Author response: Body composition status and the risk of migraine: A meta-analysis.

We thank Dr. Gupta for the interest in our article,1 in which we evaluated the relationship between body composition and migraine. The commentary raised an interesting point about biological plausibility and how observed association between low body mass index (BMI) and migraine might be attributable to underlying illness such as thyrotoxicosis. Biologically, it is well-known that there are several potential mechanisms for the BMI–migraine association.2 Adipose tissue is a neuroendocrine organ. Like the thyroid gland, too much or too little can both be associated with medical symptoms or disorders. Adipose tissue is centrally regulated by the hypothalamus and its connections, and peripherally participates in modulation of hormones, immune cells, and inflammatory-related proteins (e.g., adipocytokines).2,3 With adipose tissue expansion, or reduction, changes in macrophage recruitment, receptor expression, and the secretion of cytokines and adipokines occur.2,4 As such, it is not surprising that low BMI is associated with migraine. Is it possible that undetected illness could lead to low BMI? Yes, it is possible. However, as noted in our article, most of the studies included for meta-analysis are population-based studies.1 It is highly unlikely these rare complications can introduce bias in population-based studies. Finally, we would like to clarify that our meta-analysis was not a data-driven exercise. Rather, it was guided by a fundamental biological question with priori set procedures and consistent operational definitions of exposure and outcome variables. The details are summarized in the Methods of our article.1