‘Not biologically plausible’: questions about survey data earn fluorosis paper a flag

A recent article that offered a stark warning about the risks to children of fluoride in the nation’s water has been tagged with an expression of concern after the publication of a new paper which undermines the reliability of the original data. 

The article, “Dental fluorosis trends in US oral health surveys: 1986 to 2012,” appeared in March in JDR Clinical & Translational Research, a dental journal. The first author on the paper is Christopher Neurath, of the American Environmental Health Studies Project, which advocates against fluoridation of water.  

The article, which used data from the National Health and Nutrition Examination Survey (NHANES) for the years 2011 and 2012, reported “large increases in severity and prevalence” of fluorosis over that period — continuing a trend dating back to the mid-1980s.  

According to the researchers: 

The results of this study greatly increase the evidence base indicating that objectionable dental fluorosis has increased in the United States. Dental fluorosis is an undesirable side effect of too much fluoride ingestion during the early years of life. Policy makers and professionals can use the presented evidence to weigh the risks and benefits of water fluoridation and early exposure to fluoridated toothpaste.

But the expression of concern casts those dire conclusions in doubt:

In March 2019, the JDR CTR published the report “Dental Fluorosis Trends in US Oral Health Surveys: 1986 to 2012” (Neurath et al. 2019), in which the National Health and Nutrition Examination Survey (NHANES) data from the Centers for Disease Control and Prevention (CDC) were analyzed.

The following month, the CDC published “Data Quality Evaluation of the Dental Fluorosis Clinical Assessment Data from the National Health and Nutrition Examination Survey, 1999–2004 and 2011–2016” (US Department of Health and Human Services 2019). In its report, the CDC assessed the quality of those NHANES data and questioned their validity. The CDC reported a similar increase in dental fluorosis prevalence between 2001–2004 and 2011–2014 as that reported in Neurath et al. (2019). Because the NHANES is not longitudinal, the CDC authors constructed a synthetic cohort and concluded that this increase was not biologically plausible. In an effort to interpret this finding, the CDC examined other sources. Specifically, the CDC tested the variability of fluorosis measurements in a small population sample and found that examiner agreement was problematic. This suggested that there may have been changes in the way examiners evaluated fluorosis, resulting in intra- and interexaminer errors. 

Furthermore, they suggested that possible oversampling of certain populations was a plausible reason for the fluorosis increase. In addition, the CDC recognized that impact of geographical variables associated with exposure to water fluoridation at the time of tooth development was not available. Therefore, the CDC suggested that the above limitations, especially related to data quality assessment, should be considered for prevalence studies.

The JDR CTR is publishing this Editorial Expression of Concern to alert our readers to this issue. We urge readers to review the CDC report (US Department of Health and Human Services 2019), in conjunction with Neurath et al. (2019), for a fuller understanding of the NHANES data.

Jocelyne Feine, editor of the JDR CTR, told us that: 

Because the CDC authors believed that the results were not biologically plausible, they identified some potential reasons that errors in the NHANES data set could have arisen. However, the CDC report fell short of identifying a specific problem to explain the findings. Given that, I determined that an expression of concern was the appropriate response to raise awareness of the concerns of the CDC and to let readers consider all of the information for themselves.

Neurath told us that: 

I am glad [Feine] makes clear in her email that the EoC was not issued because of any question about the integrity of our work and was only issued because CDC raised concerns about their public data.

However, he said: 

I believe the most likely explanation is a data processing or data management error at NHANES, but from the NHANES fluorosis data alone, it is impossible to determine which cycles are incorrect and which (if any) are correct.  That is why we are urging use of the fluorosis camera images to resolve this question.

Ryne Paulose-Ram, PhD, the associate director for science in the Division of Health and Nutrition Examination Statistics at the National Center for Health Statistics, who helped write the CDC paper, said she and her colleagues “can’t really explain” why the numbers don’t add up. 

The only thing we can really see is that it’s unlikely to be the result of a biological phenomenon. 

According to Paulose-Ram, since fluorosis develops before teeth erupt from the gum, the prevalence in erupted teeth for the same birth cohort should not change over time. But the CDC report appeared to show precisely that: an increase in the prevalence of fluorosis in the same age cohort over time. 

And because: 

Everything seems to be okay when we look at the data … We think that there probably was a shift in the way that the dentists were examining [fluorosis] over time. 

Part of the problem, Paulose-Ram added, may be the tool dentists currently use to assess fluorosis: 

Newer methods being investigated — such as digital imaging — but they are not ready for prime time.

Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up for an email every time there’s a new post (look for the “follow” button at the lower right part of your screen), or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].