April 10, 2012

dental x-rays and risk of meningioma

This analysis accompanied a roundup which can be viewed here.

Title, Date of Publication & Journal

Dental X-Rays and Risk of Meningioma, Tue 10 April 2012, Cancer

Claim supported by evidence?

This paper lends broad support to the claim that exposure to dental x-rays may be associated with increased risk of intracranial meningioma, though does not use the necessary study design nor provide a sufficient depth of analysis and discussion to infer a causal link between the two. 

While there is a generally consistent finding relating increased risk to higher frequency of exposure at younger ages, the analyses do not account for the changes in X-ray dose guidelines over the years to modify this association. Given the developments in X-ray practices over the past 60 years that the authors allude to, it would be useful from a public health perspective to know whether the association is consistent across all time periods or is only limited to past practices.

Summary 

• This large population-based study allows for generalizable and precise estimation of the association between dental X-rays and meningioma, though it is difficult to rule out the possibility of bias (see below)
• Despite the high number of statistical tests performed (which increases the chance of false positive results), the magnitude and direction of associations is generally consistent
• Although the authors cite previous validation efforts claiming that biased recall of X-rays visits is minimal, this arguably remains a limitation of any case-control study like this one, given the tendency of people to overestimate previous exposure to explain their disease
• The discussion does not go into sufficient depth to explain and interpret the results by, for example, contextualising them within X-ray practices over the last 60 years. The authors suggest an ‘apparent association’, and their recommendation to limit X-ray exposure infers causality, though they do not explicitly make this claim.

Conclusions

The authors are justified in suggesting an apparent association between X-ray exposure and the risk of meningioma. Their conclusions are generally supported by their results and analyses, but given the large number of tests performed, the potential for negative findings to be omitted, and the inherent biases associated with retrospective case-control studies like this, these results should be regarded as exploratory and ‘hypothesis generating’ rather than confirmatory and ‘hypothesis proving’.

Strengths/Limitations

Study design and analysis:

• (+) Case-control design is appropriate for this type of research question, given the relatively low incidence of meningioma in the population
• (+)The analytical method used in this study was appropriate for these data
• (+) Cases were frequency matched to controls by age, sex and state of residence
• (+) Results were adjusted for age, sex, race, education and history of head CT scans, but
• (-) they do not account for changes to X-ray doses over time that may have altered the observed associations. It may be unreasonable to assume the effect of exposure in one period is the same as the effect of exposure in another, i.e. that there is no ‘period effect’ (see glossary)
• (-) Case control studies like this can yield biased estimates of association due to (1) recall bias, (2) selection bias and (3) inability to control for all known and unknown confounders in the analysis
• (-) A large number of statistical tests were conducted, increasing the chance of false positive results
• (-) There was no analysis of the potential for non-participation to influence results. This may have introduced selection bias given the smaller proportion of controls who decided to participate (52%) than cases (65%); and the possibility that people’s exposure could influence their decision to participate 

Discussion/Interpretation of results:

• (+) The conclusions are in line with the results and do not exaggerate findings, and findings generally agree with previous studies
• (+) The authors are upfront about acknowledging the issue of recall bias as a limitation
• (-) The discussion of results does not address:
  • Why bitewing and Panorex, but not full mouth series, was associated with increased risk, given previous studies have found an association with full mouth series
  • Whether the observed associations were consistent across time periods
  • The WHO estimated lag time of several decades between exposure to ionizing radiation and meningioma disease – in other words, no distinction was made between recent exposure and exposure longer ago

Glossary

• Case control study: a very practical design, but the evidence is lower than for most other study types (e.g. randomised trials or prospective cohort studies)
• Recall bias: When the recollection of past exposure is different between cases and controls; typically when cases remember, and even exaggerate their recollection of, exposures more than controls
• Selection bias: When the cases and/or controls are selected differentially on the basis of their exposure status, such that cases aren’t representative of all cases in the population and controls aren’t representative of the population that produced the cases
• Period effect: When changes occur through time, such as the introduction of safer imaging technologies or medical treatments, that render comparisons across periods inappropriate without some form of adjustment

Any specific expertise relevant to studied paper (beyond statistical)?

No.

‘Before the headlines’ is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research.  A list of contributors, including affiliations, is available here.