Maternal Toxemia Is Associated With Reduced Incidence of Germinal Matrix Hemorrhage in Premature Babies Karl C.K.
Kuban, MD; Alan Leviton, MD; Marcello Pagano, PhD; Terence Fenton, EdD; Ruth Strassfeld; Mildred Wolff, MA
Abstract To evaluate
prenatal and perinatal risk factors for development of germinal matrix hemorrhage-intraventricular hemorrhage (GMH-IVH), we conducted a prospective epidemiologic study of 449 babies whose birth weight was less than 1501 grams. This study permitted us to test our previously generated hypothesis that babies born to mothers with preeclampsia were at substantially reduced risk of developing GMH-IVH. Seventy-two (16%) of the babies in this population developed GMH-IVH. One (2.5%) of the 40 mothers with a diagnosis of preeclampsia and 71 (17.4%) of 409 mothers without preeclampsia gave birth to babies who developed GMH-IVH. GMH-IVH was seen in 6/107 (5.6%) of babies born to women with hypertension including 4/69 (5.8%) of babies born to women with pregnancy-induced hypertension, compared to 66/352 (18.8%) of babies born to mothers who did not have hypertension. Only 7.3% (8/108) of babies born to women who had proteinuria had GMH-IVH, compared to 18.3% (64/350) of babies whose mothers did not have proteinuria. GMH-IVH was seen in 5/89 (5.6%) of babies whose mothers had both hypertension and proteinuria, whereas 63/332 (19%) of babies born to mothers who lacked both factors, developed GMH-IVH. In stepwise logistic regression analysis, these significant findings were not explained by the presence of labor, postnatal acidemia, need for intubation, antenatal administration of steroids, birth weight, or gestational age. In addition, we found that maternal receipt of magnesium sulfate was associated with diminished risk of GMH-IVH even in those babies born to mothers who apparently did not have preeclampsia. We postulate that the association of preeclampsia and GVH-IVH may be related to prostaglandin effects. Reduced maternal, placental, and umbilical prostaglandin I 2 (PGI ) values are characteristic of women with preec2 lampsia. Thus, babies born to mothers with preeclampsia may have a physiologic state similar to those having received indomethacin, a cyclooxygenase blocker that causes diminished PGI 2 levels and an effective GMH-IVH prophylactic We conclude that third-trimester maternal agent. early preeclampsia/toxemia is associated with reduced risk of GMH-IVH in the
JChild Neurol 1992;7:70-76). preterm newborn. (
erminal matrix hemorrhage (GMH)Gintraventricular hemorrhage (IVH), and cerebral
parenchymal injuries,
often associated with
large IVH, in premature newborns
are
important
risk factors for the subsequent development of 1 rologic disorders, such as cerebral palsy. Received
August 8,
1991.
Accepted
for
neu-
publication August
28, 1991. From the Department of Neurology, Children’s Hospital and Harvard Medical School, and the Department of Biostatistics, Harvard School of Public Health, Boston, MA. Address correspondence to Dr Karl Kuban, Children’s Hospital, Department of Neurology, 300 Longwood Avenue, Boston, MA 02115.
In
study designed to evaluate the efficacy of phenobarbital in preventing GMH-IVH, we noted that infants born to preeclamptic women were at rea
duced risk of GMH-IVH.2 This apparent protective effect of preeclampsia could not be explained by covariates. Because we recognized that this finding could be a random occurrence, we viewed the data as hypothesis generating. Our subsequent study of born infants enrolled prematurely prospectively was to evaluate the role of prenatal designed specifically and perinatal variables in predicting the development of GMH-IVH. This study provided us with an opportunity to test the null hypothesis that preeclampsia does not reduce the risk of GMH-IVH.
70
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
Methods
ities without concomitant GMH-IVH arate
were
viewed
as a
sep-
entity.
Sample The 449 babies in this study were born at Brigham and Women’s Hospital or at Beth Israel Hospital in Boston between June 1984 and September 1987, weighed 1.5 kg or less, and had a cranial ultrasound scan performed within the first 15 days after birth. Soon after birth, their mothers were interviewed about demographic characteristics and exposures during pregnancies. Fully 95% of the eligible mothers were recruited.
Data Collected
Analysis Because many of the data items were dichotomous, data analysis consisted mainly of X2 tests. The main hypothesis to be tested with this new data set was that babies born to preeclamptic women were at lower risk of GMH than were babies born to women who were not preeclamptic. A stepwise logistic regression analysis was carried out to determine which of the indicators of preeclampsia described above conveyed the most information about risk of GMHIVH.
Definitions.
In this study, we accept that the terms toxemia and preeclampsia are inexact.3-6 We used the term toxemia or preeclampsia when obstetricians used these terms in the mothers’ labor and delivery hospitalization charts. We also collected, from the mothers’ obstetric charts, specific blood-pressure and proteinuria data-the major components of toxemia or preeclampsia. Our analyses were carried out using both component definitions of toxemia/preeclampsia and obstetrician definitions of these entities. Gestational age determination was taken from the obstetric chart and was based on a best estimate from date of the last regular menstrual period and from fetal ultra-
sonography. Antecedents.
In addition to physicians’ diagnoses of toxemia and preeclampsia, we obtained from the mothers’ obstetrical records information about prepregnancy hypertension, pregnancy-induced hypertension, labor, and use of antenatal steroids or nonsteroidal antiinflammatory medications. Pharmacy sheets were reviewed for documentation of receipt of magnesium sulfate.
Definition of GMH-IVH.
Ultrasound scans were inwho had no knowledge of the mothers’ obstetrical histories. A diagnosis of GMH-IVH required an echogenic signal in the area of the supracaudate ganglionic eminence and/or in the ventricles on either side, in more than one plane. Parenchymal echoabnormal-
terpreted by ultrasonographers
TABLE 1 Risk of GMH-IVH per 100 Babies Within Each Maternal
HTN
=
hypertension;
Prot
=
Results Fully 71% of the 72 babies with a diagnosis of GMHIVH had their ultrasonographic echodensity confined to the germinal matrix. An additional 25% had blood in the ventricles as well as in the germinal matrix.
(2.5%) of the 40 mothers preeclampsia, and 71 (17.4%) of
with a diagnosis of 409 mothers without preeclampsia gave birth to babies who developed GMH-IVH (Table 1). GMH-IVH was seen in six (5.6%) of 107 babies born to women with hypertension compared to 66 (18.8%) of 352 babies born to mothers who did not have hypertension. Among 69 babies born to women with pregnancy-induced hypertension, GMH-IVH occurred in four (5.8%). Only eight (7.3%) of 108 babies born to women who had proteinuria had GMH-IVH, compared to 64 (18.3%) of 350 babies whose mothers did not have proteinuria. GMH-IVH was seen in five (5.6%) of 89 babies whose mothers had both hypertension and proteinuria, whereas 63 (19.0%) of 332 babies born to mothers who lacked both risk factors developed GMH-IVH. One
Hypertension/Proteinuria
Stratum
proteinuria.
71
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
We evaluated the possibility that preeclampsia did not reduce the risk of GMH-IVH, but rather that some covariate accounted for the reduced risk of GMH. Babies born to women who had hypertension and proteinuria were at one half to one third the risk of GMH-IVH seen among babies born to women with neither hypertension nor proteinuria, whether or not exposed to labor, prominent acidemia during the first postnatal day, or antenatal steroids (Table 2). Heavier birth weight, reduced probability of intubation during the first postnatal day, and relatively elevated prepregnancy maternal weight-to-height ratio also did not appear to account for the reduced risk of GMH-IVH in babies born to mothers with any evidence of preeclampsia. The reduced risk was less apparent in babies delivered vaginally than in those delivered abdomi-
nally. We considered the
possibility
that in this low-
birth-weight sample, babies born to toxemic mothers would tend to be more mature than babies born to other women. If that were the case, the birthweight-specific rates of GMH-IVH would be reduced in the babies born to toxemic mothers because of their greater maturity. Indeed, in this sample, babies born to women who had a diagnosis of preeclampsia were both older and lighter than babies of women not considered preeclamptic (mean gestational ages, respectively, of 30.4 weeks and 28.9 weeks, and mean birth weights of 1049.4 g and 1121.5 g). Thus, we analyzed rates of GMH-IVH by gestational age and presence of preeclampsia regardless of birth weight (Table 3). We found that 33 of the 40 babies born to mothers with preeclampsia had gestational age less than 33 weeks, only one of whom had GMH-IVH. Among babies whose gestational age was less than 30 weeks, none of the 14 babies born to preeclamptic mothers and 48 (19.8%) of
TABLE 2 Risk of GMH-IVH for Babies of Mothers With Both Hypertension and Proteinuria and of GMH for Babies of Mothers Diagnosed With Preeclampsia
*Relative to the risk among babies born to mothers with neither condition. Odds ratio (CI). tRelative to the risk among babies born to mothers with preeclampsia. Odds ratio (CI). t48 babies did not have a pH measurement. §Ten mothers did not provide information about both height and prepregnancy weight. not calculated because there was a value of 0 in one CI 95% confidence interval; NC cell. =
=
72
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
TABLE 3 Percent of Babies who Developed GMH-IVH by Gestational Age and Presence of Maternal Preeclampsia
the 243 babies GMH-IVH.
to toxemia
unexposed
developed
Although maternal receipt of magnesium
sulfate associated with diminished risk of GMH-IVH in the entire cohort (Table 1), it did not further reduce the risk of GMH in babies born to women with hypertension and proteinuria or to preeclamptic women (Table 2). Among the 409 women who were not preeclamptic, receipt of magnesium sulfate was associated with reduced risk of GMH-IVH in virtually all subgroups except those whose gestational age exceeded 30 weeks (Table 4). The reduced risk
was
TABLE 4 Risk of GMH-IVH
Among Babies Received Magnesium Sulfate*
Whose Mothers
achieved statistical significance not only among babies born at greater risk of GMH-IVH (eg, babies with birth weight less than 1.0 kg, gestational age less than or equal to 30 weeks, delivery after a trial of labor, and/or intubation in the delivery room), but also among babies expected to be at a reduced risk of GMH-IVH (eg, those with umbilical artery pH greater than or equal to 7.2 and/or maternal receipt of steroids). In an attempt to identify which of the preeclampsia variables contributed unique and significant information about GMH risk and to control for possible confounding, we carried out a stepwise logistic regression analysis. An initial model included as potential predictors of GMH-IVH all of the preeclampsia-related variables described above. Only two variables, however, remained in the model when a .05 significance level was needed for entry. Diagnosis of preeclampsia conveyed the most information about reduced risk of GMH-IVH, followed closely by receipt of magnesium sulfate. The other preeclampsia variables (eg, pregnancyinduced hypertension, hypertension not identified as pregnancy induced, and proteinuria) did not provide additional unique information. Very similar results were obtained when covariates (eg, gestational age, birth weight, intubation, ratio of prepregnancy maternal weight to height, etc) were allowed to com-
pete. Discussion The major finding of this study is that preterm babies had prominently reduced risk of GMH-IVH if their mothers were hypertensive and had proteinuria
prior to delivery or were diagnosed as having preeclampsia. The study by van de Bor et af identified
reduced risk of GMH-IVH among babies born mothers, and in our previous study, we had generated the hypothesis that early thirdtrimester preeclampsia, and not its risk factors or therapy, reduced the risk of GMH-IVH in preterm babies. 2,7 Thus, the study reported here tested a previously generated hypothesis. The distinction between hypothesis generation and hypothesis testing is important because the first and only recognition of a
to toxemic
previously unreported relationship is more likely a random phenomenon than is a succession of independent confirmations of the relation8
a
to reflect *Relative to the risk among babies whose mothers did not receive magnesium sulfate. In both groups, mothers had no evidence of
preeclampsia. CI
=
was
95% confidence interval; NC a value of 0 in one cell.
=
not calculated because
there
ship.8
Because observational studies are not experican show only statistical associations. Hill appears to be the first to formalize the muster-
mental, they
73
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
ing of evidence
to evaluate the causal implications of the results of observational studies.9 His criterion of reproducibility is satisfied when the results of one study are replicated repeatedly in different samples. The sample reported here was confined to babies whose birth weight was less than 1.5 kg, whereas the previous observation of an inverse relationship between maternal toxemia and/or preeclampsia and the infant’s risk of GMH-IVH was in a sample of babies who weighed less than 1.75 kg and were intubated during the first half day of life. Our testing a
and
the
previously generated hypothesis confirming finding of a previous report begins to satisfy the reproducibility criterion for a causal relationship becharacteristic and a disease. exposure Biologic plausibility, another of Hill’s criteria of causality, usually refers to integrating the findings of obervational studies in humans with what is known from all other sources (including laboratory studies). For example, many of the effects of maternal preeclampsia on the fetus can be attributed to prostacyclin deficiency.l° In pregnancies complicated by toxtween
an
or a
emia, Prostaglandin 12 (PG12) (a vasodilator) is reduced in maternal, placental, and umbilical-cord vessels, as well as in the placenta.l° Reduction of
PG12 alters its ratio to Thromboxane A2 (TXA2), and thereby promotes vasoconstriction and platelet ag-
gregation.10 Furthermore, in a randomized, doubleblind, controlled study, women at high risk for toxemia did not
develop the disorder when preaspirin, an antiplatelet agent and cyclooxygenase blocker.ll Low-dose aspirin has been shown to preferentially diminish reduction of TXA2 relative to PG12, thus promoting vasodilation.l2 Treatment with indomethacin, another cyclooxgenase blocker, creates a physiologic state in the baby, treated with
in terms of PGI2:TXA2 ratio, that is similar to that of the toxemic pregnant mother. Several studies, both in animal models and in humans, have shown reduced incidence of GMH-IVH in premature new-
pretreated with indomethacin,l3-16 although other studies fail to show this beneficial effects. 17 An alternative hypothesis of the effect of preeclampsia on GMH-IVH relates to increased corticosteroid production by either the preeclamptic mother or her fetus, resulting in acceleration of fetal maturation. Exogenous corticosteroids administered antenatally to the pregnant woman appear to reduce her preterm baby’s risk of respiratory distress bronchopulmonary dysplasia,19 retinand GMH-IVH.21 The supof opathy maternal that position preeclampsia produces effects similar to those caused by increased production or borns
syndrome,18 prematurity,2°
by receipt of corticosteroids is supported by findings preterm babies born to preeclamptic mothers
that
appear to be at reduced risk of the so-called diseases of immaturity.2,22-25 Babies born to preeclamptic women also exhibit features indicative of accelerated maturation of brain and other organs. 26,27 The biologic plausibility criterion of Hill, therefore, seems to be satisfied by both the prostacyclin deficiency hypothesis and the stress/steroids hypothesis.
The
temporality criterion of a causal relationship easily satisfied because maternal preeclampsia obviously occurs well before the onset of the majority is
of GMH-IVH.28-32 Full-term babies born to women with late third-trimester preeclampsia are more likely than their peers to be growth retarded, thereby suggesting that maternal preeclampsia affects the fetus long before the disorder is clinically apparent in the mother. 10,32-35 Indeed, recent findings suggest that correlates of preeclampsia are evident in the first trimester.36 Our finding of reduced risk of GMH-IVH supports the view that the fetus is affected by maternal preeclampsia by the beginning of the third trimester. Strength of the association is another of Hill’s criteria.9 The assumption underlying this criterion is that the stronger the association, the lower the probability that bias grossly distorts the apparent relationship between an exposure or antecendent and a disorder. The risk of GMH-IVH in babies born to women who were not preeclamptic is almost three times the risk in babies born to preeclamptic women, clearly a strong association. In light of the number of Hill’s criteria satisfied, it is reasonable to conclude that preeclampsia may contribute to reducing the risk of GMH. Receipt of magnesium sulfate by women who were not preeclamptic was associated with reduced risk of GMH-IVH in their preterm newborns. Whether this reflects selection bias, confounding by indication ’37 or biologic or random phenomena remains uncertain. Some of our findings are consistent with the hypothesis that administration of magnesium sulfate to women who presumably are not preeclamptic early in the third trimester reduces the risk of GMH. We did not, however, find support for this hypothesis in our previous evaluation of pre2 eclampsia and toxemia in relation to risk of GMH.2 It is possible that obstetricians considered some of the women in the earlier study preeclamptic, even though the women did not have documented hypertension and/or proteinuria in the chart of any pregnancy hospitalization (these findings may have been
74
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
documented in outpatient medical records, not reviewed for the study). Until more support is
gathered,
our more
findings regarding viewed cautiously.
recent
mag-
nesium sulfate should be In conclusion, this is the second report that early third-trimester maternal preeclampsia and/or toxemia is associated with reduced risk of GMH-IVH in the preterm newborn. It is the first report, however, to suggest that magnesium sulfate administration independently reduces risk of GMH-IVH among babies born to women who apparently are not
14. Bada
HS, Green RS, Pourcyrous M, et al: Indomethacin reduces the risks of severe intraventricular hemorrhage. J Pediatr
1988;112:941-947. 18.
preeclamptic.
1989;115:631-637.
15. Bandstra ES, Montalvo BM, Goldberg RN, et al: Prophylactic indomethacin for prevention of intraventricular hemorrhage in premature infants. Pediatrics 1988;82:533-542. 16. Ment LR, Stewart WB, Scott DT, Duncan CC: Beagle puppy model of intraventricular hemorrhage: Randomized indomethacin prevention trial. Neurology 1983;33:179-184. 17. Hanigan WC, Kennedy G, Roemisch F, et al: Administration of indomethacin for the prevention of periventricularintraventricular hemorrhage in high-risk neonates. J Pediatr
Avery ME, Aylward G, Creasy R, et al: Update on prenatal steroid for prevention of respiratory distress: Report of a conference—September 26-28, 1985. AmJ Obstet Gynecol 1986;155:2-5.
Acknowledgments We are indebted to Dr Joseph Volpe for his critical review thoughtful suggestions, to Ms Jeanette Holmes for her diligent help in manuscript typing, to the Joint Program of Neonatol-
and
ogy for support, and to the families involved in this their willingness to be participants.
study
for
19. Van Marter LJ, Leviton A, Kuban K, et al: Maternal glucocorticoid therapy and reduced risk of bronchopulmonary dysplasia. Pediatrics 1990;86:331-336. 20. Doyle LW, Kitchen WH, Ford GW, et al: Effects of antenatal steroid therapy on mortality and morbidity in very low birth
weight infants.J Pediatr 1986;108:287-292. 21. Leviton A, Pagano M, Kuban K: Etiologic heterogeneity of intracranial hemorrhage in preterm newborns. Pediatr Neurol
References
Intraventricular hemorrhage in the premature infant. Current concepts. Part I. Ann Neurol 1989;25:1-11. 2. Leviton A, Kuban K, Pagano M: Maternal toxemia and neonatal germinal matrix hemorrhage in intubated infants less than 1750 grams. Obstet Gynecol 1988;72:571-576. 3. Sibai BM: Pitfalls in diagnosis and management of preeclamp1.
4.
Volpe JJ:
sia. Am J Obstet Gynecol 1988;159:1-5. Morgan MA, Thurnau GR: Pregnancy-induced hypertension without proteinuria: Is it true preeclampsia? South Med J
1988;81:210-213. 5. Redman CW, Jefferies M: Revised definition of pre-eclampsia. Lancet 1988;1:809-812. 6. Davey DA, MacGillivray I: The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 7.
8. 9. 10. 11.
12.
13.
1988;158:892-898. de Bor M, Verloove-Vanhorick SP, Brand R,
et al: Incidence and prediction of periventricular-intraventricular hemorrhage in very preterm infants.J Perinat Med 1987; 15:333-339. Feinstein AR: Scientific standards in epidemiologic studies of the menace of daily life. Science 1988;242:1257-1263. Hill AB: The environment and disease: Association or causation. Proc R Soc Med 1965;58:295. Romero R, Lockwood C, Oyarzun E, et al: Toxemia: New concepts in an old disease. Semin Perinatol 1988;12:302-323. Benigni A, Gregorini G, Frusca T, et al: Effect of low-dose aspirin on fetal and maternal generation of thromboxane by platelets in women at risk for pregnancy-induced hypertension. N Engl J Med 1989;321:357-362. Schiff E, Peleg E, Goldenberg M, et al: The use of aspirin to prevent pregnancy-induced hypertension and lower the ratio of thromboxane A2 to prostacyclin in relatively high risk pregnancies. N Engl J Med 1989;321:351-356. Ment LR, Duncan CC, Ehrenkranz RA, et al: Randomized low-dose Indomethacin trial for prevention of intraventricular hemorrhage in very low birth weight neonates.J Pediatr
van
1988;112:948-955.
1988;4:274-278.
’
22. Chiswick ML: Prolonged rupture of membranes, pre-eclamptic toxemia and respiratory distress syndrome. Arch Dis Child
1976;51:674-679. 23. Lee KS, Eidelman AT, Tseng PI, et al: Respiratory distress syndrome of the newborn and complications of pregnancy.
Pediatrics 1976;58:675-680. 24. Yoon JJ, Kohl S, Harper RG: The relationship between maternal hypertensive disease of pregnancy and the incidence of idiopathic respiratory distress syndrome. Pediatrics 1980; 65:735-739. 25. Purohit DM, Ellison RC, Zierler S, et al: Risk factors for retrolental fibroplasia: Experience with 3,025 premature infants. National collaborative study on patent ductus arteriosus in premature Infants. Pediatrics 1985;76:339-344. 26. Gluck L, Kulovich MV: The evaluation of functional maturity, in Gluck L (ed). Modern Perinatal Medicine. Chicago, Year Book, 1974, pp 195-207. 27. Hadi HA: Fetal cerebral maturation in hypertensive disorders of pregnancy. Obstet Gynecol 1984;63:214-219. 28. Chesley LC, Sibai BM: Clinical significance of elevated mean arterial pressure in the second trimester. AmJ Obstet Gynecol
1988;159:275-279. 29.
deCrespigny LC, MacKay R, Murton LJ, et al: Timing of neonatal cerebroventricular hemorrhage with ultrasound. Arch
Dis Child 1982;57:231-233. 30. Dolfin T, Skidmore MB, Fong KW, et al: Incidence, severity, and timing of subependymal and intraventricular hemorrhages in preterm infants born in a perinatal unit as detected by serial real-time ultrasound. Pediatrics 1983;71:541-546. 31. Beverly DW, Chance GW, Coates CF: Intraventricular haemorrhage—timing of occurrence and relationship to perinatal events.
BrJ Obstet Gynaecol 1984;91:1007-1013.
32. McDonald MM, Koops BL, Johnson ML et al: Timing and antecedents of intracranial hemorrhage in the newborn. Pediatrics 1984;74:32-36. 33. Page EW, Christianson R: Influence of blood pressure changes with and without proteinuria upon outcome of pregnancy. Am J Obstet Gynecol 1976;126:821-833.
75
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
34. Mutch LMM, Moar VA, Ounsted MK, Redman CW: Hypertension during pregnancy, with and without specific hypotensive treatment. I. Perinatal factors and neonatal morbidity. Early Hum Dev 1977;1:47-57. 35. Ounsted MK, Moar VA, Good FJ: Hypertension during pregnancy with and without specific treatment: The development of the children at the age of four years. Br J Obstet Gynaecol
36. Sibai BM, Anderson GD, Abdella TN, et al: Eclampsia. III. Neonatal outcome, growth, and development. Am J Obstet
1983;146:307-316. Gynecol 37. Miettinen OS: Efficacy of therapeutic practice: Will epidemiology provide the answers? in Melmon KL (ed): Drug Therapeutics—Concepts for Physicians. New York, Elseiver-North Holland, 1980, pp 201-208.
1980;87:19-24.
76
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
Kuban, MD; Alan Leviton, MD; Marcello Pagano, PhD; Terence Fenton, EdD; Ruth Strassfeld; Mildred Wolff, MA
Abstract To evaluate
prenatal and perinatal risk factors for development of germinal matrix hemorrhage-intraventricular hemorrhage (GMH-IVH), we conducted a prospective epidemiologic study of 449 babies whose birth weight was less than 1501 grams. This study permitted us to test our previously generated hypothesis that babies born to mothers with preeclampsia were at substantially reduced risk of developing GMH-IVH. Seventy-two (16%) of the babies in this population developed GMH-IVH. One (2.5%) of the 40 mothers with a diagnosis of preeclampsia and 71 (17.4%) of 409 mothers without preeclampsia gave birth to babies who developed GMH-IVH. GMH-IVH was seen in 6/107 (5.6%) of babies born to women with hypertension including 4/69 (5.8%) of babies born to women with pregnancy-induced hypertension, compared to 66/352 (18.8%) of babies born to mothers who did not have hypertension. Only 7.3% (8/108) of babies born to women who had proteinuria had GMH-IVH, compared to 18.3% (64/350) of babies whose mothers did not have proteinuria. GMH-IVH was seen in 5/89 (5.6%) of babies whose mothers had both hypertension and proteinuria, whereas 63/332 (19%) of babies born to mothers who lacked both factors, developed GMH-IVH. In stepwise logistic regression analysis, these significant findings were not explained by the presence of labor, postnatal acidemia, need for intubation, antenatal administration of steroids, birth weight, or gestational age. In addition, we found that maternal receipt of magnesium sulfate was associated with diminished risk of GMH-IVH even in those babies born to mothers who apparently did not have preeclampsia. We postulate that the association of preeclampsia and GVH-IVH may be related to prostaglandin effects. Reduced maternal, placental, and umbilical prostaglandin I 2 (PGI ) values are characteristic of women with preec2 lampsia. Thus, babies born to mothers with preeclampsia may have a physiologic state similar to those having received indomethacin, a cyclooxygenase blocker that causes diminished PGI 2 levels and an effective GMH-IVH prophylactic We conclude that third-trimester maternal agent. early preeclampsia/toxemia is associated with reduced risk of GMH-IVH in the
JChild Neurol 1992;7:70-76). preterm newborn. (
erminal matrix hemorrhage (GMH)Gintraventricular hemorrhage (IVH), and cerebral
parenchymal injuries,
often associated with
large IVH, in premature newborns
are
important
risk factors for the subsequent development of 1 rologic disorders, such as cerebral palsy. Received
August 8,
1991.
Accepted
for
neu-
publication August
28, 1991. From the Department of Neurology, Children’s Hospital and Harvard Medical School, and the Department of Biostatistics, Harvard School of Public Health, Boston, MA. Address correspondence to Dr Karl Kuban, Children’s Hospital, Department of Neurology, 300 Longwood Avenue, Boston, MA 02115.
In
study designed to evaluate the efficacy of phenobarbital in preventing GMH-IVH, we noted that infants born to preeclamptic women were at rea
duced risk of GMH-IVH.2 This apparent protective effect of preeclampsia could not be explained by covariates. Because we recognized that this finding could be a random occurrence, we viewed the data as hypothesis generating. Our subsequent study of born infants enrolled prematurely prospectively was to evaluate the role of prenatal designed specifically and perinatal variables in predicting the development of GMH-IVH. This study provided us with an opportunity to test the null hypothesis that preeclampsia does not reduce the risk of GMH-IVH.
70
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
Methods
ities without concomitant GMH-IVH arate
were
viewed
as a
sep-
entity.
Sample The 449 babies in this study were born at Brigham and Women’s Hospital or at Beth Israel Hospital in Boston between June 1984 and September 1987, weighed 1.5 kg or less, and had a cranial ultrasound scan performed within the first 15 days after birth. Soon after birth, their mothers were interviewed about demographic characteristics and exposures during pregnancies. Fully 95% of the eligible mothers were recruited.
Data Collected
Analysis Because many of the data items were dichotomous, data analysis consisted mainly of X2 tests. The main hypothesis to be tested with this new data set was that babies born to preeclamptic women were at lower risk of GMH than were babies born to women who were not preeclamptic. A stepwise logistic regression analysis was carried out to determine which of the indicators of preeclampsia described above conveyed the most information about risk of GMHIVH.
Definitions.
In this study, we accept that the terms toxemia and preeclampsia are inexact.3-6 We used the term toxemia or preeclampsia when obstetricians used these terms in the mothers’ labor and delivery hospitalization charts. We also collected, from the mothers’ obstetric charts, specific blood-pressure and proteinuria data-the major components of toxemia or preeclampsia. Our analyses were carried out using both component definitions of toxemia/preeclampsia and obstetrician definitions of these entities. Gestational age determination was taken from the obstetric chart and was based on a best estimate from date of the last regular menstrual period and from fetal ultra-
sonography. Antecedents.
In addition to physicians’ diagnoses of toxemia and preeclampsia, we obtained from the mothers’ obstetrical records information about prepregnancy hypertension, pregnancy-induced hypertension, labor, and use of antenatal steroids or nonsteroidal antiinflammatory medications. Pharmacy sheets were reviewed for documentation of receipt of magnesium sulfate.
Definition of GMH-IVH.
Ultrasound scans were inwho had no knowledge of the mothers’ obstetrical histories. A diagnosis of GMH-IVH required an echogenic signal in the area of the supracaudate ganglionic eminence and/or in the ventricles on either side, in more than one plane. Parenchymal echoabnormal-
terpreted by ultrasonographers
TABLE 1 Risk of GMH-IVH per 100 Babies Within Each Maternal
HTN
=
hypertension;
Prot
=
Results Fully 71% of the 72 babies with a diagnosis of GMHIVH had their ultrasonographic echodensity confined to the germinal matrix. An additional 25% had blood in the ventricles as well as in the germinal matrix.
(2.5%) of the 40 mothers preeclampsia, and 71 (17.4%) of
with a diagnosis of 409 mothers without preeclampsia gave birth to babies who developed GMH-IVH (Table 1). GMH-IVH was seen in six (5.6%) of 107 babies born to women with hypertension compared to 66 (18.8%) of 352 babies born to mothers who did not have hypertension. Among 69 babies born to women with pregnancy-induced hypertension, GMH-IVH occurred in four (5.8%). Only eight (7.3%) of 108 babies born to women who had proteinuria had GMH-IVH, compared to 64 (18.3%) of 350 babies whose mothers did not have proteinuria. GMH-IVH was seen in five (5.6%) of 89 babies whose mothers had both hypertension and proteinuria, whereas 63 (19.0%) of 332 babies born to mothers who lacked both risk factors developed GMH-IVH. One
Hypertension/Proteinuria
Stratum
proteinuria.
71
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
We evaluated the possibility that preeclampsia did not reduce the risk of GMH-IVH, but rather that some covariate accounted for the reduced risk of GMH. Babies born to women who had hypertension and proteinuria were at one half to one third the risk of GMH-IVH seen among babies born to women with neither hypertension nor proteinuria, whether or not exposed to labor, prominent acidemia during the first postnatal day, or antenatal steroids (Table 2). Heavier birth weight, reduced probability of intubation during the first postnatal day, and relatively elevated prepregnancy maternal weight-to-height ratio also did not appear to account for the reduced risk of GMH-IVH in babies born to mothers with any evidence of preeclampsia. The reduced risk was less apparent in babies delivered vaginally than in those delivered abdomi-
nally. We considered the
possibility
that in this low-
birth-weight sample, babies born to toxemic mothers would tend to be more mature than babies born to other women. If that were the case, the birthweight-specific rates of GMH-IVH would be reduced in the babies born to toxemic mothers because of their greater maturity. Indeed, in this sample, babies born to women who had a diagnosis of preeclampsia were both older and lighter than babies of women not considered preeclamptic (mean gestational ages, respectively, of 30.4 weeks and 28.9 weeks, and mean birth weights of 1049.4 g and 1121.5 g). Thus, we analyzed rates of GMH-IVH by gestational age and presence of preeclampsia regardless of birth weight (Table 3). We found that 33 of the 40 babies born to mothers with preeclampsia had gestational age less than 33 weeks, only one of whom had GMH-IVH. Among babies whose gestational age was less than 30 weeks, none of the 14 babies born to preeclamptic mothers and 48 (19.8%) of
TABLE 2 Risk of GMH-IVH for Babies of Mothers With Both Hypertension and Proteinuria and of GMH for Babies of Mothers Diagnosed With Preeclampsia
*Relative to the risk among babies born to mothers with neither condition. Odds ratio (CI). tRelative to the risk among babies born to mothers with preeclampsia. Odds ratio (CI). t48 babies did not have a pH measurement. §Ten mothers did not provide information about both height and prepregnancy weight. not calculated because there was a value of 0 in one CI 95% confidence interval; NC cell. =
=
72
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
TABLE 3 Percent of Babies who Developed GMH-IVH by Gestational Age and Presence of Maternal Preeclampsia
the 243 babies GMH-IVH.
to toxemia
unexposed
developed
Although maternal receipt of magnesium
sulfate associated with diminished risk of GMH-IVH in the entire cohort (Table 1), it did not further reduce the risk of GMH in babies born to women with hypertension and proteinuria or to preeclamptic women (Table 2). Among the 409 women who were not preeclamptic, receipt of magnesium sulfate was associated with reduced risk of GMH-IVH in virtually all subgroups except those whose gestational age exceeded 30 weeks (Table 4). The reduced risk
was
TABLE 4 Risk of GMH-IVH
Among Babies Received Magnesium Sulfate*
Whose Mothers
achieved statistical significance not only among babies born at greater risk of GMH-IVH (eg, babies with birth weight less than 1.0 kg, gestational age less than or equal to 30 weeks, delivery after a trial of labor, and/or intubation in the delivery room), but also among babies expected to be at a reduced risk of GMH-IVH (eg, those with umbilical artery pH greater than or equal to 7.2 and/or maternal receipt of steroids). In an attempt to identify which of the preeclampsia variables contributed unique and significant information about GMH risk and to control for possible confounding, we carried out a stepwise logistic regression analysis. An initial model included as potential predictors of GMH-IVH all of the preeclampsia-related variables described above. Only two variables, however, remained in the model when a .05 significance level was needed for entry. Diagnosis of preeclampsia conveyed the most information about reduced risk of GMH-IVH, followed closely by receipt of magnesium sulfate. The other preeclampsia variables (eg, pregnancyinduced hypertension, hypertension not identified as pregnancy induced, and proteinuria) did not provide additional unique information. Very similar results were obtained when covariates (eg, gestational age, birth weight, intubation, ratio of prepregnancy maternal weight to height, etc) were allowed to com-
pete. Discussion The major finding of this study is that preterm babies had prominently reduced risk of GMH-IVH if their mothers were hypertensive and had proteinuria
prior to delivery or were diagnosed as having preeclampsia. The study by van de Bor et af identified
reduced risk of GMH-IVH among babies born mothers, and in our previous study, we had generated the hypothesis that early thirdtrimester preeclampsia, and not its risk factors or therapy, reduced the risk of GMH-IVH in preterm babies. 2,7 Thus, the study reported here tested a previously generated hypothesis. The distinction between hypothesis generation and hypothesis testing is important because the first and only recognition of a
to toxemic
previously unreported relationship is more likely a random phenomenon than is a succession of independent confirmations of the relation8
a
to reflect *Relative to the risk among babies whose mothers did not receive magnesium sulfate. In both groups, mothers had no evidence of
preeclampsia. CI
=
was
95% confidence interval; NC a value of 0 in one cell.
=
not calculated because
there
ship.8
Because observational studies are not experican show only statistical associations. Hill appears to be the first to formalize the muster-
mental, they
73
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
ing of evidence
to evaluate the causal implications of the results of observational studies.9 His criterion of reproducibility is satisfied when the results of one study are replicated repeatedly in different samples. The sample reported here was confined to babies whose birth weight was less than 1.5 kg, whereas the previous observation of an inverse relationship between maternal toxemia and/or preeclampsia and the infant’s risk of GMH-IVH was in a sample of babies who weighed less than 1.75 kg and were intubated during the first half day of life. Our testing a
and
the
previously generated hypothesis confirming finding of a previous report begins to satisfy the reproducibility criterion for a causal relationship becharacteristic and a disease. exposure Biologic plausibility, another of Hill’s criteria of causality, usually refers to integrating the findings of obervational studies in humans with what is known from all other sources (including laboratory studies). For example, many of the effects of maternal preeclampsia on the fetus can be attributed to prostacyclin deficiency.l° In pregnancies complicated by toxtween
an
or a
emia, Prostaglandin 12 (PG12) (a vasodilator) is reduced in maternal, placental, and umbilical-cord vessels, as well as in the placenta.l° Reduction of
PG12 alters its ratio to Thromboxane A2 (TXA2), and thereby promotes vasoconstriction and platelet ag-
gregation.10 Furthermore, in a randomized, doubleblind, controlled study, women at high risk for toxemia did not
develop the disorder when preaspirin, an antiplatelet agent and cyclooxygenase blocker.ll Low-dose aspirin has been shown to preferentially diminish reduction of TXA2 relative to PG12, thus promoting vasodilation.l2 Treatment with indomethacin, another cyclooxgenase blocker, creates a physiologic state in the baby, treated with
in terms of PGI2:TXA2 ratio, that is similar to that of the toxemic pregnant mother. Several studies, both in animal models and in humans, have shown reduced incidence of GMH-IVH in premature new-
pretreated with indomethacin,l3-16 although other studies fail to show this beneficial effects. 17 An alternative hypothesis of the effect of preeclampsia on GMH-IVH relates to increased corticosteroid production by either the preeclamptic mother or her fetus, resulting in acceleration of fetal maturation. Exogenous corticosteroids administered antenatally to the pregnant woman appear to reduce her preterm baby’s risk of respiratory distress bronchopulmonary dysplasia,19 retinand GMH-IVH.21 The supof opathy maternal that position preeclampsia produces effects similar to those caused by increased production or borns
syndrome,18 prematurity,2°
by receipt of corticosteroids is supported by findings preterm babies born to preeclamptic mothers
that
appear to be at reduced risk of the so-called diseases of immaturity.2,22-25 Babies born to preeclamptic women also exhibit features indicative of accelerated maturation of brain and other organs. 26,27 The biologic plausibility criterion of Hill, therefore, seems to be satisfied by both the prostacyclin deficiency hypothesis and the stress/steroids hypothesis.
The
temporality criterion of a causal relationship easily satisfied because maternal preeclampsia obviously occurs well before the onset of the majority is
of GMH-IVH.28-32 Full-term babies born to women with late third-trimester preeclampsia are more likely than their peers to be growth retarded, thereby suggesting that maternal preeclampsia affects the fetus long before the disorder is clinically apparent in the mother. 10,32-35 Indeed, recent findings suggest that correlates of preeclampsia are evident in the first trimester.36 Our finding of reduced risk of GMH-IVH supports the view that the fetus is affected by maternal preeclampsia by the beginning of the third trimester. Strength of the association is another of Hill’s criteria.9 The assumption underlying this criterion is that the stronger the association, the lower the probability that bias grossly distorts the apparent relationship between an exposure or antecendent and a disorder. The risk of GMH-IVH in babies born to women who were not preeclamptic is almost three times the risk in babies born to preeclamptic women, clearly a strong association. In light of the number of Hill’s criteria satisfied, it is reasonable to conclude that preeclampsia may contribute to reducing the risk of GMH. Receipt of magnesium sulfate by women who were not preeclamptic was associated with reduced risk of GMH-IVH in their preterm newborns. Whether this reflects selection bias, confounding by indication ’37 or biologic or random phenomena remains uncertain. Some of our findings are consistent with the hypothesis that administration of magnesium sulfate to women who presumably are not preeclamptic early in the third trimester reduces the risk of GMH. We did not, however, find support for this hypothesis in our previous evaluation of pre2 eclampsia and toxemia in relation to risk of GMH.2 It is possible that obstetricians considered some of the women in the earlier study preeclamptic, even though the women did not have documented hypertension and/or proteinuria in the chart of any pregnancy hospitalization (these findings may have been
74
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
documented in outpatient medical records, not reviewed for the study). Until more support is
gathered,
our more
findings regarding viewed cautiously.
recent
mag-
nesium sulfate should be In conclusion, this is the second report that early third-trimester maternal preeclampsia and/or toxemia is associated with reduced risk of GMH-IVH in the preterm newborn. It is the first report, however, to suggest that magnesium sulfate administration independently reduces risk of GMH-IVH among babies born to women who apparently are not
14. Bada
HS, Green RS, Pourcyrous M, et al: Indomethacin reduces the risks of severe intraventricular hemorrhage. J Pediatr
1988;112:941-947. 18.
preeclamptic.
1989;115:631-637.
15. Bandstra ES, Montalvo BM, Goldberg RN, et al: Prophylactic indomethacin for prevention of intraventricular hemorrhage in premature infants. Pediatrics 1988;82:533-542. 16. Ment LR, Stewart WB, Scott DT, Duncan CC: Beagle puppy model of intraventricular hemorrhage: Randomized indomethacin prevention trial. Neurology 1983;33:179-184. 17. Hanigan WC, Kennedy G, Roemisch F, et al: Administration of indomethacin for the prevention of periventricularintraventricular hemorrhage in high-risk neonates. J Pediatr
Avery ME, Aylward G, Creasy R, et al: Update on prenatal steroid for prevention of respiratory distress: Report of a conference—September 26-28, 1985. AmJ Obstet Gynecol 1986;155:2-5.
Acknowledgments We are indebted to Dr Joseph Volpe for his critical review thoughtful suggestions, to Ms Jeanette Holmes for her diligent help in manuscript typing, to the Joint Program of Neonatol-
and
ogy for support, and to the families involved in this their willingness to be participants.
study
for
19. Van Marter LJ, Leviton A, Kuban K, et al: Maternal glucocorticoid therapy and reduced risk of bronchopulmonary dysplasia. Pediatrics 1990;86:331-336. 20. Doyle LW, Kitchen WH, Ford GW, et al: Effects of antenatal steroid therapy on mortality and morbidity in very low birth
weight infants.J Pediatr 1986;108:287-292. 21. Leviton A, Pagano M, Kuban K: Etiologic heterogeneity of intracranial hemorrhage in preterm newborns. Pediatr Neurol
References
Intraventricular hemorrhage in the premature infant. Current concepts. Part I. Ann Neurol 1989;25:1-11. 2. Leviton A, Kuban K, Pagano M: Maternal toxemia and neonatal germinal matrix hemorrhage in intubated infants less than 1750 grams. Obstet Gynecol 1988;72:571-576. 3. Sibai BM: Pitfalls in diagnosis and management of preeclamp1.
4.
Volpe JJ:
sia. Am J Obstet Gynecol 1988;159:1-5. Morgan MA, Thurnau GR: Pregnancy-induced hypertension without proteinuria: Is it true preeclampsia? South Med J
1988;81:210-213. 5. Redman CW, Jefferies M: Revised definition of pre-eclampsia. Lancet 1988;1:809-812. 6. Davey DA, MacGillivray I: The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 7.
8. 9. 10. 11.
12.
13.
1988;158:892-898. de Bor M, Verloove-Vanhorick SP, Brand R,
et al: Incidence and prediction of periventricular-intraventricular hemorrhage in very preterm infants.J Perinat Med 1987; 15:333-339. Feinstein AR: Scientific standards in epidemiologic studies of the menace of daily life. Science 1988;242:1257-1263. Hill AB: The environment and disease: Association or causation. Proc R Soc Med 1965;58:295. Romero R, Lockwood C, Oyarzun E, et al: Toxemia: New concepts in an old disease. Semin Perinatol 1988;12:302-323. Benigni A, Gregorini G, Frusca T, et al: Effect of low-dose aspirin on fetal and maternal generation of thromboxane by platelets in women at risk for pregnancy-induced hypertension. N Engl J Med 1989;321:357-362. Schiff E, Peleg E, Goldenberg M, et al: The use of aspirin to prevent pregnancy-induced hypertension and lower the ratio of thromboxane A2 to prostacyclin in relatively high risk pregnancies. N Engl J Med 1989;321:351-356. Ment LR, Duncan CC, Ehrenkranz RA, et al: Randomized low-dose Indomethacin trial for prevention of intraventricular hemorrhage in very low birth weight neonates.J Pediatr
van
1988;112:948-955.
1988;4:274-278.
’
22. Chiswick ML: Prolonged rupture of membranes, pre-eclamptic toxemia and respiratory distress syndrome. Arch Dis Child
1976;51:674-679. 23. Lee KS, Eidelman AT, Tseng PI, et al: Respiratory distress syndrome of the newborn and complications of pregnancy.
Pediatrics 1976;58:675-680. 24. Yoon JJ, Kohl S, Harper RG: The relationship between maternal hypertensive disease of pregnancy and the incidence of idiopathic respiratory distress syndrome. Pediatrics 1980; 65:735-739. 25. Purohit DM, Ellison RC, Zierler S, et al: Risk factors for retrolental fibroplasia: Experience with 3,025 premature infants. National collaborative study on patent ductus arteriosus in premature Infants. Pediatrics 1985;76:339-344. 26. Gluck L, Kulovich MV: The evaluation of functional maturity, in Gluck L (ed). Modern Perinatal Medicine. Chicago, Year Book, 1974, pp 195-207. 27. Hadi HA: Fetal cerebral maturation in hypertensive disorders of pregnancy. Obstet Gynecol 1984;63:214-219. 28. Chesley LC, Sibai BM: Clinical significance of elevated mean arterial pressure in the second trimester. AmJ Obstet Gynecol
1988;159:275-279. 29.
deCrespigny LC, MacKay R, Murton LJ, et al: Timing of neonatal cerebroventricular hemorrhage with ultrasound. Arch
Dis Child 1982;57:231-233. 30. Dolfin T, Skidmore MB, Fong KW, et al: Incidence, severity, and timing of subependymal and intraventricular hemorrhages in preterm infants born in a perinatal unit as detected by serial real-time ultrasound. Pediatrics 1983;71:541-546. 31. Beverly DW, Chance GW, Coates CF: Intraventricular haemorrhage—timing of occurrence and relationship to perinatal events.
BrJ Obstet Gynaecol 1984;91:1007-1013.
32. McDonald MM, Koops BL, Johnson ML et al: Timing and antecedents of intracranial hemorrhage in the newborn. Pediatrics 1984;74:32-36. 33. Page EW, Christianson R: Influence of blood pressure changes with and without proteinuria upon outcome of pregnancy. Am J Obstet Gynecol 1976;126:821-833.
75
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
34. Mutch LMM, Moar VA, Ounsted MK, Redman CW: Hypertension during pregnancy, with and without specific hypotensive treatment. I. Perinatal factors and neonatal morbidity. Early Hum Dev 1977;1:47-57. 35. Ounsted MK, Moar VA, Good FJ: Hypertension during pregnancy with and without specific treatment: The development of the children at the age of four years. Br J Obstet Gynaecol
36. Sibai BM, Anderson GD, Abdella TN, et al: Eclampsia. III. Neonatal outcome, growth, and development. Am J Obstet
1983;146:307-316. Gynecol 37. Miettinen OS: Efficacy of therapeutic practice: Will epidemiology provide the answers? in Melmon KL (ed): Drug Therapeutics—Concepts for Physicians. New York, Elseiver-North Holland, 1980, pp 201-208.
1980;87:19-24.
76
Downloaded from jcn.sagepub.com at University of Manchester Library on May 18, 2015
