Serum Micronutrient Concentrations and Risk of Uterine Fibroids

Abstract

Background:

Although uterine fibroids are among the most common gynecologic conditions affecting women in the United States, research on uterine fibroids is sparse. This study explored the association between micronutrients and uterine fibroids.

Methods:

We conducted a population-based, cross-sectional analysis of 887 women aged 20–49 who participated in the 2003–2004 National Health and Nutrition Examination Survey (NHANES). Serum micronutrient levels were collected during the health examinations, and information on uterine fibroids was assessed on self-reported questionnaires. Multivariate logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) while controlling for confounders.

Results:

A statistically significant dose-response relationship was observed between vitamin A and uterine fibroids, after adjustment for age, race, education, body mass index (BMI), and oral contraceptive (OC) use (middle vs. low: OR 2.43, 95% CI 1.35-4.37; high vs. low: OR 2.66, 95% CI 1.16-6.10, p for trend=0.02). After adjustment, a dose-response relationship also emerged for vitamin C, although these associations were not statistically significant.

Conclusions:

It is possible that certain micronutrients affect uterine fibroid development. If this is confirmed in large prospective studies, clinicians could advise susceptible populations on dietary changes to reduce their risk of uterine fibroids.

Introduction

Uterine fibroids are benign tumors of nonmalignant nature that develop in the smooth muscle cells of the uterus.¹ This gynecologic condition is responsible for nearly 600,000 hysterectomies annually and accounted for more than $2 billion in direct medical costs in 2000.^2,3 Uterine fibroids are typically detected during a pelvic examination, and the diagnosis is often confirmed using ultrasound.¹ Uterine fibroids typically remain asymptomatic; however, many women with uterine fibroids suffer significantly from gynecologic morbidities and pregnancy-related complications, such as excessive uterine bleeding, pelvic pain/pressure, infertility, placenta previa, abruptio placentae, premature rupture of the membranes, dysfunctional labor, breech presentation, and cesarean delivery.^1,4
–6

Research studies have associated multiple risk factors with uterine fibroids, including increasing age, premenopausal status, nulliparity, early age at menarche, and race/ethnicity.^7

–10 Despite the identification of risk factors related to uterine fibroids, research on these tumors remains sparse. The etiology of uterine fibroids is poorly understood. Findings from breast cancer research could provide insight into the causes and risk factors of uterine factors. Like uterine fibroids, the development of breast cancer is hormonally responsive, and the two conditions share similar risk factors, such as age at menarche and increasing age.^10
–12 The relationship between micronutrients and breast cancer has been studied extensively, and an association is suspected.^13

–16 Therefore, micronutrient status also might be related to the development and growth of uterine fibroids.

Micronutrients possess properties that may protect women from uterine fibroids. Vitamins C and E could protect against tumors through their antioxidant properties, which defend cell membranes and DNA from oxidative stress.¹⁷ Carotenoids also act as antioxidants that could protect against uterine fibroids by blocking damage from free radicals,¹⁸ and vitamin A is vital in epithelial cell differentiation and control of proliferation, which may help reduce tumor growth.¹⁹ B vitamins, such as folate, vitamin B₆, and vitamin B₁₂, possibly play a role in providing protection by affecting DNA methylation and synthesis.²⁰

In the few studies that have evaluated the nutritional status-uterine fibroids association, results have been inconsistent.^21

–24 Only one of these studies investigated the relationship between vitamins and uterine fibroids, and all the studies suffered from limited generalizability because of the use of populations consisting of predominantly white women or women of higher socioeconomic status (SES) and higher educational levels.

This study, therefore, explored the relationship between micronutrient levels and the odds of uterine fibroids using cross-sectional data from a large, population-based study.

Materials and Methods

Study design and population

This secondary analysis used data from the 2003–2004 National Health and Nutrition Examination Survey (NHANES). NHANES is a large, ongoing population-based, cross-sectional survey designed to estimate, monitor, and assess the prevalence of selected exposures, risk factors, and diseases.²⁵ NHANES uses a complex, multiple-stage, probability sampling technique to recruit individuals for participation. In order to produce a generalizable sample, certain groups are oversampled. From 2003–2004, low-income men and women, adolescents between 16 and 19 years of age, people >60 years of age, African Americans, and Mexican Americans were oversampled.

For the 2003–2004 NHANES, there were 10,122 participants in the survey (unweighted response rate: 79%).²⁶ Those who consented to participate in NHANES were first interviewed in their homes by trained interviewers before taking part in the health examination component of the survey. The health examination included dental, oral, and physical examinations, as well as blood, vaginal swab, and urine collection. For the purposes of this analysis, participants had to have completed both the in-home interview and the health examination. We excluded individuals for the following reasons: male (n=4970), women <20 years or >49 years of age (n=3823), women who were postmenopausal or had had a hysterectomy (n=101), and women with a history of cancer (n=33). Women were further excluded if they were missing information on uterine fibroids (n=215), micronutrient status (n=86), age at menarche (n=5), current oral contraceptive (OC) use (n=1), or smoking status (n=1). Thus, 887 women remained in the final analysis.

Measurement of outcome

Information on uterine fibroid status was self-reported during the in-home interview. Specifically, women were asked: Has a doctor or health professional ever told you that you had uterine fibroids? Women who reported “Yes” were considered to have uterine fibroids, and women who reported “No” were considered to not have the condition.

Measurement of exposure and covariates

During the health examination, serum samples were collected and certain micronutrient levels were measured. The micronutrients that were analyzed in this study included carotene (β-carotene), folate, and vitamins A (retinols), B₆ (pyridoxal 5′-phosphate, the biologically active form of vitamin B₆), B₁₂, C, and E (α-tocopherol). Micronutrient concentrations were categorized into tertiles, with low concentrations representing the referent group.

Information on demographic, lifestyle, and reproductive characteristics was self-reported on the NHANES questionnaire. Potential confounders included age in years (20–29, 30–39, 40–49), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, other race), education (less than high school graduate, high school graduate, some college, college graduate, or above), income (<$20,000, ≥$20,000, missing), marital status (married, other), age at menarche in years (≤13, >13), age at first birth in years (≤20, >20, missing), parity (0, 1–2, ≥3, missing), age at last birth in years (≤25, >25, missing), current contraceptive use (yes, no, missing), age at first use of contraceptives in years (<17, ≥17, missing), smoking status (never, current, former), female hormone use (yes, no), body mass index (BMI) (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0), and physically active (yes, no, unable to do activity).^21

–24

Statistical analysis

Summary statistics of the study population were calculated to assess characteristics of women with and without uterine fibroids. Unadjusted odds ratios (OR) and 95% confidence intervals (CI) were obtained using logistic regression to provide a crude association for each serum micronutrient concentration level and the presence of fibroids, as well as to determine other risk factors for uterine fibroids. Multivariate logistic regression was used to assess the micronutrient-uterine fibroid association while adjusting for confounders. To assess confounding, each potential confounder was entered into the model separately. If the variable changed the magnitude of the OR by at least 10%, it was included in the final model.²⁷ The following variables were confirmed as confounders according to this criterion: age (vitamin A, vitamin B₆, vitamin E, folate), BMI (β-carotene, vitamin A, vitamin B₆, vitamin B₁₂, vitamin C, folate), education (β-carotene, vitamin A, vitamin C, vitamin B₆, folate), race (vitamin A), parity (β-carotene, vitamin C, folate), age at first birth (vitamin B₆, vitamin C, folate), age at last birth (vitamin C), and OC use (vitamin A, vitamin B₁₂, folate).

For micronutrient-uterine fibroids models in which age and race/ethnicity did not meet the criteria for confounding, we still retained these variables in the multivariate models because they have been previously established as strong risk factors for uterine fibroids. As an earlier study of micronutrients and uterine fibroids identified smoking as an effect modifier, we also conducted stratified analyses to determine if smoking status (ever vs. never) was an effect modifier of any of the micronutrient-uterine fibroid associations in our population.²² Additionally, race/ethnicity (non-Hispanic white, non-Hispanic black, and Hispanic/other) and age (20–29, 30–39, 40–49) were considered as potential effect modifiers of the micronutrient-uterine fibroids associations. Tests of homogeneity were used to assess heterogeneity across the strata. All statistical analyses were performed using SAS-Callable SUDAAN statistical analysis software, version 10.0 (Research Triangle Institute) to account for the complex sampling design employed by NHANES.

Results

Characteristics of women with and without uterine fibroids

Approximately 8% (n=68) of the women in this study reported being notified by a doctor or other health professional that they had uterine fibroids (Table 1). When compared to women without uterine fibroids, a higher percentage of women reporting uterine fibroids were older (66.22% vs. 29.03% 20–29 years of age), non-Hispanic black (18.14% vs. 12.54%), married (61.59% vs. 51.77%), and overweight or obese (70.59% vs. 56.22%).

Table 1.

Number and Weighted Percentage of Baseline Characteristics of Women Included in the Serum Micronutrient and Odds of Fibroids Study, National Health and Nutrition Examination Survey

	Women with uterine fibroids (n=68)		Women without uterine fibroids (n=819)		Unadjusted OR and 95% CI
Variable	n	(%)	n	(%)	OR	95% CI
Age, years
20–29 years	8	(7.15)	352	(36.92)	1.00	Referent
30–39 years	20	(26.63)	283	(33.98)	4.04	(1.66-9.85)
40–49 years	40	(66.22)	184	(29.10)	11.74	(4.90-28.17)
Race
Non-Hispanic white	29	(65.91)	399	(67.17)	1.00	Referent
Non-Hispanic black	25	(18.14)	170	(12.55)	1.47	(0.72-3.02)
Mexican American/Hispanic	11	(11.92)	218	(15.47)	0.79	(0.28-2.20)
Other race	3	(4.03)	32	(4.81)	0.85	(0.18-4.01)
Education
Less than high school graduate	15	(12.68)	179	(13.89)	1.00	Referent
High school graduate	16	(30.67)	173	(20.55)	1.63	(0.76-3.51)
Some college	26	(41.34)	265	(36.35)	1.25	(0.47-3.33)
College graduate or above	11	(15.31)	202	(29.20)	0.57	(0.22-1.51)
Income
<$20,000	8	(8.57)	185	(16.90)	0.44	(0.17-1.11)
≥20,000	58	(90.94)	593	(78.10)	1.00	Referent
Missing	2	(0.49)	41	(5.01)	0.08	(0.01-0.52)
Marital status
Married	41	(67.59)	422	(51.73)	1.00	Referent
Other	27	(32.41)	397	(48.27)	0.51	(0.27-0.97)
Smoking status
Never	42	(58.61)	520	(60.61)	1.00	Referent
Current	15	(24.93)	184	(25.68)	1.00	(0.52-1.95)
Former	11	(16.47)	115	(13.71)	1.24	(0.60-2.55)
BMI^a
<18.5	1	(2.67)	23	(3.57)	1.12	(0.12-10.77)
18.5–24.9	16	(26.75)	286	(40.16)	1.00	Referent
25.0–29.9	18	(22.50)	215	(24.97)	1.35	(0.54-3.42)
≥30.0	33	(48.09)	295	(31.30)	2.31	(1.15-4.62)
Age at menarche
≤13	58	(86.03)	601	(74.65)	2.09	(0.96-4.54)
>13	10	(13.97)	218	(25.35)	1.00	Referent
Parity
0	4	(7.20)	49	(6.35)	0.91	(0.24-3.41)
1–2	33	(55.20)	374	(44.53)	1.00	Referent
3+	24	(30.18)	189	(19.84)	1.22	(0.60-2.51)
Missing	7	(7.31)	207	(29.28)	0.20	(0.08-0.53)
Age at first birth
≤20	22	(31.69)	211	(20.77)	1.00	Referent
>20	24	(32.29)	190	(27.71)	2.18	(0.93-5.14)
Missing	22	(36.02)	418	(51.52)	1.67	(0.81-3.42)
Age at last birth
≤25	18	(30.80)	253	(25.99)	1.00	Referent
>25	39	(54.69)	310	(38.38)	1.20	(0.56-2.58)
Missing	11	(14.51)	256	(35.63)	0.34	(0.15-0.79)
Current contraceptive use
Yes	4	(7.36)	93	(15.76)	0.36	(0.12-1.04)
No	46	(73.62)	390	(56.00)	1.00	Referent
Missing	18	(19.02)	336	(28.24)	0.51	(0.25-1.04)
Age at first contraceptive use
<17	12	(14.23)	132	(14.89)	1.00	Referent
≥17	42	(67.25)	484	(62.85)	1.12	(0.57-2.18)
Missing	14	(18.51)	203	(22.27)	0.87	(0.34-2.26)
Female hormone use
Yes	7	(11.37)	29	(4.30)	2.86	(1.04-7.88)
No	61	(88.63)	790	(95.70)	1.00	Referent
Physically active
Yes	47	(73.12)	463	(61.89)	1.71	(0.64-4.545)
No	20	(25.88)	350	(37.37)	1.00	Referent
Unable to do activity	1	(1.00)	6	(0.74)	1.94	(0.16-24.11)

Body mass index (BMI), calculated as kg/m².

CI, confidence interval; OR, odds ratio.

Unadjusted associations between select characteristics and uterine fibroids

As age increased, the odds of uterine fibroids increased. Specifically, women 30–39 years of age had nearly a 4-fold increased odds of uterine fibroids compared to women 20–29 years of age (OR 4.05, 95% CI 1.67-9.87), and women 40–49 years of age had approximately 12 times the odds of uterine fibroids (OR 11.77, 95% CI 4.97-28.21) (Table 1). Non-Hispanic black women had increased odds of uterine fibroids as compared to non-Hispanic white women; however, this result was not statistically significant (OR 1.47, 95% CI 0.72-3.02). In comparison, Hispanic race/ethnicity was found to be protective of uterine fibroids, although this finding was also not statistically significant (OR 0.79, 95% CI 0.28-2.20). There were no strong association between the other race/ethnicity category and uterine fibroids.

Unadjusted and adjusted associations between serum micronutrient concentrations and uterine fibroids

In the unadjusted model, the odds of uterine fibroids were increased for women with the middle and high concentrations of vitamin A (Table 2). After adjustment for age, race, education, BMI, and OC use, these statistically significantly elevated odds of uterine fibroids remained, but now with a dose-response relationship (middle vs. low: OR 2.43, 95% CI 1.35-4.37; high vs. low: OR 2.66, 95% CI 1.16-6.10; p for trend=0.02). After controlling for age, race, education, BMI, parity, age at first birth, and age at last birth, a dose-response relationship between vitamin C and uterine fibroids emerged, but these findings were not statistically significant (middle vs. low: OR 1.17, 95% CI 0.55-2.48; high vs. low: OR 1.62, 95% CI 0.80-3.28; p for trend=0.17). There was a dose-response relationship between vitamin E and uterine fibroids in the unadjusted model; however, after adjustment for age and race, the results were largely attenuated and remained statistically insignificant. The dose-response relationship evident in the unadjusted model for the folate-uterine fibroids association remained after adjustment for age, race, BMI, education, age at first birth, parity, and OC use, although these findings were not statistically significant. The results for the other micronutrients considered in this study were not statistically significant, and no clear patterns emerged.

Table 2.

Unadjusted and Adjusted Odds Ratios and 95% Confidence Intervals of Serum Micronutrient Concentrations (Tertiles) and Odds of Fibroids

Tertiles of serum micronutrients	Unadjusted OR	95% CI	p_trend	Adjusted OR	95% CI	p_trend
β-carotene^a
Low	1.00	Referent		1.00	Referent
Middle	1.06	(0.56-1.99)		1.30	(0.77-2.20)
High	0.82	(0.53-1.26)	0.33	1.18	(0.69-2.03)	0.66
Vitamin A^b
Low	1.00	Referent		1.00	Referent
Middle	2.25	(1.36-3.72)^*		2.43	(1.35-4.37)^*
High	2.18	(1.00-4.76)	0.07	2.66	(1.16-6.10)^*	0.02^*
Vitamin B₆ ^c
Low	1.00	Referent		1.00	Referent
Middle	1.48	(0.74-2.94)		1.37	(0.69-2.75)
High	1.39	(0.62-3.10)	0.42	1.12	(0.54-3.20)	0.65
Vitamin B₁₂ ^d
Low	1.00	Referent		1.00	Referent
Middle	0.92	(0.51-1.67)		0.95	(0.49-1.85)
High	1.62	(0.77-3.42)	0.18	1.61	(0.71-3.66)	0.22
Vitamin C^e
Low	1.00	Referent		1.00	Referent
Middle	1.05	(0.50-2.17)		1.16	(0.53-2.56)
High	1.02	(0.52-2.00)	0.95	1.79	(0.89-3.60)	0.17
Vitamin E^f
Low	1.00	Referent		1.00	Referent
Middle	1.38	(0.52-3.66)		1.02	(0.34-3.09)
High	1.61	(0.65-3.99)	0.27	1.10	(0.37-3.23)	0.83
Folate^g
Low	1.00	Referent		1.00	Referent
Middle	1.01	(0.48-2.12)		1.13	(0.50-2.53)
High	1.31	(0.67-2.56)	0.27	1.65	(0.81-3.34)	0.16

Adjusted for age, race, BMI, education, parity.

Adjusted for age, race, BMI, education, oral contraceptive use.

Adjusted for age, race, BMI, education, age at first birth.

Adjusted for age, race, BMI, oral contraceptive use.

Adjusted for age, race, BMI, education, parity, age at first birth, age at last birth.

Adjusted for age, race.

Adjusted for age, race, BMI, education, age at first birth, parity, oral contraceptive use.

Statistically significant.

Stratified analyses were conducted for smoking, race/ethnicity, and age. Smoking appeared to modify the micronutrient-uterine fibroid associations for vitamin A (test for homogeneity, p=0.008). Specifically, the associations between increasing micronutrient concentration levels and uterine fibroids were strongest for women who had ever smoked. Among never smokers, no strong association between vitamin A and uterine fibroids was observed. Race/ethnicity also emerged as an effect modifier of the association of interest, particularly for vitamin A (test for homogeneity, p=0.006). Among non-Hispanic white and non-Hispanic black women, a positive dose-response relationship emerged between vitamin A and uterine fibroids, whereas a protective effect was seen among Hispanic/other women. Age did not appear to be an effect modifier of the micronutrient-uterine fibroids association (data not shown).

Discussion

Within this exploratory, population-based study, a statistically significant positive association between vitamin A and uterine fibroids was suggested, where higher levels of vitamin A resulted in greater odds of uterine fibroids, after adjustment for age, race, education, and BMI. After controlling for confounders, a dose-response relationship emerged for vitamin C, although the findings did not reach statistical significance. No clear associations were detected for β-carotene, folate, vitamin B₆, vitamin B₁₂, or vitamin E.

As previously mentioned, we have used literature on the micronutrient-breast cancer association to guide hypothesis formulation and interpretation of results of this exploratory study. Because breast cancer and uterine fibroids share many of the same risk factors and both types of tumors are suggested to be hormonally responsive, we postulated that any observed associations might be due to similar biologic processes used to explain the micronutrient-breast cancer association.

To our knowledge, this is the first study to evaluate the association between vitamin A and uterine fibroids. As there is no literature to directly compare to our results, we consider studies of vitamin A and breast cancer. The extant literature examining the association between vitamin A and risk of breast cancer provides inconclusive results. Only a few studies have reported a positive association between retinol (i.e., preformed vitamin A) and breast cancer. A matched case-control study conducted by Malin et al.¹⁴ among women from Shanghai, China, used a food frequency questionnaire (FFQ) to assess intake of retinol and other micronutrients over the past 5 years. When comparing the highest quintile of retinol intake to the lowest quintile, the odds of breast cancer were reduced for premenopausal women by 8%. However, the odds of breast cancer were increased by >50% among postmenopausal women. None of these findings, however, were statistically significant. An additional case-control study that used FFQs to assess retinol intake also found that higher levels of this micronutrient were associated with increased, although not statistically significant, odds of breast cancer in premenopausal women.²⁸ We hypothesized that vitamin A may protect against uterine fibroids through similar mechanisms proposed in the micronutrient-breast cancer literature: by reducing the growth of tumors through control of cell differentiation and proliferation.¹⁹ We found, however, that the odds of fibroids were increased at each level of vitamin A when compared to the referent group. This finding may be attributed to a different biologic mechanism than originally hypothesized. Perhaps, as has been seen in several studies of vitamin A and breast cancer, higher levels of this micronutrient support the growth and development of uterine fibroids. To our knowledge, however, no biologic mechanism has been proposed in either the breast cancer or uterine fibroid literature to support this assertion. Alternatively, these findings may represent dietary changes that women make after a uterine fibroid diagnosis.

To date, no studies have examined the vitamin C-uterine fibroids association; thus, we consider literature of the vitamin C-breast cancer association. A positive association was found in a study assessing the role of vitamin C in breast cancer risk.²⁹ Cho et al.²⁹ used data from the Nurses' Health Study II to investigate this association. FFQs were used to assess both dietary and supplemental intake of vitamin C. A dose-response relationship was found, with higher vitamin C levels increasing the risk of breast cancer, which is consistent with the results of the present uterine fibroids study. Although vitamin C has properties that may protect against tumors, it is possible that these properties are not protective against the growth and development of uterine fibroids.

Carotenoids provide fruits and vegetables with yellow, orange, and red pigmentation.²² β-Carotene was used in the present analysis to represent the carotenoids, as many of the common vegetables consumed contain high levels of this micronutrient.¹⁴ Although not statistically significant, the results did suggest a small, positive relationship between β-carotene and uterine fibroids, which is consistent with the results of a study that used Nurses' Health Study data to evaluate the relationship between several different carotenoids and uterine fibroids.²² Specifically, the study by Terry et al.²² found that cumulative average intake of β-carotene slightly increased the risk of uterine fibroids. The authors also found an association between β-carotene and uterine fibroids among smokers; however, this finding was not confirmed in our analysis, which may be the result of our small sample sizes. Contradictory results were found in a case-control study conducted in 1987 by Chiaffarino et al.²¹ among women living in Italy. Using an FFQ to assess dietary intake, the authors found that fruits and some vegetables reduced the odds of uterine fibroids. However, this study measured the actual food and not nutrient intake, which may explain the conflicting results.

The exact biologic mechanisms involved in the micronutrient-uterine fibroid associations are unknown. Although there were a priori hypotheses and plausible biologic mechanisms regarding the role of vitamins A and C in reducing the odds of uterine fibroids, the results reported here suggest otherwise. Our findings may result from the small sample available from the NHANES dataset or from temporality issues associated with cross-sectional data. Alternatively, it is possible that peroxisome proliferator-activated receptors (PPAR) could increase the odds of uterine fibroids for women exposed to high levels of vitamin A. PPARs are a subgroup of nuclear receptors that when bound with retinoid X receptors, activate gene expression.³⁰ In a study by Tsibris et al.,³¹ an association was found among high levels of PPAR, retinoid X receptors, and uterine fibroids, suggesting that high levels of vitamin A may increase the risk of uterine fibroids. Additionally, other unknown mechanisms might explain the positive association between vitamins A and C and uterine fibroids.

The results from this study should be interpreted with caution. Women in our study were asked to self-report their uterine fibroid status, and the condition was not clinically confirmed with medical records or physical examinations. However, the accuracy of self-reported uterine fibroids has been examined in previous studies. For instance, Marshall et al.¹⁰ assessed the accuracy of self-reported uterine fibroid diagnosis in a subset of the Nurses' Health Study II cohort. Of the 116 medical records obtained, uterine fibroids were confirmed in 93% of the cases (n=108). Little variation was observed between the white and black participants, diagnosis of uterine fibroids being confirmed in 63 of the 67 medical records obtained for white participants (94%) and 45 of the 49 (92%) medical records obtained for black participants. In a separate study of a random sample of women residing in the Washington, DC, area, uterine fibroids were identified using ultrasound for 57% of women not reporting uterine fibroids during a telephone interview.⁵ Although women in the healthcare profession may be able to accurately self-report a diagnosis of uterine fibroids, other women may have difficulty in reporting this diagnosis accurately.

The prevalence of uterine fibroids in our study was 8%, which is lower than reported by other studies.^24,32 Several previous studies used ultrasound, medical records, and other clinical methods to assess uterine fibroid status rather than relying solely on self-reported data.^24,32 In using clinical methods for diagnosing uterine fibroids, asymptomatic fibroids will most likely be detected that otherwise would not have been self-reported, which could explain the low prevalence of uterine fibroids in our study. However, in a study using 1999–2002 NHANES data to examine the association of heavy metals and uterine fibroids among premenopausal women 20–49 years of age, the prevalence of uterine fibroids was similar to that in our study (8%).³³ Asymptomatic uterine fibroids may go undetected or undiagnosed, however, which could have resulted in a lower than expected prevalence of uterine fibroids in this study. Therefore, nondifferential misclassification of the outcome is possible, which would likely cause the findings to be biased toward the null value.

In addition, the overall response rate for the study population included in this analysis was >70%.²⁵ Although the sample size was adequate, it is possible that women who participated in the study differed from those who did not participate. The extent to which participation in the study would be related to both micronutrients and uterine fibroids is unknown. The 2003–2004 NHANES employed a sampling method to oversample for African Americans and Mexican Americans to increase generalizability. Although this oversampling may have been sufficient for the overall NHANES sample, it is possible that there were problems for our particular subsample. Our sample was restricted to women of a certain age group, and uterine fibroids were a relatively rare event in this sample. Thus, it is possible that our sample size was too small to produce reliable estimates for specific demographic subdomains.³⁴ Despite this potential limitation, stratified analyses revealed that race/ethnicity may be an effect modifier of the micronutrient-uterine fibroids association. In particular, protective effects were seen among women of Hispanic or other race/ethnicity. Prior research confirms that Mexican Americans, especially first-generation immigrants, have higher intakes of certain nutrients, including vitamins A, C, E, B₆, and folate.^35,36 Thus, the traditional Hispanic diet may provide protection against uterine fibroids, and this possibility should be further explored.

Micronutrient levels were assessed using serum concentrations collected during the health examination segment of NHANES. Specific procedures were undertaken in handling and assessing the serum concentrations in the respective laboratory settings in order to reduce measurement error.²⁴ Although the use of serum micronutrients may be viewed as a better method to collect micronutrient data, this assumption only applies if the serum markers are representative of the biologically relevant exposure time, measured with low variation, and correlated with dietary intake. Micronutrient status is difficult to assess in epidemiologic studies, and there is no universal method available for researchers to collect this information. The literature reports various correlation coefficients between serum micronutrients and dietary intake. Some micronutrients are not as well suited for assessing dietary intake as others.³⁷ Nevertheless, carotenoids, folate, vitamin C, and vitamin E have been reported as moderately to strongly correlated with dietary intake.^37,38 Also, according to the NHANES documentation, serum β-carotene, folate, vitamin C, and vitamin E had estimated coefficients of variation <10%.^39
–41 Although the serum micronutrients considered in this study were well correlated with dietary intake and had low estimated coefficients of variation, it is possible that there was exposure misclassification because of laboratory errors or issues with the markers reflecting the biologically relevant exposure period.

Related to this issue of the biologically relevant exposure period is temporality. In this study, women were diagnosed with uterine fibroids before collection of serum micronutrient levels. Serum micronutrient levels may change over time because of variations in diet and supplement use. A recent study that reviewed national cross-sectional survey data to evaluate changes in dietary patterns over the life span found that folate and B vitamins remain reasonably stable for women as their age increases from 20 to 49 years⁴²; however, intake of vitamin A and carotene was found to slightly increase as women aged from 20 to 49 years. Unfortunately, these women were not followed through time to determine their changes in micronutrient levels, and we are not able to determine the within-woman variation in micronutrient levels over time. Vitamin use among cancer survivors is well noted in the research literature⁴⁰; however, to our knowledge, there is no information available that explores micronutrient levels from supplement use or diet change after the diagnosis of uterine fibroids.

To further explore the potential for reverse causation between vitamin A and uterine fibroids, a sensitivity analysis was conducted using data on vitamin and mineral changes among women after a cancer diagnosis.⁴³ The conclusions remained largely unchanged over a broad range of reasonable assumptions regarding the proportion of women who would alter their vitamin and mineral use after a diagnosis of uterine fibroids. When more conservative assumptions were considered, however, the associations did become protective, suggesting a possible reverse causation. Thus, if the intake of these micronutrients increased after the time of the participants' diagnoses of uterine fibroids, it is possible that the results are an overestimate of the true association.

The assessment of confounders in this study was limited to the variables collected during the NHANES study. Although we considered numerous variables thought to be confounders of the micronutrient-uterine fibroids association, other known or unknown confounders might exist. Failure to control for these potential confounders could have resulted in either an overestimate or an underestimate of the true association.

Despite the limitations discussed, this study had several strengths. This study was the first to consider the possible association between several micronutrients (e.g., vitamins A, B₁₂, B₆, C, and E and folate) and uterine fibroids. Also, unlike prior studies that have relied on the use of FFQs, this study used serum levels to assess micronutrient levels. The use of serum micronutrient concentrations may reduce the risk of random errors associated with other dietary assessments, such as FFQs and dietary recalls, if they are representative of the biologically relevant exposure time.⁴⁴ Finally, this study used data from a large, population-based survey that uses a complex, multistage probability sampling technique to create a representative sample of the general population. Although we used only a specific subpopulation of NHANES, it is likely that the results of this study may be generalized to other premenopausal women residing in the United States.

Conclusions

Research on the micronutrient-uterine fibroids association is sparse. This study was designed to serve as an exploratory, hypothesis-generating analysis. Although the sample size of this study may have inhibited our ability to detect associations, our results did suggest associations between vitamins A and C and uterine fibroids that are important to consider in future studies. Large, prospective studies that address temporality and sampling concerns and rely on clinically ascertained uterine fibroid diagnoses are needed to confirm and expand on these results. Also, there were other micronutrient-uterine fibroids associations that appeared to have been modified by age, race/ethnicity, and smoking based on the stratified effect measures; however, the small sample size may have prevented these associations from attaining statistical significance. Thus, a large sample size would be ideal to allow for further exploration of whether associations differ by age, smoking status, and race/ethnicity. If confirmed in additional studies, clinicians and public health practitioners may be able to advise dietary changes to populations that are particularly susceptible to the development of uterine fibroids.

Footnotes

Disclosure Statement

No competing financial interests exist.

References

Stewart

. Uterine fibroids. Lancet, 2001; 357:293–298.

Merrill

. Hysterectomy surveillance in the United States, 1997 through 2005. Med Sci Monit, 2008; 14:CR24–CR31.

Flynn

, Jamison

, Datta

, Myers

. Health care resource use for uterine fibroid tumors in the United States. Am J Obstet Gynecol, 2006; 195:955–964.

Coronado

, Marshall

, Schwartz

. Complications in pregnancy, labor, and delivery with uterine leiomyomas: A population-based study. Obstet Gynecol, 2000; 95:764–769.

Wegienka

, Baird

, Hertz-Picciotto

et al. Self-reported heavy bleeding associated with uterine leiomyomata. Obstet Gynecol, 2003; 101:431–437.

Donnez

, Jadoul

. What are the implications of myomas on fertility? A need for a debate? Hum Reprod, 2002; 17:1424–1430.

Velebil

, Wingo

, Xia

, Wilcox

, Peterson

. Rate of hospitalization for gynecologic disorders among reproductive-age women in the United States. Obstet Gynecol, 1995; 86:764–769.

Samadi

, Lee

, Flanders

, Boring

III , Parris

. Risk factors for self-reported uterine fibroids: A case-control study. Am J Public Health, 1996; 86:858–862.

Baird

, Dunson

, Hill

, Cousins

, Schectman

. High cumulative incidence of uterine leiomyoma in black and white women: Ultrasound evidence. Am J Obstet Gynecol, 2003; 188:100–107.

10.

Marshall

, Spiegelman

, Barbieri

et al. Variation in the incidence of uterine leiomyoma among premenopausal women by age and race. Obstet Gynecol, 1997; 90:967–973.

11.

Henderson

, Feigelson

. Hormonal carcinogenesis. Carcinogenesis, 2000; 21:427–433.

12.

Ross

, Pike

, Vessey

, Bull

, Yeates

, Casagrande

. Risk factors for uterine fibroids: Reduced risk associated with oral contraceptives. BMJ (Clin Res Ed), 1986; 293:359–362.

13.

Adzersen

, Jess

, Freivogal

, Gerhard

, Bastert

. Raw and cooked vegetables, fruits, selected micronutrients, and breast cancer risk: A case-control study in Germany. Nutr Cancer, 2003; 46:131–137.

14.

Malin

, Qi

, Shu

et al. Intake of fruits, vegetables, and selected micronutrients in relation to the risk of breast cancer. Int J Cancer, 2003; 105:413–418.

15.

Toniolo

, Van Kappel

, Ackmedkhanov

et al. Serum carotenoids and breast cancer. Am J Epidemiol, 2001; 153:1142–1147.

16.

Lajous

, Lazcano-Ponce

, Hernandez-Avila

, Willet

, Romieu

. Folate, vitamin B₆, and vitamin B₁₂ intake and the risk of breast cancer among Mexican women. Cancer Epidemiol Biomarkers Prev, 2006; 15:443–448.

17.

Steinmetz

, Potter

. Vegetables, fruit, and cancer prevention: A review. J Am Diet Assoc, 1996; 96:1027–1039.

18.

Micozzi

, Beecher

, Taylor

, Khachlik

. Carotenoid analyses of selected raw and cooked foods associated with a lower risk for cancer. J Natl Cancer Inst, 1990; 82:282–285.

19.

Sporn

, Roberts

. Role of retinoids in differentiation and carcinogenesis. Cancer Res, 1983; 43:3034–3040.

20.

Shurbsole

, Jin

, Dai

et al. Dietary folate intake and breast cancer risk: Results from the Shanghai Breast Cancer Study. Cancer Res, 2001; 61:7136–7141.

21.

Chiaffarino

, Parazzini

, La Vecchia

, Chatenoud

, Di Cintio

, Marsico

. Diet and uterine myomas. Obstet Gynecol, 1999; 94:395–398.

22.

Terry

, Missmer

, Hankinson

, Willett

, De Vivo

. Lycopene and other carotenoid intake in relation to risk of uterine leiomyomata. Am J Obstet Gynecol, 2008; 198:37e1–8.

23.

Atkinson

, Lampe

, Scholes

, Chen

, Wahala

, Schwartz

. Lignan and isoflavone excretion in relation to uterine fibroids: A case-control study of young middle-aged women in the United States. Am J Clin Nutr, 2006; 84:587–593.

24.

Nagata

, Nakamura

, Oba

, Hayashi

, Takeda

, Yasuda

. Association of intakes of fat, dietary fibre, soya isoflavones and alcohol with uterine fibroids in Japanese women. Br J Nutr, 2009; 101:1427–1431.

25.

Centers for Disease Control and Prevention. NHANES 2003–2004 public data general release file documentation. Hyattsville, MD: US Department of Health and Human Services, Centers for Disease Control and Prevention, 2005.

26.

Centers for Disease Contol and Prevention. Unweighted response rates for NHANES 2003–2004 by gender and age. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, 2003-2004.

27.

Maldonado

, Greenland

. Simulation study of confounder-selection strategies. Am J Epidemiol, 1993; 138:923–936.

28.

Bohlke

, Spiegelman

, Trichopoulou

, Katsouyanni

, Trichopoulos

. Vitamins A, C and E and the risk of breast cancer: Results from a case-control study in Greece. Br J Cancer, 1999; 79:23–29.

29.

Cho

, Spiegelman

, Hunter

et al. Premenopausal intakes of vitamins A, C, and E, folate, and carotenoids, and risk of breast cancer. Cancer Epidemiol Biomarkers Prev, 2003; 12:713–720.

30.

Berger

, Moller

. The mechanisms of action of PPAR. Annu Rev Med, 2002; 53:409–435.

31.

Tsibris

, Porter

, Jazayeri

et al. Human uterine leiomyomata express higher levels of peroxisome proliferator-activated receptor γ, retinoid X receptor α, and all-trans retinoic acid than myometrium. Cancer Res, 1999; 59:5737–5744.

32.

Divakar

. Asymptomatic uterine fibroids. Best Pract Res Clin Obstet Gynaecol, 2008; 22:571–288.

33.

Jackson

, Zullo

, Goldberg

. The association between heavy metals, endometriosis and uterine myomas among premenopausal women: National Health and Nutrition Examination Survey 1999–2002. Hum Reprod, 2008; 23:679–687.

34.

Centers for Disease Control and Prevention. Analytic and reporting guidelines. The National Health and Nutrition Examination Survey (NHANES) Hyattsville, MD: National Center for Health Statistics. Centers for Disease Control and Prevention, 2005.

35.

Guendelman

, Abrams

. Dietary intake among Mexican American women: Generational differences and comparison with white Non-Hispanic women. Am J Public Health, 1995; 85:20–25.

36.

Dixon

, Sundquist

, Winkleby

. Differences in energy, nutrient and food intakes in a US sample of Mexican-American women and men: Findings from the Third National Health and Nutrition Examination Survey, 1988–1994. Am J Epidemiol, 2000; 152:548–557.

37.

Jacques

, Sulsky

, Sadowski

et al. Comparison of micronutrient intake measured by a dietary questionnaire and biochemical indicators of micronutrient status. Am J Clin Nutr, 1993; 57:182–189.

38.

Jenab

, Slimani

, Bictash

et al. Biomarkers in nutritional epidemiology: Applications, needs, and new horizons. Hum Genet, 2009; 125:507–525.

39.

Centers for Disease Control and Prevention. National Center for Health Statistics. Laboratory procedure manual: Fat soluble micronutrients (vitamins A, E, and carotenoids) Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, 2003–2004.

40.

Centers for Disease Control and Prevention. National Center for Health Statistics. Laboratory procedure manual: Vitamin B₁₂/folate. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, 2003–2004.

41.

Centers for Disease Control and Prevention. National Center for Health Statistics. Laboratory procedure manual: Vitamin C. Hyattsville, MD: US Department of Health and Human Services, Centers for Disease Control and Prevention, 2003–2004.

42.

Wakimoto

, Block

. Dietary intake, dietary patterns, and changes with age: an epidemiological perspective. J Gerontol A Biol Sci Med Sci, 2001; 56:65–80.

43.

Velicer

, Ulrich

. Vitamin and mineral supplement use among US adults after cancer diagnosis: A systematic review. J Clin Oncol, 2008; 26:665–673.

44.

Willett

. Nutritional epidemiology. 2nd. New York: Oxford University Press, 1998; 110–113.