1Introduction

The treatment of childhood cancer has improved substantially in recent decades [12] . Accordingly, increasing numbers of childhood cancer patients experience remission and cure. Over the past decades, however, it has become clear that childhood cancer survivors may experience, to varying degrees, a wide range of chronic health conditions resulting from previous therapeutic exposures [3] , one of which is migraine headaches [45] .

Migraine is a common and frequently disabling neurological disorder characterized by severe and recurring unilateral headaches and accompanying symptoms, known as migraine attacks [6] . Although data are still scarce, migraine headaches may be among the most common complications experienced by childhood cancer survivors. Indeed, a migraine prevalence of more than 30% among childhood cancer survivors has been reported in two previously published investigations [78] .

Until now, the assessment of migraine risk in childhood cancer survivors has predominantly relied on either self-reported information or been investigated by means of in-patient hospital discharge diagnoses, [910] and either approach can be encumbered by its own set of limitations. Studies relying on self-reported information are typically vulnerable to both participation and recall bias, and potentially to misclassification [11] . Studies based on hospital discharge diagnoses may suffer from surveillance bias as well as low sensitivity to detect migraine in the population since hospitalization due to migraine is rare [12] .

Effective acute treatment for migraine attacks is crucial to alleviate the significant personal and socio-economic burden imposed by this disorder [13] . In recent times, the acute treatment of migraine has predominantly been based on potent serotonin 5-HT1B/1D-receptor agonists, collectively termed triptans [1415] . In Denmark, triptans have mostly replaced older drug therapies for acute migraine treatment comprised of dihydroergotamines and ergotamines [16] , and can only be purchased via prescription from a doctor [17] .

Since 1994, all prescriptions dispensed by non-hospital pharmacies in Denmark have been recorded in the Danish Prescription Register [18] . In combination with other Danish health registers [1920] , this data resource enabled us to conduct a population-based study in which in-patient discharge diagnoses of migraine were supplemented by prescriptions of antimigraine medications to assess the relative risk of migraine of varying presentations among individuals diagnosed with childhood cancer.