Age-Related changes in the morphological features of medial column of the proximal humerus in the Chinese population

Background Age-related changes in the medial column (MC) of the proximal humerus have a major impact on fracture management; however, the changes in the morphological features remain unclear. This study aimed to investigate the age-related changes in the morphological features of MC and present the morphological grading. Methods One hundred computed tomography (CT) images of the proximal humerus of 100 individuals (19–95 years) were retrospectively obtained. The individuals were categorized into five age groups to quantify the differences among different ages; the youngest group (18–44 years) served as the baseline group. Parameters of the morphological features were measured on CT images with multiplanar reconstruction based on an explicit definition of MC, including length, thickness, width, oblique thickness (DSM), humeral head diameter (DHM), and ratio (RSM) of DSM to DHM. The morphological grading of MC was presented based on the value of RSM deviating different standard deviations (SD) from the mean value in the baseline group. Results Significant negative correlations were observed between age and the morphological parameters of MC (r ranged from −0.875 to −0.926; all P < 0.05), excluding DHM (r = 0.081, P = 0.422). Significant differences in the values of morphological feature parameters were detected among the five age groups (all P < 0.001). The highest mean values of morphological feature parameters were observed in the youngest group (18–44 years), which decreased gradually with increasing age until the lowest mean values were observed in the oldest group (≥90 years) (all P < 0.05). The morphological features of MC were categorized into three grades based on the value of RSM deviating 1.5 SD or 3 SD from the mean value in the baseline group. Conclusion Our study shows that the parameter values of morphological features of MC decreased with increasing age. The morphological features of MC could be categorized into three grades. Our findings may provide a more comprehensive insight into age-related changes in the morphological features of MC that facilitate risk stratification and optimize the management of proximal humeral fractures.

Background: Age-related changes in the medial column (MC) of the proximal humerus have a major impact on fracture management; however, the changes in the morphological features remain unclear. This study aimed to investigate the age-related changes in the morphological features of MC and present the morphological grading. Methods: One hundred computed tomography (CT) images of the proximal humerus of 100 individuals (19-95 years) were retrospectively obtained. The individuals were categorized into five age groups to quantify the differences among different ages; the youngest group (18-44 years) served as the baseline group. Parameters of the morphological features were measured on CT images with multiplanar reconstruction based on an explicit definition of MC, including length, thickness, width, oblique thickness (D SM ), humeral head diameter (D HM ), and ratio (R SM ) of D SM to D HM . The morphological grading of MC was presented based on the value of R SM deviating different standard deviations (SD) from the mean value in the baseline group. Results: Significant negative correlations were observed between age and the morphological parameters of MC (r ranged from −0.875 to −0.926; all P < 0.05), excluding D HM (r = 0.081, P = 0.422). Significant differences in the values of morphological feature parameters were detected among the five age groups (all P < 0.001). The highest mean values of morphological feature parameters were observed in the youngest group (18-44 years), which decreased gradually with increasing age until the lowest mean values were observed in the oldest group (≥90 years) (all P < 0.05). The morphological features of MC were categorized into three grades based on the value of R SM deviating 1.5 SD or 3 SD from the mean value in the baseline group. Conclusion: Our study shows that the parameter values of morphological features of MC decreased with increasing age. The morphological features of MC could be categorized into three grades. Our findings may provide a more comprehensive insight into age-related changes in the morphological features of MC that facilitate risk stratification and optimize the management of proximal humeral fractures.

Introduction
The medial column (MC) is considered the osseous region of the posteromedial metaphysis of the proximal humerus (PH), which plays a crucial role in fracture management (1)(2)(3). MC reportedly develops remarkable changes with increasing age owing to the development of osteopenia or osteoporosis (4, 5). These age-related changes in MC may be responsible for the increased risk of proximal humeral fractures (PHFs) and higher fracture severity (1,4,6,7). Furthermore, incompetent MC, i.e., fragile with impaired strength, due to advanced age is also associated with a higher risk of fracture complications (8)(9)(10)(11)(12). Conversely, elderly patients with severe osteopenia treated with integrity restoration of MC could achieve a satisfactory prognosis, similar to that of younger patients (13-16). Therefore, investigating the morphological features of MC and elucidating its age-related changes may be conducive to the management of PHFs. However, the specific definition of MC is inconsistent, and age-related changes in its morphological features remain unclear owing to a lack of studies.
The range of region of interest of MC varied in different studies, ranging between 20 and 40 mm below the humeral head based on subjective interpretations (4, 5, 17,18). The consequent conclusions may be limited given the subjective selection methods. Furthermore, microstructural assessments of the cortical or trabecular bone in the cadaveric bone have been used to analyze the age-related changes in MC; however, these might be insufficient to derive a comprehensive conclusion given the limited sample sizes and differences between ex and in vivo samples (4, 5, 19,20). A study on the in vivo imaging of postmenopausal females reported the regional differences in the cortical bone of MC; however, studies on individuals of other ages remain lacking (18). Moreover, no morphological grading of MC based on its age-related changes-such as the Singh index used for the proximal femur-has been proposed (21)(22)(23). Therefore, this study aimed to investigate the age-related changes in the morphological features of MC using in vivo imaging based on an objective definition of MC and individuals with a broad age range.

Study design
This is a cross-sectional study approved by the Ethics Committee of Chinese PLA General Hospital (No. S2021-021-01) and registered in Chinese Clinical Trial Register (ChiCTR2200059524). We retrospectively reviewed 681 CT images of PH of 628 adult individuals in the Picture Archiving and Communication Systems of our institution between December 2019 and December 2021. The demographic and clinical information was obtained from electronic medical records, and all personal records were anonymized prior to data analysis. Owing to the retrospective and anonymous nature of data collection, the requirement for informed consent was waived.
The CT images were obtained with a 256-slice multidetector CT scanner (Philips Healthcare, Amsterdam, Netherlands) and standardized protocol with high-resolution algorithm (120 KV, 0.675 mm slice thickness, 0.335 mm interlayer spacing, 0.8 mm reconstruction slice thickness, 1.0 mm reconstruction interlayer spacing, and 512 × 512-pixel matrix). For the acquisition of in vivo CT images of normal intact PHs, the CT images from individuals older than 18 years of age were included. Axial scanning ranged from more than 3 cm superior to the acromion to more than 3 cm inferior to the deltoid tuberosity. The exclusion criteria were as follows: (1) Prior fracture or surgery in PH; (2) bone diseases and skeletal abnormalities, including osteoarthritis, rheumatism, dysplasia, and deformity; (3) diagnosis of metabolic diseases or receiving treatment that could affect bone metabolism, such as Paget's disease and primary hyperparathyroidism; (4) diagnosis of cancer or other malignant diseases; and (5) history of smoking. Thus, 261 CT images of normal PHs from 261 individuals were obtained.
To further clarify and quantify the age-related changes in the morphological features of MC, age was categorized into five groups (24)(25)(26): Group I (18-44 years), Group II (45-59 years), Group III (60-74 years), Group IV (75-89 years), and Group V (≥90 years). Group I served as the baseline group. As the age distribution of individuals was skewed, stratified sampling was used to ensure equal sample sizes among the age groups to avoid the influence of different sample sizes among age strata on further morphological analysis. In total, 100 CT images from 100 individuals were included for analysis (20 individuals per stratum, including 10 women and 10 men; Figure 1). The mean age of individuals included in the study was 64.94 ± 21.22 years (ranges 19-95 years; 64.74 ± 21.22 years for women, and 65.14 ± 21.52 years for men).

Definitions of parameters for morphological features of the mc
CT images with multiplanar reconstruction in this study were acquired using RadiAnt DICOM Viewer (version 4.6.5; Medixant, Poznan, Poland). Previous studies showed that the changes in cortical bone and trabecular bone both have an impact on the medial supporting role of MC, and the changes in trabecular bone might be an early sign of a decrease in the mechanical properties of MC (2,4,5,18,27). Thus, MC was defined as a complex osseous region in the medial metaphysis in this study, comprising the endosteal longitudinal trabecular bone and adjacent cortical bone; the endocortical longitudinal trabecular region was used as a reference ( Figure 2). The parameters of the morphological features of MC, defined according to previous studies, were as follows ( Figure  The measurements of parameters in all individuals were obtained independently by two orthopedic researchers (ZC and WZ) with experience in CT imaging analysis. The intraobserver reliability of the measurement parameters was assessed by repeating the measurements of parameters in all individuals twice randomly by the same researcher (ZC), at least 4 weeks apart. In addition, the measurements of parameters in all individuals were obtained independently by the second researcher (W.Z.) to assess the interobserver reliability of the measurements of parameters. All measurements of parameters in all individuals were obtained thrice, and the average values were taken for final analysis to avoid researcher bias.
The morphological features of MC were graded using R SM , as R SM was presented as a ratio that facilitates comparison of the morphological features of MC with different PH sizes. The morphological features of MC were graded based on the value of R SM deviating different standard deviations (SD) from the mean value of R SM in the baseline group (Group I), which was referred to as the morphological grading method of vertebral compressive fractures and the categorization method of bone mineral density (29, 30).

Statistical analysis
The intra-and interobserver reliability of the morphological parameters were assessed using the intraclass correlation coefficient (ICC). The threshold for excellent correlation was set at 0.75 (31). The Shapiro-Wilk test was used as the normality test of continuous variables. The correlation between age and morphological parameters was assessed using Spearman's correlation coefficient. Correlation strength was assessed to be strong at r > 0.7, moderate at 0.7 > r > 0.3, and weak at r > 0.3 (32). The mean values of normally distributed continuous variables within the groups were compared using one-way analysis of variance, followed by the least-significant difference (LSD) post-hoc test for pairwise comparisons. Non-normal distribution data were analyzed using Kruskal-Wallis H test; Bonferroni correction was used in the pairwise comparisons. Categorical variables were presented as constituent ratios and analyzed using the χ 2 test or Fisher's exact test. All statistical  A priori power analysis (α = 0.05, β = 0.2, 2-tailed) was performed using PASS (version 15.0.5; NCSS, Kaysville, United States) to achieve a medium to large correlation coefficient (ρ ≥ 0.3), which was expected for the correlation between the measured parameters and age. A minimum sample size of 84 individuals was required in the present study.

Reliability of the measurement parameters of morphological features of MC
The intra-and interobserver reliability showed almost perfect agreement for all measurement parameters of the morphological features in the present study (ICC ranged from 0.814 to 0.952; all P < 0.001). Detailed results are presented in Table 1.

Age-related changes in morphological features of MC
Significant negative correlations were observed between age and the value of most parameters (including D SI , D LM , D AP , D SM , and R SM [all P < 0.001], but not D HM [P = 0.422]). The parameter values of the morphological features for the study individuals, and correlations between age and parameters, are provided in Table 2 and Figure 3. Difference in morphological features of medial column in five age groups is showed in Figure 4. Comparisons among multiple groups revealed differences in the parameter values of the morphological features of MC among the groups (including D SI , D LM , D AP , D SM , and R SM , all P < 0.001; Table 2). No significant difference was observed in D HM among the groups (P = 0.921; Table 2). Pairwise comparisons revealed that the highest mean parameter values of the morphological features were observed in the youngest group (18-44 years), while the oldest group (≥90 years) had the lowest mean parameter values (including D SI , D LM , D AP , D SM , and R SM ,

Morphological grading of MC
The highest mean value of R SM was observed in the baseline group (R SM = 51.17 ± 7.22% in Group I), and it was set as the reference value for morphological grading. Between-grades comparisons were performed to select the effective thresholds, which could distinguish between the different morphological grades. The threshold values of grading were set at different times of SD below the mean value of R SM in the baseline group (from 0.5SD to 4SD with 0.5SD interval). Finally, 1.5SD and 3SD below the mean value of R SM in Group I (R SM = 40.34%, R SM = 29.51%) were selected as the thresholds for the morphological grading of MC. The specific morphological grading was illustrated as follows: (1) Grade I: R SM > Mean -1.5SD (R SM >40.34%); (2) Grade II: Mean -1.5SD ≤ R SM ≤ Mean -3SD (29.51% ≤ R SM ≤ 40.34%); and (3) Grade III: R SM < Mean -3SD (R SM <29.51%). There were significant differences in the value of morphological features parameters of MC among the three grades (including D SI , D LM , D AP , D SM , and R SM , all P < 0.05; Table 4), while no significant difference was observed regarding sex (P = 0.329; Table 4).

Discussion
This study investigated the age-related changes in the morphological features of MC and presented morphological grading based on an objective definition of MC that covers a broader age range. We identified and quantified the age-   Frontiers in Surgery metaphysis without elucidating the anatomical rationale. However, the definitions of MC in the aforementioned studies varied by using certain distances defined subjectively as the references rather than the morphological features of MC. Additionally, the cortical or trabecular bone alone may not fully account for age-related changes in MC, as MC is affected by cortical and trabecular bone loss, which may be an early sign of a decrease in the mechanical properties of MC (4, 5, 18,27). Therefore, using the endocortical longitudinal trabecular region as a reference, we defined MC as the osseous region in the medial metaphysis combining the endocortical trabecular bone and adjacent cortical bone, that is dynamic with age. This definition, which is explicit with a reasonable anatomical rationale, could comprehensively reflect the age-related changes in MC.
In the present study, all values of the parameter of the morphological features of MC decreased with increasing age. Compared with younger individuals (18-44 years), older individuals (≥60 years) exhibited more pronounced decrease in the parameter values of the morphological features, especially in the advanced age population (≥90 years). Our findings specifically quantify the age-related changes in MC regarding the morphological features and support a previous study that showed a considerable loss of cortical and trabecular bone of the metaphysis with increasing age (5). This finding was also indirectly supported by a previous histomorphometric study, which showed that the bone density of the trabecular bone in the medial metaphysis decreased significantly in osteoporotic individuals. Additionally, we observed that parameter value decrease 'of the morphological features tended to flatten after the age of 75 years. Similarly, Helfen et al. (5) showed that decreases in the values of the microstructural parameters of the metaphysis were not visible after the age of 80 years.
Morphological grading based on morphological feature changes could provide important references for fracture management. The Singh index-which describes the morphological changes in the trabecular bone of the femoral neck and head-has an important influence on risk stratification and prognosis prediction of fracture (21,22,33). Previous studies  Significant differences in the parameters of the morphological features were detected between Grade I and Grade II (D SI , P = 0.010; D LM , P = 0.013; D AP , P = 0.024; D SM , P = 0.006; R SM , P = 0.003; All P values were adjusted with Bonferroni correction). Age and D HM are presented as mean ± standard deviation (95% confidence interval); the other parameters of morphological features (including D SI , D LM , D AP , D SM , and R SM ) are presented as median (25% quartile, 75% quartile).  (33) demonstrated that a Singh index ≤ 3 was a risk factor for in-hospital mortality (OR = 5.00). In this study, we presented morphological grading of MC which was modeled after the grading of bone mineral density and vertebral compressive fractures (29, 30), using the mean value of R SM of young adults (Group I: 18-44 years) as the reference. MC was categorized into three grades: < 1.5 SD (R SM > 40.34%), 1.5 SD -3.0 SD (29.51% ≤ R SM ≤ 40.34%), and > 3 SD (R SM < 29.51%) below the mean value of R SM of young adults. Between-grades comparison verified the validation of the morphological grading preliminarily. Our study has some limitations; first, the morphological grading presented in our study was based on the Chinese population with a limited sample size; thus, further validation is required to extend the validity of its results to other races. Second, the measurements of morphological parameters were obtained from the CT images acquired with only one CT scanner and scanning protocol. The clinical application of our method of measuring parameters needs to be verified using CT images acquired with different scanning protocols. However, this did not hinder the feasibility of the measurement method and comprehensive insight into MC in this study. Lastly, additional biomechanical and clinical studies are needed for further validation; however, this is beyond the scope of this study.

Conclusion
This study presents the age-related changes in the morphological features of MC of PH and morphological grading, based on an explicit definition of MC. Our results show that the parameter values of the morphological features of MC decreased with increasing age. Thus, the morphological features of MC could be categorized into three grades. This study may provide a more comprehensive insight into age-related changes in the morphological features of MC that can facilitate risk stratification and optimize the management of PHFs.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement
Owing to the retrospective and anonymous nature of this study, the requirement for informed consent was waived by the Ethics Committee of Chinese PLA General Hospital (No. S2021-021-01).

Author contributions
ZC writing -review and editing, formal analysis, collection and analysis of the data, and study design. ZZ and WZ formal analysis, collection and analysis of the data, and study design. HC and YL designed the study, supervised and directed the study. PT supervised and directed the study. All authors contributed to the article and approved the submitted version.