Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Temporal trends in preterm birth phenotypes by plurality: Black–White disparity over half a century



To examine trends in rates of preterm birth by race and plurality; to evaluate the association between race, plurality, and phenotypes of preterm birth.

Study design

Temporal trends analyses for preterm birth by race and plurality were performed for the years 1971–2018. Adjusted logistic regression models were utilized to evaluate the association between race, plurality, and phenotypes of preterm birth.


We observed that 1105,266 (0.7%), 1901,604 (1.2%), and 14,769,746 (9.3%) births belonged to extreme preterm, very preterm, and moderate-to-late preterm categories, respectively. We also observed that the risk of extreme preterm (RR: 2.69, 95% CI: 2.642–2.75) was highest for Black mothers as compared to White mothers.


Over the study period, preterm births disproportionately impacted Black mothers as well as pregnancies of higher plurality. With the persistence of racial disparities and growing trend of delayed childbearing and multiple pregnancies, targeted intervention is necessitated toward these vulnerable subgroups.


Preterm birth (PTB), defined as delivery before 37 weeks of gestation, is the leading cause of death in neonates and children under the age of 5 years worldwide [1]. The United States (US) is the only developed nation that ranks amongst the top-ten countries with the highest number of PTBs per year, spending ~$32.4 billion annually (2017 estimates) on PTB-related healthcare costs [2, 3]. PTB has continually risen for four consecutive years from 2015 to 2018 [4, 5].

While a substantial volume of research has been published on factors driving the upward trend of PTB in the US, it is difficult to draw consensus among studies because the etiologies of PTB are multifactorial and heterogeneous across gestational age [6, 7]. Major determinants of PTB include socio-demographic characteristics such as race and income, behavioral risk factors such as substance abuse, and maternal health factors such as preeclampsia, plurality, prior PTBs, and inadequate gestational weight gain [4, 8].

Racial disparity in PTB remains pronounced in the US despite decades of public health initiatives [9]. In 2018, PTB incidence was about 50% higher among Black (14%) than White women (9%) [4]. Recently, plurality (i.e., multiple births) has become a growing contributor of PTB in the US [10]. One hypothesis ties plurality to maternal age, as more mothers are using assisted reproductive methods to delay their pregnancies [11]. However, differential risk across orders of plurality has not been explored in larger scale temporal analyses [2, 12, 13].

To capture the evolving nature of PTB in the US and develop more comprehensive methods of obstetric risk evaluation, it is important to understand trends over long periods of time and capture nuances, such as the interplay between PTB risk factors and their heterogeneity across gestational age groups. To address this gap, we examined temporal trends for PTB, stratified by race and plurality in the US, spanning five decades i.e., 1971–2018. We also investigated race and plurality in association with different phenotypes of PTB.

Materials and methods

We conducted this population-based retrospective cohort study using the Natality data files, obtained from the National Bureau of Economic Research and compiled by the National Center for Health Statistics, and made publicly available by the US Centers for Disease Control and Prevention (CDC) [14]. The data set contained information on all births occurring within the US during the study period (1971–2018) and contained information, which were routinely abstracted from the birth certificates filed in the vital statistics offices of each state in the US. The data set contained information regarding socio-demographic and health characteristics of the child, parents’ demographics, and maternal risk factors associated with each birth. Because of the changes in birth certificates over the decades, the content of the database has also changed considerably. We included births within the gestational age of 20–42 weeks, and for a total of 48 years from 1971 to 2018. The study was approved as exempt by the Institutional Review Board of Baylor College of Medicine.

The variables, which were available for most of the study period, were included in the study. The information on mother’s age, race, parity, and plurality were available for all the years, from 1971 to 2018, whereas information on maternal gestational weight gain and comorbid conditions like diabetes, chronic hypertension, pregnancy-related hypertension, and eclampsia were available from 1989 to 2018. We used CDC’s recommendation for the definition of PTB, i.e., birth before 37 weeks of gestation after age of viability [4]. We identified three phenotypes of PTB based on the recommended classification in the literature [15]: extreme preterm: <28 weeks, very preterm: 28–31 weeks, moderate-to-late preterm: 32–36 weeks, 6 days. Term births were deliveries occurring at or after 37 weeks of gestation. We re-categorized the study variables as follows—mother’s age was subdivided as: under 15 years; 15–24 years; 25–34 years; 35–44 years; and ≥45 years; maternal race was categorized as White, Black, or other (which included all races other than White and Black, including unknown and missing); plurality of the child was classified as singleton, twin, and higher-order multiple birth (triplets, quadruplets, etc.); parity (obtained from the total birth order of the current child irrespective of whether the previous births were live or stillbirth; and whether or not any of the previous children were alive at the time of the birth of the current child) was categorized as: 1–2, 3–5, and ≥6 children, while maternal gestational weight gain was grouped as: up to 20, 21–30, 31–40, and over 40 lb. Maternal pregnancy complications that were considered included diabetes mellitus, chronic and gestational hypertension, and eclampsia.

We stratified the preterm and term births during the study period based on the birth characteristics, and calculated the incidence of PTB among each of the socio-demographic and maternal risk factors. We employed Pearson’s chi-squared test to determine the association between the birth types (preterm vs. term) and each of the maternal and infant characteristics. Next, using Joinpoint regression technique [16], we calculated the average annual percentage change (AAPC) in the rates of singleton, twin, and multiple PTB in the US by maternal race. Joinpoint regression is a novel modeling strategy that detects the change in the rate of outcomes (in this case, PTB) over time. The final model containing an optimal number of joinpoints generated the AAPC and 95% confidence interval (CI) of the rates of PTB among singletons, twin, and higher-order multiples by race over the study period.

We also conducted bivariate analyses to assess the relationship between birth characteristics and incidence of the three preterm phenotypes. Last, we generated log binomial regression models to examine the association between the various socio-demographic and gestational factors (as exposure) and PTB and each of its phenotypes (as outcomes), after removing records with race categorized as “others.” This step was performed to compare the likelihood of PTB among Black mothers to that of White mothers; and because the race category “other” was not well-defined in the data set. We conducted sensitivity analysis to examine the impact of race and plurality on PTB and on each phenotype of PTB, with and without considering maternal weight gain and comorbidities. As the results from the two models were very similar, we report the results from the model containing maternal weight gain and comorbidities, along with our socio-demographic characteristics as covariates. All tests of hypotheses were two-tailed with a type 1 error rate at 5%. All statistical analyses were performed using R version 3.5.1 (R is an open source software initially written by Robert Gentleman and Ross Ihaka of the Statistics Department of the University of Auckland, New Zealand) and Joinpoint Regression Program (National Cancer Institute, Bethesda, MD).


There were a total of 159,080,368 births in the US during the study period, of which 17,776,616 (11.17%) were PTB. Table 1 shows the incidence of PTB by birth characteristics. The incidence of PTB among various maternal age groups was “U”-shaped, with the highest incidence being among older mothers, followed by teenagers. Black mothers had almost twice the incidence of PTB as compared to White mothers (17.6% vs. 9.6%). The incidence of PTB was five times higher among twin births and ten times higher among higher-order multiples, as compared to singleton births. There was a positive dose–response relationship between parity and incidence of PTB. We observed a negative dose–response between maternal pregnancy weight gain and incidence of PTB when missing information was excluded from the analyses. The incidence of PTB was higher among persons with diabetes, chronic and pregnancy hypertension, and eclampsia as compared to those without these conditions.

Table 1 Socio-demographic characteristics of mothers who had preterm and term births in the US—1971–2018.

Figure 1 shows the trends in the rates of singleton, twin, and multiple PTB stratified by maternal race during the study period. The rates of overall singleton PTB birth were 5.3% at the beginning of the study period (4.5% for White and 9.1% for Black mothers), which then increased to 10.0% by 2018 (8.7% of White and 14.6% for Black mothers). We observed that there was a slight increment in the rate of singleton PTB among White as compared to Black mothers (AAPC: 1.4 vs. 1.0, respectively) over the study period. Among twin PTB, overall, there was an increase from 23.7 (in 1971) to 56.2% (in 2018), whereas among White mothers, it went from 23.3% in 1971 to 55.5% in 2018; and in Black mothers the rate went from 24.7 to 61.1% over the study duration. We observed that there was a similar trend of increment (AAPC: 1.9) in the rates of twin PTB, overall and among White and Black mothers during the study period. For higher-order multiples, we observed an overall increase of 1.9% (95% CI: 1.5, 2.3) in PTB rates in the population, which was similar to the increase among White mothers (AAPC: 1.8, 95% CI: 1.3, 2.3). However, among Black women, the incidence of PTB continued to rise (from 24.7% in 1971 to 61.1% in 2018), corresponding to an AAPC of 2.5% (95% CI: 2.1, 3.0) during the study period.

Fig. 1: Trends in incidence of preterm birth by plurality and race—1971–2018.
figure 1

Left panel: Trends in incidence of preterm birth among singletons by race. Middle panel: Trends in incidence of preterm birth among twins by race. Right panel: Trends in incidence of preterm birth among higher order multiples by race. An asterisk symbol denotes statistically significant change in trend. AAPC average annual percentage change.

We observed that 1105,266 (0.7%), 1901,604 (1.2%), and 14,769,746 (9.3%) births belonged to the extreme preterm, very preterm, and moderate-to-late preterm category, respectively (Table 2). The incidence of extreme and very PTB were highest in mothers <15 years old, whereas incidence of moderate-to-late PTB was highest in the ≥45 years age group. Black mothers consistently had the highest incidence of every phenotype of PTB. With increasing plurality, we observed an increment in the incidence of each phenotype of PTB. Incidence of each of the phenotypes of PTB was highest among women with six or more children. Women who gained only up to 20 lb. during their pregnancy had the highest incidence of each PTB phenotype. Among mothers with diabetes, the incidence of extreme PTB was lower than those without diabetes, whereas the incidence of very and moderate-to-late PTB was higher than those without the disease. Among women with pre-pregnancy hypertension, gestational hypertension, and eclampsia, the incidence of extreme, very, and moderate-to-late PTB was higher than among their counterparts without these diseases.

Table 2 Socio-demographic characteristics of mothers who had preterm and term births in the US—1971–2018.

Table 3 shows the adjusted association between various socio-demographic characteristics and PTB and its phenotypes. Black women were 1.66 times as likely to experience any PTB (95% CI: 1.65–1.67). We also observed that the risk of extreme preterm (RR: 2.69, 95% CI: 2.642–2.75) was highest for Black as compared to White mothers, followed by the risk of very preterm (RR: 2.02, 95% CI: 1.99–2.05) and moderate-to-late PTB (RR: 1.54, 95% CI = 1.53–1.55). As compared to singleton births, twin births were more likely to be extreme, very and moderate-to-late preterm. Higher-order multiples also had exponentially high likelihood of being preterm, as compared to singletons. Mothers under the age of 15 years were most likely to experience each phenotype of PTB. Mothers with six or more children had 36% increased likelihood of experiencing PTB (RR: 1.36, 95% CI: 1.35–1.38). Maternal comorbid conditions were associated with increased likelihood of PTB, with the highest risk being among those with eclampsia (RR: 3.96, 95% CI: 3.83–4.09), as compared to those without the disease.

Table 3 Log binomial regression model to show the association between various socio-demographic and comorbid characteristics and preterm birth (1989–2017).


Over the study period 1971–2018, PTBs constituted 11.17% of all births in the US and disproportionately impacted Black mothers as well as pregnancies of higher plurality. Our results show about twofold PTB incidence for Black compared to White women (17.6% vs. 9.6%), for the cumulative 1971–2018 period, roughly the same magnitude as estimated for 2018 alone (CDC) [14]. This emphasizes both the magnitude and long-standing persistence of racial disparity in the burden of PTB [9]. We then differentiated magnitude of risk within plurality by creating subgroups based on fetal count. Our data show that PTB incidence is about five times greater for twin pregnancies and about nine times greater for multiple pregnancies, compared to singleton pregnancies. The uterine overdistension and cervical shortening associated with carrying multiple fetuses is thought to be responsible for triggering early labor [11]. The distinction between twin and multiple pregnancies, notably, is important to make given that studies that exclude multiple pregnancies from plurality measurements tend to overestimate the contribution of iatrogenic PTBs over other etiologies [2, 11]. We also found evidence of growing overlap between race and plurality over time; the overall rates of singleton and twin PTBs over five decades was higher among Black compared to White mothers. There was a drastic increase in the PTB rates in all plurality and racial groups, but this trend has either declined or stabilized in the last decade. Although the exact reason for this unknown, it could be attributed to enforcement of controlled and stricter guidelines for multiple embryo transfer following in vitro fertilization and other infertility treatments, which usually result in multiple pregnancies [17, 18].

When examining maternal age stratified by race and plurality, our data indicated older mothers and teenage mothers to be most susceptible. PTB in older mothers, a growing concern, is largely considered a result of higher plurality rates following use of assisted reproductive techniques to delay childbearing, but there is mixed evidence whether race influences this relationship [10, 11, 19]. In teenage mothers, however, increased PTB risk is thought to coincide with both racial disparity and higher odds of plurality [20,21,22]. The mechanism of how these variables operate in conjunction to modify PTB likelihood is not well understood.

All tested maternal comorbidities (maternal eclampsia, diabetes, and hypertension), increased PTB risk, and predominantly affected Black women, with eclampsia producing the strongest association [23]. Not only are Black women at higher risk for developing comorbidities, they also tend to receive lower quality or delayed prenatal care [24, 25]. A dose-dependent association was found between inadequate gestational weight gain and PTB risk. Based on new evidence, the relationship between race and gestational weight gain may be complicated by pre-pregnancy weight class [26].

We found a positive dose-dependent relationship between increasing parity and PTB, such that mothers with six or more children had about 36% greater likelihood of experiencing PTB. This parallels the dose-dependent pattern reported by a study in the Netherlands, one of the few to control for confounding by socioeconomic status, although the authors did not analyze multiparity greater than five births [27].

We considered heterogeneity of PTB risk profiles across gestational age groups [6, 7], race, maternal age, and gestational weight gain demonstrated consistent patterns across each phenotype, with Black women, mothers under 15 years old, and mothers who gained only up to 20 lb., experiencing highest PTB incidence. Meanwhile, high multiparity, pregnancy hypertension, and eclampsia were the most dominant predictor of PTB in extreme and very preterm phenotypes. Whereas, having diabetes had a protective effect on extreme and very PTB; although it was a risk factor for experiencing moderate-to-late PTB.

There are some limitations pertaining to the large scope of our study that warrant discussion. Race categorization was limited by the data available for the desired study period; ethnicity information was not available for the entire five decades but only since 1989, which prevented us from distinguishing Non-Hispanic and Hispanic White women. Maternal pre-pregnancy BMI and other comorbidities like gestational and pre-pregnancy diabetes were not included because the information was not available for all the years. We understand that maternal weight gain during pregnancy is hugely impacted by their pre-pregnancy BMI, but because the information on BMI was not available, we could not adjust for these factors in our study.

Despite these shortcomings, our study has some noteworthy strengths. This study is unique in that it analyses PTB over five decades, the largest time-frame examined thus far to our knowledge. It is the first to differentiate between twin and multiple pregnancies in a longitudinal analysis to provide a more nuanced understanding of plurality as a PTB risk factor. Moreover, our study captures cross-interactions among variables, which may help reconcile some incongruities in previous literature caused by limitations such as short time-frames, only examining a subset of possible risk factors in isolation, or not accounting for heterogeneity among gestational age groups. With the persistence of racial disparities and growing trend of delayed childbearing and multiple pregnancies, updated models of obstetric risk assessment must account for possible risk modifiers to identify patients most vulnerable for PTB.

Code availability

The codes will be made available upon request.


  1. Harrison MS, Goldenberg RL. Global burden of prematurity. Semin Fetal Neonatal Med. 2016;21:74–9.

    Article  PubMed  Google Scholar 

  2. Purisch SE, Gyamfi-Bannerman C. Epidemiology of preterm birth. Semin Perinatol. 2017;41:387–91.

    Article  PubMed  Google Scholar 

  3. WHO. Preterm birth: key facts. World Health Organization. 2018.

  4. Centers for Disease Control and Prevention. Preterm birth | Maternal and infant health | Reproductive health | CDC. Centers for Disease Control and Prevention. 2018.

  5. Martin JA, Osterman MJK. Describing the Increase in preterm births in the United States, 2014–2016. NCHS Data Brief. 2018;312:1–8.

    Google Scholar 

  6. Frey HA, Klebanoff MA. The epidemiology, etiology, and costs of preterm birth. Semin Fetal Neonatal Med. 2016;21:68–73.

    Article  PubMed  Google Scholar 

  7. Esplin MS. Overview of spontaneous preterm birth: a complex and multifactorial phenotype. Clin Obstet Gynecol. 2014;57:518–30.

    Article  PubMed  Google Scholar 

  8. Johnson JL, Farr SL, Dietz PM, Sharma AJ, Barfield WD, Robbins CL. Trends in gestational weight gain: the pregnancy risk assessment monitoring system, 2000–2009. Am J Obstet Gynecol. 2015;212:806.e1–806.e8.

    Article  Google Scholar 

  9. Burris HH, Lorch SA, Kirpalani H, Pursley DWM, Elovitz MA, Clougherty JE. Racial disparities in preterm birth in USA: a biosensor of physical and social environmental exposures. Arch Dis Child. 2019;104:931–5.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Mathews TJ, Hamilton BE. Mean age of mothers is on the rise: United States, 2000–2014. NCHS Data Brief. 2016;1–8.

  11. Luke B. Pregnancy and birth outcomes in couples with infertility with and without assisted reproductive technology: with an emphasis on US population-based studies. Am J Obstet Gynecol. 2017;217:270–81.

    Article  PubMed  Google Scholar 

  12. Hiersch L, Rosen H, Okby R, Freeman H, Barrett J, Melamed N. The greater risk of preterm birth in triplets is mirrored by a more rapid cervical shortening along gestation. Am J Obstet Gynecol. 2016;215:357.e1–357.e6.

    Article  Google Scholar 

  13. Gyamfi-Bannerman C, Ananth CV. Trends in spontaneous and indicated preterm delivery among singleton gestations in the United States, 2005–2012. Obstet Gynecol. 2014;124:1069–74.

  14. Roth J. NCHS’ vital statistics natality birth data. The National Bureau of Economic Research. 2018. Accessed 19 Jun 2020.

  15. Dongarwar D, Aggarwal A, Barning K, Salihu HM. Trends in Stillbirths and Stillbirth phenotypes in the United States: an analysis of 131.5 million births. Int J Matern Child Heal Aids. 2020;9:146–8.

    Article  Google Scholar 

  16. Joinpoint Regression Program. 2020. Accessed 9 Oct 2020.

  17. Klitzman R. Deciding how many embryos to transfer: ongoing challenges and dilemmas. Reprod Biomed Soc Online. 2016;3:1–15.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Kissin DM, Kulkarni AD, Kushnir VA, Jamieson DJ, National ART Surveillance System Group. Number of embryos transferred after in vitro fertilization and good perinatal outcome. Obstet Gynecol. 2014;123:239–47.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Schummers L, Hacker MR, Williams PL, Hutcheon JA, Vanderweele TJ, McElrath T, et al. Variation in relationships between maternal age at first birth and pregnancy outcomes by maternal race: a population-based cohort study in the United States. BMJ Open. 2019.

  20. Balassone MLou. Multiple pregnancies among adolescents: incidence and correlates. Health Soc Work. 1988;13:266–76.

    CAS  Article  PubMed  Google Scholar 

  21. Centers for Disease Control and Prevention. About teen pregnancy | Teen pregnancy | Reproductive health | CDC. Centers for Disease Control and Prevention. 2016.

  22. Leland NL, Petersen DJ, Braddock M, Alexander GR. Variations in pregnancy outcomes by race among 10–14-year-old mothers in the United States. Public Health Rep. 1995;110:53–8.

    CAS  PubMed  PubMed Central  Google Scholar 

  23. Salihu HM, Luke S, Alio AP, Deutsch A, Marty PJ. The impact of obesity on spontaneous and medically indicated preterm birth among adolescent mothers. Arch Gynecol Obstet. 2010;282:127–34.

    Article  PubMed  Google Scholar 

  24. Metcalfe A, Wick J, Ronksley P. Racial disparities in comorbidity and severe maternal morbidity/mortality in the United States: an analysis of temporal trends. Acta Obstet Gynecol Scand. 2018.

  25. Howell EA, Egorova NN, Balb.ierz A, Zeitlin J, Hebert PL. Site of delivery contribution to black-white severe maternal morbidity disparity. Am J Obstet Gynecol. 2016;215:143–52.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Headen I, Mujahid MS, Cohen AK, Rehkopf DH, Abrams B. Racial/ethnic disparities in inadequate gestational weight gain differ by pre-pregnancy weight. Matern Child Health J. 2015;19:1672–86.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Koullali B, Van Zijl MD, Kazemier BM, et al. The association between parity and spontaneous preterm birth: a population based study. BMC Pregnancy Childbirth. 2020;20:233.

    Article  PubMed  PubMed Central  Google Scholar 

Download references


The corresponding author had full access to all the data and had final responsibility for the decision to submit for publication.

Author information

Authors and Affiliations



DD and HMS designed, planned, and conducted the study. DD and LW analyzed the data. All authors contributed to manuscript writing, revision, and approval of the final manuscript.

Corresponding author

Correspondence to Deepa Dongarwar.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Dongarwar, D., Tahseen, D., Wang, L. et al. Temporal trends in preterm birth phenotypes by plurality: Black–White disparity over half a century. J Perinatol 41, 204–211 (2021).

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

Further reading


Quick links