The Association Between Facial Ageing and Bone Mineral Density: A Mendelian Randomisation Study

By Yabo Wang^1,2, Chao Zhen³, Ziyu Wang¹, Lu Zhang¹, Yunzhen Chen¹

Affiliations

1. Department of Orthopaedics, Qilu Hospital of Shandong University, Jinan, China
2. Department of Emergency, Qilu Hospital (Qingdao), Cheeloo College of Medicine, Shandong University, Qingdao, China
3. Department of Neurology, Qingdao Municipal Hospital, University of Health and Rehabilitation Sciences, Qingdao, China

doi: 10.29271/jcpsp.2025.06.797

ABSTRACT
This article aimed to assess the causative association between facial ageing and bone mineral density (BMD) using Mendelian randomisation (MR). In this observational study, single-nucleotide polymorphisms (SNPs) linked to facial ageing and four different body regions, including total mineral density (TB-BMD, n = 56,284), femoral neck BMD (FN-BMD, n = 32,735), forearm BMD (FA-BMD, n = 8,143), and lumbar spine BMD (LS-BMD, n = 28,498), were employed as instrumental variables. A two-sample MR analysis was undertaken to scrutinise the causal association between the identified exposure and respective outcomes. The authors found there was no causal relationship between facial ageing and TB-BMD (Beta, -0.1238; 95% CI, -0.3125 to 0.0648; p = 0.1982), FN-BMD (Beta, 0.0720; 95% CI, -0.1728 to 0.3168; p = 0.5643), FA-BMD (Beta, 0.0582; 95% CI, -0.3803 to 0.4966; p = 0.7949), or LS-BMD (Beta, -0.0188; 95% CI, -0.3087 to 0.2711; p = 0.8987). Therefore, the degree of osteoporosis in a person cannot be inferred based on their facial ageing.

Key Words: Bone mineral density, Causal relationship, Mendelian randomisation, Facial ageing, Osteoporosis.

As a surprisingly complex organ, the skin serves several functions, such as maintaining homeostasis and facilitating neurosensory functions. However, similar to all other organs, the skin succumbs to the elapses of time. With ageing, facial skin often becomes less elastic, thinner, and more fragile. Osteoporosis is a skeletal disorder associated with ageing, characterised by a decline in bone mineral density (BMD).¹ As the degradation of bone tissue microarchitecture occurs, the risk of bone fragility increases. Osteoporosis has become a major concern worldwide. This article aimed to assess the causative association between facial ageing and BMD.

SNPs linked to facial ageing were used as instrumental variables to study gene-disease associations from the genome-wide association study (GWAS) database of the Medical Research Council Integrative Epidemiology Unit (MRC-IEU). Summarised data were obtained from the UK Biobank, covering 423,999 people of the European ancestry and 9,851,867 SNPs (field code 1757).²

 

 Relationships with BMD were explored using the IEU GWAS database, including total body BMD (TB-BMD, n = 56,284), femoral neck BMD (FN-BMD, n = 32,735), forearm BMD (FA-BMD, n = 8,143), and lumbar spine BMD (LS-BMD, n = 28,498).

Initially, SNPs significantly associated with facial ageing (p <5 × 10^-8) were extracted. The clumping process (using a window size of 10,000 kb and R² <0.001)³ was carried out on samples of the European origin from the 1,000 genomes project to yield robust MR estimates. SNPs with a minor allele frequency (MAF) below 0.01 were eliminated. Correlated phenotypes of these SNPs were checked using PhenoScanner V2, and those linked to confound- ing factors were excluded. Proxy SNPs with high linkage disequilibrium (LD, r2 >0.8) were utilised when the specified SNP was unavailable in the IEU GWAS database. The MR-PRESSO test was con- ducted to detect horizontal pleiotropy, and outliers were removed.

The random-effects inverse-variance weighted (IVW) method served as the primary statistical model, with additional analyses conducted using the weighted median method, MR-Egger regression, simple median, simple mode, MR robust adjusted profile score (MR. RAPS), and MR-PRESSO. Significance was set at p <0.05 for all analyses, which were two-sided. R software (version 4.2.2) was employed, along with the R packages TwoSampleMR, MRPRESSO, and Mendelian Randomisation (MR). 

According to the IVW method, there was no MR relationship between facial ageing and TB-BMD, FN-BMD, FA-BMD, or LS-BMD. Consistently, no causal association of facial ageing on TB-BMD, FN-BMD, FA-BMD, and LS-BMD was found in the results of weighted-median analyses (Figure 1).

Table I: Sensitivity analysis: Horizontal pleiotropy test and heterogeneity test.

Exposure	Outcome	Horizontal pleiotropy test (MR-Egger)			Heterogeneity test (IVW)		Heterogeneity test (MR-Egger)
		Intercept	SE	p-value	Q	p-value	Q	p-value
Facial ageing	TB-BMD	0.0017955	0.0021098	0.3982066	80.04682	0.0429055	79.07618	0.0416235
	FN-BMD	0.0029851	0.0026623	0.2667031	83.54752	0.0239469	81.80423	0.0263873
	FA-BMD	-0.0046070	0.0047620	0.3372026	70.00597	0.2009962	68.93070	0.2009647
	LS-BMD	0.00095265	0.0031959	0.7666669	90.71594	0.0080845	90.58179	0.0065524
Q: Cochran’s Q test, IVW: Inverse-variance weighted. SE: Standard error

Figure 1: Various MR estimation methods employed to evaluate the causal impact of facial ageing on bone mineral density.
TB-BMD: Total body BMD, FN-BMD: Femoral neck BMD, FA-BMD: Forearm BMD, LS-BMD: Lumbar spine BMD 

The F-statistic was ≥10 (ranging from 76 to 971) for all the facial ageing variants, and no bias stemming from instruments of insufficient strength was present.

By the MR-PRESSO method, the authors identified 3 outliers (rs12441130, rs12882664, and rs138880) for TB-BMD and 1 outlier (rs6740259) for FN-BMD. The authors noticed the existence of heterogeneity (Table I) and employed a random effects IVW model for reliable causal estimates. The final MR estimates were presented in Figure 1. MR-Egger regression was also employed to assess for horizontal pleiotropy, and little horizontal pleiotropy evidence was discovered (Table I).

There was mounting evidence that facial ageing and osteoporosis have overlapping risk factors, including genetics, lifestyle, and hormone changes.^4,5 However, the presence of these confounding factors made it difficult to effectively assess the relationship between facial ageing and osteoporosis. To address the limitations of observational studies, the authors wanted to establish causal links between facial ageing and BMD at various skeletal sites by utilising summary statistics from various GWAS and applying the two-sample MR method. To the best of the authors’ knowledge, this study initiated the first two-sample MR investigation in this area. In this MR analysis, horizontal pleiotropy was strictly controlled and possible confounders, including ever smoked, body mass index, heel BMD, weight, were minimised. These results were robust and indicated that facial ageing did not have a causal impact on BMD. Therefore, the degree of osteoporosis in a person based on their facial ageing cannot be inferred, this discovery bore immense significance for public health.

The utilisation of an MR study strategy brought many strengths to the present study. First, the large-scale GWAS was used in this study included millions of SNPs, which increased the statistical power and improved the precision of genetic estimates; second, MR represented the closest approximation to RCTs, as it allowed for a more meticulous evaluation of potential confounders; third, according to the MR-Egger regression, the identified causal associations between instrumental variables and outcomes were not affected by directional pleiotropy.

In conclusion, this study did not establish a causal relationship between facial ageing and BMD. It was implied that the process of facial ageing is not causally linked to the development of osteoporosis. However, due to the potential bias stemming from the use of self-reported data in the questionnaire, it was recommended that future studies incorporate objective scores or molecular markers to provide more accurate assessments of facial ageing.

COMPETING  INTEREST:
The authors declared no conflict of interest.

AUTHORS’ CONTRIBUTION:
YW: Performed data analysis and wrote the manuscript.
CZ: Revised the manuscript.
ZW, LZ: Clarified research ideas.
YC: Initiated the research direction.
All authors approved the final version of the manuscript to be published.

REFERENCES

1. Liu J, Curtis EM, Cooper C, Harvey NC. State of the art in osteoporosis risk assessment and treatment. J Endocrinol Invest 2019; 42(10):1149-64. doi: 10.1007/s40618-019-01041-6.
2. Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. UK biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med 2015; 12(3):e1001779. doi: 10. 1371/journal.pmed.1001779.
3. Zha LF, Dong JT, Wang JL, Chen QW, Wu JF, Zhou YC, et al. Effects of Insomnia on peptic ulcer disease using mende-lian randomisation. Oxid Med Cell Longev 2021; 2021: 2216314. doi: 10.1155/2021/2216314.
4. Evans M, Lewis ED, Zakaria N, Pelipyagina T, Guthrie N. A randomized, triple-blind, placebo-controlled, parallel study to evaluate the efficacy of a freshwater marine collagen on skin wrinkles and elasticity. J Cosmet Dermatol 2021; 20(3):825-34. doi: 10.1111/jocd.13676.
5. Goodman GD, Kaufman J, Day D, Weiss R, Kawata AK, Garcia JK, et al. Impact of smoking and alcohol use on facial ageing in women: Results of a large multinational, multiracial, cross-sectional survey. J Clin Aesthet Dermatol 2019; 12(8):28-39.