Abstract
To the Editor:
Maguire et al.(2004, 2005) investigated the bioavailability and absorption of fluoride from naturally and artificially fluoridated drinking water, looking also at the influence of water hardness.
There was a 16% greater increase in plasma fluoride (AUC at 8 hrs) when participants drank artificially fluoridated compared with naturally fluoridated water. The study, however, with 20 participants, had the power (retrospectively calculated), at the 5% level, only to detect a 33% difference in output of fluoride. The mean difference of plasma AUC at 8 hrs was 237 ng F.min/mL (95% CI: −49.7, 524.2), a standardized difference of 0.7. To have an 80% power to detect this as statistically significant at the 5% level would have needed a sample size of around 65 participants (85 participants for 90% power). The power of their study to test for any interaction effect with water hardness was even lower. A small difference in bioavailability would of course be a scientifically important discovery, given that the chemistry is thought to be identical, and thus it is surprising that such a small study was conducted.
While being potentially useful as a pilot to test the methods, the published study is grossly statistically underpowered, and potentially important findings using plasma Fp% (Maguire et al., 2004), another measure of relative bioavailability, were omitted. The study should not be used as the basis of confident or policy-useful judgments about the relative bioavailability of fluoride and interactions with the hardness of the water. Better and larger studies should be conducted.