Magnesium (Mg) is known to exert diverse actions on cardiometabolic function; suboptimal Mg intake and status may contribute significantly to adverse metabolic and cardiovascular health. While type 2 diabetes mellitus (T2DM) is the most common cause of Mg depletion, research gaps exist regarding the potential influence of transient states of impaired glycemia, such as gestational diabetes, on subsequent Mg biomarkers or Mg-glycemic associations in affected mothers or their offspring. Further, whether or not diabetes is an important effect modifier of associations between Mg biomarkers and cardiovascular endpoints, such as arrhythmias, is unknown. It has long been observed, however, that the determination of Mg status is riddled with challenges, and no simple, rapid, and accurate test has emerged in subsequent decades. Plasma (pMg) and serum Mg (sMg) remain the most commonly used biomarkers clinically and in the literature; comparatively little is known about erythrocyte Mg (rMg) and cardiometabolic outcomes. The Canadian Community Health Survey (2004) revealed that 45% of males aged 30–50 years and a similar percentage of women had Mg intakes below the estimated average requirement (EAR); 65% of men over 71 years of age had Mg intake below the EAR. Subpopulations living in Canada report especially low Mg intakes — among aboriginal populations, such as the James Bay Cree from Mistissini, 72.6% of adult men and a similar percentage of women had Mg intakes below the EAR. Canadian Inuit adults experiencing food insecurity also consume Mg below recommended levels. The main objectives of this thesis were to (i) determine if gestational diabetes history prospectively influences Mg concentrations or associations between Mg and glycemic variables in mothers and offspring 15-years post-partum (Montreal cohort); (ii) to examine associations between sMg, rMg, and cardiovascular risk profiles in adults from 2 ethnically distinct, cross-sectional studies (Cree, Inuit) and evaluate the utility of rMg as a cardiovascular risk associate; and (iii) to estimate the risk of an arrhythmia associated with mortality as well as the risk of ventricular premature beats across the sMg concentration gradient, and assess potential effect modification by T2DM status (Cree). Secondary data analysis was conducted on data collected from 3 diverse studies: (i) Diabetic pregnancies: Longitudinal follow-up study of mother-offspring pairs (multiethnic Montreal cohort); (ii) Nituuchischaayihtitaau Aschii: A multi-community environment and health longitudinal study in Iiyiyiu Aschii (Cree survey); (iii) International Polar Year (IPY) Inuit Health Survey (Inuit survey). Associations between Mg biomarkers and endpoints were examined in multivariate linear and logistic regression models. It was found that a gestational diabetes history of 15 years prior was associated with reduced pMg in mothers (p = 0.002) and elevated pMg in teenage offspring (p = 0.002) relative to mother and offspring controls without gestational diabetes history. Associations between Mg status and some glycemic variables (fasting glucose, insulin, and insulin sensitivity) were stronger in mothers and offspring with gestational diabetes history than those without. In Cree adults without T2DM, sMg was inversely associated with fasting glucose (p = 0.001), in addition to cardiovascular variables such as hsCRP (p = 0.038) and carotid-intima media thickness (p = 0.044). rMg was significantly associated with adiposity in both Cree and Inuit (p < 0.001), but no associations between rMg and fasting glucose, insulin, hsCRP, blood pressure, nor carotid intima-media thickness were observed. In Cree, hypomagnesaemia (sMg <0.70 mmol/L) was associated with an increased prevalence of ventricular premature beats (p < 0.05). T2DM was a significant effect modifier of the association between sMg and risk of this arrhythmia (p < 0.05). Transient states of impaired glycemia (gestational diabetes) may be associated with low pMg and its associations with glycemic outcome variables in affected mothers and offspring 15 years post-partum. Unlike sMg or pMg, there is no current evidence that total rMg is significantly associated with a favourable cardiovascular risk profile or adds value to cardiometabolic risk assessment. Prevalence of ventricular premature beats is more common in Cree adults with hypomagnesaemia and T2DM. Further investigations evaluating the potential utility and predictive value of pMg and sMg as markers of cardiometabolic risk are warranted.