Salt intake is highly associated with cardiac structure in patients with primary aldosteronism (PA). We investigated the association among dietary salt intake, aldosterone and left ventricular mass in patients with PA. We enrolled 158 patients with PA and 158 patients with essential hypertension. We measured 24-hour urinary sodium (UNa) and aldosterone (UAldo) level and echocardiography parameters. In patients with PA, the UAldo level was positively correlated with left ventricular mass index (LVMI; r=0.231, p=0.007). The UNa level was not linearly correlated with left ventricular structural parameters in patients with PA. To test if UNa has a non-linear relationship with LVMI among patients with PA, we categorized the participants according to the tertile of UNa (low, median, and high tertile). PA patients with medium tertile of UNa had significant lower LVMI than the other two groups (LVMI: 144.1 ± 42.9, 121.1 ± 33.4, and 136.7 ± 32.8 g/m2, from the lowest to the highest tertile of Una; analysis of variance p=0.006, post-hoc p <0.05). Multifactor analysis of variance confirmed this finding after adjustment for clinical parameters. Post-hoc analyses revealed that the high UNa tertile was associated with higher left ventricular end-diastolic volume compared with medium UNa tertile; while the low UNa tertile was associated with higher mean wall thickness compared with medium UNa tertile. The findings imply the reasons for increased LVMI may be different in patients with the highest and lowest UNa tertile. In conclusion, the medium tertile of 24-hour UNa is associated with lowest LVMI in patients with PA.