2.1 Patients

Our study identified 1553 healthy adolescents aged 10–16 years with a forearm fracture. Among them, 291 adolescents had an isolated, complete diaphyseal forearm fracture, comprising open and closed fractures. Diaphyseal forearm fractures were identified according to the AO Paediatric Comprehensive Classification of long bone fractures.^27,28 The four relevant AO types included in this study were: 22r – D/4 .1, 22u – D/4 .1, 22r – D/5 .1 and 22u – D/5 .1.²⁷ We excluded other forearm fractures: bone contusion, torus/buckle fracture, distal radius fracture, radial neck fracture, olecranon fracture, Galeazzi fracture, Monteggia fracture, re-fracture and pathological fracture. 11 cases with missing elbow radiographs were excluded. Among the 291 adolescents identified, 147 were treated conservatively with cast immobilization (‘MRC¹1MRC stands for Manipulation, Reduction and Casting.’) while 144 were treated surgically (Fig. 1). Both groups of patients were followed up for one year. Within the surgical subgroup, 128 received TENS (‘TENS’) and the remaining 16 rigid plate fixation (‘Plate’). These patients underwent removal of implants after approximately six months and one year respectively. TENS patients underwent either a closed or open nailing approach, with each following the same surgical technique of retrograde distal lateral entry point for radius and antegrade proximal lateral entry point for ulna. The implant sizes for TENS range from 1.5mm to 3.5mm in our study, and they were selected according to a 0.4 ratio of implant width to the internal diameter of the isthmus of the affected radius or ulna diaphysis.

Close

Fig. 1 Flow diagram of patients aged 10–16 years, with an isolated, complete diaphyseal fracture who were managed over the period of January 2014 to January 2018. Patients were categorised by the treatment type received and further classified according to their respective treatment outcome as defined by our functional assessment criteria. Abbreviations: MRC, manipulation, reduction and casting; Plate, rigid plate fixation TENS, Titanium Elastic Nail System.

Export to PPT

2.2 Functional assessment

The functional assessment criteria used in this study accounted for treatment complication, range of motion (ROM) and functional ability of the affected forearm (and elbow) joint as documented at the final follow-up review (after implant removal for surgically treated patients). The treatment complications defined in TENS patients consisted of compartment syndrome, delayed union, implant irritation, malunion, neuropraxia, infection, pseudoarthrosis and tendon rupture.^14–17 The normal ROM of interest included: 0–140 ± 5° for elbow flexion-extension, 70 ± 5° for forearm pronation and 85 ± 5° for forearm supination.^29–31 A goniometer was used for all instances of ROM assessments. Normal functional ability of the forearm entailed conducting normal daily activities without significant complaints. We defined a ‘good outcome’ when the affected forearm returned to normal ROM without any treatment complication or loss of baseline functional ability. A ‘poor outcome’ was obtained if any of the three criteria was not met. Of note, rehabilitation status was not considered in our analysis as the routine post-operative management in our institution does not require a referral to a physiotherapist. However, any patient with functional loss or complications at interval reviews was referred to a physiotherapist for rehabilitation.

2.3 Modified Sauvegrain method

Digitised anteroposterior and lateral radiographs of the elbow were obtained by our institution's radiological department via standard protocols and reviewed by the first author. Level of fracture in the diaphyseal forearm bones was categorised as proximal third, mid-shaft or distal third, and fractures out of normal anatomic alignment considered displaced. The modified Sauvegrain method score was assessed using a scoring diagram (Fig. 2) and matched to a corresponding skeletal age based on patient's gender (Supplemental Figure A.1). The elbow radiographs of 50 randomly selected TENS patients were independently reviewed by a senior paediatric orthopaedic surgeon.

Close

Fig. 2 The modified Sauvegrain method is a 27-point scoring system for the assessment of skeletal age based on the four anatomical landmarks of the elbow: (1) lateral condyle and epicondyle (2) trochlea (3) olecranon apophysis (4) proximal radial epiphysis (Adapted from Dimeglio et al., 2005).

Export to PPT

2.4 Statistical analyses

Analyses for this study were carried out using the Statistical Analysis System (SAS®) software (Cary, NC, USA), University Edition. We removed all identifiers from our dataset and anonymised information prior to analysis. Consistent with the primary aim of the study, TENS patients were categorised according to treatment outcomes based on the functional assessment criteria. To compare the continuous variables between the good and poor outcome groups for TENS patients, the mean was compared using a t-test and median using a Wilcoxon rank-sum test. Categorical variables were analysed using either a Chi-squared test or Fisher's exact test as appropriate. Correlation between the modified Sauvegrain method score and skeletal age was investigated via Pearson's coefficient and Spearman's rank correlation tests. Statistical significance was set at p ≤ 0.05.

Using univariate logistic regression and ROC curve analysis, we compared the effectiveness of the modified Sauvegrain method score to chronological age as independent predictors of a poor outcome in TENS patients. In each case, the Youden-J index was used to identify a statistically optimal cut-off point to discriminate between low and high-risk groups for poor outcome in TENS patients. It determined a cut-off by maximising the value of sensitivity and specificity. The larger the Youden-J index, the better the performance at the cut-off. Additional independent predictors of poor outcome in TENS patients were explored using multivariate logistic regression.

3.1 Demographics and clinical characteristics of TENS patients (Table 1)

The mean (SD) chronological age of the 128 TENS patients was 12.6 (1.49) years. There were no TENS patients aged 16 years within the study period. The median (IQR) modified Sauvegrain method score and derived skeletal age was 25.5 (18.0, 27.0) and 14.1 (13.0, 14.5) years respectively. Of the 128 TENS patients, 113 (88.3 %) were males and 15 (11.7 %) were females. Chinese (53.9 %) made up the largest proportion of the sample, with Malays (31.3 %) second, Indians (7.03 %) third and the remainder (7.81 %) from other ethnicities. There were 114 (89.1 %) both-bone diaphyseal fractures compared to 14 (10.98 %) single-bone diaphyseal fractures. More than 86 % of the fractures located at either the ulna or radius occurred at the mid-shaft. Among these 128 TENS patients, 15 (11.7 %) were open fractures and 113 (88.3 %) closed fractures. Majority had a complete displaced fracture (96.1 %) and transverse fracture configuration (45.8 %). The most common mechanism of fracture was a fall from standing height (73.4 %). The open nailing approach was used for 89 (69.3 %) of the 128 TENS patient. An approximately equal number of patients received an implant size of either less than or equal to 2mm (53.1 %) or more than 2mm (49.9 %).

3.2 Difference between TENS patients with good and poor treatment outcomes (Table 1)

Of the 128 patients who received TENS treatment, 105 (82.0 %) achieved good outcomes, while 23 (18.0 %) received poor outcomes. Among male TENS patients, 92 (81.4 %) achieved good outcomes and 21 (18.6 %) received poor outcomes as compared to female TENS patients with 13 (86.7 %) and 2 (13.3 %) respectively. All poor outcomes were attributed to a limited range of motion of the forearm with 80 % (18 of 23) demonstrating a limitation of pronation or supination between six to 10° when compared to the unaffected side. No technical or surgical complications were noted. Between the good and poor outcome groups for TENS patients, there are significant differences in mean chronological age (p = 0.008), mean modified Sauvegrain method score (p = 0.013), mean skeletal age (p = 0.005), median modified Sauvegrain method score (p = 0.006) and median skeletal age (p = 0.004). Other variables such as gender, ethnicity, fracture characteristics, mechanism of fracture, implant size and surgical technique did not differ significantly between the outcome groups.

3.3 Inter-observer correlation of modified Sauvegrain method score for TENS patients

Using Pearson correlation analysis, the inter-observer correlation coefficient between the two independent reviewers was determined to be 0.93 (p < 0.001).

3.4 Correlation between modified Sauvegrain method score and skeletal age in TENS patients by gender (Supplemental Table B.1)

The male population had a Pearson and Spearman correlation of 0.992 (p < 0.001) and 0.998 (p < 0.001) respectively between modified Sauvegrain method score and skeletal age; similarly, female population had a Pearson and Spearman correlation of 0.941 (p < 0.001) and 0.992 (p < 0.001) respectively. After pooling males and females, the Pearson and Spearman correlation were 0.825 (p < 0.001) and 0.759 (p < 0.001) respectively.

3.5 Difference between skeletal age and chronological age in TENS patients by gender (Table 2)

On average, male TENS patients had an older derived skeletal age compared to chronological age (13.8 versus 12.6 years) with a mean (SD) difference of 1.3 (1.03) years (p < 0.001). This was similar for the population of 255 male adolescents with a complete diaphyseal fracture (13.7 versus 12.5 years) with a mean (SD) difference of 1.2 (1.03) years (p < 0.001). In both of these groups, the females did not demonstrate a significant older skeletal age than chronological age.

3.6 3.6 Modified Sauvegrain method score versus chronological age as an independent predictor of poor treatment outcome in male TENS patients (Tables 3–4, Supplemental Table B.2 and Fig. 3)

In univariate logistic regression analysis, both modified Sauvegrain method score (OR 1.17, 95%CI 1.03–1.33, p = 0.019) and chronological age (OR 1.50, 95%CI 1.06–2.12, p = 0.023) were significant risk factors of poor treatment outcome in male TENS patients (Table 3). However, in ROC analysis (Table 3 and Fig. 3), modified Sauvegrain method score AUC was greater than that of chronological age (AUC 0.71, 95%CI 0.59–0.83 vs. AUC 0.67, 95%CI 0.55–0.79). The AUC difference between these two variables was not significant (p = 0.400). Similar analyses were performed for female TENS patients but due to a small sample size, reliable results were not attained.

Close

Fig. 3 Comparison of chronological age versus modified Sauvegrain method score in relation to poor outcome in male TENS patients. A. ROC curve: Chronological age, B. ROC curve: modified Sauvegrain method score, C. Histogram showing Youden cut-off point: Chronological age, D. Histogram showing Youden cut-off point: modified Sauvegrain method score Abbreviation: TENS, Titanium Elastic Nail System.

Export to PPT

The Youden J-index identified cut-offs of 26 for modified Sauvegrain method score and 13 years for chronological age in predicting high versus low-risk patients for poor treatment outcome in male TENS patients. The cut-off of 13 years for chronological age corresponds to 0.76 sensitivity, 0.52 specificity, 0.27 PPV and 0.91 NPV, whereas the modified Sauvegrain method score of 26 corresponds to 0.71 sensitivity, 0.63 specificity, 0.31 PPV and 0.91 NPV (Supplemental Table B.2).

Based on multivariate stepwise logistic regression analysis, modified Sauvegrain method score (OR 1.17, 95%CI 1.03–1.33, p = 0.019) was selected as the sole predictor of poor treatment outcome in male TENS patients over chronological age and other variables (Table 4). Thus, AUC for modified Sauvegrain method score as a univariate predictor of poor outcome remains at 0.71 (95%CI 0.59, 0.83).

4.1 Importance of accurate anatomical reduction

Achieving accurate anatomical reduction of fractures in the forearm is crucial and any residual angulation deformity post-treatment can result in impaired forearm rotation and consequently, its function.^9,15 In order to minimize angular deformity, surgical stabilisation of forearm fractures has been encouraged over closed reduction and cast immobilization.^7,15

4.2 Suitability of TENS in adolescents

Arguably, diaphyseal forearm fractures in adolescents can be better treated with TENS than plating. In our study, TENS patients categorised by treatment outcomes were well-matched with 82 % achieving a good outcome and 18 % poor outcome (23 of 128). TENS as a minimally invasive option reduces risk of surgical complications in adolescents while still ensuring acceptable fracture reduction.³² Whereas plating is more suitable for patients who are older and whose bones are wide enough to fit a plate.^9,20,22 However, the practice of correlating chronological age to skeletal suitability for TENS or plating is not always reliable. In our study, we have confirmed that skeletal age (represented by the modified Sauvegrain method score) was more reliable than chronological age in predicting the risk of poor treatment outcome in pubescent adolescents subjected to TENS treatment.

A retrospective study by Flynn et al. found that 30 % (17 of 56) of TENS adolescent patients had fair/poor outcomes (defined by forearm ROM and treatment complication), with delayed union being the most common treatment complication at six cases.⁷ Another retrospective study by Antabak et al. detected 25 % (22 of 88) of TENS adolescent patients afflicted with treatment complications among which rotational forearm restriction of 11–20° was the most common treatment complication at nine cases.³³ Similarly, the poor outcomes in our study were attributed to limited range of motion with 80 % (18 of 23) developing rotational forearm restriction of six to 10° (Refer to Supplemental Figure A.2 for a radiographic example of an affected study subject). Evidently, the reliance on chronological age to guide TENS treatment decision in these adolescents had imposed unnecessary risk of treatment failures. Therefore, this has prompted the exploration of skeletal age as an objective criterion for evaluating the suitability of TENS for adolescents.

4.3 Role of skeletal age

Variation of skeletal age from chronological age becomes more distinct during pubertal growth. Dimeglio et al. demonstrated that only 30 % of males and females investigated for the validation of the modified Sauvegrain method had identical skeletal and chronological age.²⁴ Male TENS patients in our study displayed older skeletal age than chronological age with an average difference of 1.3 years. However, the discrepancy was less apparent in female TENS patients likely due to small sample size. This trend of rapid skeletal maturation in children has been documented in multiple developed countries with increased environmental and dietary changes seen as contributing factors.³⁴ Hence, this supports skeletal age over chronological age as a better indication of the development status of a pubescent adolescent.

We demonstrated that the modified Sauvegrain method score was the sole objective predictor of poor treatment outcome in male TENS patients and a modified Sauvegrain method score of ≥26 was statistically predictive of high risk of poor treatment outcome in these male adolescents. This cut-off serves to minimize the poor outcome rate in male adolescents with a modified Sauvegrain method score <26 (91 % NPV) when considering TENS. Yet, due to potential long-term functional deficit in an adolescent when definitive treatment (i.e., rigid plate fixation) is readily available, more caution should be exercised in using TENS in male adolescents with a modified Sauvegrain method score ≥26 (31 % PPV). This result indicates that at or beyond modified Sauvegrain method score of 26, the forearm bones of adolescents start behaving similarly to that of an adult with no remodelling potential and risk of delayed healing. Although plating is not unusual in patients possessing younger skeletal maturity, decision for considering plating over TENS at modified Sauvegrain method scores ≥26 should take precedence.

The modified Sauvegrain method has a maximum score of 27. This suggests that the cut-off at or above 26 for prioritising plating for good outcomes reflects closure of majority of the physes at the elbow apart from either the olecranon apophysis or the proximal radial epiphysis at its second last stage of closure. This also implies that TENS may still be an acceptable surgical option for diaphyseal forearm fractures if there is more than one open physis at the elbow.

It is worth noting that the Youden J-Index based on the modified Sauvegrain method score cut-off at or above 26 (0.345) is greater than that for the chronological age cut-off at or above 13 years (0.284). This further supported our recommendation to rely on the modified Sauvegrain method score as an objective criterion over chronological age for evaluating the suitability of TENS for adolescents.

4.4 Study limitations and future research

The modified Sauvegrain method was selected over other skeletal age assessment tools primarily for its practicality and ease of use in the outpatient clinical setting by taking advantage of the readily available anteroposterior and lateral elbow radiographs in any paediatric forearm fracture evaluation. Digital skeletal age tools such as the Tanner-Whitehouse-III³⁵ and Greulich and Pyle³⁶ were considered but they were impractical since they were time-consuming and rely on hand-wrist radiographs which are not routinely assessed in an isolated forearm fracture. Notably, the time frame for the modified Sauvegrain method corresponds to the pubertal growth peak period but not to the subsequent pubertal growth decelerating period.^24,26 The distal radius and ulna (DRU) classification may present as a promising alternative or complementary skeletal maturity measure as it relies solely on wrist radiographs (distal radius and distal ulna) to cover the pubertal growth peak period and beyond till skeletal maturity, albeit at one-year intervals.^26,37 This contrasts with the six-monthly assessment intervals afforded by the modified Sauvegrain method.²⁴ Nevertheless, it would be valuable to explore the DRU classification in future studies to elucidate further details beyond the closure of the elbow physes to guide TENS treatment decisions in adolescents.

Several other limitations of this study ought to be acknowledged as well. This was a retrospective study with a small female adolescent population. Objective factors that may influence surgical decisions such as body weight or body mass index were not covered. Outcome of this study may not be applicable to female patients as they were incompatible with the regression analyses due to its small sample size. Although the overall study sample is heterogeneous with major ethnic groups included, there may still be potential inherent biases associated with each of the ethnic groups represented. Furthermore, the post-operative protocol for TENS patients involved interval outpatient reviews with clinical and radiological assessments where appropriate but was not influenced by skeletal age assessments. The extent of skeletal age progression after TENS may further reveal correlations with interval and eventual functional outcomes. For example, while the observed loss of forearm rotation in this study could inherently be attributed to post-operative loss of anatomical reduction since TENS provides only relative stability,¹⁸ it may be worth exploring the correlation between skeletal age and loss of reduction. Finally, it may be more beneficial to ascertain the value of skeletal age assessment in guiding operative treatment decisions by comparing TENS against plating for diaphyseal forearm fractures in a prospective longitudinal study.

4.5 Practical implications

In practice, using the modified Sauvegrain method to assess skeletal age is ideal as the elbow is routinely visualised on plain radiographs as part of the diagnostic evaluation of any paediatric forearm fracture. As such, no additional imaging is required. Compared to chronological age, the modified Sauvegrain method score can guide the treatment decision making process for diaphyseal forearm fractures in adolescents by providing objective assessment of skeletal maturity. A modified Sauvegrain method score of ≤25.5 would guide surgeons to employ TENS in adolescents, while a score of ≥26 would prompt surgeons to consider plating over TENS instead.