Mizushima S, Cappuccio F P, Nichols R, Elliott P
Department of Epidemiology and Public Health, Imperial College School of Medicine at St Mary's Campus, London, UK.
J Hum Hypertens. 1998 Jul;12(7):447-53. doi: 10.1038/sj.jhh.1000641.
Published reports of 30 separate sets of analyses from 29 observational studies relating dietary intake of magnesium to blood pressure (BP) were identified through a comprehensive search using MEDLINE and BIDS-EMBASE. Three studies were prospective, 24 cross-sectional (25 reports), of which four also contained a longitudinal component, and two were obtained from baseline data in a trial. Various dietary methodologies were used: 24-h dietary recall (n = 12), food-frequency questionnaire (8), food record (7), and duplicate diet (2). Twelve reports compared magnesium intake or BP level between subgroups. Seven showed a negative association between magnesium intake and BP level, and five reported no association. From 18 of the 30 sets of analyses either a regression estimate or a Pearson correlation coefficient was reported. Many reports also allowed identification of subgroups by sex, age and race. Ninety population samples and subgroups could thus be identified from the 30 reports. All 11 Pearson-r correlation coefficients reported for systolic BP (SBP) (three significant, P < 0.05) and 10 (out of 12) Pearson-r correlation coefficients reported for diastolic BP (DBP) (four significant) were negative. Seven reports (13 subgroups for SBP, 11 subgroups for DBP) gave partial regression coefficients after adjustment; 10 (seven significant) and eight (six significant) were negative for SBP and DBP, respectively. For 13 subgroups in five papers, Pearson-r correlation coefficients were reported after adjustment for confounding factors. Eight (out of 13) showed a negative relationship for SBP and DBP. This review points to a negative association between dietary magnesium intake and BP. A systematic quantitative overview is needed to reconcile the inconsistencies of the results of individual studies and to quantify the size of such relationship.
通过使用MEDLINE和BIDS-EMBASE进行全面检索，确定了29项观察性研究中30组独立分析的已发表报告，这些研究涉及镁的饮食摄入量与血压（BP）的关系。三项研究为前瞻性研究，24项为横断面研究（25份报告），其中四项还包含纵向研究部分，两项来自一项试验的基线数据。采用了多种饮食方法：24小时饮食回顾法（n = 12）、食物频率问卷法（8）、食物记录法（7）和双份饮食法（2）。12份报告比较了亚组之间的镁摄入量或血压水平。七份报告显示镁摄入量与血压水平之间存在负相关，五份报告称无关联。在30组分析中的18组中，报告了回归估计值或皮尔逊相关系数。许多报告还允许按性别、年龄和种族确定亚组。因此，从30份报告中可以确定90个人口样本和亚组。报告的所有11个收缩压（SBP）的皮尔逊r相关系数（三个显著，P < 0.05）和12个舒张压（DBP）的皮尔逊r相关系数中的10个（四个显著）均为负数。七份报告（SBP有13个亚组，DBP有11个亚组）在调整后给出了偏回归系数；SBP和DBP的偏回归系数分别为10个（七个显著）和八个（六个显著）为负数。对于五篇论文中的13个亚组，在对混杂因素进行调整后报告了皮尔逊r相关系数。13个中有八个显示SBP和DBP之间存在负相关。本综述指出饮食中镁摄入量与血压之间存在负相关。需要进行系统的定量综述，以协调个别研究结果的不一致之处，并量化这种关系的大小。
