Study Incapable Of Detecting IQ Loss From Fluoride.

Abstract

LETTER. FULL-TEXT.

The Broadbent et al.¹ article on fluoride and IQ has two serious weaknesses.

First, the question is not whether community water fluoridation (CWF) reduces IQ, but whether or not total fluoride intake reduces IQ. Broadbent et al. acknowledge that CWF likely represents less than 50% of total fluoride intake. Their study did not determine total fluoride exposure, nor did it simultaneously control for the three sources of fluoride intake: CWF, fluoride supplements, and fluoride toothpaste.

It is likely the Dunedin Cohort participants had a very narrow range of total fluoride exposure. This would severely limit the study’s statistical power. More importantly, their 99 participants who had no CWF were precisely the participants most likely to have taken fluoride supplements. Guha-Chowdhury et al. found that Dunedin area children without CWF who took fluoride supplements would receive greater total fluoride than CWF children.²

We calculated total fluoride intake for the CWF and non-CWF Dunedin Cohort participants using publicly available data (available as a supplement to our article at http://www.ajph.org). We estimated that lifetime CWF children had mean total fluoride intake of 0.7 milligram per day while non-CWF averaged 0.5 milligram per day. This small difference can explain why Broadbent et al. would not find a statistically significant difference in IQ, even if they had assessed total fluoride intake.

Second, although Broadbent et al. criticized previous studies for failing to control for 15 potential confounders, their study failed to control for 11 of these, including important factors with available data. This is problematic because the study’s non-CWF population came mainly from a single “satellite suburb”: Mosgiel, New Zealand.^3,4 This town used groundwater, whereas most of the CWF study population had surface water. Mosgiel’s water was among the most corrosive in New Zealand and dissolved high levels of copper from plumbing and potentially also lead.⁵ Mean blood lead measured in the Dunedin Cohort was 11.1 ?g/dl (SD ±4.91), sufficient to cause a loss of four IQ points, but was not considered in the Broadbent et al. study.^6,7 Mosgiel’s water also had high natural manganese levels, another suspected neurotoxin.^8,9

Data on the mothers’ IQ and rural versus urban is also available for the Dunedin Cohort, but the study did not control for them. Mosgiel is more rural than the fluoridated area, potentially resulting in lower IQ in its children and their mothers.¹⁰

All these confounders would bias results away from an effect of fluoride on lowering IQ.

Confounding and the lack of contrast in total fluoride exposure may explain why no difference in IQ was found.