It was with great dismay that I read the article by Claus et al regarding dental x-rays and the risk of meningioma.1 This study has a number of issues that were not adequately addressed and that may have compromised the validity of its findings.
There was no relation between dental x-rays and the tumor location (supratentorial vs infratentorial), despite the fact that there should be significant differences in radiation doses between these brain regions. This discrepancy suggests that radiation dose was not the primary determinant of tumor incidence.
There was an increased risk of meningioma found in all age categories regardless of how distant the time of exposure was. Even the category of patients aged < 10 years demonstrated an elevated risk, despite the fact that the typical participant (median age, 57 years) was being asked to recall the frequency of dental x-rays from 50 years ago. The notion that the frequency of childhood dental x-rays could be accurately self-reported 50 years later stretches credulity, and suggests that accuracy of recall is suspect.
There also was shown to be an increased risk associated with therapeutic radiation exposures to the head, including treatments for a variety of benign conditions such as tonsillitis, acne, ringworm, and ear infections. In a brain tumor study, Kleinerman et al2 found that cases were more likely than controls to report the use of appliances near their heads, including electric razors (among men) and hair dryers (among women). These researchers suggested that patients with brain tumors had concerns about their earlier head exposures that were contributing to recall bias. In addition, a study of radiation recall bias that compared self-reported questionnaires with medical reports for patients with thyroid cancer and controls indicated that the patients with thyroid cancer were nearly twice as likely as controls to report exposures even though the medical records demonstrated the exposures to be comparable.3
There was no internal assessment of the potential for recall bias. For example, questions regarding head exposures (such as cell phone use, hair dyes, head trauma, electric razors, hair dryers, etc) would have allowed for an estimation of the magnitude of recall bias related to head exposures. In addition, questions concerning radiation exposures to distant parts of the body (eg, knee or ankle x-rays) would have determined whether the cases were more aware of radiation exposures in general or just to exposures to their heads. The fact that cases and controls reported equal amounts of orthodontics and endodontics is not relevant to the issue of differential recall because everyone accurately remembers whether they have ever worn braces or had a root canal, and the determinants of braces and root canals (ie, crooked and rotten teeth) are not related to radiation exposure.
This case-control study of self-reported radiation exposure, with no dosimetry, contradicts larger and stronger cohort studies with very accurate dosimetry data. Several National Academy of Sciences Biological Effects of Ionizing Radiation (BEIR) committees have reviewed this enormous amount of data and concluded that, even under the highly conservative “linear no-threshold” model of risk assessment, the cancer risk levels for ionizing radiation are much smaller than what this study suggests.4
In short, the study by Claus et al1 had a serious threat of recall bias, which was not adequately considered. This resulted in a finding that cannot be explained without putting aside the findings of other studies with superior designs. It is true that there are still uncertainties related to the magnitude of cancer risk at low doses of radiation. However, it is virtually certain that radiation-induced meningioma risks cannot possibly be as high as suggested by this article.