013) and between LCF and SCF patients (7.7 6 4.4 vs. 11.7 6 4.7 degrees, P , 0.001). We can suggest a possible explanation for the correlation between carrying angle and the specific fracture type. Carrying angle seems to alter the transmission of traumatic force: the axial loading during trauma might be converted into valgus force in a more valgus-deviated elbow and into varus force in a more varus-deviated elbow. Figure 2 shows the schematic transmission of traumatic force when someone falls
Predisposing Effect of Elbow Alignment
onto an outstretched arm. SCF is suggested to occur due to the impact of the olecranon on the distal humerus when a vertical stress is applied during a fall onto an outstretched arm (Fig. 2A).4,5 However, in an elbow with cubitus valgus, the traumatic force (axial loading) can be converted into a compression force of the radial head and neck (Fig. 2B),4,5 whereas in an elbow with cubitus varus, the traumatic force can be converted into a distraction force on the lateral condyle from the lateral collateral ligament (Fig. 2C).4 Of course, definite varus/valgus force during trauma would be the main determinant of fracture type. We merely suggest some converting effect of elbow alignment on varus/valgus force, which could raise the chance of a specific fracture type. For statistical analysis, we used multivariate analysis and case-matched comparisons. The frequency of the type of elbow fracture may differ according to age, sex, and injured side. As in our present study, patients with RNF tend to be older and patients with LCF tend to be younger.3,5,17,23 Although not proven statistically, some previous studies, as well as our present study, showed a slight tendency of male individuals to experience more SCF or LCF than RNF,3,17,23 as well as side differences according to the fracture type.3,17 Furthermore, carrying angle increases as a child grows older and differs according to sex and side.12,13 Therefore, there are many possible bias problems when assessing the association between the carrying angle and the fracture type, and the elimination of such bias-inducing factors was important in our study setting. Multivariate analysis can not only provide the relative contributions of different causes but can also effectively eliminate confounders.19 Among the many possible multivariate analytic models, we used the baselinecategory logistic model because there were 3 (.2) fracture groups. In baseline-category logistic models, each response category is paired with a baseline category (SCF patients in
FIGURE 2. Schematic drawing showing the transmission of traumatic force when a child falls onto an outstretched arm. A, An SCF is due to impact of the olecranon on the distal humerus when a vertical stress is applied to a hyperextended elbow. B, However, in the valgus-deviated elbow, the traumatic force (axial loading) can be converted into compression force of the radial head and neck, resulting in an RNF, (C) and in the varus-deviated elbow, the traumatic force can be converted into a distraction force on the lateral condyle from the lateral collateral ligament, resulting in an LCF. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jorthotrauma.com |
e257
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
Kang and Park
the present study).24 The propensity score is the conditional probability of a defined outcome being calculated from the covariates. To reduce bias, propensity score analysis can be used to balance the confounding factors in the 2 groups.25 The results of a randomized controlled trial and propensity score analysis were reported to be very similar, and the remaining bias after propensity score analysis is regarded to be low.26 Conclusively, among patients with SCF, RNF, and LCF, a more valgus-deviated elbow predisposes the patient to RNF, whereas a more varus-deviated elbow predisposes these cases to LCF. Elbow alignment, which may influence the transmission of traumatic force during a fall onto an outstretched elbow, could be a predisposing factor for specific types of pediatric elbow fracture. The results provide the additional information about the injury mechanisms of pediatric elbow fracture and may deepen our understanding of the fractures. REFERENCES 1. Landin LA, Danielsson LG. Elbow fractures in children. An epidemiological analysis of 589 cases. Acta Orthop Scand. 1986;57:309–312. 2. Tandon T, Shaik M, Modi N. Paediatric trauma epidemiology in an urban scenario in India. J Orthop Surg. 2007;15:41–45. 3. Behdad A, Behdad S, Hosseinpour M. Pediatric elbow fractures in a major trauma center in Iran. Arch Trauma Res. 2013;1:172–175. 4. John SD, Wherry K, Swischuk LE, et al. Improving detection of pediatric elbow fractures by understanding their mechanics. Radiographics. 1996; 16:1443–1460; quiz 1463–1444. 5. Cha SM, Shin HD, Kim KC, et al. Percutaneous reduction and leverage fixation using K-wires in paediatric angulated radial neck fractures. Int Orthop. 2012;36:803–809. 6. Markolf KL, Dunbar AM, Hannani K. Mechanisms of load transfer in the cadaver forearm: role of the interosseous membrane. J Hand Surg Am. 2000;25:674–682. 7. Khare GN, Goel SC, Saraf SK, et al. New observations on carrying angle. Indian J Med Sci. 1999;53:61–67. 8. Howard AW, Macarthur C, Rothman L, et al. School playground surfacing and arm fractures in children: a cluster randomized trial comparing sand to wood chip surfaces. Plos Med. 2009;6:e1000195. 9. Igbigbi PS, Manda K. Epidemiology of humeral fractures in Malawi. Int Orthop. 2004;28:338–341.
e258
| www.jorthotrauma.com
10. Glance LG, Li Y, Osler TM, et al. Impact of obesity on mortality and complications in trauma patients. Ann Surg. 2014;259:576–581. 11. Mendiguchia J, Alentorn-Geli E, Idoate F, et al. Rectus femoris muscle injuries in football: a clinically relevant review of mechanisms of injury, risk factors and preventive strategies. Br J Sports Med. 2013;47:359–366. 12. Yilmaz E, Karakurt L, Belhan O, et al. Variation of carrying angle with age, sex, and special reference to side. Orthopedics. 2005;28:1360–1363. 13. Balasubramanian P, Madhuri V, Muliyil J. Carrying angle in children: a normative study. J Pediatr Orthop B. 2006;15:37–40. 14. Marsh JL, Slongo TF, Agel J, et al. Fracture and dislocation classification compendium - 2007: Orthopaedic Trauma Association classification, database and outcomes committee. J Orthop Trauma. 2007;21:S1–S133. 15. Pace JL, Wiater B, Schmale G, et al. Baumann angle and radial-ulnar overlap: a radiographic study to control for the angle of the x-ray beam. J Pediatr Orthop. 2012;32:467–472. 16. Beals RK. The normal carrying angle of the elbow. A radiographic study of 422 patients. Clin Orthop Relat Res. 1976;119:194–196. 17. Koh KH, Seo SW, Kim KM, et al. Clinical and radiographic results of lateral condylar fracture of distal humerus in children. J Pediatr Orthop. 2010;30:425–429. 18. Van Roy P, Baeyens JP, Fauvart D, et al. Arthro-kinematics of the elbow: study of the carrying angle. Ergonomics. 2005;48:1645–1656. 19. Katz MH. Setting up a multivariable analysis. In: Multivariable Analysis: A Practical Guide for Clinicians. 2nd ed. New York, NY: Cambridge University Press; 2006:73–95. 20. Goldfarb CA, Patterson JM, Sutter M, et al. Elbow radiographic anatomy: measurement techniques and normative data. J Shoulder Elbow Surg. 2012;21:1236–1246. 21. Chapleau J, Canet F, Petit Y, et al. Validity of goniometric elbow measurements: comparative study with a radiographic method. Clin Orthop Relat Res. 2011;469:3134–3140. 22. McLauchlan GJ, Walker CR, Cowan B, et al. Extension of the elbow and supracondylar fractures in children. J Bone Joint Surg Br. 1999;81:402– 405. 23. Tan BH, Mahadev A. Radial neck fractures in children. J Orthop Surg. 2011;19:209–212. 24. Agresti A. An Introduction to Categorical Data Analysis. 2nd ed. New York, NY: Wiley; 2007. 25. D’Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med. 1998;17:2265–2281. 26. Kuss O, Legler T, Borgermann J. Treatments effects from randomized trials and propensity score analyses were similar in similar populations in an example from cardiac surgery. J Clin Epidemiol. 2011;64:1076–1084.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Predisposing Effect of Elbow Alignment on the Elbow Fracture Type in Children Seungcheol Kang, MD and Soo-Sung Park, MD, PhD
Objectives: Under the hypothesis that the elbow alignment,
Key Words: elbow fracture, carrying angle, pediatrics, supracondylar humeral fracture, lateral condylar fracture, radial neck fracture, trauma
namely the carrying angle, could predispose individuals to a specific type of pediatric elbow fracture after a fall onto an outstretched arm, we investigated the relationship between radiographic carrying angle and elbow fracture type in children.
Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
Design: Retrospective case–control study.
(J Orthop Trauma 2015;29:e253–e258)
Setting: Level I pediatric trauma center. Patients/Participants: We reviewed 374 children who were diagnosed with supracondylar fracture (SCF, n = 208), lateral condylar fracture (LCF, n = 132), and radial neck fracture (RNF, n = 34). Intervention: The association between the radiographic carrying angle and the fracture type was investigated. Main Outcome Measurements: To adjust for bias, 2 statistical methods were used: multivariate analysis using a baseline-category logistic model and a case-matching method using propensity score analysis. Results: In the multivariate analysis, with SCF patients set as the baseline category, a more valgus-deviated elbow (increased carrying angle, P = 0.011) predisposed individuals to RNF, whereas a more varus-deviated elbow (decreased carrying angle, P , 0.001) predisposed them to LCF. In the case-matched analysis, there were also significant differences in carrying angles between RNF and casematched SCF patients (14.3 vs. 11.4 degrees, P = 0.013) and between LCF and case-matched SCF patients (7.7 vs. 11.7 degrees, P , 0.001).
Conclusions: Elbow alignment, which may influence the transmission of traumatic force during a fall onto an outstretched elbow, could be a predisposing factor for specific types of pediatric elbow fracture. The results provide the additional information about the injury mechanisms of pediatric elbow fracture and may deepen our understanding of the fractures.
Accepted for publication February 23, 2015. From the Department of Orthopaedic Surgery, Asan Medical Center Children’s Hospital, University of Ulsan College of Medicine, Seoul, South Korea. Presented in part at the 58th Annual Fall Congress of the Korean Orthopaedic Association, October 18, 2014, Seoul, South Korea. The authors report no conflict of interest. Reprints: Soo-Sung Park, MD, PhD, Department of Orthopedic Surgery, Asan Medical Center Children’s Hospital, University of Ulsan College of Medicine, 88, Olympic-ro, 43-gil, Songpa-gu, Seoul 138-736, South Korea (e-mail: [email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Orthop Trauma Volume 29, Number 8, August 2015
INTRODUCTION Elbow fracture is one of the most common traumatic problems in children, comprising 7%–20% of all pediatric fractures.1,2 Of the various types of elbow fractures, the supracondylar fracture (SCF) is generally the most common, followed by the lateral condylar fracture (LCF) and radial neck fracture (RNF).3 The typical cause of these elbow fractures in children is a fall onto an outstretched elbow4,5: when a child falls onto an outstretched arm, the injury mechanism of hyperextension with vertical stress associates with SCF, hyperextension with valgus stress associates with RNF, and hyperextension with varus stress associates with LCF.4 The degree of angulation or alignment generally affects the mechanism of load transfer.6 Thus, the degree of elbow angulation might theoretically influence the transmission of traumatic force and associate with a specific elbow fracture type. However, few studies have examined the association between elbow alignment and fracture type. In a previous experimental study, possible associations of elbow alignment with SCF and LCF were suggested,7 but that study used only 2 cadavers for each fracture. Therefore, we hypothesized that the elbow alignment in the coronal plane, namely the carrying angle, might favor a specific type of elbow fracture in children after a fall onto an outstretched arm. More specifically, a more valgus-deviated elbow might predispose individuals to RNF, whereas a more varus-deviated elbow might predispose individuals to LCF. However, there are many potential bias problems when assessing the association between carrying angle and fracture type. Some different characteristics according to the fracture type have already been reported, such as patient age and sex and the injured side.3,8,9 Patient’s height and weight or their combination would affect the injury mechanism.10,11 Moreover, the carrying angle itself differs according to age and sex.12,13 Thus, adequate adjustment of confounding factors needs to be achieved to evaluate the correct relationship. Pursuant to these concerns, under the hypothesis that elbow alignment could affect the type of pediatric elbow fracture, we measured and compared the radiographic carrying angle of children who were diagnosed with isolated SCF, LCF, or RNF after a fall www.jorthotrauma.com |
e253
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Kang and Park
onto an outstretched arm, and finally evaluated the effect of radiographic carrying angle on the fracture type after bias adjustments. To adjust for bias, we used 2 statistical methods: multivariate analysis using a baseline-category logistic model and a case-matching method using propensity score analysis.
PATIENTS AND METHODS Patients Selection Patients younger than 16 years who were diagnosed with SCF, LCF, or RNF (13-A, 13-B, or 21-A, respectively, by AO/OTA classification14) in our institution from January 2005 to December 2013 were retrospectively reviewed. Of the 240 SCF patients, those with combined fracture (n = 17) or without an outstretched injury mechanism (direct injury or undocumented injury mechanism, n = 15) were excluded. Of the 142 LCF patients, those with combined fracture (n = 2) or without an outstretched injury mechanism (n = 8) were excluded. Of the 46 RNF patients, those with combined fracture (n = 12) were excluded; all remaining RNF patients had an outstretched injury mechanism. Finally, these exclusion criteria left 374 patients for the review (208 patients with SCF, 132 with LCF, and 34 with RNF). The institutional review board of our institute approved this study.
J Orthop Trauma Volume 29, Number 8, August 2015
S.K.) performed assessments in 30 randomly selected cases. The interobserver reliabilities of the measurements were evaluated using intraclass correlation coefficients. The intraclass correlation coefficient of interobserver reliabilities for the measurements was greater than 0.9, representing satisfactory agreement. Thus, measurements taken by a single investigator (S.K.) were used in the analyses. For clarity, we refer to the “carrying angle” as the “radiographic carrying angle” throughout the Patients and Methods, and Results section.
Statistical Methods For the purpose of comparing the patients according to the fracture type, the significance of differences between frequencies was calculated using the x2 test. Comparisons of mean values among the 3 fracture groups were performed using analysis of variance. When significant differences were detected in the analyses, intergroup comparisons were performed using Scheffe post hoc tests. To evaluate the independent effect of carrying angle on the fracture type, we used 2 statistical methods to eliminate bias: multivariate analysis using
Investigated Characteristics We investigated the radiographic carrying angle; age, height, weight, and body mass index (BMI) at injury; sex; and injured side (right or left). Simple radiographs of the uninjured elbow taken at the first visit to our institution were used to assess the radiographic carrying angle. In our institution, the radiographs of the contralateral uninjured elbow were routinely checked for the children with elbow fracture at their initial visit and regarded as a reference for treatment of injured elbow. All of the first visits were within 5 days of the injury. There was no significant left–right difference of carrying angle [10.3 6 5.3 degrees for right side (n = 152, measured from uninjured left side) vs. 10.7 6 4.8 degrees for left side (n = 222, measured from uninjured right side), P = 0.457]. The elbow anteroposterior radiographs were obtained with a following standardized technique: a child seats at the end of a table; the arm is fully extended with shoulder and elbow on same horizontal plane, and the forearm is supinated; the x-ray beam is angled perpendicularly to the table centering the elbow joint. Before passing the image to our image view system, the digitalized x-ray is prechecked by a disciplined radiographer. Then, the following items are checked: the rotation of epicondyle, the center of x-ray beam, and the overlap between the proximal ulna and the radial head, radial neck, and radial tuberosity.15 If there is any insufficiency or blurring, the x-ray is repeated. The radiographic carrying angle was measured as the angle between the following 2 lines (Fig. 1): a line connecting the midpoint of the metaphysis and the diaphysis of the distal humerus and a line connecting the midpoint of the ulna at the level of the bicipital tuberosity of the radius and the midpoint of the proximal end of the ulna.16–18 To test the interobserver reliability for the assessment of the carrying angle, 2 observers (S.-S.P. and
FIGURE 1. The radiographic carrying angle was measured as the angle between the following 2 lines: a line connecting the midpoint of the metaphysis and the diaphysis of the distal humerus and a line connecting the midpoint of the ulna at the level of the bicipital tuberosity of the radius and the midpoint of the proximal end of the ulna.
e254
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
| www.jorthotrauma.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
Predisposing Effect of Elbow Alignment
a baseline-category logistic model and case-matching analysis using propensity score analysis.19 For the multivariate analysis, using the baseline-category logistic model, the SCF patients were set as a baseline or reference category. Thus, odds ratios of .1 in each category meant that the variable or its increasing value favors the categorized fracture compared with SCF. Carrying angle, age, sex, injured side, and BMI were included in the analysis, but weight and height were excluded because they interfered with BMI. Statistical analyses were conducted with SPSS for Windows statistical package (version 21; IBM Co, Armonk, NY) and P , 0.05 were considered significant. For the selection of a case-matched control group, we used propensity score analysis. Among the 208 SCF patients, 34 and 132 patients were selected as control groups (1:1 ratio) for RNF and LCF patients, respectively. The propensity score was calculated for each patient based on logistic regression analysis of the probability of experiencing RNF and LCF using patient characteristics of age (in years), sex (male or female), height (in centimeters), weight (in kilograms), and side of injured arm (right or left). These characteristics were selected based on a discussion between 2 experienced and board-certified orthopaedic surgeons (S.K. and S.-S.P.) after a literature review.3,8–13 Using the propensity scores, we created a 1:1 matching with the nearest neighbor approach. Propensity score analysis was performed with R statistical software (version 3.1.0; http://cran.r-project.org/). Then, the SPSS software was also used to compare the variables between RNF or LCF patients and their respective case-matched SCF patients. The significance of differences between mean values, including age, height, weight, BMI, and carrying angle, was calculated using the independent t test. The significance of differences between frequencies, including sex and injured side, was calculated using the x2 test; Fisher exact test was used in the case that the expected number of frequency was less than 5.
RESULTS Comparisons Among the Three Fracture Groups Comparisons among the fracture groups are shown in Table 1. The carrying angle was different in each group: the mean carrying angles were 14.3 6 4.3 degrees for the RNF
group, 11.8 6 4.5 degrees for the SCF group, and 7.7 6 4.4 degrees for the LCF group (P , 0.001). Subgroup analyses using Scheffe post hoc tests showed significant difference between each group (P = 0.010 for RNF and SCF groups, P , 0.001 for SCF and LCF groups, and P , 0.001 for RNF and LCF groups). Besides the carrying angle, there were significant differences in age, height, weight, and BMI.
Results of Multivariate Analysis Using a Baseline-Category Logistic Model The results of multivariate analysis for each type of fracture are shown in Table 2. A larger carrying angle (P = 0.011), older age (P = 0.025), and larger BMI (P = 0.020) favored RNF occurrence compared with SCF, and a smaller carrying angle (P , 0.001), younger age (P = 0.023), and larger BMI (P , 0.001) favored LCF occurrence compared with SCF.
Results of Case-Matched Comparisons Using Propensity Score Analysis The results of the case-matched comparisons are shown in Table 3. There were no significant differences in matched variables reflecting well-balancing of the confounding factors, such as age, sex, height, weight, and injured side. The carrying angle was significantly larger in RNF patients than in their case-matched SCF patients and was significantly smaller in LCF patients than in their case-matched SCF patients.
DISCUSSION Traumatic forces in outstretched arm injuries are the main determinants of the type of pediatric elbow fracture. Elbow alignment can theoretically influence the transmission of traumatic force, but its association with fracture type has scarcely been studied. Therefore, we hypothesized that elbow alignment could also result in a specific fracture type by affecting the transmission of traumatic force in an outstretched arm. The results of the present study revealed the significant relationships between carrying angle and the specific fracture type. This is the first study about the predisposing effect of carrying angle on the specific type of
TABLE 1. Characteristics of the Children in This Study Diagnosed With RNF, SCF, and LCF Variable Age (years) Sex, n (%) Male Female Height (cm) Weight (kg) Injured side, n (%) Right Left BMI (kg/m2) Carrying angle (degrees)
Total (n = 374)
RNF (n = 34)
Patients With SCF (n = 208)
Patients With LCF (n = 132)
P
6.0 6 2.7
7.8 6 2.7
6.1 6 2.7
5.5 6 2.6
,0.001* 0.100
248 (66.3) 126 (33.7) 116.1 6 17.4 23.2 6 9.1
19 (55.9) 15 (44.1) 124.2 6 16.2 27.6 6 9.7
133 (63.9) 75 (36.1) 116.4 6 17.3 22.6 6 8.4
96 (72.7) 36 (27.3) 113.6 6 17.5 23.1 6 9.7
152 (40.6) 222 (59.4) 16.7 6 2.6 10.6 6 5.0
13 (38.2) 21 (61.8) 17.4 6 2.6 14.3 6 4.3
78 (37.5) 130 (62.5) 16.2 6 2.2 11.8 6 4.5
61 (46.2) 71 (53.8) 17.3 6 3.0 7.7 6 4.4
0.007* 0.010* 0.268
,0.001* ,0.001*
*P , 0.05.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jorthotrauma.com |
e255
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
Kang and Park
TABLE 2. Results of Multivariate Analysis Using a Baseline-Category Logistic Model RNF (vs. SCF) Carrying angle (degrees) Age (years) Sex Male Female Affected side Right Left BMI (kg/m2)
LCF (vs. SCF)
OR (95% CI)
P
OR (95% CI)
P
1.122 (1.027–1.225) 1.178 (1.021–1.359)
0.011* 0.025* 0.109
0.820 (0.772–0.870) 0.885 (0.797–0.983)
,0.001* 0.023* 0.345
Variable
0.519 (0.233–1.156) 1 (reference)
1.299 (0.755–2.234) 1 (reference) 0.718
1.159 (0.521–2.578) 1 (reference) 1.207 (1.031–1.414)
0.217 1.369 (0.832–2.254) 1 (reference) 1.225 (1.098–1.366)
0.020*
,0.001*
*P , 0.05. OR, odds ratio; CI, confidence interval.
pediatric elbow fracture, with bias adjustment using both multivariate analysis and case matching. There are a few considerations when interpreting the results of this study. First, we included patients with an outstretched arm injury but did not investigate the detailed traumatic force applied during the trauma, whether varus, valgus, or only vertical stress. In our anecdotal clinical experience, many children do not remember the exact injury mechanism, and some children even regard their injury situations irrelevant to the real traumatic force when compared with injury witnesses. For this reason, we were unable to investigate the detailed traumatic force and its correlation with elbow alignment. Second, the retrospective nature and relatively small number of RNF patients may cause bias problems that limit the interpretation of our results. To overcome these limitations, we used 2 statistical methods, multivariate analysis and case matching. Although these methods can provide excellent ways of minimizing bias, no statistical process is able to completely eliminate the inherent differences. Third, although there are many possible methods used to determine the carrying angle,18 the carrying angle is
traditionally measured from the shoulder to the wrist. However, we measured the radiographic carrying angle to establish the elbow alignment because the transmission of the traumatic force through the elbow is directly affected by angulation around the elbow, not the whole arm. Furthermore, bony alignment seemed to be more objective and more appropriate for assessing the transmission of the force. Radiographic carrying angle is reported to be substantially reliable,20 and radiographic measurement is thought to be more precise than clinical measurement.21 Fourth, we did not investigate the effect of sagittal alignment. There are various possible methods for evaluating sagittal alignment, such as measuring the maximal degree of elbow extension, ligamentous laxity, and/or humeroulnar angle on lateral radiograph with elbows fully extended. However, we did not routinely evaluate the degree of elbow extension or the ligamentous laxity of the patients with fractures, and the lateral radiographs were taken with the elbow flexed at a right angle. Although the fracture types were different from our study, McLauchlan et al22 already reported no difference of elbow extension, namely sagittal alignment, between the children
TABLE 3. Results of Case-Matched Comparisons Variable Age (years) Sex, n (%) Male Female Height (cm) Weight (kg) Injured side, n (%) Right Left BMI (kg/m2) Carrying angle (degrees)
Patients With RNF (n = 34)
Control Group* (n = 34)
7.8 6 2.7
8.1 6 3.1
19 (55.9) 15 (44.1) 124.2 6 16.2 27.6 6 9.7
18 (52.9) 16 (47.1) 127.0 6 20.2 28.7 6 10.1
13 (38.2) 21 (61.8) 17.4 6 2.6 14.3 6 4.3
8 (26.5) 25 (73.5) 17.4 6 2.6 11.4 6 5.0
P 0.717 0.808
0.525 0.652 0.300
0.914 0.013‡
Patients With LCF (n = 132)
Control Group† (n = 132)
5.5 6 2.6
5.3 6 2.4
96 (72.7) 36 (27.3) 113.6 6 17.5 23.1 6 9.7
133 (63.9) 75 (36.1) 111.3 6 17.0 21.5 6 8.6
61 (46.2) 71 (53.8) 17.3 6 3.0 7.7 6 4.4
78 (37.5) 130 (62.5) 16.7 6 2.1 11.7 6 4.7
P 0.501 0.100
0.266 0.156 0.268
0.069 ,0.001‡
*SCF patients who were case matched to RNF patients. †SCF patients who were case matched to LCF patients. ‡P , 0.05.
e256
| www.jorthotrauma.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
with SCF and distal radius fracture. We also suppose the sagittal alignment would not differ among the elbow fractures, but further studies are needed. Fifth, radiographic measurement is affected by positioning of the elbow, and the positioning can be difficult especially in children.15 The mean difference of carrying angle between RNF and SCF patients is only 2.5 degrees. Therefore, the possibilities of measurement error should be considered during interpretation of the results. Finally, the correlation between the elbow alignment and the transmission of traumatic force could not be proven by the results of this study. The results of this study should be supported by a future well-designed biomechanical study. In the multivariate analysis with SCF patients set as a baseline category, a more valgus-deviated elbow (increased carrying angle, P = 0.011) and older age (P = 0.025) predisposed individuals to RNF, whereas a more varus-deviated elbow (decreased carrying angle, P , 0.001) and younger age (P = 0.023) predisposed individuals to LCF. A higher BMI was significantly associated with both RNF (P = 0.020) and LCF (P , 0.001) when compared with SCF. The significant difference in carrying angle was also confirmed by case-matching analysis using propensity score analysis; when other characteristics were distributed equally, the carrying angle was significantly different between RNF and SCF patients (14.3 6 4.3 vs. 11.4 6 5.0 degrees, P = 0.013) and between LCF and SCF patients (7.7 6 4.4 vs. 11.7 6 4.7 degrees, P , 0.001). We can suggest a possible explanation for the correlation between carrying angle and the specific fracture type. Carrying angle seems to alter the transmission of traumatic force: the axial loading during trauma might be converted into valgus force in a more valgus-deviated elbow and into varus force in a more varus-deviated elbow. Figure 2 shows the schematic transmission of traumatic force when someone falls
Predisposing Effect of Elbow Alignment
onto an outstretched arm. SCF is suggested to occur due to the impact of the olecranon on the distal humerus when a vertical stress is applied during a fall onto an outstretched arm (Fig. 2A).4,5 However, in an elbow with cubitus valgus, the traumatic force (axial loading) can be converted into a compression force of the radial head and neck (Fig. 2B),4,5 whereas in an elbow with cubitus varus, the traumatic force can be converted into a distraction force on the lateral condyle from the lateral collateral ligament (Fig. 2C).4 Of course, definite varus/valgus force during trauma would be the main determinant of fracture type. We merely suggest some converting effect of elbow alignment on varus/valgus force, which could raise the chance of a specific fracture type. For statistical analysis, we used multivariate analysis and case-matched comparisons. The frequency of the type of elbow fracture may differ according to age, sex, and injured side. As in our present study, patients with RNF tend to be older and patients with LCF tend to be younger.3,5,17,23 Although not proven statistically, some previous studies, as well as our present study, showed a slight tendency of male individuals to experience more SCF or LCF than RNF,3,17,23 as well as side differences according to the fracture type.3,17 Furthermore, carrying angle increases as a child grows older and differs according to sex and side.12,13 Therefore, there are many possible bias problems when assessing the association between the carrying angle and the fracture type, and the elimination of such bias-inducing factors was important in our study setting. Multivariate analysis can not only provide the relative contributions of different causes but can also effectively eliminate confounders.19 Among the many possible multivariate analytic models, we used the baselinecategory logistic model because there were 3 (.2) fracture groups. In baseline-category logistic models, each response category is paired with a baseline category (SCF patients in
FIGURE 2. Schematic drawing showing the transmission of traumatic force when a child falls onto an outstretched arm. A, An SCF is due to impact of the olecranon on the distal humerus when a vertical stress is applied to a hyperextended elbow. B, However, in the valgus-deviated elbow, the traumatic force (axial loading) can be converted into compression force of the radial head and neck, resulting in an RNF, (C) and in the varus-deviated elbow, the traumatic force can be converted into a distraction force on the lateral condyle from the lateral collateral ligament, resulting in an LCF. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jorthotrauma.com |
e257
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma Volume 29, Number 8, August 2015
Kang and Park
the present study).24 The propensity score is the conditional probability of a defined outcome being calculated from the covariates. To reduce bias, propensity score analysis can be used to balance the confounding factors in the 2 groups.25 The results of a randomized controlled trial and propensity score analysis were reported to be very similar, and the remaining bias after propensity score analysis is regarded to be low.26 Conclusively, among patients with SCF, RNF, and LCF, a more valgus-deviated elbow predisposes the patient to RNF, whereas a more varus-deviated elbow predisposes these cases to LCF. Elbow alignment, which may influence the transmission of traumatic force during a fall onto an outstretched elbow, could be a predisposing factor for specific types of pediatric elbow fracture. The results provide the additional information about the injury mechanisms of pediatric elbow fracture and may deepen our understanding of the fractures. REFERENCES 1. Landin LA, Danielsson LG. Elbow fractures in children. An epidemiological analysis of 589 cases. Acta Orthop Scand. 1986;57:309–312. 2. Tandon T, Shaik M, Modi N. Paediatric trauma epidemiology in an urban scenario in India. J Orthop Surg. 2007;15:41–45. 3. Behdad A, Behdad S, Hosseinpour M. Pediatric elbow fractures in a major trauma center in Iran. Arch Trauma Res. 2013;1:172–175. 4. John SD, Wherry K, Swischuk LE, et al. Improving detection of pediatric elbow fractures by understanding their mechanics. Radiographics. 1996; 16:1443–1460; quiz 1463–1444. 5. Cha SM, Shin HD, Kim KC, et al. Percutaneous reduction and leverage fixation using K-wires in paediatric angulated radial neck fractures. Int Orthop. 2012;36:803–809. 6. Markolf KL, Dunbar AM, Hannani K. Mechanisms of load transfer in the cadaver forearm: role of the interosseous membrane. J Hand Surg Am. 2000;25:674–682. 7. Khare GN, Goel SC, Saraf SK, et al. New observations on carrying angle. Indian J Med Sci. 1999;53:61–67. 8. Howard AW, Macarthur C, Rothman L, et al. School playground surfacing and arm fractures in children: a cluster randomized trial comparing sand to wood chip surfaces. Plos Med. 2009;6:e1000195. 9. Igbigbi PS, Manda K. Epidemiology of humeral fractures in Malawi. Int Orthop. 2004;28:338–341.
e258
| www.jorthotrauma.com
10. Glance LG, Li Y, Osler TM, et al. Impact of obesity on mortality and complications in trauma patients. Ann Surg. 2014;259:576–581. 11. Mendiguchia J, Alentorn-Geli E, Idoate F, et al. Rectus femoris muscle injuries in football: a clinically relevant review of mechanisms of injury, risk factors and preventive strategies. Br J Sports Med. 2013;47:359–366. 12. Yilmaz E, Karakurt L, Belhan O, et al. Variation of carrying angle with age, sex, and special reference to side. Orthopedics. 2005;28:1360–1363. 13. Balasubramanian P, Madhuri V, Muliyil J. Carrying angle in children: a normative study. J Pediatr Orthop B. 2006;15:37–40. 14. Marsh JL, Slongo TF, Agel J, et al. Fracture and dislocation classification compendium - 2007: Orthopaedic Trauma Association classification, database and outcomes committee. J Orthop Trauma. 2007;21:S1–S133. 15. Pace JL, Wiater B, Schmale G, et al. Baumann angle and radial-ulnar overlap: a radiographic study to control for the angle of the x-ray beam. J Pediatr Orthop. 2012;32:467–472. 16. Beals RK. The normal carrying angle of the elbow. A radiographic study of 422 patients. Clin Orthop Relat Res. 1976;119:194–196. 17. Koh KH, Seo SW, Kim KM, et al. Clinical and radiographic results of lateral condylar fracture of distal humerus in children. J Pediatr Orthop. 2010;30:425–429. 18. Van Roy P, Baeyens JP, Fauvart D, et al. Arthro-kinematics of the elbow: study of the carrying angle. Ergonomics. 2005;48:1645–1656. 19. Katz MH. Setting up a multivariable analysis. In: Multivariable Analysis: A Practical Guide for Clinicians. 2nd ed. New York, NY: Cambridge University Press; 2006:73–95. 20. Goldfarb CA, Patterson JM, Sutter M, et al. Elbow radiographic anatomy: measurement techniques and normative data. J Shoulder Elbow Surg. 2012;21:1236–1246. 21. Chapleau J, Canet F, Petit Y, et al. Validity of goniometric elbow measurements: comparative study with a radiographic method. Clin Orthop Relat Res. 2011;469:3134–3140. 22. McLauchlan GJ, Walker CR, Cowan B, et al. Extension of the elbow and supracondylar fractures in children. J Bone Joint Surg Br. 1999;81:402– 405. 23. Tan BH, Mahadev A. Radial neck fractures in children. J Orthop Surg. 2011;19:209–212. 24. Agresti A. An Introduction to Categorical Data Analysis. 2nd ed. New York, NY: Wiley; 2007. 25. D’Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med. 1998;17:2265–2281. 26. Kuss O, Legler T, Borgermann J. Treatments effects from randomized trials and propensity score analyses were similar in similar populations in an example from cardiac surgery. J Clin Epidemiol. 2011;64:1076–1084.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
