AbstractThe nature and magnitude of nocebo responses in primary headache disorders are still unknown. To assess the distribution and possible predictors of nocebo responses in primary headache treatments, databases, including PubMed, EMBASE, and Cochrane Library were searched from 1988 to December 31, 2020, for parallel‐group, double‐blind, randomized placebo‐controlled trials of pharmacologic treatments of primary headaches. The nocebo responses were calculated using a random effects meta‐analysis model. Subgroup and metaregression analyses were performed to determine the associations of study design and demographic characteristics with nocebo responses. A total of 178 randomized controlled trials that satisfied the inclusion criteria were included. Prophylactic treatments elicited stronger nocebo responses than acute treatments. The majority of nocebo adverse events were mild to moderate in severity, with the nervous and digestive systems being the most commonly affected. There was a strong correlation between the active medication and control groups in terms of adverse events, both quantitatively and qualitatively. Long treatment duration, a high proportion of subjects receiving active medications, multicenter design, North America, high body mass index, women, previous treatment experiences, and a high proportion of patients with migraine headache with aura were all found to be significant positive predictors of nocebo responses, whereas the year of publication was found to be inversely related to them. Nocebo effects should be noticed for their contribution to discontinuation of or lack of adherence to active treatments. Clarifying these nocebo‐related risk factors can aid in their clinical prevention and management.