Labor and Deliver Characteristics and the Risk of Germinal Matrix Hemorrhage in Low Birth Weight Infants Alan Leviton, MD; Terence Fenton, EdD; Karl C. K. Kuban, MD; Marcello Pagano, PhD
Abstract To assess the influence of labor and delivery events on the risk of germinal matrix hemorrhage in preterm newborns, we conducted a review of data collected on 449 babies who weighed 1.5 kg or less. Babies delivered vaginally were more likely to have germinal matrix hemorrhage than were babies delivered abdominally (odds ratio, 2.5; 95% confidence interval, 1.4,3.3). Among babies delivered vaginally, the risk of germinal matrix hemorrhage was increased by 39% if labor lasted more than 12 hours. Among babies delivered abdominally, the occurrence of any labor was accompanied by a 150% increased risk of germinal matrix hemorrhage. The only indication for abdominal delivery associated with an increased risk of germinal matrix hemorrhage was impending amnionitis (odds ratio, 2.6; 95% confidence interval, 1.2,5.7), whereas the only indication associated with a decreased risk was preeclampsia (odds ratio, 0.2; 95% confidence interval, 0.6). Epidural and local anesthesia were associated with a reduced risk of germinal matrix hemorrhage among babies delivered abdominally. We conclude that delivery practices, or their indications, appear to influence the risk of germinal matrix hemorrhage in low birth weight babies. (J Child Neurol 1991;6:35-40).
low Ifnally
birth
weight
babies delivered abdomi-
at lower risk of
matrix
germinal hemorrhage than are babies delivered vaginally, then perhaps obstetricians will feel under pressure to deliver preterm babies abdominally. Similarly, if the risk of germinal matrix hemorrhage increases are
with the duration of labor, then obstetricians will tend to expedite delivery of low birthweight infants, whether via the vaginal or abdominal route. Of course, the adversities associated with abdominal delivery need to counterbalance calls for more cesarean sections. Some studies have found that abdominal delivery is associated with a reduced risk of germinal maReceived Jan 3, 1990. Received revised March 19, 1990. Ac-
cepted for publication March 19, 1990. From the Neuroepidemiology Unit (Drs Leviton and Kuban), Children’s Hospital, the Department of Neurology (Drs Leviton and Kuban), Harvard Medical School, and the Department of Biostatistics (Drs Fenton and Pagano), Harvard School of Public Health, Boston, MA. Address
Hospital,
300
correspondence to Dr Alan Leviton, Children’s Longwood Avenue, Boston, MA 02115-5747.
trix
hemorrhage,1,2 especially among fetuses in the presentation, 3,4 whereas other studies have demonstrated any such relationship.5-11 Some
breech not
discrepancy might reflect differences in samples in analyses. Indeed, inadequate attention to potential confounders has plagued many studies. Route of delivery is associated with obstetric variables that might convey information about risk of intracranial hemorrhage. Among these variables are length of labor, gestational age, and indications for obstetric intervention.7 For example, the lower the gestational age, the greater the probability of of the
or
breech and transverse lie. 12 Thus, cesarean section for malpresentation provides information about gestational age. Because cesarean section for malpresentation is usually performed very shortly after labor has been unquestionably initiated, cesarean section for malpresentation also provides information about shortened stage 1 labor and the absence of stage 2 labor. The following study was designed to avoid the deficiencies of most of the previous studies evaluat-
35
Downloaded from jcn.sagepub.com at UCSF LIBRARY & CKM on March 16, 2015
the risk of germinal matrix hemorrhage associated with route of delivery and length of labor. This goal was achieved by examining each antecedent in subsamples with and without potentially confounding variables. In addition, multivariate models were also examined.
ing
squares for linear trend. Biserial correlations were obtained relationship between one continuous variable and the dichotomous outcome of germinal matrix hemorrhage present/absent. Because assumptions of normal distributions were not always appropriate, the nonparametric Wilcoxon two-sample test was also used to assess the relationship between labor duration and risk of germinal matrix hemorrhage. Analysis of variance was performed to evaluate the relationship between fetal heart patterns and both gestational age and birth weight. to assess the
Methods
Sample Description All babies bom at the Brigham and Women’s Hospital and at the Beth Israel Hospital in Boston between July 1984 and September 1987, and who weighed no more than 1500 grams, were eligible for this study. Because maternal interview was an integral part of the data to be collected about each infant, babies who died during the first days of life were excluded because their mothers were not interviewed.
Results
Risk of
germinal matrix hemorrhage increased progressively with increasing duration of stage 1 labor, a finding that was more prominent in babies delivered abdominally than in those delivered vaginally (Table 1).
A
post hoc dichotomization of labor, sepa-
rate for each route of
delivery, resulted in four levels matrix germinal hemorrhage risk. Babies born via cesarean section to women who did not go into labor had the lowest risk, and those delivered vaginally after a labor of more than 12 hours had the highest risk. The risk of germinal matrix hemorrhage among babies delivered vaginally did not diminish with of
Data Collected Antecedents. Information was obtained from the mother’s obstetric record about the duration of both stages 1 and 2 of labor, the route of delivery, and the indication for abdominal delivery if cesarean section had been performed. Indications for abdominal delivery were those mentioned in the discharge summary and/or operative note. Gestational age estimates were based on ultrasound examinations performed before the 20th week (n 410) or on physical examination of the newborn when ultrasounds were not available (n 39). Information about fetal heart patterns was obtained from check sheets contained in the clinical chart and routinely completed by the obstetrician. =
=
All of the 449 babies in the sample had a cranial ultrasonogram performed between the 1st and 15th days of life. The ultrasonograms were read by radiologists who had no knowledge of the mother’s obstetric history. Germinal matrix hemorrhage was diagnosed if the supracaudate area on either side was echogenic in more than one plane. Echogenicities were also seen in adjoining parenchyma or a lateral ventricle in 21 of the 72 babies with germinal matrix
Outcome.
hemorrhage.
increasing gestational age (whether defined by dates or physical examination) (Table 2). Among babies delivered abdominally, however, the germinal matrix hemorrhage risk declined prominently with increasing gestational age. Overall, the risk of germinal matrix hemorrhage in babies delivered vaginally is more
nally
than twice that of babies delivered abdomi(odds ratio [OR] 2.5; 95% confidence interval
[CI] 1.4,3.3). Babies with reduced or no variability of the baseline fetal heart rate were not at increased risk of germinal matrix hemorrhage (Table 3). The increased risk of germinal matrix hemorrhage in babies with baseline fetal heart rates less than 120 or greater than 160/minute occurred regardless of the route of
delivery (Table 4). Among babies
Analyses Because many of the labor/delivery items were categorical, data analysis consisted of calculating chi squares, and chi
TABLE 1 Risk of Germinal Matrix and Route of Delivery*
*Risk is
expressed
=
delivered vaginally, the only fetal heart rate response to uterine contraction consistently associated with an increased risk of germinal
Hemorrhage According to Labor Duration
per 100 babies.
36
Downloaded from jcn.sagepub.com at UCSF LIBRARY & CKM on March 16, 2015
TABLE 2 The Risk of Germinal Matrix Hemorrhage Route of Delivery and Gestational Age
who did not have an episiotTheir risk of germinal matrix hemorrhage was only marginally higher than the risk among babies born to women with a mediolateral episiotomy. Babies born to women who had a midline episiotomy had the lowest germinal matrix hemorrhage risk among those delivered via the vaginal route, with a rate that fell between the risks of babies whose mothers had a mediolateral episiotomy and babies delivered abdominally. Among babies delivered abdominally, the only indication associated with a significantly higher risk of germinal matrix hemorrhage was impending amnionitis (Table 7). Preeclampsia was the indication for abdominal delivery associated with the lowest risk of germinal matrix hemorrhage. The reduced risk of germinal matrix hemorrhage in abdominally delivered babies whose mothers were preeclamptic prompted us to reassess Table 1,
According to
matrix hemorrhage was acceleration (Table 5). This association did not reflect differential distributions of birth weight or gestational age among babies who did and did not have heart rate accelerations in utero. No fetal heart pattern was associated with increased risk of germinal- matrix hemorrhage in ba-
bies delivered abdominally. Of the 150 babies delivered
vaginally, only
TABLE 3 The Risk of Germinal Matrix Rate Variability*
*Risk is
born to
omy
(Table 6).
women
Hemorrhage Classified by Baseline Fetal Heart
expressed per 100 babies.
TABLE 4 The Risk of Germinal Matrix
*Risk is
13
were
expressed
TABLE 5 The Risk of Germinal Matrix
Hemorrhage
Classified
by Baseline
Fetal Heart Rate
per 100 babies.
Hemorrhage in
*Children could be counted in multiple strata (ie, itor records. Risk is expressed per 100 babies.
Strata Defined
once
by Fetal Heart Response
for each abnormal
to Uterine Contraction’
pattern). Only two babies had entirely normal
fetal heart
mon-
37
Downloaded from jcn.sagepub.com at UCSF LIBRARY & CKM on March 16, 2015
hemorrhage risk invariably selected a routeof-delivery variable but did not incorporate variables conveying information about labor or gestational age. This was seen in the full sample of 449 infants, as well as in subsamples that excluded babies born
TABLE 6
matrix
Germinal Matrix
Episiotomy
Hemorrhage According
to Maternal
Status*
to toxemic
*Risk is
expressed
per 100 babies.
excluding the 37 babies delivered abdominally of preeclamptic mothers. This did not change the rates appreciably. General anesthesia was associated with a statistically insignificant reduction in germinal matrix hemorrhage risk among babies delivered vaginally (Table 8). Among babies delivered via the abdominal route, general anesthesia was associated with a similar level of risk, which, however, was higher than that seen with other anesthetics. Babies delivered abdominally under local anesthesia had the lowest risk of germinal matrix hemorrhage. All of these babies were born to women who also received general or epidural anesthesia. Stepwise logistic regression models of germinal TABLE 7 Odds Ratios of Germinal Matrix
Abdominal
Hemorrhage According to Indication for
Failure to progress Previous abdominal
delivery Preeclampsia
0.2
Malpresentation Impending amnionitis
are
gestational
Odds Ratio 95% Confidence Bounds
Point Estimate 0.7 1.11 1.5 2.6 0.5 1.11
across
Discussion This study, part of a larger study designed to assess prenatal contributions to the risk of germinal matrix hemorrhage in low birth weight infants, provided an opportunity also to evaluate the contribution of obstetric practices and their indications. The role of delivery events and the occurrence of germinal matrix hemorrhage remains controversial. The conflicting findings can probably be attributed to several phenomena, including quality of data (especially of cranial ultrasound scans or their interpretation) and the magnitude of nonobstetric factors contributing to germinal matrix hemorrhage occurrence. The risk of germinal matrix hemorrhage should vary from institution to institution, reflecting the influences of differing population characteristics and medical care practices. At our institutions during this study, 16% of low birth weight babies had a
Delivery*
Indication Fetal distress Placental disorder
*Odds ratios
women.
0.3,1.5 0.4,3.4 0.7,3.2 1.2,5.7 0.1,4.0 0.2,5.0
, ’
0.0,
relative to all other babies delivered age strata.
TABLE 8 Risk of Germinal Matrix and Anesthesia*
Hemorrhage According
abdominally
to Route of
and
are
0.6 summarized
Delivery
Risk of Germinal Matrix Route of
Delivery Vaginal
Anesthetic
Local
Hemorrhage 26.0 (20/77)
General
Epidural Abdominal
Local General
Epidural *Risk is
expressed per
100 babies.
38
Downloaded from jcn.sagepub.com at UCSF LIBRARY & CKM on March 16, 2015
14.3 (1/7) (12/60) 1.2 (1/87) 15.0 (29/194) 8.0 (11/137)
20.0
cranial ultrasound scan diagnosis of germinal matrix hemorrhage. Institutions with double this rate, however, cannot be expected to report that delivery events have the same relationship to germinal matrix hemorrhage risk as do delivery events at our institutions. In a previous study at our institutions, 32.5% of germinal matrix hemorrhage was evident before the 12th postnatal hour.l3 Labor was associated with increased risk of these early instances of germinal matrix hemorrhage but not with germinal matrix hemorrhage that first occurred at an older age. Institutions that report rates of &dquo;early germinal matrix hemorrhage&dquo; that are double ours should also be expected to experience delivery/germinal matrix hemorrhage relationships that differ from ours. The mechanical hypothesis (including &dquo;battering ram&dquo; and &dquo;increased intravascular pressure&dquo; variations) have their attraction. 14 Indeed, the increased risk of germinal matrix hemorrhage associated with increased labor duration and the reduced risk in babies delivered abdominally are compatible with this hypothesis. Support for the hypothesis that labor is associated with an increased risk of intracranial hemorrhage has come from studies that failed to show a reduced risk among babies delivered abdominally.5,8 Among abdominally delivered babies, the reduced risk associated with local and epidural anesthesia may not reflect &dquo;battering ram&dquo; phenomena. Similarly, the heightened risk of germinal matrix hemorrhage associated with impending amnionitis and the prominently lowered risk in babies of preeclamptic mothers defy a mechanical explanation. The continuum-of-casualty hypothesis states that those adversities leading to death will, if less intense or of shorter duration or if the fetus is especially robust, result not in death, but rather in severe morbidity (eg, &dquo;brain damage&dquo;). 15,16 This concept entered the early fetal heart monitoring literature when some hoped that fetal heart monitoring would not only reduce the risk of intrapartum death but also of cerebral palsy, mental retardation, and seizures. 17 Results of recent studies suggest that this hope will not be achieved. 18-20 Our findings of no increased risk of germinal matrix hemorrhage among babies with reduced baseline variability or with variable and late decelerations are additional examples of the lack of relationship between &dquo;ominous&dquo; fetal heart patterns and neurologic dysfunction 21 or handicap.22 The baseline bradycardia and tachycardia, as well as the postcontraction fetal heart accelerations associated with increased risk of germinal matrix hemorrhage in babies delivered vaginally, may be the first expres-
sion of
germinal matrix hemorrhage, may indicate a physiologic disturbance contributing to an increased risk of germinal matrix hemorrhage, or may be epiphenomena. Obtaining obstetric information from hospital charts has its limitations. 23-26 One obstetrician’s &dquo;impending amnionitis&dquo; may not be the same as another’s. Thus, we urge caution in drawing inferences about those results of this study that reflect
subjective assessments. The findings reported
here should not yet serve the basis for clinical decisions about delivery and anesthesia. Replication, consensus building, and balancing potential advantages and adversities are needed even before clinical trials should be considered. as
Acknowledgments This report was made possible by a grant from the National Institute of Neurological and Communicative Disorders and Stroke (NS 20658) and a Mental Retardation Center Grant from the National Institute of Child Health and Human Development
(HD 06276).
References 1. O’Shea MD, Hage ML, Savitz D: Vaginal delivery and intraventricular hemorrhage in very-low-birthweight (VLBW) infants. Pediatr Res 1988;24:453A. 2. Strauss A, Kirz D, Modanlou HD, Freeman RK: Perinatal events and intraventricular/subependymal hemorrhage in the very low-birth weight infant. Am J Obstet Gynecol 1985; 151:1022-1027. 3. Morales WJ, Koerten J: Obstetric management and intraventricular hemorrhage in very-low-birth-weight infants. Obstet 4.
Gynecol 1986;68:35-40. Tejani N, Verma U, Shiffman R, Chayen B: Effect of route of delivery on periventricular/intraventricular hemorrhage in the low-birth-weight fetus with a breech presentation. J Reprod
Med 1987;32:911-914. 5. Anderson GD, Bada HS, Sibai BM, et al: The relationship between labor and route of delivery in the preterm infant. Am J Obstet Gynecol 1988;158:1382-1390. 6. Bada HS, Korones SB, Anderson GD, et al: Obstetric factors and relative risk of neonatal germinal layer/intraventricular
hemorrhage. Am JObstet Gynecol 1984;148:798-804. W, Ford GW, Doyle LW, et al: Cesarean section or vaginal delivery at 24 to 28 weeks’ gestation: Comparison of survival and neonatal and two-year morbidity. Obstet Gynecol
7. Kitchen
1985;66:149-157. 8. Meidell R, Marinelli P, Pettett G: Perinatal factors associated with early-onset intracranial hemorrhage in premature infants. Am J Dis Child 1985;139:160-163. 9. Newton ER, Haering WA, Kennedy JL Jr, et al: Effect of mode of delivery on morbidity and mortality of infants at early gestational age. Obstet Gynecol 1986;67:507-511. 10. Tejani N, Verma U, Hameed C, Chayen B: Method and route of delivery in the low birth weight vertex presentation corre-
39
Downloaded from jcn.sagepub.com at UCSF LIBRARY & CKM on March 16, 2015
lated with early periventricular/intraventricular hemorrhage. Obstet Gynecol 1987;69:1-4. 11. Welch RA, Bottoms SF: Reconsideration of head compression and intraventricular hemorrhage in the vertex very-low-birthweight fetus. Obstet Gynecol 1986;68:29-34. 12. Placek P: Maternal and infant health factors associated with low infant birth weight: Findings from the 1972 National Natality Survey. In Reed DM, Stanley FJ (eds): The Epidemiology of Prematurity. Baltimore, Urban & Schwarzenberg, 1977, pp 197-212. 13. Leviton A, Pagano M, Kuban KCK: Etiologic heterogeneity of intracranial hemorrhages in preterm newborns. Pediatr Neurol
1988;4:274-278. 14. Schwartz P: Birth Injuries of the Newborn, Morphology, Pathogenesis, Clinical Pathology and Prevention. New York, Hafner Publishing, 1961. 15. Pasamanick B, Knoblock H: Brain damage and reproductive casualty. Am J Orthopsychiatry 1960;30:298-305. 16. Lilienfeld AM, Pasamanick B: The association of maternal factors with the development of cerebral palsy and epilepsy. Am J Obstet Gynecol 1955;70:93-101. 17. Quilligan EJ, Paul RH: Fetal monitoring: Is it worth it? Obstet 18.
Gynecol 1975;45:96-100. Lumley J: Does continuous intrapartum fetal monitoring predict long-term neurological disorders? Paediatr Perinat Epide-
19. Grant A, O’Brien N, Joy M-T, et al: Cerebral palsy among children born during the Dublin randomised trial of intrapartum monitoring. Lancet 1989;2:1233-1236. 20. Shy KK, Luthy DA, Bennett FC, et al: Effects of electronic fetal-heart-rate monitoring, as compared with periodic auscultation, on the neurologic development of premature infants. N
EnglJ Med 1990;322:588-593. 21. Ellison P, Sheridan-Pereira M, MacDonald D, Foster M: Relations between fetal heart monitoring patterns and the neonatal neurological evaluation. Ann Neurol 1989;26:429-430. 22. Painter MJ, Scott M, Hirsch RP, et al: Fetal heart rate patterns during labor: Neurologic and cognitive development at six to nine years of age. Am J Obstet Gynecol 1988;159:854-858. 23. Tilley BC, Barnes AB, Bergstralh, et al: A comparison of pregnancy history recall and medical records: Implications for retrospective studies. Am J Epidemiol 1985;121:269-281. 24. Hewson D, Bennett A: Childbirth research data: Medical records or women’s reports? Am1987;125:484-491. J Epidemiol 25. Martin CJ: Monitoring maternity services by postal questionnaire : Congruity between mothers’ reports and their obstetric records. Stat Med 1987;6:613-627. 26. Bryant HE, Visser N, Love EJ: Records, recall loss, and recall bias in pregnancy: A comparison of interview and medical records data of pregnant and postnatal women. Am J Public
Health 1989;79:78-80.
miol 1988;2:299-307.
40
Downloaded from jcn.sagepub.com at UCSF LIBRARY & CKM on March 16, 2015