## Abstract

To investigate the effects of dietary magnesium, potassium, and sodium on children's lung function, the authors examined cross-sectional dietary data and pulmonary function tests from 2,566 children aged 11–19 years who attended schools in 12 southern California communities during 1998–1999. During school visits, each child completed a health update questionnaire, a validated food frequency questionnaire, and spirometric lung function testing. Low magnesium and potassium intakes were associated with lower lung function. Girls with low magnesium intake had lower forced expiratory flow at 75% of the forced vital capacity (FEF75) (−8.3%, 95% confidence interval: −14.8, −1.4) than did girls with higher intake; reductions were larger in girls with asthma (forced expiratory flow between 25% and 75% of the forced vital capacity (FEF25–75) (−16.2%, 95% confidence interval: −22.7, −9.1) and FEF75 (−24.9%, 95% confidence interval: −32.8, −16.1)) than in girls without asthma (FEF25–75 (−2.0%, 95% confidence interval: −7.4, 3.8) and FEF75 (−4.1%, 95% confidence interval: −11.3, 3.7)). Boys with low magnesium intake showed deficits in forced vital capacity (−2.8%, 95% confidence interval: −5.4, −0.2) compared with boys with higher intake. The effects of low magnesium intake did not vary substantially in boys with and without asthma. Among girls, low potassium intake was also associated with deficits in forced expiratory volume in 1 second (−2.7%, 95% confidence interval: −5.2, −0.1) and forced vital capacity (−2.4%, 95% confidence interval: −4.7, −0.1). In summary, low magnesium and potassium intakes were associated with lower lung volumes and flows.

As the role of diet in respiratory health receives increasing recognition, a body of evidence is emerging in support of associations between cationic minerals and lung function (15). The evidence suggests that magnesium, potassium, and sodium participate in numerous biochemical and physiologic processes that directly influence lung function and indirectly influence respiratory symptoms (3, 4, 68). The mechanisms for effects on lung function and symptoms include alteration in smooth muscle function (9), neuromuscular excitability (10), immune function (11), oxidative stress (12), DNA and RNA synthesis, and enzymatic activity (13). Furthermore, these mechanisms contribute to the pathophysiology of asthma, and their effects in cases of inadequate magnesium and potassium intakes may be particularly apparent in children with asthma (3, 4, 14, 15).

The role of magnesium intake in the variation of lung function has been studied in healthy adults and in asthma patients (3, 1618). Epidemiologic evidence from a population-based study in adults indicates that low dietary intake of magnesium is associated with impairment of lung function as measured by airway flow rates, airway hyperreactivity, and increased risk of wheezing (3). Studies of asthma patients indicate that dietary magnesium intake and serum magnesium levels are lower than those of healthy controls (19, 20). However, short-term magnesium supplementation trials to assess the effects of supplemental magnesium on lung function and symptoms among patients with asthma have had mixed results (16, 17). Studies also suggest that high sodium intake or low potassium intake may accentuate airway reactivity and reduce flows; however, the interrelations are complex and study results have been inconsistent (2). Serum potassium may be decreased in asthma patients, but the decrease appears to be related to the hypokalemia associated with the use of beta2 agonist medication (15, 2123). Although the effects of dietary intake of magnesium, potassium, and sodium on lung function have been reported in adults, the relations between the intake of these minerals and lung function have not been extensively studied in children.

The Children's Health Study offers an opportunity to further investigate the role of dietary mineral intake on children's airway flows (24). We examined cross-sectional data on dietary intake and spirometric measures of lung function from 2,566 participants in the Children's Health Study collected at follow-up visits during the 1998–1999 school year. We used regression splines to assess the effects of low magnesium, potassium, and sodium intakes on lung function.

## MATERIALS AND METHODS

### Study design

The Children's Health Study is a 10-year longitudinal study of long-term air pollution exposure health effects that includes school children who reside in 12 communities within a 200-mile (320-km) radius of Los Angeles, California. Details on the design, site selection, subject recruitment, and assessment of health effects are reported elsewhere (24, 25). At study entry in early 1993, parents or guardians of 3,681 participating children provided written informed consent and completed a self-administered questionnaire on demographics, medical and family health history, and indoor air exposures and household characteristics. In the spring of 1993, pulmonary function testing was conducted, and in each subsequent year of the ongoing study, each child completed a follow-up questionnaire and repeated the pulmonary function testing. In the fall of 1995, a second group of 2,081 fourth grade students (approximately 150 per community) was recruited and completed the same baseline and follow-up questionnaires and pulmonary function testing as the group enrolled in 1993. Beginning in 1998, dietary data were collected using a validated food frequency questionnaire; these were completed by 2,895 children who ranged in age from 11 to 19 years and were enrolled in schools in the participating communities.

### Pulmonary function testing

We focused our investigation on measures of airway flow rates based on the physiologic roles of magnesium, potassium, and sodium and studies of the effects of low magnesium, potassium, and sodium intakes on lung function in adults. We also assessed lung volume to assist in interpreting the data on airway flows. Maximal forced expiratory flow-volume maneuvers were recorded using rolling-seal spirometers (Spiroflow; P. K. Morgan, Ltd., Gillingham, United Kingdom). Spirometric calibrations and room temperatures were measured just before, during, and just after each testing session using flow-volume syringes (Jones Medical Instrument Co., Oak Brook, Illinois), and lung function measures were corrected for changes in calibration or internal temperature. Testing and data management procedures have been reported previously (24). Each subject was asked to perform three satisfactory maneuvers, and no more than seven maneuvers were attempted during any test session. A satisfactory maneuver was defined as the forced vital capacity's agreeing within 5 percent, the FEV1's agreeing within 5 percent, an extrapolated volume for FEV1 of less than 100 ml or 5 percent of forced vital capacity, an expired volume in the final 2 seconds of less than approximately 50 ml (45 ml × the standard body temperature pressure saturated correction factor), and forced expiratory time's exceeding 3 seconds. Of the 2,895 children who completed a food frequency questionnaire, 2,633 also performed an acceptable pulmonary function testing.

### Dietary information

Dietary intake was assessed using the youth/adolescent questionnaire, a validated food frequency questionnaire developed by Rockett et al. (26) and Rockett and Colditz (27) for use in older children and adolescents. The questionnaire has 131 food items including snacks and is a modified version of the validated Nurses' Health Study food frequency questionnaire (28). In a validation study, the correlation coefficients between the food frequency questionnaire and three 24-hour dietary recalls were 0.60 for magnesium, 0.56 for potassium, and 0.24 for sodium and were similar to those observed among adults (26, 27).

Nutrient and energy intakes were quantified for each person. Of the 2,633 children who completed both the food frequency questionnaire and lung function tests, 67 children with total energy intake below 500 calories or above 5,000 calories were excluded from analyses, resulting in a final sample size of 2,566 children aged 11–19 years. Because mineral intake depends on gender and total caloric intake, we used sex-specific quintiles for descriptive analyses of magnesium, potassium, and sodium intakes.

Because our a priori hypothesis was that very low magnesium intake is associated with lower lung function in our study population, we chose to compare low intake with higher intake on the basis of two considerations. First, dietary intakes of cationic minerals are known to affect airway physiology. Based on airway physiology, a linear dose-response relation over a large range of magnesium intake would not be expected. Intake of magnesium below some minimal level may increase airway tone and reactivity. Intake above that level would not be associated with as substantial a change in function. Because we live in a period in which most people's dietary intake far exceeds that needed for normal function, we reasoned that only those with very low intake would show effects on lung function. On the basis of the hypothesis that low intake is associated with reduced airway flows, we classified low intake as less than the 20th percentile in analyses of lung function and adjusted all estimates for total caloric intake.

### Sociodemographic, medical history, and exposure data

The Children's Health Study questionnaire provided information on sociodemographic factors, history of respiratory illness and associated risk factors, exposure to environmental tobacco smoke, and maternal smoking history. Ethnicity was defined as non-Hispanic White, Hispanic, African American, Asian, and other ethnicities, based on self-report. Health insurance was defined as any insurance coverage reported for the participant's family.

Baseline questionnaire responses by parents or guardians and self-report of doctor-diagnosed asthma during lung function testing were used to categorize children's asthma status at the time of food frequency questionnaire completion. A child with persistent asthma was defined as any child who had doctor-diagnosed asthma and who was symptomatic in the year before completing the food frequency questionnaire or who took any asthma medication during the 12 months before the date that the questionnaire was completed. A child with wheezing was defined as any child with any lifetime history of wheezing.

Personal smoking was defined as a history of the participant's reporting having ever smoked more than 100 cigarettes, as ascertained by a private interview during spirometry. Exposure to maternal smoking in utero was characterized using the responses from the questionnaire completed by parents or guardians. Each child's environmental tobacco smoke exposure status was based on reports of current smoking status of the mother, father, other adult household members, and regular household visitors.

Participants' height and weight were measured using a standardized protocol, and any respiratory infection within 1 month of testing and exercising within 30 minutes of testing were documented by trained field staff immediately before lung function testing. Body mass index was calculated as the weight (kg)/height (m)2 and categorized into age- and sex-specific quintiles.

### Statistical analyses

The relations between lung function and physiologic growth factors such as age and height have been found to be highly nonlinear from childhood through adolescence (29, 30). The distributions of magnesium, potassium, and sodium intakes were examined and categorized into sex-specific quintiles. Differences within categories of selected covariates were tested with analysis of variance. We then assessed the effects of magnesium, potassium, and sodium intakes on lung function by using regression splines to capture the nonlinear relation among pulmonary function, age, and height (2933). Regression splines fit piecewise polynomials that are joined smoothly at the cutpoints, known as knots. This has the advantage of allowing appropriate statistical inference while capturing the nonlinear relations in the data. Initially, a knot was placed at each integer age. The final models were fitted by using knots at ages 13 and 17 years, leading to a more parsimonious model with essentially the same results.

All models were fit separately for males and females because their smoothing shapes for the relation between lung function and age are different. The gender-specific regression model was given as:

$E\left\{\mathrm{log}\ \left(\mathrm{PFT}\right)\right\}\ {=}\ {\mu}\ {+}\ S_{1}\left(\mathrm{AGE}\right)\ {+}\ S_{2}\left(\mathrm{AGE}\right)\ {\times}\ \mathrm{log}\left(\mathrm{HT}\right)\ {+}\ \mathbf{X}{\beta},$
where PFT is a pulmonary function test like forced vital capacity or FEV1, μ is the overall mean, AGE is age at visit, HT is the residual of height at visit after smoothing height on age, and X is a vector of covariates (including mineral intake) of interest and a set of adjustment variables (including cohort, community, technician, spirometer, race/ethnicity, barometric pressure, and other possible confounders). This model is an example of the varying-coefficient modeling strategy of Hastie and Tibshirani (34). We used natural cubic splines that impose the additional constraint that the function be linear beyond the boundary knots.

Because we hypothesized that the effects of mineral intake on lung function appeared to be related to low intake, we used categories of low intake (20th percentile) and higher intake (greater than 20th percentile) in the flexible models. Flexible models were fitted that included sex-specific mineral intake and variables for cohort, community, ethnicity, spirometer, spirometer temperature, technician, and barometric pressure. Note that the models are additive on the log scale, and the results are presented as the percent differences from the reference curve at the mean age. The primary parameters of interest are the main effects for low mineral intake, which characterizes a parallel percent difference in pulmonary function compared with the baseline group of high mineral intake.

All models were adjusted for total energy intake. Based on previous analyses and reported studies, other aspects of diet such as fat, carbohydrate, fruit and vegetable intakes and vitamin intake, parental education (<12 grades, 12 grades, some college, college, and some graduate), household income (<$7,500,$7,500∼$14,999,$15,000∼$29,999,$30,000∼$49,999,$50,000∼99,999, and ≥$100,000), body mass index (weight (kg)/height (m)2) cohort- and sex-specific quintiles, insurance status (yes/no), personal smoking (yes/no), environmental tobacco smoke exposure (yes/no), and respiratory illness at lung function test (yes/no) were evaluated as potential confounders and were included in models if the adjusted estimates for intake changed by 10 percent or more compared with the unadjusted estimates. Subjects with missing data for a given covariate were excluded from the analyses that involved that covariate. To assess the modifying effect of asthma on the relation between low magnesium, potassium, and sodium intakes and lung functions, we conducted gender-specific stratified analyses for children with and without asthma. We tested the statistical significance of interaction terms between asthma status and mineral intake. All analyses were conducted by using the S-Plus statistical software package (35). ## RESULTS Selected characteristics of participants who completed both a food frequency questionnaire and lung function testing are enumerated in table 1. Participants ranged in age from 11 to 19 years and were predominantly non-Hispanic White and from middle-class families with health insurance. Few smoked, but one third were exposed to environmental tobacco smoke. Overall, 23.0 percent reported ever being diagnosed with asthma by a physician, and 14.8 percent had persistent asthma in the year previous to the date of the interview. We note that our prevalences for asthma and persistent asthma are higher than in previous studies in the United States and Australia; however, the prevalence of asthma has increased by 50 percent since 1980, and older studies may not reflect the current prevalence. TABLE 1. Selected characteristics of participants (n = 2,566), Children's Health Study, Los Angeles, California, 1998–1999 No. Sex Girls 1,389 54.1 Boys 1,177 45.9 Age (years) 11–14 1,090 42.5 15–16 617 24.0 17–19 859 33.5 Cohort Grade 4 (recruited in 1993) 461 18.0 Grade 7 (recruited in 1993) 1,000 39.0 Grade 4 (recruited in 1995) 1,105 43.1 Race/ethnicity Non-Hispanic White 1,481 57.9 Hispanic 678 26.5 Black (African American) 112 4.4 Asian 149 5.8 Others 137 5.4 Parental education Some graduate 349 14.1 College 269 10.9 Some college 1,100 44.5 12 grades 466 18.9 <12 grades 286 11.6 Income ($)
≥100,000 152 7.0
50,000∼99,999 794 36.5
30,000∼49,999 597 27.4
15,000∼29,999 332 15.2
7,500∼14,999 193 8.9
<7,500 110 5.1
Insurance
Yes 2,136 85.4
Respiratory illness at PFT*
Yes 429 18.2
Personal smoking
Yes 657 25.6
Current ETS* exposure
Yes 847 33.1
In utero exposure
Yes 377 15.3
Asthma
Yes 579 23.0
Persistent asthma
Yes 372 14.8
No.
Sex
Girls 1,389 54.1
Boys 1,177 45.9
Age (years)
11–14 1,090 42.5
15–16 617 24.0
17–19 859 33.5
Cohort
Grade 4 (recruited in 1993) 461 18.0
Grade 7 (recruited in 1993) 1,000 39.0
Grade 4 (recruited in 1995) 1,105 43.1
Race/ethnicity
Non-Hispanic White 1,481 57.9
Hispanic 678 26.5
Black (African American) 112 4.4
Asian 149 5.8
Others 137 5.4
Parental education
College 269 10.9
Some college 1,100 44.5
Income (\$)
≥100,000 152 7.0
50,000∼99,999 794 36.5
30,000∼49,999 597 27.4
15,000∼29,999 332 15.2
7,500∼14,999 193 8.9
<7,500 110 5.1
Insurance
Yes 2,136 85.4
Respiratory illness at PFT*
Yes 429 18.2
Personal smoking
Yes 657 25.6
Current ETS* exposure
Yes 847 33.1
In utero exposure
Yes 377 15.3
Asthma
Yes 579 23.0
Persistent asthma
Yes 372 14.8
*

PFT, pulmonary function test; ETS, environmental tobacco smoke.

Intake of magnesium was higher in boys than in girls; however, both had intakes that were substantially lower than the recommended daily allowance for this age group (410 mg/day for boys, 360 mg/day for girls) (table 2). Less than 14 percent of boys and 12 percent of girls had adequate intakes of magnesium. There was no difference in the mean magnesium intake between children with and without asthma. Potassium intake was also higher in boys than in girls, but intake for both genders was within the recommended range (2,000–3,500 mg/day). Sodium intake was at the upper limit of the recommended range (500–2,400 mg/day).

TABLE 2.

Mineral intake (mg/day) by participant characteristics, Children's Health Study, Los Angeles, California, 1998–1999

Mineral Participant characteristic, age (years) Boys

Girls

Mean (SD*20th percentile p value Mean (SD) 20th percentile p value
Magnesium 11–14 271.5 (120.4) 168.4 0.09 245.3 (114.8) 150.9 0.53
15–16 256.4 (109.0) 168.8  239.9 (110.4) 146.0
17–19 275.7 (117.7) 173.5  248.7 (106.1) 155.0
Potassium 11–14 2,584.3 (1,169.4) 1,584.8 0.13 2,384.2 (1,149.0) 1,427.1 0.28
15–16 2,431.8 (1,067.5) 1,566.4  2,275.9 (1,046.0) 1,416.9
17–19 2,593.2 (1,124.1) 1,639.5  2,385.1 (1,057.3) 1,454.0
Sodium 11–14 2,518.5 (1,112.3) 1,576.6 0.04 2,286.6 (1,029.3) 1,443.1 0.34
15–16 2,382.2 (986.1) 1,483.7  2,190.9 (927.0) 1,416.0
17–19 2,586.0 (1,017.0) 1,735.2  2,268.1 (920.4) 1,510.2
Mineral Participant characteristic, age (years) Boys

Girls

Mean (SD*20th percentile p value Mean (SD) 20th percentile p value
Magnesium 11–14 271.5 (120.4) 168.4 0.09 245.3 (114.8) 150.9 0.53
15–16 256.4 (109.0) 168.8  239.9 (110.4) 146.0
17–19 275.7 (117.7) 173.5  248.7 (106.1) 155.0
Potassium 11–14 2,584.3 (1,169.4) 1,584.8 0.13 2,384.2 (1,149.0) 1,427.1 0.28
15–16 2,431.8 (1,067.5) 1,566.4  2,275.9 (1,046.0) 1,416.9
17–19 2,593.2 (1,124.1) 1,639.5  2,385.1 (1,057.3) 1,454.0
Sodium 11–14 2,518.5 (1,112.3) 1,576.6 0.04 2,286.6 (1,029.3) 1,443.1 0.34
15–16 2,382.2 (986.1) 1,483.7  2,190.9 (927.0) 1,416.0
17–19 2,586.0 (1,017.0) 1,735.2  2,268.1 (920.4) 1,510.2
*

SD, standard deviation.

p value of F test.

Low magnesium intake was associated with lower measures for several lung functions (table 3). All the estimates for differences in lung function for low magnesium intake were negative and were statistically significant for forced vital capacity for boys (−2.8 percent) and forced expiratory flow at 75 percent of the forced vital capacity (FEF75) for girls (−8.3 percent). Low potassium intake was associated with lower forced vital capacity in girls, and flows tended to be lower in both boys and girls, although results were statistically significant only for FEV1 in girls. Estimates for the effects of low sodium intake on flows were in the opposite direction from those for magnesium and potassium. Low sodium intake was generally associated with higher lung function, but estimates were statistically significant only for forced vital capacity among boys. We examined the relation between the sodium/potassium ratio and lung function and found little evidence that the ratio was related to lung function. We found little evidence for confounding by socioeconomic status (income, education), other aspects of diet, or smoking and did not include these covariates in our final models.

TABLE 3.

Effect of low daily mineral intake on lung function, Children's Health Study, Los Angeles, California, 1998–1999*

Mineral Lung function Boys

Girls

% difference 95% CI % difference 95% CI
Magnesium FVC −2.84 −5.4, −0.2 −0.55 −3.1, 2.0
FEV1 −2.67 −5.4, 0.1 −1.66 −4.4, 1.1
FEF25–75 −3.63 −9.0, 2.1 −4.84 −9.8, 0.4
FEF75 −4.69 −12.0, 3.2 −8.33 −14.8, −1.4
Potassium FVC 0.05 −2.5, 2.7 −2.40 −4.7, −0.1
FEV1 −1.13 −3.8, 1.6 −2.68 −5.2, −0.1
FEF25–75 −3.51 −8.7, 2.0 −3.85 −8.5, 1.1
FEF75 −5.10 −12.2, 2.5 −6.15 −12.3, 0.5
Sodium FVC 3.03 0.1, 6.0 0.59 −2.2, 3.4
FEV1 2.81 −0.3, 6.0 0.94 −2.1, 4.1
FEF25–75 2.09 −4.0, 8.6 1.99 −3.8, 8.1
FEF75 −0.07 −8.2, 8.8 1.30 −6.4, 9.7
Mineral Lung function Boys

Girls

% difference 95% CI % difference 95% CI
Magnesium FVC −2.84 −5.4, −0.2 −0.55 −3.1, 2.0
FEV1 −2.67 −5.4, 0.1 −1.66 −4.4, 1.1
FEF25–75 −3.63 −9.0, 2.1 −4.84 −9.8, 0.4
FEF75 −4.69 −12.0, 3.2 −8.33 −14.8, −1.4
Potassium FVC 0.05 −2.5, 2.7 −2.40 −4.7, −0.1
FEV1 −1.13 −3.8, 1.6 −2.68 −5.2, −0.1
FEF25–75 −3.51 −8.7, 2.0 −3.85 −8.5, 1.1
FEF75 −5.10 −12.2, 2.5 −6.15 −12.3, 0.5
Sodium FVC 3.03 0.1, 6.0 0.59 −2.2, 3.4
FEV1 2.81 −0.3, 6.0 0.94 −2.1, 4.1
FEF25–75 2.09 −4.0, 8.6 1.99 −3.8, 8.1
FEF75 −0.07 −8.2, 8.8 1.30 −6.4, 9.7
*

Models are adjusted for cohort, community, spirometer, technician, barometric pressure, spirometer temperature, race, and total energy intake.

CI, confidence interval; FVC, forced vital capacity; FEV1, forced expiratory volume in 1 second; FEF25–75, forced expiratory flow between 25% and 75% of the forced vital capacity; FEF75, forced expiratory flow at 75% of the forced vital capacity.

Lung function was generally lower in both boys and girls with and without asthma who had low magnesium intake (table 4). Among both boys and girls without asthma, low magnesium intake was associated with lower lung function, especially flows, but the deficits did not achieve statistical significance. Children with asthma had significantly lower flow rates than did children without asthma, independent of magnesium intake. The effects of low magnesium intake on flow rates (forced expiratory flow between 25 percent and 75 percent of the forced vital capacity (FEF25–75) and FEF75) were significantly larger in girls with asthma than in girls without asthma. Among boys, there was little evidence that the effect of low magnesium intake varied by participants' asthma status. The effects of low potassium and sodium intakes did not vary by asthma status.

TABLE 4.

Joint effects of low magnesium intake and asthma* on lung function in boys and girls, Children's Health Study, Los Angeles, California, 1998–1999

Lung function Magnesium intake Boys

Girls

No asthma (n = 856)

Asthma (n = 306)

No asthma (n = 1,085)

Asthma (n = 273)

% difference 95% CI % difference 95% CI % difference 95% CI % difference 95% CI
FVC High intake Reference  1.93 0.2, 3.7 Reference  1.73 −0.2, 3.7
Low intake −2.70 −5.5, 0.2 −2.47 −5.9, 1.1 −0.22 −2.9, 2.5 0.15 −3.7, 4.1
FEV1 High intake 0.00  −2.59 −4.4, −0.8 0.00  −0.99 −3.0, 1.1
Low intake −2.63 −5.6, 0.4 −5.08 −8.6, −1.4 −0.86 −3.8, 2.2 −4.30 −8.3, −0.1
FEF25–75,§ High intake 0.00  −11.49 −14.7, −8.2 0.00  −5.83 −9.5, −2.0
Low intake −3.84 −9.5, 2.2 −11.70 −18.1, −4.8 −1.97 −7.4, 3.8 −16.17 −22.7, −9.1
FEF75 High intake 0.00  −15.12 −19.3, −8.7 0.00  −8.66 −13.5, −3.6
Low intake −4.90 −12.7, 2.6 −15.18 −23.6, −6.2 −4.08 −11.3, 3.7 −24.92 −32.8, −16.1
Lung function Magnesium intake Boys

Girls

No asthma (n = 856)

Asthma (n = 306)

No asthma (n = 1,085)

Asthma (n = 273)

% difference 95% CI % difference 95% CI % difference 95% CI % difference 95% CI
FVC High intake Reference  1.93 0.2, 3.7 Reference  1.73 −0.2, 3.7
Low intake −2.70 −5.5, 0.2 −2.47 −5.9, 1.1 −0.22 −2.9, 2.5 0.15 −3.7, 4.1
FEV1 High intake 0.00  −2.59 −4.4, −0.8 0.00  −0.99 −3.0, 1.1
Low intake −2.63 −5.6, 0.4 −5.08 −8.6, −1.4 −0.86 −3.8, 2.2 −4.30 −8.3, −0.1
FEF25–75,§ High intake 0.00  −11.49 −14.7, −8.2 0.00  −5.83 −9.5, −2.0
Low intake −3.84 −9.5, 2.2 −11.70 −18.1, −4.8 −1.97 −7.4, 3.8 −16.17 −22.7, −9.1
FEF75 High intake 0.00  −15.12 −19.3, −8.7 0.00  −8.66 −13.5, −3.6
Low intake −4.90 −12.7, 2.6 −15.18 −23.6, −6.2 −4.08 −11.3, 3.7 −24.92 −32.8, −16.1
*

Models are adjusted for cohort, community, spirometer, technician, barometric pressure, spirometer temperature, race, and total energy intake.

CI, confidence interval; FVC, forced vital capacity; FEV1, forced expiratory volume in 1 second; FEF25–75, forced expiratory flow between 25% and 75% of the forced vital capacity; FEF75, forced expiratory flow at 75% of the forced vital capacity.

§

Test for interaction of intake and asthma: p < 0.05 in girls.

## DISCUSSION

Magnesium intake varies over a substantial range in the general population, reflecting differences in both total energy intake and the intake of specific food items (36, 37). Foods vary in magnesium content. Green leafy vegetables, nuts, and whole grains have a high magnesium content, followed by milk, meats, and starches. The magnesium content of refined foods is generally low. Because the consumption of processed and refined foods has increased in the US diet, magnesium intake has decreased and is now more likely to be inadequate, especially among adolescents who have the largest daily magnesium intake requirements (1, 37). This national trend in diet was reflected in the low magnesium intake in our cohort, of whom less than 15 percent reached the recommended daily allowance for magnesium intake, which was lower than that reported by Britton et al. (3) in adults. The primary sources of magnesium in our study were meat, milk, and Mexican food such as tacos.

The low levels of magnesium intake in children may have adverse effects on lung function. In this study, low magnesium intake was associated with decreased lung function, especially for flows among girls with asthma. Boys also showed decreased forced vital capacity with low magnesium intake. Our findings are consistent with effects reported among adults (3). To our knowledge, there are no published population-based studies of magnesium intake and lung function in children.

On the basis of our cross-sectional data, we cannot determine whether the effects of low magnesium intake reflect acute reversible effects or chronic irreversible effects. The effects of magnesium on bronchial smooth muscle function might account for decreased lung function. Magnesium is essential for maintenance of electrical potential across cell membranes and therefore modulates bronchial tone and diameter (9, 10). An increase in magnesium concentration relaxes smooth muscle and decreases cholinergic neuromuscular transmission. The acute effects of magnesium on neuromuscular hyperexcitability are mediated by direct and indirect effects via changes in calcium and potassium levels. Our findings of larger effects on flows in girls with low magnesium intake and asthma are consistent with these mechanisms. A low intake of potassium could affect airway flows among children with asthma by contributing to the lower serum and intracellular levels of potassium that result from increased adrenergic tone in children with either asthma or airway hyperresponsiveness (21, 23). The mechanisms for effects on forced vital capacity may also be related to altered neuromuscular function. Although acute changes in lung function may occur with low magnesium or potassium intake, chronic effects from low magnesium intake are also possible (5, 6, 10). Chronically low magnesium intake in adults is associated with chronic adverse health outcomes such as increased cardiovascular morbidity, hypertension, decreased bone mineral density, and insulin resistance, but chronic adverse effects on lung function have not been reported in adults (5, 6, 10). However, the common effects on vascular and airway smooth muscle make chronic effects on lung function biologically plausible. Furthermore, magnesium deficiency affects immune function and responses to oxidative stress that may contribute to the immune dysfunction associated with asthma (11, 12).

We lack data to assess physiologic mechanisms for the differences in effects of mineral intake in boys and girls. Although our findings of larger effects on flow in girls may reflect biologic differences, they may also arise from sex differences in misclassification of intake. Although the validity and reproducibility of the food frequency questionnaires for children and adolescents are generally similar to those of the food frequency questionnaires for adults, the correlations of mineral intakes from the food frequency questionnaire and 24-hour diet recalls were lower in boys than in girls, suggesting that the error in measures of diet may be larger in boys than in girls (27). Because intake is measured with error, the magnitude of the flow deficits in our findings may be smaller than the true deficits associated with low magnesium and potassium intakes.

The interpretation of our findings requires consideration of the limitations in our cross-sectional study design and methods. We used a validated food frequency questionnaire designed to assess usual dietary intake in children and adolescents. The food frequency questionnaire for children and adolescents did not capture intake of minerals from other sources, such as drinking water. The mineral content of water varies with water hardness, and the intakes of magnesium, potassium, and sodium from water consumption are not accounted for in our study. Also, information on supplement use of magnesium was not available specifically for magnesium. However, the intake from supplements was likely to be low according to a study of US children who took supplements, which found a median intake of 23 mg of magnesium per day (38). We examined vitamin use but found it had little effect on the magnesium-associated deficits. Because intakes of potassium and magnesium are correlated (r = 0.98), we were unable to assess the joint effects of magnesium and potassium intakes on lung function. Our data were cross sectional, and respiratory health status may have affected dietary intake. However, deficits as were observed in children without asthma suggest that the effects of mineral intake are unlikely to be completely explained by dietary differences related to asthma. Future longitudinal analyses, as follow-up of the Children's Health Study cohort continues and additional nutrient data are collected, will allow a better assessment of the temporal relations between dietary intake and lung function.

In summary, dietary intake of magnesium was low in our cohort of children aged 11–19 years, reflecting national dietary trends. Low magnesium intake has a number of potential, but as yet undocumented, adverse health effects in children. Our findings indicate that low magnesium intake is associated with small deficits in lung function in children. Lower potassium intake may also affect lung function, but the individual effects of magnesium and potassium could not be separated because high magnesium and potassium levels occur in the same food items. Although the magnitudes of the deficits are small for individual persons, low magnesium intake is prevalent in the general population of children, and the lung function deficits associated with low magnesium may have a substantial impact on respiratory health at the population level. Based on the growing body of evidence that low magnesium intake is associated with adverse outcomes in children, consideration should be given to developing public health interventions to increase magnesium intake. Furthermore, the effects of low magnesium intake in children with asthma may be of clinical significance and warrant recommendations for increased intake of high magnesium content foods.

Reprint requests to Dr. Frank Gilliland, Department of Preventive Medicine, USC Keck School of Medicine, 1540 Alcazar Street, CHP 236, Los Angeles, CA 90033 (e-mail: gillilan@hsc.usc.edu).

This study was supported by the California Air Resources Board (contract A033-186), the National Institute of Environmental Health Sciences (grants 1P01 ES09581-05 and 5P30 ES07048-02), the US Environmental Protection Agency (grant R826708-01-0), the National Heart, Lung, and Blood Institute (grant 1 R01 HL61768-01), and the Hastings Foundation.

## REFERENCES

1.
Institute of Medicine. Dietary reference intakes for calcium, phosphorus, magnesium, vitamin D, and fluoride. Washington, DC: National Academy Press, 1999.
2.
Weiss S. Diet. In: Barnes P, Grunstein M, Leff A, et al, eds. Asthma. Philadelphia, PA: Lippincott-Raven Publishers, 1997:105–19.
3.
Britton J, Pavord I, Richards K, et al. Dietary magnesium, lung function, wheezing, and airway hyperreactivity in a random adult population sample.
Lancet

1994
;
344
:
357
–62.
4.
Burney PG, Neild JE, Twort CH, et al. Effect of changing dietary sodium on the airway response to histamine.
Thorax

1989
;
44
:
36
–41.
5.
Mathew R, Altura BM. Magnesium and the lungs.
Magnesium

1988
;
7
:
173
–87.
6.
Landon RA, Young EA. Role of magnesium in regulation of lung function.
J Am Diet Assoc

1993
;
93
:
674
–7.
7.
Zoia MC, Fanfulla F, Bruschi C, et al. Chronic respiratory symptoms, bronchial responsiveness and dietary sodium and potassium: a population-based study.
Monaldi Arch Chest Dis

1995
;
50
:
104
–8.
8.
Britton J, Pavord I, Richards K, et al. Dietary sodium intake and the risk of airway hyperreactivity in a random adult population.
Thorax

1994
;
49
:
875
–80.
9.
Spivey WH, Skobeloff EM, Levin RM. Effect of magnesium chloride on rabbit bronchial smooth muscle.
Ann Emerg Med

1990
;
19
:
1107
–12.
10.
Rude RK. Magnesium deficiency: a cause of heterogeneous disease in humans.
J Bone Miner Res

1998
;
13
:
749
–58.
11.
Weglicki WB, Dickens BF, Wagner TL, et al. Immunoregulation by neuropeptides in magnesium deficiency: ex vivo effect of enhanced substance P production on circulating T lymphocytes from magnesium-deficient mice.
Magnes Res

1996
;
9
:
3
–11.
12.
Dickens BF, Weglicki WB, Li YS, et al. Magnesium deficiency in vitro enhances free radical-induced intracellular oxidation and cytotoxicity in endothelial cells.
FEBS Lett

1992
;
311
:
187
–91.
13.
Levine BS, Coburn JW. Magnesium, the mimic/antagonist to calcium. (Editorial).
N Engl J Med

1984
;
310
:
1253
–5.
14.
Harari M, Barzillai R, Shani J. Magnesium in the management of asthma: critical review of acute and chronic treatments, and Deutsches Medizinisches Zentrum's (DMZ's) clinical experience at the Dead Sea.
J Asthma

1998
;
35
:
525
–36.
15.
Khan MA. Plasma potassium in acute severe asthma before and after treatment.
Br J Clin Pract

1989
;
43
:
363
–5.
16.
Hill J, Micklewright A, Lewis S, et al. Investigation of the effect of short-term change in dietary magnesium intake in asthma.
Eur Respir J

1997
;
10
:
2225
–9.
17.
Hill J, Lewis S, Britton J. Studies of the effects of inhaled magnesium on airway reactivity to histamine and adenosine monophosphate in asthmatic subjects.
Clin Exp Allergy

1997
;
27
:
546
–51.
18.
Hill JM, Britton J. Effect of intravenous magnesium sulphate on airway calibre and airway reactivity to histamine in asthmatic subjects.
Br J Clin Pharmacol

1996
;
42
:
629
–31.
19.
Baker JC, Tunnicliffe WS, Duncanson RC, et al. Dietary antioxidants and magnesium in type 1 brittle asthma: a case control study.
Thorax

1999
;
54
:
115
–18.
20.
Gustafson T, Boman K, Rosenhall L, et al. Skeletal muscle magnesium and potassium in asthmatics treated with oral beta 2-agonists.
Eur Respir J

1996
;
9
:
237
–40.
21.
Singhi SC, Jayashree K, Sarkar B. Hypokalaemia following nebulized salbutamol in children with acute attack of bronchial asthma.
J Paediatr Child Health

1996
;
32
:
495
–7.
22.
Millar EA, Connell JM, Thomson NC. The effect of nebulized albuterol on the activity of the renin-angiotensin system in asthma.
Chest

1997
;
111
:
71
–4.
23.
Rakhmanina NY, Kearns GL, Farrar HC 3rd. Hypokalemia in an asthmatic child from abuse of albuterol metered dose inhaler.
Pediatr Emerg Care

1998
;
14
:
145
–7.
24.
Peters JM, Avol E, Gauderman WJ, et al. A study of twelve southern California communities with differing levels and types of air pollution. II. Effects on pulmonary function.
Am J Respir Crit Care Med

1999
;
159
:
768
–75.
25.
Peters JM, Avol E, Navidi W, et al. A study of twelve southern California communities with differing levels and types of air pollution. I. Prevalence of respiratory morbidity.
Am J Respir Crit Care Med

1999
;
159
:
760
–7.
26.
Rockett HR, Breitenbach M, Frazier AL, et al. Validation of a youth/adolescent food frequency questionnaire.
Prev Med

1997
;
26
:
808
–16.
27.
Rockett HR, Colditz GA. Assessing diets of children and adolescents.
Am J Clin Nutr

1997
;
65(suppl)
:
1116S
–22S.
28.
Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a semiquantitative food frequency questionnaire.
Am J Epidemiol

1985
;
122
:
51
–65.
29.
Wypij D, Pugh M, Ware JH. Modeling pulmonary function growth with regression splines.
Statistica Sinica

1994
;
3
:
329
–50.
30.
Gold DR, Wypij D, Wang X, et al. Gender- and race-specific effects of asthma and wheeze on level and growth of lung function in children in six U.S. cities.
Am J Respir Crit Care Med

1994
;
149
:
1198
–208.
31.
Wypij D. Spline and smoothing approaches to fitting flexible models for the analysis of pulmonary function data.
Am J Respir Crit Care Med

1996
;
154
:
S223
–S8.
32.
Wang X, Wypij D, Gold DR, et al. A longitudinal study of the effects of parental smoking on pulmonary function in children 6–18 years.
Am J Respir Crit Care Med

1994
;
149
:
1420
–5.
33.
Hastie TJ, Tibshirani RJ. Generalized additive models. London, England: Chapman and Hall, 1990.
34.
Hastie TJ, Tibshirani RJ. Varying-coefficient models (with discussion).
J R Stat Soc B

1993
;
55
:
757
–96.
35.
Becker RA, Chambers JM. The new S language. Pacific Grove, CA: Wadsworth and Brooks/Cole Publishing, 1988.
36.
Briefel RR, McDowell MA, Alaimo K, et al. Total energy intake of the US population: the Third National Health and Nutrition Examination Survey, 1988–1991.
Am J Clin Nutr

1995
;
62(suppl)
:
1072S
–80S.
37.
Marier JR. Magnesium content of the food supply in the modern-day world.
Magnesium

1986
;
5
:
1
–8.
38.
Moss AJ, Levy AS, Kim I, et al. Use of vitamin and mineral supplements in the United States: current users, types of products, and nutrients. Hyattsville, MD: National Center for Health Statistics, 1989. (Advance data from vital and health statistics, no. 174).