Abstract only Introduction: The role of magnesium (Mg) in blood pressure (BP) is controversial, due to inconsistent results from prior individual trials or systematic reviews and meta-analyses. Hypothesis: We aimed to quantify the effect of oral Mg supplementation on systolic and diastolic BP by synthesizing available evidence from randomized double-blinded placebo-controlled trials. We also examined whether changes in BP were related to elevation in serum Mg levels by Mg supplementation. Methods: We searched all trials of oral Mg supplementation among healthy or hypertensive adults published before February 1, 2015 from databases of MEDLINE and EMBASE. A random-effects meta-analysis was performed to estimate weighted mean differences (WMDs) and 95% confidence intervals (CIs) of BP changes. The possible heterogeneity between studies were explored by subgroup analyses. We used restricted cubic spline curves to depict time- and dose-responses of BP in relation to changes of serum Mg levels. Results: In total, 33 eligible trials involving 2,500 participants were included. Mg supplementation at a median dose of 368 mg/day elemental Mg for a median duration of 3 months led to significantly reduced supine BPs by 2.08 mmHg (95% CI: 0.49, 3.68) for systolic BP and 1.83 mmHg (95% CI: 0.77, 2.90) for diastolic BP compared with the placebo. Overall, Mg supplementation significantly increased serum Mg concentrations by 0.05 mmol/L (95% CI: 0.03, 0.07 mmol/L). Our time- and dose-response analyses showed that Mg supplementation with 200 mg/day for 1 month was sufficient to significantly raise serum Mg and lower supine BPs (all P-values < 0.0001). However, higher doses (≥ 300 mg/day) and longer durations (≥ 2 months) were needed to achieve maximal effects. On average, per 0.1 mmol/L increment in serum Mg was associated with a supine DBP reduction of 2.26 mmHg (95% CI: 0.27, 4.26 mmHg), while the linear or curvilinear relationship with supine SBP was non-significant. In addition, there were non-significant changes in standing BPs (all P-values > 0.05). Conclusions: Our meta-analysis showed that Mg supplementation significantly lowered supine BPs which were significantly related with elevated serum Mg levels. Our findings support an antihypertensive effect of Mg, although future large well-designed randomized controlled trials with longer follow-up, selection of participants with low Mg levels, and ambulatory BP monitoring are warranted.