Changes in diagnostic criteria, for example, the various International Classification of Headache Disorders criteria, would lead to changes in the outcomes of epidemiological studies. International Classification of Headache Disorders-1 was based mainly on expert opinion, yet most of the diagnostic criteria were reliable and valid, but it did not include chronic migraine. In its second version, the classification introduced chronic migraine, but this diagnosis resembled more a high-frequency migraine rather than the actual migraine transformation process. It also introduced medication overuse headache, but it necessitated analgesic withdrawal and subsequent headache improvement to be diagnosed as such. Hence patients having medication overuse headache could only be diagnosed in retrospect, which was an awkward situation. Such restrictive criteria for chronic migraine and medication overuse headache omitted a high proportion of patients. International Classification of Headache Disorders-3 allows a diagnosis of medication overuse headache due to combination analgesics if taken for at least 10 days per month for more than three months. Hence the prevalence rate of medication overuse headache and chronic migraine can increase compared to the previous version of the headache classification. Different criteria have been used across studies to identify chronic migraine and medication overuse headache, and therefore the information acquired from previous studies using earlier criteria becomes uncertain. Hence much epidemiological research would need to be interpreted cautiously or repeated with the most updated criteria, since the subjects in studies that apply the latest criteria may be phenotypically different from those in older studies.