If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Pediatrics, Feinberg School of Medicine, Northwestern University, Chicago, ILDepartment of Preventive Medicine, Feinberg School of Medicine, Northwestern University, Chicago, ILInstitute of Healthcare Studies, Feinberg School of Medicine, Northwestern University, Chicago, IL
Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Feinberg School of Medicine, Northwestern University, Chicago, ILInstitute of Healthcare Studies, Feinberg School of Medicine, Northwestern University, Chicago, IL
The objective of the study was to determine the trends and racial differences in late preterm induction (LPI) of labor in the United States.
Data from the National Vital Statistics System were used to identify women eligible for induction between 34 and 42 weeks' gestation from 1991 to 2006. Annual LPI rates were calculated, and maternal race/ethnicity was classified into 4 groups. Changes in the frequency and odds of LPI, stratified by race/ethnicity, were assessed using logistic regression.
Among the 42.0 million eligible women, LPI rates increased from 0.46% to 1.37% (P < .01) over 16 years. LPI rates were highest for black women (P < .01) each year, and after adjusting for confounding factors, the odds of LPI were highest (P < .01) and rose most rapidly (P < .01) for black women (non-Hispanic white: odds ratio [OR], 1 [referent]; Hispanic white: OR, 0.76; black: OR, 1.31; other: OR, 0.81; P < .01).
LPI rates were persistently highest and rose most rapidly for black women.
have demonstrated that, over a recent 13 year period, the mean gestational age at birth among singletons decreased by 1 week in the United States. This decrease has been due not only to women delivering earlier at term gestations
however, the specific reasons that have contributed to the rise in preterm induction, and the resulting LPD rates remain largely unmeasured and uncertain.
Recent evidence suggests that nonmedical factors may be associated with the increase in these inductions, and the existence of a racial/ethnic disparity of labor induction supports the concept of nonmedical contributing factors.
For example, non-Hispanic white (NHW) women have been most likely to be induced at term gestation but prior to 39 weeks despite the fact that they are less likely to have medical comorbidities.
However, the extent to which there is a racial/ethnic disparity among women induced at preterm gestations remains unknown. Thus, the aims of this study were to: (1) determine national trends of late preterm induction of labor (LPI) and (2) to estimate the independent association between maternal race/ethnicity and LPI during a recent 16 year period in the United States.
Materials and Methods
Birth certificate data from the National Vital Statistics System (NVSS) were used to identify all women delivering in the United States from 1991 through 2006. Women who delivered a singleton between 340/7 and 426/7 weeks' gestation were included in the analysis. As shown in Figure 1, women with prior cesarean deliveries or with a nonvertex fetal presentation were not included because these circumstances often have been considered to be contraindications to labor induction. Those with fetal anomalies or no prenatal care (because induction of labor is generally planned during antenatal care) were excluded as well.
Also omitted, were those women whose records precluded an accurate determination of either race/ethnicity, gestational age at delivery, or whether labor induction occurred, given that the selected exposure (race/ethnicity) or primary outcome (LPI) could not be determined without these data. Lastly, women coded as having pregnancy-induced hypertension (PIH) or with premature rupture of membranes (pROM) were not considered in these analyses. The timing of onset or the severity of these conditions was unknown, which precluded the ability to control for the presence/severity of either PIH or pROM within the late preterm period. Thus, women with either PIH or pROM were all excluded to reduce the possibility of confounding bias.
LPD was defined as delivery of a live infant between 340/7 and 366/7 weeks' gestation. LPI was defined as induction of labor within this late preterm period. Induced women were identified in the NVSS dataset as women had “induction of labor,” defined in the dataset as “initiation of uterine contractions before the spontaneous onset of labor by medical and/or surgical means for the purpose of delivery.”
Women who had LPD via a cesarean delivery were not considered because the presence or onset of labor prior to their delivery could not be ascertained.
Gestational age was determined from the variable of gestational age at birth. Although some have suggested that the best clinical estimate of gestation may be a more valid variable, this variable was not collected by the state of California, which is also the state with the most annual births. Thus, to maximize the sample size and the representativeness of the national gravid population, the gestational age variable was used in the primary analysis. To ensure that the selection of the gestational age variable did not significantly affect the results, an additional sensitivity analysis that utilized the alternative variable of best clinical estimate of gestation was performed.
LPI rates were calculated as the sum of 3 week-specific induction rates within the late preterm period (ie, Σ [inductions at 34 weeks/eligible women from 34 to 42 weeks] + [inductions at 35 weeks/eligible women from 35 to 42 weeks] + [inductions at 36 weeks/eligible women from 36 to 42 weeks]). Annual and race-stratified rates of LPI are reported.
Race/ethnicity was categorized into 4 groups: NHW, Hispanic white (HW), black, and other. The black and other groups were not further subdivided by Hispanic ethnicity, given the small proportion of Hispanic women in each of these groups (3.0% and 3.3%, respectively).
Bivariable analyses were performed, using analysis of variance (ANOVA) and χ2 to determine the demographic and medical characteristics that were significantly associated with maternal race/ethnicity. Demographic factors included advanced maternal age (≥35 years old), teen pregnancy (age 13-19 y), nulliparity (yes/no), smoking status during pregnancy (yes/no), alcohol use during pregnancy (yes/no), and marital status (yes/no). The data were further stratified by medical factors (maternal diabetes mellitus [DM] or chronic hypertension [CHTN]) that are associated with labor induction and that were consistently measured from 1991 through 2006 in the birth certificate registry.
Multivariable logistic regression was used to estimate the independent association between maternal race/ethnicity and LPI (yes/no). Demographic and obstetric variables, evaluated in the regression analysis, were kept in the model if inclusion changed the estimated odds ratio of the association between race/ethnicity and LPI by at least 10%. Maternal smoking and alcohol use during pregnancy were not included in the models as potential confounding factors because of the large amount of missing data. The models also included an interaction term between maternal race/ethnicity and the calendar year of birth to estimate whether the difference in the odds of LPI by racial/ethnic group changed significantly over the 16 year study period.
All statistical tests were 2 tailed and, because of the large sample size, α = 0.01 was used to define statistical significance. Analyses were performed with STATA version 10.1 (StataCorp, College Station, TX). Birth certificate data were accessed from the National Vital Statistics System web site in April 2010. Because these data are publicly available and deidentified, the Children's Memorial Research Center Institutional Review Board exempted this study from review.
From 1991 to 2006, 63,120,211 live-born infants delivered in the United States. Among all births, 8.3% of newborns delivered during the late preterm period, and LPD increased during the 16 year study period (7.6-9.2%, an increase of 22%). After restricting the population of gravid women to those who met the inclusion criteria (Figure 1), 42,011,982 women (66.6%) were included in the analysis.
LPD and LPI occurred in 6.8% and 0.95% of eligible women, respectively, and both increased significantly over the 16 year period (LPD: 6.3-7.5%, an increase of 19%; LPI: 0.46-1.37%, an increase of 195%; P < .01). Thus, 76% of the increase in LPD was attributable to the increase in LPI. As shown in Figure 2, this increase in LPI occurred at each week (P < .01) within the late preterm period, with the majority of LPI occurring at 36 weeks' gestation (34 weeks: 15%; 35 weeks: 28%; 36 weeks: 57%) during the study period.
Table 1 illustrates the multiple differences in characteristics of the population by maternal race/ethnicity. Black women were most likely to be young or unmarried, NHW women were most likely to use tobacco or alcohol during pregnancy (although all groups had large and significantly different proportions of missing data about antenatal tobacco and alcohol exposure), and HW women were least likely to be nulliparous.
TABLE 1Maternal characteristics stratified by race/ethnicity, 1991–2006
Eligible black women had the highest rates of LPD (black: 10.2%; NHW: 5.8%; HW: 7.3%; other: 6.8%; ANOVA, P < .01). Trends for both LPD and for the proportion of these women who were induced within each racial/ethnic group are illustrated in Table 2. Black women underwent LPI most frequently (black: 1.3% vs NHW: 0.9%; HW, 0.8; other: 0.7%; P < .01), and correspondingly, the unadjusted odds ratio (OR) for LPI was highest for black women (OR, 1.36; 99% confidence interval [CI], 1.34–1.37; P < .01) and lowest for HW and women from other race/ethnicity (HW, OR: 0.80; 99% CI, 0.79–0.81; P < .01; other: OR, 0.77; 99% CI, 0.76–0.79; P < .01) compared with NHW women. Longitudinal trends in LPI, stratified by maternal race/ethnicity, are depicted in Figure 3. Annual LPI rates more than doubled for all 4 racial/ethnic groups (P < .01) during the 16 year study period.
TABLE 2Trends in LPD stratified by maternal race/ethnicity, 1991–2006
Multivariable logistic regression, which incorporated adjustments both for time trends and for multiple covariates (advanced maternal age, marital status, and presence of DM or CHTN), demonstrated that the odds to receive LPI remained significantly highest for black women (NHW: OR 1 (referent); HW: OR, 0.76; 99% CI, 0.72–0.81; black: OR, 1.31; 99% CI, 1.25–1.37; other: OR, 0.81; 99% CI, 0.73–0.89; P < .01) throughout the 16 year study period. When the sensitivity analysis with a change in the gestational age variable was conducted, the findings were similar to the base case analysis (data not shown).
Two and one half percent and 0.7% of eligible women had DM and CHTN, respectively. Women with DM had an LPI rate of 2.4%, whereas women with CHTN had an LPI rate of 6.6%. Both rates were significantly greater than the observed rate (0.9%, P < .01) for women without either DM or CHTN. Women with either DM or CHTN also exhibited a significant rise in LPI over the 16 year study period (P < .01 for DM or CHTN); however, the increase was significantly less than for women without DM or CHTN (P < .01). Similar trends were present for women of all racial/ethnic groups, as illustrated in Table 3.
TABLE 3Trends for late preterm induction stratified by maternal race/ethnicity and the presence of medical risk factors
Data are presented as the percentage of women affected (P < .01 for the difference in the increase in LPI by the presence of antenatal DM or CHTN for each racial/ethnic group during the 16 year period).
Murthy. Late preterm induction. Am J Obstet Gynecol 2011.
a Data are presented as the percentage of women affected (P < .01 for the difference in the increase in LPI by the presence of antenatal DM or CHTN for each racial/ethnic group during the 16 year period).
LPI rates in the United States have risen and nearly tripled over a recent 16 year period. Black women receive LPI most frequently compared with women from other racial/ethnic groups, even after controlling for potential confounding factors. Moreover, the rise in LPI cannot be attributed to the presence of an obstetric condition such as PIH because women with this condition were excluded from the analysis.
Although other studies have demonstrated racial/ethnic differences in rates of induction, these studies have typically been cross-sectional in design and/or have not adjusted for patient characteristics that may confound the observed associations.
In contrast, this analysis has included both potential confounding characteristics and incorporates a longitudinal analytic design. In addition, whereas other studies largely have focused on term inductions, this analysis focused on women who are induced at late preterm gestations.
This time period in pregnancy is one that has drawn increasing interest, given the contribution of LPD to the rise in preterm births as well as the recent data documenting the associated short- and long-term childhood morbidities with delivery at this gestational age.
These data suggest that the threshold for initiating LPI may have decreased in the United States. The rise in LPI persisted in the multivariable equation, even after controlling for confounding risk factors. Also, the rates of LPI rose both for women with clear medical comorbidities such as DM or CHTN as well as for women without these antenatal conditions. These results do not preclude the possibility that women with DM or CHTN may have begun to develop more severe disease over the duration of the study and thus been induced because of their condition, even with the clinical threshold for initiating LPI remaining static.
Analysis of the data by the week-specific rates of labor induction within the late preterm period demonstrates that the majority of LPI occurs because of inductions at 36 weeks' gestation. However, a significant minority of LPI (43%) is accounted for by labor induction at 34 and 35 weeks' gestation, and infants born at these gestations are more likely to incur more short-term neonatal sequelae relative to more mature newborns, even those newborns that are born at 36 weeks.
Strategies to reduce LPD through a decrease in LPI would be most effective with a greater understanding of both the indications for and the populations of women who received induction at each of these 3 weeks in pregnancy.
In addition, the consideration of preterm labor induction and the associated morbidities cannot be considered in isolation and must be balanced against the risk of stillbirth and, more globally, perinatal mortality. For at-risk fetuses in the late preterm period, published reports define the risk of stillbirth and early neonatal death at approximately 0.02-0.04%
respectively. The key dilemma comes in balancing the risks and benefits of expectant management vs preterm delivery for women at risk. This study quantifies the magnitude of the effect regarding how clinicians, over time, have chosen to balance these risks and increasingly initiated LPI.
A similar analysis has been completed for early-term (37-38 weeks' gestation) induction of labor over the same study period.
The results from both of these studies suggest that an inflection point, between the 36th and the 37th week of pregnancy, exists with regard to the racial/ethnic differences in labor induction. Between the late preterm and the early-term periods of pregnancy, the differences in labor induction by race/ethnicity change such that black women, who were most likely to undergo LPI, become less likely during the early-term period to receive labor induction compared with NHW women.
This discontinuity, at the point at which gestational age, which is a continuous variable, has a superimposed conceptual dichotomy (ie, preterm vs term) that deserves further study and suggests the interplay of the medical and nonmedical factors that affect whether a labor induction is undertaken.
In the context of preventing stillbirth mortality, this reversal in the disparity in labor induction also is significant and of interest. In the late preterm period, black women are approximately 60% more likely to endure stillbirth compared with NHW women,
and in that regard, the present results support the concept that one reason that labor inductions are performed is to avoid stillbirth. Yet after the late preterm period, NHW women receive induction more frequently,
even though black women continue to remain at highest risk for stillbirth. Further study is warranted to further explore the mechanisms that underlie these apparently discrepant trends with respect to the interactions between maternal race/ethnicity, LPI, and stillbirth to determine the optimal criteria for the selection of women who are at risk for having a stillbirth and who will avert this outcome by receiving LPI.
Several limitations of this study should be noted. Maternal race/ethnicity was classified into only 4 categories and does not adequately represent the full racial and ethnic diversity of the US gravid population. This is especially true for the classification of women as being of other race/ethnicity, a category that encompasses women from multiple different backgrounds.
As in any analysis of secondary data, unknown and unmeasured factors (eg, fetal lung maturity, fetal growth abnormalities, indication for labor induction) could have modified the observed associations. Moreover, coding error is always a concern, such that women may have been misclassified as to whether they received an induction.
the Centers for Disease Control and Prevention and numerous previous investigators have used these data to examine national trends in labor induction for both mothers and infants. These data sets remain the standard for epidemiologic analyses of US births.
This study cannot define all of the factors that may account for the rise in LPI or for the origins in the racial/ethnic differences that have persisted. Whether the racial/ethnic differences in labor induction are due to differences within or between institutions, variation across providers, patient choice (eg, selections of providers and/or birth hospitals), or combinations remains uncertain. Most importantly, the degrees to which these reported differences are associated with maternal and perinatal outcomes also are not known.
Further research will need to investigate the underlying variation within and across institutions as well as among obstetric providers and patients/families that have led to the increase and the persistent racial/ethnic differences in LPI in the United States.
Changes in the gestational age distribution among US singleton births: impact on rates of late preterm birth, 1992 to 2002.