Effect on Blood Pressure of Potassium, Calcium, and Magnesium in Women With Low Habitual Intake

Methods

Results

Three hundred of the 321 randomized participants (93%) completed baseline and midpoint measurements. Reasons for withdrawal from the study were difficulty swallowing the pills (3), pregnancy (4), illness (4), family/social problems (5), and possible adverse reactions to the supplements (5). Mean age was 39±5 (SD) years, mean body mass index was 23±3 kg/m², and whites comprised 96% of the group (Table 1⇓). There were no important differences among the treatment groups in characteristics at baseline (Table 1⇓). Ten participants withdrew between the midpoint and end and therefore did not provide final measurements. Reasons for withdrawal were pregnancy (1), possible side effects (2), development of a concomitant medical problem (2), and various personal issues (5). These subjects were included in the primary analysis of blood pressure changes. At baseline, mean daily dietary intake of potassium was 2433±792 mg (62±20 mmol), calcium 638±265 mg, and magnesium 239±79 mg; there were no appreciable differences among the groups (Table 2⇓). These intakes corresponded to the following percentiles in the Nurses Health Study II population from which the participants in the present study were selected: potassium, 10th; magnesium, 17th; and calcium, 10th. The addition of the supplements raised the mean total daily intake of potassium to 4072 mg (102 mmol), magnesium to 585 mg, and calcium to 1835 g. These intakes correspond to the following percentiles in the Nurses Health Study II population: potassium 90th, magnesium 95th, and calcium >95th.

Baseline 24-hour urinary potassium excretion was 41±16 mmol, calcium 4.8±2.2 mmol, and magnesium 3.9±1.3 mmol and was similar across the four treatment groups (Table 3⇓). Urinary excretion of potassium, calcium, and magnesium increased significantly in the groups that received these minerals as supplements, compared with the placebo group (Table 3⇓). According to pill counts, the patients in all groups took 94% to 96% of their assigned pills. The increase in potassium excretion, 29 mmol, was 76% of the ingested dose, 38 mmol (40 mmol administered × 96% compliance by pill count). The increases in the group that received all three minerals were similar to those in the groups that received a single mineral supplement. Sodium excretion was similar across treatment groups, averaging 135 mmol/d.

Mean 24-hour blood pressure at baseline was 116±8 mm Hg systolic and 73±6 mm Hg diastolic, and was similar among the four groups (Table 4⇓). Compared with the placebo group, blood pressure decreased significantly from baseline in the potassium group by 2.0 mm Hg for systolic pressure (95% confidence interval, −3.7 to −0.3), and by 1.7 mm Hg for diastolic pressure (−3.0 to −0.4). There were no significant changes in the magnesium or calcium groups. The combination of all three minerals had less effect on blood pressure than potassium and was not significantly different from placebo.

When the multiple regression analysis was used to adjust the blood pressure changes for baseline blood pressure and urinary excretion of the mineral or minerals that were supplemented, the differences were slightly diminished between treatment groups and the placebo group. These adjusted differences were for potassium −1.4 systolic (P=.09)/−1.3 diastolic (P=.04), for calcium −0.1 systolic (P=.9)/−0.4 diastolic (P=.5), for magnesium −0.5 systolic (P=.5)/−0.3 diastolic (P=.6), and for all three supplements −1.0 systolic (P=.2)/−0.8 diastolic (P=.2). Adding baseline sodium excretion to the multiple regression analysis did not affect the differences in blood pressure between the treatment groups and the placebo group. Subgroup analysis did not reveal differences in the blood pressure changes between participants who had baseline blood pressure levels, urinary cation excretion, urinary sodium excretion, or dietary cation intake above versus below the mean.

Blood pressure at baseline showed expected diurnal variation in the day-shift workers (Fig 1⇓). This diurnal pattern was similar in all groups at baseline and during supplementation. The night-shift workers had a blunted diurnal variation (data not shown). The changes in blood pressure from baseline for 2-hour intervals were analyzed in each treatment group for the full group and the day-shift workers separately. There were not enough night-shift workers in each group for separate analysis of change (4 to 5 in each group except for 8 in the magnesium group). The results for the full group (Figs 2⇓ and 3⇓) and the day-shift workers were similar. The decrease in blood pressure in the potassium group was fairly uniform throughout the 24 hours of monitoring, except in the late evening when the hourly differences tended to be less than at other times (Figs 2⇓ and 3⇓). An effect of calcium, magnesium, or the combination of minerals on blood pressure was not apparent at any interval during the 24 hours.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Ambulatory blood pressure at baseline for day-shift workers. n=274. Data points show means and standard deviations. •, systolic; ▪, diastolic.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Diastolic blood pressure changes in the potassium and placebo groups. Ambulatory blood pressure measurements were averaged for 2-hour intervals with the data points indicating the median time within each 2-hour interval. •, potassium group (n=49); ▪, placebo group (n=103). Data points show the mean changes from baseline averaging the 8- and 16-week measurements during supplementation. Error bars show standard errors of the changes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Systolic blood pressure changes in the potassium and placebo groups. See Fig 2⇑ for description.

There were no significant differences in the frequency of symptoms between the groups except for increased perception of energy level in the potassium group (8/49, 16%) compared with the control group (5/103, 5%), and loose stools in the magnesium group (20/50, 40%) compared with the control group (18/103, 17%). In each group, 1 to 2 participants withdrew from the study due to possible side effects. None were required to withdraw because of elevated blood pressure (>160 systolic or >95 diastolic).

Discussion

We determined that supplementation with potassium, but not with magnesium or calcium, significantly lowered blood pressure in a normotensive group of US nurses who reported low intakes of these minerals. Mean intakes of potassium, calcium, and magnesium in this trial corresponded to the 10th, 17th, and 10th percentiles of the Nurses Health Study II population. These mean intakes were also lower than the means for the US population of white women, ages 35 to 49 years, adjusted to the same total energy intake as the nurses: potassium 2433 versus 3077 mg,³⁴ calcium 638 versus 1059 mg,³⁴ and magnesium 239 versus 344 mg.³⁵ The overall nurse population had intake of these minerals that were similar to the means for the US population. The low reported intake in the nurses in this trial could be confirmed biochemically for potassium because of the close biological relationship between dietary intake and urinary excretion. Mean baseline 24-hour urinary excretion of potassium was 40 mmol, which corresponds to the reported daily intake of 62 mmol (2433 mg). Most other trials of potassium supplementation were of women who had a higher urinary potassium excretion, 50 to 60 mmol.^{29 36 37 38 39} Reported compliance was excellent for the potassium, calcium, and magnesium supplements, and it was confirmed by increases in urinary excretion of each supplement. Urinary creatinine excretion averaged 1.1 to 1.2 g/d, which is similar to other studies in women.^{36 40} The molar ratio of dietary potassium to urinary potassium was 1.5 compared with 1.3 in a previous study,⁴¹ also suggesting good compliance with urine collections. Finally, the quantitative relationship between changes in intake and urinary excretion of potassium demonstrated near perfect compliance in the two groups that received potassium. (Because the biological relationships between dietary and urinary calcium and magnesium are not quantitative, the significant increases in excretion during supplementation give qualititative confirmation of compliance.) Therefore, the trial achieved its goals for a low baseline mineral intake and adherence to the supplementation that were needed to test the hypotheses.

Potassium lowers blood pressure in hypertensives, −8.2/−4.5 mm Hg in one meta-analysis,¹⁶ and −5.1/−3.0 mm Hg in another.¹⁰ Potassium trials in hypertensive women³⁸ or those in which women comprised at least half of the group^{23 37 39 42 43} showed blood pressure–lowering similar to studies in hypertensive men. In contrast, potassium supplementation had little or no effect on blood pressure in normotensive persons who were not selected for low intake, even with higher doses of potassium than used in the present study.^{36 44 45} For example, in the Trials of Hypertension Prevention, Phase 1, potassium, 60 mmol, did not significantly lower blood pressure compared with placebo in the entire group of 353 normotensive persons, or in the 98 women.³⁶ Other trials in normotensive persons suggest that potassium supplementation has an enhanced effect on blood pressure in persons with low, rather than average or high, usual intake. Reducing potassium intake from an average to a very low level, 77 to 10 mmol, increased blood pressure significantly, by 6.1/3.6 mm Hg, whereas raising potassium intake from 62 to 90 mmol produced nonsignificant changes of −1.0/−3.2 mm Hg.⁴⁶ In blacks, raising potassium from a low level of 32 to 35 mmol to 112 to 115 mmol reduced blood pressure by 6.9/2.5 mm Hg.⁴⁷ However, hyper-responsiveness to potassium in blacks is an alternative explanation.^{38 42 43 48} These studies, in combination with the present study, suggest that in normotensives, dietary potassium becomes important for blood pressure regulation at low intake.

In the present trial, neither magnesium nor calcium supplements lowered blood pressure. Two recent meta-analyses found no overall effects of calcium on blood pressure in normotensive persons.^{17 18} In a recent trial in normotensive children, ages 10 to 13 years, systolic blood pressure–lowering was inversely proportional to the baseline calcium intake.⁴⁹ However, in adults with usual calcium intake less than 600 mg, calcium supplements did not lower blood pressure.^{50 51 52 53 54 55} Therefore, we conclude that increasing calcium intake by supplementation does not lower blood pressure in normotensive adults with low or average baseline intake.

Several previous trials studied combinations of cations. Trials of low-fat milk, which raised primarily calcium but also produced mild increases in magnesium and potassium intakes, did not lower blood pressure in hypertensives (mean change, +4/+1 mm Hg)⁵¹ ; the effects in normotensives were significant for systolic pressure but not diastolic blood pressure (−2.7/+0.2 mm Hg).⁵⁶ In hypertensive patients, magnesium supplements did not augment the blood pressure–lowering effect of potassium.²³ In fact, the blood pressure reduction was slightly less with the combination than with potassium. In mildly hypertensive subjects, none of the combinations, potassium and magnesium, potassium and calcium, or magnesium and calcium significantly lowered blood pressure.²⁸ These results were surprising because the two combinations with potassium did not cause the decrease in blood pressure expected from potassium itself. This raises the possibility that magnesium and calcium, ineffective as single supplements, could actually interfere with the blood pressure–lowering action of potassium. The present study also suggests such a negative interaction, since the blood pressure decrease in the group that received the combination of potassium, calcium, and magnesium was smaller in magnitude than in the group that received potassium alone and was not significant. Epidemiological studies generally have found an inverse relationship between dietary calcium and blood pressure levels.^{7 12 13} However, in certain populations, there is evidence that calcium may contribute to elevated blood pressure. In a population study, urinary calcium excretion was positively associated with blood pressure.⁵⁷ In the present and other studies,^{28 30} calcium supplements raise urinary calcium excretion. Gruchow et al⁵⁸ found that potassium had an inverse association with blood pressure but only in persons whose calcium intakes were in the lower third of the US population. In those persons with average or high calcium intakes, potassium was not associated with low blood pressure.

The results of the present trial may be used to interpret findings from epidemiological studies that found associations of dietary intakes of calcium, magnesium, or potassium with hypertension or blood pressure levels.^{1 2 3 4 5 6 7 8 9 10 11 12 13 14 15} It is difficult in multiple regression analysis to separate the individual effects of these minerals because the intakes are strongly correlated with each other. The magnitude of the blood pressure–lowering effect of potassium in the present trial is consistent with the difference in blood pressure seen from high to low potassium intake in these populations. In the large-scale Dietary Approaches to Stop Hypertension (DASH) trial,⁵⁹ a diet rich in fruits and vegetables that raised lower-than-average potassium and magnesium intakes significantly decreased 24-hour ambulatory blood pressures in persons with high-normal diastolic blood pressure or Stage 1 hypertension (−3.1/−2.1 mm Hg), and significantly decreased clinic blood pressures in the hypertensive group. The blood pressure–lowering in the DASH trial is consistent with the effects of potassium in the present and other supplementation trials. This comparison also suggests that potassium chloride, as used in the supplementation trials, has an effect on blood pressure similar to that of potassium in food, which is mainly in the form of other salts such as citrate. In conclusion, the present study points toward potassium, rather than magnesium or calcium, as the cause of lower blood pressures in populations and provides support for nutritional guidelines to increase dietary potassium to prevent hypertension.