Abstract
STUDY QUESTION
Is there an association between iron intake and ovarian reserve among women seeking fertility care?
SUMMARY ANSWER
Supplemental iron intake above 45 mg/day is associated with lower ovarian reserve among women seeking fertility care.
WHAT IS KNOWN ALREADY
Although the literature regarding iron intake in relation to ovarian reserve is scant and inconsistent, some evidence suggests that iron may have gonadotoxic effects.
STUDY DESIGN, SIZE, DURATION
This observational study included 582 female participants attending the Massachusetts General Hospital Fertility Center (2007–2019) enrolled in the Environment and Reproductive Health (EARTH) Study.
PARTICIPANTS/MATERIALS, SETTING, METHODS
Iron intake was estimated using a validated food frequency questionnaire. Markers of ovarian reserve included antral follicle count (AFC) (assessed via transvaginal ultrasound) and Day 3 FSH, both obtained during the course of an infertility evaluation.
MAIN RESULTS AND THE ROLE OF CHANCE
Participants had a median age of 35 years and median total iron intake of 29 mg/day. Total iron intake was inversely related to AFC and this association was driven by intake of supplemental iron. Compared to women with a supplemental iron intake of ≤20 mg/day, women consuming 45–64 mg/day of supplemental iron had a 17% (−35%, 0.3%) lower AFC and women consuming ≥65 mg/day of supplemental iron had a 32% (−54%, −11%) lower AFC after adjusting for potential confounders (P, linear trend = 0.003). Similarly, in a multivariable-adjusted analysis, Day 3 FSH levels were 0.9 (0.5, 1.3) IU/ml higher among women with a supplemental iron intake of ≥65 mg/day when compared to women with a supplemental iron intake of ≤20 mg/day (P, linear trend = 0.02).
LIMITATIONS, REASONS FOR CAUTION
Iron intake was estimated using a method that relies on self-report and we had no biomarkers of iron status in our participants; only 36 women consumed ≥45 mg/day of supplemental iron.
WIDER IMPLICATIONS OF THE FINDINGS
Since all study participants were seeking fertility treatment, our findings may not apply to women in the general population. Although our findings are consistent with studies of women with iron overload, given the paucity of literature on this topic, it is essential that this question is revisited in studies designed to better understand the dose–response relation of this association across the entire distribution of ovarian reserve and the risk–benefit balance of pre-conceptional iron supplementation given its many positive effects on pregnancy outcomes.
STUDY FUNDING/COMPETING INTEREST(S)
The project was funded by Grants R01ES022955, R01ES033651, R01ES009718, P30ES000002, and P30DK046200 from the National Institutes of Health. N.J.-C. was supported by a Fulbright Scholarship. N.J.-C., M.M., L.M.-A., E.O.-P., S.W., I.S., and J.E.C. declare no conflict of interest related to the work in the manuscript. R.H. has received grants from the National Institute of Environmental Health Sciences.
TRIAL REGISTRATION NUMBER
N/A.
Keywords: iron intake, iron supplements, ovarian reserve, antral follicular count, Day 3 FSH
Introduction
Iron supplementation has extensively documented benefits as a critical prenatal intervention (Sanghvi et al., 2010; Zhao et al., 2015). Pregnancy increases iron demands as a result of plasma volume expansion and increasing needs of fetoplacental unit (Fisher and Nemeth, 2017). Iron supplementation meets these demands, preventing anemia, ensuring women begin pregnancy with adequate iron levels (Bothwell, 2000), and providing important benefits including increased birth weight (Siega-Riz et al., 2006) and improved neurodevelopment (Iglesias et al., 2018).
Despite the well-documented benefits of prenatal iron supplementation, there are also concerns regarding safety of iron consumption among women with an adequate iron intake (Peña-Rosas et al., 2015). Routine universal prenatal iron supplementation is recommended only for women at risk of iron deficiency (Siega-Riz et al., 2006) and in places where severe anemia is prevalent (Bothwell, 2000) as there is evidence that in the general population iron overload is related to adverse outcomes such as increased risk of coronary heart disease (Sempos and Looker, 2001), insulin resistance and type 2 diabetes mellitus (Jahng et al., 2019). In addition, little is known about the impact of pre-conceptional iron intake on fertility. Although our group has previously reported that higher intake of non-haem iron is associated with a lower risk of infertility resulting from anovulation (Chavarro et al., 2006), this finding has not been replicated by others (Hahn et al., 2019). Furthermore, there is some evidence that iron overload may have a gonadotoxic effect in female mice (Qin et al., 2021) and in women diagnosed with beta-thalassemia major (BTM) (Singer et al., 2011; Roussou et al., 2013; Uysal et al., 2017; Talaulikar et al., 2019). In women diagnosed with BTM, infertility has been associated with ovarian iron deposition and increased oxidative stress, inducing NF-Kβ pathway (Roussou et al., 2013) and causing decreased ovarian reserve, as captured by serum anti-Mullerian hormone (AMH), estradiol, and antral follicle count (AFC) (Uysal et al., 2017; Talaulikar et al., 2019). Given the uncertainty of the relation between iron intake with fertility in general, and with ovarian reserve specifically, we evaluated the association of iron intake from food and supplements with markers of ovarian reserve among women seeking fertility evaluation and treatment at an academic fertility center. We hypothesized that women with the highest intake of iron from supplements would have lower ovarian reserve.
Materials and methods
Study population and design
Participants were women enrolled in the Environment and Reproductive Health (EARTH) Study, a prospective cohort of couples seeking fertility evaluation and treatment at the Massachusetts General Hospital (MGH) Fertility Center between 2004 and 2019 (Messerlian et al., 2018). Women between 18 and 46 years of age planning to use their own gametes over the course of infertility treatment were eligible. Starting in 2007, participants were asked to complete a validated diet questionnaire (Yuan et al., 2017, 2018) as part of the baseline assessments. AFC and Day 3 FSH levels were assessed in all patients as part of standardized infertility evaluations. Of the 1019, AFC scans conducted among study participants during the course of the study, we excluded scans performed while the woman was taking leuprolide acetate (n = 42), incomplete scans (n = 21), and scans from women with polycystic ovaries (n = 77). From the remaining 879 scans, we selected the first scan for each woman (n = 772) and excluded women who had incomplete or missing data (n = 190) leaving 582 women in the study. The study was approved by the Human Studies Institutional Review Boards of MGH and the Harvard T.H. Chan School of Public Health, and all participants gave informed consent before enrollment.
Exposure assessment
Diet was assessed using an extensively validated food frequency questionnaire (FFQ) (Yuan et al., 2017, 2018). Participants were asked to report how often on average they consumed certain amounts of more than 130 foods and beverages during the previous year. Nutrient intakes were estimated by summing the nutrient contribution of all food items in the questionnaire, and by taking into consideration the brand, type, and dose of dietary supplements used. The nutrient content of each food and specified portion size was obtained from a nutrient database derived from the US Department of Agriculture (2012) and additional information obtained from food and supplement manufacturers. To reduce extraneous variation in intake, nutrient intakes were adjusted using the nutrient residual method (Willett and Stampfer, 2009). Iron intake estimated with this questionnaire is positively correlated with intakes estimated from the average of two prospectively collected 1-week diet records (rSpearman = 0.58) (Yuan et al., 2017) and with serum ferritin levels among healthy pregnant women (rSpearman = 0.30).
Outcome assessment
AFC was assessed as part of diagnostic procedures with transvaginal ultrasound performed by a reproductive endocrinology and infertility specialist physician on the third day of an unstimulated menstrual cycle or on the third day of a progesterone withdrawal bleed. In women with menstrual dysfunction, AFC was performed only if there was evidence that patients were in early follicular phase (e.g. thin endometrial lining, no evidence of dominant follicle, and levels of estradiol and progesterone corresponding with early follicular phase). FSH was measured on serum blood samples drawn on the third day of menstrual or progesterone withdrawal bleeding using an electrochemiluminescence immunoassay (Roche Diagnostics, Indianapolis, IN, USA).
Covariate assessment
Demographics, medical history, and lifestyle data of women enrolled were collected in a staff-administered questionnaire. Height and weight data were measured by trained staff for BMI calculation (kg/m2). Clinical data, including infertility diagnosis, were abstracted from medical records. A take-home questionnaire with additional questions regarding lifestyle habits, reproductive health, and medical records was completed by participants. A validated questionnaire was used to assess time spent in leisure-time physical activities, allowing women to report the average spent time per week, during the preceding year, on 11 different activities with 13 response categories ranging from ‘never’ to ‘40 hours per week’ (Gaskins et al., 2016).
Statistical analysis
We divided women into quintiles of total iron intake and the lower quintile was used as the reference group. We tested for differences in participant characteristics across quintiles of iron intake using the Kruskal–Wallis test for continuous variables and the Χ2 test for categorical variables. We evaluated the association of iron intake with markers of ovarian reserve using Poisson regression models for count data (AFC) and quantile regression models for continuous outcomes (Day 3 FSH) because of important skewedness in outcome data. To reduce the influence of very high AFCs, we truncated the AFC measure at 30 (n = 23 women, 3.9%). We evaluated for confounding using a priori knowledge and descriptive statistics applied to directed acyclic graphs. Final multivariable models included terms for total calorie intake (continuous, kcal), carbohydrate intake (continuous, kcal), age (continuous, years), physical activity (continuous, h/week), menstrual cycle length (short cycles <24 days, regular cycles and long cycles >38 days) (Fraser et al., 2011; Munro et al., 2018) and vitamin C intake (continuous, mg/day). To describe the dose–response relationship between iron intake and study outcomes, we first fitted restricted cubic spline models (Durrleman and Simon, 1989) for total iron intake. Then, to evaluate separately the relation between dietary and supplemental iron intake, we divided our participants into quintiles of dietary iron intake and into four categories of supplemental iron intake (≤20, 21–45, 46–65, and >65 mg/day) corresponding to approximate dosing ranges available in the US market. Intakes of dietary and supplemental iron were simultaneously included in these models.
We evaluated whether participant characteristics are known to influence ovarian reserve, iron metabolism or both, including age (≥35 vs <35 years) (Park et al., 2021), BMI (≥30 vs <30) (Coimbra et al., 2013; Moslehi et al., 2018), vitamin C intake (≥173 vs <173 mg/day; median intake) (Lane and Richardson, 2014) and smoking status (current or former smoker vs never smoker) (Oladipupo et al., 2022), modified the relation between iron intake and ovarian reserve by introducing cross-product terms to the final multivariable models.
Last, we conducted a series of sensitivity analysis to evaluate the robustness of the primary analyses. First, to address the possibility that the results of analyses based on categories of supplemental iron intake were sensitive to influential values in the highest intake category, we collapsed the highest two intake categories, leaving three intake categories (≤20, >20 to ≤45, and >45 mg/day). To address the possibility that use of high-dose iron supplements was a proxy for treatment of anemia, thus resulting in spurious associations, we conducted analyses excluding women with a history of chronic non-communicable diseases (solely and jointly; hypertension n = 13, hypothyroidism n = 53, depression n = 106, anorexia/bulimia n = 22, and cancer n = 18), excluding women with obesity (BMI ≥ 30 n = 60), and excluding women with short (<24 days) or long (>38 days) cycles (n = 53). Since the study participants were mostly highly educated women with a previous subfertility diagnosis before starting the study, who may have, therefore, been aware of scientific literature linking iron intake to lower risk of infertility associated with ovulation disorders (Chavarro et al., 2006), possibly leading to using over the counter iron supplements, we conducted a sensitivity analyses to evaluate the possibility of reverse causation as a source of observed relations. Specifically, we first further adjusted our models for previous infertility treatment and previous infertility examination and, second, we conducted analysis excluding women with previous infertility treatment, a diagnosis of diminished ovarian reserve (DOR) or a diagnosis of infertility due to other ovarian pathology. All statistical analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC). Two-sided P values <0.05 were considered statistically significant.
Results
A total of 582 participants had a median (interquartile range IQR) age of 35.0 (32, 38) years. Most participants were white (82.7%) and never smokers (74.4%) and had completed a university degree (92.6%). The median (IQR) iron intake was 28.9 mg/day (14.3, 38.8) for total iron, 12.5 mg/day (9.1, 15.8) for dietary iron, and 16.0 mg/day (0, 28.0) for supplemental iron. The median (IQR) values for ovarian reserve measures were 13 (9, 18) for AFC and 6.9 IU/ml (5.9, 8.4) for FSH. Iron intake was positively related to physical activity, total caloric intake, total carbohydrate intake, vitamin C intake, and use of a multivitamins. Iron intake was unrelated to other baseline characteristics (Table 1).
Table 1.
Baseline characteristics in relation to quintiles of total iron intake; median (IQR) or n (%).
| Quintile of total iron intake |
||||||
|---|---|---|---|---|---|---|
| Total population | Q1 | Q2 | Q3 | Q4 | Q5 | |
| Intake, range, mg/day | 2.3, 156.5 | 2.3, 12.3 | 12.3, 23.2 | 23.4, 34.3 | 34.4, 41.9 | 42,156.5 |
| N | 582 | 116 | 117 | 116 | 117 | 116 |
|
| ||||||
| Personal characteristics | ||||||
|
| ||||||
| Age, years | 35 (32, 38) | 35 (32, 38) | 35 (31, 38) | 35 (32, 38) | 35 (32, 39) | 35 (32, 39) |
| BMI, kg/m2 | 23.2 (21.2, 26) | 23.2 (21.5, 25.8) | 23.6 (21.5, 26.4) | 23.2 (21, 25.9) | 23.1 (21.3, 25.2) | 23.8 (20.8, 26.6) |
| Physical activity, h/week | 5 (2.5, 9.5) | 5 (1.9, 8.4) | 5 (1.6, 9) | 5 (1.5, 9.5) | 5.7 (2.9, 9.9) | 6.5 (3.1, 12.8) |
| Ever smoker, n (%) | 148 (25.4) | 35 (30.2) | 26 (22.2) | 29 (25.0) | 29 (24.8) | 29 (25) |
| White, n (%) | 482 (82.8) | 95 (81.9) | 97 (82.9) | 96 (82.8) | 98 (83.8) | 96 (82.8) |
| Education, n (%) | ||||||
| High school or less | 42 (7.2) | 10 (8.6) | 11 (9.4) | 7 (6) | 5 (4.3) | 9 (7.8) |
| College | 196 (33.7) | 42 (36.2) | 45 (38.5) | 41 (35.3) | 32 (27.4) | 36 (31.0) |
| Graduate school | 344 (59.1) | 64 (55.2) | 61 (52.1) | 68 (58.6) | 80 (68.4) | 71 (61.2) |
|
| ||||||
| Dietary characteristics | ||||||
|
| ||||||
| Total calories, kcal/day | 1665 (1338, 2043) | 1370 (1088, 1568) | 1816 (1563, 2200) | 1774 (1350, 2219) | 1580 (1321, 1817) | 1923 (1624, 2488) |
| Carbohydrates % of energy | 48.4 (42.9, 53.6) | 46.3 (40.4, 51.6) | 48.3 (43.5, 54.6) | 48.9 (42.6, 53) | 48.6 (44.1, 53.4) | 49.9 (45, 55) |
| Fats % of energy | 33 (29.5, 37.4) | 34.3 (30.4, 38.4) | 32.6 (28.9, 36.7) | 32.9 (29.6, 39) | 33 (29.7, 37.4) | 32.4 (28.4, 36.2) |
| Protein % of kcal/d | 16.5 (14.9, 18.6) | 16.8 (14.3, 18.6) | 16.4 (15, 19.1) | 16.8 (15.2, 18.6) | 16.5 (14.9, 18.1) | 16.4 (14.7, 18.6) |
| Protein from animal source % of kcal/d | 10.5 (8.2, 12.9) | 11 (8.1, 13.5) | 10.6 (8, 12.9) | 10.5 (8.6, 12.7) | 10.4 (8.7, 12.5) | 10 (7.6, 12.3) |
| Multivitamin use, N (%) | 73 (63) | 81 (69) | 111 (96) | 116 (99) | 113 (97) | |
| Vitamin C, mg/day | 173 (116, 243) | 110 (73, 149) | 136 (90, 197) | 169 (132, 232) | 192 (172, 236) | 247 (191, 326) |
|
| ||||||
| Reproductive characteristics | ||||||
|
| ||||||
| Initial primary infertility diagnosis, n (%) | ||||||
| Diminished ovarian reserve | 51 (8.8) | 11 (9.5) | 7 (6.0) | 7 (6.0) | 13 (11.1) | 13 (11.2) |
| Ovulatory | 55 (9.5) | 10 (8.6) | 11 (9.4) | 15 (12.9) | 13 (11.1) | 6 (5.2) |
| Endometriosis | 22 (3.8) | 5 (4.3) | 5 (4.3) | 6 (5.2) | 2 (1.7) | 4 (3.5) |
| Uterine | 10 (1.7) | 2 (1.7) | 4 (3.4) | 1 (0.9) | 2 (1.7) | 1 (0.9) |
| Tubal | 28 (4.8) | 7 (6) | 5 (4.3) | 3 (2.6) | 4 (3.4) | 9 (7.8) |
| Male factor | 143 (24.6) | 21 (18.1) | 28 (23.9) | 36 (31.0) | 29 (24.8) | 29 (25.0) |
| Unexplained | 273 (46.9) | 60 (51.7) | 57 (48.7) | 48 (41.4) | 54 (46.1) | 54 (46.6) |
| Gravid, n (%) | 253 (43.5) | 55 (47.4) | 56 (47.9) | 44 (37.9) | 46 (39.3) | 52 (44.8) |
| Previous infertility exam, n (%) | 466 (80.1) | 94 (81.0) | 90 (76.9) | 98 (84.5) | 97 (82.9) | 87 (75.0) |
| Previous infertility treatment, n (%) | 269 (46.2) | 54 (46.6) | 47 (40.2) | 55 (47.4) | 61 (52.1) | 52 (44.8) |
| Menstrual cycle length, n (%) | ||||||
| <24 days | 11 (1.9) | 3 (2.6) | 4 (3.4) | 1 (0.9) | 3 (2.6) | 0 (0) |
| 24–38 days | 537 (92.3) | 107 (92.2) | 103 (88.0) | 109 (94.0) | 106 (90.6) | 112 (96.6) |
| >38 days | 34 (5.8) | 6 (5.2) | 10 (8.6) | 6 (5.2) | 8 (6.8) | 4 (3.5) |
Data are presented as median (interquartile range) or N (%).
There was a significant inverse, non-linear relation between total iron intake and AFC, whereby AFC did not vary with increasing iron intake up to an intake of approximately 45 mg/day of total iron but decreased with increasing iron intake throughout the remainder of the observed range of intake (Fig. 1A). Total iron intake was unrelated to Day 3 FSH in this analysis (Fig. 1B).
Figure 1.
Total iron intake in relation to antral follicle count and Day 3 FSH levels. (A) Antral follicle count. (B) FSH levels. Models are adjusted for age, total calories, carbohydrate intake, physical activity, vitamin C, and menstrual cycle length.
When we examined separately dietary and supplemental iron intake in relation to AFC and Day 3 FSH, we observed no relation between dietary iron intake with either outcome (Fig. 2A and B). However, supplemental iron intake was inversely related to AFC and positively related to Day 3 FSH after adjusting for potential confounders including dietary iron intake (Fig. 2C and D). Compared to women with a supplemental iron intake <20 mg/day, the adjusted difference in AFC for women in increasing categories of supplemental iron intake were 0.25% (−7.6, 8.1), −17.0% (−34.5, 0.3), and −32% (−53.9, −10.9), respectively (P, linear trend = 0.003). The corresponding adjusted differences in Day 3 FSH levels were 0.30 IU/ml (−0.08, 0.7), 0.35 IU/ml (−0.4, 1.1), and 0.86 IU/ml (0.4, 1.3), respectively (P, linear trend = 0.02). Results were similar when supplemental iron intake was modeled in three categories (Supplementary Table S1).
Figure 2.
Dietary and supplemental iron intake in relation to antral follicle count and Day 3 FSH. (A and B) Dietary iron intake and supplemental. (C and D) Supplemental iron intake. (A and C) Antral follicle count. (B and D) Day 3 FSH. Models are adjusted for age, total calories, carbohydrates intake, physical activity, vitamin C, and menstrual cycle length. AFC: antral follicle count.
The association of iron intake with AFC and FSH did not differ across strata of BMI, smoking status, or vitamin C intake. Similarly, the association between iron intake and FSH level did not differ by age. Nevertheless, statistical testing suggested that the association between dietary iron intake and AFC differed by age (P, interaction = 0.02). Dietary iron intake was unrelated to AFC among women under 35 years but was positively related to AFC among women 35 years or older (Table 2). The relation between supplemental iron intake and AFC did not differ by age, however (Table 2).
Table 2.
Dietary and supplemental iron intake in relation to AFC, stratified by age.
| Iron intake range, mg/day | <35 years, n = 269 | ≥35 years, n = 313 |
|---|---|---|
|
|
||
| Adjusted* relative difference in AFC (95% CI) | ||
| Dietary iron | ||
| Q1(2.3–8.7) | REF | REF |
| Q2(8.7–10.9) | −3.1 (−17.9, 11.8) | 2.6 (−14.7, 19.9) |
| Q3(10.9–13.4) | −1.4 (−19.0, 16.2) | 11.7 (−7.0, 30.4) |
| Q4(13.4–16.7) | 0.0 (−19.1, 19.2) | 15.5 (−3.8, 34.8) |
| Q5(16.8–47.6) | −12.8 (−39.5, 13.9) | 14.4 (−10.8, 39.6) |
| P trend | 0.61 | 0.005 |
| P, interaction | 0.02 | |
| Supplemental iron | ||
| Category 1 (≤20) | REF | REF |
| Category 2 (>20 ≤ 45) | −1.1 (−11.6, 9.4) | 2.2 (−9.6, 14.0) |
| Category 3 (>45 ≤ 65) | −22.1 (−47.8, 3.5) | −12.3 (−35.5, 10.9) |
| Category 4 (>65) | −31.1 (−53.7, −8.5) | −35.0 (−71.9, 1.9) |
| P trend | 0.009 | 0.08 |
| P, interaction | 0.13 | |
Models are adjusted for age, total calories, carbohydrate intake, physical activity, vitamin C intake, and menstrual cycle length.
AFC: antral follicle count; REF: reference.
The associations of supplemental iron intake with AFC (Supplementary Table S2) and Day 3 FSH (Supplementary Table S3) persisted in sensitivity analyses where women with a history of chronic diseases, obesity, or frequent/infrequent menses were excluded. We also found the same association pattern of supplemental iron intake with AFC and FSH after adjusting our models for previous infertility treatment and for previous infertility exam (Supplementary Table S4) and after excluding women with previous infertility treatment, infertility due to ovarian causes or DOR (Supplementary Table S5).
Discussion
We evaluated the relationship between iron intake with markers of ovarian reserve among women presenting for fertility services and found that supplemental iron intake above 45 mg/day was inversely associated with AFC and positively related to Day 3 FSH. On the other hand, dietary iron was unrelated to ovarian reserve in the entire study population, although stratified analyses suggested that dietary iron may be related to higher AFC among women over 35 years. The associations of supplemental iron with ovarian reserve persisted in a series of sensitivity analyses that addressed the possibility of reverse causation introduced by the consumption of high-dose iron supplements for specific medical needs and possible changes in dietary behavior prior to dietary assessment in the study. Similarly, the association of supplemental iron intake with ovarian reserve did not differ substantially across strata of known predictors of ovarian reserve or modifiers of iron status and metabolism. Although additional research is clearly necessary, our findings are in agreement with evidence in humans where iron overload resulting from medical treatment results in gonadotoxic effects, as is the case of transfusion-dependent beta-thalassemia (Singer et al., 2011; Roussou et al., 2013; Uysal et al., 2017; Talaulikar et al., 2019).
A gonadotoxic effect of iron is biologically plausible. Iron exhibits a U-shaped risk curve whereby in higher concentrations, non-protein bound iron can react with oxygen generating reactive oxygen species (Ozment and Turi, 2009; Imam et al., 2017), leading to damage of DNA, proteins, and lipid membranes, with subsequent metabolic stress (Zhao et al., 2017; Georgieff et al., 2019) inducing NF-Kβ pathway (Roussou et al., 2013). Gonadotoxic effects related to iron intake have been documented in animal models. In adult female mice that received intraperitoneal injection of iron dextran up to 1.0 g/kg once a week for 8 consecutive weeks, iron overload led to reduced ovarian size and weight, oxidative damage, as well as detriments in endocrine function, follicle development, and fertility (Qin et al., 2021). The authors suggested that this phenotype may be the result of disrupted ovarian steroidogenesis, alterations to the ovarian microenvironment, and blocking of the Wnt signaling pathway resulting in anovulation. The effects on the Wnt pathway are particularly informative since this pathway is related to follicle development, sex steroid production, ovulation, corpus luteum formation, and fertility (Boyer et al., 2010; Prunskaite-Hyyryläinen et al., 2014; Hernandez Gifford, 2015).
We are not aware of previous studies that have addressed the relationship between iron intake and ovarian reserve, although some studies have examined related phenotypes. In contrast to the findings of this study, our group previously reported that intake of non-haem iron was inversely related to the risk of infertility from anovulation among participants of the Nurses’ Health Study II (Chavarro et al., 2006). Nevertheless, supplemental iron intake was unrelated to fecundity in two prospective cohorts of pregnancy planners in North America and Denmark (Hahn et al., 2019). On the other hand, our findings are aligned with a report among women with transfusion-dependent BTM, who frequently develop iron overload (Chang et al., 2011) and present with lower ovarian volume, AFC, and AMH levels relative to age-matched controls; this phenotype has been interpreted as evidence of a gonadotoxic effect of iron (Singer et al., 2011; Uysal et al., 2017; Talaulikar et al., 2019). This interpretation is also consistent with the animal experiments reported above (Qin et al., 2021). Nevertheless, both beta-thalassemia and the animal models represent more extreme cases of iron overload than levels of overload that could be achieved through the intake of iron supplements alone. The apparent inconsistency of the literature may also reflect a consistent effect of iron intake on ovarian reserve that results in different patterns of association depending on the underlying distribution of ovarian reserve in a given study population. For example, if iron has a modest net negative effect on follicular count throughout the entire follicular count distribution, iron intake could have different and apparently paradoxical effects depending on the study population. Among women with very high follicular counts, as is the case in PCOS, the net negative effect could be identified as an apparent benefit that could include, in theory, improved cycle regularity, decreased anovulation, and lower risk of infertility due to anovulation among women with PCOS-like phenotypes, in agreement with previous reports from our group. At the same time, a modest deleterious effect may have no net benefit or actual harm among women in the middle of the ovarian reserve distribution in the general population, as reported in the Danish and North American pre-conception cohorts. However, in populations on the lower end of the ovarian reserve distribution, as is the case of women presenting to fertility clinics, a modest net negative effect could increase the frequency DOR and be mostly deleterious. Although the entirety of this hypothesis is not testable in this study alone, it is consistent with our results and with the apparent discordance in the literature as a whole. Clearly, the literature on this topic is scant and therefore additional research on the potential gonadotoxic effects of iron intake is needed.
It is important to interpret our findings in light of our study’s strengths and limitations. First, iron intake was estimated based on self-reported diet using an FFQ which, as is the case for any diet assessment method relying on self-report, raises concerns regarding the accuracy of reporting. Nevertheless, the questionnaire used in this study has been extensively validated (Yuan et al., 2017, 2018) including comparison against intake biomarkers conducted by other groups. Moreover, measurement error would most likely result in weaker associations. Second, we had no assessments of serum iron status in our participants and for that reason, it is not possible for us to confidently know whether women taking high-dose iron from supplements were using these supplements to correct iron deficiency anemia, thus opening the possibility of reverse causation in our findings. Although women presenting to fertility services are, generally, otherwise healthy women, and our sensitivity analyses excluding women who may have greater blood losses due to menstrual cycle abnormalities or chronic conditions were consistent with our main findings, this is still an issue that cannot be ruled out in our study and should be addressed in future studies. Third, only 36 women in our study were consuming ≥45 mg/day of supplemental iron. Therefore, it is important that this question is addressed in larger studies. Fourth, because all participants were seeking fertility treatment, our findings may not be generalizable to women in the general population. Finally, although observational studies present important challenges, including some already mentioned above, it is important to keep in mind that by using the first assessments of ovarian reserve available in the study, participants were essentially blinded to their ovarian reserve status when reporting their diet. Moreover, we have no reason to believe that women in our study would have differentially reported their iron intake in relation to their ovarian reserve status. Therefore, the greatest threats to causal inference typical of cross-sectional studies may not have been present in our study. Additional strengths of our study include the large sample size, standardized assessments of ovarian reserve status, and the availability of extensive dietary, lifestyle, demographic, and medical data that were considered in statistical adjustments when investigating the relation of iron intake with ovarian reserve.
In summary, we found that supplemental iron intake above 45 mg/day is associated with lower ovarian reserve among women presenting to an academic fertility center. Although existing literature is scant and inconsistent, our findings are in agreement with the potentially gonadotoxic effects of iron overload observed in animal models and in women with transfusion-dependent beta-thalassemia major. There are, however, important limitations of this study that make it essential that this question is further addressed in other studies that are better suited to address the dose–response relation between iron intake and ovarian reserve across the entire distribution of ovarian reserve and, critically, the risk–benefit ratio given the well-documented benefits for pregnancy and the neonatal period of prenatal iron supplementation. Importantly, while intriguing, our findings do not provide strong evidence to recommend against pre-conceptional iron supplementation in women who would otherwise benefit from this intervention.
Supplementary Material
Acknowledgements
The authors gratefully acknowledge all members of the EARTH Study Team, specifically the Harvard T.H. Chan School of Public Health research staff, Myra Keller, Ramace Dadd, and Alex Azevedo, and physicians and staff at Massachusetts General Hospital Fertility Center. A special thank goes to all of the study participants.
Contributor Information
N Jiménez-Cardozo, Grupo de Investigación en Ciencias Básicas y Clínicas de la Salud, Pontificia Universidad Javeriana, Cali, Colombia; Universidad del Valle, Cali, Colombia.
M Mitsunami, Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA.
L Mínguez-Alarcón, Department of Environmental Health, Harvard T.H. Chan School of Public Health, Boston, MA, USA; Channing Division of Network Medicine, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA.
E Ortiz-Panozo, Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA; National Institute of Public Health, Center for Population Health Research, Cuernavaca, Mexico.
S Wang, Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA.
I Souter, Fertility Center, Vincent Department of Obstetrics and Gynecology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA.
R Hauser, Department of Environmental Health, Harvard T.H. Chan School of Public Health, Boston, MA, USA; Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA; Fertility Center, Vincent Department of Obstetrics and Gynecology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA.
J E Chavarro, Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA; Channing Division of Network Medicine, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA; Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA.
Data availability
The data underlying this article cannot be shared publicly due to privacy safeguards of study participants. The data will be shared on reasonable request to the corresponding author pending approval from the reviewing IRB.
Authors’ roles
N.J.-C. conducted the statistical analyses and wrote the initial draft of the article. J.E.C. and R.H. contributed to the conception and design of the study. J.E.C., R.H., and I.S. contributed to the data acquisition. M.M., L.M.-A., and J.E.C. contributed to the data analysis. N.J.-C., M.M., L.M.-A., E.O.-P., S.W., R.H., I.S., and J.E.C. contributed to the interpretation of findings, critically revised the manuscript for important intellectual content, and gave final approval of the article.
Funding
National Institutes of Health (R01ES022955, R01ES033651, R01ES009718, P30ES000002, and P30DK046200). N.J.-C. was supported by a Fulbright Colombia Scholarship.
Conflict of interest
N.J.-C., M.M., L.M.-A., E.O.-P., S.W., I.S., and J.E.C. declare no conflict of interest related to the work in the article. R.H. has received grants from the National Institute of Environmental Health Sciences.
References
- Bothwell TH. Iron requirements in pregnancy and strategies to meet them. Am J Clin Nutr 2000;72(Suppl 1):S257–S264. [DOI] [PubMed] [Google Scholar]
- Boyer A, Lapointe E, Zheng X, Cowan RG, Li H, Quirk SM, DeMayo FJ, Richards JS, Boerboom D.. WNT4 is required for normal ovarian follicle development and female fertility. FASEB J 2010;24:3010–3025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chang H-H, Chen M-J, Lu M-Y, Chern JPS, Lu C-Y, Yang Y-L, Jou S-T, Lin D-T, Yang Y-S, Lin K-H. et al. Iron overload is associated with low anti-Müllerian hormone in women with transfusion-dependent β-thalassaemia. BJOG 2011;118:825–831. [DOI] [PubMed] [Google Scholar]
- Chavarro JE, Rich-Edwards JW, Rosner BA, Willett WC.. Iron intake and risk of ovulatory infertility. Obstet Gynecol 2006;108:1145–1152. [DOI] [PubMed] [Google Scholar]
- Coimbra S, Catarino C, Santos-Silva A.. The role of adipocytes in the modulation of iron metabolism in obesity. Obes Rev 2013;14:771–779. [DOI] [PubMed] [Google Scholar]
- Durrleman S, Simon R.. Flexible regression models with cubic splines. Stat Med 1989;8:551–561. [DOI] [PubMed] [Google Scholar]
- Fisher AL, Nemeth E.. Iron homeostasis during pregnancy. Am J Clin Nutr 2017;106:1567S–1574S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fraser IS, Critchley HO, Broder M, Munro MG.. “The FIGO recommendations on terminologies and definitions for normal and abnormal uterine bleeding. Semin Reprod Med 2011;29:383–390. [DOI] [PubMed] [Google Scholar]
- Gaskins AJ, Williams PL, Keller MG, Souter I, Hauser R, Chavarro JE; EARTH Study Team. Maternal physical and sedentary activities in relation to reproductive outcomes following IVF. Reprod Biomed Online 2016;33:513–521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Georgieff MK, Krebs NF, Cusick SE.. The benefits and risks of iron supplementation in pregnancy and childhood. Annu Rev Nutr 2019;39:121–146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hahn KA, Wesselink AK, Wise LA, Mikkelsen EM, Cueto HT, Tucker KL, Vinceti M, Rothman KJ, Sorensen HT, Hatch EE. et al. Iron consumption is not consistently associated with fecundability among North American and Danish Pregnancy Planners. J Nutr 2019;149:1585–1595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hernandez Gifford JA. “The role of WNT signaling in adult ovarian folliculogenesis. Reproduction 2015;150:R137–R148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Iglesias L, Canals J, Arija V.. Effects of prenatal iron status on child neurodevelopment and behavior: a systematic review. Crit Rev Food Sci Nutr 2018;58:1604–1614. [DOI] [PubMed] [Google Scholar]
- Imam M, Zhang S, Ma J, Wang H, Wang F.. Antioxidants mediate both iron homeostasis and oxidative stress. Nutrients 2017;9:671–619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jahng JWS, Alsaadi RM, Palanivel R, Song E, Hipolito VEB, Sung HK, Botelho RJ, Russell RC, Sweeney G.. Iron overload inhibits late stage autophagic flux leading to insulin resistance. EMBO Reports 2019;20:1–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lane DJR, Richardson DR.. The active role of vitamin c in mammalian iron metabolism: much more than just enhanced iron absorption! Free Radic Biol Med 2014;75:69–83. [DOI] [PubMed] [Google Scholar]
- Messerlian C, Williams PL, Ford JB, Chavarro JE, Mínguez-Alarcón L, Dadd R, Braun JM, Gaskins AJ, Meeker JD, James-Todd T. et al. ; for the EARTH Study Team. The Environment and Reproductive Health (EARTH) study: a prospective preconception cohort. Hum Reprod Open 2018;2018:1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moslehi N, Shab-Bidar S, Ramezani Tehrani F, Mirmiran P, Azizi F.. Is ovarian reserve associated with body mass index and obesity in reproductive aged women? A meta-analysis. Menopause 2018;25:1046–1055. [DOI] [PubMed] [Google Scholar]
- Munro MG, Critchley HOD, Fraser IS; FIGO Menstrual Disorders Committee. The two FIGO systems for normal and abnormal uterine bleeding symptoms and classification of causes of abnormal uterine bleeding in the reproductive years: 2018 revisions. Int J Gynecol Obstet 2018;143:393–408. [DOI] [PubMed] [Google Scholar]
- Oladipupo I, Ali T, Hein DW, Pagidas K, Bohler H, Doll MA, Mann ML, Gentry A, Chiang JL, Pierson RC. et al. Association between cigarette smoking and ovarian reserve among women seeking fertility care. PLoS One 2022;17:e0278998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ozment CP, Turi JL.. Iron overload following red blood cell transfusion and its impact on disease severity. Biochim Biophys Acta 2009;1790:694–701. [DOI] [PubMed] [Google Scholar]
- Park SU, Walsh L, Berkowitz KM.. Mechanisms of ovarian aging. Reproduction 2021;162:R19–R33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Peña-Rosas JP, De-Regil LM, Gomez Malave H, Flores-Urrutia MC, Dowswell T; Cochrane Pregnancy and Childbirth Group. Intermittent oral iron supplementation during pregnancy. Cochrane Database Syst Rev 2015;2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Prunskaite-Hyyryläinen R, Shan J, Railo A, Heinonen KM, Miinalainen I, Yan W, Shen B, Perreault C, Vainio SJ. Wnt4, a pleiotropic signal for controlling cell polarity, basement membrane integrity, and antimüllerian hormone expression during oocyte maturation in the female follicle. FASEB J 2014;28:1568–1581. [DOI] [PubMed] [Google Scholar]
- Qin X, Liang D, Hu M, Lv Z, Duan Z, Liu X, Tang L.. Chronic overload of concentration-dependent iron exerts different effects on ovarian function in C57BL/6J mice. Biol Reprod 2021;104:1347–1359. [DOI] [PubMed] [Google Scholar]
- Roussou P, Tsagarakis NJ, Kountouras D, Livadas S, Diamanti-Kandarakis E.. Beta-thalassemia major and female fertility: the role of iron and iron-induced oxidative stress. Anemia 2013;2013:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sanghvi TG, Harvey PW, Wainwright E.. Maternal iron-folic acid supplementation programs: evidence of impact and implementation. Food Nutr Bull 2010;31:S100–S107. [DOI] [PubMed] [Google Scholar]
- Sempos CT, Looker AC.. Iron status and the risk of coronary heart disease: an example of the use of nutritional epidemiology in chronic disease research. J Nutr Biochem 2001;12:170–182. [DOI] [PubMed] [Google Scholar]
- Siega-Riz AM, Hartzema AG, Turnbull C, Thorp J, McDonald T, Cogswell ME.. The effects of prophylactic iron given in prenatal supplements on iron status and birth outcomes: a randomized controlled trial. Am J Obstet Gynecol 2006;194:512–519. [DOI] [PubMed] [Google Scholar]
- Singer ST, Vichinsky EP, Gildengorin G, van Disseldorp J, Rosen M, Cedars MI.. Reproductive capacity in iron overloaded women with thalassemia major. Blood 2011;118:2878–2881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Talaulikar VS, Bajoria R, Ehidiamhen AJ, Mujawar E, Chatterjee R.. A 10-year longitudinal study of evaluation of ovarian reserve in women with transfusion-dependent beta thalassaemia major. Eur J Obstet Gynecol Reprod Biol 2019;238:38–43. [DOI] [PubMed] [Google Scholar]
- US Department of Agriculture. USDA National Nutrient Database for Standard Reference, 2012.
- Uysal A, Alkan G, Kurtoğlu A, Erol O, Kurtoğlu E.. Diminished ovarian reserve in women with transfusion-dependent beta-thalassemia major: is iron gonadotoxic? Eur J Obstet Gynecol Reprod Biol 2017;216:69–73. [DOI] [PubMed] [Google Scholar]
- Willett W, Stampfer M.. Implications of total energy intake for epidemiologic analyses. Nutr Epidemiol 2009;124:17–27. [DOI] [PubMed] [Google Scholar]
- Yuan C, Spiegelman D, Rimm EB, Rosner BA, Stampfer MJ, Barnett JB, Chavarro JE, Subar AF, Sampson LK, Willett WC. et al. Validity of a dietary questionnaire assessed by comparison with multiple weighed dietary records or 24-hour recalls. Am J Epidemiol 2017;185:570–584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yuan C, Spiegelman D, Rimm EB, Rosner BA, Stampfer MJ, Barnett JB, Chavarro JE, Rood JC, Harnack LJ, Sampson LK. et al. Relative validity of nutrient intakes assessed by questionnaire, 24-hour recalls, and diet records as compared with urinary recovery and plasma concentration biomarkers: findings for women. Am J Epidemiol 2018;187:1051–1063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhao G, Xu G, Zhou M, Jiang Y, Richards B, Clark KM, Kaciroti N, Georgieff MK, Zhang Z, Tardif T. et al. Prenatal iron supplementation reduces maternal anemia, iron deficiency, and iron deficiency anemia in a randomized clinical trial in rural China, but iron deficiency remains widespread in mothers and neonates. J Nutr 2015;145:1916–1923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhao Y, Hu X, Liu Y, Dong S, Wen Z, He W, Zhang S, Huang Q, Shi M.. ROS signaling under metabolic stress: cross-talk between AMPK and AKT pathway. Mol Cancer 2017;16:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article cannot be shared publicly due to privacy safeguards of study participants. The data will be shared on reasonable request to the corresponding author pending approval from the reviewing IRB.