Abstract
Background
Osteoporosis is characterised by a reduction of bone mineral density (BMD) and predisposition to fracture. Bone microarchitecture, measured by high resolution peripheral quantitative computed tomography (HR-pQCT), has been related to fragility fractures and BMD and has been the subject of large-scale genome-wide analysis. We investigated whether fracture was related to baseline values and longitudinal changes in bone microarchitecture and whether bone microarchitecture was associated with established BMD loci.
Methods
115 males and 99 females (aged 72-81 at baseline) from the Hertfordshire Cohort Study (HCS) were analysed. Fracture history was determined in 2011-2012 by self-report and vertebral fracture assessment. Participants underwent HR-pQCT scans of the distal radius and tibia in 2011-2012 and 2017. Previous fracture in relation to baseline values and changes in tibial HR-pQCT parameters was examined using sex-adjusted logistic regression with and without adjustment for age, sociodemographic, lifestyle and clinical characteristics; baseline values and changes in parameters associated with previous fracture were then examined in relation to four established BMD loci after adjustment for sex and age.
Results
Previous fracture was related to: higher trabecular area (fully-adjusted odds ratio [95% CI] per SD greater baseline value: 2.18 [1.27,3.73], p=0.005); lower total volumetric BMD (0.53 [0.34,0.84], p=0.007), cortical area (0.53 [0.30,0.95], p=0.032), cortical BMD (0.56 [0.36,0.88], p=0.011) and cortical thickness (0.45 [0.27,0.77], p=0.004); and greater declines in trabecular BMD (p=0.001). Associations were robust in sex- and fully-adjusted analysis. Relationships between BMD loci and these HR-pQCT parameters were weak: rs3801387 (WNT16) was related to decline in trabecular BMD (p=0.011) but no other associations were significant (p>0.05).
Conclusion
Baseline values of HR-pQCT parameters and greater decline in trabecular BMD were associated with fracture. Change in trabecular BMD was associated with WNT16 which has been demonstrated to influence bone health in murine models and human genome-wide association studies (GWAS).
Keywords: osteoporosis, epidemiology, fracture, quantitative computed tomography, loci
1.0. Introduction
Fragility fractures are a cause of widespread global concern with a fracture occurring (on average) every 3 seconds (1). There is a strong genetic component to fracture risk, with a history of parental fracture conferring an increased risk of fracture which is independent of bone mineral density (BMD) as measured by dual-energy x-ray absorptiometry (DXA) (2). BMD itself is highly heritable (with an h2 = 50-80%) as shown in twin studies (3) and, in 2012, Estrada and colleagues complied a genome-wide meta-analysis which identified 56 genetic loci which were associated with femoral neck and lumbar spine BMD in individuals with a European or East Asian ancestry (4).
There are fewer data investigating the genetic determinants of bone microarchitecture. In previous studies, other genetic variants in RANK/OPG have been associated with cortical volumetric BMD (5, 6) and SNPs in FMN2/GREM2 genes were associated with trabecular volumetric BMD and fracture risk (6). Furthermore, regions mapped to WNT4, ZBTB40, TNFRSF11B, AKAP11, and TNFSF11 have been associated with lumbar spine vBMD (7).
Compared to DXA, high resolution peripheral quantitative computed tomography (HR-pQCT) provides a more detailed analysis of bone microarchitecture with in-depth measurement of parameters within the trabecular and cortical compartments at the peripheral skeleton. Associations between HR-pQCT parameters and fracture have previously been observed in populations of post-menopausal females in the GERICO (8), CaMOS (9) and OFELY (10) cohorts. However, the relationship between longitudinal change in bone microarchitecture and fracture has not been investigated in a group of community-dwelling older adults comprising males, such as the Hertfordshire Cohort Study (HCS), and neither has an examination of HR-pQCT parameters in relation to established BMD loci been performed.
The aims of this study were: to describe baseline values and 5-year changes in HR-pQCT parameters; examine baseline values and changes in HR-pQCT parameters in relation to previous fracture; and examine those HR-pQCT parameters associated with previous fracture in relation to established BMD loci.
2.0. Methods
2.1. The Hertfordshire Cohort Study
The Hertfordshire Cohort Study (HCS) comprises 2997 individuals born in Hertfordshire from 1931-1939 and who still lived there in 1998-2004 where they completed a home interview and clinic visit for a detailed assessment of their health. In 2004, of the 966 participants from East Hertfordshire who had a dual-energy X-ray absorptiometry (DXA) scan at the start of the study, 642 were recruited for a musculoskeletal follow-up study. In 2011-2012, 376/642 participated in a further bone follow-up study and 224/376 took part in a 2017 bone follow-up study. The background of the HCS and further details of the follow-up studies have been described previously (11, 12).
2.2. Ascertainment of participant characteristics in 1998-2004
Dietary calcium intake was determined using a food-frequency questionnaire (13). Current or most recent full-time occupation (husband’s for ever-married females) was ascertained; social class was coded from the 1990 OPCS Standard Occupational Classification (SOC90) unit group for occupation (14). Current use of bisphosphonates at the study commencement (1998-2004), at the musculoskeletal follow-up (2004-2005) and in 2011-2012 was ascertained from details of all currently used over the counter or prescription medications.
2.3. Ascertainment of participant characteristics in 2011-2012
Smoking status, alcohol consumption (in units per week), average daily outdoor physical activity in minutes (Longitudinal Aging Study Amsterdam Physical Activity Questionnaire (LAPAQ)((15))) and any previous self-reported fractures since aged 45 years were ascertained at the home interview through nurse-administered questionnaires. Height was measured (wall-mounted SECA stadiometer) along with weight (calibrated SECA 770 digital floor scales, SECA Ltd, Hamburg) and used to derive BMI (kg/m2). Morphometric vertebral fractures were diagnosed from a lateral spine view imaged using a Lunar Prodigy Advance DXA scanner (GE Medical Systems) and graded based on the Genant semi-quantitative method of vertebral fracture assessment (16) using a single reader. Participants with a vertebral fracture or a self-reported fracture since age 45 years were regarded as having had a previous fracture.
HR-pQCT scans (XtremeCTi, Scanco Medical AG, Switzerland) of the non-dominant distal radius and tibia were performed; dominant limbs were scanned if the non-dominant limb had previously fractured. In total, 110 parallel CT slices were obtained, representing a volume of bone 9mm in axial length with a nominal resolution (voxel size) of 82μm. The scan protocol was in accordance with manufacturer’s guidelines and as described by Boutroy et al (17). Using the method of Pauchard et al (18), scans exhibiting excessive motion artefact (grade 5 scans) were excluded; scans of quality 4 and above were included in the analysis. Manufacturer standard evaluation and cortical porosity scripts were used for image analysis (8, 19–23). Extended cortical analysis was performed for all scans (24).
2.4. Ascertainment of participant characteristics at follow-up (2017)
HR-pQCT scans of the distal radius and tibia were repeated using the same devices and protocol as in 2011-2012. Quality assurance and quality control were performed as per manufacturer guidance and no significant changes were made to the machines over follow-up. Overlap between baseline and follow-up HR-pQCT scans was derived using the manufacturer’s slice-match method. Change measures for HR-pQCT parameters were set to missing if the overlap between the baseline and follow-up scan regions was less than 75% as recommended(25); median (lower quartile, upper quartile) overlap between the remaining scans were 93% (87%, 96%) for the radius and 95% (92%, 97%) for the tibia. Short term precision values (percentage root mean squared coefficient of variation) for cortical and trabecular BMD have been shown to range from 0.3 to 1.2 (26).
2.5. Selection of loci for analysis
SNP genotyping was carried out using Infinium Global Screening Array v1. Raw IDAT files were loaded into Illumina Genome Studio 2.0.0 and analysed using genotyping Module 2.0.0. Data was processed using Illumina hard cut-off technical specifications (27). bcftools and PLINK 1.9 beta (28) were used to prepare the QC genotype data ready for imputation according to SANGER specifications. Data were uploaded in VCF format to SANGER servers, pre-phased using EAGLE2 pipeline and imputed using UK10K + 1000 Genome Phase 3 (29).
Estrada et al performed a genome-wide meta-analysis and identified 64 SNPs that were associated with BMD (femoral neck or lumbar spine) or fracture (4). In total, 61/64 of these SNPs were available in HCS and were examined in relation to femoral neck and lumbar spine BMD (acquired via DXA) at HCS commencement (1998-2004). The following loci were associated with both femoral neck and lumbar spine BMD in HCS (data not shown) and, therefore, were used in this study: rs1053051 (TMEM263 (alias name: C12orf23)); rs7812088 (ABCF2); rs10226308 (TXNDC3 (also known as NME8)); and rs3801387 (WNT16).
2.6. Ethical approval and informed consent
The initial phase of the HCS (1998-2004) had ethical approval from the Hertfordshire and Bedfordshire Local Research Ethics Committee, and the 2011-2012 bone follow-up had ethical approval from the East of England - Cambridgeshire and Hertfordshire Research Ethics Committee. All participants provided written informed consent. Ethical approval was obtained for all follow-ups of the study at the time they were conducted. Investigations were conducted in accordance with the principles expressed in the Declaration of Helsinki.
2.7. Statistical analysis
In the statistical description and results, ‘baseline’ refers to the baseline HR-pQCT scan visit (2011-12). Participant characteristics, including baseline values and annual percentage changes in HR-pQCT parameters, were described using summary statistics. Standard deviation (SD) scores were coded for baseline values and annual percentage changes in HR-pQCT parameters for use in models to enable the comparison of effect sizes. Odds ratios for previous fracture per SD difference in baseline values and changes in HR-pQCT parameters were derived using logistic regression after adjustment for sex and after additional adjustment for age, height, BMI, dietary calcium, physical activity, smoking history (ever vs never), alcohol consumption and social class. Baseline values and changes in HR-pQCT parameters were then examined in relation to the four selected loci using linear regression after adjustment for sex and age.
To maintain sample size, males and females were pooled (sex-interaction effects were not statistically significant) and analyses were adjusted for sex; p<0.05 was regarded as statistically significant. Analyses were conducted using Stata, release 15.1. The analysis sample comprised 214 participants with both baseline values and change in at least one HR-pQCT parameter (tibial or radial) and had data on previous fracture or the loci of interest.
3.0. Results
3.1. Participant characteristics
The characteristics of the study population are presented in Table 1. Mean (SD) age of the 214 participants at the time of the baseline HR-pQCT scans (2011-2012) was 76.0 (2.5) years. Median (lower quartile, upper quartile) duration between the baseline and follow-up HR-pQCT scans was 5.2 (4.8, 5.4) years. Overall 27 (25.2%) males and 32 (33.3%) females had a previous fracture (vertebral or self-reported since age 45 years). The sites of fractures in this cohort have been described previously (30).
Table 1. Descriptive statistics for participant characteristics in 2011-2012.
| Participant characteristic | Mean (standard deviation), n(%) or median (lower quartile, upper quartile) | |
|---|---|---|
| Males (n=115) | Females (n=99) | |
| Age (years) | 75.8 (2.5) | 76.2 (2.6) |
| Height (cm) | 173.3 (6.5) | 160.4 (5.7) |
| Weight (kg) | 83.1 (13.0) | 72.4 (12.5) |
| BMI (kg/m2) | 27.7 (4.0) | 28.1 (4.5) |
| Weekly dietary calcium (g)* | 8.3 (2.2) | 8.0 (2.9) |
| Physical activity in last 2 weeks (min/day) ** | 195.0 (128.6, 291.4) | 207.9 (150.0, 287.1) |
| Ever smoked | 61 (56.5%) | 39 (40.2%) |
| Weekly alcohol units (M: Males; F: Females) | ||
| Very low (0/<1 M&F) | 24 (22.2%) | 48 (49.5%) |
| Low (1-10M,1-7F) | 48 (44.4%) | 39 (40.2%) |
| Moderate (11-21M,8-14F) | 17 (15.7%) | 8 (8.2%) |
| High (>21M, >14F) | 19 (17.6%) | 2 (2.1%) |
| Social class (manual)* | 59 (53.6%) | 53 (53.5%) |
| Bisphosphonate use (1998-2004 to 2012)* | 4 (3.7%) | 20 (20.6%) |
| Self-reported fracture since aged 45 years | 24 (22.4%) | 30 (30.9%) |
| Vertebral fracture | 6 (5.3%) | 6 (6.1%) |
| Any fracture (self-reported or vertebral) | 27 (25.2%) | 32 (33.3%) |
| rs1053051*: CC | 26 (24.3%) | 29 (31.5%) |
| CT | 47 (43.9%) | 33 (35.9%) |
| TT | 34 (31.8%) | 30 (32.6%) |
| rs7812088*: GG | 82 (76.6%) | 70 (76.1%) |
| AG | 24 (22.4%) | 22 (23.9%) |
| AA | 1 (0.9%) | 0 (0.0%) |
| rs10226308*: AA | 78 (72.9%) | 47 (51.1%) |
| GA | 26 (24.3%) | 40 (43.5%) |
| GG | 3 (2.8%) | 5 (5.4%) |
| rs3801387*: AA | 60 (56.1%) | 45 (48.9%) |
| GA | 39 (36.4%) | 40 (43.5%) |
| GG | 8 (7.5%) | 7 (7.6%) |
Ascertained using information at initial phase of the Hertfordshire Cohort Study (1998-2004)
Derived using the Longitudinal Ageing Study Amsterdam Physical Activity Questionnaire
Associations between tibial parameters and previous fracture were substantially stronger than those observed for radial parameters; however, the patterns of association were similar for each. Therefore, we confined the remainder of the results section for tibial measures.
Median (lower quartile, upper quartile) values for baseline values and annual percentage changes in tibial HR-pQCT parameters are presented in Table 2. Although annual percentage changes in parameters were small, many median values for these longitudinal changes were significantly different from zero (p<0.05). For example, there were decreases in trabecular density (females only), cortical area, density and thickness, and total volumetric bone density; increases in cortical pore diameter (males only), trabecular area and cortical porosity were also observed. Equivalent information for the radial HR-pQCT parameters is presented in Supplementary Table A.
Table 2. Descriptive statistics for tibial HR-pQCT parameters at baseline (2011-2012) and for changes in parameters from 2011-2012 to 2017.
| Parameter [Median (lower quartile, upper quartile) values shown] | Males (n=115) | Females (n=99) | ||||
|---|---|---|---|---|---|---|
| Baseline | Annual change (%) | P-value | Baseline | Annual change (%) | P-value | |
| Trabecular area (mm2) | 743 (648, 838) | 0.1 (0.0, 0.2) | <0.001 | 620 (542, 707) | 0.2 (0.0, 0.3) | <0.001 |
| Total volumetric bone density (mg/cm3) | 294 (262, 343) | -0.5 (-1.0, -0.2) | <0.001 | 252 (219, 276) | -0.8 (-1.5, -0.3) | <0.001 |
| Trabecular density (mg/cm3) | 196 (170, 215) | -0.1 (-0.3, 0.2) | 0.180 | 167 (149, 199) | -0.3 (-0.8, 0.2) | 0.012 |
| Trabecular number (mm-1) | 2.5 (2.2, 2.7) | -0.1 (-1.0, 1.1) | 0.444 | 2.3 (2.0, 2.5) | -0.2 (-1.7, 1.4) | 0.213 |
| Trabecular thickness (mm) | 0.066 (0.058, 0.072) | 0.0 (-1.1, 1.0) | 1.000 | 0.064 (0.054, 0.071) | 0.0 (-1.3, 1.2) | 1.000 |
| Trabecular separation (mm) | 0.34 (0.30, 0.38) | 0.1 (-1.1, 1.1) | 0.501 | 0.37 (0.34, 0.43) | 0.3 (-1.3, 2.0) | 0.300 |
| Cortical area (mm2) | 129 (116, 153) | -0.7 (-1.5, 0.0) | <0.001 | 83 (74, 98) | -1.3 (-2.4, -0.4) | <0.001 |
| Cortical density (mg/cm3) | 836 (803, 873) | -0.6 (-1.0, -0.3) | <0.001 | 763 (721, 807) | -0.7 (-1.2, -0.2) | <0.001 |
| Cortical porosity (%) | 8.8 (7.3, 10.6) | 2.5 (0.5, 4.8) | <0.001 | 9.9 (7.9, 12.2) | 1.5 (-0.1, 3.3) | <0.001 |
| Cortical thickness (mm) | 1.2 (1.1, 1.4) | -0.4 (-1.1, 0.1) | <0.001 | 0.9 (0.8, 1.1) | -1.1 (-2.0, -0.1) | <0.001 |
| Cortical pore diameter (mm) | 0.17 (0.16, 0.18) | 0.2 (-0.7, 1.3) | 0.031 | 0.18 (0.17, 0.19) | 0.1 (-0.9, 0.9) | 0.602 |
P-values correspond to tests that median annual percentage changes were zero and were calculated from sign tests
Median annual percentage changes that were significantly different from zero (p<0.05) are highlighted in bold (underlined for increases and italic for decreases)
3.2. Relationships between baseline values and changes in tibial HR-pQCT parameters and previous fracture
Associations between baseline values and annual percentage changes in tibial HR-pQCT parameters in relation to previous fracture are presented in Table 3 and Figure 1. Increased odds of previous fracture were observed for; higher trabecular area (p<0.01); lower total volumetric bone density (p<0.01); lower cortical area (p<0.04), lower cortical density (p<0.02) and lower cortical thickness (p<0.01); and greater declines in trabecular density (p<0.002). These relationships were robust in sex-adjusted and fully-adjusted models. Associations between baseline values and annual percentage changes in radial HR-pQCT parameters in relation to previous fracture are presented in Supplementary Table B.
Table 3. Odds ratios for previous fracture per standard deviation difference in both baseline values and changes in parameters.
| HR-pQCT tibia parameter (SD) | Baseline values in 2011-2012 | Annual percentage change from 2011-2012 to 2017 | ||||||
|---|---|---|---|---|---|---|---|---|
| Sex-adjusted | Fully-adjusted* | Sex-adjusted | Fully-adjusted* | |||||
| Odds ratio (95% CI) |
P-value | Odds ratio (95% CI) |
P-value | Odds ratio (95% CI) |
P-value | Odds ratio (95% CI) |
P-value | |
| Trabecular area | 1.70 (1.15,2.51) | 0.007 | 2.18 (1.27,3.73) | 0.005 | 1.11 (0.80,1.56) | 0.525 | 1.01 (0.71,1.44) | 0.945 |
| Total volumetric bone density | 0.58 (0.39,0.84) | 0.004 | 0.53 (0.34,0.84) | 0.007 | 0.72 (0.52,1.01) | 0.059 | 0.75 (0.53,1.07) | 0.117 |
| Trabecular density | 0.79 (0.56,1.10) | 0.168 | 0.72 (0.49,1.06) | 0.099 | 0.56 (0.40,0.80) | 0.001 | 0.50 (0.34,0.75) | 0.001 |
| Trabecular number | 0.92 (0.66,1.27) | 0.602 | 0.96 (0.64,1.43) | 0.836 | 0.99 (0.72,1.36) | 0.951 | 0.91 (0.64,1.29) | 0.584 |
| Trabecular thickness | 0.80 (0.58,1.10) | 0.168 | 0.69 (0.48,1.00) | 0.052 | 0.83 (0.60,1.15) | 0.256 | 0.88 (0.62,1.27) | 0.502 |
| Trabecular separation | 1.12 (0.81,1.55) | 0.504 | 1.10 (0.74,1.64) | 0.626 | 1.03 (0.75,1.41) | 0.872 | 1.13 (0.79,1.60) | 0.515 |
| Cortical area | 0.52 (0.32,0.84) | 0.008 | 0.53 (0.30,0.95) | 0.032 | 0.86 (0.62,1.20) | 0.377 | 0.98 (0.68,1.40) | 0.908 |
| Cortical density | 0.51 (0.34,0.76) | 0.001 | 0.56 (0.36,0.88) | 0.011 | 0.93 (0.67,1.28) | 0.641 | 1.00 (0.70,1.42) | 0.987 |
| Cortical porosity | 1.09 (0.79,1.50) | 0.618 | 0.96 (0.67,1.37) | 0.808 | 0.71 (0.51,1.00) | 0.048 | 0.73 (0.51,1.05) | 0.090 |
| Cortical thickness | 0.48 (0.31,0.74) | 0.001 | 0.45 (0.27,0.77) | 0.004 | 0.82 (0.59,1.15) | 0.255 | 0.91 (0.64,1.30) | 0.597 |
| Cortical pore diameter | 0.85 (0.61,1.17) | 0.317 | 0.71 (0.49,1.03) | 0.070 | 0.82 (0.59,1.15) | 0.253 | 0.82 (0.56,1.20) | 0.309 |
HR-pQCT: High resolution peripheral quantitative computed tomography; CI: Confidence interval
Adjusted for sex, age, height, BMI, dietary calcium, physical activity, smoking history (ever vs never), alcohol consumption and social class
Odds ratio of less than one for annual percentage change in parameter shows that reduced declines are related to lower risk of previous fracture
Significant associations (p<0.05) are given in italics
Figure 1. Fully-adjusted odds ratios for previous fracture per standard deviation difference in baseline values and changes in key tibial HR-pQCT parameters.
Odds ratios were adjusted for sex, age, height, BMI, dietary calcium, physical activity, smoking history (ever vs never), alcohol consumption and social class
Odds ratios greater than one: higher baseline values in 2011-2012 or reduced declines from 2011-2012 to 2017 were associated with greater risk of previous fracture
Odds ratios less than one: higher baseline values in 2011-2012 or reduced declines from 2011-2012 to 2017 were associated with lower risk of previous fracture
3.3. Selected loci in relation to tibial HR-pQCT parameters that were associated with previous fracture
Relationships between selected loci and baseline values and annual percentage changes in tibial HR-pQCT parameters that were associated with previous fracture are presented in Table 4. Few loci were associated with these measures; rs3801387 (WNT16) was related to change in trabecular density (p=0.011) and rs7812088 (ABCF2) was related to baseline values of trabecular area (p=0.072) but the remaining associations were weak (p>0.09).
Table 4. Selected loci in relation to tibial HR-pQCT parameters that were associated with previous fracture.
| HR-pQCT tibia parameter (z-scores) | rs1053051 (TMEM263): per extra T allele | rs7812088 (ABCF2): AG compared to GG | rs10226308 (TXNDC3): per extra G allele | rs3801387 (WNT16): per extra G allele | ||||
|---|---|---|---|---|---|---|---|---|
| Estimate (95% CI) | P-value | Estimate (95% CI) | P-value | Estimate (95% CI) | P-value | Estimate (95% CI) | P-value | |
| Baseline values | ||||||||
| Trabecular area | -0.07 (-0.24,0.10) | 0.426 | -0.28 (-0.59,0.03) | 0.072 | 0.07 (-0.17,0.30) | 0.584 | -0.02 (-0.23,0.18) | 0.810 |
| Total volumetric bone density | 0.13 (-0.03,0.30) | 0.112 | 0.18 (-0.13,0.48) | 0.251 | 0.02 (-0.21,0.25) | 0.856 | 0.03 (-0.17,0.23) | 0.767 |
| Cortical area | 0.06 (-0.07,0.18) | 0.385 | 0.10 (-0.13,0.33) | 0.384 | -0.03 (-0.21,0.15) | 0.734 | 0.01 (-0.14,0.16) | 0.879 |
| Cortical density | 0.04 (-0.12,0.20) | 0.636 | 0.00 (-0.28,0.29) | 0.974 | 0.06 (-0.15,0.28) | 0.560 | 0.03 (-0.16,0.21) | 0.784 |
| Cortical thickness | 0.09 (-0.06,0.24) | 0.244 | 0.21 (-0.07,0.49) | 0.136 | -0.06 (-0.27,0.15) | 0.571 | 0.01 (-0.17,0.20) | 0.874 |
| Annual percentage changes | ||||||||
| Trabecular density | 0.08 (-0.11,0.26) | 0.397 | -0.10 (-0.43,0.23) | 0.543 | 0.22 (-0.04,0.47) | 0.092 | -0.28 (-0.50,-0.07) | 0.011 |
Estimates shown are standard deviation differences in HR-pQCT parameters according to each loci
Associations were adjusted for age and sex
Significant associations (p<0.05) are given in italics
Pearson correlations between the tibial HR-pQCT parameters associated with previous fracture are presented in Supplementary Table C.
3.4. Selected loci in relation to tibial HR-pQCT parameters that were not associated with previous fracture
Of the list of tibial HR-pQCT parameters that were not associated with previous fracture, rs10226308 (TXNDC3) was related to baseline values of trabecular number and separation; rs1053051 (TMEM263) was related to change in trabecular thickness and cortical porosity; and rs3801387 (WNT16) was related to baseline values of cortical pore diameter (Supplementary Table D). For completeness, Supplementary Table D presents associations for the selected loci in relation to baseline values and annual percentage changes in both tibial and radial HR-pQCT parameters.
4.0. Discussion
Through this study, we have described the longitudinal changes in tibial bone microarchitecture which occur in older adults including decreased trabecular density (females only) and increased trabecular area, cortical porosity and pore diameter (males only) and decreased cortical area, density and thickness and total volumetric bone density. We have shown that the odds of fracture increase with larger trabecular area and decrease with higher volumetric BMD and cortical area, density and thickness. Additionally we have demonstrated that the odds of fracture increase with greater loss in trabecular density. We have also shown that a SNP within ABCF2 has a borderline association with trabecular area and a SNP within WNT16 is associated with change in trabecular density.
The description of tibial bone microarchitectural changes over 5 years in our study is defined by decreased trabecular density (in females only) and increased trabecular area, cortical porosity and pore diameter (males only) and decreased cortical area, density and thickness. These characteristics are aligned to the current theory of bone aging with trabecular bone loss and decreased cortical bone. The reason for increasing trabecular area is likely due to the process of endocortical resorption with increasing trabecular area being at the cost of cortical thinning (31).
When we consider our findings with regard to fracture prediction it is important to bear in mind the current literature regarding HR-pQCT outcomes and fracture associations. It is, of course, vital to recognise that in the studies below, the fractures described are incident whereas in this current study the fractures are prevalent. This may have a bearing on the associations observed.
Lower baseline volumetric BMD was predictive of future fracture over 5 years in the GERICO cohort of post-menopausal females (8), as was radial total BMD (Odds ratio (OR) per standard deviation (SD) lower baseline values: 2.1) and trabecular BMD (OR: 2.0) in the CaMOS cohort(9). Lower baseline cortical thickness at the tibia was also predictive of fracture in the CaMOS study (OR: 2.2) though trabecular parameters at the radius had greater predictive capacity over 10 years of follow-up in the OFELY cohort(10) and trabecular parameters had a greater association with future fracture in a group taking denosumab (32). Interestingly an additional finding from the CaMOS study was that baseline values rather than rate of change were more associated with future fractures in post-menopausal females(9) and this was certainly echoed in our study, with only change in trabecular density being associated with fracture odds. Associations between previous fracture and HR-pQCT parameters in our study were unaltered after additional adjustment for bisphosphonate use and relationships between previous fracture and changes in HR-pQCT parameters were also unaltered when adjusted for baseline values of the corresponding parameters. The reason why associations with fracture were stronger regarding baseline values of HR-pQCT parameters, compared to longitudinal changes in parameters, is potentially because genes and variants involved in building the skeleton up (to peak bone mass) are more relevant to fracture than those associated with the degradation of bone microarchitecture that is observed with ageing.
A SNP near ABCF2, the gene for a member of the ATP-binding cassette (ABC) transporter superfamily, was borderline associated with trabecular area. ABCF2 plays a role in transmembrane transportation and previous work has uncovered associations with cancer progression(33, 34), the molecular pathogenesis of Duchenne Muscular Dystrophy(35), and is down-regulated in ulcerative colitis(36). It is possible that the association between trabecular area is mediated via stem cell pathways or inflammatory mechanisms which have previously been described in this cohort (37).
We also saw that a SNP at WNT16 was associated with change in trabecular density. Wnt16 is thought to be largely secreted by osteoblasts (38) and directly suppresses osteoclastogenesis via the non-canonical JNK MAPK pathway with upregulation of osteoprotegerin. In murine models Wnt16 has been shown to be reduced by steroid administration (39) and the particular SNP rs3801387 has previously been associated with BMD, fracture and cortical bone thickness (4, 40, 41). It is surprising that we should have only observed effects on density in the trabecular compartment and also that the direction of association suggests that a greater number of G alleles was associated with a disadvantageous greater reduction of trabecular density. In the Estrada GWAS, from which the WNT16 SNP was identified, each additional A allele was disadvantageous to bone and associated with lower femoral neck BMD (SD difference per additional A allele: -0.08), lower spinal BMD (β= -0.10) and higher risk of low-trauma fracture (OR: 1.06) (4). Interestingly, we observed the same direction of association between rs3801387 and cross-sectional DXA BMD in our study at the HCS commencement (1998-2004) as was observed by Estrada et al. which therefore suggests that our finding related to change in trabecular density could be due to the different parameter of bone microarchitecture being measured. Similar deleterious effects of this SNP (rs3801387) were observed in cohort of pre-menopausal females with each additional T allele associated with lower BMD at the lumbar spine (β= -0.16) and femoral neck (-0.12) (40). It is important, therefore, that our finding is investigated in other cohorts which have HR-pQCT and bone microarchitecture available. Other GWAS have demonstrated associations between WNT16 and BMD but not at rs3801387 (42, 43).
Our study has many strengths. Firstly, the longitudinal HR-pQCT dataset in a group of older adults is a novel contribution to the literature as many previous studies have focused on post-menopausal females or included participants from a broad age range from adolescence to older age. Secondly, the opportunity to examine genotypic data in conjunction with longitudinal HR-pQCT is unique. Thirdly, the HCS has been phenotyped according to strict protocols by highly-trained fieldworkers and managed by an experienced multi-disciplinary team. The limitations of our study are that it was exploratory and used a relatively small sample size which does not allow for site-specific fracture analyses, and the fact that we have a prevalent (rather than incident) fracture history. Additionally, as the SNPs we analysed were taken from the work by Estrada and colleagues(4), these were based initially on a meta-analysis of BMD GWAS before Estrada also examined them in relation to fracture.
In conclusion, we have described the cortical and trabecular bone deterioration and endocortical bone loss associated with ageing and demonstrated that baseline cortical and BMD levels held greater associations with fracture than changes in bone microarchitecture. We have also shown borderline associations between ABCF2, a SNP previously associated with BMD, and trabecular area, and WNT16 and change in trabecular density.
Our findings contribute to the literature describing the effects of ageing on bone microarchitecture but also add to previous work teasing out the genetic determinants of trabecular and cortical characteristics.
Supplementary Material
Funding
The Hertfordshire Cohort Study was supported by the following organisations: Medical Research Council; British Heart Foundation; Versus Arthritis UK; International Osteoporosis Foundation; NIHR Southampton Biomedical Research Centre; NIHR Oxford Biomedical Research Centre; University of Southampton. NRF was supported by the Dunhill Medical Trust. The funders had no involvement in the following: the study design; the collection, analysis and interpretation of data; the writing of the report; and the decision to submit the article for publication.
Footnotes
Compliance with Ethical Standards
Declaration of interest
CC reports personal fees (outside the submitted work) from Amgen, Danone, Eli Lilly, GSK, Kyowa Kirin, Medtronic, Merck, Nestle, Novartis, Pfizer, Roche, Servier, Shire, Takeda and UCB. NCH reports consultancy, lecture fees and honoraria (outside the submitted work) from Alliance for Better Bone Health, AMGEN, MSD, Eli Lilly, Servier, Shire, UCB, Kyowa Kirin, Consilient Healthcare, Radius Health and Internis Pharma. EMD reports personal fees (outside the submitted work) from Pfizer Healthcare and from the UCB Discussion panel. NRF has received travel bursaries from Pfizer and Eli Lilly. LDW, GB, PT, MB and KAW declare that they have no conflicts of interest.
Ethical Approval
The initial phase of the Hertfordshire Cohort Study (1998-2004) had ethical approval from the Hertfordshire and Bedfordshire Local Research Ethics Committee and the 2011-2012 follow-up had ethical approval from the East of England - Cambridgeshire and Hertfordshire Research Ethics Committee.
Human and Animal Rights
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed Consent
All participants gave signed consent to participate in the study and for their health records to be accessed in the future.
Author contributions
NRF Investigation, Writing - Original Draft; LDW Methodology, Formal analysis, Writing - Original Draft; GB Investigation, Resources, Writing - Review & Editing; PT Methodology, Data Curation; MB Writing - Review & Editing; NCH Conceptualization, Writing - Review & Editing; EMD Conceptualization, Writing - Review & Editing, Supervision, Project administration; CC Conceptualization, Writing - Review & Editing, Supervision, Project administration; KAW Conceptualization, Investigation, Writing - Review & Editing, Supervision, Project administration. All authors made substantial contributions to the manuscript and approved the final version.
Contributor Information
Nicholas R Fuggle, Email: nrf@mrc.soton.ac.uk.
Leo D Westbury, Email: lw@mrc.soton.ac.uk, kaj@mrc.soton.ac.uk.
Gregorio Bevilacqua, Email: gb@mrc.soton.ac.uk.
Philip Titcombe, Email: pt@mrc.soton.ac.uk.
Mícheál Ó Breasail, Email: michealo@mrc-lmb.cam.ac.uk.
Nicholas C Harvey, Email: nch@mrc.soton.ac.uk.
Elaine M Dennison, Email: emd@mrc.soton.ac.uk.
Cyrus Cooper, Email: cc@mrc.soton.ac.uk.
Kate A Ward, Email: kw@mrc.soton.ac.uk.
References
- 1.Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int. 2006;17(12):1726–33. doi: 10.1007/s00198-006-0172-4. [DOI] [PubMed] [Google Scholar]
- 2.Kanis JA, Johansson H, Oden A, Johnell O, De Laet C, Eisman JA, et al. A family history of fracture and fracture risk: a meta-analysis. Bone. 2004;35(5):1029–37. doi: 10.1016/j.bone.2004.06.017. [DOI] [PubMed] [Google Scholar]
- 3.Arden NK, Baker J, Hogg C, Baan K, Spector TD. The heritability of bone mineral density, ultrasound of the calcaneus and hip axis length: a study of postmenopausal twins. J Bone Miner Res. 1996;11(4):530–4. doi: 10.1002/jbmr.5650110414. [DOI] [PubMed] [Google Scholar]
- 4.Estrada K, Styrkarsdottir U, Evangelou E, Hsu Y-H, Duncan EL, Ntzani EE, et al. Genome-wide meta-analysis identifies 56 bone mineral density loci and reveals 14 loci associated with risk of fracture. Nature genetics. 2012;44(5):491–501. doi: 10.1038/ng.2249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Paternoster L, Ohlsson C, Sayers A, Vandenput L, Lorentzon M, Evans DM, et al. OPG and RANK polymorphisms are both associated with cortical bone mineral density: findings from a metaanalysis of the Avon longitudinal study of parents and children and gothenburg osteoporosis and obesity determinants cohorts. J Clin Endocrinol Metab. 2010;95(8):3940–8. doi: 10.1210/jc.2010-0025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Paternoster L, Lorentzon M, Lehtimäki T, Eriksson J, Kähönen M, Raitakari O, et al. Genetic determinants of trabecular and cortical volumetric bone mineral densities and bone microstructure. PLoS Genet. 2013;9(2):e1003247. doi: 10.1371/journal.pgen.1003247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nielson CM, Liu CT, Smith AV, Ackert-Bicknell CL, Reppe S, Jakobsdottir J, et al. Novel Genetic Variants Associated With Increased Vertebral Volumetric BMD Reduced Vertebral Fracture Risk, and Increased Expression of SLC1A3 and EPHB2. J Bone Miner Res. 2016;31(12):2085–97. doi: 10.1002/jbmr.2913. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Biver E, Durosier-Izart C, Chevalley T, van Rietbergen B, Rizzoli R, Ferrari S. Evaluation of Radius Microstructure and Areal Bone Mineral Density Improves Fracture Prediction in Postmenopausal Women. J Bone Miner Res. 2018;33(2):328–37. doi: 10.1002/jbmr.3299. [DOI] [PubMed] [Google Scholar]
- 9.Burt LA, Manske SL, Hanley DA, Boyd SK. Lower Bone Density, Impaired Microarchitecture, and Strength Predict Future Fragility Fracture in Postmenopausal Women: 5-Year Follow-up of the Calgary CaMos Cohort. J Bone Miner Res. 2018;33(4):589–97. doi: 10.1002/jbmr.3347. [DOI] [PubMed] [Google Scholar]
- 10.Sornay-Rendu E, Boutroy S, Duboeuf F, Chapurlat RD. Bone Microarchitecture Assessed by HR-pQCT as Predictor of Fracture Risk in Postmenopausal Women: The OFELY Study. J Bone Miner Res. 2017;32(6):1243–51. doi: 10.1002/jbmr.3105. [DOI] [PubMed] [Google Scholar]
- 11.Syddall H, Sayer AA, Dennison E, Martin H, Barker D, Cooper C. Cohort profile: the Hertfordshire cohort study. Int J Epidemiol. 2005;34(6):1234–42. doi: 10.1093/ije/dyi127. [DOI] [PubMed] [Google Scholar]
- 12.Syddall HE, Simmonds SJ, Carter SA, Robinson SM, Dennison EM, Cooper C, et al. The Hertfordshire Cohort Study: an overview. F1000Research. 2019:8. doi: 10.12688/f1000research.17457.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Robinson S, Syddall H, Jameson K, Batelaan S, Martin H, Dennison EM, et al. Current patterns of diet in community-dwelling older men and women: results from the Hertfordshire Cohort Study. Age Ageing. 2009;38:594–9. doi: 10.1093/ageing/afp121. [DOI] [PubMed] [Google Scholar]
- 14.Office of Population Censuses and Surveys. Standard occupational classification, Vol 1 Structure and definition of major, minor and unit groups. HMSO; London: 1990. [Google Scholar]
- 15.Stel VS, Smit JH, Pluijm SM, Visser M, Deeg DJ, Lips P. Comparison of the LASA Physical Activity Questionnaire with a 7-day diary and pedometer. J Clin Epidemiol. 2004;57(3):252–8. doi: 10.1016/j.jclinepi.2003.07.008. [DOI] [PubMed] [Google Scholar]
- 16.Genant HK, Wu CY, van KC, Nevitt MC. Vertebral fracture assessment using a semiquantitative technique. JBone MinerRes. 1993;8(9):1137–48. doi: 10.1002/jbmr.5650080915. [DOI] [PubMed] [Google Scholar]
- 17.Boutroy S, Bouxsein ML, Munoz F, Delmas PD. In vivo assessment of trabecular bone microarchitecture by high-resolution peripheral quantitative computed tomography. J Clin Endocrinol Metab. 2005;90(12):6508–15. doi: 10.1210/jc.2005-1258. [DOI] [PubMed] [Google Scholar]
- 18.Pauchard Y, Liphardt A-M, Macdonald HM, Hanley DA, Boyd SK. Quality control for bone quality parameters affected by subject motion in high-resolution peripheral quantitative computed tomography. Bone. 2012;50(6):1304–10. doi: 10.1016/j.bone.2012.03.003. [DOI] [PubMed] [Google Scholar]
- 19.MacNeil JA, Boyd SK. Accuracy of high-resolution peripheral quantitative computed tomography for measurement of bone quality. Med Eng Phys. 2007;29(10):1096–105. doi: 10.1016/j.medengphy.2006.11.002. [DOI] [PubMed] [Google Scholar]
- 20.Laib A, Hauselmann HJ, Ruegsegger P. In vivo high resolution 3D-QCT of the human forearm. TechnolHealth Care. 1998;6(5-6):329–37. [PubMed] [Google Scholar]
- 21.Khosla S, Riggs BL, Atkinson EJ, Oberg AL, McDaniel LJ, Holets M, et al. Effects of sex and age on bone microstructure at the ultradistal radius: a population-based noninvasive in vivo assessment. J Bone MinerRes. 2006;21(1):124–31. doi: 10.1359/JBMR.050916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Boutroy S, Van Rietbergen B, Sornay‐Rendu E, Munoz F, Bouxsein ML, Delmas PD. Finite element analysis based on in vivo HR‐pQCT images of the distal radius is associated with wrist fracture in postmenopausal women. J Bone Miner Res. 2008;23(3):392–9. doi: 10.1359/jbmr.071108. [DOI] [PubMed] [Google Scholar]
- 23.Vilayphiou N, Boutroy S, Szulc P, van Rietbergen B, Munoz F, Delmas PD, et al. Finite element analysis performed on radius and tibia HR‐pQCT images and fragility fractures at all sites in men. J Bone Miner Res. 2011;26(5):965–73. doi: 10.1002/jbmr.297. [DOI] [PubMed] [Google Scholar]
- 24.Burghardt AJ, Buie HR, Laib A, Majumdar S, Boyd SK. Reproducibility of direct quantitative measures of cortical bone microarchitecture of the distal radius and tibia by HR-pQCT. Bone. 2010;47(3):519–28. doi: 10.1016/j.bone.2010.05.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Whittier DE, Boyd SK, Burghardt AJ, Paccou J, Ghasem-Zadeh A, Chapurlat R, et al. Guidelines for the assessment of bone density and microarchitecture in vivo using high-resolution peripheral quantitative computed tomography. 2020 doi: 10.1007/s00198-020-05438-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Paggiosi MA, Eastell R, Walsh JS. Precision of high-resolution peripheral quantitative computed tomography measurement variables: influence of gender, examination site, and age. Calcif Tissue Int. 2014;94(2):191–201. doi: 10.1007/s00223-013-9798-3. [DOI] [PubMed] [Google Scholar]
- 27.Infinium® Genotyping Data Analysis, A guide for analyzing Infinium genotyping data using the GenomeStudio® Genotyping Module. Available from: https://www.illumina.com/Documents/products/technotes/technote_infinium_genotyping_data_analysis.pdf.
- 28.Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81(3):559–75. doi: 10.1086/519795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Loh P-R, Danecek P, Palamara PF, Fuchsberger C, A Reshef Y, K Finucane H, et al. Reference-based phasing using the Haplotype Reference Consortium panel. Nature genetics. 2016;48(11):1443–8. doi: 10.1038/ng.3679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Edwards M, Robinson D, Ward K, Javaid M, Walker-Bone K, Cooper C, et al. Cluster analysis of bone microarchitecture from high resolution peripheral quantitative computed tomography demonstrates two separate phenotypes associated with high fracture risk in men and women. Bone. 2016;88:131–7. doi: 10.1016/j.bone.2016.04.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Bala Y, Bui QM, Wang XF, Iuliano S, Wang Q, Ghasem-Zadeh A, et al. Trabecular and cortical microstructure and fragility of the distal radius in women. J Bone Miner Res. 2015;30(4):621–9. doi: 10.1002/jbmr.2388. [DOI] [PubMed] [Google Scholar]
- 32.Butscheidt S, Rolvien T, Vettorazzi E, Frieling I. Trabecular bone microarchitecture predicts fragility fractures in postmenopausal women on denosumab treatment. Bone. 2018;114:246–51. doi: 10.1016/j.bone.2018.06.022. [DOI] [PubMed] [Google Scholar]
- 33.Karatas OF, Guzel E, Duz MB, Ittmann M, Ozen M. The role of ATP-binding cassette transporter genes in the progression of prostate cancer. Prostate. 2016;76(5):434–44. doi: 10.1002/pros.23137. [DOI] [PubMed] [Google Scholar]
- 34.Gao J, Dai C, Yu X, Yin XB, Zhou F. Circ-TCF4.85 silencing inhibits cancer progression through microRNA-486-5p-targeted inhibition of ABCF2 in hepatocellular carcinoma. 2020;14(2):447–61. doi: 10.1002/1878-0261.12603. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- 35.Xiu MX, Zeng B, Kuang BH. Identification of hub genes, miRNAs and regulatory factors relevant for Duchenne muscular dystrophy by bioinformatics analysis. Int J Neurosci. 2020:1–10. doi: 10.1080/00207454.2020.1810030. [DOI] [PubMed] [Google Scholar]
- 36.Verma N, Ahuja V, Paul J. Profiling of ABC transporters during active ulcerative colitis and in vitro effect of inflammatory modulators. Dig Dis Sci. 2013;58(8):2282–92. doi: 10.1007/s10620-013-2636-7. [DOI] [PubMed] [Google Scholar]
- 37.Fuggle NR, Westbury LD, Syddall HE, Duggal NA, Shaw SC, Maslin K, et al. Relationships between markers of inflammation and bone density: findings from the Hertfordshire Cohort Study. Osteoporos Int. 2018;29(7):1581–9. doi: 10.1007/s00198-018-4503-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Movérare-Skrtic S, Henning P, Liu X, Nagano K, Saito H, Börjesson AE, et al. Osteoblast-derived WNT16 represses osteoclastogenesis and prevents cortical bone fragility fractures. Nature medicine. 2014;20(11):1279–88. doi: 10.1038/nm.3654. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Hildebrandt S, Baschant U, Thiele S, Tuckermann J, Hofbauer LC, Rauner M. Glucocorticoids suppress Wnt16 expression in osteoblasts in vitro and in vivo. Scientific reports. 2018;8(1):8711. doi: 10.1038/s41598-018-26300-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Koller DL, Zheng HF, Karasik D, Yerges-Armstrong L, Liu CT, McGuigan F, et al. Meta-analysis of genome-wide studies identifies WNT16 and ESR1 SNPs associated with bone mineral density in premenopausal women. J Bone Miner Res. 2013;28(3):547–58. doi: 10.1002/jbmr.1796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Medina-Gomez C, Kemp JP, Estrada K, Eriksson J, Liu J, Reppe S, et al. Meta-analysis of genome-wide scans for total body BMD in children and adults reveals allelic heterogeneity and age-specific effects at the WNT16 locus. PLoS Genet. 2012;8(7):e1002718. doi: 10.1371/journal.pgen.1002718. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Zheng HF, Tobias JH, Duncan E, Evans DM, Eriksson J, Paternoster L, et al. WNT16 influences bone mineral density, cortical bone thickness, bone strength, and osteoporotic fracture risk. PLoS Genet. 2012;8(7):e1002745. doi: 10.1371/journal.pgen.1002745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Kemp JP, Morris JA, Medina-Gomez C, Forgetta V, Warrington NM, Youlten SE, et al. Identification of 153 new loci associated with heel bone mineral density and functional involvement of GPC6 in osteoporosis. Nature genetics. 2017;49(10):1468–75. doi: 10.1038/ng.3949. [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.