Objective
The objective of the study was to ascertain the association between fetal growth (small- [SGA], appropriate- [AGA], and large-for-gestational-age [LGA]) and early, late, and postneonatal mortality.
Study Design
Birth certificate data for nonanomalous singletons, delivered from 1996 to 2007, were obtained for Milwaukee residents. Multivariate logistic regression analyses, adjusted for 19 covariates, determined the association between fetal growth and mortality.
Results
Among the 123,383 live births, SGA was 57% higher than LGA (11% vs 7%). The infant mortality rate for SGA was 11.0, AGA, 5.3, and LGA, 2.7/1000 live births. SGA was a significant risk factor for early (adjusted odds ratio, 2.66) and late (2.06) but not postneonatal mortality. The adjusted risk of mortality for LGA was not significantly different from AGA. Over 12 years, 3 types of mortality for aberrant fetal growth did not change significantly.
Conclusion
In the city of Milwaukee, aberrant fetal growth was variably associated with early, late, and postneonatal mortality.
Aberrant fetal growth encompasses small for gestational age (SGA) or large for gestational age (LGA). Whereas SGA is defined as neonatal birthweight below 10% for the gestational age, LGA is weight above 90%, with macrosomia (weight of at least 4000 g) being a subset of accelerated growth. Fetal growth abnormalities are linked with complications. SGA, for example, is associated with fetal anomalies, oligohydramnios, stillbirth, neonatal acidosis, seizure, and death. Later in life, newborns with suboptimal growth are at increased risk of learning disabilities and cardiovascular disease. On the other hand, LGA may lower or increase the likelihood of mortality. Among diabetic mothers, accelerated growth is associated with traumatic delivery and stillbirth ; among women with pregnancy-induced hypertension, preterm LGA has significantly lower infant mortality.
The myriad complications associated with aberrant growth and the conflicting reports on mortality with LGA prompted us to inquire if SGA or LGA is associated with early (0-6 days after live birth), late (7-27 days), or postneonatal (28-364 days) mortality.
The primary purpose of this population based study was to determine whether SGA or LGA is associated with increased risk of early, late, and postneonatal mortality, and identify other risk factors; the secondary purpose was to determine whether there are temporal changes in the 3 subtypes of infant mortality.
Materials and Methods
This project was reviewed and determined exempt by the Health Sciences Institutional Review Board at the University of Wisconsin–Madison. For the years 1996-2007, we obtained vital statistics birth certificate data for the city of Milwaukee, WI, from the City of Milwaukee Health Department. The birth certificate data were linked to death certificate data, and mortality information was included. The study sample was restricted to singleton live births born to city of Milwaukee resident mothers with gestational age of 24 weeks or more.
Newborns with congenital anomaly or implausible birthweight–gestational age combinations were excluded. Implausible birthweight–gestational age data were identified by the algorithm that was developed by Alexander et al. This method was based on expert opinion concerning some questionable birthweight values that may be a result of inaccurate recording of gestational age. In addition, infant death cases with unknown time of death were also excluded.
The outcomes of interest were early, late, and postneonatal mortalities. Early mortality was defined as newborns that died during 0-6 days of life; late mortality, 7-27 days; and postneonatal mortality, 28-364 days. Infant mortality rate (IMR) was the summation of early, late, and postneonatal mortality rate. Based on Alexander et al, fetal growth was categorized as SGA, appropriate for gestational age (AGA; birthweight lies within 10-90% for that gestational age), and LGA.
Nineteen covariates obtained from vital statistics birth certificate data were included in the analyses. Maternal age was categorized into 3 groups: 19 years or less, 20-34 years, and 35 years or older. Self-reported race/ethnicity was categorized into white, African American, Hispanic, and other. Maternal educational attainment was categorized into completed high school, beyond high school, and less than high school.
Paternity was coded as having a father’s name listed on the birth certificate (yes/no). The father’s name on the record meant one of the following: the parents were married at the time of birth; there was a statement of paternity; there was a paternity adjudication by court order (meaning paternity was established by the legal system, usually after genetic testing confirms the father’s identity); or legitimation (meaning that the child is acknowledged by parents who marry after their child’s birth). We used finalized birth certificate data that included amendments made by the vital records office (such as statements of paternity, adjudications, or legitimations).
Prenatal care was categorized as adequate vs inadequate, using Kotelchuck’s Adequacy of Prenatal Care Utilization Index. Kotelchuck’s index is based on initiation of prenatal care (month prenatal care begins) and the number of visits received, and is one of the most commonly used indices of prenatal care utilization in research studies.
Maternal behaviors and obstetric complications included reported alcohol use during pregnancy (yes/no), reported cigarette use during pregnancy (yes/no), pregestational diabetes (yes/no), gestational diabetes (yes/no), hydramnios or oligohydramnios (yes/no), gestational or pregestational hypertension (yes/no), uterine bleeding (yes/no), and premature rupture of membranes of more than 12 hours (yes/no).
Gestational age at birth was defined as the clinician’s estimate of gestation (number of completed weeks) as recorded on the birth certificate. We then categorized gestational age into preterm (<37 weeks), term (37-40 weeks), and postterm (≥41 weeks). The route of delivery was categorized as spontaneous vaginal, operative vaginal, and cesarean. We also examined whether the labor was precipitous (within 3 hours vs >3 hours), abruptio placenta (yes/no), and infant sex (male/female).
Neonatal condition at birth was categorized as normal vs abnormal. Abnormal conditions included birth injury, hyaline membrane disease or respiratory distress, meconium aspiration syndrome, assisted ventilation, and seizures at birth.
We calculated corrected early, late, postneonatal, and infant mortality rate by fetal growth (SGA, AGA, and LGA). We assessed 3 year moving average mortality trends by fetal growth (SGA, AGA, and LGA) category. A χ 2 test or Fisher’s exact test was used to assess the mortality trend across years. Multivariate logistic regression analyses were performed to examine the association between fetal growth and outcomes of the 3 types of mortality and adjusting for 19 potential confounding variables. The results are presented as odds ratios (ORs) and their 95% confidence intervals (CIs). Significance level was defined as P < .05.
For all variables, the data were complete at above 99% of the study subjects. On birth certificates, for maternal medical history, a box must be checked if a mother has a certain medical condition. We assumed that a mother did not have the condition if the box was not checked. For the frequency tables, missing values were excluded and percentages were based on the number of nonmissing values. When performing the multivariate logistic regression analyses, listwise deletion was used to remove subjects if there was a missing value on any of the variables. All statistical analyses were performed using SAS 9.2 (SAS/STAT software, version 9.2; SAS Institute, Cary, NC).
Results
During the 12 years of the study period, there were 132,658 live births in Milwaukee, WI. We excluded 3879 multiple gestation births, 3559 deliveries of patients who resided outside Milwaukee, 1227 anomalous neonates, 540 newborns delivered before 24 weeks, 36 infants with implausible birthweights, 27 infant deaths with unknown time of death, and 7 others. After these exclusions, there were 123,383 (93%) live births remaining as our study population. The median number of deliveries per year was 10,257 (range, 9956–10,708). Whereas 11% of the newborns were SGA (n = 13,601), 7% were LGA (n= 8957), and the remaining 82% were AGA (n = 100,825). The 3 year moving average for the incidence of SGA (range, 10.7–11.6%; P = .586), AGA (range, 81.3–81.9%; P = .987) and for LGA (range, 6.6–7.6%; P = .140) was not significantly different across the 12 years. Table 1 presents the sample characteristics.
| Mortality | ||||
|---|---|---|---|---|
| Characteristic | Alive (n = 122,676) | Early (n = 158) | Late (n = 117) | Postneonatal (n = 432) |
| Maternal age, y | ||||
| ≤19 | 19% (23,357) | 23% (37) | 27% (31) | 28% (120) |
| 20-34 | 72% (88,428) | 65% (102) | 67% (78) | 67% (288) |
| ≥35 | 9% (10,891) | 12% (19) | 7% (8) | 6% (24) |
| Mother’s race/ethnicity | ||||
| White | 30% (36,692) | 21% (33) | 25% (29) | 19% (81) |
| Black | 47% (57,034) | 57% (90) | 62% (73) | 69% (298) |
| Hispanic | 18% (22,494) | 17% (27) | 9% (11) | 9% (37) |
| Other | 5% (6435) | 5% (8) | 3% (4) | 4% (16) |
| Mother’s education | ||||
| Less than high school | 35% (42,873) | 35% (51) | 46% (53) | 51% (218) |
| High school | 32% (39,410) | 43% (63) | 34% (39) | 32% (136) |
| Above high school | 33% (39,989) | 23% (34) | 20% (23) | 17% (74) |
| Paternity | ||||
| Has father on record | 71% (86,835) | 35% (55) | 34% (40) | 35% (153) |
| No father on record | 29% (35,841) | 65% (103) | 66% (77) | 65% (279) |
| Prenatal care | ||||
| Adequate or better | 68% (82,894) | 65% (97) | 68% (78) | 62% (266) |
| Inadequate or better | 32% (39,069) | 35% (53) | 32% (36) | 38% (166) |
| Alcohol use during pregnancy | ||||
| No | 98% (120,672) | 94% (147) | 97% (114) | 96% (415) |
| Yes | 2% (1938) | 6% (10) | 3% (3) | 4% (17) |
| Cigarette use during pregnancy | ||||
| No | 85% (103,980) | 81% (127) | 73% (85) | 69% (296) |
| Yes | 15% (18,639) | 19% (30) | 27% (32) | 32% (136) |
| Gestational diabetes | ||||
| No | 97% (118,629) | 96% (151) | 96% (112) | 97% (419) |
| Yes | 3% (4047) | 4% (7) | 4% (5) | 3% (13) |
| Pregestational diabetes | ||||
| No | 99% (121,973) | 99% (156) | 99% (116) | 100% (430) |
| Yes | 1% (703) | 1% (2) | 1% (1) | 0% (2) |
| Hydramnios/oligohydramnios | ||||
| No | 99% (121,274) | 93% (147) | 95% (111) | 97% (420) |
| Yes | 1% (1,402) | 7% (11) | 5% (6) | 3% (12) |
| Hypertensive disease | ||||
| No | 94% (115,810) | 91% (144) | 91% (106) | 91% (394) |
| Yes | 6% (6866) | 9% (14) | 9% (11) | 9% (38) |
| Uterine bleeding | ||||
| No | 99% (121,762) | 98% (154) | 99% (116) | 98% (424) |
| Yes | 1% (914) | 3% (4) | 1% (1) | 2% (8) |
| Premature rupture of membranes (>12 h) | ||||
| No | 97% (118,817) | 89% (140) | 92% (108) | 96% (415) |
| Yes | 3% (3859) | 11% (18) | 8% (9) | 4% (17) |
| Gestational age, wks | ||||
| <37 (preterm) | 9% (10,709) | 75% (118) | 62% (72) | 33% (142) |
| 37-40 (term) | 82% (100,070) | 22% (35) | 33% (38) | 60% (261) |
| ≥41 (postterm) | 10% (11,897) | 3% (5) | 6% (7) | 7% (29) |
| Precipitous labor (<3 h) | ||||
| No | 97% (118,628) | 94% (149) | 97% (114) | 96% (414) |
| Yes | 3% (4048) | 6% (9) | 3% (3) | 4% (18) |
| Delivery method | ||||
| Spontaneous vaginal | 80% (97,489) | 58% (92) | 59% (69) | 74% (321) |
| Operative vaginal | 5% (5539) | 2% (3) | 3% (3) | 5% (20) |
| Cesarean | 16% (19,648) | 40% (63) | 39% (45) | 21% (91) |
| Newborn’s sex | ||||
| Male | 51% (62,478) | 51% (80) | 52% (61) | 57% (247) |
| Female | 49% (60,198) | 49% (78) | 48% (56) | 43% (185) |
| Fetal growth | ||||
| SGA | 11% (13,451) | 32% (50) | 26% (30) | 16% (70) |
| AGA | 82% (100,292) | 64% (101) | 72% (84) | 81% (348) |
| LGA | 7% (8933) | 4% (7) | 3% (3) | 3% (14) |
| Abruptio placenta | ||||
| No | 99% (121,989) | 93% (147) | 92% (108) | 97% (418) |
| Yes | 1% (687) | 7% (11) | 8% (9) | 3% (14) |
| Newborn’s condition | ||||
| Abnormal | 7% (8749) | 36% (57) | 37% (43) | 19% (80) |
| Normal | 93% (113,927) | 64% (101) | 63% (74) | 82% (352) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
