Abstract Objective Primary bilateral macronodular adrenal hyperplasia (PBMAH) is a heterogeneous disease characterized by adrenal macronodules and variable levels of cortisol excess, with not clearly established clinical diagnostic criteria. It can be caused by ARMC5 germline pathogenic variants. In this study, we aimed to identify predictive criteria for ARMC5 variants. Methods We included 352 consecutive index patients from 12 European centers, sequenced for germline ARMC5 alteration. Clinical, biological and imaging data were collected retrospectively. Results 52 patients (14.8%) carried ARMC5 germline pathogenic variants and showed a more distinct phenotype than non-mutated patients for cortisol excess (24-h urinary free cortisol 2.32 vs 1.11-fold ULN, respectively, P < 0.001) and adrenal morphology (maximal adrenal diameter 104 vs 83 mm, respectively, P < 0.001) and were more often surgically or medically treated (67.9 vs 36.8%, respectively, P < 0.001). ARMC5-mutated patients showed a constant, bilateral adrenal involvement and at least a possible autonomous cortisol secretion (defined by a plasma cortisol after 1 mg dexamethasone suppression above 50 nmol/L), while these criteria were not systematic in WT patients (78.3%). The association of these two criteria holds a 100% sensitivity and a 100% negative predictive value for ARMC5 pathogenic variant. Conclusion We report the largest series of index patients investigated for ARMC5 and confirm that ARMC5 pathogenic variants are associated with a more severe phenotype in most cases. To minimize negative ARMC5 screening, genotyping should be limited to clear bilateral adrenal involvement and autonomous cortisol secretion, with an optimum sensitivity for routine clinical practice. These findings will also help to better define PBMAH diagnostic criteria.