Objective
We estimated the prevalence and patterns of herbal use among US women before and during pregnancy.
Study Design
The National Birth Defects Prevention Study is an ongoing, population-based, case-control study. This analysis included 4239 women from 10 centers in the United States who delivered infants without major birth defects from 1998–2004.
Results
The prevalence of reported herbal use 3 months before or during pregnancy was 10.9%. During pregnancy, prevalence was 9.4% and was highest in the first trimester. Higher prevalence was associated with age greater than 30 years and education greater than 12 years. Use varied considerably by state (5–17%). Ginger and ephedra were the most commonly reported products early in pregnancy; teas and chamomile were most commonly reported throughout pregnancy.
Conclusion
Potentially 395,000 US births annually involve antenatal exposure to herbal products. Health care providers should inquire routinely about herbal use and educate patients about what little is known regarding risks of these products.
According to the 2002 National Health Interview Survey, an estimated 38 million US adults per year use herbal therapies. The reported prevalence of herbal use is higher among women than among men. There is a lack of published US data on the frequency of herbal use among women of childbearing age in general; published estimates specifically during pregnancy range from 4.1–45.2%, but these were based on sample sizes of only 150–734 women. Studies have shown that many patients do not disclose use of herbals to their health care providers; estimates of nondisclosure range from 25–58%.
For Editors’ Commentary, see Table of Contents
Under the Dietary Supplement Health and Education Act of 1994, the responsibility for ensuring safety before marketing rests with the manufacturers, and the US Food and Drug Administration (FDA) can take regulatory action only if it can prove that a product is unsafe once it reaches the market. Therefore, in contrast to prescription and newer over-the-counter medications, herbal products usually are marketed without the benefit of clinical trials to demonstrate either efficacy or safety.
Safety concerns related to herbal products have emerged. These safety concerns have been attributed to the herbal ingredient itself (eg, ephedra and kava), interactions between a herbal product and other pharmaceuticals (eg, Ginkgo biloba and blood-thinning agents, ginseng, and insulin), and contamination of products by unlabeled toxins (eg, lead and mercury). Herbal use surrounding pregnancy raises particular concerns, because many herbals are marketed specifically for symptoms that occur commonly during pregnancy, such as nausea and vomiting. More importantly, our ignorance of the potential harm to the pregnant woman is complicated by our even greater ignorance of the potential effects on fetal safety.
Because little is known about the extent of herbal use among pregnant women, we estimated the prevalence and pattern of use among women immediately before and during pregnancy.
Materials and Methods
To estimate the prevalence of herbal use among a population-based sample of women who delivered liveborn infants with no major birth defects, we used data from the National Birth Defects Prevention Study (NBDPS). The NBDPS is an ongoing, population-based, case-control study involving case infants with major structural birth defects and control infants without such defects, conducted in 10 centers across the United States: Arkansas, California, Georgia, Iowa, Massachusetts, New Jersey, New York, North Carolina, Texas, and Utah. Each center defined a study region with between 35,000–75,000 births per year, which in some cases was statewide and in other cases was a region in a state.
The control infants are a random sample of live births drawn from the same source population as the case infants, with the selection occurring from either birth certificates or hospital birth records. About 900 mothers of control infants are interviewed each year, and only control mothers were included in this analysis. The study was approved by the institutional review boards of the Centers for Disease Control and Prevention and all participating study centers.
The study uses a computer-assisted telephone interview to collect data from mothers about exposures in the 3 months before pregnancy and throughout pregnancy, the latter defined as the time period from conception (ie, 2 weeks after the last menstrual period) to delivery. Each month was a 30-day period, and, for this analysis, trimesters were defined as 3-month periods (first trimester: pregnancy months 1–3; second trimester: pregnancy months 4–6; and third trimester: pregnancy months 7–9, which some mothers did not reach, because they delivered in the second trimester). Interviews are conducted from 6 weeks to 2 years after the estimated date of delivery (EDD), with a mean and median time to interview for control mothers of 8.9 and 7.7 months, respectively, after the EDD.
The questionnaire item related to herbal use stated: “Did you use any herbs or folk medicines to treat any medical conditions, to lose weight, or just to keep you healthy?” This question covered the time period from 3 months before pregnancy to the date of the child’s birth. Analysis was restricted to control mothers who had answered this particular question.
From among the responses provided by study participants, herbal products in this analysis were defined as “products containing a plant, plant part, or plant extract (excluding topicals)” (eg, black cohosh, chamomile oil, ephedra, and Ginkgo biloba ). Products were classified using the Slone Drug Dictionary from Boston University’s Slone Epidemiology Center, which links the reported substances to their active ingredients. We included herbal teas that were reported in response to the specific herbal question or in any other section of the questionnaire.
When herbal exposure was reported in an alternate section of the questionnaire (eg, vitamin use section or maternal illness section), the reason for use was attributed to that topic. We excluded exposure to single herbals in mainly vitamin-dominated multivitamin supplements.
The herbal question was added in mid-2000; therefore, for this analysis, we included only women interviewed after this time, which corresponds to participants with EDDs between mid-1998 and the end of 2004 (4239/5958 total controls). The participation rate in the maternal interview was approximately 69% among control mothers for this time period.
STATA 8 software (Stata Corp, College Station, TX) was used for all analyses. We cross-tabulated herbal use with each sociodemographic characteristic of interest to determine the prevalence associated with each of these factors. The 3 months prior to pregnancy and the 3 trimesters of pregnancy were each considered a distinct “period” for purposes of this analysis, and prevalence of herbal use was calculated for each of the 4 periods. Percentages of mothers taking each herbal product were tabulated and ranked to determine the most commonly reported herbals.
Results
Overall, 462 (10.9%) of the 4239 control mothers reported use of any herbal product in the 3 months before pregnancy or during pregnancy. The prevalence of herbal use by state varied considerably, with the lowest use in Iowa and North Carolina (5.4% and 6.1%, respectively) and the highest in Utah (16.5%) ( Table 1 ). Use increased with age and also was greatest among women with more than a high school education and among women with a household income of $20,000 or more per year. Hispanics reported the highest prevalence of herbal use of all racial or ethnic groups studied.
| Maternal characteristic | Reported herbal product use, n (%) | No reported herbal product use, n | Odds ratio | 95% confidence interval |
|---|---|---|---|---|
| Maternal age at delivery, y | ||||
| <20 | 39 (8.8) | 404 | 1.0 | Referent |
| 20–24 | 88 (8.8) | 907 | 1.0 | 0.7–1.5 |
| 25–29 | 120 (10.7) | 999 | 1.3 | 0.9–1.8 |
| 30–34 | 141 (13.0) | 946 | 1.5 | 1.1–2.2 |
| 35–39 | 62 (12.4) | 440 | 1.5 | 1.0–2.2 |
| ≥40 | 12 (12.9) | 81 | 1.5 | 0.8–3.1 |
| Maternal education, y | ||||
| <12 | 68 (9.2) | 671 | 1.2 | 0.9–1.7 |
| 12 (high school) | 80 (7.7) | 954 | 1.0 | Referent |
| >12 | 310 (12.7) | 2126 | 1.7 | 1.3–2.3 |
| Household income | ||||
| <$20K | 114 (8.8) | 1181 | 1.0 | Referent |
| ≥$20K | 330 (12.3) | 2351 | 1.5 | 1.2–1.8 |
| Study site | ||||
| Arkansas | 50 (9.2) | 493 | 0.7 | 0.5–1.1 |
| California | 75 (13.1) | 499 | 1.1 | 0.7–1.5 |
| Georgia | 53 (12.5) | 372 | 1.0 | Referent |
| Iowa | 24 (5.4) | 419 | 0.4 | 0.2–0.7 |
| Massachusetts | 66 (12.1) | 480 | 1.0 | 0.7–1.4 |
| New Jersey | 33 (10.3) | 286 | 0.8 | 0.5–1.3 |
| New York | 40 (12.1) | 290 | 1.0 | 0.6–1.5 |
| North Carolina | 18 (6.1) | 276 | 0.5 | 0.3–0.8 |
| Texas | 60 (11.9) | 444 | 0.9 | 0.6–1.4 |
| Utah | 43 (16.5) | 218 | 1.4 | 0.9–2.1 |
| Maternal race or ethnicity | ||||
| Non-Hispanic white | 260 (10.5) | 2208 | 1.0 | Referent |
| Non-Hispanic black | 46 (9.8) | 425 | 0.9 | 0.7–1.3 |
| Hispanic | 123 (12.5) | 863 | 1.2 | 1.0–1.5 |
| Other | 33 (11.1) | 264 | 1.1 | 0.7–1.6 |
| Maternal body mass index, kg/m 2 | ||||
| Underweight (<18.5) | 17 (7.2) | 218 | 0.6 | 0.4–1.0 |
| Normal (18.5 to <25) | 260 (11.6) | 1979 | 1.0 | Referent |
| Overweight (25 to <30) | 98 (10.6) | 825 | 0.9 | 0.7–1.2 |
| Obese (≥30) | 67 (10.1) | 597 | 0.9 | 0.6–1.1 |
| Maternal smoking in first trimester a | ||||
| Smoked | 77 (9.8) | 710 | 0.9 | 0.7–1.1 |
| Did not smoke | 384 (11.2) | 3052 | 1.0 | Referent |
| Year of estimated date of delivery | ||||
| 1998–2000 | 115 (12.1) | 839 | 1.2 | 0.9–1.6 |
| 2001 | 82 (10.3) | 711 | 1.0 | 0.7–1.4 |
| 2002 | 75 (10.6) | 631 | 1.1 | 0.8–1.5 |
| 2003 | 99 (11.1) | 791 | 1.1 | 0.8–1.5 |
| 2004 | 91 (10.2) | 805 | 1.0 | Referent |
| Time to interview b | ||||
| <6 mo | 159 (11.9) | 1183 | 1.0 | Referent |
| 6 to <12 mo | 203 (11.0) | 1635 | 0.9 | 0.7–1.2 |
| 12 to <18 mo | 77 (10.8) | 635 | 0.9 | 0.7–1.2 |
| >18 mo | 22 (6.7) | 306 | 0.5 | 0.3–0.9 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
