Introduction

Cadmium is a non-essential heavy metal with nearly ubiquitous exposure. Cadmium use is associated with refining of zinc, lead, and copper ores, mining, production of cadmium-containing soil fertilizers, plastic stabilizers, pigment production, and nickel-cadmium battery production [1]. Population exposure to cadmium primarily occurs through the diet via consumption of seafood, particularly shellfish, as well as from rice [2, 3] and some is found in leafy green and root vegetables [4]. Uptake of cadmium in tobacco plants leads to concentrated cadmium exposure in cigarette smoke [5] and consequently higher cadmium levels in smokers compared to nonsmokers. In areas where smoking rates are declining, cadmium exposure is subsequently declining [6]. Nationally representative surveys in the US show that levels in women have declined from geometric mean levels of 0.44 μg/g in 1988–1994 to 0.28 μg/g in 2003–2008 [6]. However, because cadmium comes from both natural and anthropogenic activities, exposure remains a public health risk and understanding its potential health effects is important [1]. Cadmium accumulates in the kidney and liver, and as such, both blood and urine are accepted epidemiologic biomarkers of cadmium exposure. Specifically, cadmium in urine represents cumulative exposure in the renal cortex of the kidney [7] generally reflective of exposure over a decade, although there are some recent concerns about the interpretation of urine cadmium concentrations at low-moderate levels [8, 9]. Blood cadmium levels reflect more recent exposures, on the order of several months [1, 10].

Cadmium has been identified as an endocrine disruptor [11]. The toxicological evidence with respect to cadmium’s effects on the reproductive system was recently reviewed [12–14]. Cadmium may interfere with hormonal functioning by binding at both the nuclear estrogen receptor [15] and G-protein coupled receptor 30 (GPR30) [16–18] and indirectly by P450 side-chain cleavage or through the low density lipoprotein receptor [19, 20] thus, cadmium may also be involved in the etiology of hormonally sensitive health outcomes and diseases. Experimentally, cadmium has shown effects on estrogenic activity [15, 21–24], including toxic effects to the ovary [21, 25–27]. Together with estrogen, cadmium increased estrogen receptor alpha mRNA expression beyond either estrogen or cadmium alone [20]. Further, cadmium may contribute to disease processes by altering the balance of oxidative stress [28]. The increasing evidence on human health effects among non-occupationally exposed populations, including cancer and cardiovascular disease [29, 30] hinges upon such effects at environmental levels of exposure.

Our objective was to systematically review the epidemiological evidence on the association between cadmium and women’s reproductive health, specifically considering puberty/menarche, reproductive hormones, fertility, pregnancy loss, preeclampsia, uterine leiomyoma, endometriosis, and menopause among individuals non-occupationally exposed to cadmium.

Methods

This systematic review process was conducted under the criteria provided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [31], except the evaluation for possible bias and quality of evidence, which were not evaluated in the current paper. We searched in PubMed for relevant, recent, studies published from 2003 through October 31, 2013, using the following combination of Medical Subject Heading (MeSH) search terms and text words: “cadmium” and each of the following terms separately, “puberty”, “reproductive hormones”, “fertility”, “pregnancy loss”, “preeclampsia”, “uterine leiomyoma”, “endometriosis”, or “menopause,” with additional terms as outlined in Table 1. Inclusion criteria encompassed original epidemiologic research articles that measured cadmium exposure using biomarkers.

Table 1 Search strategy for women’s reproductive health outcomes and cadmium
Full size table

Two investigators (SR and AZP) independently reviewed each of the 254 papers and selected 22 papers applying the following exclusion criteria: a) lack of original research, b) not a human study, c) case reports, and d) no cadmium exposure levels from biological tissues (i.e., environmental assessment of cadmium exposure) as outlined in Fig. 1. For studies that analyzed the same population for the same health endpoints, we selected either the most recent publication or the largest sample size. Relevant characteristics of all included studies are summarized in Tables 2, 3, and 4.

Fig. 1
figure 1

Flow diagram of the study selection process

Full size image
Table 2 Summary of cadmium exposure and reproductive hormones, fecundity, pregnancy loss, and pubertal development
Full size table
Table 3 Summary of cadmium exposure and preeclampsia
Full size table
Table 4 Summary of cadmium exposure, uterine leiomyoma, and endometriosis
Full size table

Reproductive Health

Puberty/Menarche

Earlier puberty is a risk factor for breast cancer and is a public health concern because of its association with increased risk of metabolic syndrome, diabetes, cardiovascular disease, and obesity [32]. On the other end of the spectrum, delayed pubertal development is considered a risk factor for osteoporosis [33]. Furthermore, the timing of pubertal development and menarche has shifted over the past several decades, as breast development and menarche are occurring earlier than in the past, demonstrating a secular trend [34]. Some have highlighted endocrine disrupting chemicals, including metals, as a possible contributing factor to this shift in reproductive development [35].

One study met our inclusion criteria on cadmium and puberty or menarche in girls (Table 2). National Health and Nutrition Examination Survey (NHANES) data (1988–1994) were used to evaluate the association of urinary cadmium with reproductive hormones inhibin B and luteinizing hormone among 705 girls aged 6–11 [36]. The findings suggest that in the context of elevated blood lead (>5 μg/dl), higher urinary cadmium levels (0.27–3.38 ng/ml) were associated with decreased inhibin B levels, suggesting a link to pubertal delays in adolescent girls [36]. However, this study was limited by several factors. The study was cross-sectional and therefore was unable to clearly establish the temporal relationship between metals and pubertal development. Further, the models were not adjusted for age, a major determinant of cadmium levels in urine and pubertal development and therefore. Confounding by age is thus plausible.

These cross-sectional results are intriguing and highlight the paucity of research regarding the relationship between cadmium and pubertal development in girls. There is a need for additional, particularly prospective research to understand the possible role of cadmium on pubertal development in girls.

Menstrual Cycle/Reproductive Hormones

Reproductive hormones are important biological signals and play a role in the etiology of cancer and heart disease, among other health conditions [37, 38]. Cadmium has been shown to play a role in modification of hormone levels experimentally, as recently reviewed by Iavicoli et al. [21], and may affect the hypothalamic-pituitary-gonadal axis (HPG) at multiple levels [13].

Four studies met our inclusion criteria using the search terms described in Table 1 [39–42] (Table 2). The studies were conducted in premenopausal women from New York State [39, 40], a US representative sample of perimenopausal women [41], and postmenopausal Japanese women [42]. The sample sizes ranged from 164 to 1050. Cadmium was measured in blood in the premenopausal women but in urine in the peri- and postmenopausal women. Estrogen, follicle stimulating hormone, luteinizing hormone, and progesterone as well as ovulation were the outcomes of interest in the studies in premenopausal women. Cadmium was associated with increased estrogen [39] and declines in FSH [40] in premenopausal women. However, among perimenopausal women with BMI ≥27, cadmium was associated with increased FSH [41]. Testosterone was positively associated with cadmium in postmenopausal women [42]. These studies suggest that menopausal status plays a role in the relationship between cadmium and reproductive hormones but additional research is necessary.

Additional studies at environmentally relevant levels of exposure are needed to confirm the possible role of cadmium in altering reproductive hormone function in women across their reproductive years and the menopausal transition. Although preliminary evidence suggests that cadmium could affect hormone levels, additional research with well-timed collection of reproductive hormone concentrations and exposure assessment will be necessary. Timing of measurements is particularly important given the high variability of reproductive hormones throughout the menstrual cycle in premenopausal women, as well as the current paucity of knowledge regarding the timing or threshold of exposure to cadmium (e.g., acute vs. chronic) and its effects on the various levels of the HPG axis.

Fertility

Infertility, defined as the inability of couples to become pregnant after 12 months of regular unprotected sexual intercourse [43] affects an estimated 3 to 26 % percent of the population [44–46]. Cadmium may affect fertility and fecundity via alterations in reproductive hormones, discussed above, or through oxidative stress and inflammatory pathways. Our review identified five studies that met our inclusion criteria with regard to cadmium and fertility outcomes [47••, 48, 49••, 50, 51]. Two evaluated time to pregnancy among healthy couples, and three evaluated fertilization and ongoing pregnancy rates among couples undergoing in vitro fertilization (IVF). Of note is that cadmium levels were measured in urine, blood [48, 49••], and follicular fluid [47••, 50, 51] when evaluating effects on fertility. Follicular fluid is a particularly biologically relevant measure of exposure for reproductive endpoints due to its proximity to endometrial tissues, as well as its direct contact with follicular cells responsible for synthesizing and secreting ovarian hormones which communicate in both a paracrine and endocrine manner with the entire HPG axis.

Two prospective studies in the US examined time to pregnancy among healthy couples trying to conceive [48, 49••] (Table 2). Both studies showed inconsistent findings. Among 80 women aged 18–34 from New York state who were followed prospectively while trying to conceive for up to 12 months or until becoming pregnant, blood cadmium (geometric mean level 1.63 μg/L) was not associated with time to pregnancy, although there was a non-statistically significant 13.74 % increase in probability of a positive pregnancy test for each 1.00 μg/L increase in Cd [48]. In contrast, longer time to pregnancy was associated with blood cadmium levels (fecundability odds ratio 0.78 (95 % CI 0.63-0.97) in a second prospective cohort study of 501 couples trying to conceive for up to 12 months, after adjustment for age, body mass index, cotinine, parity, serum lipids, and study site [49••]. In both of these studies, potential confounders were considered and all metals values were reported, regardless of limit of detection (LOD), a strategy that minimizes bias [52]. The conflicting evidence of these reports may be due to difference in cohort size, cohort characteristics, or unmeasured factors, but such differences highlight the need for further high quality research to identify the role of cadmium in time to pregnancy.

Three studies focused on in vitro fertilization, with mixed findings [47••, 50, 51] (Table 2). One study in Saudi Arabia reported that follicular fluid cadmium levels were associated with increased odds of fertilization while blood cadmium levels were not [50]. The mean follicular cadmium level was 0.34 μg/L (range < LOD–9.45 μg/L) in women who achieved fertilization as compared to 0.24 μg/L (range < LOD −0.90 μg/L) among those without fertilization. Blood cadmium levels were about double follicular levels and the two were highly correlated (0.59 p < 0.0001). Pregnancy was not associated with cadmium levels. A study measuring follicular fluid cadmium levels among 46 women undergoing IVF in San Francisco similarly found a positive association between oocyte fertilization and follicular fluid cadmium [51]. However, the variability of follicular cadmium was low due to nearly half of the values being below or near the LOD. Additional endpoints considered in this study included oocyte maturity, biochemical pregnancy, and clinical pregnancy, which were not significantly associated with follicular fluid cadmium levels [51]. Blood and urine cadmium levels were also measured in the same group of women. Biochemical 0.18 (95 % CI: 0.03, 0.94) and clinical pregnancy risk 0.06 (95 % CI: 0.01, 0.51) were associated with a 1 g/μL increase in blood cadmium [47••]. Blood and urinary cadmium levels in this small study were comparable to NHANES levels in the US. Collectively, the three IVF studies suggest that while cadmium in follicular fluid may be related to in vitro oocyte fertilization positively, it may act through different mechanisms systemically that ultimately result in diminished pregnancy rates. These studies, however, need to be interpreted cautiously due to their small sample sizes.

In summary, of the two prospective pregnancy studies in cohorts trying to conceive without assisted reproductive technology, one found no association between cadmium and fecundity [48] while one found a statistically significantly reduced fecundability odds ratio with increasing cadmium levels in women [49••] (Table 2). Clearly, more prospective pregnancy studies are needed to elucidate the potential association between cadmium and longer time to pregnancy among couples trying to conceive without fertility treatment. With respect to couples undergoing IVF, both available studies that measured follicular fluid cadmium levels, found a positive relationship between cadmium levels in follicular fluid and fertilization. However, cadmium in blood was associated with a reduced risk of biochemical and clinical pregnancy. These opposing findings by culture media underscore the need for future research on populations of women undergoing assisted reproductive therapy to clarify the relationship between cadmium and fertility in this important sub-population.

Pregnancy Loss

Pregnancy loss, or miscarriage, affects approximately one third of all conceptions [53] and 10-15 % of clinically recognized pregnancies [54]. Cadmium may affect pregnancy loss via effects on endocrine pathways or via the promotion of oxidative stress, which has been linked to adverse reproductive health [55]. Specifically, cadmium may reduce the body’s natural antioxidant capacity, for example, by depleting glutathione [56].

Two studies were found evaluating the association between cadmium and pregnancy loss, though one of the studies used a proxy of residence in a region polluted with cadmium as a proxy for individual exposure (Table 2). Differences in the percent of married women with pregnancy losses reported in their first or second pregnancies in the polluted as compared to the control region (10 vs. 2.8 %) were statistically significant but due to reliance on a proxy of exposure the study was not eligible for inclusion in the present review [57]. The study that met the inclusion criteria measured cadmium levels in serum in 69 pregnant women who were recruited from a clinic between 0–20 weeks of gestation in Ibadan, Nigeria [58]. Women with a self-reported history of recurrent (defined as three or more) pregnancy losses had mean blood cadmium levels of .46 μg/L while controls (women without a history of recognized pregnancy loss) had lower mean cadmium levels of 0.25 μg/L (p < 0.05). No other studies on pregnancy loss were identified.

As very little human data were available, additional prospective studies at environmentally relevant levels of cadmium exposure are needed to elucidate a possible association between cadmium and pregnancy loss. It is important to ascertain both early and clinical pregnancy losses as early losses may frequently be missed if a pregnancy test was not taken early enough. Therefore, identification of cases is problematic, and reliance on retrospective self-report of pregnancy loss is often not adequate. Therefore, prospective studies with careful evaluation of pregnancy loss and its timing in relation to cadmium exposure are needed.

Preeclampsia

Preeclampsia is characterized by hypertension and proteinuria in pregnancy and occurs in approximately 5–10 percent of all pregnancies [59]. Preeclampsia is associated with significant morbidity for the infant and mother and has largely unknown etiology. An earlier review introduced the hypothesis that cadmium was involved in the etiology of preeclampsia [60].

Our review identified five articles on preeclampsia, eclampsia, or pregnancy-induced hypertension, three of which met our inclusion criteria [61–63]. The studies were cross-sectional and limited their analysis to evaluating a difference in mean cadmium levels across women with and without preeclampsia or hypertension without adjustment for relevant confounders (Table 3). One cross-sectional study of 60 pregnant women in Belgrade, Yugoslavia measured cadmium levels in maternal blood and found that cadmium levels were more than two-fold higher in hypertensive nonsmokers compared with normotensive nonsmokers (1.9 μg/L vs. 0.8 μg/L, P = 0.001) [61]. However, among smokers, cadmium levels did not significantly differ (normotensive: 2.2 μg/L vs. hypertensive 1.9 μg/L) from one another, with both groups having levels comparable to those detected among hypertensive nonsmokers. This study did not attempt to account for potential confounding factors but rather only assessed a difference in means by smoking and hypertension status. A cross-sectional study of 145 women in Turkey measured cadmium in blood and found that mean levels were significantly higher in preeclamptic women (n = 47, 0.33 μg/L) compared with healthy pregnant women (n = 48, 0.29 μg/L) [62]. However, a case–control study in Iran measured cadmium in blood and did not find differences in mean levels between preeclampsia cases (0.54 μg/L) and controls (0.50 μg/L) [63]. Although published prior to 2003, one additional cross-sectional study measured cadmium levels in third trimester amniotic fluid in 101 normotensive pregnant women and 29 preeclamptic women [64]. The authors divided the samples into early (33–36 weeks) and late (37–40 weeks) third trimester samples and compared mean cadmium levels by preeclampsia vs. normotensive status within these groups. Amniotic fluid cadmium levels did not differ in the early third trimester group, but cadmium levels were higher in the amniotic fluid of preeclamptic women in late third trimester (106 mg/dl vs. 90 mg/dl) [64]. This stratification across the late third trimester is problematic and could induce bias because treatment for preeclampsia often involves inducing labor and delivery, hence resulting in an earlier delivery. Therefore, the severity of preeclampsia may differ across the groupings with later third trimester pregnancies likely having less severe disease than the earlier group.

All of the studies identified were cross sectional, did not utilize incident cases of preeclampsia, and were unable to evaluate the temporality of cadmium exposure levels and development of pregnancy-related hypertension. Together with the strong evidence of an association between cadmium and hypertension in adults [65] likely driven by cadmium’s pro-atherogenic activity in the circulatory system [66], these findings are suggestive that cadmium may play a role in pregnancy-related hypertensive disease, including preeclampsia, but additional research is needed to clarify these associations. In particular, prospective studies with incident identification of pre-eclampsia cases are necessary.

Uterine Fibroids

Uterine leiomyoma (fibroids) are noncancerous tumors in the myometrium and are the leading indication for gynecologic surgery. Fibroids cause reproductive dysfunction and pelvic pain [67]. Further, in the US, fibroids cost between $6 to $34 billion per year due to direct medical expenses and lost productivity [68]. Few risk factors for fibroids are known, apart from black race and age [69]. However, cadmium’s estrogenic properties are thought to contribute to the etiology of fibroids.

One large study compared cadmium levels in blood between women with and without self-reported uterine fibroids (n = 1425) and found no association [70] (Table 4). This was a nationally-representative cross-sectional study based upon self-report of uterine fibroids in NHANES. Self-report of fibroids is problematic because many women with fibroids are unaware of their condition [71]. Further, 25 % of women included in the study had cadmium levels below the limit of detection (0.3 μg/L).

Overall, the evidence with respect to uterine leiomyoma and cadmium is inadequate to draw conclusions. Additional research on this prevalent gynecological health endpoint with largely unknown etiology in relation to cadmium exposure is needed.

Endometriosis

Endometriosis is a disease that affects an estimated 6-11 % of reproductive-aged women and is often associated with infertility [72]. Endometriosis is characterized by a growth of endometrial glands and stroma outside the uterus [73]. However, while recognized as an estrogen-sensitive disease, endometriosis risk factors are poorly understood. The few established risk factors include Caucasian race, lean body type, and history of infertility [74, 75]. Cadmium may play a role in the etiology of endometriosis via its estrogenic properties [76] or through modification of oxidative stress [55].

We identified six epidemiological studies (Table 4). The studies were conducted in Belgium, Japan, Sri Lanka, and the US. Cadmium was measured in urine in two of the studies [77, 78], in blood in one study [70] and in both in three studies [79–81]. Endometriosis was defined by surgical visualization [77–80], and self-report [70]. Of the six studies, only one, a cross-sectional nationally-representative study in the US based upon data from NHANES found that cadmium was positively associated with endometriosis diagnosis. After adjustment for age, race/ethnicity, use of birth control pills prior to diagnosis, and smoking status at diagnosis, the estimated odds of endometriosis comparing the highest (0.5–8.5 μg/L) to the lowest tertile of exposure (<0.3 μg/L) was 3.39 (95 % CI 1.37-8.40) [70]. One study among 473 women from Salt Lake City, Utah and San Francisco, California, found a reduced odds of incident, surgically visualized, endometriosis associated with blood cadmium >0.37 μg/L (aOR = 0.55; 95 % CI: 0.31, 0.98) when compared with the lowest exposure tertile (<21 μg/L). However, urinary cadmium was not associated with endometriosis in this study [81]. The remaining four studies found no association between cadmium and endometriosis. Of these, three case–control studies did not attempt to adjust for potential confounding factors or utilize statistical modeling and analysis was limited to a comparison of geometric mean levels, in which no differences were detected [78–80]. One remaining case–control study was restricted to an infertile population and adjusted for average menstrual cycle length, body-mass index, and smoking status, and similarly, found no association between endometriosis and cadmium [77].

The weight of the evolving though limited body of evidence suggests that cadmium is not associated with endometriosis, with one study supporting a positive association, another study supporting a negative association, and four null studies. The one study that found a positive association between cadmium and endometriosis relied on self-report of endometriosis and measured cadmium levels in blood [70], which may not correspond to exposure at the time of endometriosis development. A negative association between blood cadmium and endometriosis [81] may be explained if cadmium lowers circulating estrogen levels, as endometriosis is an estrogen-sensitive disease. However, understanding cadmium potential effects on estrogen promotion of endometriosis requires further investigation. Most of these studies were cross-sectional without incident ascertainment of endometriosis cases. Prospective studies are needed to confirm the lack of association between cadmium exposure and incident endometriosis.

Menopause

Our review did not find any papers on cadmium and menopause specifically. A review on cadmium and menopause found a range of health effects affecting women around the time of menopause but noted the lack of data available with respect to timing of menopause and cadmium [82]. This lack of research on a potential role for cadmium in the timing of menopause underscores a need for further research on this topic.

Conclusions

Overall, cadmium may be considered a reproductive toxicant in women with respect to some reproductive health outcomes, with the greatest body of evidence existing for pregnancy-related hypertension and preeclampsia (Fig. 2). However, there is inadequate evidence to draw conclusions regarding health risks from cadmium exposure and several important endpoints. These include pubertal development, reproductive hormones, fertility, pregnancy loss, and menopause. Taken together, this review underscores the need for additional research into reproductive health effects of cadmium exposure. This is particularly important as women may be more susceptible to health effects of cadmium and consistently have higher exposure levels compared to men [83].

Fig. 2
figure 2

Summary of the evidence on women’s health endpoints in relation to cadmium exposure

Full size image

This review reveals a need for prospective studies that appropriately address limits of detection and confounding issues (e.g., current and past smoking) to better establish temporality between cadmium exposure and reproductive health outcomes in women. Improving the assessment of exposure measurement to appropriately address limits of detection issues, and to better understand how the media where cadmium is measured may play a role to better target blood, urine, or follicular fluid in future studies. Assessing women’s health outcomes can be complicated and many studies relied upon self-report or clinically ascertained samples. Such methods are limited when individuals are either not aware of their diagnosis or do not seek care. Statistical limitations are a further drawback of many of the studies included in this review. Specifically, the lack of attempt to address confounding or analysis beyond a comparison of mean levels is a shortcoming that should be improved in future studies. In sum, research on women’s health in relation to cadmium exposure is important given nearly ubiquitous exposure to cadmium in the population and some but not definitive evidence regarding cadmium exposure and women’s reproductive health related outcomes.