Abstract
Aims
There is a lack of published evidence relating to the risk factors for nonunion seen in acute scaphoid fractures. We aimed to describe the risk factors for nonunion in adult patients (aged ≥ 16 years) treated with cast immobilization for acute scaphoid fractures.
Methods
We conducted a single centre cohort study of patients with an acute scaphoid fracture and who were followed up with a CT scan to assess union. Data were gathered retrospectively. Outcomes were classified as union or nonunion based on CT scan follow-up. Descriptive and regression analyses were performed.
Results
Of 502 patients, there were 50 nonunions (10%), of which 39 (78%) were male. The risk of nonunion tripled when fracture displacement was ≥ 1 mm compared with < 1 mm displacement (odds ratio (OR) 3.10; 95% CI 1.42 to 6.71. There was strong evidence that age was a risk factor for nonunion and the relationship between age and nonunion was non-linear. The reference age was 26 years and the highest risk of nonunion was in the 26- to 36-year age group. The risk of nonunion steadily reduced as age increased above aged 36 years; for example, patients aged 40 years older (66 years old) had a 91% lower risk of nonunion respectively relative to the reference age group.
Conclusion
The two risk factors for scaphoid nonunion in those treated with cast immobilization were fracture displacement and age. Further assessment and research could help inform clinical decision making and may help guide future research to improve clinical pathways of care.
Cite this article: Bone Jt Open 2026;7(8):1001–1006.
Keywords: Wrist, Scaphoid, Trauma, Acute, Fracture, Outcome, Risk, Nonunion, cast immobilization, scaphoid nonunions, acute fractures, nonunions, CT scans, acute scaphoid fractures, regression analyses, cohort studies, scaphoid fractures, displaced fractures
Introduction
Wrist trauma represents a significant clinical and economic burden for both patients and healthcare providers.1 The scaphoid is the most fractured bone in the wrist, accounting for 2% of all fractures.2 Over 7,000 patients will sustain a scaphoid fracture per year in the UK and approximately 15% of these patients will develop nonunion of the scaphoid.3 Scaphoid fractures cause significant disability, affecting a patient’s activities of daily living, employment, and recreational activities.4 The negative life impact of scaphoid nonunion is even more considerable for patients.5 Nonunion frequently results in a painful dysfunctional wrist and a specific pattern of progressive wrist arthritis called scaphoid nonunion advanced collapse (SNAC).5
The scientific literature for nonunion in acute scaphoid fractures is surprisingly scarce. Published studies are generally small, heterogeneous in design, and inconsistent in their findings.6-9 There are also problems with the classification of acute fractures and the accuracy in assessing union.10 The ‘Surgery versus cast immobilization for adults with a bicortical fracture of the scaphoid waist’ (SWIFFT) trial reported an overall nonunion rate of around 10% but the nonunion rate was over double in patients with fracture displacement of > 1 mm compared with those with < 1 mm displacement.11 However, our recent work has shown that the nonunion rate in ‘occult’ fractures is 3.5%.12
In this context, we assessed the radiological outcomes in a cohort of patients treated with cast immobilization for acute scaphoid fractures. The aims of this study were to: 1) Describe the characteristics of patients with the outcomes of union and nonunion; and 2) report the risk factors for nonunion in patients treated with cast immobilization.
Methods
The study was part of a retrospective service evaluation project. The study was registered locally, and no ethical approval was required as stated by the Health Research Authority (HRA). We reported the study according to Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies, and a complete STROBE checklist has been included the Supplementary Material.13
Study design and patient eligibility
We conducted this single-centre study at Oxford University Hospitals NHS Trust. The radiology database was searched to obtain all wrist CT scans undertaken from January 2000 to January 2024. All patients who underwent cast treatment for an acute scaphoid fracture and who were followed up with a CT scan to assess fracture healing were included in the study. Patients presenting more than two weeks after their injury, who were aged under 16 years, who were not immobilized within two weeks of injury, who did not have a clear bicortical fracture, and who presented with nonunions were not included.
Patients who were not followed up with a CT scan for assessing healing were not included. Note that routine assessment of union by CT became routine practice at Oxford University Hospitals NHS Trust from 2019. Patients aged under 16 years were excluded as this study focused on skeletally mature adults. Scaphoid fractures in skeletally immature patients represent a very different entity, which we argue merits a separate analysis.
Data collection
Patient data were recorded, including age, sex, time from injury to presentation, type of immobilization, radiograph result (normal/bicortical fracture/isolated tubercle fracture), CT classification (displacement < 1 mm, 1 to 2 mm, > 2 mm), MRI classification (displacement < 1 mm, 1 to 2 mm, > 2 mm), long axis measurement (LAM) (for all bicortical fractures visible on radiographs), initial treatment, incidence of acute surgery (defined as surgery within four weeks of injury), and the incidence of nonunion. The LAM has been described and validated previously.14
Patient characteristics
The patient characteristics are shown in Table I.
Table I.
Characteristics of patients by outcome complications (nonunion and union), excluding patients who presented their injuries 14 days after the injury and had initial treatments except casting (n = 502).
| Variable | Nonunion (n = 50) | Union (n = 452) | Overall (n = 502) |
|---|---|---|---|
| Sex, n (%) | |||
| Female | 11 (22) | 139 (31) | 150 (30) |
| Male | 39 (78) | 331 (69) | 352 (70) |
| Age, yrs | |||
| Median (IQR) | 30 (23 to 36) | 29 (22 to 49) | 29 (22 to 48) |
| Range | (16 to 75) | (16 to 86) | (16 to 86) |
| Radiograph taken | |||
| Normal, n (%) | 5 (10) | 139 (31) | 144 (29) |
| Bicortical, n (%) | 45 (90) | 313 (69) | 358 (71) |
| Mean radiograph LAM (SD) | 0.48 (0.11) | 0.52 (0.10) | 0.51 (0.10) |
| Acute scan taken, n (%) | |||
| CT scan | 27 (54) | 162 (36) | 189 (38) |
| MRI | 9 (17) | 244 (54) | 253 (50) |
| Not done | 14 (28) | 46 (10) | 60 (12) |
| Acute scan results, n (%) | |||
| < 1 mm | 21 (42) | 345 (76) | 366 (73) |
| 1 mm to 2 mm displaced | 8 (16) | 48 (11) | 56 (11) |
| > 2 mm | 7 (14) | 13 (3) | 20 (4) |
| Not done | 14 (28) | 46 (10) | 60 (12) |
| Fracture detection, n (%) | |||
| Radiograph | 45 (90) | 313 (69) | 358 (71) |
| MRI scan | 5 (10) | 139 (31) | 144 (29) |
LAM, long axis measurement.
Outcomes
The outcome definition was informed by the SWIFFT trial.15 Union was defined as more than 50% bony bridging on CT scans. Nonunion was defined as no bony bridging on CT after a minimum of six weeks of immobilization or no progression towards union on serial imaging beyond three months, as detailed in SWIFFT.12,15 All nonunions diagnosed within the first six months had undergone CT assessment. CT scans were routinely obtained at six weeks in adults. Patients with equivocal CT scan results were followed up with serial CT scans until a clear diagnosis of nonunion or union was reached. Patient records were searched at a minimum of 12 months following presentation to ensure that all early re-presentations and complications were captured.
Statistical analysis
We described patient characteristics, overall and by union outcome (union and nonunion). We also described the characteristics of excluded patients undergoing acute surgery. All continuous data were plotted using histograms to check for normality. We reported mean and SD, median and IQR, and numbers and percentage, as appropriate. Univariable and multivariable logistic regression models were used to assess the associations between patient characteristics and the nonunion outcome. The linearity of relationship between age and risk of nonunion was assessed.
If a non-linear relationship was found, age was modelled using restricted cubic splines,16 and for interpretability, we calculated the odds ratio (OR) of nonunion per ten-year increase in age.
The acute scan values of 1 mm to 2 mm and > 2 mm displacements of the scaphoid were changed into ≥ 1 mm for the regression analysis due to the small number of patients in the > 2 mm group. A sub-group analysis was performed on patients that had a LAM recorded, given the scientific plausibility for more proximal fractures having a higher risk of nonunion.
As this study is part of a service evaluation and descriptive by nature, no sample size calculation was conducted, and sample size was determined by the available data within the study period. Statistical analysis was carried out using R statistical software package v. R-4.4.3 (R Core Team; R Foundation for Statistical Computing, Austria).
Results
We extracted data for 659 patients with an available CT scan assessment of scaphoid union. Of these, 17 patients that did not present their injuries within two weeks of injury, 59 that had an acute surgery, and 81 patients aged < 16 years, were excluded. A total of 502 patients were included in the final analysis (Figure 1).
Fig. 1.
Flow diagram for patients included in the final analysis.
Of 502 patients, 50 developed nonunion (10%), of which 39 (78%) were male and 45 had a bicortical fracture on radiograph. The median age of patients with nonunions was 30 years (IQR 23 to 36; Table I).
Those undergoing acute surgery were mostly young adult males with more proximal and more displaced fractures than those in the cast group. The characteristics of the excluded patients undergoing acute surgery are detailed in Supplementary Table ii.
Regression analysis
The relationship between patient age and the risk of nonunion was not linear (Supplementary Figure a) and was modelled using restricted cubic splines with three knots. The multivariable analysis presented in Table II shows male patients were at 41% higher risk of nonunion compared with female patients. Radiograph detected fractures had an almost two-times higher nonunion risk than those with normal x-rays (OR 1.97, 95% CI 0.75 to 6.18).
Table II.
Univariable and multivariable analysis with restricted cubic splines having three knots for age (yrs) for the risk factor of nonunion. Total development sample size for multivariable analysis (n = 442).
| Univariable analysis | Multivariable analysis | |||
|---|---|---|---|---|
| Variable | Odds ratio | 95% CI | Odds ratio | 95% CI |
| Age, yrs | ||||
| Age (spline 1) | Reference | Reference | ||
| Age (spline 2) | 1.07 | (1.00 to 1.14) | 1.08 | (1.00 to 1.18) |
| Age (spline 3) | 0.83 | (0.71 to 0.96) | 0.76 | (0.59 to 0.94) |
| Sex | ||||
| Female | Reference | Reference | ||
| Male | 1.57 | (0.81 to 3.31) | 1.41 | (0.58 to 3.98) |
| Radiograph taken | ||||
| Normal | Reference | Reference | ||
| Bicortical | 4.00 | (1.70 to 11.7) | 1.97 | (0.75 to 6.18) |
| Scan values | ||||
| < 1 mm | Reference | Reference | ||
| ≥ 1 mm | 4.04 | (1.95 to 8.23) | 3.10 | (1.42 to 6.71) |
Restricted cubic spline knots positions are aged 38, 55, and 62 years.
Further, the larger the displacement, the higher the risk of nonunion – patients with displacement of ≥ 1 mm are at over three times higher risk for nonunion (OR 3.10, 95% CI 1.42 to 6.71) than < 1 mm displacement. Note that the multivariable analysis included 442 partcicipants as displacement could not be accurately assessed in the 60 participants who did not undergoe initial cross-sectional imaging.
Using the ORs estimated from restricted cubic splines for age in the multivariable regression analysis, we plotted the risk of nonunion over age in Figure 2. We found an increased risk of nonunion per year increase in age between 16 and 38 years, which then changes to a reduced risk of nonunion per year increase in age between 38 and 55 years. Though the reduced risk of nonunion continues per year increase in age after 55 years, the rate of reduced risk decreases.
Fig. 2.
The effect of age for the risk of nonunion scaphoid fracture using restricted cubic splines with three notes and three cutting points at the age of 38, 55, and 62 years for linearity of risk distribution. The blue line is the point estimate and green dash lines are the cutting points.
Table III presents the non-linear relationship between age and risk of nonunion and provides ten-year age band adjusted odds ratios for nonunion scaphoid fractures. Risk of nonunion increases by 20% for patients aged 36 years compared with those aged ten years younger (26 years) (OR 1.20, 95% CI 1.09 to 1.33); however, all other ten-year ages showed reduced risk of nonunion when compared with patients aged 26 years, for example patients aged 16 or 46 years had 52% and 27% lower risk of nonunion, respectively, than patients aged 26 years.
Table III.
Adjusted odds ratio of nonunion scaphoid fractures over ten-year age intervals.
| Age, yrs | Odds ratio | 95% CI |
|---|---|---|
| 26 | Reference | Reference |
| 16 | 0.48 | (0.34 to 0.69) |
| 36 | 1.20 | (1.09 to 1.33) |
| 46 | 0.73 | (0.60 to 0.89) |
| 56 | 0.29 | (0.15 to 0.57) |
| 66 | 0.09 | (0.02 to 0.42) |
| 76 | 0.03 | (0.00 to 0.33) |
| 86 | 0.01 | (0.00 to 0.27) |
Sub-group analysis
Of the 502 included patients, 350 patients had a recorded LAM. A description of these patients is presented in Supplementary Table i. A sub-group analysis of these patients is presented in Table IV which found similar but slightly reduced strengths of the estimated odds ratios compared with the main analysis.
Table IV.
Univariable and multivariable analysis with restricted cubic splines having three notes age (year) for risk factors of nonunion scaphoid fractures with sub-group analysis for radiograph long axis measurement. Total sample size used for multivariable model development (n = 290).
| Univariable analysis | Multivariable analysis | |||
|---|---|---|---|---|
| Variable | Odds ratio | 95% CI | Odds ratio | 95% CI |
| Age, yrs | ||||
| Age (spline 1) | Reference | Reference | ||
| Age (spline 2) | 1.09 | (1.00 to 1.02) | 1.09 | (0.99 to 1.20) |
| Age (spline 3) | 0.98 | (0.97 to 0.99) | 0.81 | (0.64 to 0.98) |
| Sex | ||||
| Female | Reference | Reference | ||
| Male | 1.35 | (0.67 to 2.98) | 1.34 | (0.51 to 4.24) |
| Radiograph LAM | 0.02 | (0.01 to 0.50) | 0.15 | (0.001 to 8.00) |
| Scan values | ||||
| < 1 mm (ref) | Reference | Reference | ||
| ≥ 1 mm | 6.84 | (2.32 to 19.2) | 3.03 | (1.39 to 6.62) |
Restricted cubic splines positions: age at 31, 38, 57 years.
LAM, long axis measurement.
Risk of nonunion decreased by 85% per unit increase of radiograph LAM (OR 0.15, 95% CI 0.001 to 8.00). The displacement of the scaphoid fracture by ≥ 1 mm showed increased risk of nonunion compared with< 1 mm displacement (OR 3.03, 95% CI 1.39 to 6.62).
Discussion
We found two key risk factors for scaphoid nonunion: age and fracture displacement. The risk of nonunion tripled when fracture displacement was ≥ 1 mm compared with < 1 mm displacement. The highest risk of nonunion was found in young skeletally mature adults.
These findings are particularly meaningful given the inconsistent conclusions of previous smaller studies. For example, Desai et al6 concluded that the radiological features of acute scaphoid fractures cannot be used to predict the likelihood of fracture union. However, Bhat et al7 concluded that fracture displacement did appear to be a predictor of nonunion. Notably, Bhat et al7 acknowledged that the small number of nonunions did limit the power of their study, which appears a common theme to many previous cohort studies. Geoghegan et al8 also demonstrated a higher nonunion rate of 31% in displaced fractures compared with 2% in non-displaced fractures; again the number of patients followed up is an acknowledged limitation of this study. However, a body of evidence is emerging, demonstrating that more proximal scaphoid nonunions have a higher risk of failing to unite after nonunion surgery.17,18
The strengths of this study include the large number of patients relative to previous studies, the routine use of serial CT scans in the assessment of union, and the inclusion of ‘occult’ scaphoid fractures detected on MRI. The limitations of the study include the selection bias because of excluding those patients who underwent acute surgery and those with delayed presentations. The higher risk of nonunion with delayed presentation has been well described previously; it is therefore important to note that the study findings do not apply directly to those with a delayed presentation. We have therefore included the descriptive data of those patients who underwent acute surgery so that this important caveat is not ignored. There is an obvious predominance of young adult males with more proximal and more displaced fractures in this acute surgical group, meaning that relative risk is likely to have been underestimated in this young adult male group with more proximal fractures. We would therefore note extreme caution in interpreting the results with regards to the age risk factor; it is our opinion that the skeletally mature age group from aged 16 to 25 years is likely extremely high risk and this has been underestimated (though the direction of the effect is as would be expected).
As the study was a part of a service evaluation, no sample size was conducted, and the sample size was dictated by the number of records available during the study period. However, though this is larger than previous studies in this research area, it may still be underpowered to detect the effect for each risk factor. For example, several risk factors including sex, radiograph detection, and fracture location showed trends towards influencing nonunion risk but were not statistically significant. With greater numbers and more statistical power, it may be that statistical significance would be reached. We conducted this study as an exploratory analysis of risk factors for nonunion of scaphoid fractures and with intent to recommend further research into these risk factors.
In conclusion, the two risk factors for scaphoid nonunion were fracture displacement and age. Further assessment and research could help inform clinical decision making and may help guide future research to improve clinical pathways of care.
Take home message
- The risk factors for scaphoid nonunion are displacement and age. The risk of nonunion tripled when fracture displacement was ≥ 1 mm compared with < 1 mm displacement.
- The relationship between age and nonunion risk is non-linear, with risk steadily reducing with increasing age, and highest in young skeletally mature adults.
- A limitation of this study is that the necessary exclusion patients undergoing acute surgery means that the relative risk is likely to have been underestimated in the young adult male group with more proximal fractures.
Author contributions
B. J. F. Dean: Methodology, Project administration, Visualization, Writing – original draft, Writing – review and editing, Writing – original draft
F. Saedi: Investigation, Methodology, Writing – review & editing
N. D. Riley: Investigation, Methodology, Writing – review & editing
S. Ather: Investigation, Methodology, Writing – original draft, Writing – original draft
B. Tsegaye: Conceptualization, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing
P. Dhiman: Conceptualization, Investigation, Methodology, Writing – review & editing
Funding statement
The author(s) disclose receipt of the following financial or material support for the research, authorship, and/or publication of this article: B. J. F. Dean is part-funded by AO UK, and the publication fee for this work has been funded by AO UK.
ICMJE COI statement
B. J. F. Dean is part-funded by AO UK. S. Ather reports unrelated employment and board membership for RAIQC; grants/grants pending with Lunit, Qure AI, and Deeplook Medical; payment for development of educational presentations for Pfizer; and travel/accommodations/meeting expenses for GE Healthcare and Lunit. N. Riley discloses consulting fees and payments or honoraria for lectures, presentations, speaker bureaus, manuscript writing, or educational events for Acumed, Arthrex, Sovereign, and Medartis/Kerimedical, which are unrelated to this study. B. Tsegaye reports unrelated funding from Cancer Research UK (project grant: PRCPJT-28 Nov21\100021) and funding from the National Institute for Health and Care Research (NIHR) Blood and Transplant Research 30 Unit in Data Driven Transfusion Practice (NIHR203334).
Data sharing
The data that support the findings for this study are available to other researchers from the corresponding author upon reasonable request.
Acknowledgements
Oxford Scaphoid Study Group:
Christopher Little, Oxford University Hospitals NHS Trust, Oxford, UK.
Warren Sheehan, Oxford University Hospitals NHS Trust, Oxford, UK.
Matt L. Costa, Trauma Unit, John Radcliffe Hospital, Oxford, UK.
All collaborators have been involved in Methodology, Project administration, Writing – original draft, and Writing – review & editing.
Ethical review statement
This study was registered locally as a service evaluation project.
Open access funding
The open access fee was funded by AO UK.
Supplementary material
STROBE checklist for observational studies and a supplementary figure.
© 2026 Dean et al. This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives (CC BY-NC-ND 4.0) licence, which permits the copying and redistribution of the work only, and provided the original author and source are credited. See https://creativecommons.org/licenses/by-nc-nd/4.0/
Contributor Information
Benjamin J. F. Dean, Email: bendean1979@gmail.com.
Collaborators: Christopher Little, Warren Sheehan, and Matt L. Costa
Data Availability
The data that support the findings for this study are available to other researchers from the corresponding author upon reasonable request.
References
- 1. de Putter CE, Selles RW, Polinder S, Panneman MJM, Hovius SER, van Beeck EF. Economic impact of hand and wrist injuries: health-care costs and productivity costs in a population-based study. J Bone Joint Surg Am. 2012;94-A(9):e56. doi: 10.2106/JBJS.K.00561. [DOI] [PubMed] [Google Scholar]
- 2. Garala K, Taub NA, Dias JJ. The epidemiology of fractures of the scaphoid: impact of age, gender, deprivation and seasonality. Bone Joint J. 2016;98-B(5):654–659. doi: 10.1302/0301-620X.98B5.36938. [DOI] [PubMed] [Google Scholar]
- 3. Zura R, Xiong Z, Einhorn T, et al. Epidemiology of fracture nonunion in 18 human bones. JAMA Surg. 2016;151(11):e162775. doi: 10.1001/jamasurg.2016.2775. [DOI] [PubMed] [Google Scholar]
- 4. Leighton PA, Brealey SD, Dias JJ. Patient experiences of scaphoid waist fractures and their treatment: a qualitative investigation. Bone Jt Open. 2022;3(8):641–647. doi: 10.1302/2633-1462.38.BJO-2022-0042.R1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Dias JJ, Kheiran A, Ngo DN, et al. The onset, progress and factors influencing degenerative arthritis of the wrist following scaphoid fracture non-union. Injury. 2023;54(3):930–939. doi: 10.1016/j.injury.2022.12.025. [DOI] [PubMed] [Google Scholar]
- 6. Desai VV, Davis TRC, Barton NJ. The prognostic value and reproducibility of the radiological features of the fractured scaphoid. J Hand Surg Br. 1999;24(5):586–590. doi: 10.1054/JHSB.1999.0197. [DOI] [PubMed] [Google Scholar]
- 7. Bhat M, McCarthy M, Davis TRC, Oni JA, Dawson S. MRI and plain radiography in the assessment of displaced fractures of the waist of the carpal scaphoid. J Bone Joint Surg Br. 2004;86(5):705–713. doi: 10.1302/0301-620x.86b5.14374. [DOI] [PubMed] [Google Scholar]
- 8. Geoghegan JM, Woodruff MJ, Bhatia R, et al. Undisplaced scaphoid waist fractures: is 4 weeks’ immobilisation in a below-elbow cast sufficient if a week 4 CT scan suggests fracture union? J Hand Surg Eur Vol. 2009;34(5):631–637. doi: 10.1177/1753193409105189. [DOI] [PubMed] [Google Scholar]
- 9. Amirfeyz R, Bebbington A, Downing ND, Oni JA, Davis TRC. Displaced scaphoid waist fractures: the use of a week 4 CT scan to predict the likelihood of union with nonoperative treatment. J Hand Surg Eur Vol. 2011;36(6):498–502. doi: 10.1177/1753193411403092. [DOI] [PubMed] [Google Scholar]
- 10. Smith G, Teng WH, Riley ND, et al. The diagnostic characteristics and reliability of radiological methods used in the assessment of scaphoid fracture union. Bone Jt Open. 6(3):246–253. doi: 10.1302/2633-1462.63.BJO-2024-0211.R1. n.d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Johnson NA, Morris H, Dias JJ. Questions regarding the evidence guiding treatment of displaced scaphoid fractures. J Hand Surg Eur Vol. 2021;46(2):213–218. doi: 10.1177/1753193420977241. [DOI] [PubMed] [Google Scholar]
- 12. Dean BJF, Riley N, Little C, et al. The rate of nonunion in the MRI-detected occult scaphoid fracture. Bone Joint J. 2024;106-B(4):387–393. doi: 10.1302/0301-620X.106B4.BJJ-2023-1171.R1. [DOI] [PubMed] [Google Scholar]
- 13.No authors listed STROBE guidelines. [7 July 2026]. https://www.strobe-statement.org/download/strobe-checklist-cohort-studies-pdf date last. accessed.
- 14. Dean BJF, Riley ND, McCulloch ER, Lane JCE, Touzell AB, Graham AJ. A new acute scaphoid fracture assessment method: a reliability study of the “long axis” measurement. BMC Musculoskelet Disord. 2018;19(1):310. doi: 10.1186/s12891-018-2236-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Dias JJ, Brealey SD, Fairhurst C, et al. Surgery versus cast immobilisation for adults with a bicortical fracture of the scaphoid waist (SWIFFT): a pragmatic, multicentre, open-label, randomised superiority trial. The Lancet. 2020;396(10248):390–401. doi: 10.1016/S0140-6736(20)30931-4. [DOI] [PubMed] [Google Scholar]
- 16. Gauthier J, Wu QV, Gooley TA. Cubic splines to model relationships between continuous variables and outcomes: a guide for clinicians. Bone Marrow Transplant. 2020;55(4):675–680. doi: 10.1038/s41409-019-0679-x. [DOI] [PubMed] [Google Scholar]
- 17. Ramamurthy C, Cutler L, Nuttall D, Simison AJM, Trail IA, Stanley JK. The factors affecting outcome after non-vascular bone grafting and internal fixation for nonunion of the scaphoid. J Bone Joint Surg Br. 2007;89-B(5):627–632. doi: 10.1302/0301-620X.89B5.18183. [DOI] [PubMed] [Google Scholar]
- 18. Ammori MB, Elvey M, Mahmoud SS, et al. The outcome of bone graft surgery for nonunion of fractures of the scaphoid. J Hand Surg Eur Vol. 2019;44(7):676–684. doi: 10.1177/1753193419841278. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings for this study are available to other researchers from the corresponding author upon reasonable request.