Background: Although adrenal vein sampling (AVS) is the gold standard for subtyping primary aldosteronism (PA), consensus regarding the optimal protocol and data interpretation is lacking. Advantages of cosyntropin stimulation during AVS include enhancing the rate and confidence of successful cannulation, and circumventing the intra-procedure hormonal fluctuations. While it has been suggested that cosyntropin might stimulate aldosterone secretion from non-autonomous areas, aldosterone-producing adenomas are also responsive to cosyntropin. The relative effect of cosyntropin on autonomous vs. normal aldosterone-producing cells remains poorly understood. Objectives: To define the major patterns of time-dependent AVS lateralization response to cosyntropin and the determinants of these patterns. Methods: We retrospectively studied all patients with confirmed PA who underwent AVS in our institution between 2009-2018. Simultaneous adrenal vein (AV) and peripheral samples were obtained before, 10 and 20 minutes after cosyntropin administration. Only patients who had successful catheterization on all 3 time points were included. Unilateral (U) or bilateral (B) PA were defined based on a lateralization index (LI) ≥ or < 4, respectively. Demographic, laboratory, imaging, and pathological data were reviewed. Aldosterone synthase (CYP11B2)-guided genetic analysis was performed on available adrenal tissue by next generation sequencing. Linear mixed effects model was used to identify factors influencing LI patterns across the three time points. Results: Study criteria were met in 220 patients. Based on pre-and post-cosyntropin data, AVS lateralization was concordant in 163 (74%) patients (UU, 104; BB, 59) and discordant in 57 (26%) patients (UB, 35; BU, 22). Baseline and stimulated LI in the UU group were significantly higher than in the UB and BU group, respectively (~2.3 fold, p<0.001). Three major patterns of LI response to cosytropin were identified: increasing, decreasing, and stable. Baseline peripheral and dominant AV aldosterone concentrations were strong predictors for LI pattern across the 3 time points (p=0.03 and p<0.001, respectively) while cortisol, age, sex or race did not predict LI response to cosyntropin. The mutation status has been analyzed in 50 patients who underwent adrenalectomy and revealed: 21 KCNJ5, 17 CACNA1D, 8 ATP1A1 and 4 ATP2B3 somatic mutations. For tumors harboring KCNJ5 mutations, LI was highest at baseline compared with all other groups (35.5 vs.12.9; p<0.001), and LI declined across the three time points studied. In contrast, for tumors with ATP1A1 and ATP2B3 mutations, LI increased with cosyntropin stimulation. Conclusions: The cosyntropin impact on aldosterone secretion and LI varies in PA patients. Predictors of LI patterns of response to cosyntropin include baseline aldosterone concentrations and aldosterone-driver mutations.