We investigated the association between fluoride exposure and several sleep outcomes in a large population-based sample of older adolescents and adults, using Cycle 3 (2012–2013) of the Canadian Health Measures Survey. As of 2017, approximately 39% of Canadians had access to fluoridated water [3]; in our sample, approximately 32% lived in regions with community water fluoridation, 30% lived in regions without water fluoridation, and 38% lived in areas of mixed community water fluoridation or had missing data.

We found that for every 0.5 mg/L higher water fluoride concentration, there is a 34% increased relative risk of reporting sleeping less than the recommended amount. This result suggests that higher fluoride exposure is associated with sleep deficits in older adolescents and adults. The association between higher water fluoride concentration and sleeping less remained consistent if we restrict the sample to those aged 18 and older (RRR = 1.32). Using weighted data, we found an even stronger, though less precise, association between water fluoride concentration and sleeping less (RRR = 1.96). Among individuals aged 16 and older, we found that approximately 29% of individuals reported sleeping less than the recommended duration. This proportion of individuals reporting low habitual sleep durations is consistent with a larger sample of 10,976 adult respondents of the CHMS of which 32% reported sleeping less than the recommended duration [36]. Insufficient sleep is associated with various adverse outcomes, including changes in mood [37,38,39,40], cognition [37, 40,41,42], and reaction time [40, 43, 44], higher rates of motor vehicle accidents [45], hypertension, cardiovascular disease [46,47,48], and diabetes [47].

While the direction of causality cannot be discerned due to the cross-sectional nature of the survey, we can provide some hypotheses to explain the finding. Given the tendency of fluoride to accumulate in the pineal gland, the association between fluoride exposure and shorter sleep duration may be explained by an effect on melatonin production and subsequently the timing of sleep. In humans, the pineal gland is subject to calcification and forms concretions. This calcification varies between individuals, but generally increases with age [9, 49]. Sympathetic nervous system innervation forms the major communication between the pineal gland and the superior cervical ganglion [9, 50] and may also serve to maintain these concretions [49]. Additionally, these concretions may reflect prior biosynthetic metabolic activity of the gland, resulting from dark exposure, for example [49]. Fluoride accumulates in the pineal gland to a similar degree as in teeth [10] and the pineal gland in older individuals contains more fluoride than any other soft tissue [49]. This accumulation is likely due to fluoride’s high affinity for hydroxyapatite [12], as pineal fluoride concentration is directly correlated with pineal calcium concentration [10, 49], as well as the fact that it sits outside of the blood brain barrier, has a substantial blood supply [13], and may be ‘sampling’ the blood in circulation [49]. One study found that the association between pineal fluoride and calcium was very strong (r²?=?0.92) only for pineal glands with high levels of fluoride, which implies that high pineal fluoride is associated with increased calcification [12, 13]. The deposition of fluoride in calcified tissues, such as the pineal gland, bones, and teeth, may represent a defense mechanism against potential fluoride toxicity (in other tissues), which may begin in the prenatal period [13].

Fluoride deposition in the pineal gland and its calcification would most likely exert effects on sleep via changes to pinealocytes and subsequently melatonin output. The pineal gland is composed primarily of pinealocytes, which synthesize melatonin [13, 50]. Pinealocyte calcification is not directly correlated with decreased plasma melatonin; however, it is associated with a decreased number of pinealocytes [13]. Therefore, pinealocyte calcification likely has indirect effects on melatonin biosynthesis, as the degree of uncalcified pineal tissue is associated with plasma and saliva melatonin [13, 16, 17]. Calcification of the pineal gland may also lead to decreased levels of melatonin in the cerebrospinal fluid [13]. Fluoride has been shown to affect pinealocytes and melatonin in animal models; for example, adult rats showed a 73% increase in these cells after 8 weeks of a fluoride-free diet compared to rats consuming standard fluoridated food and water [14], and a doctoral dissertation showed that prepubescent gerbils receiving a diet high in fluoride had lower melatonin production than those receiving a low-fluoride diet [49]. To our knowledge, there are no studies in living humans on the direct effects of fluoride exposure on the pineal gland or melatonin production and secretion.

These changes in calcification and melatonin excretion are a credible means through which fluoride could impact sleep and circadian outcomes. Calcification of the pineal gland has been shown to be associated directly with sleep complaints, including daytime tiredness and sleep disturbance [19], as well as decreased REM sleep percentage, total sleep time, and sleep efficiency on polysomnographic measurement [18]. Higher uncalcified pineal volume has been associated with less sleep rhythm disturbance, even in otherwise ‘good sleepers’ [16]. Changes in melatonin onset time, peak time, and peak concentration have been associated with sleep disturbances, specifically sleep efficiency [51], and a delay in endogenous melatonin rhythms compared to normal individuals is what typically leads to the clinical manifestation of delayed sleep phase syndrome (DSPS). For individuals with DSPS, exogenous melatonin administration can be an effective means of ameliorating symptoms, by advancing sleep onset and wake times [52]. Alternatively, it is possible that fluoride could directly inhibit enzymes required for melatonin synthesis [13]. Together, these findings imply that a smaller or more calcified pineal gland and subsequent decreased melatonin output could be a risk factor for circadian instability and/or sleep and sleep rhythm disturbances.

Another possible mechanism through which fluoride could exert effects on the central nervous system, including sleep outcomes, is through oxidative stress. Fluoride can be a source of oxidative stress and may induce generation of reactive oxygen species (ROS) and lipid peroxidization [13, 53, 54]. Melatonin and other proteins from the pineal gland act as antioxidants [54]. Therefore, as the pineal gland calcifies and there is decreased melatonin output, there may be increased activity of these ROS and subsequent neural damage [13]. Further research should be conducted to elucidate this potential pathway. Alternatively, mechanisms outside of the pineal gland may be responsible for our findings; for example, hypothyroidism [55, 56] or hypertension [57, 58] could potentially influence the relationship between fluoride exposure and short sleep duration. These health conditions could possibly be on the causal pathway that mediates the association between fluoride exposure and sleep duration.

As far as we are aware, this is only the second human study investigating the effects of fluoride exposure on sleep outcomes, and the first study with an adult sample. Fluoride exposure has been associated with later bed and wake times in 16-to-19 year olds [20]; if participants went to bed later due to a delayed onset of nighttime sleepiness, but were constrained from a later wake time due to conflicting commitments (work, school, etc.), this situation could contribute to a shortened sleep duration as seen in the present study. However, given that our sample included mainly adults (and only adults when we restricted the sample to those aged 18 and older), this generalization from results found in older adolescents should be made cautiously. Additionally, the present study did not include a measure of sleep timing, so we were unable to test this hypothesis. Future studies should continue to investigate effects of fluoride exposure on sleep timing in adults.

Although we found an association between water fluoride and sleep duration, we did not find any associations between measures of fluoride exposure and our other sleep outcomes (frequency of daytime sleepiness or sleep problems). There are several potential reasons for these findings. It is possible that fluoride primarily leads to a disturbance in melatonin secretion in the form of a delay, as outlined above, or to a direct shortening of total sleep time, either of which could manifest as a shortened sleep duration but without accompanying daytime sleepiness or nighttime sleep disturbance. Alternatively, our measurements of sleep outcomes were rudimentary (comprised of a single question capturing each sleep outcome), and it is possible that they were not sensitive enough to capture various sleep pathologies. For example, the question assessing daytime sleepiness asks about difficulty staying awake; participants may have experienced more moderate daytime sleepiness, which would not necessarily have been captured in our data. Similarly, participants would likely not have been aware of disturbances in their sleep architecture (i.e. sleep problems), and our dataset would therefore not include this information. Future research should investigate the relationship between fluoride exposure and sleep using both objective (e.g. polysomnography, actigraphy, dim light melatonin onset) and validated subjective measures (e.g. Pittsburgh Sleep Quality Index [59], Stanford Sleepiness Scale [60]) to better characterize the nature of any sleep and circadian disturbances.

One limitation of this study is that we were unable to account for the presence of sleep disorders or other behaviors that could impact sleep (e.g. exercise, diet, use of melatonin supplements, sleeping pills, or medications). Future studies should investigate the relationship between fluoride exposure, sleep, and these factors; the relationship with physical activity is of particular interest given that a pilot study found that vigorous exercise in adults is associated with a reduction in renal fluoride clearance and an increased trend in plasma fluoride concentration [61]. This could potentially lead to greater uptake by body tissue [62], though this study of mice found no link between chronic exercise and bone fluoride levels. While we did not account for physical activity in our study, we did include BMI as a covariate, which is associated with physical activity levels [63]. Additionally, because our study is cross-sectional, the direction of the association between fluoride exposure and sleep outcomes is not established. Given these limitations and the limited evidence on the topic to date, the results of our study should be viewed as hypothesis-generating.

A strength of our study is that we included two measurements of fluoride exposure: tap water fluoride and specific-gravity adjusted urinary fluoride levels. Our results demonstrate that while water fluoride level was significantly associated with a lower than recommended sleep duration, UF_SG was not. This finding is consistent with two other population-based fluoride studies [20, 24], both of which found that water fluoride levels but not fluoride biomarkers (i.e. plasma, urine) were associated with adverse health outcomes. Consistent with their explanation, it is likely that water fluoride measurements act as a proxy for chronic or historical fluoride exposure (assuming the individual’s residence in a community with water fluoridation remains stable), whereas single urinary spot samples represent short-term (contemporaneous) exposure that may fluctuate between subsequent measurements [20]. In terms of risk to pineal gland calcification, fluoride may need time to accumulate in the pineal gland before exerting an impact on sleep. Although it is known that pineal gland calcification increases with age [9, 49], it is not clear whether lifetime exposure drives this relationship. One report [49] suggested that pineal fluoride does not reflect cumulative fluoride exposure of the individual due to a lack of correlation between fluoride in the pineal gland and in bone ash. However, this finding was potentially confounded by other methodological factors, such as the inclusion of bone samples from different types of bone, which have been shown to contain differing levels of fluoride [9, 10, 49]. Moreover, we would not necessarily expect pineal and bone fluoride levels to be correlated given that fluoride is reversibly bound to bone and can be re-released into plasma [64]. Prospective studies are needed to investigate the potential impacts of cumulative fluoride exposure across varying stages of development on sleep outcomes.

Higher water fluoride concentration was significantly associated with increased risk of reporting fewer than the recommended hours of sleep. This finding suggests that fluoride exposure may contribute to clinically meaningful reductions in sleep duration among individuals living in areas with optimal water fluoridation. These findings should be interpreted in the context of the demonstrated benefits of fluoride exposure for dental health [1, 4, 5, 8, 65].

The data from the Canadian Health Measures Survey (CHMS) Cycle 3 (2012–2013) analyzed during the current study are publicly available here: https://www.statcan.gc.ca/eng/statistical-programs/document/5071_D5_T9_V1