AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
March 1991
SERUM MAGNESIUM LEVELS IN PREGNANCY AND PRETERM LABOR Richard B. Kurzel, M.D., Ph.D.
ABSTRACT
The variation of serum magnesium levels during normal pregnancy, as well as with various disease states in pregnancy, has been the subject of continued interest. Such variation, and how it relates to uterine activity and vascular reactivity, is especially of interest. The role played by magnesium in the mechanism of the contractile response of uterine and other smooth muscle is only now becoming appreciated and understood. Steer and Petrie1 demonstrated that raising serum magnesium levels serves to relax uterine smooth muscle, thereby providing the basis for the use of magnesium sulfate as a tocolytic agent. How does serum magnesium vary in states with uterine irritability such as preterm labor (PTL) and term labor, or even as one approaches term without labor? These questions, as well as how does serum magnesium vary during normal gestation, with preeclampsia, and with diabetes in pregnancy, are the objects of this study.
Downloaded by: NYU. Copyrighted material.
Pregnancy is marked by a state of hypomagnesemia. The serum magnesium level shows no gestational dependence (mean, 1.79 ± 0.44 mg/dl) until 33 weeks, at which point it continuously declines. Serum magnesium is not depressed further with the onset of labor at term. Patients in preterm labor have a significantly depressed serum magnesium level (mean, 1.60 ± 0.46 mg/dl; 21 to 33 weeks; p < 0.0005). This level was not dependent on whether the etiology for the preterm labor was premature rupture of the membranes (PROM), twin gestation, abruption, placenta previa with bleeding, orchorioamnionitis. With PROM, the serum magnesium level was not depressed prior to the initiation of preterm labor. However, observation of hypomagnesemia for this and other etiologies just prior to the initiation of preterm labor were not available. Possible mechanisms by which hypomagnesemia induces uterine irritability are explored, including inhibition of adenyl cyclase with resultant increase in cytoplasmic calcium levels. Patients with diabetes mellitus appeared to have slightly reduced serum magnesium levels, but the results were not statistically significant. Magnesium levels in patients with preeclampsia were not significantly different from controls. Hypomagnesemia (magnesium 1.4 mg/dl or less) may be a marker for true preterm labor.
MATERIALS AND METHODS
Over the period from April 1, 1988, through April 1989, venous serum samples were obtained from pregnant patients consenting to participate in this study. Serum samples were assayed for total magnesium, total calcium, phosphate, and albumin. In the normal control group of 128 patients 154 samples were obtained at different points in gestation. Gestations in this group ranged from 10 through 41 weeks, as determined by the last menstrual period and confirmed by ultrasound examination of each patient. Excluded from this group and from the study were patients with known renal disease or serum creatinine levels 0.9 mg/dl or higher, diabetics (insulin and noninsulin requiring; see later), patients with endocrinopathies such as hypoparathyroidism, and patients with known malabsorption disorders.
Department of Obstetrics and Gynecology, St. Louis University School of Medicine, St. Mary's Health Center, Division of Maternal Fetal Medicine, St. Louis, Missouri Reprint requests: Dr. Kurzel, Department of Obstetrics and Gynecology, St. Mary's Health Center, 6420 Clayton Rd., St. Louis, MO 63117 Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
119
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
120
tant ethylene glycol-bis-N,N'-tetraacetic acid and potassium cyanide. Absorption at 548 nm was determined on the Abbott Spectrum System. Total serum calcium was determined by the reaction with o-cresolphthalein complexon in alkaline solution, eliminating interference by magnesium with 8-quinolinol sulfate. Absorption was measured by bichromatic analysis on the Abbott Spectrum System at 572/660 nm. Serum phosphate was determined by the reaction with molybdate in acid solution; absorption was measured at 340/380 nm. Serum albumin was determined by the reaction with bromcresol purple; absorption was measured at 600 to 610 nm. Serum creatinine was determined by the reaction with alkaline picrate; absorption was measured by bichromatic analysis at 516/604 nm. Statistical analysis utilized the paired Student's t test. RESULTS
For normal pregnant patients not in labor, the mean serum magnesium was calculated for each week of gestation and is plotted in Figure 1 with its range of values. It was found that serum magnesium showed a linear relationship without gestational dependence up until 33 weeks' gestation, with a mean magnesium of 1.79 ± 0.44 mg/dl (range, 1.4 to 2.3 mg/dl). From 34 to 42 weeks, magnesium concentrations showed a continuously decreasing trend, down to 1.57 mg/dl at 42 weeks. For patients at term (37 to 42 weeks) not in labor, the mean serum magnesium of 1.68 ± 0.55 mg/dl was significantly lower than values at 21 to 33 weeks (p < 0.025). For patients at term in labor compared with those at term not in labor, no significant difference in values could be seen (see Fig. 1). Patients from 21 to 33 weeks in PTL were compared to the normal controls over the same gestations. The mean magnesium value for patients in PTL was 1.60 ± 0.46 mg/dl (range, 1.1 to 2.3 mg/dl) and was significantly different from the normal controls (p < 0.0005). Similarly over the same gestational span, patients with known etiologies for PTL were compared with the normal controls. Patients with PROM in labor had a mean magnesium (1.57 ± 0.59 mg/dl; range, 1.1 to 2.2 mg/dl) significantly lower than that for patients with PROM without labor (1.82 ± 0.34 mg/dl; range, 1.4 to 2.1 mg/dl) (p < 0.005). The values for patients with PROM without labor was not significantly different from those of the normal controls (p > 0.25). Patients with twins in PTL (mean magnesium level of 1.56 ± 0.46 mg/dl; range, 1.2 to 2.0 mg/dl) and patients with either abruption or placenta previa in PTL (mean magnesium level of 1.57 ± 0.21 mg/dl; range, 1.5 to 1.8 mg/ dl) had means significantly lower than the normal controls (p < 0.005 for both). When patients in PTL with PROM, twins, or bleeding disorders were compared, their mean magnesium levels were not significantly different (p > 0.25) (Figs. 2, 3). Since all
Downloaded by: NYU. Copyrighted material.
Patients with known diabetes mellitus of any Priscilla White class were considered separately (16 patients were in class A, six in B, four in C, one in D, two in F, and four in R), with 23 determinations in the non-insulin-requiring group, and 20 in the insulin-dependent group (IDDM). Gestations ranged from 9 to 41 weeks. The 23 patients with preeclampsia or eclampsia were considered as a separate group. Their gestational ages ranged from 24 to 40 weeks, and the diagnosis was made by standard criteria.2 Patients with premature rupture of the membranes (PROM) were considered separately as two groups: those without and those with uterine irritability or contractions. Twenty-six patients from 22 to 36 weeks' gestation constituted the first group. Only patients who had absent contractions or no evidence of uterine activity for at least 48 hours after the serum sample was obtained were entered into this group. Uterine activity was considered to include patients who perceived cramping or contractions, as well as those who had contractions that were palpable or could be recorded on an external tocodynamometer of frequency six or more per hour or who had cervical effacement and dilation on examination. Sixteen patients with PROM and uterine activity constituted the second group (21 to 35 weeks' gestation). PROM was considered gross rupture of the membranes confirmed by positive Nitrazine and fern testing. Serum samples were likewise obtained from 71 patients in PTL (gestations 21 to 36 weeks). PTL was defined as uterine contractions with accompanied cervical effacement or dilation, or regular uterine contractions of frequency six or more per hour. Included in this group were the 16 patients with PROM, three with abruptions, four with placenta previa, eight had twins, one had a cervical circlage, and three had documented chorioamnionitis. The remaining patients in this group had undetermined etiologies for their PTL. All of these patients were transferred in from other hospitals to our tertiary care facility. Patients were excluded from this study if they had received intravenous magnesium sulfate tocolysis prior to arrival at our institution. Thirteen normal term patients (37 to 42 weeks) in active labor constituted another study group. All patients in the study had taken iron and prenatal vitamin supplements.3 Patients were excluded from the study if they had taken oral magnesium or calcium supplements, diuretics, or had aminoglycoside therapy.4 Likewise excluded from the preeclamptic, PROM, and PTL study groups were patients with diabetes, renal disease, and hypoparathyroidism. Total serum calcium, total magnesium, phosphate, albumin, and creatinine were determined by colorimetric methods and quantitated absorptiometrically. Total serum magnesium was determined by the reaction with the metallochrome dye calmagite, in an alkaline solution, eliminating interference with calcium and heavy metals by including the surfac-
March 1991
MAGNESIUM LEVELS IN PREGNANCY/Kurzel ° •
NORMAL/NOT IN LABOR TERM/LABOR
IT 2.5-
i
2.0-
III
I
IT
IT
,r-J-
U4 -1-
I
1.5-
u
1.0
20
40
30
Downloaded by: NYU. Copyrighted material.
10
GESTATION ( weeks ) Figure 1. Serum magnesium concentrations (mg/dl) for normal pregnant women, o = not in labor, and * = term patients (37 to 42 weeks) in labor. Plotted are mean values for each week of gestation and the range of values. There is no gestational dependence up until 33 weeks, at which point magnesium levels drop.
PRE
2.5-
TERM LABOR
-
o • •
•
•
*
•o M
E
*
2.0 —
a o
0
•
TOTAL PROM BLEED TWINS CHORIO NO CVX CHANGE
2 u
z o
o • o*
• •
o • * o o
1.5 — - 8 — • ¥* •
0 -
*
—
—
•
•
' o *
o
o o .—o-• * -fcT*< a
—
0
o
*
0 D
* •
20
1
1 30
GESTATION
40
( weeks )
Figure 2. Serum magnesium levels for patients in PTL. Mean values for each week gestation are shown, and the least squares plot describing these data (y = 1.25 + 0.0114 x). Shown are the means for total patients in PTL(o), patients with PROM in labor (*), those with PTL due to abruption and placenta previa (•), those with twin gestation in PTL {+), those with chorioamnionitis in PTL (•), and those in PTL of unknown etiology and with no cervical change (0). Irrespective of etiology, all points cluster about the same curve.
121
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
March 1991
° PROM NOT IN LABOR • PROM IN LABOR 2.5 •o
2.0 U
z o 1.5 OS bJ
40
20
GESTATION
(
weeks
)
Figure 3. Serum magnesium levels for patients whose PTL is due to PROM. Shown are the mean values at given gestations for patients with PROM without labor (o), and the least squares curve describing these data (top curve: y = 1.45 + 0.0125x). Also shown are the means for patients with PROM in PTL (*) and the corresponding least squares plot (lowest curve: y = 1.20 + 0.012x). Given for reference is the curve for normal controls, not in labor (bold curve).
patients with PTL were transferred from other hospitals, with the exception of those with PROM, all PTL patients were seen here already with uterine irritability or contractions. We therefore could not observe patients temporally prior to the development of uterine contractions to determine whether magnesium levels were depressed prior to development of PTL, as was suggested by Martin and Morrison.5 For the special case of PROM, we observed this not to be the case. More work needs to be done to study this point, since it may have practical bearing on the issue of oral magnesium prophylaxis for patients at risk of preterm delivery. Patients with non-insulin-requiring diabetes mellitus (mean magnesium level of 1.75 ± 0.42 mg/ dl; range, 1.4 to 2.2 mg/dl) and IDDM of all White classes (mean magnesium level of 1.75 ± 0.51 mg/dl; range, 1.4 to 2.4 mg/dl) had means that appeared to have a tendency to be lower than the controls, but the difference was not significant (p = 0.25) (Fig. 4). Patients with preeclampsia or eclampsia were compared to normal controls for the gestational spans 24 to 33 weeks and 37 to 42 weeks. For 24 to 33 weeks, patients with preeclampsia had a mean magnesium level of 1.72 ± 0.41 mg/dl (range, 1.4 to 2.0 mg/dl), which was not significantly different from the controls (p > 0.1). At 37 to 42 weeks' gestation, the mean magnesium of 1.70 ± 0.60 mg/dl (range, 122 1.4 to 2.3 mg/dl) for patients with preeclampsia was
also not significantly different from controls (p > 0.25), whose values were also dropping over this gestational span (Fig. 5). Total serum magnesium did not have a significant dependence on total serum calcium, phosphate, or albumin, as shown by the scatter plots in Figures 6 and 7. Total serum calcium was found to decline slowly to a minimum at 34 weeks, and then rise again approaching term, as has previously been reported by Bodansky and Duff,6 and Pitkin et al7 For 21 to 33 weeks' gestation, the mean total serum calcium was 9.16 ± 0.50 mg/dl compared with 9.14 ± 0.37 mg/dl for patients at this gestation in PTL, which is not statistically different (p > 0.25). Similarly, the mean total calcium for patients not in labor at 37 to 42 weeks was 9.17 ± 0.84 mg/dl compared with 9.14 ± 1.06 mg/dl for term patients in labor, which is not statistically significant (p > 0.25). DISCUSSION
The average serum level of magnesium in nonpregnant healthy women of reproductive age is 2.4 ± 0.49 mg/dl.5 That pregnancy is marked by a state of hypomagnesemia was noted by Hall in 1957.8 Numerous studies have since confirmed this finding, although the exact gestational dependence of serum magnesium is still unsettled, with various studies
Downloaded by: NYU. Copyrighted material.
1.0
MAGNESIUM LEVELS IN PREGNANCY/Kurzel
DIABETES CLASS
A B C D,
2.5
o » a F, R
•
2.01
*
•
o oo a
1.5
o
o
1.0' 30
20
GE STAT I O N
(
40
w e e k s
)
Figure 4. Serum magnesium levels for pregnant diabetics, by the White classification. Given are the mean values at each gestation (class A = o, B = *, C = •, D, F, R = •), and the curve for normal controls not in labor is presented for reference. No significant difference in levels was found for either non-insulin requiring or IDDM patients relative to normal controls.
2.5-
PREECLAMPSIA
\
2.0o o
o
o
o
Z
o o
o
1.5 UJ
1.0 20
I 30
GESTATION
40
( weeks )
Figure 5. Mean serum magnesium levels for patients with preeclampsia who were not in labor (o). The curve for normal controls not in labor is shown for reference. No statistically significant difference could be seen between the two groups for the entire span of gestations. 123
Downloaded by: NYU. Copyrighted material.
10
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
2.5-
2.5-
2.0-
2.0
Mg
March 1991
-
1.5-
1.5 -
1.0-
1.0-
3.0
2.0
PO
I 4.0
I 5.0
2.0
4.0
3.0
ALB
Figure 6. A: Scatter plot for total serum magnesium versus serum phosphate for normal pregnant patients not in labor (gestations, 10 to 42 weeks). Nocorrelation isseen(r2 = 0.097). B: Scatter plotfortotal serum magnesium versus serum albumin for normal pregnant patients not in labor (gestations, 10 to 42 weeks). No correlation is seen (r2 = 0.047).
-
NOT IN LABOR
-
PRETERM LABOR ;
10.0 —
10.0-
bis
• 9.0 -
9.0-
2
u < o
•
8.0-
8.0 -
7.0-
7.0-
U
4 A
2.0
1.0
"
1
1
2.0
SERUM MAGNESIUM ( mg/dl ) Figure 7. Scatter plots for total serum calcium versus total serum magnesium for patients not in labor (21 to 33 weeks) 124 (left), and for those in preterm labor (right).
Downloaded by: NYU. Copyrighted material.
Mg
MAGNESIUM LEVELS IN PREGNANCY/Kurzel
gestation and not in labor, with which to compare this mean level. Similarly, Elliott et al20 noted that patients in PTL in his study had a mean serum magnesium level of 1.80 ± 0.04 mg/dl. Again there was no control group to compare these levels to, and the gestational dependence was not stated. Along these lines, hypomagnesemia accompanying PTL was implied but not demonstrated by three clinical studies: an epidemiologic study by Kuti,21 and double-blinded clinical trials by Conradt et al22 and Spading et al.23 These three studies demonstrated that prophylactic oral magnesium supplementation to patients at risk for preterm delivery was successful in lowering the preterm delivery rate. Abnormally low serum magnesium with PTL was not documented, but nonetheless these authors proposed that a certain percentage of preterm deliveries could be attributed to a magnesium deficiency. In a Russian study, Penev et al24 studied 201 patients in PTL and noted low serum magnesium levels in 21.9% of their patients. In our study, from 21 to 33 weeks' gestation, a statistically significant reduction in serum magnesium was seen in patients with PTL, which appears to be independent of the etiology for the PTL. The depressed serum magnesium level therefore reflects the tendency for PTL, or the initiation of PTL, and does not characterize the particular predisposing factor for PTL. The number of patients in each of these individual groups is small, however, and this question warrants further study in the future. The question, which is not answered by this study, is whether hypomagnesemia precedes the development of uterine irritability, as is the case for our normal patients prior to labor. What is needed to answer this question is a longitudinal study of patients at risk for PTL. Interestingly for the particular subgroup with PROM, magnesium levels for those patients not in PTL were not different from the normal controls at matched gestations. Magnesium levels dropped, however, only once PTL ensued. Whether magnesium levels drop in the period immediately preceding the development of PTL is unknown. Hypomagnesemia therefore may either be a causative factor or simply reflect the process involved with the development of uterine irritability for both term labor and PTL. Evidence has accumulated in recent years from in vitro studies to explain the connection between hypomagnesemia and the development of uterine irritability. Reduced extracellular magnesium levels have been shown to increase smooth muscle contractility in vitro. Altura et al25 and Sjogren and Edvinsson26 have demonstrated that extracellular magnesium influences the release of calcium from intracellular depots; a decrease in extracellular magnesium results in an increase in the total exchangeable and intracellular calcium fractions, with a resultant increase in smooth muscle tension. In addition, Somlyo et al27 showed that decreased extracellular magnesium can result in depletion of the magnesium content of smooth muscle cells. Magnesium is a required cofactor in the myometrial contractile re-
Downloaded by: NYU. Copyrighted material.
giving different trends. Hall8 noted that serum magnesium progressively declined during pregnancy, reaching a nadir at 34 to 38 weeks, and then began to rise approaching term. Pitkin et al,9 De Jorge et al,10 and Sheldon et al3 described a similar trend, but the nadir was at 28 to 32 weeks. Pitkin's9 curve showed a very shallow decline before rising approaching term. De Jorge's10 curve rose from the nadir at 28 to 32 weeks, peaked at 36 weeks, and then declined again in the last month approaching term. By contrast, Roelofsen et al11 found serum magnesium to drop and reach a steady state in the second trimester, only to remain constant until term. Almuna et al12 observed no variation at all in magnesium levels from 28 to 42 weeks (mean 1.74 ± 0.18 mg/dl). The results of our study showed a reduced magnesium level in pregnancy (mean 1.79 ± 0.44 mg/dl) showing no gestational dependence until 33 weeks. Beyond 33 weeks, magnesium concentrations progressively declined (Fig. 1). This is in agreement with the trend noted by Olatunbosun et al,13 Bartl and Riss,14 Rusu etal, 15 Ray etal, 16 and Martin and Morrison.5 Kolasa et al17 also noted a progressive decline post-term compared with levels at term. For patients in labor at term, Bogert and Plass18 as early as 1923 suggested that there "seems to be a tendency for the serum magnesium to be lowered in labor." Although the mean serum magnesium was 2.1 mg/dl, they noted that 65% of the values were under 2.0 mg/dl (range, 1.4 to 3.2 mg/dl). Results of Rusu et al15 and Ray et al16 seem to agree with this finding, but plots of values relative to gestations were not given. The study of Olatunbosun et al13 actually found levels for patients in labor to be higher than for those at term not in labor. Our results do not confirm the findings of Bogert and Plass.18 It would appear that serum magnesium levels for term patients in labor (37 to 42 weeks) are not significantly different from the values of term patients not in labor (see Fig. 1). Serum magnesium therefore drops as a prodrome preceding labor, but is not lowered further by labor. The reason for hypomagnesemia during pregnancy is unknown, but it is unlikely to be explained solely by hemodilution, as suggested by Hall.8 De Jorge et al10 and Sheldon et al3 have demonstrated that by taking into consideration plasma volume expansion, total circulating magnesium may be calculated and shown to increase steadily through the pregnancy until the last month, when it drops. Since magnesium may exert a smooth muscle relaxing effect, this progressive rise may exert a quieting effect on the myometrium, while the drop in levels in the 1 to 2 months preceding term may contribute to increased uterine irritability. What initiates this drop in serum magnesium concentration from 33 weeks onward is unknown. If a drop in serum magnesium levels is a prodrome to the initiation of term labor, might the same be also true for PTL? In a small study of only 17 patients in PTL, Martin et al19 found that the mean serum magnesium level was 1.44 ± 0.22 mg/dl. There was no control group of patients matched for
125
sponse,28 and low extracellular magnesium levels may result in increased contractility in any or all of the following ways: 1. Magnesium is an activator of adenylate cyclase. Reduced magnesium correlates with decreased synthesis of cyclic adenosine monophosphyate (c-AMP), and thereby reduced c-AMP-mediated calcium sequestration.25 Cytoplasmic concentration of calcium is therefore increased, resulting in muscle contraction. In addition, decreased c-AMP-dependent protein kinase activity would contribute toward increased myosin light chain kinase activity and uterine contractility. 2. Cytoplasmic calcium concentration is also controlled by the sarcoplasmic reticulum. The magnesium-adenosine triphosphate (ATP) complex acts as the substrate for calcium-dependent ATPase, stimulating calcium pump activity. Reduced magnesium results in inhibition of the calcium pump, increased cytoplasmic calcium, and increased myocyte contractility.25-29 3. Magnesium may act on the cell membrane to regulate entry and exit of calcium into the cell, by influencing membrane permeability. Calcium and magnesium may share the same carrier on the membrane and compete for transport, hence affect cytoplasmic concentrations. This transport is also c-AMP dependent, acting via the Na+/K+ pump.25-30 4. Magnesium-ATP is bound to the myosin light chains to weaken the actin-myosin interaction. Decreased magnesium concentration may therefore favor contraction.31^33 5. Magnesium can bind to calmodulin and may have an effect on the calcium-calmodulin complex.^ As noted earlier, Conradt et al22 and Spading et 23 al have attributed PTL to a magnesium deficiency. This is unlikely. Hsueh and Zuspan35 in reporting a case of hypomagnesemia in pregnancy secondary to nutritional causes noted that true magnesium deficiency is rare (whereas PTL is not). During pregnancy, hypomagnesemia is most commonly seen with diabetes mellitus, followed by renal insufficiency, alcoholism, and with hypoparathyroidism. Low serum magnesium levels with diabetes have been reported by Cruikshank et al,36 Durlach et al,37 and Mimouni et al38 In our study where the patients were in good glycemic control, a tendency for lower magnesium levels was seen for all diabetics, but did not reach statistical significance. Cruikshank et al36 and Durlach et al37 have shown that the lowest magnesium levels correlated with increasing severity of diabetes (although the magnesium levels were reduced by only approximately 4% for IDDM). Mimouni et al38 explained this by increased urinary magnesium losses with severe uncontrolled diabetes, versus an effect of insulin itself on lowering the serum magnesium. Interestingly, patients with IDDM are no more prone to PTL in spite of their hypomagnesemia, and only one patient with IDDM in this 126 study was noted in PTL.
March 1991
Similarly, patients with renal insufficiency were excluded from this study. The kidney is the major regulator of serum magnesium, and with renal insufficiency, urinary magnesium losses result in hypomagnesemia.39 Finally, patients with preeclampsia were studied as a separate group. Speculation40 on the occurrence of hypomagnesemia with preeclampsia is based on such observations by Kolasa et al17 and Altura et al25 in patients near term or at term. Hypomagnesemia has been hypothesized to play a role in the increased vascular reactivity seen in preeclampsia. Since preeclampsia is most common as term is approached, it occurs when serum magnesium levels are normally declining. Our results do not show a tendency for patients with preeclampsia to have serum magnesium levels different from normal. This is in agreement with the findings of Roelofsen et al,11 Almuna et al,i2 and Hall.8 It is interesting that in 1939 Bodansky and Duff6 noted that hypocalcemia developed preceding labor in the rat. We and others 79 have shown that this does not hold true at term or with PTL in humans, where serum calcium is stringently regulated by parathyroid hormone. Instead, hypomagnesemia is seen, which in ways still to be determined, reflects a process resulting in increased uterine irritability/ contractility. A potential use of these results may be the determination of a marker to differentiate patients truly in PTL, and those patients at greatest risk for progressing on to preterm delivery. For certain from 21 to 33 weeks gestation, patients with a serum magnesium 1.4 mg/dl or less (^ 2 SD below the mean) should be viewed as being at risk for a preterm delivery.
REFERENCES
1. Steer CM, Petrie RH: A comparison of MgSO4 and alcohol for the prevention of premature labor. Am J Obstet Gynecol 129:1-4, 1977 2. Pritchard JA, MacDonald P, Gant NF: Williams Obstetrics. Norwalk, CT: Appleton-Century-Crofts, 1985, pp 525-527 3. Sheldon WL, Aspillaga MO, Smith PA, Lind T: The effect of oral iron supplementation on zinc and magnesium levels during pregnancy. Br J Obstet Gynaecol 92:892-898, 1985 4. Dirks JH, Wong NL: Renal magnesium wasting disorders. Adv Exp Med Biol 208:193-197, 1986 5. Martin RW, Morrison JC: Oral magnesium for tocolysis. Contemp OB/GYN 30:111-118, 1987 6. Bodansky M, Duff VB: Regulation of the level of calcium in the serum during pregnancy. JAMA 112:223-229, 1939 7. Pitkin RM: Calcium metabolism in pregnancy: A review. Am J Obstet Gynecol 121:724-737, 1975 8. Hall DG: Serum magnesium in pregnancy. Obstet Gynecol 9:158-162, 1957 9. Pitkin RM, Reynolds WA, Williams GA, Hargis GK: Calcium metabolism in normal pregnancy: A longitudinal study. Am J Obstet Gynecol 133:781-790, 1979 10. De Jorge FB, Delascio D, de Ulhoa Cintra AB, Antunes ML: Magnesium concentration in the blood serum of normal pregnant women. Obstet Gynecol 25:253-254, 1965 11. Roelofsen JMT, Berkel GM, Uttendorfsky OT, Siegers JFG: Urinary excretion rates of calcium and magnesium in
Downloaded by: NYU. Copyrighted material.
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
MAGNESIUM LEVELS IN PREGNANCY/Kurzel
13.
14. 15. 16. 17.
18. 19. 20. 21.
22.
23. 24.
25.
26.
27. 28.
29. 30. 31. 32. 33. 34. 35. 36.
37. 38.
39. 40.
channel blockers on myogenic tone and agonist-induced responsiveness of blood vessels. Can J Physiol Pharmacol 65:729-745, 1987 Sjogren A, Edvinsson L: The influence of magnesium on the release of calcium from intracellular depots in vascular smooth muscle cells. Pharmacol Toxicol 62:17—21, 1988 Somlyo AP, Somlyo AV: Effects and subcellular distribution of magnesium in smooth and striated muscle. Fed Proc 40:2667-2671, 1981 D'Ocon MP, Anselmi E, Villar A: Effects of MgCl2 on the contractile response of uterus to several agonists in calcium-free solution. J Pharm Pharmacol 39:444—448, 1987 Watras J: Effects of Mg+2 on calcium accumulation by two fractions of sarcoplasmic reticulum from rabbit skeletal muscle. Biochim Biophys Acta 812:333-344, 1985 Flatman PW: Magnesium transport across cell membranes. J Membrane Biol 80:1-14, 1984 Cooke R, Pate E: The effects of pyrophosphate on myosin heads and on fibre stiffness. Adv Exp Med Biol 226:255265, 1988 Davis JS: Myosin heads in muscle thick filament assembly. Nature 333:807, 1988 Potter JD, Robertson SP, Johnson JD: Magnesium and the regulation of muscle contraction. Fed Proc 40:26532656, 1981 Tsai MD, Drakenberg T, Thulin E, Forsen S: Is the binding of Mg(II) to calmodulin significant? An investigation by Mg-25 NMR. Biochemistry 26:3635-3643, 1987 Hseuh GSC, Zuspan FP: Magnesium deficiency in pregnancy: A case report. J Reprod Med 27:479-483, 1982 Cruikshank DP, Pitkin RM, Reynolds WA, Williams GA, Hargis GK: Altered maternal calcium homeostasis in diabetic pregnancy. J Clin Endocrinol Metab 50:264— 267, 1980 Durlach J, Collery P: Magnesium and potassium in diabetes and carbohydrate metabolism. Magnesium 3:315—323, 1984 Mimouni F, Miodovnik M, Tsang RC, Holroyde J, Dignan PS, Siddiqi TA: Decreased maternal serum magnesium concentration and adverse fetal outcome in insulindependent diabetic women. Obstet Gynecol 70:85—88, 1987 Agus, ZS, Wasserstein A, Goldfarb S: Disorders of calcium and magnesium homeostasis. Am J Med 72:473—488, 1982 Conradt A: New concepts in the pathogenesis of gestosis under special consideration of magnesium deficiency. Z Geburtshilfe Perinatol 188:49-58, 1984
Downloaded by: NYU. Copyrighted material.
12.
normal and complicated pregnancies. Eur J Obstet Gynecol Reprod Biol 27:227-236, 1988 Almuna RV, Trivino XB, Radrigan GA, Moreno EG, Herrera RZ, Espinoza IC, Galindo ES, Aguilar PO: Comportamiento del Magnesio en embarazadas. I. Estudio comparativo entre embarazadas normales y con patologia. Rev Chil Obstet Ginecol 48:65-70, 1983 Olatunbosun DA, Adeniyi FA, Adadevoh BK: Serum calcium, phosphorus, and magnesium levels in pregnant and non-pregnant Nigerians. Br } Obstet Gynaecol 82: 568-571, 1975 Bartl W, Riss P: Magnesium serum levels and magnesium deficiency in pregnancy. Z Geburtshille Perinatol 186: 335-337, 1982 Rusu O, Lupan C, Baltescu V: Magneziul seric in sarcina normala la termen si nasterea prematura. Obstet Ginecol 14:215-219, 1966 Ray KK, De MK, Roy N: Serum magnesium in different trimesters of pregnancy. J Indian Med Assoc 72:28—32, 1979 Kolasa F, Krajewski J, Szpakowski M: Serum magnesium levels in pregnant women with symptoms of EPH-Gestosis and with post-mature pregnancy. Wiad Lek 36:897— 899, 1983 Bogert LJ, Plass ED: Placental transmission: I. The calcium and magnesium content of fetal and maternal blood serum. J Biol Chem 56:297-307, 1923 Martin RW, Gaddy DK, Martin JN Jr, Lucas JA, Wiser WL, Morrison JC: Tocolysis with oral magnesium. Am J Obstet Gynecol 156:433-434, 1987 Elliott JP: Magnesium sulfate as a tocolytic agent. Am J Obstet Gynecol 147:277-284, 1983 Kuti V, Balazs M, Morvay F, Varenka Z, Szekely A, Szucs N: Effect of maternal magnesium supply on spontaneous abortion and premature birth and on intrauterine fetal development: Experimental epidemiological study. Magnesium Bull 3:73-79, 1981 Conradt A, Weidinger H, Algayer H: Magnesium therapy decreased the rate of intrauterine fetal retardation, PROM, and premature delivery in risk pregnancies treated with betamimetics. Magnesium 4:20-28, 1985 Spatling L, Spading G: Magnesium supplementation in pregnancy. A double blind study. Br J Obstet Gynaecol 95:120-125, 1988 Penev IV, Ruseva R, Pavlova F, Naumov IV, Stereva K, Serdev V: Threatened preterm delivery at low attached placenta and magnesium insufficiency. Akush Ginekol (Sofiia) 23: 97-102, 1984 Altura BM, Altura BT, Carella A, Gebrewold A, Murakawa T, Nishio A: Mg+2 —Ca+2 interaction in contractility of vascular smooth muscle: Mg+2 versus organic calcium
127
March 1991
SERUM MAGNESIUM LEVELS IN PREGNANCY AND PRETERM LABOR Richard B. Kurzel, M.D., Ph.D.
ABSTRACT
The variation of serum magnesium levels during normal pregnancy, as well as with various disease states in pregnancy, has been the subject of continued interest. Such variation, and how it relates to uterine activity and vascular reactivity, is especially of interest. The role played by magnesium in the mechanism of the contractile response of uterine and other smooth muscle is only now becoming appreciated and understood. Steer and Petrie1 demonstrated that raising serum magnesium levels serves to relax uterine smooth muscle, thereby providing the basis for the use of magnesium sulfate as a tocolytic agent. How does serum magnesium vary in states with uterine irritability such as preterm labor (PTL) and term labor, or even as one approaches term without labor? These questions, as well as how does serum magnesium vary during normal gestation, with preeclampsia, and with diabetes in pregnancy, are the objects of this study.
Downloaded by: NYU. Copyrighted material.
Pregnancy is marked by a state of hypomagnesemia. The serum magnesium level shows no gestational dependence (mean, 1.79 ± 0.44 mg/dl) until 33 weeks, at which point it continuously declines. Serum magnesium is not depressed further with the onset of labor at term. Patients in preterm labor have a significantly depressed serum magnesium level (mean, 1.60 ± 0.46 mg/dl; 21 to 33 weeks; p < 0.0005). This level was not dependent on whether the etiology for the preterm labor was premature rupture of the membranes (PROM), twin gestation, abruption, placenta previa with bleeding, orchorioamnionitis. With PROM, the serum magnesium level was not depressed prior to the initiation of preterm labor. However, observation of hypomagnesemia for this and other etiologies just prior to the initiation of preterm labor were not available. Possible mechanisms by which hypomagnesemia induces uterine irritability are explored, including inhibition of adenyl cyclase with resultant increase in cytoplasmic calcium levels. Patients with diabetes mellitus appeared to have slightly reduced serum magnesium levels, but the results were not statistically significant. Magnesium levels in patients with preeclampsia were not significantly different from controls. Hypomagnesemia (magnesium 1.4 mg/dl or less) may be a marker for true preterm labor.
MATERIALS AND METHODS
Over the period from April 1, 1988, through April 1989, venous serum samples were obtained from pregnant patients consenting to participate in this study. Serum samples were assayed for total magnesium, total calcium, phosphate, and albumin. In the normal control group of 128 patients 154 samples were obtained at different points in gestation. Gestations in this group ranged from 10 through 41 weeks, as determined by the last menstrual period and confirmed by ultrasound examination of each patient. Excluded from this group and from the study were patients with known renal disease or serum creatinine levels 0.9 mg/dl or higher, diabetics (insulin and noninsulin requiring; see later), patients with endocrinopathies such as hypoparathyroidism, and patients with known malabsorption disorders.
Department of Obstetrics and Gynecology, St. Louis University School of Medicine, St. Mary's Health Center, Division of Maternal Fetal Medicine, St. Louis, Missouri Reprint requests: Dr. Kurzel, Department of Obstetrics and Gynecology, St. Mary's Health Center, 6420 Clayton Rd., St. Louis, MO 63117 Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
119
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
120
tant ethylene glycol-bis-N,N'-tetraacetic acid and potassium cyanide. Absorption at 548 nm was determined on the Abbott Spectrum System. Total serum calcium was determined by the reaction with o-cresolphthalein complexon in alkaline solution, eliminating interference by magnesium with 8-quinolinol sulfate. Absorption was measured by bichromatic analysis on the Abbott Spectrum System at 572/660 nm. Serum phosphate was determined by the reaction with molybdate in acid solution; absorption was measured at 340/380 nm. Serum albumin was determined by the reaction with bromcresol purple; absorption was measured at 600 to 610 nm. Serum creatinine was determined by the reaction with alkaline picrate; absorption was measured by bichromatic analysis at 516/604 nm. Statistical analysis utilized the paired Student's t test. RESULTS
For normal pregnant patients not in labor, the mean serum magnesium was calculated for each week of gestation and is plotted in Figure 1 with its range of values. It was found that serum magnesium showed a linear relationship without gestational dependence up until 33 weeks' gestation, with a mean magnesium of 1.79 ± 0.44 mg/dl (range, 1.4 to 2.3 mg/dl). From 34 to 42 weeks, magnesium concentrations showed a continuously decreasing trend, down to 1.57 mg/dl at 42 weeks. For patients at term (37 to 42 weeks) not in labor, the mean serum magnesium of 1.68 ± 0.55 mg/dl was significantly lower than values at 21 to 33 weeks (p < 0.025). For patients at term in labor compared with those at term not in labor, no significant difference in values could be seen (see Fig. 1). Patients from 21 to 33 weeks in PTL were compared to the normal controls over the same gestations. The mean magnesium value for patients in PTL was 1.60 ± 0.46 mg/dl (range, 1.1 to 2.3 mg/dl) and was significantly different from the normal controls (p < 0.0005). Similarly over the same gestational span, patients with known etiologies for PTL were compared with the normal controls. Patients with PROM in labor had a mean magnesium (1.57 ± 0.59 mg/dl; range, 1.1 to 2.2 mg/dl) significantly lower than that for patients with PROM without labor (1.82 ± 0.34 mg/dl; range, 1.4 to 2.1 mg/dl) (p < 0.005). The values for patients with PROM without labor was not significantly different from those of the normal controls (p > 0.25). Patients with twins in PTL (mean magnesium level of 1.56 ± 0.46 mg/dl; range, 1.2 to 2.0 mg/dl) and patients with either abruption or placenta previa in PTL (mean magnesium level of 1.57 ± 0.21 mg/dl; range, 1.5 to 1.8 mg/ dl) had means significantly lower than the normal controls (p < 0.005 for both). When patients in PTL with PROM, twins, or bleeding disorders were compared, their mean magnesium levels were not significantly different (p > 0.25) (Figs. 2, 3). Since all
Downloaded by: NYU. Copyrighted material.
Patients with known diabetes mellitus of any Priscilla White class were considered separately (16 patients were in class A, six in B, four in C, one in D, two in F, and four in R), with 23 determinations in the non-insulin-requiring group, and 20 in the insulin-dependent group (IDDM). Gestations ranged from 9 to 41 weeks. The 23 patients with preeclampsia or eclampsia were considered as a separate group. Their gestational ages ranged from 24 to 40 weeks, and the diagnosis was made by standard criteria.2 Patients with premature rupture of the membranes (PROM) were considered separately as two groups: those without and those with uterine irritability or contractions. Twenty-six patients from 22 to 36 weeks' gestation constituted the first group. Only patients who had absent contractions or no evidence of uterine activity for at least 48 hours after the serum sample was obtained were entered into this group. Uterine activity was considered to include patients who perceived cramping or contractions, as well as those who had contractions that were palpable or could be recorded on an external tocodynamometer of frequency six or more per hour or who had cervical effacement and dilation on examination. Sixteen patients with PROM and uterine activity constituted the second group (21 to 35 weeks' gestation). PROM was considered gross rupture of the membranes confirmed by positive Nitrazine and fern testing. Serum samples were likewise obtained from 71 patients in PTL (gestations 21 to 36 weeks). PTL was defined as uterine contractions with accompanied cervical effacement or dilation, or regular uterine contractions of frequency six or more per hour. Included in this group were the 16 patients with PROM, three with abruptions, four with placenta previa, eight had twins, one had a cervical circlage, and three had documented chorioamnionitis. The remaining patients in this group had undetermined etiologies for their PTL. All of these patients were transferred in from other hospitals to our tertiary care facility. Patients were excluded from this study if they had received intravenous magnesium sulfate tocolysis prior to arrival at our institution. Thirteen normal term patients (37 to 42 weeks) in active labor constituted another study group. All patients in the study had taken iron and prenatal vitamin supplements.3 Patients were excluded from the study if they had taken oral magnesium or calcium supplements, diuretics, or had aminoglycoside therapy.4 Likewise excluded from the preeclamptic, PROM, and PTL study groups were patients with diabetes, renal disease, and hypoparathyroidism. Total serum calcium, total magnesium, phosphate, albumin, and creatinine were determined by colorimetric methods and quantitated absorptiometrically. Total serum magnesium was determined by the reaction with the metallochrome dye calmagite, in an alkaline solution, eliminating interference with calcium and heavy metals by including the surfac-
March 1991
MAGNESIUM LEVELS IN PREGNANCY/Kurzel ° •
NORMAL/NOT IN LABOR TERM/LABOR
IT 2.5-
i
2.0-
III
I
IT
IT
,r-J-
U4 -1-
I
1.5-
u
1.0
20
40
30
Downloaded by: NYU. Copyrighted material.
10
GESTATION ( weeks ) Figure 1. Serum magnesium concentrations (mg/dl) for normal pregnant women, o = not in labor, and * = term patients (37 to 42 weeks) in labor. Plotted are mean values for each week of gestation and the range of values. There is no gestational dependence up until 33 weeks, at which point magnesium levels drop.
PRE
2.5-
TERM LABOR
-
o • •
•
•
*
•o M
E
*
2.0 —
a o
0
•
TOTAL PROM BLEED TWINS CHORIO NO CVX CHANGE
2 u
z o
o • o*
• •
o • * o o
1.5 — - 8 — • ¥* •
0 -
*
—
—
•
•
' o *
o
o o .—o-• * -fcT*< a
—
0
o
*
0 D
* •
20
1
1 30
GESTATION
40
( weeks )
Figure 2. Serum magnesium levels for patients in PTL. Mean values for each week gestation are shown, and the least squares plot describing these data (y = 1.25 + 0.0114 x). Shown are the means for total patients in PTL(o), patients with PROM in labor (*), those with PTL due to abruption and placenta previa (•), those with twin gestation in PTL {+), those with chorioamnionitis in PTL (•), and those in PTL of unknown etiology and with no cervical change (0). Irrespective of etiology, all points cluster about the same curve.
121
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
March 1991
° PROM NOT IN LABOR • PROM IN LABOR 2.5 •o
2.0 U
z o 1.5 OS bJ
40
20
GESTATION
(
weeks
)
Figure 3. Serum magnesium levels for patients whose PTL is due to PROM. Shown are the mean values at given gestations for patients with PROM without labor (o), and the least squares curve describing these data (top curve: y = 1.45 + 0.0125x). Also shown are the means for patients with PROM in PTL (*) and the corresponding least squares plot (lowest curve: y = 1.20 + 0.012x). Given for reference is the curve for normal controls, not in labor (bold curve).
patients with PTL were transferred from other hospitals, with the exception of those with PROM, all PTL patients were seen here already with uterine irritability or contractions. We therefore could not observe patients temporally prior to the development of uterine contractions to determine whether magnesium levels were depressed prior to development of PTL, as was suggested by Martin and Morrison.5 For the special case of PROM, we observed this not to be the case. More work needs to be done to study this point, since it may have practical bearing on the issue of oral magnesium prophylaxis for patients at risk of preterm delivery. Patients with non-insulin-requiring diabetes mellitus (mean magnesium level of 1.75 ± 0.42 mg/ dl; range, 1.4 to 2.2 mg/dl) and IDDM of all White classes (mean magnesium level of 1.75 ± 0.51 mg/dl; range, 1.4 to 2.4 mg/dl) had means that appeared to have a tendency to be lower than the controls, but the difference was not significant (p = 0.25) (Fig. 4). Patients with preeclampsia or eclampsia were compared to normal controls for the gestational spans 24 to 33 weeks and 37 to 42 weeks. For 24 to 33 weeks, patients with preeclampsia had a mean magnesium level of 1.72 ± 0.41 mg/dl (range, 1.4 to 2.0 mg/dl), which was not significantly different from the controls (p > 0.1). At 37 to 42 weeks' gestation, the mean magnesium of 1.70 ± 0.60 mg/dl (range, 122 1.4 to 2.3 mg/dl) for patients with preeclampsia was
also not significantly different from controls (p > 0.25), whose values were also dropping over this gestational span (Fig. 5). Total serum magnesium did not have a significant dependence on total serum calcium, phosphate, or albumin, as shown by the scatter plots in Figures 6 and 7. Total serum calcium was found to decline slowly to a minimum at 34 weeks, and then rise again approaching term, as has previously been reported by Bodansky and Duff,6 and Pitkin et al7 For 21 to 33 weeks' gestation, the mean total serum calcium was 9.16 ± 0.50 mg/dl compared with 9.14 ± 0.37 mg/dl for patients at this gestation in PTL, which is not statistically different (p > 0.25). Similarly, the mean total calcium for patients not in labor at 37 to 42 weeks was 9.17 ± 0.84 mg/dl compared with 9.14 ± 1.06 mg/dl for term patients in labor, which is not statistically significant (p > 0.25). DISCUSSION
The average serum level of magnesium in nonpregnant healthy women of reproductive age is 2.4 ± 0.49 mg/dl.5 That pregnancy is marked by a state of hypomagnesemia was noted by Hall in 1957.8 Numerous studies have since confirmed this finding, although the exact gestational dependence of serum magnesium is still unsettled, with various studies
Downloaded by: NYU. Copyrighted material.
1.0
MAGNESIUM LEVELS IN PREGNANCY/Kurzel
DIABETES CLASS
A B C D,
2.5
o » a F, R
•
2.01
*
•
o oo a
1.5
o
o
1.0' 30
20
GE STAT I O N
(
40
w e e k s
)
Figure 4. Serum magnesium levels for pregnant diabetics, by the White classification. Given are the mean values at each gestation (class A = o, B = *, C = •, D, F, R = •), and the curve for normal controls not in labor is presented for reference. No significant difference in levels was found for either non-insulin requiring or IDDM patients relative to normal controls.
2.5-
PREECLAMPSIA
\
2.0o o
o
o
o
Z
o o
o
1.5 UJ
1.0 20
I 30
GESTATION
40
( weeks )
Figure 5. Mean serum magnesium levels for patients with preeclampsia who were not in labor (o). The curve for normal controls not in labor is shown for reference. No statistically significant difference could be seen between the two groups for the entire span of gestations. 123
Downloaded by: NYU. Copyrighted material.
10
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
2.5-
2.5-
2.0-
2.0
Mg
March 1991
-
1.5-
1.5 -
1.0-
1.0-
3.0
2.0
PO
I 4.0
I 5.0
2.0
4.0
3.0
ALB
Figure 6. A: Scatter plot for total serum magnesium versus serum phosphate for normal pregnant patients not in labor (gestations, 10 to 42 weeks). Nocorrelation isseen(r2 = 0.097). B: Scatter plotfortotal serum magnesium versus serum albumin for normal pregnant patients not in labor (gestations, 10 to 42 weeks). No correlation is seen (r2 = 0.047).
-
NOT IN LABOR
-
PRETERM LABOR ;
10.0 —
10.0-
bis
• 9.0 -
9.0-
2
u < o
•
8.0-
8.0 -
7.0-
7.0-
U
4 A
2.0
1.0
"
1
1
2.0
SERUM MAGNESIUM ( mg/dl ) Figure 7. Scatter plots for total serum calcium versus total serum magnesium for patients not in labor (21 to 33 weeks) 124 (left), and for those in preterm labor (right).
Downloaded by: NYU. Copyrighted material.
Mg
MAGNESIUM LEVELS IN PREGNANCY/Kurzel
gestation and not in labor, with which to compare this mean level. Similarly, Elliott et al20 noted that patients in PTL in his study had a mean serum magnesium level of 1.80 ± 0.04 mg/dl. Again there was no control group to compare these levels to, and the gestational dependence was not stated. Along these lines, hypomagnesemia accompanying PTL was implied but not demonstrated by three clinical studies: an epidemiologic study by Kuti,21 and double-blinded clinical trials by Conradt et al22 and Spading et al.23 These three studies demonstrated that prophylactic oral magnesium supplementation to patients at risk for preterm delivery was successful in lowering the preterm delivery rate. Abnormally low serum magnesium with PTL was not documented, but nonetheless these authors proposed that a certain percentage of preterm deliveries could be attributed to a magnesium deficiency. In a Russian study, Penev et al24 studied 201 patients in PTL and noted low serum magnesium levels in 21.9% of their patients. In our study, from 21 to 33 weeks' gestation, a statistically significant reduction in serum magnesium was seen in patients with PTL, which appears to be independent of the etiology for the PTL. The depressed serum magnesium level therefore reflects the tendency for PTL, or the initiation of PTL, and does not characterize the particular predisposing factor for PTL. The number of patients in each of these individual groups is small, however, and this question warrants further study in the future. The question, which is not answered by this study, is whether hypomagnesemia precedes the development of uterine irritability, as is the case for our normal patients prior to labor. What is needed to answer this question is a longitudinal study of patients at risk for PTL. Interestingly for the particular subgroup with PROM, magnesium levels for those patients not in PTL were not different from the normal controls at matched gestations. Magnesium levels dropped, however, only once PTL ensued. Whether magnesium levels drop in the period immediately preceding the development of PTL is unknown. Hypomagnesemia therefore may either be a causative factor or simply reflect the process involved with the development of uterine irritability for both term labor and PTL. Evidence has accumulated in recent years from in vitro studies to explain the connection between hypomagnesemia and the development of uterine irritability. Reduced extracellular magnesium levels have been shown to increase smooth muscle contractility in vitro. Altura et al25 and Sjogren and Edvinsson26 have demonstrated that extracellular magnesium influences the release of calcium from intracellular depots; a decrease in extracellular magnesium results in an increase in the total exchangeable and intracellular calcium fractions, with a resultant increase in smooth muscle tension. In addition, Somlyo et al27 showed that decreased extracellular magnesium can result in depletion of the magnesium content of smooth muscle cells. Magnesium is a required cofactor in the myometrial contractile re-
Downloaded by: NYU. Copyrighted material.
giving different trends. Hall8 noted that serum magnesium progressively declined during pregnancy, reaching a nadir at 34 to 38 weeks, and then began to rise approaching term. Pitkin et al,9 De Jorge et al,10 and Sheldon et al3 described a similar trend, but the nadir was at 28 to 32 weeks. Pitkin's9 curve showed a very shallow decline before rising approaching term. De Jorge's10 curve rose from the nadir at 28 to 32 weeks, peaked at 36 weeks, and then declined again in the last month approaching term. By contrast, Roelofsen et al11 found serum magnesium to drop and reach a steady state in the second trimester, only to remain constant until term. Almuna et al12 observed no variation at all in magnesium levels from 28 to 42 weeks (mean 1.74 ± 0.18 mg/dl). The results of our study showed a reduced magnesium level in pregnancy (mean 1.79 ± 0.44 mg/dl) showing no gestational dependence until 33 weeks. Beyond 33 weeks, magnesium concentrations progressively declined (Fig. 1). This is in agreement with the trend noted by Olatunbosun et al,13 Bartl and Riss,14 Rusu etal, 15 Ray etal, 16 and Martin and Morrison.5 Kolasa et al17 also noted a progressive decline post-term compared with levels at term. For patients in labor at term, Bogert and Plass18 as early as 1923 suggested that there "seems to be a tendency for the serum magnesium to be lowered in labor." Although the mean serum magnesium was 2.1 mg/dl, they noted that 65% of the values were under 2.0 mg/dl (range, 1.4 to 3.2 mg/dl). Results of Rusu et al15 and Ray et al16 seem to agree with this finding, but plots of values relative to gestations were not given. The study of Olatunbosun et al13 actually found levels for patients in labor to be higher than for those at term not in labor. Our results do not confirm the findings of Bogert and Plass.18 It would appear that serum magnesium levels for term patients in labor (37 to 42 weeks) are not significantly different from the values of term patients not in labor (see Fig. 1). Serum magnesium therefore drops as a prodrome preceding labor, but is not lowered further by labor. The reason for hypomagnesemia during pregnancy is unknown, but it is unlikely to be explained solely by hemodilution, as suggested by Hall.8 De Jorge et al10 and Sheldon et al3 have demonstrated that by taking into consideration plasma volume expansion, total circulating magnesium may be calculated and shown to increase steadily through the pregnancy until the last month, when it drops. Since magnesium may exert a smooth muscle relaxing effect, this progressive rise may exert a quieting effect on the myometrium, while the drop in levels in the 1 to 2 months preceding term may contribute to increased uterine irritability. What initiates this drop in serum magnesium concentration from 33 weeks onward is unknown. If a drop in serum magnesium levels is a prodrome to the initiation of term labor, might the same be also true for PTL? In a small study of only 17 patients in PTL, Martin et al19 found that the mean serum magnesium level was 1.44 ± 0.22 mg/dl. There was no control group of patients matched for
125
sponse,28 and low extracellular magnesium levels may result in increased contractility in any or all of the following ways: 1. Magnesium is an activator of adenylate cyclase. Reduced magnesium correlates with decreased synthesis of cyclic adenosine monophosphyate (c-AMP), and thereby reduced c-AMP-mediated calcium sequestration.25 Cytoplasmic concentration of calcium is therefore increased, resulting in muscle contraction. In addition, decreased c-AMP-dependent protein kinase activity would contribute toward increased myosin light chain kinase activity and uterine contractility. 2. Cytoplasmic calcium concentration is also controlled by the sarcoplasmic reticulum. The magnesium-adenosine triphosphate (ATP) complex acts as the substrate for calcium-dependent ATPase, stimulating calcium pump activity. Reduced magnesium results in inhibition of the calcium pump, increased cytoplasmic calcium, and increased myocyte contractility.25-29 3. Magnesium may act on the cell membrane to regulate entry and exit of calcium into the cell, by influencing membrane permeability. Calcium and magnesium may share the same carrier on the membrane and compete for transport, hence affect cytoplasmic concentrations. This transport is also c-AMP dependent, acting via the Na+/K+ pump.25-30 4. Magnesium-ATP is bound to the myosin light chains to weaken the actin-myosin interaction. Decreased magnesium concentration may therefore favor contraction.31^33 5. Magnesium can bind to calmodulin and may have an effect on the calcium-calmodulin complex.^ As noted earlier, Conradt et al22 and Spading et 23 al have attributed PTL to a magnesium deficiency. This is unlikely. Hsueh and Zuspan35 in reporting a case of hypomagnesemia in pregnancy secondary to nutritional causes noted that true magnesium deficiency is rare (whereas PTL is not). During pregnancy, hypomagnesemia is most commonly seen with diabetes mellitus, followed by renal insufficiency, alcoholism, and with hypoparathyroidism. Low serum magnesium levels with diabetes have been reported by Cruikshank et al,36 Durlach et al,37 and Mimouni et al38 In our study where the patients were in good glycemic control, a tendency for lower magnesium levels was seen for all diabetics, but did not reach statistical significance. Cruikshank et al36 and Durlach et al37 have shown that the lowest magnesium levels correlated with increasing severity of diabetes (although the magnesium levels were reduced by only approximately 4% for IDDM). Mimouni et al38 explained this by increased urinary magnesium losses with severe uncontrolled diabetes, versus an effect of insulin itself on lowering the serum magnesium. Interestingly, patients with IDDM are no more prone to PTL in spite of their hypomagnesemia, and only one patient with IDDM in this 126 study was noted in PTL.
March 1991
Similarly, patients with renal insufficiency were excluded from this study. The kidney is the major regulator of serum magnesium, and with renal insufficiency, urinary magnesium losses result in hypomagnesemia.39 Finally, patients with preeclampsia were studied as a separate group. Speculation40 on the occurrence of hypomagnesemia with preeclampsia is based on such observations by Kolasa et al17 and Altura et al25 in patients near term or at term. Hypomagnesemia has been hypothesized to play a role in the increased vascular reactivity seen in preeclampsia. Since preeclampsia is most common as term is approached, it occurs when serum magnesium levels are normally declining. Our results do not show a tendency for patients with preeclampsia to have serum magnesium levels different from normal. This is in agreement with the findings of Roelofsen et al,11 Almuna et al,i2 and Hall.8 It is interesting that in 1939 Bodansky and Duff6 noted that hypocalcemia developed preceding labor in the rat. We and others 79 have shown that this does not hold true at term or with PTL in humans, where serum calcium is stringently regulated by parathyroid hormone. Instead, hypomagnesemia is seen, which in ways still to be determined, reflects a process resulting in increased uterine irritability/ contractility. A potential use of these results may be the determination of a marker to differentiate patients truly in PTL, and those patients at greatest risk for progressing on to preterm delivery. For certain from 21 to 33 weeks gestation, patients with a serum magnesium 1.4 mg/dl or less (^ 2 SD below the mean) should be viewed as being at risk for a preterm delivery.
REFERENCES
1. Steer CM, Petrie RH: A comparison of MgSO4 and alcohol for the prevention of premature labor. Am J Obstet Gynecol 129:1-4, 1977 2. Pritchard JA, MacDonald P, Gant NF: Williams Obstetrics. Norwalk, CT: Appleton-Century-Crofts, 1985, pp 525-527 3. Sheldon WL, Aspillaga MO, Smith PA, Lind T: The effect of oral iron supplementation on zinc and magnesium levels during pregnancy. Br J Obstet Gynaecol 92:892-898, 1985 4. Dirks JH, Wong NL: Renal magnesium wasting disorders. Adv Exp Med Biol 208:193-197, 1986 5. Martin RW, Morrison JC: Oral magnesium for tocolysis. Contemp OB/GYN 30:111-118, 1987 6. Bodansky M, Duff VB: Regulation of the level of calcium in the serum during pregnancy. JAMA 112:223-229, 1939 7. Pitkin RM: Calcium metabolism in pregnancy: A review. Am J Obstet Gynecol 121:724-737, 1975 8. Hall DG: Serum magnesium in pregnancy. Obstet Gynecol 9:158-162, 1957 9. Pitkin RM, Reynolds WA, Williams GA, Hargis GK: Calcium metabolism in normal pregnancy: A longitudinal study. Am J Obstet Gynecol 133:781-790, 1979 10. De Jorge FB, Delascio D, de Ulhoa Cintra AB, Antunes ML: Magnesium concentration in the blood serum of normal pregnant women. Obstet Gynecol 25:253-254, 1965 11. Roelofsen JMT, Berkel GM, Uttendorfsky OT, Siegers JFG: Urinary excretion rates of calcium and magnesium in
Downloaded by: NYU. Copyrighted material.
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 8, NUMBER 2
MAGNESIUM LEVELS IN PREGNANCY/Kurzel
13.
14. 15. 16. 17.
18. 19. 20. 21.
22.
23. 24.
25.
26.
27. 28.
29. 30. 31. 32. 33. 34. 35. 36.
37. 38.
39. 40.
channel blockers on myogenic tone and agonist-induced responsiveness of blood vessels. Can J Physiol Pharmacol 65:729-745, 1987 Sjogren A, Edvinsson L: The influence of magnesium on the release of calcium from intracellular depots in vascular smooth muscle cells. Pharmacol Toxicol 62:17—21, 1988 Somlyo AP, Somlyo AV: Effects and subcellular distribution of magnesium in smooth and striated muscle. Fed Proc 40:2667-2671, 1981 D'Ocon MP, Anselmi E, Villar A: Effects of MgCl2 on the contractile response of uterus to several agonists in calcium-free solution. J Pharm Pharmacol 39:444—448, 1987 Watras J: Effects of Mg+2 on calcium accumulation by two fractions of sarcoplasmic reticulum from rabbit skeletal muscle. Biochim Biophys Acta 812:333-344, 1985 Flatman PW: Magnesium transport across cell membranes. J Membrane Biol 80:1-14, 1984 Cooke R, Pate E: The effects of pyrophosphate on myosin heads and on fibre stiffness. Adv Exp Med Biol 226:255265, 1988 Davis JS: Myosin heads in muscle thick filament assembly. Nature 333:807, 1988 Potter JD, Robertson SP, Johnson JD: Magnesium and the regulation of muscle contraction. Fed Proc 40:26532656, 1981 Tsai MD, Drakenberg T, Thulin E, Forsen S: Is the binding of Mg(II) to calmodulin significant? An investigation by Mg-25 NMR. Biochemistry 26:3635-3643, 1987 Hseuh GSC, Zuspan FP: Magnesium deficiency in pregnancy: A case report. J Reprod Med 27:479-483, 1982 Cruikshank DP, Pitkin RM, Reynolds WA, Williams GA, Hargis GK: Altered maternal calcium homeostasis in diabetic pregnancy. J Clin Endocrinol Metab 50:264— 267, 1980 Durlach J, Collery P: Magnesium and potassium in diabetes and carbohydrate metabolism. Magnesium 3:315—323, 1984 Mimouni F, Miodovnik M, Tsang RC, Holroyde J, Dignan PS, Siddiqi TA: Decreased maternal serum magnesium concentration and adverse fetal outcome in insulindependent diabetic women. Obstet Gynecol 70:85—88, 1987 Agus, ZS, Wasserstein A, Goldfarb S: Disorders of calcium and magnesium homeostasis. Am J Med 72:473—488, 1982 Conradt A: New concepts in the pathogenesis of gestosis under special consideration of magnesium deficiency. Z Geburtshilfe Perinatol 188:49-58, 1984
Downloaded by: NYU. Copyrighted material.
12.
normal and complicated pregnancies. Eur J Obstet Gynecol Reprod Biol 27:227-236, 1988 Almuna RV, Trivino XB, Radrigan GA, Moreno EG, Herrera RZ, Espinoza IC, Galindo ES, Aguilar PO: Comportamiento del Magnesio en embarazadas. I. Estudio comparativo entre embarazadas normales y con patologia. Rev Chil Obstet Ginecol 48:65-70, 1983 Olatunbosun DA, Adeniyi FA, Adadevoh BK: Serum calcium, phosphorus, and magnesium levels in pregnant and non-pregnant Nigerians. Br } Obstet Gynaecol 82: 568-571, 1975 Bartl W, Riss P: Magnesium serum levels and magnesium deficiency in pregnancy. Z Geburtshille Perinatol 186: 335-337, 1982 Rusu O, Lupan C, Baltescu V: Magneziul seric in sarcina normala la termen si nasterea prematura. Obstet Ginecol 14:215-219, 1966 Ray KK, De MK, Roy N: Serum magnesium in different trimesters of pregnancy. J Indian Med Assoc 72:28—32, 1979 Kolasa F, Krajewski J, Szpakowski M: Serum magnesium levels in pregnant women with symptoms of EPH-Gestosis and with post-mature pregnancy. Wiad Lek 36:897— 899, 1983 Bogert LJ, Plass ED: Placental transmission: I. The calcium and magnesium content of fetal and maternal blood serum. J Biol Chem 56:297-307, 1923 Martin RW, Gaddy DK, Martin JN Jr, Lucas JA, Wiser WL, Morrison JC: Tocolysis with oral magnesium. Am J Obstet Gynecol 156:433-434, 1987 Elliott JP: Magnesium sulfate as a tocolytic agent. Am J Obstet Gynecol 147:277-284, 1983 Kuti V, Balazs M, Morvay F, Varenka Z, Szekely A, Szucs N: Effect of maternal magnesium supply on spontaneous abortion and premature birth and on intrauterine fetal development: Experimental epidemiological study. Magnesium Bull 3:73-79, 1981 Conradt A, Weidinger H, Algayer H: Magnesium therapy decreased the rate of intrauterine fetal retardation, PROM, and premature delivery in risk pregnancies treated with betamimetics. Magnesium 4:20-28, 1985 Spatling L, Spading G: Magnesium supplementation in pregnancy. A double blind study. Br J Obstet Gynaecol 95:120-125, 1988 Penev IV, Ruseva R, Pavlova F, Naumov IV, Stereva K, Serdev V: Threatened preterm delivery at low attached placenta and magnesium insufficiency. Akush Ginekol (Sofiia) 23: 97-102, 1984 Altura BM, Altura BT, Carella A, Gebrewold A, Murakawa T, Nishio A: Mg+2 —Ca+2 interaction in contractility of vascular smooth muscle: Mg+2 versus organic calcium
127
