Introduction

It is well known that distal forearm fractures are the most common fractures in childhood, accounting for 74% of all fractures in the upper limb [1]. The incidence of these fractures has increased in recent years due to increased participation in different and competitive sports, rising rates of overweight children and adolescents, and decreased bone mineral density [2, 3]. Antero-posterior and lateral X-ray views of the injured bone remain the gold standard for the diagnosis of a fracture.

In the last years, there has been growing concern about children’s exposure to ionizing radiation and its potential adverse effects in the future [4]. A correct diagnosis of a fracture typically requires at least two radiographic examinations of the injured area (f + profile views). The number of X-rays may need to be increased if the clinical examination alone is insufficient, especially in young children [5, 6]. Furthermore, radiography has limitations in visualizing unossified and cartilaginous parts of bones in the pediatric population, and alternative diagnostic modalities like MRI or CT scans are either expensive, not always available, or entail additional radiation exposure [7,8,9]. It is also worth noting that only a small fraction of children who present in the emergency room due to accidental injuries will actually have a fracture [10].

The use of ultrasound for diagnosing fractures in children has gained significant attention over the past decade, emerging as a viable alternative [11,12,13]. Moreover, published reviews and studies on the diagnostic accuracy of ultrasound in detecting bone fractures have consistently demonstrated high levels of sensitivity and specificity [14,15,16,17]. Additionally, ultrasound offers several advantages, including real-time assessment at the bedside and the capability to examine multiple planes [18, 19].

As distal forearm injuries are the most common in children, this systematic review focuses exclusively on assessing the diagnostic accuracy of ultrasound for distal wrist fractures in children. By narrowing our scope in this way, we aim to avoid the variability that comes with reviewing diagnostic accuracy across different bone types and mixed populations. We will also collect and analyze information regarding the operator’s training, the ultrasound machine, the type of probe used, the age and gender of the patients, and the scanning method employed for the forearm. Additionally, we will discuss the advantages of using ultrasonography in pediatric orthopedic trauma cases.

Patients and methods

Study design–search strategy–data source

We conducted a systematic literature review to assess the diagnostic accuracy of ultrasound in detecting distal forearm fractures in children, comparing it with conventional X-ray. We utilized the electronic databases Pubmed and Cochrane CENTRAL for our search. This review followed the guidelines and recommendations outlined in PRISMA (Supplemental Material 1) [20]. The terms “ultrasound OR ultrasonography OR diagnostic imaging OR ultrasonics” AND “forearm fracture/injury OR radius fracture/injury OR wrist fracture/injury” AND “children OR child OR pediatric OR juvenile OR adolescent” were scanned. Full details of the search strategy are reported in Supplemental Material 2.

The research aimed to investigate whether ultrasound (index test) is both suitable and accurate in detecting fractures in children and adolescents when compared with X-ray (reference test). The focus was on the distal forearm, as it represents the most common location for fractures in this age group. The title and abstract of potential articles were reviewed for possible inclusion in our analysis. Subsequently, full-text reviews were conducted on articles deemed potentially suitable. Additionally, the bibliographies of the included studies and other systematic reviews were also screened. The initial study selection was carried out by a single author (I.D.). Full-text screening, final decision on inclusion/exclusion, data extraction, quality assessment, and risk of bias assessment were conducted by two authors (I.D. and V.B.). The study protocol was not registered.

Our inclusion and exclusion criteria were as follows:

  1. a)

    Studies published between January 2010 to August 2023 (to focus on current trends and enhance result reliability).

  2. b)

    Studies published exclusively in English or German.

  3. c)

    Studies comparing the diagnostic accuracy of ultrasound (index test) with X-ray (reference test), for distal forearm fractures in children/adolescents. Studies using ultrasound for purposes other than fracture diagnosis (e.g., reduction, treatment without X-ray) were excluded.

  4. d)

    Studies comparing ultrasound with X-ray in both children and adults were included only if over 50% of the patients were under 18 years old.

  5. e)

    Studies comparing ultrasound with X-ray in multiple long bones in children were included only if they reported the number of forearm fractures detected by both ultrasound and X-ray.

  6. f)

    Studies with fewer than 5 patients and case reports were excluded.

  7. g)

    Editorials, comments, and studies not conducted on humans were excluded.

  8. h)

    Studies sharing the same population in different journals were excluded.

Data extraction–quality assessment of studies–risk of bias assessment

Data extraction, quality assessment, and risk of bias assessment were conducted by two authors (I.D. and V.B.). From each included study, we gathered the following data: (1) year of publication and the country where the study was conducted, (2) study design, (3) specialty of the sonographer (e.g., emergency doctor, pediatrician, trauma surgeon), (4) sonographer’s training level (untrained, trained, type of training), (5) ultrasound machine used, (6) type of transducer and frequency used, (7) protocol for scanning the forearm (e.g., 6-view, 4-view, longitudinal, transverse planes), (8) sample size, (9) number of ultrasound scans performed, (10) age of the patients, (11) male-to-female ratio, (12) total number of fractures identified, and (13) fractures detected via ultrasound and X-ray (true positive, false positive, true negative, and false negative).

To evaluate the methodological quality of the included studies, we employed the QUADAS-2 tool, designed specifically for assessing the quality of diagnostic accuracy studies [21]. This tool comprises four domains: patient selection, index test, reference standard, and flow and timing. Each domain is evaluated to determine the risk of bias. When it comes to applicability, only the first three domains are considered. For assessing the risk of bias, judgements are categorized as “low,” “high,” or “unclear” based on the responses “yes,” “no,” or “unclear” to the signaling questions (Supplemental Material 3) [21].

Effect measures and synthesis methods

In order to examine the accuracy of ultrasound for the detection of fractures in the distal forearm in children, the authors extracted data in the form of frequencies regarding the four vectors of true positive (TP), true negative (TN), false positive (FP), and false negative (FN) and recorded outcomes in a 2 × 2 table. We thereafter calculated the sensitivity and specificity as TP/(TP + FN) and TN/(FN + FP) and pooled the effect estimates utilizing the logistic-normal bivariate random-effects meta-analysis model (BRMA) and presented outcomes as percentages with their corresponding 95% confidence intervals (CIs) [22]. A formal test for heterogeneity utilizing the I2 statistic by Zhou and Dendukuri was implemented. Outcomes were visualized through forest plots and summary receiver operating characteristics (SROC) plots [23].

Pooled diagnostic odds ratios, defined as the ratio of the odds of a positive test result in subjects with a fracture to the odds of a positive test result in subjects without a fracture, pooled positive likelihood ratios, calculated as the probability of a positive test result in individuals with a fracture divided by the probability of a positive test result in individuals without a fracture, and pooled negative likelihood ratios, computed as the probability of a negative test result in individuals with a fracture divided by the probability of a negative test result in individuals without a fracture, were determined using the sampling-based approach proposed by Zwinderman and Bossuyt [24]. The results were presented as the mean with a 95% CI.

We conducted all statistical analyses using R software (version 4.2.0, R Foundation for Statistical Computing, Vienna, Austria) and STATA statistical software (version 18, Stata Corp LP, College Station, Texas).

Results

Literature search

The flow diagram of our study is illustrated in Fig. 1. Initially, 724 studies were identified. After removing duplicates, 657 studies were screened for eligibility. Among the potentially suitable studies, there were 66. Following a thorough review and screening process, 16 studies were included. An additional study was incorporated based on references found within the included studies [32]. In total, our systematic review encompassed 17 studies [25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41]. One study was excluded due to insufficient data, making it impossible to construct a 2 × 2 table [42].

Fig. 1
figure 1

Flow diagram of the included and excluded studies

Full size image

Characteristics of the included studies

Within the studies incorporated into our review, a total of 2003 patients presented to the emergency department with potential (distal) forearm fractures. In cases where multiple different long bones were scanned (such as the fibula, tibia, humerus, hand, feet, skull), we considered only the number of patients with distal forearm injuries [25, 29]. In total, 2546 ultrasound scans were conducted. Some studies performed separate ultrasound scans for the radius and ulna [26, 30, 31, 39], while two studies also conducted ultrasound scans on the uninjured forearm [33, 41]. Our systematic review identified 1203 cases of distal forearm fractures, confirmed using X-ray as the gold standard (reference index).

Out of seventeen studies included in our analysis, sixteen are prospective studies, and one is a cross-sectional study [36]. Five studies were conducted in North America (three in the USA and two in Canada) [25, 28, 33, 36, 41], eight in Europe [26, 27, 30, 31, 34, 37,38,39], two in Asia [29, 32], and two in Australia [35, 40]. All of these studies were focused on children with a mean age of 9.8 years (ranging from 2 to 16 years). Only two studies included adults older than 18 years: the first study involved a sub-group of skeletally mature patients aged 18 to 24 years, and another sub-group, where the patients were under 18 years [25]. The second study included patients up to 55 years old, with over 70% of them being under 18 years [32]. Male patients constituted 56% of the study population, while female patients accounted for 44%. Detailed characteristics of each study are summarized in Table 1 and 2.

Table 1 Characteristics of included studies in our systematic review
Full size table
Table 2 More characteristics of included studies. Table 2 focuses on true positive, false positive, true negative, and false negative results
Full size table

In all the studies, medium to high-frequency linear transducers (ranging from 7.5 to 15 MHz) were employed, regardless of the varying ultrasound machine types. Furthermore, nearly all studies adhered to a standardized scanning protocol, commonly referred to as the “6-views” of the distal forearm. This protocol includes dorsal, lateral, and palmar views of the radius, along with dorsal, medial, and palmar views of the ulna, as originally described by Ackermann et al. [26]. Only 3 studies did not specify their ultrasound scanning method [27, 29].

Quality assessment of the included studies (QUADAS-2)

We assessed the quality of the included studies using the QUADAS-2 tool, and the results are presented in Fig. 2. There was a moderate to low risk of bias and applicability concerns regarding patient selection, index test, reference test, and flow and timing. It is worth noting that many of the studies were convenience studies, which implies a potential bias in patient selection [25, 28, 29, 35, 36, 38,39,40,41]. All patients enrolled in the studies would normally undergo X-ray evaluation for diagnosing distal forearm fractures. With the exception of one study that used either X-ray or computed tomography as the reference standard [25], all other studies relied on conventional X-ray as the reference standard.

Fig. 2
figure 2

Quality assessment of the included studies–QUADAS-2

Full size image

In almost all studies, the sonographers and the radiologists were blinded to each other results. However, in three studies, it remained unclear whether the sonographer was blinded to the X-ray findings, indicating a potential risk of bias [27, 31, 33]. Both examinations were typically performed on the same day or within a short period, typically 2–3 days. Most ultrasound scans were carried out by emergency physicians or pediatric physicians. In three studies, it was unclear who performed the ultrasound examination [26, 27, 30], in one study it was performed by sports medicine doctors [33], in one study (which was finally excluded due to insufficient data, making it impossible to construct 2 × 2 table) by radiologists [42], in another one by nurses [40], and one study had medical students performing the examinations [41]. The level of training across most studies was relatively modest, typically consisting of 30-min to 1 or 2-h lecture training combined with hands-on training (see Table 1). However, in three studies, details regarding the sonographer’s training remained unclear [26, 27, 30].

Among the seventeen studies, fifteen defined ultrasound findings indicative of a fracture, which typically included (i) the presence of any cortical disruption, gap, or displacement and (ii) the presence of any cortical irregularities such as torus formation or bulging. However, 2 studies clearly mentioned that the presence of (subperiosteal) hematoma in most ultrasound views was also considered as an indicator of a fracture [28, 31].

Overall analysis of the diagnostic accuracy of ultrasound for distal forearm fractures in children

In our systematic review, we analyzed 17 studies that compared the diagnostic accuracy of ultrasound in detecting distal forearm fractures in children with X-ray. Figure 3 illustrates the forest plots for sensitivity and specificity. The pooled sensitivity and specificity were 0.96 (95% CI: 0.93–0.98) and 0.96 (95% CI: 0.89–0.98), respectively. Figure 4 presents the HSROC curve, demonstrating the high accuracy of the ultrasound in detecting distal forearm fractures in children when compared to X-ray. The pooled positive likelihood ratio was 13.40 (95% CI: 7.97–21.50), the pooled negative likelihood ratio was 0.06 (95% CI: 0.04–0.1) and the pooled diagnostic odds ratio was 209 (95% CI: 92.20–412.00). Since our systematic review focused on (distal) forearm fractures in children, we expected low heterogeneity. Indeed, we calculated I2 values to assess heterogeneity among the included studies, and they were found to be 38.38% for sensitivity and 39.22% for specificity, indicating low to moderate heterogeneity. The observed heterogeneity can be attributed to various factors. Two studies compared ultrasound to X-ray in pediatric patients, encompassing multiple injury sites [25, 29]. Another study focused on evaluating the efficacy of ultrasound in diagnosing distal radius fractures across all age groups [32]. Additionally, it is well established that ultrasound, as an imaging modality, is highly operator-dependent.

Fig. 3
figure 3

Pooled sensitivity and specificity of ultrasound compared with the reference standard (X-ray), for the diagnosis of distal forearm fractures in children

Full size image
Fig. 4
figure 4

HSROC curve of ultrasound for the diagnosis of distal forearm fractures in children

Full size image

Discussion

This study systematically reviewed the diagnostic accuracy of ultrasound for detecting distal forearm fractures in children. The findings clearly indicate that ultrasound scans of the distal forearm in children following acute trauma can safely serve as the primary diagnostic method.

Advantages of ultrasound in pediatric orthopedic trauma cases

As previously mentioned, there is a growing concern regarding radiation exposure in young children during X-ray examinations. Conducting a clinical examination in young children after trauma is often challenging, and pinpointing the exact location of maximum pain can be particularly difficult [43]. The standard approach involves a minimum of two radiological examinations (f + profile views of the injured area), and in cases of clinical ambiguity, additional bones may need radiographic evaluation [6, 43]. The use of ultrasound as a viable alternative has been proposed in the literature. Ultrasonography offers advantages such as speed, real-time information, bedside applicability, the ability to scan the contralateral limb, and, notably, it is radiation-free [19, 44].

In addition, ultrasound imaging has been found to be less painful when compared to X-rays [28]. The need to reposition the injured limb for X-ray imaging often causes more discomfort for the patient [13]. Furthermore, sonography provides valuable insights into soft-tissue injuries and the condition of unossified bones [7, 45]. One study conducted a cost analysis comparing ultrasound and X-ray [46]. It found that fracture treatment involving ultrasound costs €20.54, whereas treatment involving an X-ray examination costs €26.60, thus concluding that ultrasound usage can result in cost savings [46].

Above all, it is crucial to minimize unnecessary radiation exposure. Child tissues are more radio-sensitive due to their increased mitotic rate, and children have a long life expectancy during which malignancies can develop following irradiation [35, 47, 48]. Ackermann et al. conducted a study on the efficacy of the Wrist SAFE Algorithm in fracture sonography of distal forearm fractures in children and found that 81% of X-rays could be avoided, which translates to a significant reduction of 2.8 million X-ray exposures in G10 member states [49].

Literature regarding the diagnostic accuracy of ultrasound in detecting fractures

Our findings align with previously published reviews. In a meta-analysis by Tsou PY et al., which assessed the diagnostic accuracy of ultrasound for upper extremity fractures in children, a sensitivity of 0.95 (95% CI: 0.93–0.97) and a specificity of 0.95 (95% CI: 0.91–0.98) were reported [14]. It should be mentioned that this meta-analysis encompassed studies involving elbow, humerus, and forearm fractures, not limited to forearm injuries alone [14]. Another systematic review of Schmid GL et al. compared ultrasound with conventional imaging and reported a pooled sensitivity of 0.91 (95% CI: 0.90; 0.92) and specificity of 0.94 (95% CI: 0.93; 0.95) [15]. However, the analyzed studies displayed marked heterogeneity, including adults and various bone types [15]. In a literature review focusing on ultrasound as a triage tool for diagnosing fractures in children (clavicle, metacarpals, metatarsals, forearm, etc.), ultrasound demonstrated a sensitivity of nearly 100% when combined with clinical suspicion of fracture [16].

Douma-den Hamer et al. published a diagnostic meta-analysis on ultrasound for distal forearm fractures, encompassing all age groups [50]. Their findings revealed a sensitivity of 97% (CI 95%, 93–99%) and a specificity of 95% (CI 95%, 89–98%) [50]. In a recent large, prospective, multi-center study focused on the accuracy of ultrasound in detecting fractures in children, high-skill providers (with 18 months of training and at least 200 examinations) demonstrated a sensitivity of 91.67% (95% CI, 76.41–97.82%) and a specificity of 88.89% (95% CI, 73.00–96.34%) [17]. Standard-skill providers (with 6 months training and 50 examinations) exhibited a sensitivity of 71.50% (95% CI, 64.75–77.43%) and a specificity of 82.91% (95% CI, 77.82–87.06%) [17]. It is worth noting that this study introduced heterogeneity to the results, as it examined multiple bone types, including long bones, hands, feet, and the skull [17]. Similarly, another systematic review included heterogeneous studies and concluded that their results do not support replacing radiography with ultrasound for diagnosing extremity fractures (encompassing all extremity fractures) [51]. However, it did acknowledge that ultrasound can be a valid alternative under specific conditions [51].

The previously mentioned reviews are characterized by the inclusion of studies that compared ultrasound’s diagnostic accuracy either across multiple (long) bones within similar age groups or within a specific “injured area” but across different age groups, and sometimes both. In our study, our aim was to minimize heterogeneity as much as possible. Given that distal forearm fractures are the most prevalent in childhood, we sought to include studies that specifically compared the diagnostic accuracy of ultrasound for fractures within this “area” and within this age group (0–16 years old). Under these criteria, we found only one review in the literature, which included just 8 studies [46]. It reported a sensitivity ranging from 64 to 100% and a specificity between 73 and 100%, suggesting indications of comparable diagnostic accuracy between X-ray and ultrasound. However, it emphasized the need for more studies with adequate sample size calculation [46]. Conversely, a review addressing ultrasound in diagnosing pediatric distal radius fractures concluded that ultrasound cannot adequately replace conventional X-rays [52]. It is important to note that this review included only a few studies, lacked statistical analysis, and primarily provided an orthopedic perspective on treatment policy [52].

Methods of scanning–definition of fracture–training

While most studies employ a consistent ultrasound examination protocol for the distal forearm, some variations exist (see Table 1). The six-view method, as described by Ackermann et al., appears to be the most accurate approach [26]. This method involves examining the radius and ulna from the dorsal, lateral, and palmar sides, resulting in a total of six images (2 bones × 3 scans = 6 images) [26]. Some studies focus on scanning the radius only (which increases the risk of overlooking an isolated ulna fracture), and some other studies do not describe the exact method of ultrasound examination or perform the 4-view method instead of the six-view method [27, 34, 41, 53]

Furthermore, there is no unanimous consensus regarding the ultrasound definition of a fracture. While most studies agree that the presence of cortical irregularity or disruption on ultrasound indicates a fracture, some studies also consider the presence of subperiosteal hematoma [28, 31]. Additionally, recent research has explored specific ultrasound signs. For example, a recent study assessed the ultrasound pronator quadratus hematoma sign and found that it effectively distinguished children with and without cortical breach fractures [54]. Another recent study, which evaluated the diagnostic accuracy of ultrasound in children diagnosed with a distal radius fracture on X-ray, measured the fracture-physis distance and suggested that a cutoff of 1 cm differentiated Salter-Harris II fractures from other cortical breach fractures, though not buckle fractures [55].

In nearly all studies, sonographer training comprises a combination of a lecture-based training and hands-on practice. This training duration varies widely, from as short as a 4-min training video [36] to 30-min sessions [31], 1-h sessions [37, 38], or an 80-min online learning module with quiz [35]. These variations indicate minimal emphasis on theoretical training. Hands-on training is similarly minimal, ranging from 30-min practical training [32] to 2-h hands-on training sessions [35] or the requirement to perform at least 20–25 supervised scans [29, 39]. A meta-analysis revealed no significant difference in diagnostic accuracy between trained and untrained personnel, suggesting that ultrasound can be effectively conducted after minimal training [50]. Conversely, in a prospective, multi-center study on the accuracy of ultrasound in detecting fractures in children, high-skill providers (with 18 months training period and at least 200 examinations) demonstrated higher sensitivity and specificity, while standard-skill providers (with 6 months training period and 50 examinations) exhibited lower sensitivity [17]. A recent study described a learning curve for diagnosing pediatric distal forearm fractures using ultrasound [56]. This concluded that a brief training course, coupled with approximately 15 scans, was sufficient [56].

The variations in results among different studies can be attributed to several factors: (a) differences in the definition of fracture, (b) variances in the ultrasound scanning method (whether the six-view or 4-view method was employed), and (c) disparities in sonographer training.

Nevertheless, it is imperative to establish a universal consensus regarding the definition of a fracture based on ultrasound findings, an agreed-upon scanning protocol, and a standardized sonographer training. Future studies should aim to develop globally accepted protocols and guidelines encompassing the aforementioned aspects: fracture definition, scanning methods, and sonographer training. Furthermore, these protocols should undergo critical evaluation for their applicability to different joints or bones. For instance, the complex anatomy of the elbow joint and the variation in ossification centers in children may necessitate distinct ultrasound scanning protocols and methods [14].

It is important to note that the ultimate decision regarding a child’s treatment is contingent upon a combination of clinical assessment and the findings from a diagnostic modality (such as X-ray, ultrasound, or computed tomography). This means that the sonographer may not have the final decision on patients’ treatment. Our study investigates the possibility that the ultrasound may replace the X-ray (under certain circumstances) and should not replace the final judgement of the treating physician.

Strengths and limitations

To the best of our knowledge, our systematic review encompasses the most extensive collection of studies dedicated to assessing the diagnostic accuracy of ultrasound for distal forearm injuries in the pediatric population, totaling 17 studies. Our intention was to minimize heterogeneity in our results to the greatest extent possible, a goal that is evident in Fig. 2 (QUADAS-2-quality assessment of included studies). Our review’s strengths lie in the substantial number of included studies and our efforts to mitigate heterogeneity in the results.

One limitation of our review (which can paradoxically be regarded as a strength) is the exclusion of certain studies that examined the distal forearm with ultrasound across all age groups (including adults) or as part of assessing the diagnostic accuracy of ultrasound in various anatomical regions (such as tibia, fibula, humerus, metacarpals, metatarsals, skull, ribs). Among the seventeen studies, two studies compared ultrasound to X-ray in pediatric patients across multiple injury sites, but they were included in our systematic review because they clearly reported the number of fractures detected by ultrasound and X-ray in the radius and ulna [25, 29]. Additionally, one study assessed the efficacy of ultrasound in diagnosing distal radius fractures across all age groups but was included in our study due to the substantial proportion (72%) of patients under 18 years [32]. That being said, we excluded studies that evaluated the ultrasound’s ability to detect fractures when they did not precisely specify the number of distal forearm fractures found or when the study population consisted of over 50% adults [57,58,59]. This decision may have resulted in the omission of some pediatric distal forearm fractures and introduced a selection bias. However, it likely had a positive impact on the overall quality of our meta-analysis.

Another limitation of our study pertains to the varying levels of experience and training among sonographers, disparities in the definition of a fracture using ultrasound, and differences in scanning methods, all of which have been thoroughly discussed previously. Consequently, the widespread applicability of ultrasound for detecting distal forearm fractures cannot be universally assumed until a consensus is achieved in these areas.

Another limitation to consider is that our study did not investigate the efficacy of ultrasound in diagnosing fractures in other bones or joints. As a result, the findings of our study are specifically relevant to distal forearm injuries in children and adolescents.

Furthermore, we have to acknowledge that all studies included in our analysis focused on patients who arrived at the emergency department with a reported history of distal forearm injury. This introduces the possibility of selection bias since only those patients with a conventional history of injury were included across all studies.

Finally, our exclusion criteria encompassed studies in which patients did not undergo both ultrasound and conventional X-ray examinations, and we restricted our analysis to studies available in English or German. As a result, there is the potential for the omission of relevant studies that did not meet these criteria.

Although full-text screening, final decisions regarding inclusion/exclusion, and quality assessment were conducted collaboratively by two authors (I.D. and V.B.), the initial study selection was performed by a single author (I.D.). The initial screening may have led to some studies being overlooked.

Conclusions

Our systematic review concludes that ultrasonography can safely contribute to the detection of distal forearm fractures in children and adolescents. While it is not intended to entirely replace conventional X-ray examinations, ultrasonography holds the potential to effectively rule in or rule out a fracture. This capability can significantly reduce unnecessary radiation exposure for a considerable number of children who present in the emergency department with potential distal forearm fracture.

Future studies should focus on two critical objectives: (a) establishing a universally accepted protocol for training and scanning methods to eliminate disparities and (b) evaluating the long-term functional outcomes and arm function of children who were treated based on ultrasound findings alone.