Abstract Background and aims Eosinophilia-associated coronary vasospasm is an under-recognized and scarcely reported life-threatening condition for whose management is undefined. Preliminary data suggest that eosinophil-targeted therapies could be beneficial. Our aim is to provide an in-depth report of the clinical picture, outcomes and predictors of relapse in patients with eosinophilia-associated coronary vasospasm. Methods European multicentric retrospective study, gathering twenty centers across France, Italy and Spain. Patients of 15 years or more with definite or suspected coronary vasospasm according to the 2017 International standardization of diagnostic criteria for vasospastic and concomitant (± 48 hours) eosinophilia (threshold: 0.5 x 109/L) were included in this study. Outcomes included vasospasm relapse and other major cardiovascular events (MACE) during follow-up. Predictors of relapse were assessed using bidirectional stepwise regression analysis. Results Thirty-seven patients (median age: 52 [42-64] years, 43% women) were included, among which 16 (43%) had aspirin-exacerbated respiratory disease, 16 (43%) and 13 (35%) fulfilled classification criteria for Eosinophilic Granulomatosis with Polyangiitis (all with negative antineutrophil cytoplasmic antibodies) and Hypereosinophilic syndrome, respectively. There was no predominant coronary territory involved. A iatrogenic trigger was evidenced in 7 (19%) patients, consisting in all cases but one of non-steroidal anti-inflammatory drugs. After a median follow-up of 32 [12-63] months, 20 (54%) patients relapsed (median time to first relapse: 10 [5−29] months), despite vasodilators regimen in 95% of cases. Besides coronary vasospasm(s), 13 (35%) patients experienced an additional MACE. In multivariate analysis, persistent eosinophilia (i.e. absolute eosinophil count > 0.5 x 109/L at the time of coronary vasospasm relapse or at last follow-up for relapsing and non-relapsing patients respectively) was the sole indepedendent predictor of relapse (Odds Ratio 8.87 [3.28-24.82]; p<0.001). All six patients treated with interleukin-5 biologics remained relapse-free under treatment. Conclusions Eosinophilia-associated coronary vasospasm can occur in patients with type 2 inflammation. Long-term normalization of absolute eosinophil counts appears crucial in preventing the recurrence of eosinophilia-associated coronary vasospasm, while conventional use of vasodilators seems ineficient.