Reddy, Uma M. MD, MPH; Bettegowda, Vani R. MHS; Dias, Todd MS; Yamada-Kushnir, Tomoko MPH, MS; Ko, Chia-Wen PhD; Willinger, Marian PhD
Term pregnancy, defined as 37–41 weeks of gestation (260–294 days), is generally regarded as a period of homogeneous pregnancy risk. Studies that investigate perinatal outcomes often use deliveries that occur over the entire length of the term period as the reference group. This definition of term gestation, however, was determined in a relatively arbitrary fashion.1
There is limited evidence that neonates born between 37 0/7 and 38 6/7 weeks of gestation, referred to as “early”-term births, have increased mortality and neonatal morbidity as compared with neonates born at later gestational age, referred to as “full” term.2,3 An analysis of U.S. singleton term nonanomalous live births among non-Hispanic white women between 1995 and 2001 showed that the neonatal mortality rate decreased with increasing gestational age from 0.66 per 1,000 live births at 37 weeks to 0.33 per 1,000 live births at 39 weeks and then remained stable until 40 weeks.3 The increased risk of neonatal, postneonatal, and infant mortality associated with late preterm deliveries (34–36 weeks of gestation) persisted in the early term period (37–38 weeks of gestation) when compared with deliveries at 40 weeks of gestation.3 Similar observations were made in a study of births from 1999–2004 in Utah.4 However, there is a public perception that the early-term and full-term periods are equivalent, a homogeneously low-risk period. In a recent survey (n=650), over half of insured first-time mothers who delivered within the past 18 months believed that full term was reached at 37–38 weeks of gestation and most believed it is safe to deliver before 39 weeks of gestation when there are no other medical complications requiring early delivery.5
The purpose of this study was to estimate the trend over the past decade for maternal racial and ethnic differences in neonatal, postneonatal, and infant mortality for early-term (37 0/7 to 38 6/7 weeks of gestation) when compared with full-term births (39 0/7 to 41 6/7 weeks of gestation). Furthermore, we estimated if the distribution of the causes of neonatal and postneonatal death differ for early-term and full-term births by maternal race and ethnicity.
MATERIALS AND METHODS
We analyzed National Center for Health Statistics U.S. period-linked birth and infant death data for 1995–2006. Beginning in 1995, linked files were produced by the National Center for Health Statistics using the period format and consist of all infant deaths in a specific year that had been linked to their corresponding birth certificates in the same or preceding year. The National Center for Health Statistics applies a weight to records in the data file to account for those infant deaths that could not be linked to the corresponding birth certificate.6 The linkage rate was greater than 97% during the study period of 1995–2006.
Analysis was limited to singleton live births between 37 and 41 completed weeks of gestation. Gestational age in the period-linked file was reported by the National Center for Health Statistics as completed weeks of gestation and measured as the interval between the first day of the mother's last menstrual period and the date of birth, except when gestational age is inconsistent with birth weight and plurality. In these cases, if only the month and year of the last menstrual period are available, the National Center for Health Statistics imputes the gestational age by assigning the weeks of gestation of the previous completed record in the file with a similar race and birth weight.7 In cases in which the month, year, or entire last menstrual period is missing, or when the imputed gestational age appears to be inconsistent with birth weight, the clinical estimate of gestation is substituted, occurring in approximately 5% of live births primarily as a result of missing last menstrual period.6
Term infants were defined as those born between 37 0/7 and 41 6/7 weeks of gestation and categorized as early-term and full-term. Early-term infants were defined as those born between 37 0/7 and 38 6/7 weeks of gestation, and full-term neonates were defined as those born between 39 0/7 and 41 6/7 weeks of gestation.
We examined infant deaths separating the neonatal and postneonatal period. Infant deaths were all those between 0 and 364 days of life. Neonatal deaths were those between 0 and 27 days of life; postneonatal deaths were those between 28 and 364 days. Infant mortality rates were calculated as the number of infant deaths by age at time of death per 1,000 live births.
To estimate differences in cause-specific infant mortality rates by maternal race and ethnicity and age at time of infant death, we aggregated data for 2000–2006. Cause-specific infant mortality rates were defined as the number of infant deaths resulting from a specific cause of death per 100,000 live births. Rates for specific causes of death are based solely on the underlying cause of death. Multiple conditions or causes reported in the cause of death section of the death certificate are converted to a single underlying cause of death following the World Health Organization rules. Underlying causes of death were coded according to International Classification of Diseases, 10th Revision for 2000–2006. The International Classification of Diseases, 10th Revision 130 selected causes of infant death were assigned 1 of 71 rankable causes of infant death as defined by the National Center for Health Statistics.8
Rates of infant mortality by age at time of infant death between 1995 and 2006 were calculated for each gestational week for term Hispanic, non- Hispanic white, and non-Hispanic black infants. Relative risk ratios by age at time of death and maternal race and ethnicity were calculated for infants born at 37, 38, 39, and 41 completed weeks of gestation compared with gestational age of 40 completed weeks for all deaths and deaths excluding birth defects. Relative risk ratios of infant mortality by age at time of death and early-term and full-term gestational age groups for 2006 were calculated for Hispanic and non-Hispanic black infants compared with non-Hispanic white infants. Five leading neonatal and postneonatal cause-specific mortality rates for 2000–2006 by maternal race and ethnicity were calculated for early-term and full-term infants.
Data analyses were conducted with SAS 9.1.3. P values were computed by chi-square test for association between maternal characteristics and gestational age groups. We used Poisson regression to estimate the trend in mortality rates between 1995 and 2006 by using the number of deaths as the response variable with a population offset and added years to the model as an independent variable. To assess significance of the relative risk ratios, confidence intervals (CIs) were constructed. Statistical significance was set a priori at a P<.05.
During the study period, 1995–2006, there were 46,779,901 singleton live births. A total of 46,329,018 (99.0%) had gestational age reported, of which 11,717,596 (25.3%) were early term and 26,450,038 (57.1%) were full term. In 1995, term deliveries represented 81.3% of singleton births; early term comprised 21.8% and full term comprised 59.5% of singleton deliveries. On the other hand, there was a shift in 2006 with term deliveries representing 83.1% of singleton births; early term comprised 28.9% and full term comprised 54.2% of singleton deliveries.
The distribution of births for all racial or ethnic groups increased at 37–39 weeks of gestation and decreased at 40–41 weeks. The proportion of early-term births out of all term births is consistently higher for non-Hispanic blacks compared with the other racial and ethnic groups of 31.4% in 1995 increasing to 38.3% in 2006. The proportion of early-term births out of all term births was 27.4% for Hispanics and 25% for non-Hispanic whites in 1995 increasing to 34.2% for both groups in 2006. Across all race and ethnicity groups, early-term births were more likely to be associated with advanced maternal age (older than 35 years old), multiparity, and married status compared with full-term births (Table 1).
Early-term infants (37 and 38 weeks of gestation) had higher infant mortality rates when compared with full-term infants (39–41 weeks of gestation) for every year of the study period (Fig. 1). Between 1995 and 2006, the infant mortality rate was highest at 37 weeks of gestation with a rate of 5.1 per 1,000 in 1995 declining to 3.9 per 1,000 in 2006. The infant mortality rate at 38 weeks of gestation was 3.4 per 1,000 in 1995 declining to 2.5 per 1,000 in 2006. These rates compare with an infant mortality rate at 40 weeks of gestation of 2.6 per 1,000 in 1995 declining to 1.9 per 1,000 in 2006. Mortality for infants born at 41 weeks of gestation decreased the least over this time period.
When examining early-term infant mortality rate by race and ethnicity, there are differences in trends over the study period. Among births at 37 weeks of gestation, Hispanics had the greatest decline in infant mortality rate of 35.4% followed by a 22.4% decline for non-Hispanic whites (Table 2; also, Appendix 1, available online at http://links.lww.com/AOG/A241). In contrast, the infant mortality rate among non-Hispanic blacks declined only 6.8%. This disparity is attributable to the divergent trends in neonatal mortality rates. The neonatal mortality rates at 37 weeks of gestation among Hispanics and non-Hispanic whites declined 34.8% and 33.3%, respectively, whereas the neonatal mortality rate among non-Hispanic blacks increased by 15.8% (Table 2). Although the increase is not statistically significant, there is clearly no reduction in the rate over time. In fact, the neonatal mortality rate for non-Hispanic blacks infants born at 37 weeks of gestation in 2006 is similar to the neonatal mortality rate in 1995 for non-Hispanic whites and Hispanics at 37 weeks of gestation. The pattern is somewhat different for postneonatal mortality. At 37 weeks of gestation, the postneonatal mortality rate declined between 1995 and 2006 by 36.0% for Hispanics followed by a decline of 22.0% for non-Hispanic blacks and 15.4% for non-Hispanic whites (Table 2).
For infants born at 38 weeks of gestation, the decline in neonatal mortality rate among non-Hispanic blacks (14.3%) is less than that of non-Hispanic whites (35.7%) and Hispanics (23.1%). Similarly, the decline for postneonatal mortality rate is less for non-Hispanic blacks (14.3%) compared with non-Hispanic whites (26.3%) and Hispanics (25.0%) (Table 2).
Among all races and ethnicities, early term is a period of consistently higher neonatal mortality rates when compared with full term. Figure 2 demonstrates that neonatal mortality rate is highest at 37 weeks of gestation, 1.5, 1.6, and 2.2 per 1,000 live births in 2006 for Hispanics, non-Hispanic whites, and non-Hispanic blacks, respectively. Likewise, the neonatal mortality rate is lowest at 40 weeks of gestation, 0.6, 0.6, and 0.8 per 1,000 live births for Hispanics, non-Hispanic whites, and non-Hispanic blacks, respectively.
The increased risk of neonatal death at 37 weeks compared with 40 weeks of gestation is as follows: Hispanics—relative risk, 2.6 (95% CI 2.0–3.3); non-Hispanic whites—relative risk, 2.6 (95% CI 2.2–3.1); and non-Hispanic blacks—relative risk, 2.9 (95% CI 2.2–3.8) (Table 3). The excess risk in neonatal mortality at 38 weeks compared with 40 weeks of gestation declines significantly: Hispanics—relative risk, 1.7 (95% CI 1.3–2.1); non-Hispanic whites—relative risk, 1.5 (95% CI 1.3–1.8); and non-Hispanic blacks—relative risk, 1.5 (95% CI 1.2–2.0).
When deaths resulting from birth defects were excluded, the magnitude of the increased risk of neonatal mortality for infants born at 37 weeks compared with 40 weeks remained the same for Hispanic and non-Hispanic black infants but declined from a relative risk of 2.6 to a relative risk of 1.8 (95% CI 1.5–2.3) for non-Hispanic white infants. When deaths resulting from birth defects were excluded, there was still an excess risk of neonatal mortality for neonates born at 38 weeks compared with 40 weeks: Hispanics—relative risk, 1.6 (95% CI 1.1–2.3); non-Hispanic whites—relative risk, 1.2 (95% CI 1.0–1.5); and non-Hispanic blacks—relative risk, 1.3 (95% CI 0.9–2.0) (data not shown).
Early term is also a period of consistently higher postneonatal mortality rates when compared with 39–41 weeks of gestation for all races and ethnicities (Table 3). The increased relative risk of an infant born at 37 weeks dying in the postneonatal period compared with 40 weeks is as follows: Hispanics—relative risk, 1.7 (95% CI 1.3–2.1); non-Hispanic whites—relative risk, 1.8 (95% CI 1.6–2.1); and non-Hispanic blacks—relative risk, 1.6 (95% CI 1.3–1.9). The excess risk in postneonatal mortality at 38 weeks compared with 40 weeks is less: Hispanics—relative risk, 1.2 (95% CI 1.0–1.5); non-Hispanic whites—relative risk, 1.2 (95% CI 1.0–1.3); and non-Hispanic blacks—relative risk, 1.4 (95% CI 1.2–1.7). The magnitude of the excess risk of postneonatal mortality for infants born at 37 or 38 weeks compared with 40 weeks remained the same when deaths resulting from birth defects were excluded.
Table 4 examines racial disparity in neonatal mortality rate in 2006 by gestational age. The risk of neonatal mortality is 40% higher for non-Hispanic blacks compared with non-Hispanic whites at both 37–38 and 39–41 weeks of gestation (relative risk, 1.4; 95% CI 1.2–1.6). Hispanic neonatal mortality is equivalent to that of non-Hispanic whites for these intervals of gestation (relative risk, 1.0; 95% CI 0.9–1.2 at 37–38 weeks of gestation and relative risk, 1.0; 95% CI 0.9–1.1 at 39–41 weeks of gestation).
There is also racial disparity in postneonatal mortality rate by gestational age as seen in Table 4. The risk of postneonatal mortality is 80% higher among non-Hispanic blacks compared with non-Hispanic whites (relative risk, 1.8; 95% CI 1.6–2.0) at both 37–38 weeks of gestation and 39–41 weeks of gestation. Hispanics have a reduced risk of postneonatal mortality (relative risk, 0.8; 95% CI 0.7–0.9) at both 37–38 weeks of gestation and 39–41 weeks of gestation when compared with non-Hispanic whites.
When cause of neonatal mortality is examined from 2000–2006 in singleton gestations, the distribution of causes of death are relatively similar across races and ethnicities (Table 5). Birth defects (congenital malformations, deformations, and congenital anomalies) are the leading cause of neonatal death among all races and ethnicities between 37–41 weeks of gestation with the highest proportion among Hispanic neonatal deaths. The rate of neonatal death attributable to birth defects declines substantially between early term and full term for all races and ethnicities. For all races and ethnicities, intrauterine hypoxia, bacterial sepsis, and sudden infant death syndrome are in the top five causes of neonatal death. Accidents account for one of the top five causes of death for non-Hispanic whites and non-Hispanic blacks at 37–38 and 39–41; however, it is absent from the top five causes of neonatal death for Hispanics.
For early-term and full-term infants, birth defects are the leading cause of postneonatal death for Hispanics, whereas sudden infant death syndrome is the leading cause of postneonatal death for non-Hispanic whites and non-Hispanic blacks. Accidents are the third leading cause of postneonatal death regardless of gestational age or race and ethnicity. Assault accounts for one of the top five causes of postneonatal death in early- and full-term gestations for all racial and ethnic groups.
Using the most recently available U.S. period-linked birth and infant death data, we demonstrated that the period of term gestation (37–41 weeks of gestation) is more heterogeneous in mortality risk than previously recognized. Although births at 37 and 38 completed weeks are considered term, these early-term births are consistently associated with significantly higher neonatal and infant mortality rates when compared with births at 39 through 41 weeks of gestation over time.
When estimating the time trend, infant mortality for 37–41 completed weeks of gestation has decreased in the past decade across all race and ethnicities. However, the non-Hispanic black infant mortality rate has experienced the smallest decrease at 37 weeks of gestation when compared with improvements for non-Hispanic whites and Hispanics. This is because the neonatal mortality rate for non-Hispanic blacks at 37 weeks of gestation has not sustained any improvement with an increase of 15.8% over the past decade. In addition, the declines in mortality for infants born at 38 weeks of gestation were less for non-Hispanic black infants compared with the other groups.
Alexander and colleagues analyzed singleton live births from 1995–2000 linked birth infant death cohort files and found for infants born at less than 34 weeks of gestation, black had a survival advantage in infant mortality over whites. For births at 34–35 weeks of gestation, the infant mortality rates were equivalent. For infants born at 36 weeks or greater, the black and white disparity increased with increasing gestation through 41 weeks.9 In our study, non-Hispanic black infants were 40% more likely die in the neonatal period and 80% in postneonatal period in 2006 compared with whites if they were born at either early-term or full-term gestations. This is another reflection of the increasingly poor outcome over time for non-Hispanic black early-term infants relative to non-Hispanic whites. The stagnation in the rate of these early-term neonatal deaths requires further study to devise interventions to improve outcome.
The distribution of the leading five causes of neonatal death is relatively uniform across races and ethnicities with birth defects being the leading cause of death. However, the causes of postneonatal death vary by race and ethnicity. Accidents, assault, and sudden infant death syndrome require further targeting because these outcomes are amenable to intervention and are major contributors to the black and white disparity in postneonatal mortality in the United States.
The strength of the study is the use of a large sample size with recent data collected over one decade allowing for analysis of time trend and high power to detect differences in mortality rates. Although there may be concerns regarding the inaccuracies of gestational age estimates in birth certificates, they are less frequent for term births.10
Studies using vital statistics have tried to assess the independent risk of mortality for an infant born at 37 and 38 weeks compared with 40 weeks age after adjusting for maternal medical and pregnancy conditions, obstetric complications, and history of prior preterm birth.3,11 There appears to be no change in the increased risk after adjustment. However, the limitation of vital statistics is the quality and completeness of data on medical history and obstetric or fetal complications.12–14 Studies are needed in databases that have high-quality medical record information to be able to understand the contribution of complications to the increased risk for infant mortality among births in the early-term period and to the racial disparity in mortality rates. In particular, because births at 37 weeks have an exceptionally higher mortality rate, studies should be designed to examine the origin of the increased risk.
These data demonstrate that “term” pregnancy is not a period of uniform risk with early-term deliveries (37 and 38 weeks of gestation) experiencing higher neonatal, postneonatal, and infant mortality rates than full-term deliveries (39–41 weeks of gestation). Because 40 weeks of gestation has the lowest infant mortality rates across all race and ethnicities, it should be regarded as the optimal gestational age to use as a control group rather than analyzing infants born over the entire term period. In addition, although there have been improvements in overall neonatal, postneonatal, and infant mortality rates across the term period in the past decade, the unacceptable disparity in infant mortality remains for non-Hispanic blacks and must be targeted by intervention to decrease the mortality rate for this high-risk group.
1.Fleischman AR, Oinuma M, Clark SL. Rethinking the definition of ‘term pregnancy.' Obstet Gynecol 2010;116:136–9.
2.Reddy UM, Ko CW, Willinger M. ‘Early' term births (37–38 weeks) are associated with increased mortality. Am J Obstet Gynecol 2006;195:S202.
3.Zhang X, Kramer MS. Variations in mortality and morbidity by gestational age among infants born at term. J Pediatr 2009;154:358–62, 362.e1.
4.Young PC, Glasgow TS, Li X, Guest-Warnick G, Stoddard G. Mortality of late-preterm (near-term) newborns in Utah. Pediatrics 2007;119:e659–65.
5.Goldenberg RL, McClure EM, Bhattacharya A, Groat TD, Stahl PJ. Women's perceptions regarding the safety of births at various gestational ages. Obstet Gynecol 2009;114:1254–8.
6.Mathews TJ, MacDorman MF. Infant mortality statistics from the 2006 period linked birth/infant death data set. Natl Vital Stat Rep 2010;58:1–31.
7.Taffel A, Johnson D, Heuser R. A method of imputing length of gestation on birth certificates. Vital Health Stat 2 1982:1–11.
8.Anderson RN, Smith BL. Deaths: leading causes for 2002. Natl Vital Stat Rep 2005;53:1–89.
9.Alexander GR, Wingate MS, Bader D, Kogan MD. The increasing racial disparity in infant mortality rates: composition and contributors to recent US trends. Am J Obstet Gynecol 2008;198:51.e1–9.
10.Mustafa G, David RJ. Comparative accuracy of clinical estimate versus menstrual gestational age in computerized birth certificates. Public Health Rep 2001;116:15–21.
11.Donovan EF, Besl J, Paulson J, Rose B, Iams J. Infant death among Ohio residents born at 32 to 41 weeks of gestation. Am J Obstet Gynecol 2010;203:58.e1–5.
12.Piper JM, Mitchel EF Jr, Snowden M, Hall C, Adams M, Taylor P. Validation of 1989 Tennessee birth certificates using maternal and newborn hospital records. Am J Epidemiol 1993;137:758–68.
13.DiGiuseppe DL, Aron DC, Ranbom L, Harper DL, Rosenthal GE. Reliability of birth certificate data: a multi-hospital comparison to medical records information. Matern Child Health J 2002;6:169–79.
14.Lydon-Rochelle MT, Holt VL, Cardenas V, Nelson JC, Easterling TR, Gardella C, et al. The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certificates and in hospital discharge data. Am J Obstet Gynecol 2005;193: 125–34.
Figure. No caption available.