Increased cardiac left ventricular mass (LVM) is a significant predictor of adverse cardiovascular events. Experimental studies suggest a pathophysiological role of magnesium (Mg ++ ) in the development of arterial hypertension and left ventricular hypertrophy (LVH). Subjects, aged 45 – 79 years, from the population-based, epidemiological, longitudinal “Study of Health in Pomerania” with complete echocardiographic data (n = 1 348 after exclusion of aortic valve disease) were grouped into five quintiles according to serum Mg ++ at baseline (0.790±0.003 mmol/l, mean±SEM). In the lowest Mg ++ -quintile (Mg ++ <<26>0.73 mmol/l), LVM (187.4±3.1 g at baseline) increased by 14.9±1.2 g over the following five years, while in the highest Mg ++ -quintile (Mg ++ <<0.85 mmol/l) LVM (186.7±3.4 g at baseline) decreased by −0.5±2.8 g (p<0.0001 between quintiles). After five years, LVM was significantly higher in the lowest Mg ++ -quintile (LVM: 202.2±3.4 g) in comparison to the highest Mg ++ -quintile (187.2±3.5 g, p<0.003 between quintiles). Mg ++ inversely correlated with the difference in LVM over five years (p<0.0001, females: p<0.002, males: p<0.024) and also with differences in Sokolow-Lyon-index over five years (p<0.01). Unlike age, gender, presence of hypertension or diabetes mellitus, body mass index, smoking status, renal function, high-sensitive C-reactive protein and use of various antihypertensive drugs including diuretics, only Mg ++ (p<0.0001), pulse pressure (p<0.002) and use of β-blockers (p<0.04) were identified as independent predictors of difference in LVM over five years. Hypomagnesemia is one of the strongest predictors of increase in LVM over the following five years. It should initiate decisive preventive measures in subjects prone to developing LVH.