RETRACTED ARTICLE: Fluoride exposure and intelligence in school-age children: evidence from different windows of exposure susceptibility

Fluoride can prevent dental caries [1], and it is beneficial to bone metabolism as an essential trace element in the body [2]. However, the safe dose range of fluoride is limited [3]. Increasing evidences have shown that long-term exposure to excessive fluoride will not only increase the risk of dental fluorosis [4] and skeletal fluorosis [5], but also impair neural development by influencing gene and protein expression, enzyme activity and inducing oxidative stress [6].

The effects of fluoride on the nervous system have been manifested in intellectual development of rats [7]. Chronic high fluoride exposure of maternal mice in drinking water during pregnancy and lactation may have harmful influences on learning and memory of young mice [8]. Studies indicated that fluoride ingested by the mother can cross the placental barrier [9] and the blood-brain barrier [10], to affect the offspring’s cognitive function development [11]. Green et al. (2019) found that prenatal exposure to excessive fluoride was associated with children’s lower IQ scores. Surprisingly, the mean fluoride concentrations of drinking water in the fluoridated and non-fluoridated communities were 0.59 and 0.13 mg/L, respectively [12], which were lower than 1.5 mg/L suggested by the World Health Organization [3]. Elsewhere, a cross-sectional study in India showed that children’s intelligence quotient (IQ) was inversely associated with fluoride exposure via drinking water and maternal diet during pregnancy [13]. And, another Mexican study suggested that maternal exposure to excessive fluoride during pregnancy reduced children’s cognition and IQ scores, and the fluoride exposure range was 0.15–1.38 mg/L [14]. Some contradictory results have been found in the association between childhood fluoride exposure and intellectual development. Among such instances, a Chinese study found that exposure to moderately high fluoride in childhood could cause the loss of excellent intelligence [15]. A cohort study in Canada also reported that water fluoride concentration was associated with intelligence loss of formula-fed children, and the mean fluoride concentrations in the drinking water in the fluoridated and non-fluoridated communities appeared same as the study of Green et al. (2019) [16]. Meanwhile, no association between excessive fluoride and IQ scores were observed in a community water fluoridation program in New Zealand, where the fluoride concentration ranged from 0.7 to 1.0 mg/L [17]. Adding to the above, there was no significant correlation between fluoride exposure and lower IQ in adolescents [18]. Due to the different study populations and areas, the conclusion that excessive fluoride causes loss of children’s IQ still lacks strong evidence. Furthermore, compared with children without tobacco smoke exposure history, the relative risk of behavior problems of children exposed only in prenatal period was 90% higher, while that of children exposed only in postnatal period was 30% according to a prospective birth cohort study [19], indicating that the effects of prenatal and postnatal exposure on children’s neurodevelopment are different.

On the basis of the above analyses, it is an urgent need to distinguish the health effects of fluoride exposure in different susceptibility windows during development. Here, we selected Tongxu County, Kaifeng, Henan Province, a typical drinking-water-born fluorosis area, as the research region (no industrial sources of neurotoxic environmental pollutants such as arsenic, lead, mercury, etc.) to assess the association between excessive fluoride exposure in prenatal or childhood periods and children’s intelligence.

Our study examined the association between excessive fluoride exposure in prenatal and childhood periods and the intelligence of school-age children. We found that prenatal excessive fluoride exposure could cause lower IQ scores, especially the decreased odds of developing excellent intelligence. Meanwhile, a negative association between fluoride exposure and children’s IQ scores was observed in children without prenatal exposure.

Several epidemiological studies have shown that excessive prenatal fluoride exposure could reduce offspring’s IQ scores [12, 14], which are consistent with our findings. Fluoride ingested by the mother during pregnancy can cross the placental barrier and enter the embryo [24], and affect the brain development of children [25]. In addition, we found that the odds of developing excellent intelligence for children in PFG decreased by 51.1% compared with nonexposed children. Lower intelligence has become a new public health and social problem, which not only endangers children’s physical and mental health, but also increases family burden [26, 27]. Hence, it is necessary to consider limiting the intake of excessive fluoride by pregnant women.

For all participants, we observed that a decrease of 4.965 points in children’s IQ scores was associated with each 1.0 mg/L increase in the CUF concentration of ≥1.7 mg/L. A cross-sectional study found that moderate fluoride exposure was inversely correlated with children’s IQ scores, and suggested a UF threshold of 1.6 mg/L for intelligence damage [15]. The reason for the difference between the two thresholds may be that we adjusted the UF concentration by UCr. We further found that every 1.0 mg/L increment in the CUF concentration of ≥2.1 mg/L was related to a reduction of 11.4 points of children’s IQ scores in prenatal control group. These results suggested that excessive fluoride exposure in childhood was negatively associated with children’s IQ scores. Fluoride can penetrate into the brain through the blood-brain barrier [10]. Several researches reported that exposure to fluoride could damage the function of hippocampus and thyroid, thus impairing intellectual development of school children [28, 29]. Contrary, no findings were obtained in a New Zealand study to support that fluoride exposure in community water fluoridation could affect neurologic development or IQ of children aged 7–13 years old [17]. Also, a study in India showed that fluorosis was not significantly associated with adolescent intelligence in a non-endemic fluorosis area with lower level of fluoride [18]. The differences of exposure level, ethnicity and study areas may explain these results which are inconsistent with ours.

An animal experiment found that fluoride may elevate fluoride tolerance by down-regulating ribosomes, pancreatic secretion, steroid biosynthesis, glutathione metabolism, steroid biosynthesis, and glycerolipid metabolism gene expression [30]. Furthermore, another animal experiment suggested that the decrease of the susceptibility to fluoride was related to the changes in expression level of glutathione S-transferase gene after fluoride stimulation [31]. In the current population study, we found some similar results. Children with prenatal fluoride exposure have lower IQ scores, and therefore the reduction of IQ scores at higher CUF levels in some range during childhood did not become that evident. In addition, the IQ scores of children without prenatal exposure decreased evidently with the increase of CUF concentrations of ≥2.1 mg/L (Fig. 3). However, the mechanism still needs to be further studies at the population level.

Our study has several limitations. Firstly, urine samples were collected in the morning, not 24 h. However, a study has reported that for population studies, morning urine is a good representation of 24-h urine samples and the results are reliable [32]. Secondly, considering the influence of urine dilution on UF, UCr was used to adjust the UF. However, due to the instability of UCr in growing children or experiencing puberty, this may also be one of the limitations. Since the subjects in our study are all boarding students, their same diet and sleep schedule could alleviate the limitation to some extent. Thirdly, it was very difficult to obtain UF concentrations data from a pregnant mother, thus, we used the fluoride level in local drinking water during mother’s pregnancy. This is because fluoride level in drinking water is highly correlated with UF level [33]. On point, all the mothers in the study were local residents, and local drinking water fluoride level in the village could well represent maternal fluoride exposure level. Fourthly, the project was a cross-sectional study which lasted 1 month, and only one-time sampling was conducted, and so the inferential causality was weak. Long term large-scale epidemiological studies should be conducted to provide more abundant evidences. Finally, we tried to reduce the influence on the results through adjusting for children’s age, gender, gestational weeks, parental education levels and children’s BMI, though some unknown confounding factors may still.

In summary, prenatal and childhood exposure to excessive fluoride may impair the intelligence development of school children. Furthermore, children with prenatal fluoride exposure had lower IQ scores than children who were not prenatally exposed; therefore the reduction of IQ scores at higher levels of fluoride exposure in childhood does not become that evident. Our results support the necessary restriction of fluoride intake for pregnant women and school-age children to protect children’s intellectual development.