Diabetes modifies the association between magnesium and insomnia, but not depressive symptomatology, in Puerto Rican adults: a prospective cohort study.

BACKGROUND

Evidence on magnesium (Mg), depression, and sleep quality has been controversial.

OBJECTIVES

We aimed to examine associations of serum and dietary Mg with depressive and insomnia symptoms in a prospective longitudinal cohort.

METHODS

Adults from the Boston Puerto Rican Health Study who completed the Centre for Epidemiologic Studies Depression Scale (CES-D) assessment at any visit from baseline to the third visit (∼6-y follow-up) were included in analyses on Mg and depressive symptomatology (n = 1465 with Mg intake; n = 1116 with serum Mg). Linear mixed-effects models were used to estimate β coefficients and 95% confidence intervals (CI). Mg measures at visit 2 were used to predict incident insomnia symptoms at visit 3 (n = 613 with Mg intake; n = 587 with serum Mg; median follow-up: ∼4 y). Logistic regression was used to estimate odds ratios (ORs). Diabetes, sex, baseline Mg status, and calcium:Mg intake ratio were evaluated as modifiers.

RESULTS

No clear association was observed between Mg and CES-D score after adjustment for confounders. However, higher Mg intake from food was associated with less low positive affect, a dimension of CES-D [β (95% CI): -0.28 (-0.45, -0.11)]. Mg was not associated with insomnia in the full sample. However, higher Mg intake and supplement use were associated with lower odds of insomnia in participants without diabetes [per 100 mg/d total Mg: OR (95% CI): 0.42 (0.23, 0.79); per 100 mg/d Mg from food: 0.41 (0.21, 0.79); Mg supplement users compared with nonusers: 0.37 (0.17, 0.81)], which was not found in participants with diabetes (P-interaction < 0.05). Serum Mg was not associated with either outcome. Limited interaction was detected for other potential modifiers, despite some subgroup variation in association direction or magnitude.

CONCLUSIONS

Higher Mg intake was associated with lower odds of insomnia symptoms in Puerto Rican adults without diabetes. Further research is needed to investigate the role of diabetes in the Mg-sleep association.