SERUM MAGNESIUM LEVELS 1N THE NEWBORN Constantine S. Anast, M.D. Department of Pedkitrics, University of Missouri School of Medicine, Columbia, Mlssouri K NOWLEDGE of magnesium has lagged levels in the newborn infant, there are very behind that of other electrolytes limited data concerning serum magnesium largely due to difficulties in methodology. concentrations in this age group. The pur- However, interest in magnesium metabolism pose of the present study was to carry out is indicated by the increasing number of serial determinations of serum magnesium articles on the subject that have appeared levels in the newborn period in an effort to: in the recent literature. Low serum mag- (1) establish the normal values in this age nesium levels have been observed in states group; (2) compare the values found in new- of malnutrition with loss of body fluids such born infants with those of older children ( as may be found in postoperative patients and adults; and 3) compare the values in with nasogastric suction who are being breast-fed and artificially fed newborn in- maintained on magnesium-free parenteral fants. fluids. Low serum magnesium levels have CASE MATERIAL also been found in patients with hyper- and hypoparathvroidism, alcoholism with de- A study was made of 72 normal, full-term lirium tremens, hyperaldosteronism, and infants who were the products of uncompli- diabetic acidosis. In renal disease the serum cated pregnancies and deliveries. Thirty- magnesium levels may either be elevated or four of the infants were breast fed and 38 ( depressed, hypermagnesemia being found received evaporated milk 13 oz of evapo- in the presence of significantly depressed rated milk with 400 U. Vit. D, 24 oz of glomerular filtration rates. In some patients water, and 3 tablespoons of Karo syrup). low levels of magnesium have been associ- Milk feedings were started after 24 hours of ated with tetany, generalized convulsions, age and no supplemental vitamins were athetoid motion of the extremities, and offered. Birth weights of the infants ranged marked reaction to mechanical and auditory from 2,500 gm to 4,250 gm. stimulation. The mental status varies from Sixty-six older children, ages 2 months to complete lucidity to semi-coma. Increased 16 years, and 47 adults were also studied. neuromuscular irritability with convulsions The adults were all normal hospital per- is known to occur in the magnesium-dc- sonnel. The children were either normal or ficient animal. admitted to the hospital for elective surgical Similarities exist between calcium and procedures. In no instance was there any magnesium metabolism. Both are divalent history or evidence of metabolic disease, ions and a decrease in the serum level of fluid and electrolyte problems, or convulsive either is associated with an increase in disorders. neuromuscular irritability. Phosphate infu- METHOD OF STUDY sion in dogs results in decreased serum levels of calcium and magnesium.1 Both are Samples of blood were collected from the incompletely absorbed from the gastroin- heel of the infants during the first 5 days testinal tract and bone is the chief reservoir of life and from the finger tips of the older for each in the body. Although there are children and adults in Hinton capillary several reported studies of serum calcium tubes. In all cases blood was collected in (Submitted Sept. 13, 1983: accepted for publication February 3, 1964.) Supported by research grant (AM-01351) from the National Institutes of Health, Public Health Serv- ice, Bethesda, Maryland. ADDRESS: Department of Pediatrics, University Hospital, Columbia, Missouri. PEDIATRICS, June 1964 969 970 SERUM MAGNESIUM TABLE I One limitation of the Titan yellow method is the variation that may be found when the SERUM MAGNESIUM LEVELS same sample is analyzed on different days. Determi- Therefore, in order to compare serial sam- . Mean+S.D. Newborn Infants nations Range ples more accurately all of the sera from a (mg/100 ml) (no.) given infant were analyzed on the same day in the same run. Included with each run Day I 56 1.94 ± .26 1.40-2.90 Day 2 56 1.87± .27 1.36-2.90 were one or two samples obtained from DayS 55 1.95±.24 1.41-2.59 older children and adults. This allowed for Day4 49 1.91±27 1.36-2.62 more accurate comparison of the newborn Day 5 22 2.01± .28 1.49-2.73 sera with that of the control group. Total5days 238 1.92±27 1.36-2.90 In a previous study we found strikingly Cordblood 41 1.89±26 1.43-2.45 false low serum magnesium levels in new- born infants receiving intravenous calcium Adults and older gluconate.(cid:1) In no case in the present study children 111 1.96±24 1.20-2.58 were any of the subjects receiving calcium gluconate or any other medication. the morning midway between feedings and RESULTS the serum samples were stored in a freezer. In most cases, a minimum of three samples The mean value of 238 determinations of from each infant was analyzed, for a total serum magnesium levels during the first 5 of 238 determinations. In 41 cases, cord days of life was 1.92 mg/100 ml with a blood was also examined. range of 1.36 to 2.90 mg/100 ml. The mean Magnesium determinations were done by value, the standard deviation, and the range a modification of the Titan yellow method for this group as well as for 41 cord bloods of Orange and Rhine.2 The reaction con- and for 111 older children and adults are sists of the combining of magnesium hy- listed in Table I. No statistically significant droxide with the dye Titan yellow in the difference was found among the mean presence of the dispersing agent polyvinyl values of the three groups. Neither was any alcohol. The magnesium hydroxide-Titan significant difference found when the con- yellow lake results in a pink color, the in- trol group was broken down by age to tensity of which is measured spectrophoto- groups of less than one year, 1 to 16 years metrically. The method was modified to de- and 16 to 35 years. termine magnesium concentrations in 0.2 The values determined for each of the ml of serum. first 5 days of life also are listed in Table I. The precision of the method was tested Differences in mean values between any of by making a total of 20 analyses of samples the days were not statistically significant of the same serum on five different days. nor were any significant differences found The values found ranged from 1.60 to 1.88 when the values of each day were compared mg/100 ml with a mean of 1.70 and an aver- with the mean value of older children and age deviation of plus or minus 0.07 mg/100 adults or the mean value of the cord bloods. ml. No differences were found that could be The validity of the method was further related to the birth weight, sex, or race of tested by carrying out recovery studies. Dc- the infant. terminations were performed on 0.2 ml of In Table II the magnesium levels have sera to which varying quantities of mag- been tabulated according to the type of nesium ranging from 50 to 300 micrograms feeding (breast or evaporated milk) and the were added. The average recovery in 10 age of the infant in days. In addition the analyses was 100.64 ± 3.4% showing that 5-day mean for each feeding schedule is recovery is complete. listed. The mean serum magnesium levels ARTICLES 971 TABLE II SERUM MAGNESIUM LEVELS (mg/100 ml) ACCORDING TO TYPE OF FEEDING Day 1 Day 2 Day 3 Day 4 Day 5 Total Evaporated milk Mean±S.D 1.96±26 1.87±2.5 1.88±25 1.85±24 1.95±23 1.89±25 Range 1.62-2.90 1.36-2.32 1.41-2.51 1.36-2.32 1.49-2.27 1.36-2.90 Number (27) (32) (29) (30) (13) (131) Breast milk Mean±S.D. 1.92±26 1.88±31 2.03±21 2.01±27 2.11±34 1.97±.28 Range 1.40-2.59 1.44-2.90 1.67-2.59 1.56-2.62 1.72-2.73 1.40-2.90 Number (29) (24) (26) (19) (9) (107) pvalue <0.60 <0.90 <0.02 <0.05 <0.20 <0.02 of the evaporated-milk-fed infants were pared to day 1 were found in 5 of these 6 found to be lower than those of the breast- breast-fed infants while decreasing levels fed infants on days 3, 4, and 5. The differ- were found in 5 of the 6 corresponding ences on days 3 and 4 are significant at the evaporated-milk-fed infants. The serum 98 and 95% confidence limits respectively. magnesium levels of the 2 groups on day 1 The lack of statistical significance on day 5 were comparable. In 2 pairs the level was may reflect the smaller number of determi- higher in the breast-fed infant, in 2 pairs it nations for this day. The smaller number of was higher in the infants fed evaporated determinations on day 5 is a result of the milk, and in 2 pairs there was no appreci- routine discharge of many healthy infants able difference in the initial serum mag- from our nursery before this day. nesium concentration. Eight of the breast-fed infants and 10 of Further analysis of serum magnesium the evaporated-milk-fed infants had serum levels as related to type of feeding is found magnesium levels measured on each of the in Table III. The number of infants in first 4 days of life. Plotting these results which there was a rise or fall in serum mag- (Fig. 1) allows a comparison of values for nesium concentrations of at least 0.1 mg/ the same infants on each of the first 4 days. 100 ml on day 3 and on day 4 (day 5 not A tendency for increasing values for the included because of small number of de- breast-fed infants and decreasing values for terminations) when compared to day 1 in the evaporated-milk-fed infants after day 2 is graphically demonstrated. This group in- eludes the last 12 babies studied in which 2.6 2.5 there were 6 breast-fed and 6 evaporated- milk-fed infants. Each breast-fed infant was paired with an evaporated-milk-fed in- \ 2.2 fant who was born at approximately the (cid:1)2I same time (within a week). All of the sera (cid:1) 2.0 collected from each pair were analyzed on (cid:1) ‘.9 (cid:1) .8 the same day in the same run. In addition a (cid:1) reference standard of magnesium was anal- yzed in each of the 6 paired runs. The 2 3 4 I 2 3 4 results for the standard reference samples DAYS OF LIFE DAYS OF LIFE varied from 1.90 to 2.04 mg/100 ml during Fic. 1. Serum magnesium levels on each of the the study of the 6 pairs. Increasing serum first 4 days of life in 8 breast-fed and 10 evapo- magnesium levels on days 3 and 4 as com- rated-milk-fed infants. 972 SERUM MAGNESIUM TABLE III breast-fed infants and decreasing values in evaporated-milk-fed infants is again demon- INCREASE OR DECREASE IN SERUM MAGNESIUM ON DAYS 3 AND 4 AS COMPARED TO LEVEL ON DAY 1 IN strated. THE SAME INFANT ACCORDING TO TYPE OF FEEDING COMMENT No Change We believe that the major contribution Feeding Increase Decrease (<0.1 mgI 100 ml) of the present study is that it presents nor- mal values for serum magnesium concentra- Day 3 lions during the first 5 days of life. The re- suits represent analysis of 238 serum sam- Breast 11 3 7 Evaporated milk S 9 9 pies from this age group. The tendency for X2=7.82 p=O.O2 increasing levels in breast-fed infants and decreasing levels in evaporated-milk-fed in- Day 4 fants in this study is of interest. In some in- fants the changes were small. Because of Breast 8 2 5 Evaporated milk S 9 6 this and because of the limitations of the X2=6.6 p<O.O5 Titan yellow method it is probably best to view the differences in formula and breast- the same infant is listed according to the fed infants with some reservation until they type of feeding. Also listed are the number are confirmed. of intervals during which the difference Previously reported studies of serum mag- from day 1 was less than 0.1 mg/100 ml. nesium levels in the newborn period are The tendency for increasing values in the summarized in Table IV. When available, TABLE IV SUMMARY OF THE LITERATURE OF SERUM MAGNESIUM LEVELS IN TILE NEWBORN PERIOD ( The mean in mg/i#{174} ml, the range, and the number of determinations are listed for each group) Autker Method 3Iater,(cid:1)at Cord First Week 1-12 Jfo Older Children Adults Bogert&Plass’ PhosphatePpt. 2.0 2.1 2.3 1.4-3.2 1.4-3.4 1.9-2.7 (23) (23) (8) Salmi’ 2.68 Art. 2.76 2.98 1.88 1.7-3.5 1.5-4.0 2.3-4.0 1.3-2.3 1.7-2.9 (32) ((cid:1)8) (14) (20) (cid:1) (10) Yen. 2.74 1.5-4.1 (37) Orange & Rhine2 Titan yellow i.18 2.16 2.27 1.6-2.3 1.9-2.4 1.9-2.5 (12) (12) (45) Breton el d.6 Titan yellow 3.26 2. 57’ 2.57 2.16-4.0 1.8-3.2 2.0-3.6 Mays&Keele’ EDTA 1.68 I.3(cid:1) 1.63 1.64 (23) (18) (193) (15) Marioni at al.(cid:1) Phosphate Ppt. I.80 1.84 1.20-2.20 1.25-2.35 (30) (31) Present series Titan yellow 1.89 1.92 2.04 i.94 .98 1.28-2.45 1.36-2.90 1.80-2.40 1.51-2.58 1.36-2.36 (41) (238) (11) (52) (47) SIncluded infants up to 24 months of age. ARTICLES 973 the values determined by the authors for cow’s milk is .013% as compared to .004% older children and adults also are listed. It for human milk.9 However, the ratio of can be seen that our values are of the same phosphorus to magnesium in human milk general magnitude as those found by Bogert is 4 to 1 and in cow’s milk 7.6 to 1. Gardner and Plass,4 Orange and Rhine,2 and Marioni et al.b0 demonstrated a rise in serum in- et al.s Differences from results found by organic phosphate and a fall in serum cal- other authors5’ ‘(cid:1) may be due in large part cium and magnesium in a newborn infant to methodology and indicate the importance receiving a cow’s milk formula. When cal- of each laboratory establishing its own nor- cium was added to the formula they ob- ma! values. served a diminution in the serum inorganic Of prime interest is the comparison of phosphorus level with the return of the values for different age groups within each ionized calcium and total magnesium to author’s series. No significant difference was normal levels. found between cord blood and maternal Bruck and Weintraub,11 Graham et al.,12 serum magnesium concentrations in the and Gittleman and Pincusl3 have all demon- three studies where both were deter- strated that the intake of cow’s milk is fre- 45 8 J(cid:1) addition, Salmi(cid:1) did not find quently associated with a rise in serum any difference in magnesium levels in blood phosphorus concentration in the newborn obtained from the umbilical artery as com- infant, while in breast-fed infants the level pared to the umbilical vein. Both Salmi and tends usually not to vary or to decrease Breton et al.(cid:1) found cord blood levels to somewhat. The serum phosphorus concen- be higher than those of older children and tration in a number of evaporated-milk-fed adults. The number of determinations in infants was seen to approximate the values Breton’s study is not listed and therefore for breast-fed infants by the second week of the significance of his findings is difficult to life.12 In the study of Gittleman and Pincus, evaluate. The studies of Bogart and Piass4 infants fed evaporated milk showed a tend- and of Mays and Kcele(cid:1) are in agreement ency to hypocalcemia while the calcium with ours in that no significant difference level in the breast-fed infants did not de- was found between magnesium levels in crease. Neither Bruck and Weintraub nor cord blood and in the blood of older chil- Graham et al. were able to demonstrate any dren and adults. definite correlation between serum calcium It can be seen from Table IV that only a and phosphorus levels in individual infants. relatively few determinations have previ- However, Bruck and Weintraub studying ously been performed in infants during the premature and full-term infants found that first week of life. In neither of the two stud- a falling calcium level was usually associ- ies (cid:1) was the type of infant feed- ated with a rising phosphorus level. Bak- ing published. Of interest is Salmi’s study in win14 observed a fall in serum calcium fol- which the serum magnesium levels found lowing the oral loading of phosphate to during the first week of life tended to be newborn infants. Sulvesen, Hastings, and higher than the level of the cord blood of Mclntoshl fed neutral phosphate to dogs the same infant. This tendency of rising by stomach tube in daily amounts such that levels during the first week is in keeping the phosphate intake closely approximated with our findings in breast-fed infants. the intake of formula-fed infants when cal- If the differences reported here in mag- culated on the basis of surface area. With nesium levels of the breast-fed and formula- the ingestion of phosphate the dogs showed fed infants are confirmed by additional data an elevation of serum inorganic phosphorus they may be explainable by differences in and a reduction in serum magnesium and the ratios of dietary phosphorus and mag- calcium. nesium. The magnesium concentration of It is, therefore, possible that the observed 974 SERUM MAGNESIUM differences in serum magnesium levels in Acknowledgment breast-fed and evaporated-milk-fed infants The technical assistance of Mrs. Joanne Bunge is related to the phosphate content of cow’s and Mrs. Marsha Schweiss is gratefully acknowl- milk as compared to human milk. The rising edged. serum phosphorus level in infants fed cow’s REFERENCES milk may account for the tendency demon- 1. Salvesen, H. A., Hastings, A. B., and McIntosh, strated in the study for lower serum mag- J. F. : Blood changes and clinical symptoms following oral administration of phosphates. nesium levels in this group as compared to J. Biol. Chem., 60:311, 1924. breast-fed infants during the third, fourth, 2. Orange, M., and Rhine, H. C. : Nlicroestima- and fifth day of life. tion of magnesium in body fluids. J. Biol. Consideration must also be given to the Chem., 56:297, 1951. serum proteins. Approximately 35% of the 3. Anast, C. S. : The unreliability of the Titan serum magnesium is bound to protein. It is yellow method for the determination of magnesium in patients receiving intravenous possible that variations in protein levels in calcium gluconate. Clin. Chem., 9:544, the two groups may account for the ob- 1963. served differences in serum magnesium 4. Bogert, L. J., and Plass, E. D. : Placental levels. This study will be extended to de- transmission : the calcium and magnesium termine serum magnesium, phosphate, and content of fetal and maternal blood serum. J. Biol. Chem., 56:297, 1951. protein levels simultaneously in the new- 5. Salmi, I. : On influence of anoxia on plasma born period. calcium. Ann. Paedat. Fenniae (Suppl. 2), 1:40, 1954-1955. 6. Breton, A., Gardner, B., Lefebvre, C., and SUMMARY Trainsel, M. : Etude de la magnesemi chez Two hundred and thirty-eight determina- le nouveau-ne. Scm. Hop. (Paris), 36:2298, tions of serum magnesium levels during the 1960. first 4 days of life were carried out on 78 7. May’s, J. E., and Keele, D. K. : Serum mag- nesium levels in healthy children and in infants. No statistically significant differ- various disease states. Amer. J. Dis. Child., ences were found when these values were 102:235, 1961. compared to those determined in 111 older 8. Marioni, A., Heredia, J. L., and Semino, 0. E.: children and adults. Differences in mean Calcio, fosforo, magnesio y fosfatasa durante values between any of the first 5 days were ci trabajo del parto y en el cordon umbilical del nino. Rev. Fac. Cienc. Med. Univ. not statistically significant nor were any Cordoba 9:109, 1951. significant differences found when the 9. Macy, I. C., Kelly, H. J., and Sloan, R. E.: values of each day were compared with The composition of milks, Nat. Res. Council the mean value of older children and adults Pub., p. 250, 1953. or the mean value of the cord bloods. The 10. Gardner, L. I., MacLachlan, E. A., Pick, W., mean values on days 3, 4, and 5 were higher et al.: Etiologic factors in tetany of the newborn. PEDIATRICS, 5:228, 1950. in breast-fed infants than in infants fed 11. Bruck, E., and Weintraub, D. H. : Serum evaporated milk. Higher values in breast- calcium and phosphorus in premature and fed infants and lower values in evaporated- full term infants. Amer. J. Dis. Child., 90: milk-fed infants on days 3 and 4 when com. 653, 1955. 12. Graham, G. G., Barness, L. A., and Gyorgy, P.: pared to day 1 in the same infant were Serum calcium and inorganic phosphate in found in a significant number of cases. The the newborn infant, and their relation to possibility that the observed differences in different feedings. J. Pediat., 42:401, 1953. these two groups of infants may be related 13. Gittleman, I. F., and Pincus, J. B. : Influence to the difference in phosphate to magnesium of diet on the occurrence of hyperphos- ratio in cow’s milk as compared to human phatemia and hypocalcemia in the newborn milk is discussed. Further investigation of infant. PEDIAmIcS, 8:778, 1951. 14. Bakwin, H. : Pathogenesis of tetany of the this problem is needed before definite con- newborn. Amer. J. Dis. Child., 54:1211, clusions can be drawn. 1937.