Epidemiology and Incidence of Upper Limb Fractures: A UK Level 1 Trauma Center Perspective

Introduction Understanding the incidence and epidemiology can inform clinicians and policymakers about the population's needs. Our study reports on upper limb fractures treated at a major trauma center over 7.5 years. Methods We collected data on fracture locations, age, gender, Charlson Comorbidity index (CCI), and treatment options of all upper limb fractures treated at a Level I Trauma Centre from January 1, 2015 to June 30, 2022. Humerus, radius, and ulna fractures were each classified as proximal, diaphyseal, and distal. Results About 9,915 patients sustained 12,790 fractures, given an overall incidence of 303.2 fractures per 100,000 patients per year. The most common fracture site was the distal radius (60.1 fractures per 100,000 patients per year, whereas carpal and metacarpal bones had the lowest incidence. The mean age and CCI were 46.4 years and 1.54, respectively. 58.1% of patients were male. All bone fractures distal to the elbow were associated with an age younger than the mean (all p<0.001), with humerus fracture patients having the oldest mean age (54.6 years). Compared to the mean gender ratio, except for ulna (no association), humerus (55% female), and radius (51% female), all other locations showed significantly higher incidences of males (all p<0.001). When plotting the incidence based on the age of injury, the entire cohort, along with radius and ulna fracture subgroups, demonstrated a bi-peak distribution. This pattern revealed that younger males and older postmenopausal females had the highest incidence rates. Conclusion To our knowledge, this represents the first study of this type in the UK since 2006. We sought to elucidate relative incidence and demographic associations with fractures to highlight changing population needs and allow policymakers and services at a regional and national level to operate with up-to-date information.


Introduction
Understanding the demographics, patterns, and underlying trends in fractures in the population allows for a better understanding of the current and evolving needs of the population.There has yet to be a recent, large UK-based study on the epidemiology of upper limb fractures.The latest research paper that delved into the epidemiology of upper limb fractures in the UK dates back to 2006 [1]; however, this paper also examined fractures in other locations.A more recent study in 2015 explored fracture epidemiology in the UK, but it categorized fractures based on generalized body areas rather than specific upper limb fractures [2].Notably, both papers only covered a one-year time span.Moreover, a significant disparity exists in the reported incidence of fractures in the UK, ranging from 98 to 132.7 per 10,000 population [3].The underlying reasons for this considerable fluctuation still need to be determined, although they may stem from variations in data sources.For instance, orthopedic fracture clinic data frequently fails to capture all fractures since not all fractures necessitate referral there.Conversely, data from Accident and Emergency departments might lead to overestimating fracture incidence due to potential over-diagnosis of certain types of fractures [1].Additionally, the methodology employed to calculate the overall population for determining incidence rates contributes to the observed variations in fracture incidence.
Analyzing data that delineates the epidemiological patterns of fractures holds significant importance, as it provides valuable insights to clinicians and policymakers regarding the population's requirements.This, in turn, aids in the strategic planning of healthcare delivery.Such an analysis can also unveil which genders are susceptible to specific fractures, along with correlations involving age, BMI, and comorbidities in relation to fracture patterns.Notably, the demographic analysis aspect gains heightened significance within aging populations, as seen in the case of the UK, where it helps pinpoint fractures that could be categorized as osteoporotic.It is noteworthy that in 2019, osteoporotic fractures accounted for approximately 2.4% of healthcare expenditures in the UK [4].The identification of osteoporotic fractures and individuals with an elevated risk of fracture could play a pivotal role in fracture prevention, subsequently leading to a reduction in the overall expenses associated with osteoporotic fractures within the UK healthcare budget.
To comprehensively define the epidemiology of upper limb fractures, we conducted a retrospective study encompassing a 7.5-year timeframe, focusing on upper limb fractures treated at a Major Trauma Center within the UK.

Materials And Methods
Situated in the Cambridgeshire region, Addenbrooke's Hospital is a Major Trauma Centre catering to a population of around 570,000 patients.This study used the Epic Systems TM database to identify patients who visited Accident and Emergency (A&E) and orthopedic outpatient departments.Inclusion criteria encompassed patients with specific upper limb fracture diagnoses, as indicated by ICD-10 clinical coding.
The study's timeline extended over seven years, from January 1, 2015 to June 30, 2022.All fractures located distal to and including the scapula and clavicle were included.Furthermore, each fracture category underwent a subsequent classification based on whether it was an open or closed fracture type.The scope of included bones comprised the clavicle, scapula, humerus, ulna, radius, carpal, metacarpal, and phalanges.The humerus, ulna, and radius were subjected to further subdivision, and categorized into their proximal, shaft, and distal segments.Instances, where multiple fractures of varying bone types occurred within the same patient and admission, were treated as distinct occurrences.
Demographic features such as age at admission, Body Mass Index (BMI), gender, ASA grade (for surgically managed patients), and Charlson Comorbidity Index (CCI) score were calculated for each patient.The study then analyzed the demographic breakdown for each fracture sublocation and type.Graphs were generated to display the frequency of fractures for each bone location in relation to patient age, and separate graphs for each gender to identify trends in the fracture data.Additionally, the quarterly frequency of fractures over the seven-and-a-half-year period was calculated and graphed, along with location subtype graphs, to examine the impact of the COVID-19 pandemic and subsequent lockdowns on acute upper limb injuries.Furthermore, an in-depth analysis was conducted to investigate potential differences in patient demographics and the percentage of surgically managed cases over time.A one-way ANOVA (analysis of variance) was performed, treating each year (2015-2022) as a separate group.This statistical test allows for comparisons between multiple groups simultaneously, similar to a T-test comparing two groups.A significance level of p<0.05 was considered statistically significant.
Variables such as the percentage of surgically managed cases, patient age, gender, ASA grade, BMI, and Charleston Comorbidity Index were compared across the years.If a variable showed significant differences (p<0.05),post-hoc analysis using the Bonferroni post-hoc test with Bonferroni correction was conducted to identify specific significant comparisons.This test ensures that when there are multiple T-tests performed across an array of data groups, the p-value threshold is accordingly adjusted and reduced, to ensure the increase in type I error inherent to multiple statistical tests is offset.The manuscript includes means for each investigated category in each year to provide context for significant findings, ANOVA significance levels, and post-hoc test results displayed in tables.Additional supplementary tables present mean differences, standard deviations/errors, and confidence intervals.Data analysis and graph construction were performed using SPSS v28 (IBM Corp., Armonk, NY) and Microsoft Excel TM .

Results
A total of 12,790 fractures were recorded from 9,915 patients over the study's time frame.In our study, the average age was 48.2 years, with 56.4% of patients being male.The overall incidence of upper limb fractures was 303.2 fractures per 100,000 patients per year.This value is considerably lower than quoted in previous studies [1] and is discussed in the limitations.

Location
The breakdown by location and open vs. closed nature of fractures are displayed in Table 1.The most common fracture site was the distal radius, with 20.1% of all fractures being of that location, representing an incidence of 60.1 distal radial fractures per 100,000 patients per year.Carpal and metacarpal bones were of lowest incidence, with the humerus and ulna being roughly similar in incidence, at 18.6% and 18.8%, respectively, when combining fractures occurring proximally, at the shaft, and distally.Phalangeal fractures were the most likely to be open injuries, with 50.7% and 59.8% of thumb and other phalangeal fractures presenting in this fashion, whereas only

Demographic breakdown
The demographic breakdown of each fracture location and subtype is displayed in Table 2.The oldest average age of fracture by location was proximal humeral (69.83 years old), with the youngest average age for the radial shaft (24.42 years old) followed by the ulnar shaft (26.51 years old).Most locations showed a higher incidence among males, with the first metacarpal being the most skewed (84% male), whereas the proximal humerus had the highest incidence of females (63% female).The BMI measurements across different locations showed very little overall trend in this population.For P-values and standard deviations, see Supplementary Table A.

ASA grade, surgical management
During the study, 3,993 surgeries were performed on 12790 fractures.Surgeries included external fixation, intramedullary nails, or plate fixation.The shaft of the ulna and radius had the highest number of operatively treated fractures, while metacarpal and phalangeal fractures had the lowest surgical rates.

Charlson comorbidity index
Finally, we calculated each patient's CCI.We found that patients with fractures of the shaft of the ulna and radius showed the lowest CCI means, indicating younger and more active patients sustaining these location fractures.In contrast, patients with proximal humerus fractures were the most frail (highest CCI).
Fracture locations and the entire cohort were graphed by age and gender in Figures

Upper limb fractures over time
The incidence of each fracture sub-location by year is shown in Figure 3, and the overall incidence by quarter is shown in Figure 4.For ages, 2020 and 2021, patients had significantly higher ages compared to 2015.For ASA grades, 2019, 2020, and 2021 patients had higher grades compared to 2015 and 2016 patients.For CCI score, 2020 and 2021 patients had greater comorbidities compared to 2015, 2017, and 2018.

Discussion
Our study investigated all upper limb fractures presenting at a major trauma center over the span of seven years.We sought to elucidate demographic trends and patterns, sub-location breakdown insights, and temporal differences.The study showed that the distal radius was the most common site for fractures; however, the shaft of the ulna and radius had the highest number of operatively treated fractures.A 2006 paper by Court-Brown et al. noted that distal radial fractures were the most prevalent, whereas scapula fractures had the lowest incidence [1].Similarly, a study conducted in 2009 in the US found radial fractures to be the most frequent, while clavicle fractures were the least common [6].These large-scale epidemiology studies indicate a consensus on the most prevalent upper limb fracture.However, the least common upper limb fracture remains to have a concurrence.Fractures of the radius and ulna are often the result of highenergy trauma, such as road traffic accidents in young individuals and adults.Conversely, low-energy trauma is a common cause of forearm fractures, usually attributed to compromised bone quality.Pediatric forearm fractures seldom require surgical intervention, unlike adults who frequently need reduction and immobilization prior to surgery to improve patient outcomes [7].These observations support the high surgery rates for radius and ulna shaft fractures.
The American Society of Anesthesiologists (ASA) physical status grading predicts operative risk by considering patients' comorbidities [8].Generally, higher ASA grades were associated with proximal fractures at presentation or during the initial operation, with the scapula having the highest average grade of 3.00.While limited literature directly links ASA grade to fracture location, it is worth noting that ASA grade can independently forecast the likelihood of readmission after a fracture [9].

Body Mass Index
Our BMI measurements across different locations showed very little overall trend in this population.However, obesity is a growing issue in the UK, with estimates projecting that the percentage of the population in England who are morbidly obese (BMI > 40) will reach 8% in 2035 [10].A previous study found a positive relationship between obesity and proximal humerus fractures in males and a positive relationship between BMI and shaft of humerus fractures in females.A negative relationship was found between BMI and clavicle fractures in both genders.However, the paper concludes that there are likely to be many factors involved in fracture epidemiology, of which BMI is a minor factor [11].This sentiment is echoed in a study from 2020, showing that there is an increased incidence of humeral fractures in obese patients compared to those of the wrist in the same patient group [12].

Charlson Comorbidity Index
Proximal humeral occurred in the eldest age demographic with the highest CCI means, with the radial and ulnar shaft most commonly occurring in the third decade of life for patients.Despite this, CCI was shown in a previous study to be a poor predictor in calculating the risk of hip fractures and osteoporotic fractures.A new Charlson Fracture Index (CFI) was created to predict hip fractures in men and women.This tool could identify patients at an increased risk of hip fracture [13].It would be intriguing to explore whether the CFI could serve as a predictive measure for upper limb fractures, potentially forming the basis for future research endeavors.Despite earlier indications that CCI is not a robust predictor for calculating fracture vulnerability, it is important to highlight that CCI has demonstrated efficacy as a predictor of mortality, particularly in cases of proximal humerus fractures [14].

Age, Gender, and Upper Limb Fracture Patterns
The overall graph and graphs for the humerus, ulna, and radius fractures show gender-specific peaks for younger males and older females.The rise in the incidence of fractures in females occurs around menopause, reflecting the increased fracture risk due to estrogen deficiency [15].The peak incidence of fractures in younger males aligns with Court-Brown et al.'s findings, although they also note a gradual increase after age 60 [1].This initial peak in males likely stems from younger men engaging in more physical risks and activities, leading to a higher likelihood of experiencing high-energy trauma [16].Similarly, clavicle, scapula, and hand fractures, graphed by age, display distinct shapes within each gender, but the relative incidence between genders differs significantly.Males have a much higher incidence across various age ranges, with a crossover occurring around the 80-90 years old range.This discrepancy may be attributed to the scapula, clavicle, and hand fractures often resulting from traumatic injuries rather than fragility and osteoporosis, unlike humeral, ulnar, and radial fractures.

Fractures and Osteoporosis
The UK has an aging population, and osteoporotic fractures account for 2.4% of UK healthcare spending [4].
Osteoporosis is characterized by "compromised bone strength predisposing a person to an increased risk of fracture" [17].On a cellular level, it can be considered a metabolic bone disease resulting from an imbalance between osteoclastic bone resorption and osteoblastic bone formation, leaving bones weak and fragile, thus increasing the risk of fracture.Factors contributing to the development of osteoporosis include estrogen or vitamin D deficiency, secondary hyperparathyroidism, and even the gut microbiome [17].Traditional osteoporotic fractures were initially thought to be fractures of the thoracolumbar vertebrae, distal radius, proximal femur, and proximal humerus [1].This helps to explain the graph for radial and humerus fractures.
As the shape and relative trends of the ulna fractures are similar to that of the radius and humerus, this could suggest that previously unidentified osteoporotic fractures were responsible.In the future, as the UK population demographics age further, more fractures may be defined as osteoporotic, which will have implications for the prevention, detection, treatment, and management of these osteoporotic fractures.

Upper Limb Fractures Over Time
From 2015 to 2019, upper limb fractures remained relatively constant, however, in 2020 and 2021, fewer fractures were recorded, particularly for ulna and radius fractures.This decrease coincided with the second and third UK COVID lockdowns, November to December 2020 and January to February 2021 [18].Lim et al. reported a similar decrease in fractures but observed an increase in the proportion of hand fractures, which our study did not [19].When splitting each year into quarters, the number of fractures oscillates, with increases in numbers generally seen in the second and third quarters of each year.This pattern could be driven by the summer months.MacDermid et al. found an increase in clavicle, radial head, and other fractures in regions of the upper extremity during the summer months, possibly due to increased participation in outdoor sports and recreational activities due to the warmer weather [20].
Only age, ASA grade, and CCI score reached significance in the ANOVA model.This highlights that there were no significant changes in the percentage of fractures surgically managed, gender, and BMI of patients presenting with upper limb fractures over this time period.The post-hoc analysis using the Bonferroni method showed that 2020 and 2021 patients had significantly higher ages, grades, and comorbidities than previous years.This indicates some limited evidence to support our previous results that over the two pandemic and lockdown-affected years, the patients presenting with fractures were more likely to be older and frailer due to the overall decrease in activity amongst the younger population; however, the lack of widespread statistical significance means this could nonetheless be a spurious finding.Our negative findings of no significant changes over time for surgical management percentage, gender, and BMI also contribute to the literature and give insight into the population investigated.

Limitations and further directions
Our overall incidence of upper limb fractures was considerably lower than reported in previous studies [1].This could be attributed to several factors, mainly the uncertainty surrounding the population served by the trauma center.Although we examined the catchment area of the Level 1 Major Trauma Center, the incidence of fractures reported in this study may be influenced by individuals' decisions on which local A&E to visit and complex ambulance dispatch areas and routes.This limitation was also acknowledged in the 2006 study by Court-Brown et al.A comprehensive fracture epidemiology study covering the entire UK would be necessary to determine the true incidence.However, this would be a highly challenging endeavor, requiring coordination among multiple trusts across the country, but it could provide insights into potential regional differences in fracture incidence.Another factor contributing to our lower reported fracture incidence is the recording of multiple fractures on the same bone and side of the body as a single fracture.For instance, multiple metacarpal fractures or a comminuted proximal humerus fracture would each be counted as a single fracture.
Furthermore, we did not utilize the AO OTA classification, which is commonly used in research but not in clinical practice.Instead, we relied on ICD-10 coding at the time of admission.However, it is worth noting that most epidemiology fracture papers also do not employ the AO OTA fracture classification.We anticipate readers will utilize our results to inform policy decisions and gain insights into population needs and epidemiology rather than conducting further research on fracture types and outcome associations.
A systematic review or meta-analysis examining fracture incidence across different countries would be beneficial to identify additional factors that could contribute to variations in fracture incidence for the same types of fractures.Climate, study methodologies, healthcare structures, and the likelihood of patients seeking emergency care for minor injuries are all potential factors that may impact reported fracture incidence.

Conclusions
Our study sought to report upon, define and analyze the epidemiology of upper limb fractures in the United Kingdom.We have presented a large cohort, in a tertiary trauma center, and the subsequent detailed exploration of trends and associations.Finally, this has elucidated up-to-date information that equips clinicians and policymakers with the tools to meet and tailor healthcare services to the needs of a dynamic population, fostering more effective and responsive healthcare practices.

FIGURE 3 :FIGURE 4 :
FIGURE 3: Incidence of fractures over the seven-year time period, stratified by location

Open vs Closed Number Location Percentage (%) Location Incidence (/100,000 patients per year) Percentage Open (%)
1.38% of clavicular fractures were open.Overall, 17.5% of all fractures were open.

Fracture locations incidence by gender graphed with respect to age FIGURE 2: Forearm fracture locations incidence by gender graphed with respect to age
1, 2, following the methods introduced by Burh et al. in 1959 and later utilized by Court-Brown et al. to characterize fracture patterns [1,5].The overall graph and humerus, ulna, and radius fractures exhibit a pattern where younger males and older females have gender-specific peaks.FIGURE 1:

TABLE 3 : One-way ANOVA analysis of various demographic and management factors across each year included in the study
Only the factors of age, ASA grade, and CCI score reached significance in this model.This highlights that there were no significant changes in the percentage of fractures surgically managed, gender, and BMI of patients presenting with upper limb fractures over this time period.Subsequently, post-hoc analysis with the Bonferroni method was conducted for age, ASA grade, and CCI score, shown in Table4.

TABLE 4 : Post-hoc analysis with the Bonferroni method was conducted for age, ASA grade, and CCI score
*Significance at p<0.05.Please see Supplementary Table C for further breakdown of the Bonferroni method results.

TABLE 5 : Extended to include P-values and standard deviations for the further demographic breakdown of each fracture location and subtype
.B. Table split in two for formatting and ease of reading, the first two columns are repeated.