Associations of urinary caffeine and caffeine metabolites with metabolic syndrome in U.S. adults

doi:10.21203/rs.3.rs-2953101/v1

Download PDF

Research Article

Associations of urinary caffeine and caffeine metabolites with metabolic syndrome in U.S. adults

https://doi.org/10.21203/rs.3.rs-2953101/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

The relationship between caffeine and metabolic syndrome (MetS) has only been evaluated from the perspective of caffeine intake, however, the association between urinary caffeine and MetS is still unclear. This study examines the associations between urinary caffeine and its metabolites and MetS and its components among adults. Data from the US National Health and Nutrition Examination Survey (NHANES) 2011–2014 was analyzed. NHANES is a stratified, multi-stage survey of all non-institutionalized persons in the United States. A total of 2394 subjects without missing data and aged ≥18 years were selected in this study. Urinary caffeine and 14 of its metabolite levels are quantified by the use of high-performance liquid chromatography-electrospray ionization-tandem quadrupole mass spectrometry (HPLC-ESI-MS/MS) with stable isotope-labeled internal standards. We performed principal components analysis (PCA) to investigate the underlying correlation structure of these fifteen features of urinary caffeine and its metabolites and used these principal components (PCs) as independent variables to conduct logistic regression analysis with or without restricted cubic splines (RCS) terms to explore the associations between caffeine metabolites and MetS. Two main PCs that derived from PCA explained 90.67% of the total variance of caffeine and its metabolites. The first PC (PC1, strongly correlated with 1-MU, 1,3-DMU, 1,7-DMU, 1,3,7-TMU, 1-MX, 1,3-DMX, 1,7-DMX, 1,3,7-TMX and AAMU) was positively correlated with risk of MetS (OR=1.27, P<0.001) and all its components (all ORs>1, all P-values<0.001) in the unadjusted models, while in the adjusted models, it was positively correlated with MetS (OR=1.16, P=0.042) and central obesity (OR=1.22, P<0.001). In the unadjusted model, there were significant associations between the second PC (PC2, correlated with 3-MU, 7-MU, 3,7-DMU, 3-MX, 7-MX and 3,7-DMX) and MetS (OR=1.11, P=0.030) and central obesity (OR=1.16, P<0.001), while in the adjusted models, PC2 was positively associated with MetS (OR=1.15, P=0.035) and central obesity (OR=1.15, P=0.005), and negatively associated with raised triglycerides (TG) (OR=0.84, P=0.008). Moreover, we observed U-shaped associations between PC1 and the risk of raised TG both in unadjusted (P_non-linear = 0.017) and adjusted (P_non-linear = 0.014) models. Urinary caffeine metabolites were positively associated with the risk of MetS and its components through a variety of linear or non-linear patterns. Excessive consumption of coffee increases the risk of developing MetS and its related diseases.

Caffeine

Logistic regression

Metabolic syndrome

Principal components analysis

Restricted cubic splines

Metabolic syndrome (MetS) is a multifactorial pathological condition with high prevalence among adult populations (Eckel et al. 2005, Moore et al. 2017), these pathological conditions include central obesity, elevated triglycerides (TG), elevated blood pressure (BP), elevated fasting plasma glucose (FPG), and decreased high-density lipoprotein-cholesterol (HDL-C) (Eckel et al. 2005, Lam &LeRoith 2000, Lemieux &Després 2020). MetS is a major global public health problem and is associated with an increased risk of many diseases (Després &Lemieux 2006). Some studies have shown that MetS increases the risk of type 2 diabetes, cardiovascular disease, stroke, and myocardial infarction by 2–5 times (Alberti et al. 2009, Lopez-Candales et al. 2017). Therefore, understanding the risk factors of MetS can decrease costs for public health systems worldwide and is of great importance to human health.

Coffee is one of the most popular drinks in the world, with an estimated 165 million 60 kg bags consumed per year (Ridder 2023). Up to now, several studies have explored the relationship between coffee intake and MetS, but the conclusions are inconsistent. For example, in 2016, a meta-analysis demonstrated that coffee consumption was associated with a low risk of MetS (Shang et al. 2016), and another study also demonstrated that coffee consumption was associated with less likelihood of having MetS (Marventano et al. 2016). However, another meta-analysis found that the overall sex-adjusted association of the coffee consumption level was not significant with MetS (Wong et al. 2021). Besides, a non-linear relationship was found between coffee consumption and the MetS in a previous study (Shang et al. 2016). However, few studies have investigated the caffeine and caffeine metabolites and MetS. More than 70% of the caffeine in vivo comes from coffee intake, and 70–80% of caffeine is formed from 7-dimethylxanthine (84%), theobromine (12%) and theophylline (4%) through 3-N demethylation reaction catalyzed by CYP1A2 enzyme (Guessous et al. 2015, Nehlig 2018, Oñatibia-Astibia et al. 2016). Although there was no study had reported the relationships between caffeine and its metabolites and MetS, some studies have shown that caffeine and caffeine metabolite were correlated with the risk of hypertension, insulin resistance cardiovascular disease (Guessous et al. 2015, Lee et al. 2020, Ngueta 2020, Weng et al. 2021), all of which may appear in people with MetS. For instance, a study has reported that ambulatory systolic BP was inversely associated with urinary caffeine and its metabolites (Guessous et al. 2015), and another study found that metabolites of caffeine, but not caffeine, significantly reduce the odds of hypertension in the U.S. population (Ngueta 2020). In addition, a previous study indicated that levels of urinary caffeine, paraxanthine, and theophylline were associated with decreased parameters of arterial stiffness (Ponte et al. 2018). A study of U.S. showed that caffeine and caffeine metabolites were positively related to insulin resistance and beta cell function (Lee et al. 2020). Some researchers found that urinary theophylline and caffeine were inversely associated with CVDs in women (Weng et al. 2021). Given the above inconsistent research and the nonlinear relationship between coffee intake and MetS that previous study has proved, we speculate that caffeine and its metabolites have linear or nonlinear relationships with the risk of MetS and its components. In the present study, we sought to determine the associations of caffeine and 14 caffeine metabolites with Mets and its components using data from the National Health and Nutrition Examination Survey (NHANES) 2011–2014.

2.1 Study population

The National Health and Nutrition Examination Survey (NHANES) is a stratified, multi-stage survey of all non-institutionalized persons in the United States. For our study, we abstracted data from two cycles of continuous 2011–2014 NHANES examination. Individuals aged < 18 years and missing data on one or more caffeine metabolites were excluded, resulting in a total of 2394 subjects being included from both cycles. Missing data was not selected for the analysis. The present study was exempt from formal ethics review as a secondary analysis of existing NHANES public data under the US Health & Human Services (HHS) regulations at 45 CFR 46.101 (b). Ethical approval was given by the National Center for Health Statistics Ethics Review Board and all survey participants provided informed written consent.

2.2 Measurement of urinary caffeine and caffeine metabolites levels

Urine Caffeine and 14 of its metabolite levels are quantified by the use of high-performance liquid chromatography-electrospray ionization-tandem quadrupole mass spectrometry (HPLC-ESI-MS/MS) with stable isotope-labeled internal standards. A total of 15 metabolites were measured:1-methyluric acid (1-MU), 3-methyluric acid (3-MU), 7-methyluric acid (7-MU), 1,3-dimethyluric acid (1,3-DMU), 1,7-dimethyluric acid (1,7-DMU), 3,7-dimethyluric acid (3,7-DMU), 1,3,7-trimethyluric acid (1,3,7-TMU), 1-methylxanthine (1-MX), 3-methylxanthine (3-MX), 7-methylxanthine (7-MX), theophylline (1,3-dimethylxanthine, 1,3-DMX), paraxanthine (1,7-dimethylxanthine, 1,7-DMX), theobromine (3,7-dimethylxanthine, 3,7-DMX), caffeine (1,3,7-trimethylxanthine, 1,3,7-TMX), and 5-acetylamino-6-amino-3-methyluracil (AAMU). The details of laboratory methodology were described at:

https://wwwn.cdc.gov/nchs/data/nhanes/2011-2012/labmethods/CAFE_G_MET.pdf .

2.3 Definition of MetS

We defined MetS according to the new International Diabetes Federation definition as: central obesity plus any two of the following four factors: (1) raised TG ≥ 150 mg/dL (1.7 mmol/L) or specific treatment for this lipid abnormality; (2) reduced HDL-C < 40 mg/dL (1.03 mmol/L) in male, < 50 mg/dL (1.29 mmol/L) in female, or specific treatment for this lipid abnormality; (3) raised BP: systolic BP ≥ 130 or diastolic BP ≥ 85 mm Hg, or treatment of previously diagnosed hypertension; (4) raised FPG: ≥ 100 mg/dL (5.6 mmol/L), or previously diagnosed type 2 diabetes.

2.4 Covariates

Covariates were obtained from the NHANES 2011–2014 and included demographic characteristic factors (gender, age, race/ethnicity, education level and poverty-income ratio) and health-related behavior factors (smoking, drinking and physical activity, BMI), Alcohol drinking over past 12 months was used to assess alcohol consumption (CDC/NCHS, 2014). The details of the definition of covariates information were described in our previous study (Zhou et al. 2022).

2.5 Statistical analysis

Levels of urinary caffeine and caffeine metabolites were Log10-transformed to correct distribution skewness. Median and quartiles were used to describe continuous data with a non-normal distribution, frequency and percentage were used to describe categorical data. Spearman correlations were employed to test the relationship between 15 urine caffeine and caffeine metabolites. To investigate the underlying correlation structure of the fifteen features, we used principal components analysis (PCA) as an exploratory analysis. Orthogonal varimax rotation to the factors was applied, and retained those with eigenvalues higher than 1(Yeomans &Golder 1982). Principal components (PC) scores (linear combinations of the caffeine and caffeine metabolites multiplied by their respective loadings) for every of the retained components were calculated for every individual. Then we further use these PC as independent variables and MetS (and its components) as dependent variables for univariate logistic regression and multiple logistic regression analysis. Restricted cubic spline (RCS) logistic regression with three knots was employed to explore the nonlinear associations of the risk of MetS and the levels of urinary caffeine and caffeine metabolites, and the median values of levels of urinary caffeine and caffeine metabolites (Jin et al. 2018). Both the multiple logistic regression analysis and the RCS logistic regression were adjusted for the confounding demographic characteristic factors (gender, age, race/ethnicity, education level and poverty-income ratio) and health-related behavior factors (smoking, drinking and physical activity, BMI), if the dependent variable was MetS or obesity, BMI was not considered as a confounding factor. The ‘anova’ function in the ‘rms’ package (version 6.1-0) of R to estimate P_overall and P_non−linear was used to assess the statistical significance of dose-response associations. If both P-values were lower than 0.05, it indicated a non-linear association of dose-response between the probability of MetS and the concentration of caffeine and caffeine metabolites in the urinary. R Core Team (2023). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. and SPSS version 26 (SPSS Inc., Chicago, IL, USA) were used for all statistical analyses.

3.1. Basic characteristics of the study population

Table 1

Table 1

Characteristics of participants included in the study.
Characteristics	N (%) or Median (P25 ~ P75)
Gender	N (%)
Male	1282(53.55%)
Female	1112(46.45%)
Race/ethnicity
Mexican American	273(11.40%)
Other Hispanic	232(09.69%)
Non-Hispanic white	948(39.60%)
Non-Hispanic black	505(21.09%)
Other	436(18.21%)
Education
< 12 y	517(21.60%)
12 y	495(20.68%)
> 12 y	1382(57.73%)
Smoking status
Never	1346(56.22%)
Current	502(20.97%)
Former	546(22.81%)
Physical Activity
None	1496(62.49%)
Moderate	452(18.88%)
Vigorous	446(18.63%)
Continuous (median [IQR])
Age, years	48.00 [33.00 ~ 62.00]
Poverty-income ratio	2.16 [1.07 ~ 4.16]
Drinking (times/year)	2.00 [1.00 ~ 4.00]
BMI	26.30 [23.10 ~ 30.10]
1-MU (umol/L)	56.85 [21.40 ~ 124.00]
3-MU (umol/L)	0.43 [0.14 ~ 1.17]
7-MU (umol/L)	11.10 [3.82 ~ 27.50]
1,3-DMU (umol/L)	6.28 [2.09 ~ 14.50]
1,7-DMU (umol/L)	24.40 [6.49 ~ 56.60]
3,7-DMU (umol/L)	0.73 [0.28 ~ 1.78]
1,3,7-TMU (umol/L)	1.24 [0.28 ~ 3.32]
1-MX (umol/L)	26.30 [9.13 ~ 64.00]
3-MX (umol/L)	22.95 [9.61 ~ 53.40]
7-MX (umol/L)	37.55 [15.50 ~ 87.90]
1,3-DMX (umol/L)	1.65 [0.53 ~ 3.45]
1,7-DMX (umol/L)	16.10 [4.83 ~ 33.60]
3,7-DMX (umol/L)	13.45 [5.46 ~ 30.40]
1,3,7-TMX (umol/L)	3.08 [0.77 ~ 8.24]
AAMU (uM/L)	58.90 [18.90 ~ 133.00]
1-MU: 1-methyluric acid, 3-MU: 3-methyluric acid; 7-MU: 7-methyluric acid; 1,3-DMU: 1,3-dimethyluric acid; 1,7-DMU: 1,7-dimethyluric acid; 3,7-DMU: 3,7-dimethyluric acid; 1,3,7-TMU: 1,3,7-trimethyluric acid; 1-MX: 1-methylxanthine; 3-MX: 3-methylxanthine; 7-MX: 7-methylxanthine, 1,3-DMX: 1,3-dimethylxanthine, theophylline; 1,7-DMX: 1,7-dimethylxanthine, paraxanthine; 3,7-DMX: 3,7-dimethylxanthine, theobromine; 1,3,7-TMX: 1,3,7-trimethylxanthine, caffeine; AAMU: 5-acetylamino-6-amino-3-methyluracil

The characteristics of this study population are presented in Table 1. Males represented 53.55% of the overall sample. A majority of the participants were non-Hispanic white (39.60%). More than half (57.73%) of participants had an educational level above 12 y. The median (P25 ~ P75) age of participants was 48.00 [33.00 ~ 62.00] years. The median of non-log-transformed levels of urinary 1-MU, 3-MU, 7-MU, 1,3-DMU, 1,7-DMU, 3,7-DMU, 1,3,7-TMU, 1-MX, 3-MX, 7-MX, 1,3-DMX, 1,7-DMX, 3,7-DMX, 1,3,7-TMX, were 56.85, 0.43, 11.10, 6.28, 24.40, 0.73, 1.24, 26.30, 22.95, 37.55, 1.65, 16.10, 13.45, and 3.08 umol/L respectively, and levels of urinary AAMU was 58.90 uM/L.

3.2 Principal component of caffeine and caffeine metabolites

Figure 1

Spearman correlation analysis showed the relationship between urinary caffeine and caffeine metabolites, all coefficients (rs) are significant (Fig. 1). The main PCs derived from the PCA are presented in Table 2. The first two components had eigenvalues ≥ 1 and accounted for 90.67% of the variance. Table 3 shows the rotated component matrix of 15 urinary caffeine and caffeine metabolites with correlations greater than 0.5. The first component (PC1) was strongly correlated with 1-MU, 1,3-DMU, 1,7-DMU, 1,3,7-TMU, 1-MX, 1,3-DMX, 1,7-DMX, 1,3,7-TMX and AAMU. The second component (PC2) correlated with 3-MU, 7-MU, 3,7-DMU, 3-MX, 7-MX and 3,7-DMX.

Table 2

Total variance showed by components of principal component analysis
Component	Initial eigenvalues
Component	Total	% of variance	Cumulative %
1	11.07	73.77	73.77
2	2.54	16.91	90.67
3	0.59	3.92	94.59
4	0.23	1.56	96.14
5	0.14	0.90	97.04
6	0.10	0.65	97.69
7	0.09	0.59	98.27
8	0.08	0.52	98.79
9	0.05	0.36	99.15
10	0.04	0.28	99.43
11	0.03	0.17	99.60
12	0.02	0.14	99.74
13	0.02	0.11	99.85
14	0.02	0.10	99.95
15	0.01	0.05	100.00

Table 3

Rotated component matrix of 15 urinary caffeine and caffeine metabolites
Caffeine metabolites	PC1	PC2
1-MU (umol/L)	0.830
3-MU (umol/L)		0.845
7-MU (umol/L)		0.925
1,3-DMU (umol/L)	0.914
1,7-DMU (umol/L)	0.922
3,7-DMU (umol/L)		0.93
1,3,7-TMU (umol/L)	0.919
1-MX (umol/L)	0.884
3-MX (umol/L)		0.914
7-MX (umol/L)		0.942
1,3-DMX umol/L	0.933
1,7-DMX umol/L	0.901
3,7-DMX umol/L		0.857
1,3,7-TMX umol/L	0.910
AAMU (uM/L)	0.872
1-MU: 1-methyluric acid, 3-MU: 3-methyluric acid; 7-MU: 7-methyluric acid; 1,3-DMU: 1,3-dimethyluric acid; 1,7-DMU: 1,7-dimethyluric acid; 3,7-DMU: 3,7-dimethyluric acid; 1,3,7-TMU: 1,3,7-trimethyluric acid; 1-MX: 1-methylxanthine; 3-MX: 3-methylxanthine; 7-MX: 7-methylxanthine, 1,3-DMX: 1,3-dimethylxanthine, theophylline; 1,7-DMX: 1,7-dimethylxanthine, paraxanthine; 3,7-DMX: 3,7-dimethylxanthine, theobromine; 1,3,7-TMX: 1,3,7-trimethylxanthine, caffeine; AAMU: 5-acetylamino-6-amino-3-methyluracil

Table 2

Table 3

3.3 Univariate logistic regression associations between MetS with PC scores of urinary caffeine metabolites levels

We used PC1 and PC2 as independent variables, and test their associations with MetS and its components (Table 4). Effect sizes were different for the determinants across the two PCs. PC1 had significant positive associations with MetS and all components (all P values < 0.001), and the models indicate that an increase of PC1 leads to a 27%, 25%, 17%, 23%, 28% and 13% higher likelihood of MetS, raised FPG, raised BP, central obesity, raised TG and reduced HDL-C, respectively. While PC2 had only significant positive associations with MetS and its components of central obesity (MetS: OR = 1.11, 95% CI: 1.01–1.22; central obesity: OR = 1.16, 95% CI: 1.08–1.25).

Table 4

Univariate logistic regression analysis of PC scores of Log10-transformed urinary caffeine and caffeine metabolites with MetS and its components
Variables	OR (95% CI)	P value
MetS
PC 1	1.27 (1.15 ~ 1.41)	< 0.001
PC 2	1.11 (1.01 ~ 1.22)	0.030
Raised FPG
PC 1	1.25 (1.12 ~ 1.39)	< 0.001
PC 2	0.96 (0.87 ~ 1.05)	0.352
Raised BP
PC 1	1.17 (1.08 ~ 1.26)	< 0.001
PC 2	0.96 (0.89 ~ 1.03)	0.268
Central obesity
PC 1	1.23 (1.14 ~ 1.33)	< 0.001
PC 2	1.16 (1.08 ~ 1.25)	< 0.001
Raised TG
PC 1	1.28 (1.16 ~ 1.42)	< 0.001
PC 2	0.94 (0.86 ~ 1.03)	0.192
Reduced HDL-C
PC 1	1.13 (1.04 ~ 1.21)	0.002
PC 2	1.03 (0.96 ~ 1.11)	0.347
PC1: 1-MU; 1,3-DMU; 1,7-DMU; 1,3,7-TMU; 1-MX; 1,3-DMX; 1,7-DMX; 1,3,7-TMX; AAMU. PC2: 3-MU; 7-MU; 3,7-DMU; 3-MX; 7-MX; 3,7-DMX

Table 4

3.4 Multiple logistic regression associations between MetS with PC scores of urinary caffeine metabolites levels

In the multiple logistic regression models, effect sizes for the association of PC1 and PC2 with MetS and its components are given in Table 5. Positive associations were present between the PC1 and MetS (OR = 1.16, 95% CI: 1.01–1.34), and central obesity (OR = 1.22, 95% CI: 1.10–1.36). PC2 had significant positive associations with MetS and central obesity, but had significant negative association raised TG (all P values < 0.001).

Table 5

Multiple logistic regression analysis of PC scores of log-transformed urinary caffeine and caffeine metabolites with MetS and its components
Variables	OR (95% CI)	P value
MetS
PC1	1.16 (1.01 ~ 1.34)	0.042
PC2	1.15 (1.01 ~ 1.31)	0.035
Raised FPG
PC1	1.15 (0.98 ~ 1.34)	0.094
PC2	0.90 (0.78 ~ 1.04)	0.155
Raised BP
PC1	1.02 (0.90 ~ 1.15)	0.795
PC2	0.94 (0.84 ~ 1.05)	0.285
Central obesity
PC1	1.22 (1.10 ~ 1.36)	< 0.001
PC2	1.15 (1.04 ~ 1.27)	0.005
Raised TG
PC1	1.07 (0.92 ~ 1.24)	0.375
PC2	0.84 (0.73 ~ 0.95)	0.008
Reduced HDL-C
PC1	0.93 (0.84 ~ 1.03)	0.173
PC2	1.00 (0.91 ~ 1.11)	0.949
PC1: 1-MU; 1,3-DMU; 1,7-DMU; 1,3,7-TMU; 1-MX; 1,3-DMX; 1,7-DMX; 1,3,7-TMX; AAMU. PC2: 3-MU; 7-MU; 3,7-DMU; 3-MX; 7-MX; 3,7-DMX

Table 5

3.5 Dose-response associations between MetS with PC scores of urinary caffeine metabolites levels

Figure 2 shows the non-linear associations between the risk of PC1 and MetS and its components. A to F show the RCS curves without adjusting for covariates and G to J show the RCS curves with adjusting for covariates between PC1 and the risk of MetS, raised FPG, raised BP, central obesity, raised TG and reduced HDL-C, respectively. We observed U-shaped associations between the risk of PC1 and raised TG in univariate and multiple RCS logistic regression models (Fig. 2E and K, all values of P_overall and P_non−linear < 0.05). No statistically significant non-linear associations between PC1 and the risk of MetS and its other components in univariate and multiple RCS logistic regression models were observed.

Figure 2

Figure 3 shows the non-linear associations between PC2 and the risk of MetS and its components. A to F show the RCS curves without adjusting for covariates and G to J show the RCS curves with adjusting for covariates for PC2 and the risk of MetS and its components. In the univariate and multiple RCS logistic regression models, no non-linear statistical associations between PC2 and the risk of MetS and its components were observed (Fig. 3A-L).

Figure 3

Caffeine is a xanthine alkaloid compound, which mainly exists in coffee, tea, cola, dark chocolate and some analgesics (Górecki &Hallmann 2020, Jee et al. 2020). Because of its effects of reducing fatigue, invigorating the spirit, strengthening the stomach and strengthening the heart, caffeine-containing diet is popular worldwide (Ősz et al. 2022). Caffeine is a widely used psychoactive substance, often used as a stimulant of the central nervous system (YM et al. 2022). However, in recent years, more and more research and epidemiological data have confirmed that caffeine also plays an important role in energy metabolism (Du et al. 2018, Tabrizi et al. 2019). In this study, we conducted a principal component analysis of 15 urinary caffeine and caffeine metabolites, and explored the linear and nonlinear associations between PCs and the risk of Mets and its components among adults using data from NHANES 2011–2014. The PCA of 15 urinary caffeine and caffeine metabolites identified two main components that explained 90.67% of the total variance. Our study found that in the univariate logistic regression models, PC1 (strongly correlated with 1-MU, 1,3-DMU, 1,7-DMU, 1,3,7-TMU, 1-MX, 1,3-DMX, 1,7-DMX, 1,3,7-TMX and AAMU) positively correlated with the risk of MetS and all its components, while in the multiple logistic regression models, and it only positively correlated with the risk of MetS and central obesity. In the univariate logistic regression models, there were significant associations between PC2 (correlated with 3-MU, 7-MU, 3,7-DMU, 3-MX, 7-MX and 3,7-DMX) and MetS and central obesity, while in the multiple logistic regression models, PC2 was positive correlated with the risk of MetS and central obesity, and negative correlated with raised TG. Moreover, we observed U-shaped associations between PC1 and raised TG in univariate and multiple logistic regression models. Our study further clarified the associations between caffeine exposure and the risk of MetS and its components, provided clues for the follow-up cohort study and mechanism study, and also provided a theoretical basis for the prevention and control of MetS.

To date, there has been little agreement on the associations between caffeine intake and Mets. It has previously been observed that in univariate and multiple logistic regression analyses, compared with non-coffee drinkers, men who drink ≥ 4 cups of coffee a day have a lower risk of developing MetS in Japanese (Matsuura et al. 2012), this finding is consistent with another Japanese study that also found coffee consumption was inversely correlated with the risk of MetS (Takami et al. 2013). These findings are contrary to another Korean study, which demonstrated that coffee consumption (particularly instant coffee mix) may have harmful effects on MetS (Kim et al. 2014). Our research shows that PC1 and PC2 were positively correlated with the risk of MetS in both univariate and multiple logistic models, this is consistent with the results of a Korean research. However, no previous study has considered urinary caffeine and caffeine metabolites in relation to MetS. As a result, it is difficult to compare our results with those of previous studies.

MetS is a group of symptoms characterized by obesity, dyslipidemia, elevated BP and impaired glucose regulation, the relationships between caffeine (and caffeine metabolites) and MetS are affected by the components of MetS. At present, some studies have explored the relationships between caffeine and the risk of hypertension (Guessous et al. 2015, Hartley et al. 2000, Robertson et al. 1981), but the results are still controversial. Studies such as that conducted by Idris Guessous have shown that inverse associations were observed for caffeine, 1,7-DMX and 1,3-DMX with 24-hour and night-time systolic blood pressure (SBP), no associations of 3,7-DMX levels with 24-h or night-time ambulatory SBP were observed (Guessous et al. 2015). It has previously been observed that total plasma caffeine and 1,7-DMX at 10 to 13 weeks were inversely associated with glucose, this finding was contrary to our study which had suggested that in the univariate model, PC1 is positively correlated with raised BP, our study is consistent with T R Hartley’s results which showed that caffeine raised both systolic and diastolic BP (Hartley et al. 2000). We speculate that since caffeine is a methylxanthine, and myocardial contractility can be enhanced by methylxanthines (Robertson et al. 1981), this will increase cardiac output, thus raising BP. However, in the multiple logistic regression models, our study did not find statistical associations between caffeine and raised BP both in PC1 and PC2.

Data from the US adult study suggest that caffeine and its metabolites were positively related to insulin resistance (Lee et al. 2020). This is consistent with our results which demonstrated that PC1 was positively associated with raised FPG in univariate logistic regression analysis. Previous studies have shown that caffeine intake reduces insulin sensitivity in the short term (e.g., a 15% reduction after intaking a dose of 3 mg per kilogram of body weight) (Keijzers et al. 2002). This may reflect a partly promoting effect of caffeine on the increased epinephrine release which may decrease the storage of glucose as glycogen in muscle (van Dam et al. 2020).

Prior study has shown that coffee consumption was not associated with the incidence of central obesity (Wong et al. 2022). In contrast to this finding, a Mendelian randomization study showed that high coffee intake was associated with a low risk of obesity (Nordestgaard et al. 2015). A Korean study suggested that the elderly who consume less than one cup of coffee per day had a greater risk of sarcopenic obesity than those who consume more than three cups per day (Lee &Shin 2023). Up to now, far too little attention has been paid to the associations between caffeine and caffeine metabolites and central obesity. However, high levels of coffee consumption often indicate higher levels of caffeine and caffeine metabolites in the body, suggesting that this result is contrary to the results of our study. Our study observed that there were positive associations between PC1/PC2 and central obesity in both univariate and multiple logistic regression models. Although previous studies suggest that caffeine may improve energy balance by reducing appetite and increasing the basal metabolic rate and food-induced thermogenesis (Harpaz et al. 2017). However, caffeine intake is often accompanied by high energy intake, which may lead to excessive weight gain.

Some epidemiological studies were fuscous on the relationships between coffee and TG and HDL-C. A study conducted by Australian researchers suggested that coffee consumption was not associated with the risk of high TG or low HDL-C (Wong et al. 2022), and another study indicated no significant association between the consumption of coffee and serum lipid levels (Karabudak et al. 2015). However, in a Japan multi-institutional collaborative cohort study showed that coffee consumption was associated with lower serum TG levels (Takami et al. 2013). Besides, a Korean adult study demonstrated that in women, the prevalence of elevated TG and reduced HDL-C were significantly lower compared to non-coffee consumers (Kim &Shin 2019). An ELSA-Brasil study study indicated that more than 3 cups per day of coffee consumption was associated with a TG level increase (Miranda et al. 2022). Our study found a negative association between PC2 (strongly correlated with 3-MU, 7-MU, 3,7-DMU, 3-MX, 7-MX and 3,7-DMX) and raised TG in multiple logistic regression analysis. Regarding PC1(strongly correlated with 1-MU, 1,3-DMU, 1,7-DMU, 1,3,7-TMU, 1-MX, 1,3-DMX, 1,7-DMX, 1,3,7-TMX and AAMU), although no linear relationship between PC1 and raised TG was observed, we observed a nonlinear positive relationship between PC1 and raised TG in the univariate and multiple logistic regression models. This indicates that the associations between caffeine metabolites and raised TG were not simple upward or downward relationships, but complex nonlinear relationships. In the previous randomized trial, high consumption of unfiltered coffee (median, 6 cups per day) increased low-density lipoprotein cholesterol levels by 17.8 mg per deciliter as compared with filtered coffee (Jee et al. 2001), and unfiltered coffee contains cafestol and kahweol, while cafestol and kahweol raise the serum concentration of cholesterol and TG in humans (Urgert &Katan 1997), this may explain why caffeine and its metabolites positively associated with raise TG and reduced HDL-C.

The main limitations of this study are as follows: firstly, as a cross-sectional study, the causal relationship between urinary caffeine and its metabolites and MetS is difficult to elucidate, and further prospective research is needed to verify. Secondly, the participants in our study are all Americans, so the results of this study are not generally applicable to people in Asia or other parts of the world, but they still have certain reference values because the NHANES cycle also includes some non-Hispanic Asians.

Strengths of our study are included as follows: first, different types of coffee could substantially vary in the content of biological compounds, and urinary caffeine and caffeine metabolites are effective measures of caffeine intake (Guessous et al. 2015, Nehlig 2018, Oñatibia-Astibia et al. 2016), therefore, our study reflected the relationships between coffee and the risk of MetS more accurately compared with previous studies. Second, we use PCs as independent variables and MetS (and also its components) as dependent variables for the analysis, this to some extent could avoid the collinearity between caffeine and 14 kinds of caffeine metabolites. Third, besides exploring the linear associations between caffeine and its metabolites and the risk of MetS, our study further explored the nonlinear relationships between caffeine and the risk of MetS.

To conclude, the elements of PC1(1-MU, 1,3-DMU, 1,7-DMU, 1,3,7-TMU, 1-MX, 1,3-DMX, 1,7-DMX, 1,3,7-TMX and AAMU) were positively correlated with MetS and central obesity. The elements of PC2 (3-MU, 7-MU, 3,7-DMU, 3-MX, 7-MX and 3,7-DMX) were positively correlated with the risk of MetS, central obesity and negatively correlated with raised TG. The U-shaped associations exist between PC1 and the risk of elevated TG in univariate and multiple RCS logistic regression models. Further prospective studies are required to confirm the causal associations between caffeine metabolites and MetS. Our study provides empirical support for formulating an acceptable concentration range of caffeine metabolites in urine in the future and offers a new concept for preventing the occurrence of MetS.

Ethics approval and consent to participate: This study uses only secondary data analyses without any personal information identified using statistical data from the NHANES website, no further ethical approval for conducting the present study is required.

Consent for publication: Not applicable.

Authors' contributions: Linyuan Qin: conceptualization, methodology, formal analysis, resources, data curation, writing—original draft, writing—review and editing, and visualization. Jianli Zhou: conceptualization, methodology, formal analysis, validation, writing—review and editing, and visualization. All authors read and approved the final manuscript.

Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Competing interests: The authors declare that they have no conflict of interest.

Availability of data and material: The datasets analyzed during the current study are available in the NHANES repository. (https://www.cdc.gov/nchs/nhanes/index.htm).

Acknowledgments: None

Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, Fruchart JC, James WP, Loria CM, Smith SC, Jr. (2009) Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation. 120: 1640-5
Després JP, Lemieux I (2006) Abdominal obesity and metabolic syndrome. Nature. 444: 881-7
Du X, Huang Q, Guan Y, Lv M, He X, Fang C, Wang X, Sheng J (2018) Caffeine Promotes Conversion of Palmitic Acid to Palmitoleic Acid by Inducing Expression of fat-5 in Caenorhabditis elegans and scd1 in Mice. Front Pharmacol. 9: 321
Eckel RH, Grundy SM, Zimmet PZ (2005) The metabolic syndrome. Lancet. 365: 1415-28
Górecki M, Hallmann E (2020) The Antioxidant Content of Coffee and Its In Vitro Activity as an Effect of Its Production Method and Roasting and Brewing Time. Antioxidants (Basel). 9
Guessous I, Pruijm M, Ponte B, Ackermann D, Ehret G, Ansermot N, Vuistiner P, Staessen J, Gu Y, Paccaud F, Mohaupt M, Vogt B, Pechère-Bertschi A, Martin PY, Burnier M, Eap CB, Bochud M (2015) Associations of ambulatory blood pressure with urinary caffeine and caffeine metabolite excretions. Hypertension. 65: 691-6
Harpaz E, Tamir S, Weinstein A, Weinstein Y (2017) The effect of caffeine on energy balance. J Basic Clin Physiol Pharmacol. 28: 1-10
Hartley TR, Sung BH, Pincomb GA, Whitsett TL, Wilson MF, Lovallo WR (2000) Hypertension risk status and effect of caffeine on blood pressure. Hypertension. 36: 137-41
Jee HJ, Lee SG, Bormate KJ, Jung YS (2020) Effect of Caffeine Consumption on the Risk for Neurological and Psychiatric Disorders: Sex Differences in Human. Nutrients. 12
Jee SH, He J, Appel LJ, Whelton PK, Suh I, Klag MJ (2001) Coffee consumption and serum lipids: a meta-analysis of randomized controlled clinical trials. Am J Epidemiol. 153: 353-62
Jin R, Zhu X, Shrubsole MJ, Yu C, Xia Z, Dai Q (2018) Associations of renal function with urinary excretion of metals: Evidence from NHANES 2003-2012. Environ Int. 121: 1355-1362
Karabudak E, Türközü D, Köksal E (2015) Association between coffee consumption and serum lipid profile. Exp Ther Med. 9: 1841-1846
Keijzers GB, De Galan BE, Tack CJ, Smits P (2002) Caffeine can decrease insulin sensitivity in humans. Diabetes Care. 25: 364-9
Kim HJ, Cho S, Jacobs DR, Jr., Park K (2014) Instant coffee consumption may be associated with higher risk of metabolic syndrome in Korean adults. Diabetes Res Clin Pract. 106: 145-53
Kim SA, Shin S (2019) The Association between Coffee Consumption Pattern and Prevalence of Metabolic Syndrome in Korean Adults. Nutrients. 11
Lam DW, LeRoith D (2000) Metabolic Syndrome. Circulation
Lee DY, Shin S (2023) Sarcopenic obesity is associated with coffee intake in elderly Koreans. Front Public Health. 11: 990029
Lee S, Min JY, Min KB (2020) Caffeine and Caffeine Metabolites in Relation to Insulin Resistance and Beta Cell Function in U.S. Adults. Nutrients. 12
Lemieux I, Després JP (2020) Metabolic Syndrome: Past, Present and Future. Nutrients. 12
Lopez-Candales A, Hernández Burgos PM, Hernandez-Suarez DF, Harris D (2017) Linking Chronic Inflammation with Cardiovascular Disease: From Normal Aging to the Metabolic Syndrome. J Nat Sci. 3
Marventano S, Salomone F, Godos J, Pluchinotta F, Del Rio D, Mistretta A, Grosso G (2016) Coffee and tea consumption in relation with non-alcoholic fatty liver and metabolic syndrome: A systematic review and meta-analysis of observational studies. Clin Nutr. 35: 1269-1281
Matsuura H, Mure K, Nishio N, Kitano N, Nagai N, Takeshita T (2012) Relationship between coffee consumption and prevalence of metabolic syndrome among Japanese civil servants. J Epidemiol. 22: 160-6
Miranda AM, Goulart AC, Generoso G, Bittencourt MS, Santos RD, Toth PP, Jones SR, Benseñor IM, Lotufo PA, Marchioni DM (2022) Association between coffee consumption with serum lipid profile in ELSA-Brasil study: a metabolomic approach. Eur J Nutr. 61: 4205-4214
Moore JX, Chaudhary N, Akinyemiju T (2017) Metabolic Syndrome Prevalence by Race/Ethnicity and Sex in the United States, National Health and Nutrition Examination Survey, 1988-2012. Prev Chronic Dis. 14: E24
Nehlig A (2018) Interindividual Differences in Caffeine Metabolism and Factors Driving Caffeine Consumption. Pharmacol Rev. 70: 384-411
Ngueta G (2020) Caffeine and caffeine metabolites in relation to hypertension in U.S. adults. Eur J Clin Nutr. 74: 77-86
Nordestgaard AT, Thomsen M, Nordestgaard BG (2015) Coffee intake and risk of obesity, metabolic syndrome and type 2 diabetes: a Mendelian randomization study. Int J Epidemiol. 44: 551-65
Oñatibia-Astibia A, Martínez-Pinilla E, Franco R (2016) The potential of methylxanthine-based therapies in pediatric respiratory tract diseases. Respir Med. 112: 1-9
Ősz BE, Jîtcă G, Ștefănescu RE, Pușcaș A, Tero-Vescan A, Vari CE (2022) Caffeine and Its Antioxidant Properties-It Is All about Dose and Source. Int J Mol Sci. 23
Ponte B, Pruijm M, Ackermann D, Ehret G, Ansermot N, Staessen JA, Vogt B, Pechère-Bertschi A, Burnier M, Martin PY, Eap CB, Bochud M, Guessous I (2018) Associations of Urinary Caffeine and Caffeine Metabolites With Arterial Stiffness in a Large Population-Based Study. Mayo Clin Proc. 93: 586-596
Ridder M (2023): Global coffee consumption 2012/13-2020/21. statista. https://www.statista.com/statistics/292595/global-coffee-consumption/. Accessed 18 May 2023
Robertson D, Wade D, Workman R, Woosley RL, Oates JA (1981) Tolerance to the humoral and hemodynamic effects of caffeine in man. J Clin Invest. 67: 1111-7
Shang F, Li X, Jiang X (2016) Coffee consumption and risk of the metabolic syndrome: A meta-analysis. Diabetes Metab. 42: 80-7
Tabrizi R, Saneei P, Lankarani KB, Akbari M, Kolahdooz F, Esmaillzadeh A, Nadi-Ravandi S, Mazoochi M, Asemi Z (2019) The effects of caffeine intake on weight loss: a systematic review and dos-response meta-analysis of randomized controlled trials. Crit Rev Food Sci Nutr. 59: 2688-2696
Takami H, Nakamoto M, Uemura H, Katsuura S, Yamaguchi M, Hiyoshi M, Sawachika F, Juta T, Arisawa K (2013) Inverse correlation between coffee consumption and prevalence of metabolic syndrome: baseline survey of the Japan Multi-Institutional Collaborative Cohort (J-MICC) Study in Tokushima, Japan. J Epidemiol. 23: 12-20
Urgert R, Katan MB (1997) The cholesterol-raising factor from coffee beans. Annu Rev Nutr. 17: 305-24
Van Dam RM, Hu FB, Willett WC (2020) Coffee, Caffeine, and Health. N Engl J Med. 383: 369-378
Weng Z, Xu C, Xu J, Jiang Z, Liu Q, Liang J, Gu A (2021) Association of urinary caffeine and caffeine metabolites with cardiovascular disease risk in adults. Nutrition. 84: 111121
Wong THT, Wong CH, Zhang X, Zhou Y, Xu J, Yuen KC, Wan JMF, Louie JCY (2021) The Association Between Coffee Consumption and Metabolic Syndrome in Adults: A Systematic Review and Meta-Analysis. Adv Nutr. 12: 708-721
Wong THT, Burlutsky G, Gopinath B, Flood VM, Mitchell P, Louie JCY (2022) The longitudinal association between coffee and tea consumption and the risk of metabolic syndrome and its component conditions in an older adult population. J Nutr Sci. 11: e79
Yeomans KA, Golder PA (1982) The Guttman-Kaiser Criterion as a Predictor of the Number of Common Factors. Journal of the Royal Statistical Society. Series D (The Statistician). 31: 221-229
YM MY, Waldvogel HJ, Faull RLM, Kwakowsky A (2022) Neuroprotective Effect of Caffeine in Alzheimer's Disease. Molecules. 27
Zhou J, Meng X, Deng L, Liu N (2022) Non-linear associations between metabolic syndrome and four typical heavy metals: Data from NHANES 2011-2018. Chemosphere. 291: 132953

Download PDF

Editorial decision: Major Revision
18 Jul, 2023
Reviewers agreed at journal
27 Jun, 2023
Reviewers invited by journal
27 Jun, 2023
Editor invited by journal
23 Jun, 2023
Editor assigned by journal
24 May, 2023
First submitted to journal
18 May, 2023

You are reading this latest preprint version

Associations of urinary caffeine and caffeine metabolites with metabolic syndrome in U.S. adults

Status:

Version 1

Abstract

Figures

1. Introduction

2. Materials and methods

2.1 Study population

2.2 Measurement of urinary caffeine and caffeine metabolites levels

2.3 Definition of MetS

2.4 Covariates

2.5 Statistical analysis

3. Results

3.1. Basic characteristics of the study population

3.2 Principal component of caffeine and caffeine metabolites

3.3 Univariate logistic regression associations between MetS with PC scores of urinary caffeine metabolites levels

3.4 Multiple logistic regression associations between MetS with PC scores of urinary caffeine metabolites levels

3.5 Dose-response associations between MetS with PC scores of urinary caffeine metabolites levels

4. Discussion

Declarations

References

Status:

Version 1