Beverage consumption and facial skin aging: Evidence from Mendelian randomization analysis

Observational studies have linked coffee, alcohol, tea, and sugar‐sweetened beverage (SSB) consumption to facial skin aging. However, confounding factors may influence these studies. The present two‐sample Mendelian randomization (MR) investigated the potential causal association between beverage consumption and facial skin aging.


| INTRODUC TI ON
Facial skin aging is an ongoing and irreversible process caused by a complex interaction between intrinsic and extrinsic factors. 1 Gender, ethnicity, and genetic variations are intrinsic factors that contribute to facial skin aging.Among the extrinsic factors, sun exposure is the most significant and well-established risk factor. 2 Lifestyles such as smoking and nutrition are significant extrinsic factors as well. 3spite this, the precise etiology and pathogenesis of skin aging remain unclear.
Coffee, alcohol, tea, and sugar-sweetened beverage (SSB) are commonly considered dietary components.Studies have investigated their health benefits and health risks. 4,5Daily coffee consumption may slow facial skin aging and improve skin health. 4,6cording to a recent systematic review, tea may have potential skin-protective properties by delaying skin aging. 7][10] Furthermore, cross-sectional studies have shown a correlation between elevated glucose levels and premature skin aging. 11,12However, current evidence linking beverage consumption to facial skin aging has mainly originated from observational studies, which are susceptible to confounding variables.
Mendelian randomization (MR) employs genetic variants as instrumental variables (IVs) to identify causal relationships between exposures and outcomes while effectively diminishing confounding factors and reverse causation through the inherent random allocation of alleles, thus avoiding bias. 135][16][17][18] To our knowledge, there has been no MR analysis investigating the causal association between beverage consumption and facial skin aging.Given these findings, a comprehensive two-sample MR analysis was conducted to investigate the potential causal association between beverage consumption (coffee, alcohol, tea, and SSB) and facial skin aging.

| Study design
Two-sample MR analysis was performed with publicly available genome-wide association study (GWAS) data.As only European individuals were included in our study, population stratification was minimized.Single nucleotide polymorphisms (SNPs) were employed as IVs.IVs should meet the following MR assumptions 1 : they are strongly associated with exposures, 2 they exhibit no correlation with any potential confounder factors and 3 they affect the outcome exclusively via exposure pathways rather than alternative pathways (Figure 1). 19

| Data sources of exposures
SNPs associated with coffee intake, alcohol intake frequency and tea intake were derived from the IEU (https:// gwas.mrcieu.ac.uk/ ) (Table 1).The UK Biobank study provided genetic and phenotypic data regarding 500 000 individuals (aged 40-69) throughout the United Kingdom. 20Specifically, coffee intake-associated SNPs were whereas tea intake was reported as a continuous variable.SSB intake-associated SNPs were obtained from a recent publicly available GWAS that involved 85 852 participants of European ancestry. 21

| Data sources of outcomes
Genetic variants associated with facial skin aging, characterized as an ordered categorical variable, were derived from a previous GWAS (Table 1).The present study involved 423 992 participants from the UK Biobank who completed questionnaires on their perceived age.
The participants were asked whether they believed they appeared younger, older, or their actual age or preferred not to answer. 22

| IVs selection
For the analysis of coffee, alcohol and tea intake, SNPs with p < 5 × 10 −8 were identified as candidate SNPs. 23Candidate SNPs were clumped based on a cutoff of r 2 < 0.001 within a 10 000 kb window to guarantee independence.All SNPs were harmonized in exposure and outcome datasets to reflect the same effect allele.Palindromic SNPs were deleted to prevent allele coding or strand orientation distortion.Linkage disequilibrium (LD) proxies (r 2 > 0.8) were applied for missing SNP instruments in the outcome dataset.
SNPs for SSB intake were selected using a lenient threshold of p < 1× 10 −5 to achieve adequate IVs and removed SNPs in LD with r 2 > 0.01 to ensure independence. 24The strength of IV was evaluated by calculating the F statistic, and all included SNPs demonstrated F statistic values exceeding 10.
The MR pleiotropy residual sum and outlier (MR-PRESSO) method was implemented to rectify horizontal pleiotropy (Table 2). 25rst, the pleiotropy significance p-value for each SNP was calculated using the MR-PRESSO outlier test.The overall horizontal pleiotropy was then calculated using the MR-PRESSO global test.The SNPs were ranked ascendingly based on their MR-PRESSO outlier test p-values.Some SNPs were sequentially removed until no significant p-value (p > 0.05) was reached for the MR-PRESSO global test.
The remaining SNPs were then used in subsequent MR analysis. 26

| MR analysis
The random-effect model in the inverse variance weighted (IVW) served as the main method for examining the causal effect of beverage intake on facial skin aging, with other methods (MR Egger, weighted median, simple mode, and weighted mode) providing complementary information. 27,28To ensure robustness, the weighted median method was employed as a supplementary tool, provided that at least 50% of the SNPs met the MR assumptions. 29Similarly, the weighted mode method was used to estimate the causal relationship even if most IVs were pleiotropic. 30The MR-Egger approach was used to estimate causal effects even when all genetic variants were invalid IVs. 31 The MR-Egger approach allows for the presence of pleiotropy, so it can provide a valid causal estimate even in the presence of pleiotropy. 31e MR-Egger intercept analysis and the MR-PRESSO method were performed in sensitivity analysis, whereas the heterogeneity was assessed by Cochran's Q test. 32The intercept did not significantly deviate from zero, indicating no directional pleiotropy (p > 0.05). 31Furthermore, the leave-one-out analysis was conducted to assess the influence of individual SNP.
The causal relationship between beverage consumption (coffee, tea, alcohol, and SSB) and facial skin aging was assessed using the "TwoSampleMR" 33 and "MRPRESSO" 25  multiple testing, statistical significance was determined at p < 0.0125 (0.05/4) after Bonferroni correction.The Bonferroni correction is considered conservative, which magnifies the risk of Type 2 statistical error. 34A p < 0.0125 was deemed as substantial evidence for causal effects, while a p-value ranging from 0.0125 to 0.05 suggested a potential causal effect.

| Replication analysis
To enhance the reliability of the findings, a replicated analysis was performed, utilizing self-reported coffee consumption data from a distinct GWAS dataset comprising 375 833 individuals of European  3).The funnel plot analysis demonstrated no evidence of an asymmetrical distribution of data (Figure S2).
Notably, the outcomes derived from the replicated MR analysis were consistent with the initial analysis results (Figure S3).The IVW analysis revealed that coffee consumption (OR = 0.997; 95% CI = 0.996-0.999;p = 0.003) exhibited a causal impact on facial skin aging (Table 3).

| DISCUSS ION
This study is the first to examine the causal relationship of beverage intake with facial skin aging in UK Biobank populations.The key finding of this MR analysis is that coffee consumption was inversely associated with facial skin aging.However, no evidence was found that higher genetically determined intake levels of tea, alcohol, or SSB led to skin aging.
This study demonstrated that coffee intake had a causal effect on a lower risk of facial skin aging, consistent with most previous studies. 4,6,35,36To verify the identified causal relationship in the initial dataset, we employed a sample comprising coffee consumption data from a different European population.The subsequent replication analysis yielded comparable outcomes.However, several retrospective studies reported no impact of coffee intake on the presence of skin aging. 5,37There are several possible explanations for the positive association between coffee intake and facial skin aging.The process of facial skin aging is heavily influenced by oxidative stress.The antioxidant and anti-inflammatory properties of coffee may protect skin senescence as a dietary factor. 38In vivo studies have proved the beneficial effects of the coffee extract on skin moisture, elasticity and collagen content. 4The extracellular matrix (ECM) is mostly composed of collagen, elastin and glycosaminoglycans.The expression of matrix metalloproteinases (MMPs), which are induced by reactive oxygen species (ROS), leads to the degradation of the ECM. 39,40As collagen degrades, wrinkles deepen and skin elasticity decreases. 41Additionally, it has been clarified that green coffee oil (GCO) stimulates human skin fibroblasts to synthesize ECM compounds. 42This effect is attributed to the upregulation of TGFβ and GM-CSF.Moreover, GCO upregulates the expression of AQP-3 mRNA in cultured keratinocytes, which serves as an essential indicator of skin hydration and contributes to the prevention of water loss. 42asted coffee is the primary dietary source of polyphenols, known for potent antioxidant properties. 43Research has shown that consuming polyphenol-rich plants can effectively delay cellular senescence and mitigate skin aging.These positive effects are believed to be linked to the downregulation expression of the MMP and increased collagen synthesis in fibroblasts. 4,35,36ffee also contains caffeine, another bioactive compound, that has anti-senescent properties because of its ability to scavenge free radicals and eliminate damaged keratinocytes. 44Caffeine has been proven to elongate telomeres, thereby reducing cellular senescence. 45Additionally, studies revealed that chlorogenic acid isolated from coffee has the potential to reduce skin aging. 36In summary, the findings imply that coffee could delay the process of facial skin aging. 36Nevertheless, several unexplored coffee extracts warrant further investigation.Subsequent research should focus on elucidating the correlation between coffee extracts and facial skin aging, thereby facilitating the identification and exploration of underlying molecular mechanisms.
7][48][49][50][51] Tea extracts have been found to improve wrinkles and attenuate light-induced skin aging in mouse models. 46,47wever, a randomized controlled trial in vivo found no statistically significant differences in skin aging after 2 years of supplementation with oral tea polyphenols. 52Tea contains tea polyphenols and caffeine, known to exhibit antioxidant and anti-inflammatory properties. 7,4753][54][55] Some studies revealed that  especially with heavy alcohol intake. 9,10,53,55However, a twin study found no significant effect of alcohol intake on facial aging, 54 and a cohort-based survey found that drinking alcohol significantly reduced skin aging risk. 853][54][55][56][57][58][59][60] In our study, alcohol intake was considered as an exposure but lacked stratified data such as the specific types of alcoholic beverages and amount of alcohol consumed.Future studies with more comprehensive GWAS data are needed to perform subgroup analysis to investigate whether alcohol consumption and facial skin aging are causally related.
SSBs are a major source of added sugar, and high SSB intake was linked to several adverse health outcomes according to an umbrella review. 61Mammalian cells rely on glucose as their main energy source, 62 and previous research has established a significant correlation between high-sugar diets and elevated blood sugar levels in both the systemic circulation and skin.Conversely, adhering to low-sugar diets holds the potential to reduce cutaneous blood sugar levels. 63The formation of advanced glycation end products (AGEs), resulting from the bonding of glucose or fructose with collagen or elastin, leads to the manifestation of various dermatological signs such as wrinkles, sagging, and reduced elasticity.This mechanism serves as a pivotal contributor to the aging process of the skin. 64wever, our study found no association between SSBs and facial skin aging.For possible reasons, the SSB-associated IVs did not meet the recommended threshold, p-value <5 × 10 −8 , which reduces the robustness of MR results.Furthermore, prior studies primarily focused on evaluating the effects of SSB intake on type 2 diabetes and elderly patients, while lack of investigation into its impact on the general population.This highlights the importance of conducting randomized controlled trials to further investigate the relationship between SSB intake and facial skin aging.
This study has several notable strengths.Initially, our investigation used MR analysis to assess the causal relationship between beverage intake and facial skin aging, which was less susceptible to confounding bias.Additionally, only European individuals were included to reduce the bias resulting from population stratification.
This study had some limitations.First, our results could not be generalized to other populations, as our sample was limited to Europeans.Therefore, it is crucial to incorporate a more diverse range of pedigree populations to validate the results.Secondly, the utilization of self-reported data for facial skin aging assessment affected the objective and quantitative validation of outcomes due to the absence of a standardized evaluation system.Thirdly, the investigation of the effects of beverage consumption on facial skin aging was constrained.Because of the limited availability of GWAS datasets, we were unable to conduct a comprehensive analysis stratified by type, brewing method, or volume.Lastly, the methodology employed to evaluate beverage consumption through the utilization of food frequency questionnaires or food records may exhibit inconsistencies.The questionnaire surveys may introduce biases when recalling beverage intake.Relying solely on self-reported beverage intake via questionnaires may yield inaccurate results.To mitigate potential biases, it is advisable to conduct further research utilizing intake biomarkers as IVs.

| CON CLUS ION
This MR analysis revealed that coffee intake decelerated facial skin aging.This study also found that alcoholic intake frequency, tea intake and SSB intake were not associated with facial skin aging, which may motivate individuals to lead healthy dietary habits.

AUTH O R CO NTR I B UTI O N S
obtained from the MRC-IEU (428 860 individuals, Dataset: ukbb-5237), alcohol intake-associated SNPs were obtained from the same consortium (462 346 individuals, Dataset: ukb-b-5779) and tea intake-associated SNPs were obtained from the MRC-IEU Traits consortium (447 485 individuals, Dataset: ukb-k-6066).Coffee and alcohol intake were reported as categorical ordered variables, F I G U R E 1 Overview and assumptions of the Mendelian randomization (MR) study.

descent. 21 3
| RE SULTS In this study, the causal relationship between four types of beverage consumption and facial skin aging was first investigated.The number of SNPs used as IVs ranged from 29 to 85.These instruments explained 3.2%-15.4% of the variance for their respective exposures.The MR estimates (IVW, MR Egger, weighted median, simple mode and weighted mode) of beverage consumption on the risk of facial skin aging were illustrated in Figure 2. Additionally, as shown in Figure S1, the scatter plots depicted the estimated effect of SNPs on beverage intake and facial skin aging.The X-axis represents the SNP effects on beverage intake, while the Y-axis represents the SNP effects on facial skin aging.Each point represents a single SNP, and the gray line on each point represents 95% CI.The slopes of each colored line represent the causal association for each method.As shown in Figure 2 and Figure S1, only one causality was identified in this study.The IVW method revealed that coffee consumption (OR: 0.852; 95% CI: 0.753-0.964;p = 0.011) had a protective effect on facial skin aging, which was supported by the weighted median method (OR: 0.845; 95% CI: 0.721-0.990;p = 0.037).No statistically significant results were observed in other three methods (p > 0.05).This finding should be interpreted cautiously and validated through replication studies.No heterogeneity and directional pleiotropy were discovered by Cochrane's Q test (p > 0.05) (Table

F I G U R E 2
alcohol consumption is linked to increased facial wrinkles and midface volume loss, leading to the older perceived age of individuals, Forest plot of the causal effects of beverage intake-associated SNPs on facial skin aging.TA B L E 2 Results of MR-Egger intercept test and MR-PRESSO for pleiotropy.
Details of the GWAS summary-level datasets.
R packages.Considering TA B L E 1

Exposure Outcome MR-Egger intercept test MR-PRESSO global test
Results of Cochrane's Q test for heterogeneity.IVW, inverse variance weighted; SNPs, single-nucleotide polymorphisms; SSB, sugar-sweetened beverage.
Liu: Conception and design of study, data collection and analysis, manuscript writing; Xin Li: Conception and design of study, TA B L E 3