Congenital anomalies are an independent risk factor for neonatal morbidity and perinatal mortality in preterm birth

European Journal of Obstetrics & Gynecology and Reproductive Biology 90 (2000) 43–49

www.elsevier.com / locate / ejogrb

Original Article

Congenital anomalies are an independent risk factor for neonatal morbidity and perinatal mortality in preterm birth q a, a a b a b Yifat Linhart *, Asher Bashiri , Eli Maymon , Ilana Shoham-Vardi , Boris Furman , Hillel Vardi , a Moshe Mazor a

Department of Obstetrics and Gynecology, Soroka University Medical Center, P.O. Box 151, Beer-Sheva 84101, Israel b Epidemiology Unit, Faculty of Health Sciences, Ben Gurion University of the Negev, Beer-Sheva, Israel Received 18 January 1999; received in revised form 29 June 1999; accepted 23 July 1999

Abstract Objective: To determine whether congenital anomalies are associated with a high rate of neonatal morbidity in preterm birth. Study design: 312 singletons (22–36 wk) with congenital anomalies that were delivered preterm were compared with a random sample of 936 preterm singleton without congenital anomalies. Data was obtained using the computerized birth discharge records. Statistical analysis included univariate and multivariate logistic regression analyses. Results: Three thousand five hundred and seventy-eight (3578) women with preterm births met the inclusion criteria (singleton with prenatal care). The prevalence of congenital anomalies in the study population was 8.7% (312 / 3578). Gestational age at delivery was significantly lower in the congenital anomaly group compared with the control (32.063.7 SD vs. 34.462.7 SD; p,0.001). The following pregnancy complications were higher in the group with congenital anomalies than in those without anomalies: severe pregnancy induced hypertension (PIH), hydramnions, oligohydramnion, intrauterine growth restriction (IUGR), fetal distress, cesarean section, malpresentation and mal position, abruption placenta, meconium stained amniotic fluid, 1 min Apgar score (,2), 5 min Apgar score (,7). Perinatal mortality rates in 28–32 wk and 33–36 wk were significantly higher in the group with congenital anomalies than in the control group. Neonatal morbidity data (necrotizing enterocolitis, respiratory distress syndrome, bronchopulmonary dysplasia, intraventricular hemorrhage, and sepsis) was available for 909 neonates (239 with congenital anomalies and 670 without congenital anomalies). After adjusting for gestational age, the presence of congenital anomalies remained strongly associated with neonatal morbidity (having one or more of the above mentioned conditions) (adjusted OR: 5.3, 95% CI 3.4–9.2). When adjusting for other confounding variables, congenital anomalies were strongly associated with neonatal morbidity (OR: 6.44, 95% CI 3.94–10.51), and perinatal mortality (OR: 3.08, 95% CI 2.04–4.65). In terms of attributable fraction in our population of preterm births, the proportion of neonatal morbidity and the proportion of perinatal mortality attributable to congenital malformation is 32% and 15%, respectively. Conclusion: Congenital anomalies in preterm birth are associated with a higher rate of pregnancy complications and are an independent risk factor for neonatal morbidity and perinatal mortality.  2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Congenital anomalies; Preterm birth; Perinatal mortality; Neonatal morbidity

1. Introduction In all developed countries the two leading causes of neonatal morbidity and perinatal mortality are preterm q

This study is part of J. Linhart’s MD requirements. *Corresponding author. Tel.: 1972-7-640-0774; fax: 1972-7-6275338.

birth and congenital malformation. According to the Center of Disease Control (CDC) report from 1986, congenital anomalies and prematurity are the leading causes for neonatal mortality (20.5% and 17.5%, respectively) [1,2]. In the United States in 1991, perinatal mortality rates associated with congenital anomalies numbered 186.9 in 100 000 livebirths, and in cases where prematurity was the cause of death, 100.7 in 100 000

0301-2115 / 00 / $ – see front matter  2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S0301-2115( 99 )00196-7

44

Y. Linhart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 90 (2000) 43 – 49

livebirths [3]. In Israel, the Central Bureau of Statistics report from 1997 indicates that congenital anomalies were the leading cause of infant deaths (2.4 in 1000 livebirths) [4]. Several authors have demonstrated that congenital anomalies are more prevalent in preterm birth. Chen et al. [5] and Hartikainen-Sorri et al. [6] reported that congenital anomalies are strongly associated with preterm birth (OR: 6.65 and 5.2, respectively). Herceg et al. [7] have shown that low birth weight (,2500) is associated with congenital anomalies. Neonatal morbidity including Respiratory Distress Syndrome (RDS), Broncho-Pulmonary Dysplasia (BPD), Intra Ventricular Hemorrhage (IVH), Necrotizing Enterocolitis (NEC) and sepsis are associated mainly with prematurity [8–10]. Congenital anomalies are known to be a major cause of morbidity in term deliveries [11], however their contribution to morbidity among infants in preterm births is not well established [12]. The present study was undertaken to test the hypothesis that congenital anomaly in preterm births is an independent risk factor for neonatal morbidity. In addition, maternal characteristics, pregnancy complications and perinatal mortality were determined.

2. Materials and methods The study population consists of women who delivered preterm infants at Soroka University Medical Center between January 1, 1990 and December 31, 1995. During the study period, a total of 59 215 women delivered at our institution, of whom 5088 deliveries were preterm (,37 weeks). After excluding women who had no prenatal care and those with multiple pregnancies, 3578 women with singleton pregnancy and preterm delivery remained the study population. Among these were 312 singleton births with congenital anomalies (one or more). Of the remaining 3266, a sample of 936 births, with no diagnosis of congenital anomaly, was made at birth or during the postnatal period hospitalization, was selected as a comparison group. In general, in our institution the use of prenatal diagnosis of congenital anomalies and selective abortion before 22 weeks of pregnancy is 80%–90%. For each birth with malformation, three births without malformations were randomly selected within each one of the two ethnic groups giving birth in our hospital: Jews and Bedouins. The reason for this selection is that Jews and Bedouins in our area differ in many aspects relevant to the question under study, including etiology and prevalence of congenital malformations, utilization of prenatal diagnostic services, as well as in the range of sociocultural and sociodemographic characteristics [13]. Neonatal morbidity data was obtained through record linkage with the computerized database of the hospital for 909 newborns (239 with congenital anomaly and 670 without congenital

anomaly). For the other births, 73(23%) in the congenital anomalies group and 266(28%) in the group with no congenital anomalies, no hospitalization records were found. The mean follow up period for a newborn, for which morbidity data was collected (from birth date until December 31, 1995), was 732.0 days. A delivery was considered preterm if it occurred before 37 weeks of gestation. Gestational age was determined by a reliable recollection of last menstrual period (with a history of regular cycles, a rise in basal body temperature, a positive urine bHCG test before 6 weeks’ gestation, pelvic examination findings in the first trimester consistent with the stated length of amenorrhea, fetal heart rate determined by continuous wave Doppler ultrasonography before 14 weeks of gestation), confirmed by an ultrasonographic examination within 20 weeks’ gestation or by first trimester sonographic measurement of crown–rump length. The results of the sonographic examinations were available to the clinicians managing the patients. Congenital anomalies were identified in the computerized file of birth discharge record. Linkage with infants’ hospitalization records was performed to obtain information on the specific type of malformation. Bad obstetric history was defined as any of the following: 1) three or more spontaneous first trimester abortions; 2) one or more spontaneous midtrimester abortions; 3) one or more preterm deliveries; 4) one or more stillbirths; 5) one or more low birth weight infants; 6) one or more neonatal deaths [14]. Moderate and severe pregnancy induced hypertension (PIH) were defined as a systolic blood pressure of $140 mmHg or a diastolic blood pressure $90 mmHg, and systolic blood pressure $160 mmHg or a diastolic blood pressure $110 mmHg, respectively on two occasions at least 6 h apart after 20 weeks’ gestation [15]. Patients were defined as having diabetes if gestational diabetes (class A) or insulin dependent diabetes (class B–R), as defined by White [16], were observed. Oligohydramnion was defined as an AFI lower than 5 cm or a subjective estimation of severely decreased amniotic fluid volume. Hydramnions was defined as an AFI greater than 25 cm or a single pocket greater than 8 cm, or as a subjective estimation of increased amniotic fluid volume [17]. Clinical chorioamnionitis was diagnosed in the presence of a body temperature elevation 37.88C and two or more of the following criteria: uterine tenderness, malodorous vaginal discharge, fetal tachycardia, and maternal leukocytosis (white blood cell count .15 000 cell / mm 3 ) [18]. Induction of labor was performed by Foley catheter, oxytocin, prostaglandins, and / or surgical induction. Newborns were considered large for gestational age or small for gestational age when the birth weight was above the 90th or below the 10th percentile, respectively, according to the gestational age at delivery [19]. Neonatal morbidity was defined as the presence of at least one of the following conditions: necrotizing en-

Y. Linhart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 90 (2000) 43 – 49

terocolitis (NEC), respiratory distress syndrome (RDS), bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH), or sepsis. Perinatal mortality was defined as the occurrence of a stillbirth or neonatal death within 28 days of life.

3. Statistical analysis The first stage of analysis consisted of comparing, within each ethnic group, births with and births without congenital anomalies. The following parameters were compared: maternal and obstetric characteristics, birth complications and birth outcome, perinatal mortality and neonatal morbidity. As the association between those parameters and congenital anomalies were consistent in Jews and Bedouins, analysis was repeated for the two ethnic groups combined. Statistical tests to compare proportions were x 2 , or Fisher exact test if required. For continuous variables, group means and standard deviations were computed and the difference was assessed by t-test. To assess possible effect modification by gestational age on the association between the presence of a congenital anomaly and the two major outcomes of interest (perinatal mortality and neonatal morbidity), stratified analysis was performed and Mantel-Haenszel adjusted odds ratios computed. The final stage of the analysis consisted of multivariate analyses. Logistic regression models were computed to assess 1) perinatal mortality, and 2) neonatal morbidity.

4. Results Preterm neonates had a significantly higher rate of congenital anomalies than those who delivered at term (8.7% vs. 2.1%, OR: 4.55, 95% CI 3.98–5.21). The congenital anomalies found in the study population were grouped by anatomical sites: neural tube defects (NTD) – 32(13.4%), head and neck – 11(4.6%), cardiovascular defects (CVD) – 133(55.6%), respiratory system – 8(3.3%), GI system – 19(7.9%), GU system – 20(8.4%), musculoskeletal system – 20(8.4%), chromosomal – 11(4.6%), miscellaneous – 13(5.4%). As some infants had more than one anomaly, the percents exceed 100%. Maternal characteristics and pregnancy complications in the study population and the control group are presented in Table 1. There were no statistically significant differences between the two groups regarding mothers’ age, gravidity and parity, (including grand multiparity rates), as well as in rates of previous cesarean section (CS), previous post partum death (PPD), and habitual abortion. Patients with congenital anomalies were associated with a statistically significantly higher rate of severe PIH (9% vs. 5.7%, p,0.05), IUGR (15.1% vs. 8%, p,0.001), hydramnions

45

(21.2% vs. 7.9%, p,0.001), oligohydramnions (9.3% vs. 3.6%, p,0.001), than the control group. Lower rates of mild / moderate PIH, chronic hypertension, GDM class A, as well as DM class B–R, were noted in the study population, but they were not statistically significant. Gestations with congenital anomaly had statistically significantly higher rates of fetal distress (13.1% vs. 6.5%, p,0.001), CS (37.5% vs. 23%, p,0.001), malpresentation (23.1% vs. 10.3%, p,0.001), meconium stained amniotic fluid (9.6% vs. 6.1%, p,0.05), and abruption placenta (7.7% vs. 4.5%, p,0.05), and were shorter than those without congenital anomalies (32.063.7 wk vs. 34.462.7 wk, p,0.001). No statistical significant differences between the two groups were observed regarding PROM, vacuum delivery, induction of labor, chorioamnionitis, and placenta previa. Table 2 displays a comparison of perinatal outcomes according to the presence or absence of congenital anomalies. Mortality rates were significantly higher in the congenital anomalies group than in the control group: APD (13.1% vs. 4.3%, p,0.001), IPD (2.6% vs. 0.7%, p, 0.05), and PPD (22.1% vs. 2.9%, p,0.001). The mean birth weight was significantly lower in newborns with congenital anomalies than in those without anomalies (1734.06860.5 g vs. 2317.66642.5 g, p,0.001). To test for possible confounding or effect modification of gestational age, a stratified analysis was conducted and odds ratios for neonatal morbidity and perinatal mortality were determined according to gestational age groups. Gestational age was found to modify the effect of congenital anomalies on mortality, but not of morbidity. Fig. 1 presents the perinatal mortality rates by gestational age groups. Perinatal mortality rate 22–27 wk was lower in the congenital anomaly group than in those without anomalies, but did not reach statistical significance (OR: 0.35, 95% CI 0.12–1.06). The perinatal mortality rates in the gestational age groups 28–32 wk and 33–36 wk were higher when associated with congenital anomalies (OR: 2.69, 95% CI 1.36–5.34, p,0.05; and OR: 15.57, 95% CI 8.69–28.07, p,0.001, respectively). Table 3 presents the neonatal morbidity by gestational age groups. Morbidity rates in all gestational age groups were higher in the congenital anomaly group. The adjusted Mantel-Haenszel OR was 5.3, 95% CI 3.4–9.2, p,0.001. Table 4 presents the multivariate adjusted OR’s of congenital anomalies obtained from multiple logistic models including ethnicity, birth weight, severe PIH, polyhydramnion, IUGR, fetal distress, abruption placenta, meconium, SGA, and malpresentation. The adjusted OR’s for congenital anomalies were 6.44 (95% CI 3.94–10.51, p,0.001), when at least one of the neonatal morbidity was present; and 3.08 (95% CI 2.04–4.65, p,0.001), for perinatal mortality. In terms of attributable fraction in our population of preterm births, the proportion of neonatal morbidity and the proportion of perinatal mortality, attributable to

Y. Linhart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 90 (2000) 43 – 49

46

Table 1 Maternal characteristics and pregnancy complications according to the presence of congenital anomalies a

Maternal characteristics: Ethnicity b – Bedouin Jews Maternal age (yr.) – ,19 20–34 351 Mean maternal age (yr.) Gravidity Parity Grandmultiparity ($5 deliveries) Previous PPD Other bad obstetric history Previous cesarean section Pregnancy characteristics: Chronic hypertension Mild / moderate PIH Severe PIH DM class B–R GDM class A Polyhydramnion Oligohydramnion IUGR Pregnancy outcome: PROM Fetal distress Chorioamnionitis Vacuum extraction Induction of labor Cesarean section Gestational age (wk) – 22–27 28–32 33–36 Mean Malpresentation Abruption placenta Placenta previa MSAF

With congenital anomalies N5312 (%)

Without congenital anomalies N5936 (%)

P-value

160(51.3) 152(48.7)

480(51.3) 456(48.7)

25(8.0) 224(72.0) 62(19.9) 28.1166.8 SD 4.063.1 SD 3.362.6 SD 58(18.6)

60(6.4) 689(73.8) 184(19.7) 28.0466.7 SD 4.263.2 SD 3.562.7 SD 165(17.6)

0.873 0.376 0.225 0.701

11(3.5) 38(12.2) 53(17.0)

26(2.8) 107(11.4) 138(14.7)

0.500 0.721 0.340

17(5.4) 10(3.2) 28(9.0) 9(2.9) 21(6.7) 66(21.2) 29(9.3) 47(15.1)

56(6.0) 52(5.6) 53(5.7) 29(3.1) 70(7.5) 74(7.9) 34(3.6) 75(8.0)

0.727 0.097 0.039 0.849 0.659 0.000 0.000 0.000

38(12.2) 41(13.1) 19(6.1) 0(0.0) 49(15.7) 117(37.5)

140(15.0) 61(6.5) 51(5.4) 11(1.2) 118(12.6) 215(23.0)

0.224 0.000 0.669 0.075 0.163 0.000 0.000

49(15.7) 105(33.7) 158(50.6) 32.063.7 SD 72(23.1) 24(7.7) 10(3.2) 30(9.6)

41(4.4) 102(10.9) 793(84.7) 34.462.7 SD 96(10.3) 42(4.5) 21(2.2) 57(6.1)

0.608

0.000 0.000 0.028 0.344 0.034

a

PPD, Post partum death; PIH, Pregnancy induced hypertension; DM, Diabetes mellitus; GDM, gestational diabetes mellitus; IUGR, Intrauterine growth restriction; PROM, Premature rupture of membrane; MSAF, Meconium stained amniotic fluid; SD, standard deviation. b Ethnicity was a matching factor.

congenital malformations are 32.13% and 15.35%, respectively.

5. Comment The prevalence of congenital anomalies in preterm deliveries in this study was 8.7% and it is four-fold higher than in term deliveries. These findings are not surprising and are consistent with other reports [5,6]. The study is the first to address the question of whether

preterm birth with congenital anomalies is associated with higher expression of neonatal morbidity traditionally associated with preterm deliveries (RDS, BPD, NEC, IVH and sepsis). Our findings clearly indicate that preterm neonates who have also congenital anomalies are more prone to develop at least one of these morbidities than infants who do not have anomalies. The additive effect of the presence of anomalies for neonatal morbidity in preterm neonates is surprising and implies that congenital anomaly is not a structural defect only, but may have a more profound exposing effect. The anatomical site of the anomaly may

Y. Linhart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 90 (2000) 43 – 49

47

Table 2 Perinatal outcome of newborns with or without congenital anomalies a With congenital anomalies N5312 (%)

Without congenital anomalies N5936 (%)

P-value

Perinatal mortality: APD IPD PPD Total

41(13.1) 8(2.6) 69(22.1) 118(37.8)

40(4.3) 7(0.7) 27(2.9) 74(7.9)

0.000 0.016 0.000 0.000

LGA SGA 1 min Apgar (,2) 5 min Apgar (,7)

18(5.8) 54(17.3) 45(14.4) 50(16.0)

73(7.8) 63(6.7) 42(4.5) 32(3.4)

0.232 0.000 0.000 0.000

Birth weight (g) – 500–750 750–1000 1000–1500 1500–2000 2000–2500 Mean

21(8.1) 52(20.1) 66(25.5) 62(23.9) 58(22.4) 1734.06860.5 SD

29(5.3) 15(2.7) 50(9.1) 136(24.7) 321(58.3) 2317.66642.5 SD

Sex – Male Female

156(50.0) 156(50.0)

515(55.0) 421(45.0)

a

0.000

0.000 0.123

APD, Antepartum death; IPD, Intrapartum death; PPD, Postpartum death; LGA, Large for gestational age; SGA, Small for gestational age.

play a role in exposing preterm neonates with anomalies to one of the morbidities associated with prematurity, however this study did not look into this aspect. We found that several pregnancy complication rates

were significantly higher in women who delivered preterm infants with congenital anomalies: severe PIH, polyhydramnion, oligohydramnion, and IUGR. Each one of these complications was reported to be associated with preterm

Fig. 1. Perinatal mortality by gestational age groups.

Y. Linhart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 90 (2000) 43 – 49

48

Table 3 Neonatal morbidity by gestational age groups Gestational age (wk)

With congenital anomalies N5239 (%)

Without congenital anomalies N5670 (%)

OR

P-value

95% CI a

22–27 28–32 33–36

14(41.2) 45(57.7) 24(18.9)

2(8.0) 15(28.8) 18(3.0)

8.05 3.36 7.44

0.004 0.001 0.000

1.51–78.91 1.49–7.65 3.73–14.91

a

CI, Confidence interval.

birth (spontaneous or induced) and / or congenital anomalies, but our findings showed that the combination of preterm birth with congenital anomalies has a higher risk of pregnancy complications than preterm delivery or congenital anomalies alone. Higher rates of fetal distress, cesarean section, very short gestations (22–32 wk), malpresentation, abruption placenta, and meconium, were also observed in the congenital anomalies group. These findings contribute to

the uniqueness of this group, and to the fact that pregnancies with congenital anomalies are at a higher risk of delivery in addition to higher risk pregnancy. Because the study did not include women with no prenatal care, all the women in this study were followed regularly and were screened as needed. The study population includes women who delivered preterm infants (22– 36 wk). The indications for active management (caesarian section, induction of labor) in the case of prenatal diag-

Table 4 Multivariate logistic regression for neonatal morbidity and perinatal mortality a Variable

Morbidity b OR (95% CI)

P-value

Mortality c OR (95% CI)

P-value

Congenital anomalies (y / n)

6.44 (3.94–10.51)

0.000

3.08 (2.04–4.65)

0.000

Ethnicity (Jew / Bedouin)

1.23 (0.77–1.96)

0.377

2.08 (1.38–3.15)

0.000

Birth weights (100 g)

0.88 (0.85–0.92)

0.000

0.81 (0.78–0.85)

0.000

SGA (y / n)

0.33 (0.15–0.71)

0.004

2.31 (1.23–4.33)

0.008

Severe PIH (y / n)

2.71 (1.22–6.04)

0.014

0.33 (0.14–0.77)

0.011

Polyhydramnion (y / n)

0.65 (0.33–1.31)

0.234

3.55 (2.04–6.19)

0.000

IUGR (y / n)

0.60 (0.26–1.36)

0.224

0.47 (0.22–0.97)

0.042

Fetal distress (y / n)

1.59 (0.75–3.38)

0.219

0.37 (0.18–0.76)

0.006

Abruption placenta (y / n)

1.34 (0.58–3.10)

0.488

0.98 (0.46–2.06)

0.962

MSAF (y / n)

0.95 (0.40–2.25)

0.909

2.44 (1.23–4.81)

0.010

Malpresentation (y / n)

0.60 (0.32–1.12)

0.110

1.49 (0.90–2.45)

0.116

a SGA, Small for gestational age; PIH, Pregnancy induced hypertension; IUGR, Intrauterine growth restriction; MSAF, Meconium, stained amniotic fluid; CI, confidence interval. b N5909. c N51248.

Y. Linhart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 90 (2000) 43 – 49

nosis of anomalies in our institute are based only on obstetric needs, and not on the diagnosis of an anomalies alone. Induced abortions are not included in this study, since abortions are caring out before 22 weeks of gestation. Late abortion (.22 wk) were done by feticide and were not included in the study. In relation to induction of labor, we found no significant difference between the two groups, which indicate that the anomaly itself did not influence the decision of induction. We found that premature infants with congenital anomalies had poor perinatal outcome that includes: lower mean birth weight, SGA, lower Apgar scores, and were at higher risk for perinatal mortality. These findings were partially reported by other investigators [12]. After adjusting for gestational age at delivery, in order to control for the influence of prematurity, we found that perinatal mortality remained higher in the congenital anomalies group in 28–36 wk. This means that the very premature infants (22–28 wk) are at a higher risk for mortality regardless of having a congenital anomaly. In contrast, congenital anomalies are a risk factor for perinatal mortality in the more mature group (28–36 wk). When the presence of congenital anomalies was adjusted for other conditions usually associated with adverse perinatal outcome, which were found to differ between the two groups, the risk for perinatal mortality associated with congenital anomalies in preterm delivery was more than 3-fold compared with those without congenital anomalies. In relation to neonatal morbidity, newborns with congenital anomalies in preterm birth are at higher risk for neonatal morbidity after adjusting for gestational age. These findings were not previously described. Others reported morbidity more in general terms and in relation to the hospitalization period of neonates [12]. As for perinatal mortality, after adjusting for conditions that are usually associated with poor neonatal outcome, the risk of neonatal morbidity associated with congenital anomalies was 6-fold higher than infants without congenital anomalies in preterm birth. This means that congenital anomalies in patients with preterm birth are an independent risk factor for neonatal morbidity. When we conducted the multivariate logistic regression model, we found that severe PIH and fetal distress had a protective influence on perinatal mortality. These two entities are indications for active management, as the risk for the mother or the infant justifies an immediate management. Moreover, fetal distress promotes lung maturity in the fetus, which lowers the risk of respiratory failure in the infant. Another surprising finding is the protective influence that SGA and polyhydramnions have on neonatal

49

morbidity. We could not find an explanation for this and further studies should be done to investigate it.

References [1] Contribution of birth defects to infant mortality – United States, 1986. In: MMWR, Vol. 38, 1989, pp. 633–5. [2] Cordero JF. Finding the causes of birth defect. N Engl J Med 1994;331:48–9, Editorial. [3] Singh GK, Yu SM. Infant mortality in the United States: Trends, differentials, and projections, 1950 through 2010. Am J Public Health 1995;85:957–64. [4] Statistical abstract of Israel 1997;48:144. [5] Chen CP, Wang KG, Yang YC, See LC. Risk factors for preterm birth in an upper middle class Chinese population. Eur J Obstet Gynecol Reprod Biol 1996;70:53–9. [6] Hartikainen-Sorri AL, Sorri M. Occupational and sociomedical factors in preterm birth. Obstet Gynecol 1989;74:13–6. [7] Herceg A, Simpson JM, Thompson JF. Risk factors and outcome associated with low birth weight delivery in the Australian Capital Teritory 1989–90. J Paediatr Child Health 1994;30:331–5. [8] Berkowitz GS, Papiernik E. Epidemiology of preterm birth. Epidemiol Rev 1993;15:414–43. [9] Mupanemunda RH, Azzopardi D. Recent advances in the management of the preterm infant. In: Elder MG, Lamont RF, Romero R, editors, Preterm Labor, 1997, pp. 493–516. [10] Robertson PA, Sniderman SH, Laros Jr RK et al. Neonatal morbidity to gestational age and birth weight from five tertiary care centers in the United States 1983 through 1986. Am J Obstet Gynecol 1992;166:1629–45. [11] Rosenthal N, Abramowsky CR. The causes of morbidity and mortality among infants born at term. Arch Pathol Lab Med 1988;112:178–81. [12] Ling EW, Sosuan LC, Hall JC. Congenital anomalies: an increasingly important cause of mortality and workload in a neonatal intensive care unit. Am J Perinatol 1991;8:164–9. [13] Weitzman D., Shoham-Vardi I., Belmaker I., Siton Y., Carmi R. Factors affecting utilization of prenatal procedures in a traditional society. Proceeding of the 15 th Annual Meeting of Israel Epidemiology Society, 1997, p. 5. [14] Arias F, Tomich P. Etiology and outcome of low birth weight and preterm infants. Obstet Gynecol 1982;60:277–81. [15] Davey D, MacGillivary I. The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 1988;158:892–8. [16] White P. Pregnancy complicating diabetes. Am J Med 1949;7:609– 15. [17] Mazor M, Ghezzi F, Maymon E et al. Polyhydramnions is an independent risk factor for perinatal mortality and intrapartum morbidity in preterm delivery. Eur J Obstet Gynecol Reprod Biol 1996;70:41–7. [18] Gibbs RS, Blanco JD, St Clair PJ, Castaneda YS. Quantitative bacteriology of amniotic fluid from women with clinical intraamniotic infection at term. J Infect Dis 1982;145:1–8. [19] Leiberman JR, Fraser D, Weitzman S, Glezerman M. Birth weight curves in southern Israel populations. Isr J Med Sci 1993;29:198– 203.