Study cohorts
In total, 983 participants in the ‘pregnant women’ cohort and 9444 participants in the ‘women with hormone-related conditions’ cohort were eligible for analysis. Women with hormone-related conditions included the following subgroups: women receiving hormonal contraception (n=1562; among them, only 74 women were taking progestin-only formulations), women with premenstrual syndrome (n=1549), women with climacteric syndrome not receiving HRT (n=1618), women receiving HRT, including surgical menopause (n=1555), women with osteoporosis (n=1625) and women with other hormone-related conditions, such as endometriosis, polycystic ovarian disease, uterine leiomyoma, dysmenorrhea, endometrial hyperplasia (n=1535) (Figure 1).
Participants in the ‘pregnant women’ cohort had median age 28.0 (min–max 18–52), median body mass index (BMI) 23.1 kg/m2 (min–max 15.9–50.3 kg/m2) and mean (standard deviation [SD]) serum magnesium level 0.714 (0.125) mmol/L (Table 1).
Participants in the ‘women with hormone-related conditions’ cohort had median age 44.0 years (min–max 18–60 years), median BMI 25.0 kg/m2 (min–max 13.8–63.4 kg/m2) and mean (SD) serum magnesium level 0.776 (0.198) mmol/L (Table 1). Among different subgroups, women with osteoporosis, women with climacteric syndrome and women receiving HRT were on average older and had higher BMI than women in the other subgroups.
Magnesium levels and prevalence of magnesium deficiency
Participants in the ‘pregnant women’ cohort had lower mean serum magnesium levels than ‘women with hormone-related conditions’ (0.714 mmol/L [SD=0.125 mmol/L] vs 0.776 mmol/L [SD=0.198 mmol/L], p<0.0001). In the ‘women with hormone-related conditions’ cohort, the highest mean total serum magnesium level was found in the subgroup of women receiving hormonal contraception (0.789 mmol/L [SD=0.197 mmol/L]), and the lowest – in women with osteoporosis (0.758 mmol/L [SD=0.199 mmol/L]) (Table 1). The differences across subgroups were statistically significant (p<0.0001).
Prevalence of magnesium deficiency assessed by serum levels in ‘pregnant women’ cohort was 34.0% or 78.9% when using 0.66 mmol/L or 0.8 mmol/L as the cut-off, respectively (Figure 2). Prevalence of magnesium deficiency assessed by serum blood levels in ‘women with hormone-related conditions’ cohort was 24.1% or 54.8% when using 0.66 mmol/L or 0.8 mmol/L as the cut-off, respectively (Figure 2). Among all subgroups, the highest prevalence of magnesium deficiency was observed among women with osteoporosis (28.0% using cut-off <0.66 mmol/L and 58.3% using cut-off <0.8 mmol/L) and women with climacteric syndrome (27.4% and 58.4%, respectively).
Potential risk factors associated with low serum magnesium levels in pregnant women
In the ‘pregnant women’ cohort, several risk factors and comorbidities showed a statistically significant association with low serum magnesium levels (p≤0.0005; Supplementary Table 1). For the cut-off of <0.66 mmol/L, significant associations included increase in systolic blood pressure (p=0.0003), increase in diastolic blood pressure (p<0.0001), endocrine disorders (p=0.0002), previous pregnancy complications (such as placental insufficiency [p<0.0001]), and complaints (edema [p<0.0001] and pelvic girdle pain [p=0.0004]). Edema was the only risk factor significantly associated with hypomagnesemia defined by the cut-off of <0.8 mmol/L (p<0.0001) (Figure 3; Supplementary Table 1).
Potential risk factors associated with low serum magnesium levels in women with hormone-related conditions
In the ‘women with hormone-related conditions’ cohort, a large number of risk factors and comorbidities showed a statistically significant association with low serum magnesium (26 factors for cut-off <0.66 mmol/L and 38 factors for cut-off <0.8 mmol/L; Supplementary Table 2). Factors associated with both cut-offs included age and BMI (p<0.0001 for all). All studied symptoms of magnesium deficiency were significantly associated with low serum magnesium (p<0.0001 for all). Furthermore, the potential risk factors included various comorbidities and previous obstetric and gynecological conditions outlined below.
The association was significant for the following comorbidities: gastrointestinal diseases (hepatitis [p<0.0001 for both cut-offs] and cholelithiasis [p<0.0001 for both cut-offs]), urolithiasis [p<0.0001 for both cut-offs], cardiovascular diseases (pathology of heart valves [p≤0.0001 for both cut-offs]), and hypothyroidism (p<0.0001 for both cut-offs) (Supplementary Table 2).
The association was also significant for previous obstetric and gynecological conditions, such as endometriosis (p=0.0001 for both cut-offs) and complications of pregnancy and childbirth (including preeclampsia [p<0.0001 for both cut-offs] and feto-placental insufficiency [p<0.0001 for both cut-offs]), (Figure 4).
Among factors that defined the subgroups of ‘women with hormone-related conditions’, osteoporosis was significantly associated with hypomagnesemia using the cut-off of 0.66 mmol/L (p<0.0001). Associations for all risk factors and both cut-offs are listed in the Supplementary Table 2.
Magnesium supplementation
Participants of observational studies MAGIC, MAGIC2, MAGYN and MAGYN2 received magnesium supplements prescribed by treating physicians in accordance to routine clinical practice; a proportion of participants received magnesium-vitamin B6 combination (Magne B6/Magne B6 Forte) [18-21]. The effectiveness results are presented in detail elsewhere[22]. Briefly, after taking magnesium supplements for four weeks, 293 of 318 (92.1%) pregnant women with initial level <0.66 mmol/L achieved magnesium level ≥0.66 mmol/L, and 529 of 717 (73.8%) pregnant women with initial level <0.8 mmol/L achieved magnesium level ≥0.8 mmol/L. The corresponding proportions among women with hormone-related conditions were 78.4% (283 or 361 women) and 58.9% (452 of 767 women), respectively.