Abstract
Introduction:
Proximal third diaphyseal fractures of the radius and ulna represent an onerous fracture pattern due to difficulty maintaining acceptable alignment with nonoperative management. Our aim was to identify the factors that increase the odds for a surgical treatment of these fractures. Recognizing these factors can raise awareness to patients who are more likely to require additional care and assist clinicians in counseling families, targeting treatment plans, and constructing follow-up protocols. We hypothesized that the age of the patient, the amount of initial fracture displacement, and the angulation of the fracture would predict the need for operative treatment.
Methods:
We retrospectively reviewed 276 proximal third diaphyseal forearm fractures at a single tertiary care institution. All patients underwent a nonoperative treatment trial, and if failed continued to surgery. Following a univariate analysis, we constructed a binary multivariate logistic regression model that included age, initial translation, and initial angulation to assess the association between the tested variables.
Results:
A regression model revealed that age (10 years and older, odds ratio: 8.2, 95% confidence interval: 3.9–17.24, p < 0.001) and radius translation of more than 100% (odds ratio: 7.06, 95% confidence interval: 2.69–18.52, p < 0.001) were associated with the need for surgical treatment. Initial fracture angulation lacked an association with a surgical treatment (odds ratio: 0.81 95% confidence interval: 0.38–1.74, p = 0.59).
Conclusion:
Age above 10 years and 100% initial translation of the radius fracture increased the odds for an ultimate decision to perform a surgery. Initial angulation, although often being the most remarkable radiographic feature, was not associated with a nonoperative treatment failure. We recommend an initial reduction attempt after counseling patients and their families that there is a high rate of conversion to operative treatment when the above features are met.
Level of evidence:
level III.
Introduction
Forearm fractures are common in children, with an incidence of 3%–6% of all pediatric fractures.1–4 Traditional treatment of these fractures involves closed reduction (CR) and immobilization with close radiographic follow-up to evaluate for re-displacement to discern the need for potential operative intervention. The failure of nonoperative management has ranged from 16% to 51% depending on the definition of successful treatment.5–7 Fracture location and patient age influence the success of nonoperative treatment in the pediatric patient population.6,7 Distal fractures and fractures in younger patients have been shown to have greater remodeling potential and subsequently greater success with nonoperative management given the larger degree of acceptable variation in alignment.8,9 Conversely, proximal fractures—being further away from the more active distal radius growth plate—have been shown to have decreased remodeling potential and thus require tighter parameters for successful closed treatment. 10 In addition, the muscle bulk of the proximal forearm limits the ability for molded splints and casts to adequately maintain fracture alignment. As such, proximal forearm malunions are more common than middle and distal malunions and have been shown to contribute to decreased pronation and supination.9,11,12
Studies have also shown that patients aged above 10 years with proximal third radius fractures and minimally angulated ulnar fractures had an increased incidence of conversion to surgical intervention due to interval loss of reduction.6,7 Although previous literature has highlighted the increased incidence of surgical treatment in proximal third forearm fractures compared with midshaft or distal fractures, only one study to date has evaluated what factors contribute to this trend, citing patients aged above 10 years have increased incidence of surgical management; however, this study only included eight extra-articular proximal third fractures, with the remainder being Monteggia fractures or Monteggia variants. 13 Determining which factors, if any, are associated with unsuccessful CR attempts in this subset of patients could shed light on those patients who are more prone to require surgical treatment, predict the difficulty level and success rates for CR attempts, promote a case-specific counseling, and guide more efficient follow-up appointments following an apparent successful CR. Our aim was to identify the factors that increase the odds of surgical treatment of these fractures. Our hypothesis was that older pediatric patients and those with increased initial fracture displacement and angulation of the proximal radius are more likely to be unstable and would thus demonstrate an increased incidence of malalignment that would require an early conversion to surgical management.
Materials and methods
Following institutional review board approval, a cohort of proximal third diaphyseal forearm fractures (PTDF) was retrospectively reviewed at a single tertiary care institution. All diaphyseal forearm fractures treated between September 2014 and December 2018 were identified by the International Classification of Diseases (ICD), Ninth Edition (ICD-9), and ICD-10 codes from an electronic medical record (EMR) database. Images were reviewed by team members consisting of an attending pediatric orthopedic hand surgeon, pediatric orthopedic fellows, and residents. All clinical decision-making occurred under the supervision of 11 fellowship-trained attending pediatric orthopedic providers. The criteria for surgical fixation were unacceptable alignment following the initial reduction attempt or a loss of acceptable alignment during routine follow-up of attempted nonoperative treatment. Based on current literature, there is no consensus on operative indications for PTDF fractures.1,6,7 Therefore operative indications were created on a case-by-case basis by the attending surgeon treating the fracture using the criteria above. All patients underwent a CR attempt under a sedation or anesthesia prior to surgical fixation.
Patients were included if they were between the ages of 5 and 16 years and were skeletally immature. They were identified as having a fracture in the proximal third of the radius by dividing the diaphysis into thirds using Picture Archive and Communication System (PACS) software. The diaphysis ends were defined proximally as the proximal end of the bicipital tuberosity and distally by measuring the physical diameter and plotting this distance from the physis proximally.5,7
Patients were excluded if they demonstrated skeletal maturity with closed distal radius and ulnar physes, had less than 8 weeks of radiographic follow-up, or sustained ipsilateral upper extremity fractures. In addition, six (2%) patients who proceeded straight to operative intervention without any CR attempt were excluded from the analysis. Radiographic measurements (Figure 1) were performed by the research team on initial injury X-ray films and subsequent follow-up radiographs until healing was identified, with a minimum of 8 weeks between injury and final follow-up films.
(a, b) A 7-year-old female with a proximal third diaphyseal radius fracture. The fracture was angulated on the lateral view (X) but had no angulation on the coronal view (Y). The fracture translation was <50% (group A). (c, d) The patient was treated successfully with a closed reduction and a 6-week casting.
The EMR was queried for age, sex, injured side, mechanism of injury, prior related injuries, and any related diagnoses. We also recorded reduction attempts, episodes of cast removal and placement of a molded cast, and failures of nonoperative treatment, which were defined as patients requiring operative treatment with internal fixation. Our primary measurements consisted of angulation and translation at the fracture site as depicted in Figures 1–3. Initial angulation was measured on both coronal and sagittal views and was presented as both continuous and categorical variables (more or less than 15°). We categorized translation for patients into three categories, including <50% (Figure 1), 50%–100% (Figure 2), and >100% (Figure 3). This was defined by the maximal translation value measured on either the coronal or sagittal view of the radius. We did not categorize the translation of the ulna. Age was presented both as a continuous and as a categorical variable, divided into three groups (<7 years, 7–10 years, and >10 years). These cutoffs were chosen following our primary univariate analysis that revealed the age differences between patients who required surgery and those who did not (Table 1).
(a, b) An 11.9-year-old female with a proximal third diaphyseal radius fracture. The fracture was angulated mainly on the sagittal plane and had a 50% translation of radius (group B), along with a significant angulation of an ulna greenstick fracture. (c, d) The patient was treated successfully with a closed reduction and casting, and had a satisfying healing and remodeling at 8-month follow-up.
(a, b) A 14-year-old female with a proximal third diaphyseal radius fracture. The fracture was angulated on both the coronal and sagittal planes and had a >100% translation of radius (group C). After failing and attempt for a closed reduction and casting (c and d), the patient was treated successfully with intramedullary elastic nails (e and f).
Demographic and fracture element depiction of pediatric patients with radius or ulna proximal diaphyseal fractures.
SD: standard deviation; AP: anteroposterior.
Calculated with excluding group A, patients who have been treated with no reduction at all.
Angulation was divided into two groups: <15° and ≥15° (data presented for the former).
Statistical analysis
We used descriptive statistics to present raw data. Data analysis used an unpaired Student’s t-test for numerical data and a Chi-square test for categorical data. We constructed a multivariate logistic regression model to assess the association between tested variables and the ultimate need for a surgical treatment. Variables that were found to be statistically significant in the univariate analysis were included in the multivariate analysis. A p-value of 0.05 was considered statistically significant. SPSS 28.0 software (Chicago, IL, USA) was used for data collection and analysis.
Results
We identified 2153 consecutive radius and ulna diaphyseal fractures, from which 276 (12.82%) PTDFs were further reviewed. Of those patients, 210 (76.08%) completed a successful nonoperative treatment protocol and 66 (23.92%) were ultimately treated operatively. A univariate analysis revealed demographic and fracture pattern differences between these two groups (Table 1). Patients who were ultimately treated operatively were more likely to be 10 years or older compared to patients who were successfully treated by closed management. Similarly, this group was more probable to present with a significant fracture translation (Table 1). However, when looking at fracture angulation, there were no significant differences in the pre-reduction angulation in patients who were successfully treated nonoperatively when compared to patients treated operatively (Table 1). Out of the total 276 cases in which union could be defined by the final X-ray imaging study, 263 (95.29%) cases had X-ray images with proper anteroposterior and lateral views on which we could reliably measure the final fracture angulation. The depiction of the last follow-up fracture angulation is presented in Table 2. Among 22 patients with a translation of at least 100% who had measurable final fracture angulation, the difference in the radiographic outcome between 11 operated and 11 non-operated patients was not statistically significant (p > 0.05; Table 2).
Fracture angulation upon union of patients who were treated for radius proximal diaphyseal fractures.
Calculated with excluding group A, patients who have been treated with no reduction at all.
A binary multivariate logistic regression model that included age (categorical) and initial translation (categorical) was constructed. This model was able to predict 76.1% of cases and found age (10 years and older, odds ratio (OR): 8.2, 95% confidence interval (CI): 3.9–17.24, p < 0.001) and translation of more than 100% (OR: 7.06, 95% CI: 2.69–18.52, p < 0.001) as being associated with an ultimate need for surgery. When the initial fracture angulation was included in the model, it lacked an association with surgery (OR: 0.81, 95% CI: 0.38–1.74, p = 0.59), which was consistent with the univariate analysis.
Discussion
The successful treatment of PTDFs represents a challenging problem faced by pediatric and general orthopedic surgeons taking pediatric trauma calls. While only comprising 13% of the forearm fractures reported in our preliminary cohort, they present unique challenges in management. 14 Several previous studies have identified fractures of the proximal third of the radius and ulna as “bad-actors” among forearm fractures, leading to a higher rate of failed nonoperative management. With this in mind, we sought to identify key characteristics within this particularly problematic fracture pattern that would predict the need for operative intervention after preliminary closed reduction. Our evaluation of 276 consecutive PTDFs found that about one in four patients (23.92%) ultimately required operative intervention to maintain alignment (Table 1). We also found that patients with increased age and initial translation at the time of injury were more likely to require operative treatment. While post-reduction angulation is a common surgical indication, the initial angulation, although often being the most remarkable radiographic feature, was not associated with an eventual need for surgery.
Evidence supporting the difficulty of treating PTDFs nonoperatively has previously been reported in the literature. Wacker et al. 5 evaluated 309 complete radial shaft fractures and found that proximal third fractures were significantly more likely to fail nonoperative treatment and exceeded angulation criteria 70% of the time compared with more distal fractures (33%). 15 Failure of closed reduction and casting was 4.6 times higher (95% CI, 2.3–9.1) in proximal third fractures and 2.4 times greater (95% CI, 1.5–3.9) in proximal half fracture. Similarly, Bowman et al. 7 demonstrated nonoperatively treated both-bone forearm fractures involving a proximal third radial shaft fracture had a 6.8 times greater risk of failing nonoperative treatment. This has also been reported by Thomas et al. 16 who showed that half of the proximal third fractures treated nonoperatively resulted in poor outcomes. When looking at angular malalignment defined as greater than or equal to 10°, Price et al. 17 found 80% of nonoperatively treated proximal third forearm shaft fractures exceeded these criteria. However, these studies did not detail what measurable factors led to increased risks of failure of nonoperative treatment and thus can aid the orthopedic surgeon in counseling families more accurately on the risks of subsequent intervention. In the current study, all patients deemed appropriate for reduction underwent a closed reduction before proceeding to surgical intervention. Only those patients for whom a satisfactory alignment and stabilization could not be achieved by nonoperative means were treated operatively. In this study, 210 (76.08%) patients have been treated nonoperatively and reached final angulation that was similar to and sometimes even better than those who have been treated operatively (Table 2). Accordingly, this study supports the aforementioned body of literature, although the percentage of those requiring subsequent operative intervention was significantly lower at 23.92% (Table 1).
In this study, initial fracture translation, rather than initial angulation, was found to be associated with an ultimate need for operative care. This differs from a previous study by Bowman et al., 7 who did find angulation to be a predictive criterion for nonoperative treatment failure. An initial translation likely represents higher energy trauma and a significant disruption of the periosteal sleeve, leading to greater instability of the fracture. The initial angulation, on the contrary, is often more amenable to reduction by nonoperative means when compared to translation and does not necessarily represent a complete periosteal disruption. Thus, initial translation rather than initial angulation appears to be a more reliable indicator of fracture instability and ultimate failure of nonoperative treatment. This notion has been supported in this study by both univariate and multivariate analyses. We could not find previous studies that evaluated the association between the primary fracture translation and a failure of nonoperative treatment in PTDFs. Our finding that patients with loss of alignment were older at the time of injury (10.64 vs 7.51 years, p < 0.001) supports previous results from Bowman et al., 7 demonstrating that patients >10 years old were at higher risk of failing angular alignment criteria and therefore failing nonoperative management. Nevertheless, even in patients aged 10 years and above, or when the fracture translation was 100% and more, if CR and casting enabled to obtain an acceptable alignment, the nonoperative mode of treatment provided a satisfactory radiographic outcome (Table 2). This modality should therefore not be overlooked even when predisposing factors for an ultimate surgical treatment exist.
Our study demonstrates that nearly a quarter of PTDFs initially treated with closed reduction required an additional operative intervention. Wacker et al. 5 demonstrated that PTDFs failed nonoperative treatment in terms of angulation at an alarmingly high rate of 70% and suggested that earlier consideration of surgical management is reasonable. 15 While our study did not find such a high rate, we did find that 210 (76.08%) out of 276 patients who underwent a primary nonoperative treatment did not proceed to surgery. Therefore, like most other authors, we recommend an initial reduction attempt for all fractures with unacceptable angulation and displacement prior to recommending operative intervention. Fractures that prove to be unstable during reduction attempts, or during close follow-up, can be then indicated for surgical treatment.18,19
The current study outcomes must be interpreted within the context of the study design. The retrospective nature of our study did lend itself to selection bias, however; we used a large cohort of consecutive fractures and applied rigorous criteria for inclusion and exclusion to minimize this risk. One inherent limitation exists due to the tendency for biplanar radiographs to underestimate translation and angulation measurements as this is affected by the rotation of the forearm at the time of the radiograph. Furthermore, this study focused on the radius fracture, and fracture variables such as angulation and translation have been measured only on the early imaging studies. Accordingly, although we measured angulation, we did not measure or analyze the ulna fracture translation, which might have added additional information. In addition, because our study was not meant to be a clinical outcome study, we were unable to comment on patient clinical outcomes such as pain, return to activities, or persistent range of motion limitations after healing. Finally, the operative indications in our study were attending surgeon–specific, given the lack of consensus for operative treatment indications for this fracture pattern in pediatric patients, ultimately leading to this study’s conception.
Overall, in pediatric patients with a PTDF, operative intervention for loss of acceptable alignment is frequently encountered. Older patients with a greater initial translation of the radius were more likely to require operative intervention (Tables 1 and 2). Nevertheless, many patients who underwent an initial trial of nonoperative treatment were able to be managed successfully without surgery. We therefore recommend a reduction attempt for all PTDFs with an unacceptable alignment. We carefully counsel patients and their families that there is a high rate of conversion to operative treatment, especially in older patients with a substantial initial fracture translation, likely indicating unstable fracture patterns. A significant initial fracture angulation does not indicate an expected nonoperative treatment failure and should not deter surgeons from a trial of nonoperative care.
Footnotes
Author contributions
Kevin Williams: Data curation; Methodology; Writing—original draft; Writing review & editing.
Noelle Whyte: Conceptualization; Data curation; Methodology; Writing original draft; Writing—review & editing.
Jacob R Carl: Data curation; Formal analysis; Writing—original draft; Writing review & editing.
David Segal: Data curation; Formal analysis; Methodology; Software; Writing—original draft; Writing—review & editing.
Kevin J Little: Conceptualization; Data curation; Formal analysis; Methodology; Project administration; Supervision; Writing—original draft; Writing review & editing.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Ethical approval
This study had been conducted following the hospital ethical board committee and under compliance with Helsinki Declaration.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Informed consent
A waiver of informed consent was granted.
