- Email: [email protected]
Anticonvulsant drug therapy in human pregnancy: Effects on serum concentrations of vitamin D metabolites in maternal and cord blood
Chamberlain et al.
tion: development of a fetal biophysical profile score. AM J 0BSTET GYNECOL 1980; 136:787. 6. Haworth JA, Pussey VA. The relationship between birth weight and gestational age of a Winnipeg Hospital population. Can Med AssocJ 1969;100:842-845. 7. Gohari P, Berkowitz RL, Hobbins JC. Prediction of intrauterine growth retardation by determination of total intrauterine volume. AM j 0BSTET GYNECOL 1977; 123:255. 8. Hobbins JC, Grannum PAT, Berkowitz RL, et al. Ultrasound in the diagnosis of congenital anomalies. AM J 0BSTET GYNECOL 1979;134:331.
October I, 1984 Am J Obstet Gynecol
9. Queenan J, Godow E. Amniography for the detection of congenital anomalies. Obstet Gynecol 1970;35:648. 10. Gresham EL, RankinJH, Makowski EL, et al. Evaluation of fetal renal function in a chronic sheep preparation. J Clin Invest 1972;51: 149. 11. Pritchard JA. Deglutition by normal and anencephalic fetuses. Obstet Gynecol 1965;25:289. 12. Dawes GS, Fox HE, Leduc BM, et al. Respiratory movements and rapid eye movement sleep in the fetal lamb. J Physiol (Lond) 1972;220: 119.
Anticonvulsant drug therapy in human pregnancy: Effects on serum concentrations of vitamin D metabolites in maternal and cord blood Trond Markestad, M.D., Magnar Ulstein, M.D., Roald E. Strandjord, M.D., Lage Aksnes, Ph.D., and Dagfinn Aarskog, M.D. Bergen, Norway Serum concentrations of the main vitamin D metabolites and of calcium, phosphate, and alkaline phosphatase were determined in each of the three trimesters of pregnancy and in simultaneously obtained maternal and cord blood at delivery in 22 epileptic women treated with diphenylhydantoin or carbamazepine alone or with a combination with one other drug. The results were compared with similarly obtained data from 22 normal pregnancies. Women in both groups received supplements of 400 IU vitamin 0 3 per day. All the women had 25-hydroxyvitamin D levels within the normal range for healthy adults (> 12 ng/ml) throughout pregnancy. The epileptic women had, however, significantly (p < 0.05) lower median 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D levels and higher median 25,26-dihydroxyvitamin D values than the reference group. The 24,25-dihydroxyvitamin D concentrations did not differ significantly, but the median ratio of 24,25-dihydroxyvitamin D to 25-hydroxyvitamin D was higher in the epileptic women at the end of pregnancy (p = 0.05). The respective differences in cord serum concentrations reflected those of the mothers at delivery. Serum calcium tended to be lower during epileptic pregnancy, but none were hypocalcemic. The alkaline phosphatase and phosphate values did not consistently differ from those of the reference women. The median alkaline phosphatase level of cord serum was slightly higher in the epileptic group, but the calcium and phosphate levels were similar to the reference values. The various biochemical parameters of the carbamazepine-treated women tended to be intermediate between those of the healthy and diphenylhydantoin-treated groups. Antiepileptic drug therapy appears to affect vitamin D metabolism and calcium homeostasis during pregnancy. The derangements may not be of major clinical significance, however, in vitamin D-supplemented and normally functioning women on long-term low-dose therapy. (AM J 0BSTET GYNECOL 1984;150:254-8.)
Long-term anticonvulsant drug treatment has been
more commonly, laboratory abnormalities indicative of
associated with clinical osteomalacia and rickets or,
disturbed mineral metabolism such as hypocalcemia, elevated serum alkaline phosphatase levels, radiographic evidence of osteopenia, decreased bone min-
From the Departments of Pediatrics, Obstetrics and Gynecology, and Neurology, University of Bergen. Supported !Jy The Nor-wegian Research Council for Science and the Humanities. Received for publication October 7, 1983; revised March 20, 1984; accepted April25, 1984. Reprint requests: Dr. Trond Markestad, Department of Pediatrics, University of Bergen, N-5016 Bergen, Norway.
eral density by photon absorptiometry, and osteomalacia on bone biopsy. 1
254
The exact pathogenesis of these disorders has not been established, but altered vitamin D metabolism appears to be of major importance. 1 In particular, it has been suggested that antiepileptic drugs may promote
Anticonvulsant drug therapy in pregnancy
Volume 150 Number 3
hepatic degradation of vitamin D and 25-hydroxyvitamin D with subsequent depletion of vitamin stores. 1 • 2 The role of vitamin D deficiency in anticonvulsant osteomalacia is also supported by observations that vitamin D treatment may heal the dlsorder. 3 It has also been suggested that these drugs may modify the synthesis and metabolism of dihydroxy-vitamin D metabolites,4-6 or inhibit intestinal mineral absorption by mechanisms which are independent of their effects on vitamin D metabolism. 7 During human pregnancy the intestinal absorption of calcium and phosphate is normally increased to meet the requirements of the growing fetus, 8 and this increase appears to be mediated, at least partly, by increased synthesis of active vitamin D metabolites. 9 Pregnancy thus poses additional stresses on the vitamin D endocrine system. The purpose of this investigation was, therefore, to study the effect of long-term anticonvulsant therapy on vitamin D and mineral homeostasis in pregnant women and in the cord blood of their neonates. Material and methods
Twenty-two epileptic women were studied longitudinally throughout pregnancy. Fourtee:n of the women were treated with diphenylhydantoin alone (n = 4) or in combination with phenobarbital (n = 6), clonazepam (n = 2), carbamazepine (n = 1), or primidone (n = 1); the group was collectively referred to as the diphenylhydantoin group. Eight patients were treated with carbamazepine alone (n = 6) or in combination with clonazepam (n = 2) and were defined as the clonazepam group. The median age was 27 years (range, 17 to 39 years), median order of pregnancy was 2 (range, 1 to 5), median duration of epilepsy was 12.5 years (range, 0.5 to 37 years), and median duration of presently used drugs was 9 years (range, 0.5 to 24 years). The median daily dose of diphenylhydantoin was 200 mg (range, 100 to 350 mg) in the first trimester and 225 mg (range, 100 to 350 mg) in the subsequent stages of pregnancy. The median serum concentration of this drug was 11~-tg/ml (range, 3 to 30 ~-tg/ml) in the first, 3 ~-tg/ml (range, 0 to 20 ~-tg/ml) in the second, and 4.5 ~-tg/ml (range, 1 to 26 ~-tg/ml) in the third trimester of pregnancy. The therapeutic range was 10 to 20 ~-tg/ml. Four of the six patients on phenobarbital received 100 mg of the drug per day while the others received 150 and 200 mg, respectively. The median serum level of phenobarbital was 16~-tg/ml (range, 10 to 17 ~-tg/ml) in the second, and 13~-tg/ml (range, 11 to 24~-tg/ml) in the third trimester (therapeutic range, 11.6 to 30.2 ~-tg/ml). The median carbamazepine dose was 400 mg/day with a range of 200 to 1200 mg/day in all stages of pregnancy, and the median serum concentration of the drug was 7 ~-tg/ml (range, 0 to 9 ~-tg/ml) in the first, 5 ~-tg/ml (range, 4 to I 1 ~-tg/ml) in the second, and 6
255
(range, 2 to 7 ~-tg/ml) in the third trimester (therapeutic range, 4.8 to 9.5 ~-tg/ml). Clonazepam was given in doses of 1, 4, 6, and 10 mg during the course of the pregnancy. Serum levels of this drug were not monitored. Primidone was given in a daily dose of 250 mg to one patient, and the serum level was stable at 5 ~-tg/ml (therapeutic range, 5.4 to 10.9 ~-tg/ml). All the women received a daily supplement of 400 IU vitamin D3 after the blood sampling in the first trimester of pregnancy. Twelve of the women (seven of the diphenylhydantoin group and five of the carbamazepine group) delivered during summer (May to October). Median birth weight of the infants was 3480 gm (range 2730 to 4470 gm). One infant suffered severe birth asphyxia, two had a combination of malformations compatible with the fetal hydantoin syndrome, and two infants of carbamazepine-treated mothers had coronal hypospadias. Twenty-two healthy pregnant women were studied longitudinally at the same time-intervals of pregnancy as the epileptic women. Like the latter group they received 400 IU vitamin D 3 per day after the first trimester, and they did not differ from the epileptic women with respect to age (median, 27.5 years; range, 23 to 40 years), order of pregnancy (median, 2; range, 1 to 5), seasonal distribution ( 11 delivered during summer and 11 during winter), or social class. Blood was collected during the first trimester (for practical purposes defined as before the end of the fourteenth week following the first day of the last menstrual period), during the second trimester (week 15 to 28), during the third trimester (week 29 to 40), and simultaneously from maternal and cord blood at delivery. The collections were, however, incomplete. Among the diphenylhydantoin group, blood was available from seven women in the first trimester, from 11 in the second, and from 13 in the third, and from 12 maternal-cord pairs at delivery. The respective number of samples from the carbamazepine group was four, seven, six, five, and six. The missing samples were not related to severity of disease, patients compliance, or socioeconomic factors and, therefore, probably had no significant effects on the final results. The serum was stored at -20° C until analysis. The vitamin D metabolites 25-hydroxyvitamin D, 1,25dihydroxyvitamin D, 24,25-dihydroxyvitamin D, and 25,26-dihydroxyvitamin D were determined by competitive protein-binding assays after extensive purification and separation procedures. 10 The intraassay and interassay coefficients of variation were 6.3% and 8.5%, respectively, for 25-hydroxyvitamin D, 9.7% and 12.1% for 1,25-dihydroxyvitamin D, 7.0% and 9.7% for 24,25-dihydroxyvitamin D, and 7.3% and 10.2% for 25,26-dihydroxyvitamin D. Vitamin D-binding protein was measured by quantitative immunoelectro~-tg/ml
256
Markestad et al.
0.014
October 1, 1984 Am J Obstet Gynecol
0.002
phoresis, and calcium, phosphate, albumin, and alkaline phosphatase levels were determined by standard autoanalyzer techniques. Calcium values were corrected for differences in corresponding albumin levels. The results were presented as medians with a 50% interquartile range. Differences were tested for statistical significance with use of a two-sided Wilcoxon test. Associations were calculated by means of Kendall's tau (T).
0.055
0.009
60 0
5 I l.{'l
40 20
N
Results
E '
~
0
_N
5
0.010
0.016
<0.001
<0.001
100 5Q
.Q L----r-----r-----r~·~-r~~
E
~
0.028
0
3
5
2
_N
I
l.{'l
N
~-
0.050
8 6
0.005
L.
2 ~----~----~r------o--7;~'--~-;~'~-,--
~
0 J:" I 0
0.026
0.81
2
0 006
0.003
0.007
,...____..,.~~
0.6 0.4 0.2
0.005
1ST
<0.001
<0001
2 NO 3 RD TRIMESTER
<0.001
/
I
<0.001
/1--.-
DELIVERY CORD
Fig. I. Serum concentrations of vitamin D metabolites (medians and 50% interquartile ranges) in normal (e--•) and epileptic (o-o) pregnancy; p values for differences under or at the 5% significance level are plotted on the graphs.
None of the epileptic women or their babies had clinical symptoms suggestive of hypocalcemia or bone disease. The mothers of the two infants with fetal hydantoin syndrome received a combination of diphenylhydantoin and phenobarbital, and the doses of both drugs were the highest of the group. Biochemical data were not available for the first trimester, but the serum concentrations of the drugs were later within the therapeutic range, and the data on vitamin D and mineral homeostasis did not differ from those of the rest of the group. The mother of one of the infants with hypospadia received the highest carbamazepine dose (1200 mg/day plus 1 mg of clonazepam) while the other received only 400 mg carbamazepine. Both, however, had carbamazepine levels within the therapeutic range, and the other biochemical parameters did not differ from those of the other women. Single or combined drug treatment, or differences in dose or serum levels of therapeutic agents, were not associated with apparent variations in the biochemical parameters in either the diphenylhydantoin or the carbamazepine group. All the epileptic and healthy women had 25-hydroxyvitamin D concentrations above the low normal limit of 12 ng/ml established for nonpregnant and unsupplemented adults in our laboratory. The epileptic women, however, had lower median concentrations of 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D, and higher 25;26-dihydroxyvitamin D levels than their healthy peers throughout pregnancy (Fig. 1). The absolute concentrations of 24,25-dihydroxyvitamin D did not differ consistently between the two groups, but the median 24,25-dihydroxyvitamin D-to-25-hydroxyvitamin D ratio was higher in the epileptic women at delivery (Fig. 1). The variation in vitamin D metabolite concentrations in cord serum reflected the differences in maternal levels at delivery (Fig. 1). There were no significant relationships between corresponding levels of 1,25-dihydroxyvitamin D and any of the other vitamin D metabolites. The 25-hydroxyvitamin D concentrations were, however, positively related to the concomitant 24,25-dihydroxyvitamin D levels in all stages of pregnancy and in cord serum for
Anticonvulsant drug therapy in pregnancy
Volume 150 Number 3
both groups (T = 0.33 to 0. 75; p = 0.05 for the lowest value, otherwise p < 0.0 1) and to a lesser extent to the 25,26-dihydroxyvitamin D values. The latter relationship was only significant, however, after the first trimester in the epileptics (T = 0.35-0.51, p < 0.04), and only at delivery and in cord serum for the healthy group (T = 0.27, p = 0.05, and T = 0.44, p = 0.004, respectively). For both groups there was a positive and comparable correlation between paired maternal-cord serum levels of 25-hydroxyvitamin D (control subjects: T = 0.54, p < 0.001; epileptics: T = 0.75, p < 0.001), 24,25-dihydroxyvitamin D (T = 0.70 and 0.72, respectively, p < 0.001), and 25,26-dihydroxyvitamin D (T = 0.58 and 0.67, p < 0.001). The corresponding levels of 1,25dihydroxyvitamin D were significantly associated only for the epileptics (T = 0.69, p < 0.001; control subjects: T = 0.16, p = 0.15). The median serum calcium concentration was generally lower in the epileptics throughout pregnancy, while the phosphate and alkaline phosphatase values were not consistently different from those of the reference group (Fig. 2). The cord levels of calcium and phosphate were similar for the two groups, but the median alkaline phosphatase value was somewhat higher in the epileptic group (Fig. 2). The serum levels of albumin, total protein, and vitamin D-binding protein did not vary between the groups. None of the individual biochemical parameters were consistently different when the diphenylhydantoin and carbamazepine groups were compared, although the median values of the carbamazepine-treated patients tended to be intermediate between those of the diphenylhydantoin group and the control group. Comment
Vitamin Dis first hydroxylated to 25-hydroxyvitamin D in the liver and subsequently to 1,25-dihydroxyvitamin D in the kidneys to become metabolically active. 11 Other metabolites, such as 24,25-dihydroxyvitamin D and 25,26-dihydroxyvitamin D may be synthesized from 25-hydroxyvitamin D alternatively to 1,25-dihydroxyvitamin D, but their biologic significance, if any, is unclearY The circulating 1,25-dihydroxyvitamin D concentration is normally rigidly regulated according to the calcium and phosphate needs of the body and to the dietary supply of these mineralsY 24,25-Dihydroxyvitamin D synthesis is also regulated to a certain extent and then reciprocally to the 1,25-dihydroxyvitamin D production. 11 • 12 The serum concentration of this metabolite, as well as that of 25,26-dihydroxyvitamin D, appears, however, to be determined primarily by the concentration of the precursor 25-hydroxyvitamin D. 13
E
0.024
12
0.005
!~
0
!2 11
--E' 10 8
9
1
!~~ '
,
p
/~
6
E 0
5
Cl
4
a..
3
!2
--E
~
H-=t9f ''
::::400
257
'~ 0.026
<0.001
~~
::l
~200 <(
~-----r------~-----,,-~~~.--~.~~
1ST
2 NO 3 RD TRIMESTER
DELIVERY
CORD
Fig. 2. Serum concentrations of calcium (Ca), phosphate (P), and alkaline phosphates (ALP). The calcium concentrations have been adjusted to an albumin level of 46 mg/100 mi. (Symbols as in Fig. 1.)
It is therefore necessary to correct for differences in precursor levels (e.g., by calculating the molar ratios of these dihydroxymetabolites and 25-hydroxyvitamin D) before meaningful comparisons of 24,25-dihydroxyvitamin D and 25,26-dihydroxyvitamin D levels, respectively, can be made. The serum concentration of 25-hydroxyvitamin D is an accepted index of vitamin D nutritional status. 11 In agreement with previous reports the present study suggests that treatment with antiepileptic drugs tends to lower serum concentrations of 25-hydroxyvitamin D. 2 • 5 • !4-ts A daily supplement of 400 IU vitamin D 3 was, however, adequate to maintain normal serum levels of this metabolite during pregnancy. Reported effects of antiepileptic drug therapy on serum concentrations of other vitamin D metabolites are conflicting. Elevated, normal, and decreased 1,25dihydroxyvitamin D levels, and normal and decreased 24,25-dihydroxyvitamin D values have been observed in nonpregnant adults and children. 4 - 6 • 14 - 16 These confusing results may, however, reflect not only the direct effects of the various drugs on vitamin D metabolism but also differences in calcium and phosphate homeostasis secondary to various unspecific factors related to the severity of neurological disease, such as physical mobility, dietary content of minerals, and vitamin D supply from sunshine exposure and diet. The epileptic women of the present study did not differ from the reference group with regard to physical, so-
258
October I, I 984 Am J Obstet Gynecol
Markestad et al.
cia!, or mental function, and the observed variations in vitamin D metabolite levels were, therefore, probably related to drug effects. The lower l ,25-dihydroxyvitamin D and higher ratio of 24,25-dihydroxy vitamin D and 25,26-dihydroxy vitamin D to 25-hydroxyvitamin Din the treated group suggest that the regulation of dihydroxyvitam in D synthesis, or metabolism, is altered during epileptic pregnancy. It is unlikely that the decreased 1,25-dihydroxyv itamin D levels were caused by substrate deficiency, since there was no correlation between 25-hydroxyvitam in D and l ,25-dihydroxyvit amin D concentrations, and since all had 25-hydroxyvitamin D levels within the accepted normal range. The epileptic women had a slightly, but significantly, lower median serum calcium level than their healthy counterparts. None were, however, hypocalcemic according to the standards of the laboratory, and the phosphate and alkaline phosphatase values were normal, suggesting that the altered calcium homeostasis possibly were without major clinical significance. The mineral homeostasis of cord serum did not differ between the healthy and epileptic groups, despite the variance in vitamin D metabolite levels, indicating that other factors than active vitamin D compounds may be of prime importance for transplacental calcium and phosphate transport. It has been suggested that carbamazepine may have less untoward effects on vitamin D and calcium metabolism than diphenylhydanto in. 16 A similar tendency was observed in the present study, but this finding should be interpreted with caution since the groups were small, and since most members of the diphenylhydantoin group received two drugs whereas all but two of the carbamazepine- treated women were on this compound alone. Although antiepileptic drug therapy appears to affect vitamin D and calcium homeostasis in human pregnancy, the present study suggests that these derangements possibly are of limited clinical significance in normally functioning women on long-term low-dose therapy and that a daily vitamin D supplement of 400 IU satisfies the increased vitamin D requirement of such patients. Autoanalyzer tests were carried out at The Laboratory of Clinical Biochemistry, University of Bergen, with the technical assistance of Mr. Kaare S0nstab0.
The statistical analyses were carried out by cand. real. Trond Haider, Nordic Statistical Center, Ciba-Geigy Pharma A/S, Oslo, Norway. REFERENCES I. Hahn TJ. Drug-induced disorders of vitamin D and min-
eral metabolism. Clin Endocrinol Metab 1980;9: 107. 2. Hahn TJ, Birge SJ, Scharp CR, Avioli LV. Phenobarbitalinduced alterations in vitamin D metabolism.] Clin Invest 1972;51:741. 3. Christiansen C, R0dbro P, Lund M. Incidence of anticonvulsant osteomalacia and effect of vitamin D: controlled therapeutic trial. Br Med J 1973;4:695. 4. Weisman Y, Fattal A, Zeisenberg Z, Hare! S, Spirer Z, Harell A. Decreased serum 24,25-dihydroxy vitamin D concentrations in children receiving chronic anticonvulsant therapy. Br Med J 1979;2:521. 5. Christensen CK, Lund B, Lund BJ, Sq,rensen OH, Nielsen HE, Mosekilde L. Reduced 1,25-dihydroxyvit amin D and 24,25-dihydroxyv itamin D in epileptic patients receiving chronic combined anticonvulsant therapy. Metab Bone Dis Relat Res 1981;3:17. 6. Zerwekh JE, Homan R, Tindall R, Pak CYC. Decreased serum 24,25-dihydroxyv itamin D concentration during long-term anticonvulsant therapy in adult epileptics. Ann Neurol1982;12:18 4. 7. Koch H-U, Kraft D, Von Herrath D, Schaeffer K. Influence of diphenylhydantoi n and phenobarbital on intestinal calcium transport in the rat. Epilepsia 1972;13:829. 8. Heaney RP, Skillman TG. Calcium metabolism in normal human pregnancy. J Clin Endocrinol 1971;33:661. 9. Kumar R, Cohen WR, Silva P, Epstein FH. Elevated 1,25-dihydroxyvit amin D plasma levels in normal human pregnancy and lactation. J Clin Invest 1979;63:342. I 0. Aksnes L. Quantitation of the main metabolites of vitamin D in a single serum sample. II. Determination by UVabsorption and competitive protein binding assays. Clin Chim Acta 1980;104:147. 11. DeLuca HF. Vitamin D: metabolism and function. Monographs on Endocrinology, vol. 13. New York: SpringerVerlag, 1979. 12. Aksnes L, Aarskog D. Plasma concentrations of vitamin D metabolites in puberty: effect of sexual maturation and implications for growth. J Clin Endocrinol Metab 1982;55:94. 13. Stanbury SW, Mawer EB. The metabolism of a physiological dose of radioactive cholecalciferol (vitamin D3 ) to its hydroxylated metabolites in man. Clin Sci 1980;58:523. 14. Bell RD, Pak CYC, Zerwekh J, Barilla DE, Vasko M. Effect of phenytoin on bone and vitamin D metabolism. Ann Neurol 1979;5:374. 15. Jubitz W, Haussler MR, McCain T A, Tolman KG. Plasma 1,25-dihydroxyvit amin D levels in patients receiving anticonvulsant drugs. J Clin Endocrinol Metab 1977; 44:617. 16. Tjellesen L, Christiansen C. Serum vitamin D metabolites in epileptic patients treated with two different anticonvulsants. Acta Neurol Scand 1982;66:335.
tion: development of a fetal biophysical profile score. AM J 0BSTET GYNECOL 1980; 136:787. 6. Haworth JA, Pussey VA. The relationship between birth weight and gestational age of a Winnipeg Hospital population. Can Med AssocJ 1969;100:842-845. 7. Gohari P, Berkowitz RL, Hobbins JC. Prediction of intrauterine growth retardation by determination of total intrauterine volume. AM j 0BSTET GYNECOL 1977; 123:255. 8. Hobbins JC, Grannum PAT, Berkowitz RL, et al. Ultrasound in the diagnosis of congenital anomalies. AM J 0BSTET GYNECOL 1979;134:331.
October I, 1984 Am J Obstet Gynecol
9. Queenan J, Godow E. Amniography for the detection of congenital anomalies. Obstet Gynecol 1970;35:648. 10. Gresham EL, RankinJH, Makowski EL, et al. Evaluation of fetal renal function in a chronic sheep preparation. J Clin Invest 1972;51: 149. 11. Pritchard JA. Deglutition by normal and anencephalic fetuses. Obstet Gynecol 1965;25:289. 12. Dawes GS, Fox HE, Leduc BM, et al. Respiratory movements and rapid eye movement sleep in the fetal lamb. J Physiol (Lond) 1972;220: 119.
Anticonvulsant drug therapy in human pregnancy: Effects on serum concentrations of vitamin D metabolites in maternal and cord blood Trond Markestad, M.D., Magnar Ulstein, M.D., Roald E. Strandjord, M.D., Lage Aksnes, Ph.D., and Dagfinn Aarskog, M.D. Bergen, Norway Serum concentrations of the main vitamin D metabolites and of calcium, phosphate, and alkaline phosphatase were determined in each of the three trimesters of pregnancy and in simultaneously obtained maternal and cord blood at delivery in 22 epileptic women treated with diphenylhydantoin or carbamazepine alone or with a combination with one other drug. The results were compared with similarly obtained data from 22 normal pregnancies. Women in both groups received supplements of 400 IU vitamin 0 3 per day. All the women had 25-hydroxyvitamin D levels within the normal range for healthy adults (> 12 ng/ml) throughout pregnancy. The epileptic women had, however, significantly (p < 0.05) lower median 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D levels and higher median 25,26-dihydroxyvitamin D values than the reference group. The 24,25-dihydroxyvitamin D concentrations did not differ significantly, but the median ratio of 24,25-dihydroxyvitamin D to 25-hydroxyvitamin D was higher in the epileptic women at the end of pregnancy (p = 0.05). The respective differences in cord serum concentrations reflected those of the mothers at delivery. Serum calcium tended to be lower during epileptic pregnancy, but none were hypocalcemic. The alkaline phosphatase and phosphate values did not consistently differ from those of the reference women. The median alkaline phosphatase level of cord serum was slightly higher in the epileptic group, but the calcium and phosphate levels were similar to the reference values. The various biochemical parameters of the carbamazepine-treated women tended to be intermediate between those of the healthy and diphenylhydantoin-treated groups. Antiepileptic drug therapy appears to affect vitamin D metabolism and calcium homeostasis during pregnancy. The derangements may not be of major clinical significance, however, in vitamin D-supplemented and normally functioning women on long-term low-dose therapy. (AM J 0BSTET GYNECOL 1984;150:254-8.)
Long-term anticonvulsant drug treatment has been
more commonly, laboratory abnormalities indicative of
associated with clinical osteomalacia and rickets or,
disturbed mineral metabolism such as hypocalcemia, elevated serum alkaline phosphatase levels, radiographic evidence of osteopenia, decreased bone min-
From the Departments of Pediatrics, Obstetrics and Gynecology, and Neurology, University of Bergen. Supported !Jy The Nor-wegian Research Council for Science and the Humanities. Received for publication October 7, 1983; revised March 20, 1984; accepted April25, 1984. Reprint requests: Dr. Trond Markestad, Department of Pediatrics, University of Bergen, N-5016 Bergen, Norway.
eral density by photon absorptiometry, and osteomalacia on bone biopsy. 1
254
The exact pathogenesis of these disorders has not been established, but altered vitamin D metabolism appears to be of major importance. 1 In particular, it has been suggested that antiepileptic drugs may promote
Anticonvulsant drug therapy in pregnancy
Volume 150 Number 3
hepatic degradation of vitamin D and 25-hydroxyvitamin D with subsequent depletion of vitamin stores. 1 • 2 The role of vitamin D deficiency in anticonvulsant osteomalacia is also supported by observations that vitamin D treatment may heal the dlsorder. 3 It has also been suggested that these drugs may modify the synthesis and metabolism of dihydroxy-vitamin D metabolites,4-6 or inhibit intestinal mineral absorption by mechanisms which are independent of their effects on vitamin D metabolism. 7 During human pregnancy the intestinal absorption of calcium and phosphate is normally increased to meet the requirements of the growing fetus, 8 and this increase appears to be mediated, at least partly, by increased synthesis of active vitamin D metabolites. 9 Pregnancy thus poses additional stresses on the vitamin D endocrine system. The purpose of this investigation was, therefore, to study the effect of long-term anticonvulsant therapy on vitamin D and mineral homeostasis in pregnant women and in the cord blood of their neonates. Material and methods
Twenty-two epileptic women were studied longitudinally throughout pregnancy. Fourtee:n of the women were treated with diphenylhydantoin alone (n = 4) or in combination with phenobarbital (n = 6), clonazepam (n = 2), carbamazepine (n = 1), or primidone (n = 1); the group was collectively referred to as the diphenylhydantoin group. Eight patients were treated with carbamazepine alone (n = 6) or in combination with clonazepam (n = 2) and were defined as the clonazepam group. The median age was 27 years (range, 17 to 39 years), median order of pregnancy was 2 (range, 1 to 5), median duration of epilepsy was 12.5 years (range, 0.5 to 37 years), and median duration of presently used drugs was 9 years (range, 0.5 to 24 years). The median daily dose of diphenylhydantoin was 200 mg (range, 100 to 350 mg) in the first trimester and 225 mg (range, 100 to 350 mg) in the subsequent stages of pregnancy. The median serum concentration of this drug was 11~-tg/ml (range, 3 to 30 ~-tg/ml) in the first, 3 ~-tg/ml (range, 0 to 20 ~-tg/ml) in the second, and 4.5 ~-tg/ml (range, 1 to 26 ~-tg/ml) in the third trimester of pregnancy. The therapeutic range was 10 to 20 ~-tg/ml. Four of the six patients on phenobarbital received 100 mg of the drug per day while the others received 150 and 200 mg, respectively. The median serum level of phenobarbital was 16~-tg/ml (range, 10 to 17 ~-tg/ml) in the second, and 13~-tg/ml (range, 11 to 24~-tg/ml) in the third trimester (therapeutic range, 11.6 to 30.2 ~-tg/ml). The median carbamazepine dose was 400 mg/day with a range of 200 to 1200 mg/day in all stages of pregnancy, and the median serum concentration of the drug was 7 ~-tg/ml (range, 0 to 9 ~-tg/ml) in the first, 5 ~-tg/ml (range, 4 to I 1 ~-tg/ml) in the second, and 6
255
(range, 2 to 7 ~-tg/ml) in the third trimester (therapeutic range, 4.8 to 9.5 ~-tg/ml). Clonazepam was given in doses of 1, 4, 6, and 10 mg during the course of the pregnancy. Serum levels of this drug were not monitored. Primidone was given in a daily dose of 250 mg to one patient, and the serum level was stable at 5 ~-tg/ml (therapeutic range, 5.4 to 10.9 ~-tg/ml). All the women received a daily supplement of 400 IU vitamin D3 after the blood sampling in the first trimester of pregnancy. Twelve of the women (seven of the diphenylhydantoin group and five of the carbamazepine group) delivered during summer (May to October). Median birth weight of the infants was 3480 gm (range 2730 to 4470 gm). One infant suffered severe birth asphyxia, two had a combination of malformations compatible with the fetal hydantoin syndrome, and two infants of carbamazepine-treated mothers had coronal hypospadias. Twenty-two healthy pregnant women were studied longitudinally at the same time-intervals of pregnancy as the epileptic women. Like the latter group they received 400 IU vitamin D 3 per day after the first trimester, and they did not differ from the epileptic women with respect to age (median, 27.5 years; range, 23 to 40 years), order of pregnancy (median, 2; range, 1 to 5), seasonal distribution ( 11 delivered during summer and 11 during winter), or social class. Blood was collected during the first trimester (for practical purposes defined as before the end of the fourteenth week following the first day of the last menstrual period), during the second trimester (week 15 to 28), during the third trimester (week 29 to 40), and simultaneously from maternal and cord blood at delivery. The collections were, however, incomplete. Among the diphenylhydantoin group, blood was available from seven women in the first trimester, from 11 in the second, and from 13 in the third, and from 12 maternal-cord pairs at delivery. The respective number of samples from the carbamazepine group was four, seven, six, five, and six. The missing samples were not related to severity of disease, patients compliance, or socioeconomic factors and, therefore, probably had no significant effects on the final results. The serum was stored at -20° C until analysis. The vitamin D metabolites 25-hydroxyvitamin D, 1,25dihydroxyvitamin D, 24,25-dihydroxyvitamin D, and 25,26-dihydroxyvitamin D were determined by competitive protein-binding assays after extensive purification and separation procedures. 10 The intraassay and interassay coefficients of variation were 6.3% and 8.5%, respectively, for 25-hydroxyvitamin D, 9.7% and 12.1% for 1,25-dihydroxyvitamin D, 7.0% and 9.7% for 24,25-dihydroxyvitamin D, and 7.3% and 10.2% for 25,26-dihydroxyvitamin D. Vitamin D-binding protein was measured by quantitative immunoelectro~-tg/ml
256
Markestad et al.
0.014
October 1, 1984 Am J Obstet Gynecol
0.002
phoresis, and calcium, phosphate, albumin, and alkaline phosphatase levels were determined by standard autoanalyzer techniques. Calcium values were corrected for differences in corresponding albumin levels. The results were presented as medians with a 50% interquartile range. Differences were tested for statistical significance with use of a two-sided Wilcoxon test. Associations were calculated by means of Kendall's tau (T).
0.055
0.009
60 0
5 I l.{'l
40 20
N
Results
E '
~
0
_N
5
0.010
0.016
<0.001
<0.001
100 5Q
.Q L----r-----r-----r~·~-r~~
E
~
0.028
0
3
5
2
_N
I
l.{'l
N
~-
0.050
8 6
0.005
L.
2 ~----~----~r------o--7;~'--~-;~'~-,--
~
0 J:" I 0
0.026
0.81
2
0 006
0.003
0.007
,...____..,.~~
0.6 0.4 0.2
0.005
1ST
<0.001
<0001
2 NO 3 RD TRIMESTER
<0.001
/
I
<0.001
/1--.-
DELIVERY CORD
Fig. I. Serum concentrations of vitamin D metabolites (medians and 50% interquartile ranges) in normal (e--•) and epileptic (o-o) pregnancy; p values for differences under or at the 5% significance level are plotted on the graphs.
None of the epileptic women or their babies had clinical symptoms suggestive of hypocalcemia or bone disease. The mothers of the two infants with fetal hydantoin syndrome received a combination of diphenylhydantoin and phenobarbital, and the doses of both drugs were the highest of the group. Biochemical data were not available for the first trimester, but the serum concentrations of the drugs were later within the therapeutic range, and the data on vitamin D and mineral homeostasis did not differ from those of the rest of the group. The mother of one of the infants with hypospadia received the highest carbamazepine dose (1200 mg/day plus 1 mg of clonazepam) while the other received only 400 mg carbamazepine. Both, however, had carbamazepine levels within the therapeutic range, and the other biochemical parameters did not differ from those of the other women. Single or combined drug treatment, or differences in dose or serum levels of therapeutic agents, were not associated with apparent variations in the biochemical parameters in either the diphenylhydantoin or the carbamazepine group. All the epileptic and healthy women had 25-hydroxyvitamin D concentrations above the low normal limit of 12 ng/ml established for nonpregnant and unsupplemented adults in our laboratory. The epileptic women, however, had lower median concentrations of 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D, and higher 25;26-dihydroxyvitamin D levels than their healthy peers throughout pregnancy (Fig. 1). The absolute concentrations of 24,25-dihydroxyvitamin D did not differ consistently between the two groups, but the median 24,25-dihydroxyvitamin D-to-25-hydroxyvitamin D ratio was higher in the epileptic women at delivery (Fig. 1). The variation in vitamin D metabolite concentrations in cord serum reflected the differences in maternal levels at delivery (Fig. 1). There were no significant relationships between corresponding levels of 1,25-dihydroxyvitamin D and any of the other vitamin D metabolites. The 25-hydroxyvitamin D concentrations were, however, positively related to the concomitant 24,25-dihydroxyvitamin D levels in all stages of pregnancy and in cord serum for
Anticonvulsant drug therapy in pregnancy
Volume 150 Number 3
both groups (T = 0.33 to 0. 75; p = 0.05 for the lowest value, otherwise p < 0.0 1) and to a lesser extent to the 25,26-dihydroxyvitamin D values. The latter relationship was only significant, however, after the first trimester in the epileptics (T = 0.35-0.51, p < 0.04), and only at delivery and in cord serum for the healthy group (T = 0.27, p = 0.05, and T = 0.44, p = 0.004, respectively). For both groups there was a positive and comparable correlation between paired maternal-cord serum levels of 25-hydroxyvitamin D (control subjects: T = 0.54, p < 0.001; epileptics: T = 0.75, p < 0.001), 24,25-dihydroxyvitamin D (T = 0.70 and 0.72, respectively, p < 0.001), and 25,26-dihydroxyvitamin D (T = 0.58 and 0.67, p < 0.001). The corresponding levels of 1,25dihydroxyvitamin D were significantly associated only for the epileptics (T = 0.69, p < 0.001; control subjects: T = 0.16, p = 0.15). The median serum calcium concentration was generally lower in the epileptics throughout pregnancy, while the phosphate and alkaline phosphatase values were not consistently different from those of the reference group (Fig. 2). The cord levels of calcium and phosphate were similar for the two groups, but the median alkaline phosphatase value was somewhat higher in the epileptic group (Fig. 2). The serum levels of albumin, total protein, and vitamin D-binding protein did not vary between the groups. None of the individual biochemical parameters were consistently different when the diphenylhydantoin and carbamazepine groups were compared, although the median values of the carbamazepine-treated patients tended to be intermediate between those of the diphenylhydantoin group and the control group. Comment
Vitamin Dis first hydroxylated to 25-hydroxyvitamin D in the liver and subsequently to 1,25-dihydroxyvitamin D in the kidneys to become metabolically active. 11 Other metabolites, such as 24,25-dihydroxyvitamin D and 25,26-dihydroxyvitamin D may be synthesized from 25-hydroxyvitamin D alternatively to 1,25-dihydroxyvitamin D, but their biologic significance, if any, is unclearY The circulating 1,25-dihydroxyvitamin D concentration is normally rigidly regulated according to the calcium and phosphate needs of the body and to the dietary supply of these mineralsY 24,25-Dihydroxyvitamin D synthesis is also regulated to a certain extent and then reciprocally to the 1,25-dihydroxyvitamin D production. 11 • 12 The serum concentration of this metabolite, as well as that of 25,26-dihydroxyvitamin D, appears, however, to be determined primarily by the concentration of the precursor 25-hydroxyvitamin D. 13
E
0.024
12
0.005
!~
0
!2 11
--E' 10 8
9
1
!~~ '
,
p
/~
6
E 0
5
Cl
4
a..
3
!2
--E
~
H-=t9f ''
::::400
257
'~ 0.026
<0.001
~~
::l
~200 <(
~-----r------~-----,,-~~~.--~.~~
1ST
2 NO 3 RD TRIMESTER
DELIVERY
CORD
Fig. 2. Serum concentrations of calcium (Ca), phosphate (P), and alkaline phosphates (ALP). The calcium concentrations have been adjusted to an albumin level of 46 mg/100 mi. (Symbols as in Fig. 1.)
It is therefore necessary to correct for differences in precursor levels (e.g., by calculating the molar ratios of these dihydroxymetabolites and 25-hydroxyvitamin D) before meaningful comparisons of 24,25-dihydroxyvitamin D and 25,26-dihydroxyvitamin D levels, respectively, can be made. The serum concentration of 25-hydroxyvitamin D is an accepted index of vitamin D nutritional status. 11 In agreement with previous reports the present study suggests that treatment with antiepileptic drugs tends to lower serum concentrations of 25-hydroxyvitamin D. 2 • 5 • !4-ts A daily supplement of 400 IU vitamin D 3 was, however, adequate to maintain normal serum levels of this metabolite during pregnancy. Reported effects of antiepileptic drug therapy on serum concentrations of other vitamin D metabolites are conflicting. Elevated, normal, and decreased 1,25dihydroxyvitamin D levels, and normal and decreased 24,25-dihydroxyvitamin D values have been observed in nonpregnant adults and children. 4 - 6 • 14 - 16 These confusing results may, however, reflect not only the direct effects of the various drugs on vitamin D metabolism but also differences in calcium and phosphate homeostasis secondary to various unspecific factors related to the severity of neurological disease, such as physical mobility, dietary content of minerals, and vitamin D supply from sunshine exposure and diet. The epileptic women of the present study did not differ from the reference group with regard to physical, so-
258
October I, I 984 Am J Obstet Gynecol
Markestad et al.
cia!, or mental function, and the observed variations in vitamin D metabolite levels were, therefore, probably related to drug effects. The lower l ,25-dihydroxyvitamin D and higher ratio of 24,25-dihydroxy vitamin D and 25,26-dihydroxy vitamin D to 25-hydroxyvitamin Din the treated group suggest that the regulation of dihydroxyvitam in D synthesis, or metabolism, is altered during epileptic pregnancy. It is unlikely that the decreased 1,25-dihydroxyv itamin D levels were caused by substrate deficiency, since there was no correlation between 25-hydroxyvitam in D and l ,25-dihydroxyvit amin D concentrations, and since all had 25-hydroxyvitamin D levels within the accepted normal range. The epileptic women had a slightly, but significantly, lower median serum calcium level than their healthy counterparts. None were, however, hypocalcemic according to the standards of the laboratory, and the phosphate and alkaline phosphatase values were normal, suggesting that the altered calcium homeostasis possibly were without major clinical significance. The mineral homeostasis of cord serum did not differ between the healthy and epileptic groups, despite the variance in vitamin D metabolite levels, indicating that other factors than active vitamin D compounds may be of prime importance for transplacental calcium and phosphate transport. It has been suggested that carbamazepine may have less untoward effects on vitamin D and calcium metabolism than diphenylhydanto in. 16 A similar tendency was observed in the present study, but this finding should be interpreted with caution since the groups were small, and since most members of the diphenylhydantoin group received two drugs whereas all but two of the carbamazepine- treated women were on this compound alone. Although antiepileptic drug therapy appears to affect vitamin D and calcium homeostasis in human pregnancy, the present study suggests that these derangements possibly are of limited clinical significance in normally functioning women on long-term low-dose therapy and that a daily vitamin D supplement of 400 IU satisfies the increased vitamin D requirement of such patients. Autoanalyzer tests were carried out at The Laboratory of Clinical Biochemistry, University of Bergen, with the technical assistance of Mr. Kaare S0nstab0.
The statistical analyses were carried out by cand. real. Trond Haider, Nordic Statistical Center, Ciba-Geigy Pharma A/S, Oslo, Norway. REFERENCES I. Hahn TJ. Drug-induced disorders of vitamin D and min-
eral metabolism. Clin Endocrinol Metab 1980;9: 107. 2. Hahn TJ, Birge SJ, Scharp CR, Avioli LV. Phenobarbitalinduced alterations in vitamin D metabolism.] Clin Invest 1972;51:741. 3. Christiansen C, R0dbro P, Lund M. Incidence of anticonvulsant osteomalacia and effect of vitamin D: controlled therapeutic trial. Br Med J 1973;4:695. 4. Weisman Y, Fattal A, Zeisenberg Z, Hare! S, Spirer Z, Harell A. Decreased serum 24,25-dihydroxy vitamin D concentrations in children receiving chronic anticonvulsant therapy. Br Med J 1979;2:521. 5. Christensen CK, Lund B, Lund BJ, Sq,rensen OH, Nielsen HE, Mosekilde L. Reduced 1,25-dihydroxyvit amin D and 24,25-dihydroxyv itamin D in epileptic patients receiving chronic combined anticonvulsant therapy. Metab Bone Dis Relat Res 1981;3:17. 6. Zerwekh JE, Homan R, Tindall R, Pak CYC. Decreased serum 24,25-dihydroxyv itamin D concentration during long-term anticonvulsant therapy in adult epileptics. Ann Neurol1982;12:18 4. 7. Koch H-U, Kraft D, Von Herrath D, Schaeffer K. Influence of diphenylhydantoi n and phenobarbital on intestinal calcium transport in the rat. Epilepsia 1972;13:829. 8. Heaney RP, Skillman TG. Calcium metabolism in normal human pregnancy. J Clin Endocrinol 1971;33:661. 9. Kumar R, Cohen WR, Silva P, Epstein FH. Elevated 1,25-dihydroxyvit amin D plasma levels in normal human pregnancy and lactation. J Clin Invest 1979;63:342. I 0. Aksnes L. Quantitation of the main metabolites of vitamin D in a single serum sample. II. Determination by UVabsorption and competitive protein binding assays. Clin Chim Acta 1980;104:147. 11. DeLuca HF. Vitamin D: metabolism and function. Monographs on Endocrinology, vol. 13. New York: SpringerVerlag, 1979. 12. Aksnes L, Aarskog D. Plasma concentrations of vitamin D metabolites in puberty: effect of sexual maturation and implications for growth. J Clin Endocrinol Metab 1982;55:94. 13. Stanbury SW, Mawer EB. The metabolism of a physiological dose of radioactive cholecalciferol (vitamin D3 ) to its hydroxylated metabolites in man. Clin Sci 1980;58:523. 14. Bell RD, Pak CYC, Zerwekh J, Barilla DE, Vasko M. Effect of phenytoin on bone and vitamin D metabolism. Ann Neurol 1979;5:374. 15. Jubitz W, Haussler MR, McCain T A, Tolman KG. Plasma 1,25-dihydroxyvit amin D levels in patients receiving anticonvulsant drugs. J Clin Endocrinol Metab 1977; 44:617. 16. Tjellesen L, Christiansen C. Serum vitamin D metabolites in epileptic patients treated with two different anticonvulsants. Acta Neurol Scand 1982;66:335.