Pitfalls for readers of “Special Issue on Pearls and Pitfalls in Headache Research”

Dear Editor,

The special issue of Cephalalgia on Pearls and Pitfalls in Headache Research is commendable. Its laudable goal of guiding the reader to understand limitations of a given study, however, demands that the editor is aware of the pitfalls awaiting readers of imprecise or biased review, not to mention disservice to the research, particularly because access to letters of clarification as I intend here is uncommon when reading review articles.

In his partial review of neuroimaging in headache (1) as a potentially important mechanism of explaining the progression of episodic to chronic migraine, May thoughtfully singled out our studies of iron accumulation over time in the central structure of the brain’s nociceptive system, the periaqueductal gray (PAG) (2). Yet he dismisses the import of this finding based on two subsequent papers from other researchers that, to use his words, “did not confirm migraine-specific significant changes in the PAG” (3,4). Indeed not, but only because they didn’t study it! In fact, Kruit et al. in their paper (incorrectly referenced to its editorial in the review) correctly pointed out they were unable to study the PAG because their manual selection of region of interest (ROI) would have included the hyperintense signal from cerebrospinal fluid (CSF) in the aqueduct, causing inaccurate iron measurement (3). Not credited by the reviewer, this pitfall was avoided in our own work by using automated Iterative Self-Organizing Data Analysis Technique (ISODATA) segmentation to accurately delineate and identify the entire area of the PAG excluding aqueduct CSF (3). Even with methods of high sensitivity to iron content, Tepper et al. did not study the PAG either, presumably because again the ROI was manually selected (4). Thus the statement that these studies failed to confirm PAG abnormality in migraine is unjustly misleading, opening the reviewer to the challenge of bias.

Even so, to avoid the same pitfall of misleading your readers I should emphasize that iron accumulation in the PAG still remains to be confirmed by replication with such methods as we used (2), requiring high magnetic fields in combination with specialized computer-based selection for sensitive measurements in discrete brainstem centers. So far this has not been accomplished.