Dietary Magnesium and Migraine in Adults

A Cross-sectional Analysis of the National Health and Nutrition Examination Survey 2001-2004

Margaret Slavin PhD, RDN; Huilun Li MS; Manisha Khatri BPS; Cara Frankenfeld PhD


Headache. 2021;61(2):276-286. 

In This Article


Study Population

A total of 5528 individuals aged 20- to 50-years old participated in NHANES 2001–2002 and 2003–2004 cycles. After exclusions, a total of 3626 individuals were included in the analysis of this study. A full list of exclusions and associated counts is presented in Figure 1.

Figure 1.

Inclusion and exclusion criteria


Of the included 3626 participants, 905 were classified in the migraine group, and 2721 were in the control group. The demographic characteristics of the study participants are summarized in Table 1. Mean age did not differ between the migraine and control groups, whereas significant differences were observed between groups for gender, BMI, alcohol intake, and PIR.

Magnesium Intake

Table 2 shows the mean magnesium intake by dietary or supplement source, migraine status, and gender. The overall mean dietary, supplement, and total magnesium intakes were 289.9, 28.4, and 318.3 mg/day, respectively. While a statistically significant difference in the daily dietary and total intake was found between migraine and control groups when assessing all participants (p < 0.001), females consumed less dietary magnesium than males and are known to experience migraine at higher rates. When observing intake within genders, for males there was no difference in magnesium intake from any source between migraine and control groups. For females, a statistically significant difference was detected between migraine and control groups for both dietary and total magnesium; the migraine group consumed about 20–25 mg/day less magnesium than the control group. No difference was detected in mean supplement magnesium intake in the total sample or by gender.

Fewer than one in four participants (23.6%) met their RDA through diet alone. Those who achieved their magnesium RDA through diet had 13% lower adjusted odds of migraine, which approached but did not achieve statistical significance (see Table 3). When assessing for total magnesium intake, nearly 1 in 3 participants (30.2% or) met their magnesium RDA. Rates of achievement of the RDA by diet were similar across genders (23.6% of females vs. 23.7% of males), whereas females had a slightly higher achievement of the RDA when supplements were included (31.4% vs. 29.1%). Those who achieved their magnesium RDA through a combination of diet and supplements had 17% lower odds of migraine, adjusted for potential confounders (OR = 0.83; p = 0.035).

As presented in Table 4, statistically significant differences in odds of migraine were detected between the highest and lowest Qs of magnesium consumption from both dietary and total (diet + supplement) sources in the adjusted models. The p trend was also significant for both dietary and total magnesium intake, indicating decreasing odds of migraine across increasing consumption levels of magnesium. Furthermore, in this model, no significant interaction was detected between magnesium and gender for dietary, supplemental, or total intake (p = 0.376, p = 0.619, p = 0.244, respectively).

A total of 958 individuals in the sample (495 males and 463 females), or 26% of study participants, reported taking any magnesium as a dietary supplement and were also included in the supplement intake analysis. Of those individuals taking a supplement, the average magnesium intake from supplements was 91.7 mg/day, and the average magnesium intake from supplements of the subset of adults with migraine was 89.7 mg/day. Similar rates of magnesium supplement use (26%) were seen in both the migraine and control groups.

The inclusion of all supplement users might decrease the representativeness of the total magnesium consumption because the consumption of magnesium supplements might be temporary among some users. To address this issue, an additional analysis was performed (not shown), which included only magnesium supplement users who reported taking the supplement for more than 90 days (n = 742), the typical minimum duration of magnesium prophylaxis supplement trials. In this 90-day supplement user analysis, the adjusted logistic regression p value for total magnesium intake Q4 across the overall sample decreased to 0.021 and the p-trend dropped to 0.030, but the supplement only models remained above the threshold for significance.

A total of 1637 females and 1989 males were included in the stratified gender analysis (Table 5). Significance of migraine OR was not detected in adjusted models when stratifying analysis by gender. However, a decreasing trend of OR in females for total magnesium intake that mimicked the trend in the overall analysis was noted; similarly, the p trend for this female total magnesium intake analysis neared but did not reach statistical significance.