Smoking, Alcohol and Caffeine and Ovarian Age During the Reproductive Years
Smoking, Alcohol and Caffeine and Ovarian Age During the Reproductive Years
Background: We sought to determine whether smoking, alcohol and caffeine are related to four indicators of ovarian age: antral follicle count (AFC), follicle stimulating hormone (FSH), inhibin B and estradiol.
Methods: Analyses drew on ultrasound scans and sera from 188 women, aged 2249. We used least squares regression to estimate differences in AFC and hormone levels for women who smoke cigarettes or who drink alcohol or caffeine.
Results: Current smoking is related to elevated FSH (β for ln(FSH) = 0.21, 95% CI 0.04, 0.39), but not to AFC, inhibin B or estradiol. Neither alcohol nor caffeine is related to any ovarian age indicator. Exploratory analyses suggest that the association of current smoking with FSH varies with age: comparing current with never smokers, at ages 30, 35, 40 and 45, estimated differences in mean FSH are 0.3, 1.3, 3.2 and 6.9 mIU/ml.
Conclusions: The association of current smoking with FSH may reflect accelerated oocyte atresia, impaired follicle quality or dysregulation of the hypothalamicpituitaryovarian axis. Identification of the causal mechanism has implications for prevention or treatment of conception delay, infertility and morbidity associated with early menopause.
The ability to conceive and to carry a pregnancy to term lessens as women grow older. At any given age, however, women vary in their fecundability and fertility. For women in their 30s and 40s, follicular and ovarian, rather than uterine, factors may be the primary determinants of fecundity (CDC, 2000; Toner et al., 2002). It seems reasonable to infer, then, that variability in fecundability derives from differences in ovarian age that are tied to, but not entirely determined by, chronologic age. This paper examines whether or not cigarette smoking and, secondarily, two common exposures often correlated with smoking, alcohol and caffeine intake, are related to three indicators of ovarian age, namely, antral follicle count (AFC) and levels of follicle stimulating hormone (FSH) and inhibin B. We also examine associations with estradiol.
Several observations suggest that cigarette smoking is associated with older ovarian age. Current smoking is associated with earlier age at menopause (reviewed by Harlow and Signorello, 2000) and a shortened menopausal transition (McKinlay et al., 1992; Cooper et al., 1995). In a sample of women aged 3554, the number of ovarian follicles is reduced among women who ever smoked (Westhoff et al., 2000). Smoking has been linked to increased levels of FSH (Cramer et al., 1994; Cooper et al., 1995; Backer et al., 1999; Cramer et al., 2002) and decreased levels of estrogen (MacMahon et al., 1982; Barbieri et al., 1986; Westhoff et al., 1996). Of two studies examining a possible relation between smoking and inhibin B, one (Freeman et al., 2005) showed no association, whereas the other (Lambert-Messerlian and Harlow, 2006) showed decreased levels among current smokers. Almost all of these observations derive from samples of women nearing the end of their reproductive prime (mid30s or older). It is unclear whether or not the findings extend to younger women.
The observed associations between current smoking and indicators of advanced ovarian age might stem from effects on: oocyte quantity, e.g. accelerated oocyte atresia (Mattison and Thorgeirsson, 1978); oocyte quality, e.g. intrafollicular oxidative stress (Paszkowski et al., 2002) or disruption of endocrine function, e.g. activation of the aryl hydrocarbon receptor (reviewed by Valdez and Petroff, 2004). If smoking diminishes the size of the underlying oocyte pool, we might expect the number of recruited antral follicles to decrease accordingly, with a concomitant decrease in levels of inhibin B, a product of the antral follicles. Alternatively, if smoking alters oocyte or endocrine function, impairment might manifest in increased levels of FSH or decreased production of inhibin B even if no change is evident in the number of antral follicles. In either scenario, a critical threshold of the oocyte pool might confine associations to women of older chronologic age.
Evidence relating alcohol and caffeine to ovarian age is sparse. For alcohol, primary observations relate to a possible association with increased estrogen levels (Valimaki et al., 1983; Mendelson et al., 1987, 1988, 1989; Gavaler et al., 1987; Reichman et al., 1993; Muti et al., 1998), a result consistent with studies suggesting that age at menopause may be later among women who drink alcohol (Torgerson et al., 1997; Cooper et al., 2001; Brett and Cooper, 2003; Kinney et al., 2006). For caffeine, one study (Lucero et al., 2001) of women aged 3645 shows an association between caffeine and increased estradiol levels. Other evidence derives from studies (Wilcox et al., 1988; Christiansonet al., 1989; Williams et al., 1990; Grodstein et al., 1993; Hatch and Bracken, 1993; Spinelli et al., 1997) that show an association with infertility [although not all studies (Joesoef et al., 1990; Florack et al., 1994; Curtis et al., 1997) confirm this finding].
Background: We sought to determine whether smoking, alcohol and caffeine are related to four indicators of ovarian age: antral follicle count (AFC), follicle stimulating hormone (FSH), inhibin B and estradiol.
Methods: Analyses drew on ultrasound scans and sera from 188 women, aged 2249. We used least squares regression to estimate differences in AFC and hormone levels for women who smoke cigarettes or who drink alcohol or caffeine.
Results: Current smoking is related to elevated FSH (β for ln(FSH) = 0.21, 95% CI 0.04, 0.39), but not to AFC, inhibin B or estradiol. Neither alcohol nor caffeine is related to any ovarian age indicator. Exploratory analyses suggest that the association of current smoking with FSH varies with age: comparing current with never smokers, at ages 30, 35, 40 and 45, estimated differences in mean FSH are 0.3, 1.3, 3.2 and 6.9 mIU/ml.
Conclusions: The association of current smoking with FSH may reflect accelerated oocyte atresia, impaired follicle quality or dysregulation of the hypothalamicpituitaryovarian axis. Identification of the causal mechanism has implications for prevention or treatment of conception delay, infertility and morbidity associated with early menopause.
The ability to conceive and to carry a pregnancy to term lessens as women grow older. At any given age, however, women vary in their fecundability and fertility. For women in their 30s and 40s, follicular and ovarian, rather than uterine, factors may be the primary determinants of fecundity (CDC, 2000; Toner et al., 2002). It seems reasonable to infer, then, that variability in fecundability derives from differences in ovarian age that are tied to, but not entirely determined by, chronologic age. This paper examines whether or not cigarette smoking and, secondarily, two common exposures often correlated with smoking, alcohol and caffeine intake, are related to three indicators of ovarian age, namely, antral follicle count (AFC) and levels of follicle stimulating hormone (FSH) and inhibin B. We also examine associations with estradiol.
Several observations suggest that cigarette smoking is associated with older ovarian age. Current smoking is associated with earlier age at menopause (reviewed by Harlow and Signorello, 2000) and a shortened menopausal transition (McKinlay et al., 1992; Cooper et al., 1995). In a sample of women aged 3554, the number of ovarian follicles is reduced among women who ever smoked (Westhoff et al., 2000). Smoking has been linked to increased levels of FSH (Cramer et al., 1994; Cooper et al., 1995; Backer et al., 1999; Cramer et al., 2002) and decreased levels of estrogen (MacMahon et al., 1982; Barbieri et al., 1986; Westhoff et al., 1996). Of two studies examining a possible relation between smoking and inhibin B, one (Freeman et al., 2005) showed no association, whereas the other (Lambert-Messerlian and Harlow, 2006) showed decreased levels among current smokers. Almost all of these observations derive from samples of women nearing the end of their reproductive prime (mid30s or older). It is unclear whether or not the findings extend to younger women.
The observed associations between current smoking and indicators of advanced ovarian age might stem from effects on: oocyte quantity, e.g. accelerated oocyte atresia (Mattison and Thorgeirsson, 1978); oocyte quality, e.g. intrafollicular oxidative stress (Paszkowski et al., 2002) or disruption of endocrine function, e.g. activation of the aryl hydrocarbon receptor (reviewed by Valdez and Petroff, 2004). If smoking diminishes the size of the underlying oocyte pool, we might expect the number of recruited antral follicles to decrease accordingly, with a concomitant decrease in levels of inhibin B, a product of the antral follicles. Alternatively, if smoking alters oocyte or endocrine function, impairment might manifest in increased levels of FSH or decreased production of inhibin B even if no change is evident in the number of antral follicles. In either scenario, a critical threshold of the oocyte pool might confine associations to women of older chronologic age.
Evidence relating alcohol and caffeine to ovarian age is sparse. For alcohol, primary observations relate to a possible association with increased estrogen levels (Valimaki et al., 1983; Mendelson et al., 1987, 1988, 1989; Gavaler et al., 1987; Reichman et al., 1993; Muti et al., 1998), a result consistent with studies suggesting that age at menopause may be later among women who drink alcohol (Torgerson et al., 1997; Cooper et al., 2001; Brett and Cooper, 2003; Kinney et al., 2006). For caffeine, one study (Lucero et al., 2001) of women aged 3645 shows an association between caffeine and increased estradiol levels. Other evidence derives from studies (Wilcox et al., 1988; Christiansonet al., 1989; Williams et al., 1990; Grodstein et al., 1993; Hatch and Bracken, 1993; Spinelli et al., 1997) that show an association with infertility [although not all studies (Joesoef et al., 1990; Florack et al., 1994; Curtis et al., 1997) confirm this finding].