Scaphoid fractures are one of the most common adult carpal bone fractures, but are less common in the skeletally immature population, representing 2.9% of hand fractures and less than 1% of all fractures [1–3]. However, with increasing youth sports participation, and increasing participation in extreme sports, the incidence of scaphoid fracture is rising in the pediatric population. There has been some controversy regarding the most common scaphoid fracture pattern in the skeletally immature; historically, distal pole fractures appear to be more common in the younger ages [2,4,5], and there is support for this pattern within the theory of the scaphoid’s predictable ossification pattern [6,7]. However, more recent publications demonstrate an escalating prevalence of scaphoid waist fractures in the skeletally immature, similar to the prevalence in the adult population [1,4,8].
The majority of pediatric scaphoid fractures are treated with immobilization with high rates of union in most fracture patterns (Fig. 1) [1,7,9–11]. There has not been any consensus on length of immobilization nor the type of immobilization within the literature [1–4,7–17]. Gholson et al.  reported longer time to union in proximal pole fractures, displaced fractures, chronic fractures, and fractures with evidence of avascular necrosis (AVN). However, there is a paucity of literature regarding fracture characteristics that risk conversion from conservative management to surgical management. Therefore, the purpose of this study was to evaluate the outcomes of management for all fracture types, in pediatric scaphoid fractures. The primary aim was to assess the risk of conversion to surgical intervention in children initially managed with casting for an isolated scaphoid fracture. The secondary aim was to assess the incidence and management of occult scaphoid fractures without radiographic signs of fracture on initial plain radiographs. In combination, the results of these two aims should provide both an understanding of the importance of immobilization of presumed fracture, as well as the index findings that predict failure of immobilization.
A retrospective review of patients aged less than 18 years old, who presented to our institution between 2009 and 2017, and underwent initial conservative treatment for a suspected isolated scaphoid fracture was performed, after Institutional Review Board approval. A total of 905 scaphoid fractures were identified. The inclusion criteria for the first aim of the study were radiographic findings of a scaphoid fracture in typically developing patients, under the age of 18 years, treated by our orthopedic department with a cast as their initial treatment. All notes had to be available in our electronic medical record, and patients had to be compliant with cast treatment. For the cohort who met inclusion criteria, we collected the presenting information, treatment details, and final radiographic outcomes. Exclusion criteria are listed in Table 1, and 521 of 905 scaphoid fractures identified were excluded.
|Treated but no evidence of fracture
|Not treated by ortho
|Initial treatment was surgery
|Family requested surgery
The presenting data included: patient demographics, mechanism of injury, and time from injury to presentation. The patients’ presenting clinical exam and location of tenderness (particularly the presence of snuffbox tenderness) at the initial visit were recorded. Similar to Gholson et al. , we defined a delay in presentation as at least 6 weeks from the date of injury to the date of treatment. All subjects included in this study were treated with a thumb spica cast; the type of thumb spica cast (above or below the elbow) was recorded and was based on the preference of the provider. Patients who failed nonoperative management were defined as conversion to surgical management. These conversions were due to unacceptable displacement or malalignment, lack of evidence of radiographic healing or signs of AVN on follow-up radiographs. Healing was defined as any signs of callus formation or bridging, whereas AVN was defined as relatively increased sclerosis of the more proximal fracture fragment. Conversion to surgical intervention was recorded for analysis.
The initial presenting plain radiographic information regarding fracture pattern, location of fracture, and displacement was collected. Displacement was defined as any displacement at the fracture site and included in our analysis as ‘displaced’ or ‘non displaced’. In regards to the primary aim of our study, the follow-up radiographs were evaluated for union status and signs of AVN. For the second aim of the study, those patients with normal appearing index radiographs but with positive snuffbox tenderness on exam were documented. These patients were treated with a thumb spica cast, and a radiograph was taken at the time the cast was removed. Children with callus formation or evidence of a scaphoid fracture on the follow-up radiograph were tabulated and included for the primary aim of the study (Fig. 2). Those with no evidence of fracture (callus/healing) at the follow-up radiograph were dropped from the study’s primary aim. However, this subset of children without radiographic evidence of fracture were included as a denominator for the secondary aim of the study and the evaluation of occult scaphoid fracture risk.
Basic descriptive statistics are presented. The wrist was used as the unit of analysis. Due to the difference in the proportion of subjects that converted to surgery compared with those that did not, comparisons among those two cohorts were analyzed with the Mann–Whitney U test. Categorical data were evaluated with Pearson’s Chi-square or Fisher’s exact test. Statistical significance was defined as P < 0.05. No a priori power analysis was performed. Statistical analysis was performed using SPSS (version 27; IBM, New York, New York, USA).
Three hundred and eighty-four scaphoid fractures (381 patients) were included in the first aim of this study. The most common mechanism of injury was a fall onto an outstretched hand (91.9%), which most commonly occurred from either a skateboard (n = 82), bicycle (n = 35), basketball (n = 30), or football (n = 28). Mechanism of injury information can be found in Table 2, in which we consolidated all ‘fall on outstretched hand’ injuries into one group, regardless of the activity they were engaged in when they fell, the remaining activities listed in Table 2 were activities that the patients were engaged in when they acquired their injury, but these did not occur from a fall on an outstretched hand. The majority of our cohort was male (85%), with 100% of the operative subjects being male. Additional cohort characteristics can be found in Table 3.
Table 2. -
Mechanism of injury
|Fall on outstretched hand
|Fell on by another person
|Motor vehicle collision
|Tug of war
MOI, mechanism of injury.
Table 3. -
|Delay in presentation (≥6 weeks)
|Long-arm thumb spica
|Short-arm thumb spica
|Avascular necrosis (AVN) at presentation
|Union status at presentation
|Non- or delayed-union
|Skin wound from cast
There were a total of 21 (5.5%) scaphoid fractures that failed nonoperative management and subsequently went on to surgery. None of the surgically treated scaphoid fractures in this study required a revision procedure. Children that went on to require surgery were older at injury (15.9 ± 1.8 years old compared with 14.0 ± 2.2 years old; P < 0.001) and older at initial treatment (16.1 ± 1.9 years compared with 14.1 ± 2.2 years; P < 0.001). Subjects that went on to require surgery took an average of 4× longer to seek treatment than those that did not require surgery (96.3 ± 148.8 days compared with 20.6 ± 62.1 days; P = 0.003) (Table 4). Children with a delay in presentation (n = 40) were 8.0 times more likely to require surgical treatment [95% confidence interval (CI), 3.1–20.6] than those that did not have a delay in treatment (P < 0.001). We also observed a relationship between a delay in presentation and union status at presentation. Children with a delay in presentation were 37.6 times more likely to have a delay in union, or nonunion (95% CI, 16.7–84.9), than those with no delay in presentation (P < 0.001).
Table 4. -
||Mean ± SD
|Age at injury (years) (P < 0.001)
||14.0 ± 2.2
||15.9 ± 1.8
|Age at presentation (years) (P < 0.001)
||14.1 ± 2.2
||16.1 ± 1.9
|Days from injury to presentation (P = 0.003)
||20.6 ± 62.1
||96.3 ± 148.8
|Days in cast (P = 0.248)
||53.5 ± 25.4
||59.0 ± 27.5
|Follow-up from initial presentation (months) (P < 0.001)
||2.8 ± 3.0
||12.5 ± 7.6
aDays in cast prior to surgery.
We found no significant difference in time in cast among subjects treated with a cast only and those that went on to require surgery (P = 0.248), with the mean time in cast for the whole cohort being 53.8 ± 25.5 days. We found that those managed with a long-arm thumb spica were at slightly increased odds of requiring surgery than those treated with a short-arm thumb spica [odds ratio, 3.5 (95% CI, 1.4–8.7); P = 0.007].
Subjects with displaced fractures were 9.3 times (95% CI, 2.6–33.1) more likely to require surgery than subjects presenting with nondisplaced fractures (P = 0.003). We also observed a difference in the proportion of subjects that required surgery based on fracture location (P < 0.001), with distal pole fractures requiring surgery 0.7% of the time, waist fractures requiring surgery 7.4% of the time, and proximal pole fractures requiring surgery 23.5% of the time. Fracture location was related to age with distal pole fractures occurring at a mean age of 13.3 ± 2.1 years, waist fractures occurring at a mean age of 14.6 ± 2.1 years, and proximal pole fractures occurring at a mean age of 15.6 ± 2.0 years (P < 0.001). Subjects presenting with AVN were 11.9 times (95% CI, 2.6–53.9) more likely to require surgery than subjects that did not present with AVN (P = 0.007).
Age at injury, delay in treatment, location of the fracture, union status at presentation, AVN at presentation, displacement, and cast type were included in a binary logistic regression model to determine predictors of conservative treatment failure. Our model indicated that the only significant predictor of conservative treatment failure was a delay in presentation [Exp(B) = 0.233 (95% CI, 0.061–0.888) (P = 0.033)].
Seven hundred and twenty-eight children with reported snuffbox point tenderness were treated with a cast. Of these, 400 had no radiographic evidence of a scaphoid fracture at their initial visit. On follow-up plain radiographs, 56 of these subsequently demonstrated evidence of a healing scaphoid fracture. Therefore, regarding our second aim, the relative risk of having a scaphoid fracture with only snuffbox point tenderness was 14%. In contrast, all children that were thought to have radiographic evidence of a scaphoid fracture at the initial visit continued to have radiographic evidence of a fracture at their follow-up X-ray (there were no false-positive radiographs at the index assessment). None of the children treated via immobilization with snuffbox tenderness in the absence of radiographic evidence of fracture, at initial visit, but with radiographic evidence of fracture at a subsequent visit required surgical intervention. Of note, two children who presented to clinic with only snuffbox tenderness (and no radiographic evidence of a fracture) chose to forego cast treatment despite recommendations. These children subsequently returned to clinic approximately 2 months later with continued snuffbox pain. One patient had an X-ray positive for a scaphoid fracture, and cast treatment was initiated. This patient ultimately failed cast treatment and went on to require surgical intervention (Fig. 3). For the second patient, advanced imaging was obtained, which demonstrated a scaphoid fracture nonunion. This patient was placed in a splint and scheduled for surgery.
The goal of scaphoid fracture management is to achieve bony union in order to restore the native carpal alignment essential for painless motion at the wrist. Prior literature in the adult population suggests that the natural history of a scaphoid nonunion can lead to pain and dysfunction secondary to the development of dorsal intercalated segmental instability and scaphoid nonunion advanced collapse [7,18]. The majority of pediatric scaphoid fractures are amenable to cast immobilization when initiated early after injury [1,7]. However, there are specific conditions when surgical intervention is more frequently recommended such as proximal fractures  and displaced fractures [1,7,11,13]. Our study looked at the potential predictors for failure of conservative management in those pediatric scaphoid fractures treated initially by casting, as well as the incidence of fracture in the setting of snuffbox tenderness but with normal initial radiographs.
Distal pole fractures have historically been considered to be the more common location of pediatric scaphoid fractures. But, Gholson et al.  found that waist fractures occurred in 71% of scaphoid fractures compared with distal pole fractures that occurred at a rate of 23%. The findings in the current study support more recent literature [1,8] that waist fractures (56.5%) were more common than distal pole fractures (39.1%). However, Nguyen et al.  determined that the location of the fracture was highly dependent on age at the time of injury, where the younger the patient, the more distal the fracture. This is also supported by the current study with the distal pole fractures occurring in younger patients (13.3 ± 2.1 years) than both the waist (14.6 ± 2.1 years) and the proximal pole (15.6 ± 2.0 years) fractures.
Bae et al.  demonstrated that in acutely treated pediatric scaphoid fractures, there was no difference in functional outcomes (disability of the arm, shoulder, and hand inventory and Modified Mayo Wrist Score and 100% union rate between casting and surgery). Thus, suggesting that surgery does not foretell a worse outcome but instead that union is key to successful clinical outcome. Therefore, the ability to achieve union becomes the primary outcome measure for most studies on the subject.
In order to reduce the risk of a scaphoid nonunion, physicians must remain diligent to appropriately manage patients with acute wrist injuries. The diagnosis of a scaphoid fracture is difficult in part due to the challenges with interpreting radiographs, particularly in subjects younger than 8 years old , and the variation in standard radiographic carpal measurements during development up to age 12 years old . It has been reported that plain radiographs miss a scaphoid fracture up to 6% of the time in the general population , but in the skeletally immature population, the diagnosis is missed 12–30% on initial presenting radiographs [12,20,21]. In the current study, 14% of patients with snuffbox point tenderness but negative initial radiographs demonstrated evidence of a healing scaphoid fracture upon follow-up radiograph. Therefore, we recommend that children presenting to clinic with a history of a fall on an outstretched hand, and a clinical exam finding of snuffbox tenderness, should be treated as a potential scaphoid fracture, even in the absence of radiographic evidence of a scaphoid fracture.
When treated without delay via cast immobilization, all these fractures previously not seen on plain radiographs in our present study went on to union without complication. However, we noted that two subjects who declined cast treatment at the first visit subsequently went on to require surgical intervention to achieve union. Moreover, in the current study, patients who presented greater than 6 weeks from injury to initial treatment were eight times more likely to require surgery. A delay in presentation was related to a 68% rate of delayed-/nonunion of the scaphoid compared with only 5% in children with no delay in presentation. Therefore, healthcare professionals managing patients with presumed wrist sprains, negative plain radiographs, but tenderness at the snuffbox must remain attentive with close radiographic follow-up or pursue advanced imaging.
Historically, case series of delayed union or nonunion of scaphoid in the skeletally immature suggested prolonged immobilization will eventually achieve union [22–24]. However, there is a paucity of literature to guide the length of immobilization once a delayed union or nonunion has been diagnosed, and such prolonged casting may not be tolerable to the patient or family . With improving surgical technique, there has been successful healing following nonunion surgery consisting of open reduction with or without grafting [11,19,24,25]. Jauregui et al.  demonstrated that both grafted and nongrafted scaphoid nonunions after rigid internal fixation achieved 95% union rate in both groups demonstrating significant improvement in grip strength and wrist range of motion. Gholson et al.  found that 82% of the chronic fractures that were initially treated with casting failed and were 29.7 times less likely to achieve union with only casting. While surgery for scaphoid nonunion yields a high rate of achieving union, Bae et al.  demonstrated that chronic fractures and AVN were independent risk factors for worse outcomes in those patients with persistent functional limitations. Identifying scaphoid fractures early and initiating treatment seems to be the best predictor for successful outcomes.
There were a number of limitations to this study. The determination for surgery was at the discretion of the treating surgeon and family. As stated previously, there is no standard length of immobilization recommended before recommending surgery. In addition, due to timing around sports participation, and family preference, the length of immobilization and time to surgery was inconsistent. We found a higher proportion of surgical intervention in children treated with a long-arm thumb spica compared with a short-arm thumb spica, upon further review we discovered that the primary treating surgeon does have a bias towards long-arm thumb spica casts for ‘at risk’ fractures such as those found in patients that were thought to be unlikely to follow activity restrictions while in the cast. This bias toward long-arm thumb spicas may account for this finding. The interpretation of radiographic union of scaphoid fractures is difficult on plain radiographs alone. While a CT scan is not standard of care at this institution, advanced imaging would have provided a more accurate time to union. However, the avoidance of the required additional radiation was felt to be appropriate in this circumstance. We had difficulty with follow-up compliance, particularly in subjects treated with a cast that did not go on to require surgery.
In conclusion, when evaluated as individual variables, the children that required surgery tended to be older, have a displaced fracture, or already have evidence of AVN at their index presentation. A non- or delayed-union also had a higher proportion of surgical conversion. However, when these variables were entered into a regression model, the only significant predictor of failing conservative treatment of scaphoid fractures in this young cohort was a delay in that initial treatment from the time of injury. Although we found that our institution leans toward aggressively immobilizing patients presenting to clinic with snuffbox tenderness, we also identified that a delay in managing pediatric scaphoid fractures is the most important predictor of converting to surgical management. Therefore, given the potential sequelae of missing a scaphoid fracture in children with snuffbox tenderness without radiographic evidence of a fracture on index films, we recommend remaining vigilant in children with appropriate mechanisms of injury and clinical examination via early immobilization of suspected fractures and repeat imaging to confirm or refute the suspected diagnosis.
This study was supported by the Division of Orthopedics, Rady Children’s Hospital, San Diego.
Conflicts of interest
There are no conflicts of interest.
1. Gholson JJ, Bae DS, Zurakowski D, Waters PM. Scaphoid fractures in children and adolescents: contemporary injury patterns and factors influencing time to union. J Bone Joint Surg Am. 2011; 93:1210–1219.
2. Nguyen JC, Nguyen MK, Arkader A, Guariento A, Sze A, Moore ZR, Chang B. Age-dependent changes in pediatric scaphoid fracture
pattern on radiographs. Skeletal Radiol. 2020; 49:2011-2018.
3. Shaterian A, Santos PJF, Lee CJ, Evans GRD, Leis A. Management modalities and outcomes following acute scaphoid fractures in children: a quantitative review and meta-analysis. Hand (N Y). 2019; 14:305–310.
4. Lempesis V, Rosengren BE, Landin L, Tiderius CJ, Karlsson MK. Hand fracture epidemiology and etiology in children-time trends in Malmö, Sweden, during six decades. J Orthop Surg Res. 2019; 14:213.
5. Van Tassel DC, Owens BD, Wolf JM. Incidence estimates and demographics of scaphoid fracture
in the U.S. population. J Hand Surg Am. 2010; 35:1242–1245.
6. Sallam AA, Briffa N, Mahmoud SS, Imam MA. Normal wrist development in children and adolescents: a geometrical observational analysis based on plain radiographs. J Pediatr Orthop. 2020; 40:e860–e872.
7. Anz AW, Bushnell BD, Bynum DK, Chloros GD, Wiesler ER. Pediatric scaphoid fractures. J Am Acad Orthop Surg. 2009; 17:77–87.
8. Stanciu C, Dumont A. Changing patterns of scaphoid fractures in adolescents. Can J Surg. 1994; 37:214–216.
9. Vahvanen V, Westerlund M. Fracture of the carpal scaphoid in children. A clinical and roentgenological study of 108 cases. Acta Orthop Scand. 1980; 51:909–913.
10. D’Arienzo M. Scaphoid fractures in children. J Hand Surg Br. 2002; 27:424–426.
11. Toh S, Miura H, Arai K, Yasumura M, Wada M, Tsubo K. Scaphoid fractures in children: problems and treatment. J Pediatr Orthop. 2003; 23:216–221.
12. Evenski AJ, Adamczyk MJ, Steiner RP, Morscher MA, Riley PM. Clinically suspected scaphoid fractures in children. J Pediatr Orthop. 2009; 29:352–355.
13. Waters PM, Stewart SL. Surgical treatment of nonunion and avascular necrosis of the proximal part of the scaphoid in adolescents. J Bone Joint Surg Am. 2002; 84:915–920.
14. Gutow AP. Percutaneous fixation of scaphoid fractures. J Am Acad Orthop Surg. 2007; 15:474–485.
15. Chloros GD, Themistocleous GS, Wiesler ER, Benetos IS, Efstathopoulos DG, Soucacos PN. Pediatric scaphoid nonunion. J Hand Surg Am. 2007; 32:172–176.
16. Khouri JS, Shin AY. Pediatric scaphoid fractures. In: Buijze GA, Jupiter JB, editors. Scaphoid fractures: evidence-based management; Elsevier; 2018. pp. 189-197.
17. Weber DM, Fricker R, Ramseier LE. Conservative treatment of scaphoid nonunion in children and adolescents. J Bone Joint Surg Br. 2009; 91:1213–1216.
18. Ruby LK, Stinson J, Belsky MR. The natural history of scaphoid non-union. A review of fifty-five cases. J Bone Joint Surg Am. 1985; 67:428–432.
19. Bae DS, Gholson JJ, Zurakowski D, Waters PM. Functional outcomes after treatment of scaphoid fractures in children and adolescents. J Pediatr Orthop. 2016; 36:13–18.
20. Christodoulou AG, Colton CL. Scaphoid fractures in children. J Pediatr Orthop. 1986; 6:37–39.
21. Nafie SA. Fractures of the carpal bones in children. Injury. 1987; 18:117–119.
22. Fabre O, De Boeck H, Haentjens P. Fractures and nonunions of the carpal scaphoid in children. Acta Orthop Belg. 2001; 67:121–125.
23. Wilson-MacDonald J. Delayed union of the distal scaphoid in a child. J Hand Surg Am. 1987; 12:520–522.
24. Henderson B, Letts M. Operative management of pediatric scaphoid fracture
nonunion. J Pediatr Orthop. 2003; 23:402–406.
25. Jauregui JJ, Seger EW, Hesham K, Walker SE, Abraham R, Abzug JM. Operative management for pediatric and adolescent scaphoid nonunions: a meta-analysis. J Pediatr Orthop. 2019; 39:e130–e133.