Abstract
Purpose
The treatment of scaphoid nonunions remains a challenge, with multiple surgical techniques described. The purpose of this retrospective study was to evaluate the union rates in patients undergoing treatment of scaphoid nonunion with a single headless compression screw and the risk factors associated with failure to achieve union.
Materials and Methods
Using the CPT code for the repair of a scaphoid nonunion (25440), 167 patients were identified as having operative repair of a scaphoid fracture nonunion using headless compression screw fixation performed at a single institution between the years of 2002 and 2024. All patients were followed clinically and with serial radiographic analysis postoperatively to determine if union was achieved. The primary outcome of fracture union was analyzed for each patient. Demographics of individual patients, fracture characteristics, and different autograft types were collected as a secondary analysis to evaluate for risk factors for scaphoid nonunion repair failure.
Results
The overall union rate was 75.5% (126/167). Female patients trended toward a higher rate of union following repair compared with male patients, although statistical significance was not fully met ( p = 0.07). Age, time from injury to surgery, fracture location, bone graft source, bone graft type, avascular necrosis, tobacco use, and workers' compensation did not demonstrate a significant impact on union rates. No difference was found in the union rate of scaphoid nonunion repairs based on the graft harvest location or graft type.
Conclusion
Fixation with headless compression screws and autologous bone graft is a reliable, effective method for treating scaphoid nonunions, with union rates similar to those previously reported. Female patients trended toward having a higher rate of union following repair, although statistical significance was not reached.
Level of Evidence
Therapeutic Level IV.
Keywords: autograft, nonunion, repair, scaphoid, headless screw
An unstable scaphoid nonunion disrupts wrist kinematics, leading eventually to collapse and progressive arthrosis. 1 2 The effects of and progression resulting from scaphoid nonunion advanced collapse on the wrist are well described, and the recommendation for surgical fixation for any displacement underscores this principle. 3 As such, scaphoid nonunions are most commonly managed with surgical fixation in the absence of preexisting arthritis. 4 Although ample literature exists documenting treatment strategies for scaphoid nonunions, definitive management remains difficult, and choice of both bone graft source and method of fixation remains controversial.
Multiple strategies for the surgical fixation of scaphoid nonunions have been described, including the use of Kirschner wires (K-wires), headless compression screws, and locking plates, among others. The headless compression screw as initially proposed by Herbert and Fisher in 1984 is the most commonly utilized fixation for acute scaphoid injuries, but there remains ongoing debate between modalities for nonunion repair. 5 Plate fixation has also been shown to be very successful for the treatment of scaphoid nonunions, debatably better than screws fixation, but does often require eventual removal of hardware. 6 7 Prior studies have also shown success with fixation alone without bone grafting, particularly isolated screw fixation in nondisplaced fibrous nonunions. 8 9 In nonunions with displacement, bone loss, or avascular necrosis (AVN), the use of bone grafting with internal fixation is accepted as an adjuvant to restore biologic healing potential and address defects. 10 11 12 While there are multiple retrospective studies evaluating different fixation techniques and bone graft types and sources, no consensus exists. Even with an adequate biologic environment, further biomechanical factors, most importantly optimal screw positioning, can influence union rates. 13 14 Even as newer techniques such as vascularized bone grafting have been described for treatment of nonunions, union rates varying from 27 to 100% have been reported. 15 16 17 Because of the wide variety in treatment strategies and union rates, we sought to evaluate a consecutive series of patients to establish a standard upon which patients may be counseled regarding union rate of this challenging injury.
The objective of this study was to evaluate all patients with a scaphoid nonunion, regardless of nonunion location, time from presentation, and presence of bone loss or AVN to provide a realistic assessment of healing potential and evaluate factors predictive of radiographic union. This study represents a large, retrospective single-center review of scaphoid nonunions treated with a uniform method of fixation, a single standard headless compression screw. We sought to establish a rate of union and to identify patient variables that may influence outcomes in our patient cohort. We hypothesize that the use of headless compression screw fixation along with judicious application of graft source and type will result in acceptable rates of radiographic union in all types of scaphoid nonunions, regardless of patient demographics.
Materials and Methods
Patient Population
The CPT code 25440 was used to identify patients who underwent repair of a scaphoid nonunion at a single academic institution between the years 2002 and 2024. Diagnosis of scaphoid nonunion was determined by two fellowship-trained hand surgeons. Patients who underwent primary scaphoid nonunion repair using a headless compression screw as the primary mode of fixation were included. Exclusion criteria were revision surgeries/previously operated scaphoids, cases utilizing primary types of fixation other than a headless compression screw, cases with intraoperative use of bone morphogenic protein, inadequate operative/postoperative records, and less than 8 months of documented clinical follow-up without confirmed union. This study was performed with institutional review board exemption and in accordance with Strengthening the Reporting of Observational Studies in Epidemiology guidelines.
Data Collection
Demographic and clinical/perioperative variables were recorded via direct review of the medical record. These included patient age, sex, tobacco use, workers' compensation payment, fracture location/morphology, presence of AVN, mechanism of injury, time from reported injury to operative treatment (in days), bone graft type, bone graft harvest location, scaphoid union, and time from surgery to union (in months). Bone graft types included no graft, packed/cancellous, structural/corticocancellous, and vascularized pedicle graft. Bone graft harvest sites were iliac crest, distal radius, and olecranon process.
Patients were considered to have achieved union with radiographic confirmation of full bony consolidation (progression of union visualized using the standard scaphoid view series, without progressive peri-implant lucency/screw migration, or obliteration of the fracture lines), signs of bony healing with full painless range of motion and return to preinjury activities without restriction, or with computed tomography (CT) imaging confirmation. CT imaging was only utilized to evaluate union when patients remained symptomatic at 3 or more months postoperatively and at the discretion of the attending surgeon.
Surgical Technique/Clinical Protocol
All nonunion repairs were performed by one of two fellowship-trained hand surgeons. The treating surgeon independently chose the surgical approach, need for bone graft, and graft type based on fracture characteristics and individual surgical expertise with the common factor being the use of a single fixation method. The treating surgeons typically utilized a dorsal approach for proximal one-third nonunions ( Fig. 1 ) and a volar approach for waist nonunions ( Fig. 2 ) to obtain adequate exposure for preparation of the nonunion site and then insertion of the fixation screw. The longest standard headless compression screw that could be completely countersunk below the articular surface was chosen. Preference was given to placement perpendicular to the axis of the nonunion over absolute length when the fracture lines were not perpendicular to the anatomic scaphoid axis. In the absence of any significant cortical defect or deformity, packed cancellous autograft was preferentially utilized. When significant defects secondary to cystic degeneration or AVN were found after thorough curettage of the nonunion site, structural corticocancellous autograft was utilized to support and fill the defect with or without the addition of retained K-wires to prevent displacement of the implanted cortical graft. This was performed as the preferred surgical technique of one surgeon when using structural graft, comprising 23 of the 48 patients treated with structural graft. This technique required return to surgery around 6 to 8 weeks postoperatively for removal of the retained wires ( Fig. 3 ). Vascularized distal radius graft was typically utilized where there was high concern for avascularity of the proximal pole (severe sclerosis or fragmentation of the proximal pole) in younger patients, although not exclusively, and was performed by harvesting a 4 to 5 extensor compartment artery vascularized distal radius graft and fixating it to the proximal pole of the scaphoid with a mini-Acutrak screw. Patients with negligible bone loss and fibrous nonunion were treated with nonunion site debridement and internal fixation without bone grafting.
Fig. 1.
Preoperative anteroposterior (AP) (A) and lateral (B) radiographs of a patient with a proximal one-third scaphoid nonunion. Final postoperative AP (C) , oblique (D) , and lateral (E) radiographs demonstrating definitive healing of the scaphoid nonunion following nonunion repair with packed cancellous distal radius bone graft.
Fig. 2.
Preoperative anteroposterior (AP) (A) and lateral (B) radiographs of a patient with a scaphoid waist nonunion. Final postoperative AP (C) , oblique (D) , and lateral (E) radiographs demonstrate definitive healing of the scaphoid nonunion following nonunion repair with packed cancellous distal radius bone graft.
Fig. 3.
Early postoperative anteroposterior (AP) radiograph demonstrating a Kirschner wire placed temporarily to prevent displacement of cortical graft during nonunion repair of a scaphoid waist fracture with structural corticocancellous distal radius bone graft.
Following surgical fixation, all patients were immobilized with a thumb spica splint for the first 2 weeks, followed by a thumb spica cast for typically 5 to 7 additional weeks. The need for additional cast immobilization was based on a review of the patient's clinical symptoms, subjective discomfort, and radiographic findings. Patients were followed postoperatively at the discretion of the surgeon, typically with follow-up appointments at 2 weeks, 6 to 8 weeks, and 3 months postoperatively. Further follow-up was made if warranted after that. An example of persistent scaphoid nonunion following nonunion repair is shown in Fig. 4 .
Fig. 4.
Preoperative anteroposterior (AP) (A) and lateral (B) radiographs of a patient with a scaphoid waist nonunion. Early postoperative AP ( C ) and lateral (C) radiographs following nonunion repair with packed cancellous distal radius bone graft. Final postoperative AP (D) and lateral (E) radiographs demonstrating persistent nonunion 9-month status after nonunion repair.
Statistical Analysis
Achievement of scaphoid union was the primary outcome of interest. Patients were grouped between those that were documented to achieve union and those who did not (nonunion). Demographic and perioperative characteristics were compared between groups using chi-square for categorical variables and Mann–Whitney U test for continuous variables, respectively. Fisher's exact test was used in place of chi-square when any category had ≤5 occurrences for a given variable. Patients were then grouped according to the specific type of bone graft utilized, as follows: no graft, packed cancellous iliac crest, packed cancellous distal radius, structural corticocancellous iliac crest, structural corticocancellous distal radius, and vascularized distal radius. Demographic and clinical characteristics were then compared among bone graft conditions using chi-square/Fisher's exact test for categorical variables and Kruskal–Wallis H test for continuous variables.
Binomial multiple logistic regression analysis was then performed to assess the odds of a nonunion outcome associated with individual demographic/perioperative characteristics while holding constant the effect of all other included variables. These variables included bone grafting location and type, age, sex, tobacco use, workers' compensation, fracture location, and AVN. Statistical testing was performed using Stata version 18.0 (StataCorp, College Station, TX). A p -value ≤ 0.05 was considered the threshold for statistical significance.
Results
A total of 231 patients were initially identified. Of these, 6 did not undergo fixation with a headless compression screw, 20 had undergone previous scaphoid surgery, and 38 did not have adequate clinical follow-up data available and were excluded from this study. Following exclusion, 167 patients were included, of whom 126 (75.4%) achieved union of their scaphoid fracture and 41 (24.6%) went on to nonunion. Mean patient age was 23.2 years, a majority were male (85.6%) and nonsmokers (83.8%), and 8.4% paid through workers' compensation. Fractures were located at the proximal pole in 42.3% of cases and the scaphoid waist in 58.7%. AVN was present in 29.9% of patients. The most common mechanism of injury was sport-related (45.2%). Most cases utilized packed cancellous bone graft (62.3%) and harvested graft from the distal radius (78.4%). Mean time from injury to surgery was 332.9 days. None of these demographic and clinical characteristics significantly differed between union and nonunion groups ( Table 1 ). Time to union was not significantly associated with any demographic or perioperative characteristic. Sex ( p = 0.033), fracture location ( p < 0.001), and time to treatment ( p = 0.002) significantly differed across bone grafting conditions ( Table 2 ). Structural grafts tended to be used more frequently for waist fractures and more vascularized grafts were used to treat proximal pole fractures. Multiple regression analysis did not demonstrate any statistically significant, independent association between selected demographic or clinical variables and odds of persistent scaphoid nonunion ( Table 3 ).
Table 1. Association of baseline characteristics with eventual scaphoid union.
| Characteristic | Total ( n = 167) |
Union ( n = 126) |
Nonunion ( n = 41) |
p -Value |
|---|---|---|---|---|
| Age (mean ± SD, y) | 23.2 ± 8.0 | 23.1 ± 9.6 | 23.4 ± 8.0 | c 0.281 |
| Age category | a 0.372 | |||
| < 18-y old | 29.9% (50) | 31.8% (40) | 24.4% (10) | |
| ≥18-y old | 70.1% (117) | 68.3% (86) | 75.6% (31) | |
| Sex | b 0.070 | |||
| Male | 85.6% (143) | 82.5% (104) | 95.1% (39) | |
| Female | 14.4% (24) | 17.5% (22) | 4.9% (2) | |
| Smoking status | a 0.856 | |||
| Nonsmoker | 83.8% (140) | 84.1% (106) | 82.9% (34) | |
| Smoker | 16.2% (27) | 15.9% (20) | 17.1% (7) | |
| Payor | b 0.748 | |||
| Other | 91.6% (153) | 92.1% (116) | 90.2% (37) | |
| Workers' comp | 8.4% (14) | 7.9% (10) | 9.8% (4) | |
| Fracture location | a 0.699 | |||
| Proximal pole | 42.3% (69) | 40.5% (51) | 43.9% (18) | |
| Waist | 58.7% (98) | 59.5% (75) | 56.1% (23) | |
| Avascular necrosis | a 0.285 | |||
| Not present | 70.1% (117) | 72.2% (91) | 63.4% (26) | |
| Present | 29.9% (50) | 27.8% (35) | 36.6% (15) | |
| Mechanism of injury | a 0.903 | |||
| Fall | 21.1% (35) | 19.8% (25) | 25.0% (10) | |
| Sport | 45.2% (75) | 46.8% (59) | 40.0% (16) | |
| MVC | 16.3% (27) | 15.9% (20) | 17.5% (7) | |
| Fall from height | 6.0% (10) | 5.6% (7) | 7.5% (3) | |
| Impact/strike | 4.8% (8) | 5.6% (7) | 2.5% (1) | |
| Unknown | 6.6% (11) | 6.4% (8) | 7.5% (3) | |
| Time to treatment (d) | 332.9 ± 439.1 | 318.8 ± 408.5 | 376.1 ± 525.5 | c 0.478 |
| Graft type | b 0.269 | |||
| None | 3.6% (6) | 3.2% (4) | 4.9% (2) | |
| Packed | 62.3% (104) | 58.7% (74) | 73.2% (30) | |
| Structural | 28.7% (48) | 31.8% (40) | 19.5% (8) | |
| Vascularized | 5.4% (9) | 6.4% (8) | 2.4% (1) | |
| Graft harvest location | b 0.776 | |||
| None | 3.6% (6) | 3.2% (4) | 4.9% (2) | |
| Iliac crest | 17.4% (29) | 16.7% (21) | 19.5% (8) | |
| Distal radius | 78.4% (131) | 79.4% (100) | 75.6% (31) | |
| Olecranon | 0.6% (1) | 0.8% (1) | 0% (0) |
Abbreviations: MVC, motorized vehicle collision; SD, standard deviation.
p -Values calculated using chi-square test.
p -Values calculated using Fisher's exact test.
p -Value calculated using Mann–Whitney U test.
Table 2. Association of bone graft with demographic and perioperative characteristics.
| Characteristic | None ( n = 6) |
Packed iliac ( n = 5) |
Packed radius ( n = 99) | Structural iliac ( n = 24) | Structural radius ( n = 24) d | Vascularized radius ( n = 9) | p -Value |
|---|---|---|---|---|---|---|---|
| Union outcome | a 0.222 | ||||||
| Union | 66.7% (4) | 40.0% (2) | 72.7% (72) | 79.2% (19) | 87.5% (21) | 88.9% (8) | |
| Nonunion | 33.3% (2) | 60.0% (3) | 27.3% (27) | 20.8% (5) | 12.5% (3) | 11.1% (1) | |
| Time to union (mean ± SD, mo) | 5.8 ± 3.2 | 7.0 ± 1.4 | 5.4 ± 2.1 | 6.2 ± 1.8 | 4.7 ± 2.2 | 5.4 ± 2.2 | c 0.183 |
| Age (y) | 23.2 ± 8.0 | 23.4 ± 10.3 | 22.7 ± 9.6 | 22.8 ± 7.2 | 26.5 ± 10.1 | 20.4 ± 8.1 | c 0.168 |
| Age category | a 0.127 | ||||||
| < 18 y old | 4.0% (2) | 4.0% (2) | 66.0% (33) | 14.0% (7) | 4.0% (2) | 8.0% (4) | |
| ≥18 y old | 3.4% (4) | 2.6% (3) | 56.4% (66) | 14.5% (17) | 18.8% (22) | 4.3% (5) | |
| Sex | a 0.033 | ||||||
| Male | 4.2% (6) | 2.8% (4) | 62.2% (89) | 14.7% (21) | 10.5% (15) | 5.6% (8) | |
| Female | 0% (0) | 4.2% (1) | 41.7% (10) | 12.5% (3) | 37.5% (9) | 4.2% (1) | |
| Smoking status | c 0.520 | ||||||
| Nonsmoker | 2.9% (4) | 3.6% (5) | 60.7% (85) | 14.3% (20) | 12.9% (18) | 5.7% (8) | |
| Smoker | 7.4% (2) | 0% (0) | 51.9% (14) | 14.8% (4) | 22.2% (6) | 3.7% (1) | |
| Payor | a 0.751 | ||||||
| Other | 3.9% (6) | 2.6% (4) | 58.8% (90) | 14.4% (22) | 15.0% (23) | 5.2% (8) | |
| Workers' comp | 0% (0) | 7.1% (1) | 64.3% (9) | 14.3% (2) | 7.1% (1) | 7.1% (1) | |
| Fracture location | a < 0.001 | ||||||
| Proximal pole | 5.8% (4) | 0% (0) | 73.9% (51) | 7.3% (5) | 2.9% (2) | 10.1% (7) | |
| Waist | 2.0% (2) | 5.1% (5) | 49.0% (48) | 19.4% (19) | 22.5% (22) | 2.0% (2) | |
| Avascular necrosis | a 0.085 | ||||||
| Not Present | 5.1% (6) | 2.6% (3) | 58.1% (68) | 15.4% (18) | 16.2% (19) | 2.6% (3) | |
| Present | 0% (0) | 4.0% (2) | 62.0% (31) | 12.0% (6) | 10.0% (5) | 12.0% (6) | |
| Mechanism of injury | b 0.093 | ||||||
| Fall | 2.9% (1) | 5.7% (2) | 60.0% (21) | 14.3% (5) | 14.3% (5) | 2.9% (1) | |
| Sport | 2.7% (2) | 1.3% (1) | 65.3% (49) | 10.7% (8) | 10.7% (8) | 9.3% (7) | |
| MVC | 7.4% (2) | 7.4% (2) | 51.9% (14) | 18.5% (5) | 14.8% (4) | 0% (0) | |
| Fall from height | 10.0% (1) | 0% (0) | 40.0% (4) | 0% (0) | 50.0% (5) | 0% (0) | |
| Impact/strike | 0% (0) | 0% (0) | 75.0% (6) | 12.5% (1) | 12.5% (1) | 0% (0) | |
| Unknown | 0% (0) | 0% (0) | 36.4% (4) | 45.5% (5) | 9.1% (1) | 9.1% (1) | |
| Time to treatment (d) | 148.5 ± 174.7 | 118.6 ± 1126.2 | 265.5 ± 301.3 | 604.6 ± 769.1 | 453.9 ± 485.1 | 268.4 ± 323.6 | c 0.002 |
Abbreviations: MVC, motorized vehicle collision; SD, standard deviation.
Note: Bold indicates statistical significance.
p -Value calculated using Fisher's exact test.
p -Value calculated using chi-square.
p -Value calculated using Kruskal–Wallis H test.
One case utilized a structural olecranon graft.
Table 3. Multiple regression analysis of predictors of scaphoid nonunion.
| Characteristic | OR | SE | 95% CI | ap -Value |
|---|---|---|---|---|
| Graft condition | ||||
| None | Reference | – | – | – |
| Packed iliac | 3.9 | 4.9 | [0.3, 44.8] | 0.270 |
| Packed radius | 0.8 | 0.7 | [0.1, 4.7] | 0.809 |
| Structural iliac | 0.6 | 0.6 | [0.1, 4.1] | 0.577 |
| Structural radius | 0.4 | 0.4 | [0.04, 3.4] | 0.391 |
| Vascularized radius | 0.3 | 0.4 | [0.01, 4.6] | 0.350 |
| Age | 1.0 b | 0.02 | [0.98, 1.1] | 0.366 |
| Sex | ||||
| Male | Reference | – | – | – |
| Female | 0.2 | 0.2 | [0.04, 1.2] | 0.076 |
| Smoking status | ||||
| Nonsmoker | Reference | – | – | – |
| Smoker | 1.3 | 0.7 | 0.5, 3.7] | 0.561 |
| Payor | ||||
| Other | Reference | – | – | – |
| Workers' comp | 1.3 | 0.9 | [0.3, 5.2] | 0.698 |
| Fracture location | ||||
| Proximal pole | Reference | – | – | – |
| Waist | 1.0 | 0.4 | [0.4, 2.3] | 0.988 |
| Avascular necrosis | ||||
| Not present | Reference | – | – | – |
| Present | 1.3 | 0.5 | [0.6, 2.9] | 0.515 |
Abbreviations: CI, confidence interval; OR, odds ratio; SE, standard error.
p -Values calculated using binomial multiple logistic regression.
Relative to each 1-year increase.
Discussion
The overall union rates of 75.5% in our study of scaphoid nonunion repairs at a single institution was comparable to previously published retrospective reviews in recent literature. 18 19 20 21 22 We found no significant difference when comparing union rates across different types of graft harvest sites and the types of graft used in the nonunion repairs, regardless of fracture site location or characteristics. Patient sex trended toward significance with female sex having higher rates of union; however, the statistical significance threshold was not met. Age (analyzed as a continuous variable and by a cutoff of 18 years of age), time to treatment, mechanism of injury, tobacco use, and workers' compensation claims associated with the injury did not affect the rate of unions following repair. Subgroup analysis unexpectedly showed a significantly higher rate of structural/corticocancellous graft use in female patients compared with the higher rate of packed cancellous graft in males, although the importance of this is uncertain. Structural/corticocancellous graft was also found to be more commonly used in waist fractures, which was expected due to their use in the correction of the humpback deformity that occurs in chronic waist fractures. 10 11 12
We do understand the inherent weaknesses of our study. While our study includes a large cohort of patients, our study is retrospective in nature, and therefore more prone to confounding and bias when compared with a prospective study. Tobacco use and the time between initial surgery are additionally limited by the patients' ability to accurately report this information. We also did not stratify our series based on nonunion morphology, including all patients regardless of presence or absence of deformity or collapse. Additionally, these morphologies contribute to determining the preferred surgical treatment, which was left at the discretion of the operative surgeon and adds further limitations. A final weakness of our study was that determination of union was made using serial radiographic evaluation for up to 1 year following surgery, with only 72.1% of patients undergoing postoperative CT scan. There remains mixed literature on the interobserver reliability utilizing only radiographs for union determination. 23 Our analytical protocol was similar to prior studies, and the statistical power afforded by the additional radiograph-only subjects was not inconsequential, despite needing to be asymptomatic as an additional qualification for radiograph-only union.
The 75.5% union rate demonstrated in our cohort establishes a strong reference for a single common method of fixation. Large-scale literature reviews suggest that overall union rates following surgical treatment are 69 to 84%, with lower rates expected with proximal pole nonunions. 18 19 20 21 22 These union rates are supported by a more recent retrospective comparison study between one and two screw fixation, which identified an 89% union rate with a single screw. 17 Our study included patients with significant delays in presentation, AVN, deformity, and carpal collapse. By analyzing a large cohort of scaphoid nonunions of varied severity treated with a single type of fixation with all types of bone grafting, our findings are generalizable for a broad range of patients. The results present a reference standard that allows surgeons to counsel patients on the factors that influence union rate of surgical treatment of scaphoid nonunion. Patients requiring structural corticocancellous bone graft tended to demonstrate a longer delay from injury to definitive surgery. This is likely secondary to encountered cystic degeneration and humpback deformity associated with nonunions of the scaphoid waist and segmental cortical bone loss that occurs with motion and necrosis over time. Despite this trend, corticocancellous graft showed no lower rate of union following repair in our study.
Prior studies have additionally noted that prolonged delay in surgery were significant predictors of union rates and functional outcome as well. 20 21 22 24 25 26 27 The role of early surgical fixation following diagnosis of nonunion is supported by existing literature; however, our data did not show a significant difference in union based on duration to surgery. Inaparthy and Nicholl reported that all patients treated within 6 months from injury in their series united; on the contrary, only 25% of those with a remote injury more than 5 years prior to surgery demonstrated union. 24 Similarly, other studies have demonstrated that shorter duration to surgical treatment in fractures closer to the proximal pole of the scaphoid with less structural compromise unite more predictably. 26 Our results did not demonstrate a significant difference related to timing from initial injury to surgery, but this phenomenon may be attributed to inconsistencies between perceived dates of injuries and the actual dates. It was not uncommon to have patients note onset of pain following a vague history or after frequent falls/trauma, with radiographic changes to the scaphoid, which would suggest a much older injury. To limit biased removal of these timings, all reported injury dates were recorded as stated. Furthermore, the attending surgeons' discretion to use of different types of grafts aimed to treat deformities and radiographic changes associated with more chronic injuries could have further confounded this result.
Several other studies have been published evaluating union rates in the treatment of this difficult condition. A large systematic review of surgical treatment with bone grafting evaluated 5,464 scaphoid nonunions, with overall union rates of 84 and 80% for subjects undergoing vascularized and nonvascularized bone grafting, respectively. 20 In the presence of AVN, union rates dropped to 74% with vascularized grafts and 62% with nonvascularized. Reported union rates varied considerably among included studies. While AVN was defined as presence of proximal fragment sclerosis or fragmentation on plain radiographs in our study, this has been shown to be less reliable compared with magnetic resonance imaging with contrast or histopathology of intraoperative specimen. This could partially explain why AVN did not significantly affect union rates in our study. Furthermore, Rancy and colleagues have demonstrated in their histopathological study that true infarction of the proximal pole of the scaphoid is rare, which explains their equivocal outcome differences between vascularized and nonvascularized bone grafts. 28 29 However, other retrospective studies have demonstrated an improvement in union rates with vascularized bone grafts in selected patients, and so higher-level studies are required to reach consensus. 30
More recently, Ammori and colleagues published their findings on 462 patients undergoing bone graft surgery for nonunion of the scaphoid. 21 In their large, multicenter study, an overall union rate of 69% was documented. In cases with indeterminate union, presumed union yielded an overall union rate of 78% that was similar to the 75.5% union rate demonstrated in our series. Their series likewise preferentially utilized nonvascularized bone grafting. While there remains ongoing debate over the clinical effect of vascularized bone graft in scaphoid nonunion repair, all vascularized grafts in our study were done in the setting of radiographic evidence of AVN, with our rate of union being higher than our general rate of union at 87.5%. The low number of vascularized grafts limits the statistical significance of this finding but may suggest an underutilization in our patient population, considering a higher rate of union in those select patients with AVN compared with the typically higher rate of those without AVN.
Regarding demographic factors, sex was the only demographic variable in our series that trended toward statistical significance, affecting union rates. Females trended toward higher union rates compared with their male counterparts. Larger patient populations would be needed to delineate whether this phenomenon represented an actual difference or was a result of normal variance. Future studies are needed to determine if this trend is significant and to determine what the exact cause may have been.
Our findings were also intriguing as smoking was not predictive of nonunion. Prior literature has suggested that tobacco use can adversely affect union. 21 27 31 In a similar series to ours, Little et al. concluded that smoking was strongly associated with failure of union following screw fixation and nonvascularized bone grafting used in fixation of scaphoid nonunion or delayed union. 27 Tobacco use was self-reported in our patient population, which may have confounded the true rate of tobacco use in our study.
Our study represents a large, complete cohort of scaphoid nonunions treated with a single method of fixation at a single institution. The results generated show that a union rate consistent with that of previous retrospective studies can be achieved with headless compression screw fixation and accepted bone grafting techniques in a way that is generalizable to the entire population.
Conflict of Interest The authors declare that they have no conflict of interest.
Informed Consent
This study was performed under the guidance of the SUNY Upstate Institutional Review Board and was granted an exempt status according to their standards and review process.
References
- 1.Szabo R M, Manske D. Displaced fractures of the scaphoid. Clin Orthop Relat Res. 1988;(230):30–38. [PubMed] [Google Scholar]
- 2.Mack G R, Bosse M J, Gelberman R H, Yu E. The natural history of scaphoid non-union. J Bone Joint Surg Am. 1984;66(04):504–509. [PubMed] [Google Scholar]
- 3.Ruby L K, Leslie B M. Wrist arthritis associated with scaphoid nonunion. Hand Clin. 1987;3(04):529–539. [PubMed] [Google Scholar]
- 4.Moon E S, Dy C J, Derman P, Vance M C, Carlson M G. Management of nonunion following surgical management of scaphoid fractures: current concepts. J Am Acad Orthop Surg. 2013;21(09):548–557. doi: 10.5435/JAAOS-21-09-548. [DOI] [PubMed] [Google Scholar]
- 5.Rogers M J, Ohlsen S M, Huang J I. Fixation techniques for scaphoid nonunion. J Am Acad Orthop Surg. 2023;31(15):783–792. doi: 10.5435/JAAOS-D-23-00287. [DOI] [PubMed] [Google Scholar]
- 6.Putnam J G, Mitchell S M, DiGiovanni R M, Stockwell E L, Edwards S G. Outcomes of unstable scaphoid nonunion with segmental defect treated with plate fixation and autogenous cancellous graft. J Hand Surg Am. 2019;44(02):1600–1.6E9. doi: 10.1016/j.jhsa.2018.05.023. [DOI] [PubMed] [Google Scholar]
- 7.Dodds S D, Williams J B, Seiter M, Chen C. Lessons learned from volar plate fixation of scaphoid fracture nonunions. J Hand Surg Eur Vol. 2018;43(01):57–65. doi: 10.1177/1753193417743636. [DOI] [PubMed] [Google Scholar]
- 8.Somerson J S, Fletcher D J, Srinivasan R C, Green D P. Compression screw fixation without bone grafting for scaphoid fibrous nonunion. Hand (N Y) 2015;10(03):450–453. doi: 10.1007/s11552-015-9741-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Slade J F, III, Geissler W B, Gutow A P, Merrell G A. Percutaneous internal fixation of selected scaphoid nonunions with an arthroscopically assisted dorsal approach. J Bone Joint Surg Am. 2003;85-A 04:20–32. doi: 10.2106/00004623-200300004-00003. [DOI] [PubMed] [Google Scholar]
- 10.Robbins R R, Ridge O, Carter P R. Iliac crest bone grafting and Herbert screw fixation of nonunions of the scaphoid with avascular proximal poles. J Hand Surg Am. 1995;20(05):818–831. doi: 10.1016/s0363-5023(05)80438-1. [DOI] [PubMed] [Google Scholar]
- 11.Bunker T D, McNamee P B, Scott T D. The Herbert screw for scaphoid fractures. A multicentre study. J Bone Joint Surg Br. 1987;69(04):631–634. doi: 10.1302/0301-620X.69B4.3611171. [DOI] [PubMed] [Google Scholar]
- 12.Pinder R M, Brkljac M, Rix L, Muir L, Brewster M. Treatment of scaphoid nonunion: a systematic review of the existing evidence. J Hand Surg Am. 2015;40(09):1797–1.805E6. doi: 10.1016/j.jhsa.2015.05.003. [DOI] [PubMed] [Google Scholar]
- 13.Trumble T E, Clarke T, Kreder H J. Non-union of the scaphoid. Treatment with cannulated screws compared with treatment with Herbert screws. J Bone Joint Surg Am. 1996;78(12):1829–1837. doi: 10.2106/00004623-199612000-00005. [DOI] [PubMed] [Google Scholar]
- 14.Dodds S D, Panjabi M M, Slade J F., III Screw fixation of scaphoid fractures: a biomechanical assessment of screw length and screw augmentation. J Hand Surg Am. 2006;31(03):405–413. doi: 10.1016/j.jhsa.2005.09.014. [DOI] [PubMed] [Google Scholar]
- 15.Straw R G, Davis T R, Dias J J. Scaphoid nonunion: treatment with a pedicled vascularized bone graft based on the 1,2 intercompartmental supraretinacular branch of the radial artery. J Hand Surg [Br] 2002;27(05):413. doi: 10.1054/jhsb.2002.0808. [DOI] [PubMed] [Google Scholar]
- 16.Steinmann S P, Bishop A T, Berger R A. Use of the 1,2 intercompartmental supraretinacular artery as a vascularized pedicle bone graft for difficult scaphoid nonunion. J Hand Surg Am. 2002;27(03):391–401. doi: 10.1053/jhsu.2002.32077. [DOI] [PubMed] [Google Scholar]
- 17.Quadlbauer S, Beer T, Pezzei C et al. Stabilization of scaphoid type B2 fractures with one or two headless compression screws. Arch Orthop Trauma Surg. 2017;137(11):1587–1595. doi: 10.1007/s00402-017-2786-8. [DOI] [PubMed] [Google Scholar]
- 18.Munk B, Larsen C F. Bone grafting the scaphoid nonunion: a systematic review of 147 publications including 5,246 cases of scaphoid nonunion. Acta Orthop Scand. 2004;75(05):618–629. doi: 10.1080/00016470410001529. [DOI] [PubMed] [Google Scholar]
- 19.Merrell G A, Wolfe S W, Slade J F., III Treatment of scaphoid nonunions: quantitative meta-analysis of the literature. J Hand Surg Am. 2002;27(04):685–691. doi: 10.1053/jhsu.2002.34372. [DOI] [PubMed] [Google Scholar]
- 20.Ferguson D O, Shanbhag V, Hedley H, Reichert I, Lipscombe S, Davis T R. Scaphoid fracture non-union: a systematic review of surgical treatment using bone graft. J Hand Surg Eur Vol. 2016;41(05):492–500. doi: 10.1177/1753193415604778. [DOI] [PubMed] [Google Scholar]
- 21.British Society for Surgery of the Hand Scaphoid Nonunion Group . Ammori M B, Elvey M, Mahmoud S S et al. The outcome of bone graft surgery for nonunion of fractures of the scaphoid. J Hand Surg Eur Vol. 2019;44(07):676–684. doi: 10.1177/1753193419841278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Inoue G, Shionoya K, Kuwahata Y. Herbert screw fixation for scaphoid nonunions. An analysis of factors influencing outcome. Clin Orthop Relat Res. 1997;(343):99–106. [PubMed] [Google Scholar]
- 23.Dias J J, Taylor M, Thompson J, Brenkel I J, Gregg P J. Radiographic signs of union of scaphoid fractures. An analysis of inter-observer agreement and reproducibility. J Bone Joint Surg Br. 1988;70(02):299–301. doi: 10.1302/0301-620X.70B2.3346310. [DOI] [PubMed] [Google Scholar]
- 24.Inaparthy P K, Nicholl J E. Treatment of delayed/nonunion of scaphoid waist with Synthes cannulated scaphoid screw and bone graft. Hand (N Y) 2008;3(04):292–296. doi: 10.1007/s11552-008-9112-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wong K, von Schroeder H P. Delays and poor management of scaphoid fractures: factors contributing to nonunion. J Hand Surg Am. 2011;36(09):1471–1474. doi: 10.1016/j.jhsa.2011.06.016. [DOI] [PubMed] [Google Scholar]
- 26.Green D P. The effect of avascular necrosis on Russe bone grafting for scaphoid nonunion. J Hand Surg Am. 1985;10(05):597–605. doi: 10.1016/s0363-5023(85)80191-x. [DOI] [PubMed] [Google Scholar]
- 27.Little C P, Burston B J, Hopkinson-Woolley J, Burge P. Failure of surgery for scaphoid non-union is associated with smoking. J Hand Surg [Br] 2006;31(03):252–255. doi: 10.1016/j.jhsb.2005.12.010. [DOI] [PubMed] [Google Scholar]
- 28.Scaphoid Nonunion Consortium . Rancy S K, Swanstrom M M, DiCarlo E F, Sneag D B, Lee S K, Wolfe S W. Success of scaphoid nonunion surgery is independent of proximal pole vascularity. J Hand Surg Eur Vol. 2018;43(01):32–40. doi: 10.1177/1753193417732003. [DOI] [PubMed] [Google Scholar]
- 29.Rancy S K, Schmidle G, Wolfe S W. Does anyone need a vascularized graft? Hand Clin. 2019;35(03):323–344. doi: 10.1016/j.hcl.2019.03.005. [DOI] [PubMed] [Google Scholar]
- 30.Aibinder W R, Wagner E R, Bishop A T, Shin A Y. Bone grafting for scaphoid nonunions: is free vascularized bone grafting superior for scaphoid nonunion? Hand (N Y) 2019;14(02):217–222. doi: 10.1177/1558944717736397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Dinah A F, Vickers R H. Smoking increases failure rate of operation for established non-union of the scaphoid bone. Int Orthop. 2007;31(04):503–505. doi: 10.1007/s00264-006-0231-7. [DOI] [PMC free article] [PubMed] [Google Scholar]