| Sign In to gain access to subscriptions and/or personal tools. |
Bioavailability of Fluoride in Drinking Water: a Human Experimental StudyFaculty of Medical Sciences, University of Newcastle, Framlington Place, Newcastle upon Tyne, NE2 4BW, UK; Correspondence: * corresponding author, A.Maguire{at}ncl.ac.uk
It has been suggested that systemic fluoride absorption from drinking water may be influenced by the type of fluoride compound in the water and by water hardness. Using a human double-blind cross-over trial, we conducted this study to measure cmax, Tmax, and Area Under the Curve (AUC) for plasma F concentration against time, following the ingestion of naturally fluoridated hard and soft waters, artificially fluoridated hard and soft waters, and a reference water. Mean AUC over 0 to 8 hours was 1330, 1440, 1679, 1566, and 1328 ng F•min•mL–1 for naturally fluoridated soft, naturally fluoridated hard, artificially fluoridated soft, artificially fluoridated hard, and reference waters, respectively, with no statistically significant differences among waters for AUC, cmax, or Tmax. Any differences in fluoride bioavailability between drinking waters in which fluoride is present naturally or added artificially, or the waters are hard or soft, were small compared with large within- and between-subject variations in F absorption. Abbreviations used: F, fluoride; AUC, Area under the Curve for plasma F concentration against time; AUC(0-3), Area under the Curve for plasma F concentration against time for 0 to 3 hours following water ingestion; AUC(0-8), Area under the Curve for plasma F concentration against time for 0 to 8 hours following water ingestion; cmax, maximum plasma F concentration corrected for baseline plasma F and dose (i.e., F concentration of individual waters); Tmax, time of cmax.
Key Words: bioavailability • drinking water • fluoride • water hardness.
Fluoride occurs naturally in tap drinking-water supplies in the United Kingdom (UK), at concentrations up to 1.5 milligrams per liter. As a means of reducing the prevalence of dental caries, low F concentrations in drinking-water in selected areas of the UK may be increased to a target concentration of 1 milligram per liter by the addition of hexafluorosilicic acid or its sodium salt. However, most of the evidence relating to the long-term general health effects of F is based on studies of populations living in areas receiving F occurring naturally in water supplies. A recent systematic review on the health effects of F, conducted by the University of York Centre for Reviews and Dissemination (McDonagh et al., 2000), concluded that "the assessment of natural versus artificial water fluoridation effects is greatly limited due to lack of studies making a comparison". A subsequent MRC Working Group, established to determine what further research was required to improve knowledge on F and health, identified bioavailability of naturally fluoridated water vs. artificially fluoridated water and of hard vs. soft water, as important areas of uncertainty (Medical Research Council, 2002). Higher natural levels of F occur typically (but not exclusively) in hard water, while artificial fluoridation usually involves softer water sources. Water hardness is principally determined by the concentrations of calcium and magnesium ions in water and may be expressed as the equivalent concentration of calcium carbonate. Fluoride is present in water supplies as F ions, due to the almost complete dissociation of the parent F compounds, occurring either naturally (predominantly from minerals such as fluorspar [calcium fluoride]) or added as hexafluorosilicic acid or its sodium salt (Department of the Environment, 1987; Her Majesty’s Government, 1991). The undissociated proportion is between 10–18 and 10–30 of the dissociated F ions over the pH range of 6–9 usually found in water supplies, and when the impact of water hardness on dissociation is included, the proportion of free F ion ranges from 91% for water with very high hardness, to more than 99% for soft water (Jackson et al., 2002). It is thought that the absorption of soluble F from the gut into the bloodstream is essentially complete at approximately 95% (Whitford, 1996). However, it has been suggested that absorption may be influenced by the type of F compound in the water and by the water hardness, with the presence of calcium decreasing uptake of F from the gut into the bloodstream, perhaps through ion-pairing, although the importance of any such effect is unclear (Medical Research Council, 2002). A review of the chemistry and bioavailability aspects of F in drinking-water (Jackson et al., 2002), concluded that, "In terms of chemistry and bioavailability there is absolutely no difference between added and ‘natural’ fluoride", and that, "The effect of major cations—calcium and magnesium (hardness) and sodium—on the chemical speciation and hence bioavailability of fluoride is very small". This review was based on in vitro work and a small number of animal studies. With the exception of mineral water (Trautner and Einwag, 1986) and a dilute sodium fluoride solution (Ekstrand et al.,1994), there have been no human studies reported on the bioavailability of F from tap drinking-waters. The aim of this study was to compare the bioavailability of fluoride in natural and artificially fluoridated drinking-waters, and soft and hard drinking waters, using a human experimental model.
Subjects Ethical approval for the study was obtained from the Newcastle and North Tyneside Local Research Ethics Committee. Thirty-six healthy adults, aged between 20 and 35 yrs, volunteered to take part in the study. Twenty subjects with no history of metabolic disease or acid-base disturbance, and who were not receiving a therapeutic diet, were recruited, having given written informed consent. The subjects were provided with fluoride-free toothpaste and were asked to avoid using any significant fluoride products one week before, as well as during, the whole six-week experimental period. They were also asked to avoid drinking tea and beer and eating seafood during the washout and experimental periods and were provided with low-fluoride bottled waters (< 0.02 mg F/L) for drinking and cooking. For the 11 males and nine females who completed all aspects of the study, the mean (± SD) age, weight, height, and body mass index were 25.7 (± 4.5) yrs, 62.7 (± 12.1) kg, 163.1 (± 9.0) cm, and 23.3 (± 2.3) kg/m2, respectively, for females, and 25.4 (± 2.5) yrs, 75.5 (± 6.2) kg, 175.8 (± 6.3) cm, and 24.5 (± 2.4) kg/m2 for males.
Experimental Design
All test waters came from UK tap water supplies apart from NFS, which was an Italian mineral water (Acqua Panna, Tione Spring, Italy, bottled by Panna SpA, Italy).
Sample Collection During each session, each subject consumed a specified low-F drink, meal, and snack at 11.00, 12.00, and 15.00 hours, respectively (fluoride content per serving was 10 µg F, < 10 µg F, and < 1 µg F, respectively). The same food and drink items were consumed in each of the 5 sessions.
Analytical Procedure The validity and reliability of the methods used in this study have been tested and described elsewhere (Ekstrand, 1977; Zohouri and Rugg-Gunn, 1999). The Coefficient of Variation (CV%) for the "known addition technique" method used was 9.7%. Results of re-analysis of plasma samples were within 0.64 ng F/mL.
Pharmacokinetic Analysis We "dose-corrected" the AUCs for the F concentration by dividing the AUC value by the concentration of fluoride (mg/L) in the test or reference water, to derive baseline and dose-corrected AUCs for 0–3 and 0–8 hrs, AUC(0-3) and AUC(0-8). In addition, maximum plasma F concentration (cmax) and lag time to maximum F concentration (Tmax) were calculated for each subject for each session.
Statistical Analysis
For the total of 100 experimental sessions undertaken overall by the 20 subjects, the mean baseline fasting plasma concentration of fluoride was 19.76 ng•mL–1 (SD 3.23 ng•mL–1).
Pharmacokinetic variables were determined for each type of drinking water (Table 1
Analysis of covariance (Table 2  ) showed that, for Area Under the Curve for 0 to 3 hrs [AUC(0-3)], the mean difference between artificially and naturally fluoridated waters was 176.3 ng F•min•mL–1 (95% CI = –29.9, 382.5), and that between hard and soft waters was 67.5 ng F•min•mL–1 (95% CI = –138.8, 273.7).
For Area Under the Curve for 0 to 8 hrs [AUC(0-8)], the mean difference between artificially and naturally fluoridated waters was 237.3 ng F•min•mL–1 (95% CI = –49.7, 524.2), and between hard and soft waters, it was –1.0 ng F•min•mL–1 (95% CI = –288.0, 285.9). No statistically significant differences between waters for Tmax, cmax, or Area Under the Curve for the periods 0 to 8 and 0 to 3 hrs were detected by analysis of covariance. A noticeable feature observed in the data was the large variation in fluoride absorption both within and between subjects; there were large differences among subjects and, for individual subjects, there were large differences among the different types of water. For three-hour AUC, the estimated standard deviation of the between-subject random errors, based on the analysis of covariance model, was 166 ng F•min•mL–1; the estimated standard deviation of the within-subject random errors was 470 ng F•min•mL–1. The corresponding figures for eight-hour AUC were 207 and 655 ng F•min•mL–1. Across all 20 subjects, the observed differences between the two types of fluoride and between hard and soft water were not large when considered against these two sources of variation and were not statistically significant.
This study provides the first data on F pharmacokinetics and bioavailability of F from naturally and artificially fluoridated tap drinking-waters with different degrees of hardness. The results suggest that any differences between waters for these variables are small. The analytical technique used in this study to measure plasma F concentration has been reported to provide sensitive and reliable determination of F in small volumes of plasma samples (Ekstrand et al., 1977). Since plasma F concentrations are influenced by the relative rates of bone accretion and dissolution, we chose the age range of subjects (21 to 35 yrs) to ensure the inclusion of individuals whose skeletal mass was still increasing. At the same time, the relatively narrow age range selected minimized inter-person variability, thus maximizing the potential of the study to detect treatment-related differences. The results show similarity in time to reach maximum plasma F concentration for different types of water, which suggests that the rate of F absorption is independent of the type of water with regard to hardness (soft or hard) or fluoride (i.e., artificially or naturally fluoridated). The relatively wide variations in AUCs for plasma F may have arisen because the subjects were not prevented from moving around during their eight-hour experimental period once the cannula was in place, which may have reduced the rate and extent of gastric emptying, with a resultant impact on F absorption from the stomach. Fluoride absorption from the stomach occurs by simple diffusion of HF, with the rate of absorption dependent upon the concentration gradient of HF and inversely related to the pH of the stomach contents (Whitford, 1990). Absorption from the stomach may account for up to 50% of the amount ingested, with the remainder being absorbed from the upper small intestine (Whitford, 1996). The mean values for cmax and AUC for artificially fluoridated waters were slightly, but (given the wide 95% Confidence Intervals) not significantly, higher than those for naturally fluoridated waters. Any differences between F bioavailability of these waters are likely to be small. With regard to water hardness, although the observed mean AUC for the hard waters was slightly higher than that for the soft waters, any differences between F bioavailability of these waters are also small. This is consistent with results from a study involving administration of F simultaneously with a calcium-rich breakfast and calcium-poor lunch in which the calcium content of the food did not influence F bioavailability (Arnold et al., 1989), although other studies have shown the bioavailability of F to be affected when taken with various foods. Foods containing appreciable amounts of divalent or trivalent cations can lead to the formation of insoluble complexes and precipitates that reduce F absorption (Ekstrand and Ehrnebo, 1979; Spak et al., 1982; Whitford, 1996). Across all 20 subjects, there were two main sources of variation, with large differences in Area Under the Curve between subjects and, for individual subjects, large differences in Area under the Curve among the 5 different waters. The observed differences between the two types of fluoride and between hard and soft waters were not large when considered against these two sources of variation and were not statistically significant. In young or middle-aged adults, once F is absorbed from the gastro-intestinal tract, approximately 50% is excreted in the urine (World Health Organization, 1994), although this proportion can vary considerably, depending on numerous variables, including F intake, acid-base balance, and urinary pH (Ekstrand et al., 1978; Whitford, 1996). Some of the relatively wide variation in AUCs between subjects observed in this study might be explained by inter-individual differences in physiological variables such as volume and pH of gastric secretions, gastro-intestinal motility, plasma volume, and urinary pH. This study adds to the understanding of F pharmacokinetics of naturally fluoridated compared with artificially fluoridated waters, and hard compared with soft waters. The findings suggest that, measured by Area under the Curve, Tmax, and cmax following water ingestion, any differences in bioavailability of fluoride between drinking waters in which the fluoride is present naturally or added artificially, or the waters are hard or soft, are small compared with the large within- and between-subject variation in F absorption following ingestion of drinking waters with F concentrations close to 1 part per million.
The authors are most grateful to: Emeritus Professor A.J. Rugg-Gunn for his advice regarding the design of the study; Mr. A. Bruce, Water Quality Scientist at Northumbria Water, for his valuable help and advice; the Severn Trent, Northumbrian, and Anglian water companies for supplying the waters; the sedation nurses and Dr. N. Girdler at the School of Dental Sciences, Newcastle University, for arranging cannulation of subjects; Ms. E. Jackson for help in laboratory analyses of the samples; and the study volunteers for their participation and cooperation. This work was undertaken by the University of Newcastle (Faculty of Medical Sciences), which received funding from the Department of Health (# 0290049). The views expressed in the publication are those of the authors and not necessarily those of the Department of Health. Received for publication June 20, 2004. Revision received April 21, 2005. Accepted for publication September 15, 2005.
Journal of Dental Research, Vol. 84, No. 11,
989-993 (2005)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION