Melin and Bondy Respond to “E Pluribus Unum for Epidemiology”

We very much appreciate the commentary by Drs. Wang and Lacey (1). Wang and Lacey describe the Genetic Epidemiology of Glioma International Consortium (GLIOGENE Consortium) and our latest study, the Glioma International Case-Control (GICC) Study (2), as a melting pot, unifying many different subject sources with a common questionnaire and specimen collection protocols.

We have a few additional suggestions that should be taken into consideration for future consortium studies of relatively rare diseases. One important issue raised is what to do if it is not possible to follow similar or identical protocols for case and control recruitment, given local variation in resources and infrastructure, as was the situation in the GICC Study. Although it may be possible to perform multisite international case-control studies with even more standardized protocols (as has been done in some clinical trials), such standardization will generally be associated with much higher costs than can be incurred in the current funding era. The compromise of allowing different source populations for cases (clinic vs. population-based) and controls (clinic vs. visitor vs. population-based) is not especially problematic, because we will be able to analyze the data to assess whether there is heterogeneity for each exposure by case and control type and by questionnaire administration method. These types of comparisons will be of methodological interest for the entire epidemiology community.

For future epidemiologic endeavors, studies of both rare and common diseases are also facing problems with cost-effective ways of conducting population-based research in light of the changes in telephone technology and distribution, which has rendered random-digit dialing a problematic approach for control selection. Web-based approaches are becoming increasingly popular for subject recruitment, but they may involve types of selection bias which are difficult to quantify.

A limitation of the GICC Study is that the field of molecular pathological classification of glioma has been progressing quickly. Thus, the new World Health Organization classification will require information on tumor molecular markers, which was not systematically collected in our study. It was not routine clinical practice to evaluate such markers when the study began in 2010. Additionally, recently published studies now show that it is possible to classify grade II–IV gliomas into relatively homogeneous etiological and prognostic groups according to molecular markers, such as 1p/19q co-deletion, isocitrate dehydrogenase gene (IDH), and telomerase reverse transcriptase gene (TERT) promoter mutations (3, 4). The GLIOGENE Consortium will be considering these markers in our future studies, as many research centers have the ability to collect tumor tissue for molecular analyses of these markers. Although epidemiologic studies of rare outcomes often suffer from a lack of precision, having a more open protocol allowed us accrue more patients for our study, providing greater statistical power and more reliable results than we would have had with a smaller set of cases and a stricter protocol.