Research Studies
Study Tracker
Water Fluoridation and Birth Outcomes in California.
| Environmental Health Perspectives | Goin DE, Padula AM, Woodruff TJ, Sherris A, Charbonneau K, Morello-Frosch R. | 132(5):057004.
California NIEHS grant U.S.
View PDF

Abstract

Full-text original study online at
https://pmc.ncbi.nlm.nih.gov/articles/PMC11098007/

Background:

There is a lack of research on the relationship between water fluoridation and pregnancy outcomes.

Objectives:

We assessed whether hypothetical interventions to reduce fluoride levels would improve birth outcomes in California.

Methods:

We linked California birth records from 2000 to 2018 to annual average fluoride levels by community water system. Fluoride levels were collected from consumer confidence reports using publicly available data and public record requests. We estimated the effects of a hypothetical intervention reducing water fluoride levels to 0.7 ppm (the current level recommended by the US Department of Health and Human Services) and 0.5 ppm (below the current recommendation) on birth weight, birth-weight-for-gestational age z-scores, gestational age, preterm birth, small-for-gestational age, large-for-gestational age, and macrosomia using linear regression with natural cubic splines and G-computation. Inference was calculated using a clustered bootstrap with Wald-type confidence intervals. We evaluated race/ethnicity, health insurance type, fetal sex, and arsenic levels as potential effect modifiers.

Results:

Fluoride levels ranged from 0 to 2.5 ppm, with a median of 0.51 ppm. There was a small negative association on birth weight with the hypothetical intervention to reduce fluoride levels to 0.7 ppm [-2.2g; 95% confidence interval (CI): -4.4, 0.0] and to 0.5 ppm (-5.8g; 95% CI: -10.0, -1.6). There were small negative associations with birth-weight-for-gestational-age z-scores for both hypothetical interventions (0.7 ppm: -0.004; 95% CI: -0.007, 0.000 and 0.5 ppm: -0.006; 95% CI: -0.013, 0.000). We also observed small negative associations for risk of large-for-gestational age for both the hypothetical interventions to 0.7 ppm [risk difference (RD)= -0.001; 95% CI: -0.002, 0.000 and 0.5 ppm (-0.001; 95% CI: -0.003, 0.000)]. We did not observe any associations with preterm birth or with being small for gestational age for either hypothetical intervention. We did not observe any associations with risk of preterm birth or small-for-gestational age for either hypothetical intervention.

Conclusion:

We estimated that a reduction in water fluoride levels would modestly decrease birth weight and birth-weight-for-gestational-age z-scores in California. https://doi.org/10.1289/EHP13732

EXCERPTS FROM ARTICLE DUE TO USE OF SCIENCE SYMBOLS

Introduction

Fluoridation of community water systems in the United States began in 1945 in Grand Rapids, Michigan.1 At this time, researchers initiated what was supposed to be a 15-y epidemiological study of water fluoridation and rates of dental decay in Grand Rapids and Muskegon, a nearby town, serving as a control community. Researchers observed that rates of dental decay declined in Grand Rapids after fluoridation but stayed the same in Muskegon. Based on these findings, city officials in Muskegon decided to fluoridate their water in 1951, ending the study 9 y early. Since then, many water systems across the world have fluoridated their water supplies. Fluoride is added to the water supply at the treatment plant, and a sodium fluoride saturator system, fluorosilicic acid system, or dry fluoride feed system can be used.2 In the case excess fluoride is added, water systems shut down the equipment and flush out the water lines; they are also required to notify the public.2 Individuals who wish to remove fluoride must use a reverse osmosis filter.

Although the dental health benefits are well substantiated,3 evidence suggests fluoride consumption in early life may adversely affect child neurodevelopment,4–9 although some recent studies have also found positive or null associations.10,11 Given existing evidence, the National Toxicology Program has concluded that “fluoride is presumed to be a cognitive neurodevelopmental hazard to humans.”12 Because birth outcomes can also affect neurodevelopment,13,14 adverse birth outcomes may be on the pathway between in utero fluoride exposure and potential effects on neurodevelopment. However, only recently has the impact of water fluoridation on pregnancy outcomes been considered in the United States,15,16 and studies in the United States and elsewhere that do exist show mixed results. For instance, some studies found higher fluoride levels were associated with lower birth weight,15 whereas others found associations with higher birth weight.17,18 Some studies have found increased risk of preterm birth,19 others have found the opposite,16,20 and one study found no association with either gestational age or fetal growth.21 To our knowledge, no studies have been done at the population level, which is an important next step for establishing risk in a general population.

Despite the paucity of evidence surrounding the effects of water fluoridation on reproductive health, some animal and epidemiological studies in places with naturally high levels of fluoride in ground water point to potential mechanisms and suggest there may be adverse effects. Evidence indicates fluoride exposure may affect endocrine systems via alterations in thyroid hormone levels.22–25 Many other endocrine-disrupting chemicals have been linked to adverse birth outcomes because of the key role of hormones in regulating the normal processes of growth during gestation and parturition.26,27 Drinking water fluoride levels exceeding the World Health Organization (WHO) guideline of 1.5 ppm have also been linked to anemia,28 hypertension,29 and other adverse cardiovascular outcomes,30 which may increase risk of adverse birth outcomes due to systemic inflammation and oxidative stress.31–33 Animal studies have shown fluoride can affect antioxidant enzyme activity, including reactive oxygen species and NADPH oxidase, mechanisms related to inflammation and oxidative stress.34,35 These mechanisms related to inflammation and oxidative stress have also been implicated in the pathogenesis of adverse birth outcomes.36 However, given the lack of research investigating effects of water fluoridation on human reproductive outcomes, specific mechanistic pathways are not yet known.

California passed Assembly Bill 733 in 1995 requiring all public water systems with at least 10,000 service connections to install a water fluoridation system once the funds became available.37 Since then, water districts have complied at different times, resulting in spatial and temporal variability in water fluoridation practices. The Metropolitan Water District of South California, which provides surface water to 18 million Southern Californians, began fluoridating in 2007.38 Furthermore, in 2015, the US Department of Health and Human Services reduced the recommended level of fluoride in drinking water to 0.7 ppm, a change from previous standards that were dependent on ambient temperature and ranged from 0.7 to 1.2 ppm.39 This policy change introduced additional variability in levels of water fluoridation within water systems over time in California, especially among those water districts that have fluoridated their water for decades. Some water districts also vary the proportion of ground vs. imported surface water they use over time, which creates variability in the amount of fluoride in the blended water distributed to consumers. Our goal was to use this variability to examine the effects of water fluoridation on birth outcomes. To evaluate the potential impacts of policy change regarding community water fluoride levels on birth outcomes, we estimated the effects of two hypothetical interventions to reduce water fluoride levels to 0.7 ppm and 0.5 ppm on birth outcomes in California. We chose these thresholds to represent reduction to the current recommended level (0.7 ppm) and to a level below the current recommendation (0.5 ppm).

Data and Methods

Study Population

We used California birth records to capture births that occurred in areas served by community water systems during the period 2000–2018. Birth records were geocoded to the pregnant person’s address at the time of delivery and linked to water system boundaries using the UC Berkeley Water Equity Science Shop Community Water System boundaries.40 We began with N =9,805,694 births and limited our analyses to live singleton births, where the birthing parent’s address was able to be geocoded, the parent was living in an area served by a community water system, and the outcome variables, covariates, and fluoride levels had nonmissing values (n=5,008,915) (Supplemental Figure 1).

This study was approved by the institutional review board of the University of California, San Francisco (IRB No. 21-34762 and CDPH IRB No. 13-05-1231).

Birth Outcomes

The outcomes we considered in this study were birth weight in grams, birth-weight-for-gestational-age z-scores, gestational age in completed weeks, preterm birth, small-for-gestational age, large-for-gestational age, and macrosomia. We used a population reference to calculate birth weight-for-gestational-age z-scores, small-for-gestational age, and large-for-gestational age.41 Small-for-gestational age was defined as below the 10th percentile of sex-specific birth-weight-for-gestational age based on a population reference,41 and large-for-gestational age was defined as above the 90th percentile. Gestational age was determined based on the obstetric estimate of gestational age from 2007 forward; before 2007 the obstetrical estimate was not recorded on birth records so last menstrual period was used to estimate gestational age. Preterm birth was defined as delivery before 37 completed weeks of gestation. Macrosomia was defined as birth weight >4,000g.

Fluoride Levels

Comprehensive historical data on fluoride levels for community water systems across California are not publicly available. However, since 1998, water systems have been required to test and report levels of several chemicals, including fluoride to consumers as part of the Consumer Confidence Report Rule amendment to the Safe Drinking Water Act. This requirement usually means a hard copy brochure of the consumer confidence report is mailed to customers. The data are not compiled into electronic databases, and record keeping among water system administrators varies substantially. The US EPA compiles consumer confidence reports in the Safe Drinking Water Information System (SDWIS) system; however, only the most recent years are available. To compile a database of historical water fluoride levels from 2000 to 2018, we used all publicly available data, and we filed Freedom of Information Act requests to individual water systems for the consumer confidence reports for years that were missing data.

Consumer confidence reports provide fluoride levels measured at specific water sources (e.g., a water intake or treatment plant) and/or system average concentrations. They also distinguish between naturally occurring and treatment-related fluoride levels. To determine the average level of fluoride in a given water system and year, we used the level of fluoride reported in the consumer confidence report if only one water source was reported. If only the treatment-related level of fluoride was provided, we used this level. If only the naturally occurring level of fluoride was provided, we used this level. If more than one water source was listed, but an average for the system was listed, we used the average level reported. If more than one water source was listed and no system average was reported, but there were percentages listed for each water source, we created weighted averages using the percentages reported. If more than one water source was listed, no system average was reported, and no percentages were listed for each water source, we created a simple average between the sources. If only a range was reported, we used the median of the range. If only a range or no data were available, but a value of average annual fluoride levels was listed on the data provided on the California Water Board website,42 we used that level. If levels were missing for a year but prior year levels were available, we carried those levels forward. We assumed nondetected fluoride levels were the limit of detection for the purpose of reporting (0.1 ppm) divided by the square root of 2. Of the 5,861 water systems and years we included in the study, fluoride levels for 893 (15.3%) were nondetected or were an average of a detected and nondetected level, and 3,408 (58.1%) had fluoride levels directly from the consumer confidence report, 984 (16.8%) had a simple average of fluoride levels from two or more sources in the consumer confidence report, 835 (14.2%) had a weighted average of source levels from the consumer confidence report, 184 (3.1%) had consumer confidence reports that did not report a fluoride level (which, given the current regulations, we determined to be a nondetect), 96 (1.6%) used the California Water Board level, 79 (1.3%) used a median of a range reported on the consumer confidence report, 15 (0.3%) used a weighted average of sources from a personal communication with someone who worked at the water system, and 260 (4.4%) had some other method, which sometimes represented when a water system purchased data from another system or if the method was not recorded. Of the 546 unique water systems included in this analysis, there were 349 small or medium water systems (<10,000 connections) and 196 large water systems (10,000 or more connections). The median service area for small or medium water systems was 9.3 km2, and the median service area for large water systems was 57.5 km2 (Supplemental Table 1).

Birth records were linked to the average level of fluoride that was present in the community water system that served the pregnant person at delivery. We assigned each pregnancy the fluoride level in the year of the pregnancy, and if the pregnancy spanned 2 y, we created a weighted average for which the weights were the proportions of the pregnancy spent in each year. We previously conducted a study linking community water system fluoride levels to maternal serum, urine, and amniotic fluid collected during midgestation and found modest but statistically significant correlations between the community water system fluoride and fluoride levels in biological samples.43 For example, the correlations between community water fluoride and the fluoride measured in maternal serum and amniotic fluid were 0.39 and 0.41, respectively. This finding provides some confidence that water fluoride level is a decent proxy for actual exposure levels.

Arsenic Levels

To control for arsenic levels, which often covary with fluoride levels (either because of geology or industrial pollution) and thus may cause coexposure confounding,44 we linked birth records to the Water Quality Database Files from California’s State Water board,45 which provide historical public water monitoring data for community water systems. These data represent testing levels by water systems for each water source but do not include comprehensive data on treatment-related fluoride. To estimate the contaminant levels in the distribution system, we connected each source to a receiving source to avoid double counting of testing levels and remove duplicates as described in previous work.46 We then created yearly averages of arsenic levels for each water source. If levels were missing for a year, but prior year levels were available, we carried those levels forward.

Statistical Analyses

Our goal was to estimate the effects of two hypothetical interventions of reducing water fluoride levels to 0.7 ppm and 0.5 ppm on birth outcomes. To do this, we used G-computation47 with a linear regression model that included a natural cubic spline to allow for nonlinearity in the effects of fluoride. G-computation is a standardization approach that allows the investigator to model the effects of a hypothetical intervention.48,49 We selected the number of knots for each outcome regression separately using the Aikake Information Criteria. We then predicted the outcomes if fluoride levels that exceeded 0.7 ppm were reduced to 0.7 ppm and did the same for 0.5 ppm. For inference, we used a clustered bootstrap with 200 bootstrapped samples to calculate Wald-type CIs at the water system level. All analyses were conducted on the additive scale.

Furthermore, we wanted to assess whether small changes in fluoride levels were associated with differences in birth outcomes across the distribution. Thus, we evaluated the association of fluoride levels with birth outcomes using generalized additive models to identify potential nonlinearities.

We selected covariates based on the literature and a directed acyclic graph (Supplemental Figure 2). We adjusted for individual-level characteristics including continuous birthing parent age, race/ethnicity, health insurance type, educational attainment, and the month and year of conception. We included race/ethnicity as a confounder and potential effect modifier because of the way racialized access to wealth and racial residential segregation influences where people live in California, as well as how racism can affect birth outcomes. The categories of race on the birth record included White, Black, American Indian/Native American, Asian-Specified, Asian-Chinese, Asian-Japanese, Asian-Korean, Asian-Vietnamese, Asian-Cambodian, Asian-Thai, Asian-Hmong, Other-Specified, Indian (Excludes American Indian, Aleut, and Eskimo), Filipino, Hawaiian, Guamanian, Samoan, Eskimo, Aleut, Pacific Islander (Excludes Hawaiian, Guamanian, Samoan), Withheld, and Unknown. The categories of Hispanic ancestry or origin were Not Hispanic, Mexican, Puerto Rican, Cuban, Other Hispanic, or Unknown/Withheld. We summarized these into the following categories: White, Black, Asian, American Indian/Alaska Native, Hawaiian/Pacific Islander, multirace, other race/ethnicity, and Hispanic/Latinx. Any participant with Hispanic/Latinx ethnicity was coded as Hispanic/Latinx, except for the multirace group. Insurance categories were public, private, and other payment. Public health insurance included Medicaid, other government insurance, and Indian Health Service. Before 2005, public insurance also included Medicare, worker’s compensation, and Title V. Private insurance included private insurance and CHAMPUS/TRICARE. Before 2005, private insurance also included categories for Blue Cross–Blue Shield and health maintenance organization (HMO). Other payment source includes a medically unattended delivery, self-pay, and other payment. Before 2005, other options included no charge, other nongovernmental insurance, and medically indigent. Educational attainment was categorized as less than high school attainment, high school graduate, some college, college graduate, and graduate school graduate, which included a master’s degree, doctorate, or professional degree. All individual characteristics were collected from the birth record.

We also adjusted for the total population served by the water system, drinking water arsenic levels, county-level annual average temperature from the National Oceanic and Atmospheric Administration,50 and urbanicity from the National Center for Health Statistics.51 The urbanicity measure is based on the Office of Management and Budget’s classification of metropolitan and micropolitan statistical areas and is designed to characterize health differences across urban and rural areas. Ambient temperature has historically been considered in fluoride recommendations, because it was assumed that people living in warmer climates drink more water, and therefore fluoride levels should be lower to prevent excess consumption.52,53 We included county-level unemployment from the Bureau of Labor Statistics54 and income inequality, represented as the ratio of the 80th to the 20th percentiles of household income, from the County Health Rankings55 as proxies for public finances and area-level socioeconomic status.

Effect Modification

We performed various stratified analyses to assess whether there were any populations for whom fluoride exposure was particularly harmful. We were particularly interested in identifying whether socially or economically marginalized groups experience larger impacts that may raise environmental justice concerns. Thus, we estimated associations within each subgroup of race/ethnicity and health insurance status and assessed differences in associations with each hypothetical intervention across groups. Given the effects of in-utero exposure to chemicals including fluoride can differ by fetal sex,6–8,56,57 we also evaluated stratified by sex. Finally, given the potential for synergistic effects with arsenic,44 we stratified analyses by whether the water system had arsenic levels that exceeded the maximum contaminant level (MCL) of 10 ppb or exceeded the 75th percentile of arsenic levels (2.53 ppb) observed across California births during the study period. To evaluate whether associations differed meaningfully across groups, we used a Wald test for heterogeneity.58

Sensitivity Analyses …

Results

The fluoride levels varied within and across water systems over the study period, 2000–2018, and patterns differed based on the urbanicity of the water system (Figure 1). There is a large increase in water fluoridation levels around 2007, which is when the Metropolitan Water District started fluoridating water it delivered to Southern California districts.38 There is also a decrease starting in 2015, which is when the recommended level of fluoride changed to 0.7 ppm for all districts, whereas previously it ranged from 0.7–1.2 ppm, depending on the ambient temperature.

Figure 1.

Figure 1 is a set of six line graphs titled Large central metro, Large fringe metro, Medium metro, Small metro, Micropolitan, Noncore, plotting Average fluoride (parts per million), ranging from 0.0 to 1.2 in increments of 0.2 (y-axis) across Year, ranging as 2000 to 2018 in increments of 5 (x-axis) for Alameda, Los Angeles, Orange, Riverside, Sacramento, San Diego, San Francisco, Santa Clara; Contra Costa, El Dorado, Marin, Placer, San Benito, San Bernardino, San Mateo, Yolo; Fresno, Kern, Merced, Monterey, San Joaquin, San Luis Obispo, Santa Barbara, Santa Cruz, Solano, Sonoma, Stanislaus, Tulare, Ventura; Butte, Imperial, Kings, Madera, Napa, Shasta, Sutter, Yuba; Del Norte, Humboldt, Lake, Lassen, Mendocino, Nevada, Tehama, Tuolumne; and Amador, Calaveras, Colusa, Glenn, Inyo, Mono, Plumas, and Siskiyou, respectively.

Open in a new tab
Average drinking water fluoride levels in California community water systems from 2000 to 2018 by county urbanicity. Note: These county urbanicity categories are defined using the National Center for Health Statistics Urban-Rural Classification Scheme.

After limiting to our inclusion criteria, we had a sample size of 5,008,915 births (Supplemental Figure 1). The demographics of all births compared with those included in the study during the 2000–2018 period are shown in Supplemental Table 2. The median fluoride level among participants in our study was 0.51 ppm (Table 1). Black, Asian, and Hawaiian/Pacific Islander participants had higher median fluoride levels (0.61, 0.59, and 0.60 ppm, respectively) in comparison with White (0.52 ppm) and Hispanic/Latinx participants (0.49 ppm). Those with college or graduate level education had higher median exposure levels (0.59 and 0.61 ppm, respectively) in comparison with those with a high school education or less (0.49 ppm).

Table 1.

Descriptive statistics of drinking water fluoride levels (in ppm) among California births from 2000 to 2018 by individual-level and area-level covariates.

Births n % Min Max Mean Median SD
All 5,008,915 100.0% 0.00 2.50 0.49 0.51 0.28
Race/ethnicity
White 1,243,166 24.8% 0.00 2.50 0.49 0.52 0.29
Black 290,392 5.8% 0.00 1.95 0.57 0.61 0.28
Asian 657,290 13.1% 0.00 2.50 0.54 0.59 0.29
American Indian/Alaska Native 12,411 0.2% 0.00 2.50 0.42 0.38 0.31
Hawaiian or Pacific Islander 21,828 0.4% 0.00 1.52 0.56 0.60 0.30
Other 3,832 0.1% 0.00 1.80 0.54 0.60 0.28
Multirace 143,044 2.9% 0.00 2.50 0.51 0.55 0.30
Hispanic/Latinx 2,636,952 52.6% 0.00 2.50 0.47 0.49 0.27
Education
Less than high school 1,187,518 23.7% 0.00 2.50 0.47 0.49 0.28
High school graduate 1,450,109 29.0% 0.00 2.50 0.47 0.47 0.28
Some college 1,122,950 22.4% 0.00 2.50 0.48 0.50 0.29
College graduate 799,154 16.0% 0.00 2.50 0.53 0.59 0.28
Graduate school 449,184 9.0% 0.00 2.28 0.56 0.61 0.27
Insurance type
Public insurance 2,505,737 50.0% 0.00 2.50 0.48 0.49 0.28
Private insurance 2,306,198 46.0% 0.00 2.50 0.51 0.54 0.29
Other payment 190,877 3.8% 0.00 2.50 0.52 0.56 0.25
Birthing parent age
<25 1,427,768 28.5% 0.00 2.50 0.47 0.47 0.28
25–34 2,651,528 52.9% 0.00 2.50 0.49 0.51 0.28
>35 929,619 18.6% 0.00 2.50 0.53 0.58 0.28
Month of conception
January 437,130 8.7% 0.00 2.50 0.49 0.51 0.28
February 398,438 8.0% 0.00 2.50 0.49 0.51 0.28
March 421,137 8.4% 0.00 2.50 0.49 0.51 0.28
April 399,449 8.0% 0.00 2.50 0.49 0.50 0.28
May 415,146 8.3% 0.00 2.50 0.49 0.50 0.28
June 401,076 8.0% 0.00 2.50 0.49 0.50 0.28
July 402,113 8.0% 0.00 2.50 0.49 0.51 0.28
August 408,741 8.2% 0.00 2.50 0.49 0.51 0.28
September 401,093 8.0% 0.00 2.50 0.49 0.51 0.28
October 431,542 8.6% 0.00 2.50 0.49 0.51 0.28
November 433,389 8.7% 0.00 2.50 0.49 0.51 0.28
December 459,661 9.2% 0.00 2.50 0.49 0.51 0.28
Year of conception
2000 89,053 1.8% 0.00 1.50 0.39 0.30 0.27
2001 97,473 1.9% 0.00 1.50 0.39 0.30 0.27
2002 102,387 2.0% 0.00 1.50 0.40 0.30 0.28
2003 125,512 2.5% 0.00 1.50 0.44 0.32 0.28
2004 143,420 2.9% 0.00 1.80 0.45 0.35 0.28
2005 239,799 4.8% 0.00 1.80 0.43 0.35 0.28
2006 306,139 6.1% 0.00 1.80 0.43 0.39 0.28
2007 335,161 6.7% 0.00 1.94 0.49 0.50 0.27
2008 338,484 6.8% 0.01 1.95 0.55 0.58 0.27
2009 328,656 6.6% 0.01 1.95 0.55 0.56 0.29
2010 342,261 6.8% 0.00 1.96 0.54 0.55 0.29
2011 346,236 6.9% 0.00 1.90 0.52 0.56 0.29
2012 351,588 7.0% 0.00 1.85 0.52 0.58 0.29
2013 358,933 7.2% 0.00 1.85 0.51 0.59 0.29
2014 358,997 7.2% 0.00 2.50 0.51 0.59 0.28
2015 359,640 7.2% 0.00 2.50 0.49 0.56 0.28
2016 355,186 7.1% 0.00 2.50 0.47 0.53 0.26
2017 342,152 6.8% 0.00 2.00 0.48 0.51 0.26
2018 87,838 1.8% 0.00 1.94 0.47 0.51 0.26
Total population served by the water system
<10,000 102,860 2.1% 0.00 2.50 0.30 0.20 0.29
10,000 or more 4,906,055 97.9% 0.00 1.94 0.50 0.52 0.28
Annual average temperature
Less than or equal to 12.1°C (53.7°F) 23,000 0.5% 0.01 0.90 0.31 0.21 0.27
More than 12.1°C (53.7°F)–14.6°C (58.3°F) 601,582 12.0% 0.00 2.50 0.35 0.17 0.35
More than 14.6°C (58.3°F)–17.7°C (63.8°F) 2,797,716 55.9% 0.00 2.50 0.52 0.58 0.29
More than 17.7°C (63.8°F)–21.4°C (70.6°F) 1,525,026 30.4% 0.00 1.85 0.50 0.50 0.23
More than 21.4°C (70.6°F) 61,591 1.2% 0.04 1.38 0.48 0.49 0.21
Urbanicity
Large central metro 3,551,807 70.9% 0.00 1.96 0.55 0.59 0.25
Large fringe metro 542,521 10.8% 0.00 1.53 0.47 0.46 0.29
Medium metro 777,224 15.5% 0.00 2.50 0.27 0.15 0.29
Small metro 113,477 2.3% 0.00 1.85 0.36 0.14 0.41
Micropolitan 17,944 0.4% 0.00 0.93 0.16 0.10 0.17
Noncore 5,942 0.1% 0.02 0.49 0.17 0.09 0.12
Unemployment
<5% 1,119,058 22.3% 0.00 1.80 0.49 0.49 0.26
5% to <10% 2,573,112 51.4% 0.00 2.50 0.50 0.54 0.28
10% or more 1,316,745 26.3% 0.00 2.50 0.47 0.49 0.31
Income inequality
<5 2,907,873 58.1% 0.00 2.00 0.42 0.41 0.28
5 or more 2,101,042 41.9% 0.00 2.50 0.58 0.65 0.26
Drinking water arsenic
Below MCL 4,932,121 98.5% 0.00 2.50 0.49 0.51 0.28
At or above MCL 76,794 1.5% 0.02 1.85 0.55 0.68 0.32
Open in a new tab

Note: We have created categories of birthing parent age, total population served by the water system, annual average temperature, unemployment, income inequality, and drinking water arsenic for this table but they are included as continuous covariates in the models. The categories for annual average temperature are based on the recommended control limits for fluoridation from the 1962 Drinking Water Standards. The MCL for arsenic is 10 ppb. MCL, maximum contaminant level; ppb, parts per billion; ppm, parts per million; SD, standard deviation.

The associations of fluoride with birth weight, gestational age, and birth-weight-for-gestational-age z-scores were nonlinear (Supplemental Figures 3–5). Hypothetical interventions to reduce fluoride levels were not associated with changes in the risk of preterm birth or small-for-gestational age (Figure 2; Supplemental Excel Table S1). We estimated small reductions in birth weight associated with the hypothetical intervention to 0.5 ppm (-5.8g; 95% CI: -10.0, -1.6) and for the hypothetical intervention to 0.7 ppm (-2.2g; 95% CI: -4.4, 0.0). We also estimated small reductions in birth-weight-for-gestational-age z-scores; the association was slightly larger for the hypothetical intervention to 0.5 ppm (-0.006; 95% CI: -0.013, 0.000) in comparison with the hypothetical intervention to 0.7 ppm (-0.004; 95% CI: -0.007, 0.000). Differences in fetal growth may have occurred by changes toward the right side of the distribution, because we estimated small reductions in the risk of large-for-gestational-age for both the hypothetical intervention to 0.5 ppm (RD=-0.001; 95% CI: -0.003, 0.000) and to 0.7 ppm (RD=-0.001; -0.002, 0.000). We also estimated reductions in the risk of macrosomia for the hypothetical intervention to 0.5 ppm (RD=-0.002; 95% CI: -0.004, -0.001) and to 0.7 ppm (RD=-0.001; 95% CI: -0.002, 0.000). We estimated small reductions in gestational age associated with the hypothetical intervention to 0.5 ppm (-0.01 wk; 95% CI: -0.03, 0.00) and 0.7 ppm (-0.01 wk; 95% CI: -0.02, 0.00). For each outcome, the direction of the association changed (lower birth weight, birth-weight-for-gestational-age z-scores, and gestational age) when fluoride levels were above 1.0 ppm (Supplemental Figures 3–5), although CIs were wide. When removing implausible combinations of birth weight and gestational age, we saw minimal differences in the findings (Supplemental Table 3).

Figure 2.

Figure 2 is a set of seven graphs titled Birthweight, Gestational age, Birthweight z?score, Preterm birth, SGA, LGA, Macrosomia, plotting Change in birth weight (gram), ranging from negative 10.0 to 2.5 in increments of 2.5; Change in gestational age (weeks), ranging from negative 0.04 to 0.02 in increments of 0.02; Change in birth weight z?score, ranging from negative 0.04 to 0.02 in increments of 0.02; Change in risk of preterm birth, ranging from negative 0.00050 to 0.0050 in increments of 0.0025; Change in risk of S G A, ranging from negative 0.0050 to 0.0050 in increments of 0.0025; Change in risk of L G A, ranging from negative 0.0050 to 0.0050 in increments of 0.0025; Change in risk of macrosomia, ranging from negative 0.0050 to 0.0050 in increments of 0.0025 (y-axis) across reduce to 0.5 part per million and reduce to 0.7 part per million (x-axis), respectively.

Open in a new tab

Associations of hypothetical interventions to reduce drinking water fluoride among California births from 2000 to 2018 with birth weight, gestational age, birth-weight-for-gestational-age z-scores, and risk of preterm birth, small-for-gestational age, large-for-gestational age, and macrosomia (n=5,008,915). Note: To estimate the effects of the hypothetical interventions, we used G-computation with linear regression and a spline for the fluoride levels with a non-parametric bootstrap clustered by water system for inference. All models were adjusted for birthing parent age, race/ethnicity, health insurance type, educational attainment, month and year of conception, total population served by the water system, county-level annual average temperature, urbanicity, unemployment, income inequality, and drinking water arsenic levels. The point estimates are represented by diamond shapes, and the 95% CIs are represented by the error bars. These results are available in Excel Table S1. CI, confidence level.

When stratifying by race/ethnicity, we saw no consistent differences in birth weight, gestational age, or birth weight-for-gestational age z-scores across racial/ethnic groups for either of the hypothetical interventions (Figure 3). We also did not observe differences across racial/ethnic groups for preterm birth (Supplemental Figure 6), small-for-gestational age (Supplemental Figure 7), or large-for-gestational age (Supplemental Figure 8). We did observe some differences in the estimated effect of hypothetical fluoride interventions by insurance status. For example, the association of birth weight with the hypothetical intervention to 0.7 ppm was in the opposite direction for those with private insurance (-2.6g; 95% CI: -5.1, -0.2) and public insurance (-2.3g; 95% CI: -4.3, -0.3) in comparison with those with another source of payment (2.6g; 95% CI: -0.6, 5.8) (Figure 4). The association for the hypothetical intervention to 0.5 ppm was -5.5g (95% CI: -9.9, -1.1) for those with private insurance, -6.3g (95% CI: -11.0, -1.5) for those with public insurance, and -1.7g (95% CI: -8.8, 5.4) for those with other insurance. The patterns were similar for gestational age and birth-weight-for-gestational-age z-scores (Supplemental Figures 9 and 10). We did not see consistent differences across insurance status for preterm birth (Supplemental Figure 11), small-for-gestational age (Supplemental Figure 12), or large-for-gestational age (Supplemental Figure 13). We did not find any differences in the estimated effects of the hypothetical interventions by fetal sex for any of the pregnancy outcomes studied (Supplemental Figures 14–19).

Figure 3.

Figure 3 is a set of three graphs titled Birthweight, Gestational age, and Birthweight?for?gestational age z?score, plotting Change in birth weight (gram), ranging from negative 30 to 30 in increments of 10; Change in gestational age (weeks), ranging from negative 0.1 to 0.1 in increments of 0.1; Change in birth weight z?scores, ranging from negative 0.05 to 0.05 in increments of 0.05 (y-axis) across All, White, Hispanic or Latinx, Black, Asian, American Indian, Hawaiian or Pacific Islander, Multirace, and Other (x-axis) for Hypothetical intervention, including 0.5 part per million and 0.7 part per million, respectively

Open in a new tab
Associations of hypothetical interventions to reduce drinking water fluoride among California births from 2000 to 2018 with birth weight, gestational age, and birth-weight-for-gestational age z-scores by racial/ethnic groups. Note: To estimate the effects of the hypothetical interventions, we used G-computation with linear regression and a spline for the fluoride levels with a non-parametric bootstrap clustered by water system for inference. All models were adjusted for birthing parent age, race/ethnicity, health insurance type, educational attainment, month and year of conception, total population served by the water system, county-level annual average temperature, urbanicity, unemployment, income inequality, and drinking water arsenic levels. Race-specific models did not include race/ethnicity as a covariate. The point estimates are represented by diamond shapes, and the 95% CIs are represented by the error bars. The Wald test for heterogeneity across racial groups showed no difference for birth weight (p=0.77 for intervention to 0.5 ppm, p=0.27 for intervention to 0.7 ppm), no difference for gestational age (p=0.69 for intervention to 0.5 ppm, p=0.48 for intervention to 0.7 ppm), and no difference for birth-weight-for-gestational age z-scores (p=0.58 for intervention to 0.5 ppm), but there was a potential difference for the intervention to 0.7 ppm (p=0.02). These results are available in Excel Table S1. The sample sizes for each group are as follows: All n=5,008,915; White n=1,243,166; Hispanic/Latinx n=2,636,952; Black n=290,392; Asian n=657,290; American Indian n=12,411; Hawaiian/Pacific Islander n=21,828; Multirace n=143,044; Other race n=3,832. CI, confidence interval.

Figure 4.

Figure 4 is a set of two graphs titled Hypothetical intervention to 0.5 parts per million and Hypothetical intervention to 0.7 parts per million, plotting Change in birth weight (gram), ranging from negative 15 to 10 in increments of 5 (y-axis) across Private insurance, Public insurance, and Other payment (x-axis), respectively.

Open in a new tab
Associations of hypothetical interventions to reduce drinking water fluoride among California births from 2000 to 2018 with birth weight by insurance status. Note: To estimate the effects of the hypothetical interventions, we used G-computation with linear regression and a spline for the fluoride levels with a non-parametric bootstrap clustered by water system for inference. Models were adjusted for birthing parent age, race/ethnicity, educational attainment, month and year of conception, total population served by the water system, county-level annual average temperature, urbanicity, unemployment, income inequality, and drinking water arsenic levels. The point estimates are represented by diamond shapes, and the 95% CIs are represented by the error bars. The Wald test for heterogeneity across insurance groups showed no difference for the intervention to 0.5 ppm (p=0.51), whereas there was a potential difference between groups for the intervention to 0.7 ppm (p=0.01). These results are available in Excel Table S1. The sample sizes for each group are as follows: Private insurance n=2,306,198; Public insurance n=2,505,737; Other payment n=190,877. CI, confidence interval.

When stratifying by arsenic concentrations in drinking water, we did not observe any differences in the effects of a hypothetical intervention to 0.5 ppm for people with arsenic above or below the MCL (Supplemental Figure 20). However, when we analyze the effects for the hypothetical intervention to 0.7 ppm, there are some potential differences by arsenic levels (Supplemental Figure 21). For example, we did not observe an association for small-for-gestational age in the group below the arsenic MCL (RD=0.000; 95% CI: 0.000, 0.001), but we saw a negative association for those above the arsenic MCL (RD=-0.002; 95% CI: -0.005, 0.000). Among those below the arsenic MCL, we also saw a small negative association with large-for-gestational age (RD=-0.001; 95% CI: -0.002, 0.000), whereas the association was in the opposite direction for those above the arsenic MCL (RD=0.005; 95% CI: -0.002, 0.011). Similarly, among those below the arsenic MCL, there was no association with preterm birth (RD=0.000; 95% CI: 0.000, 0.001), but we observed a positive association for those above the MCL (RD=0.005; 95% CI: -0.002, 0.011). However, these differences were not consistent with the results of the hypothetical intervention to 0.5 ppm, nor with the results that stratified above and below the 75th percentile of arsenic levels, rather than the MCL (Supplemental Figures 22 and 23).

Finally, we analyzed the associations of a 0.1-ppm change in fluoride levels on each of the perinatal outcomes among siblings in California. Due to exclusion of the year 2000 data, single children, and missing data among any of the siblings, our sample size was reduced to 1,386,826 (Supplemental Table 4). We also compared the results of a 0.1-ppm change in fluoride among all births in California and among the unmatched sibling data (Table 2). For both the overall analysis and the unmatched sibling analysis, we found that higher fluoride levels of 0.1 ppm were associated with an increase in birth weight (all births RD=1.51g; 95% CI: 1.32, 1.71 and unmatched siblings RD=1.68g; 95% CI: 1.32, 2.05). However, when we analyzed the matched siblings, we did not see an association (RD=-0.03; 95% CI: -0.38, 0.33). We observed a similar pattern for birth-weight-for-gestational-age z-scores. We did observe small consistent positive associations with gestational age and negative associations with risk of preterm birth across all births, the matched sibling, and unmatched sibling populations. For instance, a 0.1-ppm increase in fluoride was associated with a reduction in risk of preterm birth of -0.0002 (95% CI: -0.0003, -0.0001) among all births, -0.0007 (95% CI: -0.0009, -0.0006) among the matched siblings, and -0.0004 (95% CI: -0.0005, -0.0002) among the unmatched siblings. Associations for small-for-gestational age and large-for-gestational age were inconsistent across the study populations. However, we did observe positive associations of a 0.1-ppm increase in fluoride levels with the risk of macrosomia.

Table 2.

Associations of a 0.1-ppm increase in drinking water fluoride with birth weight, gestational age, birth-weight-for-gestational-age z-scores, preterm birth, small for gestational age, large for gestational age, and macrosomia among the full sample of births from 2000 to 2018 and matched and unmatched sibling births from 2001 to 2018 in California.

Outcome All births n=5,008,915 Matched siblings
n =1,386,826		 Unmatched siblings n=1,386,826
Estimate
(95% CI)		 Estimate
(95% CI)		 Estimate
(95% CI)
Birth weight (grams) 1.51
(1.32, 1.71)		 -0.03
(-0.38, 0.33)		 1.68
(1.32, 2.05)
Gestational age (wk) 0.006
(0.006, 0.007)		 0.007
(0.006, 0.008)		 0.007
(0.006, 0.009)
Birth-weight-for-gestational-age z-scores 0.002
(0.002, 0.002)		 -0.003
(-0.003, -0.002)		 0.002
(0.001, 0.003)
Preterm birth -0.0002
(-0.0003, -0.0001)		 -0.0007
(-0.0009, -0.0006)		 -0.0004
(-0.0005, -0.0002)
Small for gestational age -0.0001
(-0.0002, 0.0000)		 0.0001
(-0.0001, 0.0003)		 -0.0001
(-0.0003, 0.0001)
Large for gestational age 0.0005
(0.0004, 0.0006)		 -0.0006
(-0.0008, -0.0004)		 0.0006
(0.0004, 0.0008)
Macrosomia 0.0006
(0.0005, 0.0007)		 0.0003
(0.0001, 0.0005)		 0.0007
(0.0005, 0.0009)
Open in a new tab

Note: The overall analysis uses a linear regression model to estimate the effect of a 0.1-ppm increase in water fluoride levels across all births in California from 2000 to 2018. The matched sibling analysis estimates the effect of a 0.1-ppm change in fluoride levels within siblings from 2001 to 2018. The unmatched sibling analysis estimates the effect of a 0.1-ppm increase in fluoride across all sibling births in California from 2001 to 2018, without taking into account the sibling relationships. The estimates for preterm birth, small-for-gestational age, large-for-gestational age, and macrosomia represent risk differences. CI, confidence level; ppm, parts per million.

To further assess whether a difference in population composition or confounding is driving the differences in results, we compared the demographic characteristics and distribution of fluoride levels for each group between the siblings and the main analysis (Supplemental Table 4). Despite the reduction in sample size from 5,008,915 to 1,386,826, we did not observe many demographic differences between the populations, although the sibling group was slightly more likely to be Hispanic or Latinx, have less than a high school education, and receive public insurance. The distribution of fluoride levels was also similar between the main analysis and the sibling group.

Discussion

We conducted a population-level study to characterize drinking water fluoride exposure among pregnant people and to evaluate the association between water fluoridation and pregnancy outcomes in the United States. We found small negative associations with birth weight, gestational age, birth-weight-for-gestational-age z-scores, risk for large-for-gestational age, and macrosomia associated with hypothetical interventions to reduce fluoride levels to 0.5 and 0.7 ppm, but no associations with risk of preterm birth or small-for-gestational age. This finding suggests that higher fluoride levels as observed in our study are unlikely to increase risk for the most severe adverse birth outcomes, like preterm birth and small-for-gestational age, but levels above may slightly increase fetal growth beyond what is ideal. Large-for-gestational-age babies are at greater risk for birth injury, including fracture, shoulder dystocia, and Erb’s palsy.61,62 Very macrosomic babies with birth weights over 4,500g are at higher risk for birth complications and death.63 Large-for-gestational-age and macrosomic babies can also have increased risk for cardiometabolic disorders later in life.64–66 However, the risk differences we observed were very small, and although the findings contribute to data on etiological investigations of adverse birth outcomes in relation to fluoride on a population level, the results were not strong enough to impact changes in clinical guidance or practice. Furthermore, we observed only increased risk of macrosomia but not large-for-gestational age, with higher levels of fluoride in the sibling-matched analysis. Nevertheless, gestational diabetes is an important risk factor for large-for-gestational age and has been increasing in prevalence since 2016.67 Water fluoridation has been linked to alterations in thyroid hormone levels,68 and thyroid disease and diabetes are physiologically related,69 which suggest potential mechanisms linking fluoride levels with neonate size. Examining the relationship of fluoride with pregnancy health complications, including gestational diabetes, preeclampsia, and pregnancy-related hypertensive disorders may be important for future inquiry.

Our main findings are consistent with a cohort study in Sweden, which found that higher urinary fluoride levels during pregnancy were associated with higher birth weight, birth-weight-for-gestational-age z-scores, and risk of large-for-gestational age births.17 In this study, median urinary fluoride levels were 0.71mg/L, and an increase of 1mg/L was associated with an increase in birth weight (RD=84g; 95% CI: 30, 138) and increased odds of large-for-gestational age (OR=1.39; 95% CI: 1.03, 1.89). In addition, a study in Mexico found urinary fluoride levels were positively associated with birth-weight-for-gestational-age z-scores, but only in the first trimester,18 where an increase of 1mg/L in maternal urinary fluoride above 0.99mg/L was associated with an increase in birth-weight z-scores (RD=0.79; 95% CI: 0.10, 1.48). A study in Canada with median maternal urinary fluoride of 0.5mg/L also found no associations of urinary fluoride levels with the risk of preterm birth or small-for-gestational age.21 These studies used urinary fluoride concentrations to evaluate the reproductive impacts of exposure to fluoride from all sources, whereas our study estimated the impact of exposure to drinking water fluoride specifically.

Evidence evaluating associations with drinking water fluoride, rather than urinary fluoride, is less consistent. For instance, living in a ZIP code with fluoridated water was associated with increased risk of preterm birth in New York State,19 and one study using National Health and Nutrition Examination Survey (NHANES) data found that higher fluoride levels in drinking water were associated with low birth weight, but only among Hispanic mothers. One study in Massachusetts found community water fluoridation was protective for preterm birth, especially when combined with dental cleaning during pregnancy,16 and another Massachusetts study found higher water fluoride levels were protective against major malformations and neonatal deaths.20

Studies in Africa and South Asia have tended to find associations between higher drinking water fluoride levels with adverse birth outcomes. For instance, a study in India found 8.7 higher odds of preterm birth and 10.6 higher odds of low birth weight among mothers with serum fluoride levels >1 ppm,70 and a study in Senegal found increased risk of low birth weight among mothers with dental fluorosis and those who consumed well water during pregnancy in areas of endemic fluoride contamination.71 Pregnant women in India with serum fluoride above 1 ppm were found to have high rates of anemia and congenital abnormalities in offspring.28 Differences between these findings and the current study may be due to exposure levels, because mothers in India and Senegal were exposed to much higher fluoride levels on average in comparison with California mothers. In addition, these studies represent areas in which fluoride occurs naturally in high amounts, whereas in California there is a mix of naturally occurring and treatment-related fluoride.

This work had several important limitations. The water fluoride levels likely included measurement error because we used annual averages, which are the only estimates available from the consumer confidence reports from community water systems. In addition, these measures tend to round to one decimal place, which may reduce precision, especially for lower levels of exposure. We used the detection limit for the purpose of reporting (0.1 ppm) divided by the square root of 2 when fluoride levels were reported to be nondetected. Given the limited number of nondetects in the data (7.5% of water system-years), we expect this approach introduced only minimal bias.72 The reports sometimes did not clarify the proportion of water used from different sources, and in this case, we used a simple average of the sources. We geocoded participants based on their address at delivery, but people may have moved during pregnancy, which we were not able to capture. We also excluded participants living outside of water boundary service areas from our analysis because we did not have estimates of fluoride levels for them. In addition, people relying on domestic wells who live within water system service boundaries may have been misclassified. Previous studies have shown some groups are less likely to drink tap water and thus may have different levels of exposure,15,73 but we were not able to measure drinking water consumption behaviors in our study. However, our previous study evaluated the relationship between community water system annual fluoride levels and fluoride levels in maternal urine, serum, and amniotic fluid and found significant correlations,43 suggesting the water system levels do capture variation in fluoride exposure and reflect prenatal exposures, including fetal exposures. Additional studies outside the United States also suggest water fluoride levels are an important predictor of fluoride measures in urine and plasma of pregnant people.74 However, we did not measure fluoride within individuals and were unable to include sources of fluoride other than drinking water, including tea and dental products.

There was also some missingness in the fluoride data. Many water systems did not keep records back to 2000, and several systems did not respond to our public records requests. Specifically, the California American Water Company and California Water Service Company did not respond to our multiple public records requests. Thus, we were able to include only data on their water systems for recent years, which were publicly available. We were also unable to assess whether there are critical windows of exposure during the prenatal period because the exposure information was only available annually. Previous studies in Mexico have suggested minimal changes in urinary and plasma fluoride levels over the course of pregnancy,76 although there may be differential effects of fluoride on fetal growth depending on the trimester of exposure.18 We were unable to evaluate whether associations differed by the type of chemical used for fluoridation due to lack of data. Previous studies have suggested potential effects of water fluoridation on fertility.77 For instance, an ecological study of water fluoridation and birth rates found reduced fertility linked to water fluoride levels in the United States,77 and toxicological evidence suggests fluoride exposure can inhibit both male78,79 and female fertility.80 It is also possible our results were biased downward due to live birth bias and the exclusion of miscarriages and stillbirths in the population.81 Both reductions in fertility and higher risk of fetal loss can mask associations of chemicals with birth outcomes due to selection bias.

More than 200 million Americans receive fluoridated water, and no previous studies to our knowledge had evaluated the effect on birth outcomes at the population level. This study built a novel database of fluoride levels in California community water systems using consumer confidence reports collected using public records requests and linked those records to more than 5 million births. We did not find associations with either preterm birth or being small for gestational age, both birth outcomes that have the strongest associations with later morbidity and mortality. However, we did estimate small reductions in birth weight, gestational age, and risk for large-for-gestational age associated with hypothetical interventions to reduce fluoride levels. These results suggest that fluoride levels above 0.5 and 0.7 ppm are associated with higher birth weights, longer gestational ages, and greater risk for large-for-gestational age and macrosomia. We did not observe differences in associations by race/ethnicity or insurance status, which in combination with our mostly null results suggests that water fluoridation does not contribute to inequities in adverse birth outcomes. Our hypothesis that water fluoride would be associated with birth outcomes was motivated by the associations of in utero and early-life fluoride consumption with lower IQ,12 because birth outcomes like preterm birth and being small for gestational age are linked to later neurodevelopment. However, given we did not find associations with these perinatal health outcomes (and if anything found small protective effects for preterm birth in the sibling analyses), the effects of fluoride on neurodevelopment likely operate through alternate mechanistic pathways.

The US Environmental Protection Agency (US EPA) regulates the level of fluoride in drinking water, and the current maximum contaminant level goal is 4.0 ppm, which establishes the level to prevent adverse health effects in the general population. However, it is difficult to claim this standard protects maternal and infant health, because very limited evidence exists. In our study the maximum fluoride level was 2.5 ppm, and although the association of fluoride with each outcome was nonlinear, the directions of effect were largely negative for birth weight, birth-weight-for-gestational-age z-scores, and gestational age when fluoride levels exceeded 1 ppm. Additional research to evaluate associations with other measures of reproductive health and child development will help ensure the benefits of fluoridation for caries prevention do not come at the cost of other adverse health outcomes.

Supplementary Material

ehp13732.smcontents.508.pdf (17.4KB, pdf)
ehp13732.s001.acco.pdf (699.4KB, pdf)
ehp13732.s002.codeanddata.acco.zip (23.5KB, zip)

Acknowledgments

This work was supported by National Institute of Environmental Health Sciences (NIEHS) K99ES033274 (D.E.G.), National Institutes of Health (NIH) P30ES030284 (T.J.W.), and NIEHS Superfund Research Program P42ES004705 (R.M.F.).

Conclusions and opinions are those of the individual authors and do not necessarily reflect the policies or views of EHP Publishing or the National Institute of Environmental Health Sciences.

References

  • 1.US CDC (US Centers for Disease Control and Protection). 1999. Achievements in public health, 1900–1999: fluoridation of drinking water to prevent dental caries. MMWR 48(41):933–940. [Google Scholar]
  • 2.US CDC. 1995. Engineering and Administrative Recommendations for Water Fluoridation, 1995. https://www.cdc.gov/mmwr/preview/mmwrhtml/00039178.htm [accessed 12 November 2023].
  • 3.McDonagh MS, Whiting PF, Wilson PM, Sutton AJ, Chestnutt I, Cooper J, et al. 2000. Systematic review of water fluoridation. BMJ 321(7265):855–859, PMID: 11021861, 10.1136/bmj.321.7265.855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Valdez Jiménez L, López Guzmán OD, Cervantes Flores M, Costilla-Salazar R, Calderón Hernández J, Alcaraz Contreras Y, et al. 2017. In utero exposure to fluoride and cognitive development delay in infants. Neurotoxicology 59:65–70, PMID: 28077305, 10.1016/j.neuro.2016.12.011. [DOI] [PubMed] [Google Scholar]
  • 5.Choi AL, Sun G, Zhang Y, Grandjean P. 2012. Developmental fluoride neurotoxicity: a systematic review and meta-analysis. Environ Health Perspect 120(10):1362–1368, PMID: 22820538, 10.1289/ehp.1104912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Green R, Lanphear B, Hornung R, Flora D, Martinez-Mier EA, Neufeld R, et al. 2019. Association between maternal fluoride exposure during pregnancy and IQ scores in offspring in Canada. JAMA Pediatr 173(10):940–948, PMID: 31424532, 10.1001/jamapediatrics.2019.1729. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Cantoral A, Téllez-Rojo MM, Malin AJ, Schnaas L, Osorio-Valencia E, Mercado A, et al. 2021. Dietary fluoride intake during pregnancy and neurodevelopment in toddlers: a prospective study in the progress cohort. Neurotoxicology 87:86–93, PMID: 34478773, 10.1016/j.neuro.2021.08.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Grandjean P, Hu H, Till C, Green R, Bashash M, Flora D, et al. 2022. A benchmark dose analysis for maternal pregnancy urine-fluoride and IQ in children. Risk Anal 42(3):439–449, PMID: 34101876, 10.1111/risa.13767. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Grandjean P. 2019. Developmental fluoride neurotoxicity: an updated review. Environ Health 18(1):110, PMID: 31856837, 10.1186/s12940-019-0551-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ibarluzea J, Gallastegi M, Santa-Marina L, Jiménez Zabala A, Arranz E, Molinuevo A, et al. 2022. Prenatal exposure to fluoride and neuropsychological development in early childhood: 1-to 4 years old children. Environ Res 207:112181, PMID: 34627799, 10.1016/j.envres.2021.112181. [DOI] [PubMed] [Google Scholar]
  • 11.Aggeborn L, Öhman M. 2021. The effects of fluoride in drinking water. J Pol Econ 129(2):465–491, 10.1086/711915. [DOI] [Google Scholar]
  • 12.NASEM (National Academies of Sciences, Engineering and Medicine). 2020. Review of the Draft NTP Monograph: Systematic Review of Fluoride Exposure and Neurodevelopmental and Cognitive Health Effects. Washington, DC: National Academies Press. [PubMed] [Google Scholar]
  • 13.Cheong JL, Doyle LW, Burnett AC, Lee KJ, Walsh JM, Potter CR, et al. 2017. Association between moderate and late preterm birth and neurodevelopment and social-emotional development at age 2 years. JAMA Pediatr 171(4):e164805, PMID: 28152144, 10.1001/jamapediatrics.2016.4805. [DOI] [PubMed] [Google Scholar]
  • 14.Ferguson KK, Sammallahti S, Rosen E, van den Dries M, Pronk A, Spaan S, et al. 2021. Fetal growth trajectories among small for gestational age babies and child neurodevelopment. Epidemiology 32(5):664–671, PMID: 34086648, 10.1097/EDE.0000000000001387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Arun AK, Rustveld L, Sunny A. 2022. Association between water fluoride levels and low birth weight: National Health and Nutrition Examination Survey (NHANES) 2013–2016. Int J Environ Res Public Health 19(15):8956, PMID: 35897326, 10.3390/ijerph19158956. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Zhang X, Lu E, Stone SL, Diop H. 2019. Dental cleaning, community water fluoridation and preterm birth, Massachusetts: 2009–2016. Matern Child Health J 23(4):451–458, PMID: 30542985, 10.1007/s10995-018-2659-y. [DOI] [PubMed] [Google Scholar]
  • 17.Kampouri M, Gustin K, Stråvik M, Barman M, Levi M, Daraki V, et al. 2022. Association of maternal urinary fluoride concentrations during pregnancy with size at birth and the potential mediation effect by maternal thyroid hormones: the Swedish NICE birth cohort. Environ Res 214(pt 4):114129, PMID: 35998692, 10.1016/j.envres.2022.114129. [DOI] [PubMed] [Google Scholar]
  • 18.Ortíz-García SG, Torres-Sánchez LE, Muñoz-Rocha TV, Mercado-García A, Peterson KE, Hu H, et al. 2022. Maternal urinary fluoride during pregnancy and birth weight and length: results from ELEMENT cohort study. Sci Total Environ 838(pt 3):156459, PMID: 35660617, 10.1016/j.scitotenv.2022.156459. [DOI] [PubMed] [Google Scholar]
  • 19.Hart R, Gray C, Lodise T, Patel N, Wymer S, McNutt LA. 2009. Relationship between municipal water fluoridation and preterm birth in upstate New York. https://apha.confex.com/apha/137am/webprogram/Paper197468.html [accessed 10 May 2024].
  • 20.Aschengrau A, Zierler S, Cohen A. 1993. Quality of community drinking water and the occurrence of late adverse pregnancy outcomes. Arch Environ Health 48(2):105–113, PMID: 8476301, 10.1080/00039896.1993.9938403. [DOI] [PubMed] [Google Scholar]
  • 21.Goodman C, Hall M, Green R, Hornung R, Martinez-Mier EA, Lanphear B, et al. 2022. Maternal fluoride exposure, fertility and birth outcomes: the MIREC cohort. Environ Adv 7:100135, PMID: 36644332, 10.1016/j.envadv.2021.100135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Singh N, Verma KG, Verma P, Sidhu GK, Sachdeva S. 2014. A comparative study of fluoride ingestion levels, serum thyroid hormone & TSH level derangements, dental fluorosis status among school children from endemic and non-endemic fluorosis areas. SpringerPlus 3(1):7, PMID: 24455464, 10.1186/2193-1801-3-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Xiang Q, Chen L, Liang Y, Wu M, Chen B. 2009. Fluoride and thyroid function in children in two villages in China. J Toxicol Environ Health Sci 1(3):54–59. [Google Scholar]
  • 24.Griebel-Thompson AK, Sands S, Chollet-Hinton L, Christifano D, Sullivan DK, Hull H, et al. 2023. A scoping review of iodine and fluoride in pregnancy in relation to maternal thyroid function and offspring neurodevelopment. Adv Nutr 14(2):317–338, PMID: 36796438, 10.1016/j.advnut.2023.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Hall M, Lanphear B, Chevrier J, Hornung R, Green R, Goodman C, et al. 2023. Fluoride exposure and hypothyroidism in a Canadian pregnancy cohort. Sci Total Environ 869:161149, PMID: 36764861, 10.1016/j.scitotenv.2022.161149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Zlatnik MG. 2016. Endocrine-disrupting chemicals and reproductive health. J Midwifery Womens Health 61(4):442–455, PMID: 27391253, 10.1111/jmwh.12500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Woodruff TJ. 2011. Bridging epidemiology and model organisms to increase understanding of endocrine disrupting chemicals and human health effects. J Steroid Biochem Mol Biol 127(1–2):108–117, PMID: 21112393, 10.1016/j.jsbmb.2010.11.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Goyal LD, Bakshi DK, Arora JK, Manchanda A, Singh P. 2020. Assessment of fluoride levels during pregnancy and its association with early adverse pregnancy outcomes. J Family Med Prim Care 9(6):2693–2698, PMID: 32984109, 10.4103/jfmpc.jfmpc_213_20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Sun L, Gao Y, Liu H, Zhang W, Ding Y, Li B, et al. 2013. An assessment of the relationship between excess fluoride intake from drinking water and essential hypertension in adults residing in fluoride endemic areas. Sci Total Environ 443:864–869, PMID: 23246666, 10.1016/j.scitotenv.2012.11.021. [DOI] [PubMed] [Google Scholar]
  • 30.Liu H, Gao Y, Sun L, Li M, Li B, Sun D. 2014. Assessment of relationship on excess fluoride intake from drinking water and carotid atherosclerosis development in adults in fluoride endemic areas, China. Int J Hyg Environ Health 217(2–3):413–420, PMID: 24012047, 10.1016/j.ijheh.2013.08.001. [DOI] [PubMed] [Google Scholar]
  • 31.Ma Y, Niu R, Sun Z, Wang J, Luo G, Zhang J, et al. 2012. Inflammatory responses induced by fluoride and arsenic at toxic concentration in rabbit aorta. Arch Toxicol 86(6):849–856, PMID: 22422340, 10.1007/s00204-012-0803-9. [DOI] [PubMed] [Google Scholar]
  • 32.Melila M, Rajendran R, Lumo AK, Arumugam G, Kpemissi M, Sadikou A, et al. 2019. Cardiovascular dysfunction and oxidative stress following human contamination by fluoride along with environmental xenobiotics (Cd & Pb) in the phosphate treatment area of Togo, West Africa. J Trace Elem Med Biol 56:13–20, PMID: 31442949, 10.1016/j.jtemb.2019.07.002. [DOI] [PubMed] [Google Scholar]
  • 33.Ferreira MKM, Aragão WAB, Bittencourt LO, Puty B, Dionizio A, Souza MPCde, et al. 2021. Fluoride exposure during pregnancy and lactation triggers oxidative stress and molecular changes in hippocampus of offspring rats. Ecotoxicol Environ Saf 208:111437, PMID: 33096359, 10.1016/j.ecoenv.2020.111437. [DOI] [PubMed] [Google Scholar]
  • 34.Dec K, ?ukomska A, Baranowska-Bosiacka I, Pilutin A, Maciejewska D, Skonieczna-?ydecka K, et al. 2018. Pre-and postnatal exposition to fluorides induce changes in rats liver morphology by impairment of antioxidant defense mechanisms and COX induction. Chemosphere 211:112–119, PMID: 30071422, 10.1016/j.chemosphere.2018.07.145. [DOI] [PubMed] [Google Scholar]
  • 35.Yan L, Liu S, Wang C, Wang F, Song Y, Yan N, et al. 2013. JNK and NADPH oxidase involved in fluoride-induced oxidative stress in BV-2 microglia cells. Mediators Inflamm 2013:e895975, PMID: 24072958, 10.1155/2013/895975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Menon R. 2014. Oxidative stress damage as a detrimental factor in preterm birth pathology. Front Immunol 5:567, PMID: 25429290, 10.3389/fimmu.2014.00567. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.State of California Assembly. AB 733 Assembly Bill–Chaptered Bill Text. http://www.leginfo.ca.gov/pub/95-96/bill/asm/ab_0701-0750/ab_733_bill_951010_chaptered.html [accessed 16 October 2023].
  • 38.Newbrun E. 2019. A history of water fluoridation in California: lessons learned. J California Dental Assoc 47(11):705–711, 10.1080/19424396.2019.12220848. [DOI] [Google Scholar]
  • 39.US Department of Health and Human Services Federal Panel on Community Water Fluoridation. 2015. U.S. Public health service recommendation for fluoride concentration in drinking water for the prevention of dental caries. Public Health Rep 130(4):318–331, 10.1177/003335491513000408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Pace C, Balazs C, Cushing L, Morello-Frosch R. 2019. UC Berkeley Water Equity Science Shop Community Water System Boundaries. https://drinkingwatertool.communitywatercenter.org/wp-content/uploads/2020/02/Interactive_CWS_geography_v1_Metadata_2.3.2020.pdf [accessed 10 May 2024].
  • 41.Talge NM, Mudd LM, Sikorskii A, Basso O. 2014. United States birth weight reference corrected for implausible gestational age estimates. Pediatrics 133(5):844–853, PMID: 24777216, 10.1542/peds.2013-3285. [DOI] [PubMed] [Google Scholar]
  • 42.California State Water Resources Control Board. Fluoridation by Public Water Systems. https://www.waterboards.ca.gov/drinking_water/certlic/drinkingwater/Fluoridation.html [accessed 26 October 2023].
  • 43.Abduweli Uyghurturk D, Goin DE, Martinez-Mier EA, Woodruff TJ, DenBesten PK. 2020. Maternal and fetal exposures to fluoride during mid-gestation among pregnant women in Northern California. Environ Health 19(1):38, PMID: 32248806, 10.1186/s12940-020-00581-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Mondal P, Chattopadhyay A. 2020. Environmental exposure of arsenic and fluoride and their combined toxicity: a recent update. J Appl Toxicol 40(5):552–566, PMID: 31867774, 10.1002/jat.3931. [DOI] [PubMed] [Google Scholar]
  • 45.California Water Resources Control Board. n.d. EDT Library and Water Quality Analyses Data and Download Page. https://www.waterboards.ca.gov/drinking_water/certlic/drinkingwater/EDTlibrary.html [accessed 26 October 2023].
  • 46.Sherris AR, Baiocchi M, Fendorf S, Luby SP, Yang W, Shaw GM. Nitrate in drinking water during pregnancy and spontaneous preterm birth: a retrospective within-mother analysis in California. Environ Health Perspect 129(5):57001, PMID: 33949893, 10.1289/EHP8205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Robins J. 1986. A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect. Mathematical Modelling 7(9–12):1393–1512, 10.1016/0270-0255(86)90088-6. [DOI] [Google Scholar]
  • 48.Snowden JM, Reid CE, Tager IB. 2015. Framing air pollution epidemiology in terms of population interventions, with applications to multi-pollutant modeling. Epidemiology 26(2):271–279, PMID: 25643106, 10.1097/EDE.0000000000000236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Snowden JM, Rose S, Mortimer KM. 2011. Implementation of G-computation on a simulated data set: demonstration of a causal inference technique. Am J Epidemiol 173(7):731–738, PMID: 21415029, 10.1093/aje/kwq472. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.National Oceanic and Atmospheric Administration National Centers for Environmental Information. 2024. County Mapping: Climate at a Glance. https://www.ncei.noaa.gov/access/monitoring/climate-at-a-glance/county/mapping [accessed 26 October 2023].
  • 51.Data Access–Urban Rural Classification Scheme for Counties. https://www.cdc.gov/nchs/data_access/urban_rural.htm [accessed 26 October 2023].
  • 52.Jha SK, Singh RK, Damodaran T, Mishra VK, Sharma DK, Rai D. 2013. Fluoride in groundwater: toxicological exposure and remedies. J Toxicol Environ Health B Crit Rev 16(1):52–66, PMID: 23573940, 10.1080/10937404.2013.769420. [DOI] [PubMed] [Google Scholar]
  • 53.Grimaldo M, Borja-Aburto VH, Ramírez AL, Ponce M, Rosas M, Díaz-Barriga F. 1995. Endemic fluorosis in San-Luis-Potosi, Mexico. 1. Identification of risk-factors associated with human exposure to fluoride. Environ Res 68(1):25–30, PMID: 7729383, 10.1006/enrs.1995.1004. [DOI] [PubMed] [Google Scholar]
  • 54.US Bureau of Labor Statistics. n.d. Local Area Unemployment Statistics. https://www.bls.gov/lau/ [accessed 26 October 2023].
  • 55.Robert Wood Johnson Foundation, University of Wisconsin Population Health Institute. n.d. County Health Rankings 2015: California. https://www.countyhealthrankings.org/sites/default/files/media/document/state/downloads/CHR2015_CA_0.pdf [accessed 10 May 2024].
  • 56.Green R, Rubenstein J, Popoli R, Capulong R, Till C. 2020. Sex-specific neurotoxic effects of early-life exposure to fluoride: a review of the epidemiologic and animal literature. Curr Epidemiol Rep 7(4):263–273, PMID: 33816056, 10.1007/s40471-020-00246-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Farmus L, Till C, Green R. 2021. Critical windows of fluoride neurotoxicity in Canadian children. Environ Res 200:115202, PMID: 34051202, 10.1016/j.envres.2021.111315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Kaufman JS, MacLehose RF. 2013. Which of these things is not like the others? Cancer 119(24):4216–4222, PMID: 24022386, 10.1002/cncr.28359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Basso O, Wilcox A. 2010. Mortality risk among preterm babies: immaturity vs. underlying pathology. Epidemiology 21(4):521–527, PMID: 20407380, 10.1097/EDE.0b013e3181debe5e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Currie J, Schwandt H. 2013. Within-mother analysis of seasonal patterns in health at birth. Proc Natl Acad Sci USA 110(30):12265–12270, PMID: 23836632, 10.1073/pnas.1307582110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Chauhan SP, Rice MM, Grobman WA, Bailit J, Reddy UM, Wapner RJ, et al. 2017. Neonatal morbidity of small- and large-for-gestational-age neonates born at term in uncomplicated pregnancies. Obstet Gynecol 130(3):511–519, PMID: 28796674, 10.1097/AOG.0000000000002199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Ju H, Chadha Y, Donovan T, O’Rourke P. 2009. Fetal macrosomia and pregnancy outcomes. Aust N Z J Obstet Gynaecol 49(5):504–509, PMID: 19780734, 10.1111/j.1479-828X.2009.01052.x. [DOI] [PubMed] [Google Scholar]
  • 63.Zhang X, Decker A, Platt RW, Kramer MS. 2008. How big is too big? The perinatal consequences of fetal macrosomia. Am J Obstet Gynecol 198(5):517.e1–517.e6, PMID: 18455528, 10.1016/j.ajog.2007.12.005. [DOI] [PubMed] [Google Scholar]
  • 64.Chiavaroli V, Derraik JGB, Hofman PL, Cutfield WS. 2016. Born large for gestational age: bigger is not always better. J Pediatr 170:307–311, PMID: 26707580, 10.1016/j.jpeds.2015.11.043. [DOI] [PubMed] [Google Scholar]
  • 65.Padmanabhan V, Cardoso RC, Puttabyatappa M. 2016. Developmental programming, a pathway to disease. Endocrinology 157(4):1328–1340, PMID: 26859334, 10.1210/en.2016-1003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Vohr BR, Boney CM. 2008. Gestational diabetes: the forerunner for the development of maternal and childhood obesity and metabolic syndrome? J Matern Fetal Neonatal Med 21(3):149–157, PMID: 18297569, 10.1080/14767050801929430. [DOI] [PubMed] [Google Scholar]
  • 67.Gregory CWE, Danielle ME. 2022. Trends and Characteristics in Gestational Diabetes: United States, 2016–2020. Hyattsville, MD: National Center for Health Statistics. 10.15620/cdc:118018. [DOI] [PubMed] [Google Scholar]
  • 68.Kheradpisheh Z, Mirzaei M, Mahvi AH, Mokhtari M, Azizi R, Fallahzadeh H, et al. 2018. Impact of drinking water fluoride on human thyroid hormones: a case-control study. Sci Rep 8(1):2674, PMID: 29422493, 10.1038/s41598-018-20696-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69.Duntas LH, Orgiazzi J, Brabant G. 2011. The interface between thyroid and diabetes mellitus. Clin Endocrinol (Oxf) 75(1):1–9, PMID: 21521298, 10.1111/j.1365-2265.2011.04029.x. [DOI] [PubMed] [Google Scholar]
  • 70.Sastry MG, Mohanty S, Varma A, Mishra AK, Rao P. 2011. Association of higher maternal serum fluoride with adverse fetal outcomes. Int J Med Public Health 1(2):13–17, 10.5530/ijmedph.2.2011.4. [DOI] [Google Scholar]
  • 71.Diouf M, Cisse D, Lo CMM, Ly M, Faye D, Ndiaye O. 2012. Femme enceinte vivant en zone de fluorose endémique au Sénégal et faible poids du nouveau-né à la naissance: étude cas–témoins. Rev Epidemiol Sante Publique 60(2):103–108, PMID: 22424749, 10.1016/j.respe.2011.09.009. [DOI] [PubMed] [Google Scholar]
  • 72.Lubin JH, Colt JS, Camann D, Davis S, Cerhan JR, Severson RK, et al. 2004. Epidemiologic evaluation of measurement data in the presence of detection limits. Environ Health Perspect 112(17):1691–1696, PMID: 15579415, 10.1289/ehp.7199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 73.Rosinger AY. 2022. Using water intake dietary recall data to provide a window into US water insecurity. J Nutr 152(5):1263–1273, PMID: 35102375, 10.1093/jn/nxac017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74.Till C, Green R, Grundy JG, Hornung R, Neufeld R, Martinez-Mier EA, et al. Community water fluoridation and urinary fluoride concentrations in a national sample of pregnant women in Canada. Environ Health Perspect 126(10):107001, PMID: 30392399, 10.1289/EHP3546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 75.Kabir H, Gupta AK, Tripathy S. 2020. Fluoride and human health: systematic appraisal of sources, exposures, metabolism, and toxicity. Crit Rev Environ Sci Tech 50(11):1116–1193, 10.1080/10643389.2019.1647028. [DOI] [Google Scholar]
  • 76.Thomas DB, Basu N, Martinez-Mier EA, Sánchez BN, Zhang Z, Liu Y, et al. 2016. Urinary and plasma fluoride levels in pregnant women from Mexico City. Environ Res 150:489–495, PMID: 27423051, 10.1016/j.envres.2016.06.046. [DOI] [PubMed] [Google Scholar]
  • 77.Freni SC. 1994. Exposure to high fluoride concentrations in drinking water is associated with decreased birth rates. J Toxicol Environ Health 42(1):109–121, PMID: 8169995, 10.1080/15287399409531866. [DOI] [PubMed] [Google Scholar]
  • 78.Ortiz-Pérez D, Rodríguez-Martínez M, Martínez F, Borja-Aburto VH, Castelo J, Grimaldo JI, et al. 2003. Fluoride-induced disruption of reproductive hormones in men. Environ Res 93(1):20–30, PMID: 12865044, 10.1016/s0013-9351(03)00059-8. [DOI] [PubMed] [Google Scholar]
  • 79.Zhang S, Niu Q, Gao H, Ma R, Lei R, Zhang C, et al. 2016. Excessive apoptosis and defective autophagy contribute to developmental testicular toxicity induced by fluoride. Environ Pollut 212:97–104, PMID: 26840522, 10.1016/j.envpol.2016.01.059. [DOI] [PubMed] [Google Scholar]
  • 80.Zhou Y, Qiu Y, He J, Chen X, Ding Y, Wang Y, et al. 2013. The toxicity mechanism of sodium fluoride on fertility in female rats. Food Chem Toxicol 62:566–572, PMID: 24071475, 10.1016/j.fct.2013.09.023. [DOI] [PubMed] [Google Scholar]
  • 81.Goin DE, Casey JA, Kioumourtzoglou MA, Cushing LJ, Morello-Frosch R. 2021. Environmental hazards, social inequality, and fetal loss: implications of live-birth bias for estimation of disparities in birth outcomes. Environ Epidemiol 5(2):e131, PMID: 33870007, 10.1097/EE9.0000000000000131. [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

ehp13732.smcontents.508.pdf (17.4KB, pdf)
ehp13732.s001.acco.pdf (699.4KB, pdf)
ehp13732.s002.codeanddata.acco.zip (23.5KB, zip)

Articles from Environmental Health Perspectives are provided here courtesy of National Institute of Environmental Health Sciences