The causal association between smoking initiation, alcohol and coffee consumption, and women's reproductive health: A two-sample Mendelian randomization analysis

Abstract

Objective: A number of epidemiological studies have demonstrated that smoking initiation and alcohol and coffee consumption were closely related to women's reproductive health. However, there was still insufficient evidence supporting their direct causality effect. Methods: We utilized two-sample Mendelian randomization (TSMR) analysis with summary datasets from genome-wide association study (GWAS) to investigate the causal relationship between smoking initiation, alcohol and coffee consumption, and women's reproductive health-related traits. Exposure genetic instruments were used as variants significantly related to traits. The inverse-variance weighted (IVW) method was used as the main analysis approach, and we also performed MR-PRESSO, MR-Egger, weighted median, and weighted mode to supplement the sensitivity test. Then, the horizontal pleiotropy was detected by using MRE intercept and MR-PRESSO methods, and the heterogeneity was assessed using Cochran's Q statistics. Results: We found evidence that smoking women showed a significant inverse causal association with the sex hormone-binding globulin (SHBG) levels (corrected β = -0.033, p = 9.05E-06) and age at menopause (corrected β = -0.477, p = 6.60E-09) and a potential positive correlation with the total testosterone (TT) levels (corrected β = 0.033, p = 1.01E-02). In addition, there was suggestive evidence for the alcohol drinking effect on the elevated TT levels (corrected β = 0.117, p = 5.93E-03) and earlier age at menopause (corrected β = -0.502, p = 4.14E-02) among women, while coffee consumption might decrease the female SHBG levels (corrected β = -0.034, p = 1.33E-03). Conclusion: Our findings suggested that smoking in women significantly decreased their SHBG concentration, promoted earlier menopause, and possibly reduced the TT levels. Alcohol drinking had a potential effect on female higher TT levels and earlier menopause, while coffee consumption might lead to lower female SHBG levels.

Keywords: Mendelian randomization; alcohol drinking; coffee consumption; menstrual health; reproduction-related hormones; smoking initiation.

FIGURE 1

MR results of causal effects of smoking initiation SNPs on related outcome risks. (The estimates were derived from the random-effects inverse-variance weighted (IVW) model, the MR pleiotropy residual sum and outlier (MR-PRESSO) approach, MR-Egger (MRE) regression, weighted median (WM), and weighted mode (Wm). (A–H): (A) sex hormone-binding globulin, SHBG; (B) bioavailable testosterone, bio-T; (C) total testosterone, TT; (D) estradiol, E₂; (E) anti-Mullerian hormone, AMH; (F) age at menopause; (G) irregular menstrual cycle/bleeding; and (H) dysmenorrhea. Outliers of MR-PRESSO: (A) rs56208390, rs3115418, rs13066050, rs7631379, rs3934797, rs1116690, rs7743165, rs9331343, rs76841737, rs11192347, rs644740, rs76460663, rs1106363, rs11611651, rs62052916, rs4476253, and rs7505855; (B) rs56208390, rs3115418, rs13066050, rs1549979, rs1116690, rs2173019, rs7024924, rs10969352, rs11192347, rs76460663, rs62052916, and rs4476253; (C) rs1518393, rs3115418, rs13066050, rs1549979, rs2173019, rs359431, rs13261666, rs7024924, rs10966092, rs10969352, rs11192347, and rs76460663; (D) N; (E) N; (F) rs160631, rs9331343, rs10966092, and rs56902655; (G) N; and (H) N. p-values < 5.00E-02 are presented in red, and red and bold font means p-values < 2.08E-03. β, beta; CI, confidence interval; OR, odds ratio; N, none).

FIGURE 2

MR results of causal effects of alcohol drinking SNPs on related outcome risks. (The estimates were derived from the random-effects IVW model, the MR-PRESSO approach, MRE regression, WM, and Wm. (A–H): (A) SHBG; (B) bio-T; (C) TT; (D) E₂; (E) AMH; (F) age at menopause; (G) irregular menstrual cycle/bleeding; and (H) dysmenorrhea. Outliers of MR-PRESSO: (A) rs705687, rs56337305, rs13094887, rs11940694, rs6460047, rs35034355, rs1217091, rs55932213, rs7074871, rs62044525, rs7185555, and rs2532276; (B) rs705687, rs56337305, rs13094887, rs4501255, rs35034355, rs1217091, rs28601761, rs55932213, rs11625650, and rs62044525; (C) rs705687, rs13094887, rs4501255, rs35034355, rs28601761, and rs11625650; (D) N; (E) N; (F) rs13094887; (G) N; and (H) N. p-values < 5.00E-02 are presented in red, and red and bold font means p-values < 2.08E-03. β, beta; CI, confidence interval; OR, odds ratio; N, none).

FIGURE 3

MR results of causal effects of coffee consumption SNPs on related outcome risks. (The estimates were derived from the random-effects IVW model, the MR-PRESSO approach, MRE regression, WM, and Wm. (A–H): (A) SHBG; (B) bio-T; (C) TT; (D) E₂; (E) AMH; (F) age at menopause; (G) irregular menstrual cycle/bleeding; and (H) dysmenorrhea. Outliers of MR-PRESSO: none or not available. p-values < 5.00E-02 are presented in red, and red and bold font means p-values < 2.08E-03. β, beta; CI, confidence interval; OR, odds ratio).