Endocrine Effects of Tobacco Smoking

Konstantinos Tziomalos; Faidon Charsoulis


Clin Endocrinol. 2004;61(6) 

In This Article

Gonadal and Reproductive Function

Tobacco smoking in adulthood has a marginal impact on spermatogenesis. Male fertility and spermatogenesis in adult men is surprisingly resistant to deleterious effects of tobacco smoke (Bonde & Storgaard, 2002). Male smoking has no strong impact on the probability of conception in a menstrual cycle (Bolumar et al., 1996). Nevertheless, according to a meta-analysis of 20 different study populations worldwide, the sperm count is reduced by 13–17% in smokers (Vine et al., 1994). Other investigators found significantly lower sperm motility and percentage of morphologically normal spermatozoa, as well as altered sperm morphometric parameters and sperm function tests in smokers compared to nonsmokers (Sofikitis et al., 1995). Furthermore, tobacco smoking might still be of importance if smoking during pregnancy has impact on the development of the foetal gonads. In utero exposure to the highly toxic polyaromatic hydrocarbons could destroy Sertoli cells or impair the replication of these cells. A reduced number of Sertoli cells is believed to be associated with reduced sperm output in adult life (Bonde & Storgaard, 2002). However, direct experimental evidence demonstrating the effects on gonadal development of tobacco smoke is lacking (Jensen et al., 1998).

Compared with nonsmokers and independent of relative weight and age, middle-aged male smokers have increased serum levels of DHEA, DHEAS, androstenedione, oestradiol, and SHBG, but normal total testosterone compared to nonsmokers (Attia et al., 1989; Field et al., 1994). Raised oestradiol may be among the mechanisms through which cigarette smoking impairs male reproduction. Of interest is the finding that smoking elevates oestradiol in men, while it lowers oestradiol in women. The finding of a lack of difference between smokers and nonsmokers in testosterone levels suggests that the steroidogenic function of the testis is not affected by smoking; on the other hand, smoking may affect the free fraction of testosterone (Attia et al., 1989).

Present evidence supports an adverse effect of smoking on ovarian function, which is prolonged and dose dependent (Shiverick & Salafia, 1999). Several mechanisms have been proposed to explain this effect. Extracts of both cigarette smoke and nicotine produce a direct inhibition of granulosa cell aromatase activity in vitro, and indirectly as well because of an associated reduction in adiposity (Byrne et al., 1991; Shiverick & Salafia, 1999). The observation that the 2-hydroxylation of oestradiol is increased in smokers has led to the proposal that increased catechol oestrogen formation is another mechanism for the anti-oestrogenic effect of smoking (Michnovicz et al., 1986; Baron et al., 1990; Shulman et al., 1990; Shiverick & Salafia, 1999). Furthermore, polyaromatic hydrocarbons in cigarette smoke may induce microsomal cytochrome P-450, which metabolizes steroid hormones, possibly enhancing the formation of catechol metabolites of oestradiol (Shiverick & Salafia, 1999). Cigarette alkaloids also inhibit progesterone synthesis both by inhibiting progesterone synthesis and by causing less specific cytotoxic effects (Gocze et al., 1999).

Smoking may also alter fertility through effects on uterine–fallopian tube functions, which mediate gamete and conceptus transport (Shiverick & Salafia, 1999). Hughes and Brennan (1996) conducted a meta-analysis to assess the effects of female and male smoking on natural and assisted fertilization. In 13 studies of natural conception, all but one demonstrated a negative association between smoking and time to conception. Smoking one pack of cigarettes per day and starting to smoke before 18 years of age were further associated with an increased risk of infertility, providing evidence of dose- and age-related effects on natural fertility.

Important new information is being learned from clinical in vitro fertilization and assisted reproduction technologies regarding the effects of smoking on fertility. In the previously quoted meta-analysis, seven studies of subfertile women undergoing assisted reproductive technologies showed small but consistent reductions in pregnancy rates among smokers. Women, in particular younger women, who smoke have a higher mean basal serum FSH concentration and require significantly higher mean dosage of gonadotrophins for ovarian stimulation than nonsmokers (Van Voorhis et al., 1996; El-Nemr et al., 1998). A lower mean number of oocytes is obtained in smokers than nonsmokers and the rate of abandoned cycles and total fertilization failure is higher (Van Voorhis et al., 1992, 1996; El-Nemr et al., 1998). Women who smoked during their treatment cycle had approximately a 50% reduction in implantation rate and ongoing pregnancy rate compared with women who had never smoked (Pattinson et al., 1991; Van Voorhis et al., 1996). In conclusion, cigarette smoking in women appears to significantly reduce their ovarian reserve and lead to poor response to ovarian stimulation at an earlier age (Sharara et al., 1994; Van Voorhis et al., 1996; El-Nemr et al., 1998). The relative hyperandrogenism of ovarian origin, which is apparent in smokers, with higher serum androstendione, DHEAS and testosterone/SHBG ratio, may contribute to the lower pregnancy rate at in vitro fertilization in this group (Byrne et al., 1991; Gustafson et al., 1996). Nevertheless, these adverse effects appear to be reversible in women who quit smoking before initiating treatment; these women have the same pregnancy rate as nonsmokers (Van Voorhis et al., 1996).

Maternal smoking during pregnancy is known to be associated with adverse pregnancy outcomes, including low birthweight, intrauterine growth retardation, premature delivery, spontaneous abortion, placental abruption, placenta praevia, perinatal mortality and ectopic pregnancy, especially in older mothers (Kistin et al., 1996; Ahluwalia et al., 1997; Shiverick & Salafia, 1999). There is also increased postnatal morbidity and mortality relating to deficits in pulmonary function and neurocognitive development (Shiverick & Salafia, 1999). These effects are proportionate to dose, starting with passive smoke (Shulman et al., 1990; Shiverick & Salafia, 1999). Furthermore, it seems that the exposure of the pregnant woman to passive smoking by her partner is also detrimental, as it results in a significant passage of the metabolites of nicotine through the placenta (Shulman et al., 1990). In addition, nicotine may be less rapidly metabolized in pregnant women. By contrast, an intriguing aspect of the association between smoking and reduced birthweight is the dose relationship and reversibility, in that women who cease smoking before 30 weeks gestation have heavier infants (Shiverick & Salafia, 1999).

The increased miscarriage rate among mothers who smoke may be related to direct adverse effects of smoke components such as nicotine, cadmium and polyaromatic hydrocarbons on trophoblast invasion and proliferation (Shiverick & Salafia, 1999). In addition, Jauniaux and Burton (1992) have described increased syncytial necrosis and increased thickness of the syncytio/cytotrophoblast membrane in early pregnancy in mothers who smoke. Syncytial damage may also explain the striking reduction in midtrimester maternal serum chorionic gonadotrophin levels reported in a study of 23 668 pregnancies (Palomaki et al., 1993).

During early pregnancy, smoking is associated with significantly depressed levels of oestriol, oestradiol, human chorionic gonadotrophin and human placental lactogen, which may explain certain adverse effects of smoking; there appears to be a steady decline in these values with increasing cigarette consumption (Bernstein et al., 1989; Shiverick & Salafia, 1999). It has also been proposed that the smoking-induced corpus luteal deficiency could underlie the increase in early pregnancy loss observed in smokers (Shiverick & Salafia, 1999). In addition, placental microsomes of smokers have increased 2- and 4-hydroxylation of oestradiol (Juchau et al., 1982).

Tobacco smoking during pregnancy enhances the increase in thyroid volume caused by the latter, particularly when combined with iodine deficiency (Knudsen et al., 2002a). Thyroglobulin and thiocyanate concentrations at birth and at 1 year of age in infants of smoking parents are greater than in infants with nonsmoking parents. These results indicate that the change in thyroid function observed at birth can persist for at least 1 year if the exposure to passive smoking from both parents is continued (Gasparoni et al., 1998). Furthermore, cord serum thyrotrophin, thyroxine and free thyroxine index are also decreased and free thyroxine index/thyrotrophin ratio increased in the smoking group compared to infants of nonsmokers. Infants of smoking mothers may have a hyperfunction of the thyroid gland at birth compared to infants of nonsmokers, with a negative feedback on thyrotrophin production from the pituitary gland. Increased metabolic rate and oxygen consumption caused by foetal thyroid hyperfunction may be pathogenetic factors for the foetal growth retardation caused by maternal smoking (Meberg & Marstein, 1986).

It is well accepted that menopause occurs at least 1–1·5 years earlier in current smokers than in women who have never smoked (Baron et al., 1990; Cooper et al., 1999). By contrast, there is no evidence of a decrease in age at natural menopause in former or passive smokers, or of a dose–response among current smokers. Thus the effect of smoking on ovarian senescence is limited to active smoking during the menopausal transition (Cooper et al., 1999).

It is of interest that smoking is associated with a decreased incidence of uterine fibroids, endometriosis and uterine cancer, which may reflect inhibitory effects of smoke constituents on uterine cell proliferation and extracellular matrix interactions (Shiverick & Salafia, 1999). Moreover, women who smoke may have a reduced risk of hyperemesis gravidarum and benign breast disease, possibly due to an altered oestradiol metabolism (Baron et al., 1990). By contrast, the relative frequency of polycystic ovary syndrome and idiopathic hirsutism is similar in smokers and nonsmokers (Byrne et al., 1991).