The objective of this study is to determine the prevalence of placenta previa among different racial and ethnic groups.
We conducted a retrospective cohort study to examine the prevalence of placenta previa among five major racial and ethnic groups: African American, Asian, Caucasian, Hispanic and Native American. We included all deliveries 20 weeks gestation from a large northern Californian Health Maintenance Organization from 1995–2006. A multivariable logistic regression model was used to control for potential confounders.
Of the 394 083 deliveries in our cohort, 1580 (0.40%) were complicated by placenta previa. The prevalence of placenta previa was: Asian 0.64%, Native American 0.60%, African American 0.44%, Caucasian 0.36%, Hispanic 0.34% and unknown 0.31% (P<0.001). In our multivariable logistic regression model, only Asians (odds ratio (OR) 1.73, 95% confidence intervals (CI) 1.53–1.95) and African Americans (OR 1.43, 95% CI 1.19–1.72) were at increased risk for having placenta previa, compared with Caucasians.
Asian women have the highest prevalence of placenta previa.
Placenta previa, in which the placenta completely or partially covers the cervix, complicates approximately four out of every 1000 pregnancies.1 The presence of a placenta previa requires a cesarean delivery and is potentially life-threatening, with significantly associated maternal and neonatal morbidity related to maternal hemorrhage and preterm birth.2, 3, 4, 5, 6 It also increases the risk of placenta accreta, especially in the setting of prior cesarean delivery.7
Known risk factors for placenta previa include advancing maternal age, multiparity, assisted reproductive technology, multiple gestation, prior cesarean delivery, prior abortion and tobacco use.8, 9, 10, 11, 12, 13 However, there is conflicting data regarding whether specific racial or ethnic groups are associated with an increased risk for placenta previa and the biologic plausibility of such an association is unclear. Several studies have found an increased risk of placenta previa among Asian women compared with Caucasian women,14, 15, 16 and among African-American and/or other minority women pooled into a mixed group than for Caucasian women.2, 8, 13, 16, 17, 18 In contrast, other studies failed to show a relationship between maternal race and placenta previa.8, 9, 11, 19, 20, 21, 22, 23 The two largest studies to date on race and placenta previa were population-based retrospective cohort studies with data derived from US national databases, and although both agreed that Asian women had a higher risk of placenta previa compared with Caucasian women, they reached opposite conclusions regarding African-American women.15, 16 (See Appendix 1, Supplementary Information, for a summary of the literature.)
Given these discrepancies in the literature regarding the relationship between maternal race and placenta previa, we sought to determine the prevalence of placenta previa stratified by race and ethnicity, using a large northern Californian Health Maintenance Organization birth cohort from 1995 to 2006, and to identify whether any of the five major racial and ethnic groups studied (African American, Asian, Caucasian, Hispanic and Native American) were associated with an increased risk of placenta previa.
We conducted a retrospective cohort study of all women who delivered from 1995 to 2006 at Kaiser Permanente of northern California, a large Health Maintenance Organization. We included all deliveries 20 weeks gestation and compared the characteristics of women with pregnancies complicated by placenta previa to those without placenta previa. The Kaiser Permanente Institutional Review Board approved this study. We analyzed all data using Stata statistical software version 7.1 (StataCorp, College Station, TX, USA).
We defined placenta previa as the complete or partial covering of the cervix by the placenta. First, we identified women with placenta previa by ICD-9 code and included them in the study if they had a cesarean delivery. Second, a board certified obstetrician, who was blinded to the race and ethnicity of the patients, personally confirmed the diagnosis of placenta previa by individual chart review. We considered the diagnosis of placenta previa to be correct if there was a follow-up third trimester ultrasound showing persistent placenta previa and/or written documentation in the hospital chart confirming the presence of placenta previa at the time of cesarean delivery. In order to confirm that Kaiser Permanente did not underreport the diagnosis of placenta previa, we also conducted a chart review of a random selection of 240 cesarean deliveries that did not have an ICD-9 code for placenta previa.
Patients self-reported their racial and ethnic identification and we abstracted the data from the Kaiser Permanente Northern California electronic records. We categorized race and ethnicity into five major groups: African American, Asian, Caucasian, Hispanic and Native American. If a patient did not report a racial or ethnic group, or if a particular racial or ethnic group did not fall into one of these five major categories, we listed it as ‘unknown’.
We compared the prevalence of placenta previa for each of these five major racial and ethnic groups using the Pearson χ2-test. We also stratified the prevalence of placenta previa among women with twin pregnancies and among women with a history of cesarean delivery by racial/ethnic categories using the Pearson χ2-test. We further examined the prevalence of placenta previa by maternal age from 17 to 45 years using the Pearson χ2-test.
In order to estimate the unadjusted and adjusted odds ratios (OR) of having a pregnancy complicated by placenta previa with 95% confidence intervals (CI), we performed both bivariate and multivariable logistic regression analyses. In the multivariable model, we controlled for maternal race/ethnicity (reference group was Caucasian), maternal age (reference age of 20–29), multiple gestation (reference group was singleton gestation) and prior cesarean delivery (reference group was no history of cesarean).
We included 394 083 deliveries in our cohort for the period of 1995-2006. Of these deliveries, we confirmed that 1580 pregnancies were complicated by placenta previa with an overall prevalence of 0.40% (Table 1). The overall cesarean delivery rate among the entire cohort was 20.8%, and all women with a placenta previa underwent a cesarean delivery. Table 1 shows comparisons of characteristics between the major racial/ethnic groups, as well as for the entire cohort.
From a random chart review of 240 cesarean deliveries without an ICD-9 code for placenta previa, we found that there were no cases of unreported placenta previa. The upper confidence for 0% with a sample size of 240 is four cases. Assuming there were four cases of a missed diagnosis of placenta previa, the sensitivity of a placenta previa diagnosis in this study is 99.70% (95% CI 99.35–99.90%), and the specificity is 99.77% (95% CI 99.76–99.79%).
We also compared the following characteristics of the 1580 women with placenta previa to the 392 503 women without placenta previa (see Appendix 2, Supplementary Information, for a summary of this comparison). Both groups of women were racially diverse. Among the women with a placenta previa compared with no previa, 8.6 versus 7.9% were African American, 27.6 versus 17.3% Asian, 38.6 versus 43.4% Caucasian, 9.8 versus 11.7% Hispanic, 0.6 versus 0.4% Native American and 14.8 versus 19.3% unknown (P<0.001). Women with a placenta previa were older compared with their non-previa counterparts, with a mean age of 33.5±5.4 s.d. versus 29.4±6.0 s.d. More women with a placenta previa had a history of cesarean delivery, compared with women without previa, 21.5 versus 11.2% (P<0.001). Also, more women with a placenta previa compared with no previa had a current multiple gestation, 4.9 versus 3.0% (P<0.001).
The percent prevalence of placenta previa by race and ethnicity were as follows: African American 0.44%, Asian 0.64%, Caucasian 0.36%, Hispanic 0.34%, Native American 0.60% and unknown 0.31% (P<0.001). We further stratified our data by maternal age group categories. As expected, we found a universal rise in the prevalence of placenta previa by increasing maternal age for each of the major racial/ethnic groups. We also stratified our cohort by multiple gestation and history of cesarean delivery (Figure 1). With respect to twins, we found that there were no cases of placenta previa among African-American women (0/1170, P=0.021), whereas Asian women had the highest prevalence of previa in twins of 1.37% (22/1609, P<0.001). Asian women again had the highest prevalence of placenta previa, 1.1% (79/7120, P<0.001), among women with a prior cesarean delivery.
Increasing maternal age was associated with progressively higher rates of placenta previa (Figure 2). When compared with a baseline age group of 20–29 years, women older than a threshold of 30 years had progressively higher significant OR of a pregnancy complicated by placenta previa, whereas the opposite was true for women under 20 years old (Table 2).
In our multivariable logistic regression model, only Asians (OR 1.73, P<0.001) and African Americans (OR 1.43, P<0.001) had a significant adjusted OR for having placenta previa, compared with Caucasians when controlling for the confounders of maternal age, multiple gestation and prior cesarean delivery (Table 2). We calculated the area under the receiver operator characteristic (AUROC, or c-statistic) to depict model discrimination (a c-statistic=0.5 shows that results are completely random, whereas a c-statistic=1.0 indicates perfect discrimination). The c-statistic of our multivariable model is 0.69. To test the goodness-of-fit between observed and expected outcomes in our multivariable model, we used the Hosmer–Lemeshow technique (a P-value >0.005 is consistent with no significant difference between the observed and expected values, meaning that the model fit is acceptable, with a higher P-value indicating a better goodness-of-fit). The Hosmer–Lemeshow corresponding P-value for our model is 0.49.
We also calculated the relative contribution of each variable in the logistic regression model to the odds of having a placenta previa and found that maternal age had the greatest contribution (75.2%), followed by maternal race/ethnicity (13.2%), history of cesarean delivery (10.3%) and current multiple gestation (1.3%).
To assess differences in neonatal outcome by race and ethnicity among the cohort of women with a placenta previa, we calculated the rates of preterm delivery <37 weeks, as well as composite neonatal morbidity (consisting of respiratory distress syndrome, necrotizing enterocolitis, intraventricular hemorrhage and/or sepsis). The percent of women with a placenta previa, who experienced preterm delivery <37 weeks for each race and ethnicity were: African American 39%, Asian 47%, Caucasian 46%, Hispanic 35%, Native American 78% and unknown 55% (P<0.001). The percent of women by race and ethnicity with a placenta previa, who had at least one of the components of our composite neonatal morbidity were: African American 10%, Asian 8%, Caucasian 14%, Hispanic 9%, Native American 33% and unknown 11% (P=0.011).
Our large retrospective cohort study that included 1580 women with placenta previa shows that Asian women, followed by African-American women, had a higher prevalence of placenta previa than Caucasian women. When controlling for potential confounders such as maternal age, multiple gestation and prior cesarean delivery, Asian and African-American women were still significantly more likely to have a pregnancy complicated by placenta previa, compared with Caucasian women.
Novel to the literature on placenta previa, we report differences in neonatal outcomes by race and ethnicity. Interestingly, African Americans with a placenta previa had lower rates of preterm delivery than Caucasians or Asians. This is in contrast to the fact that African Americans have higher preterm delivery rates overall, compared with other racial/ethnic groups.24 In our study, African-American women with placenta previa also had less neonatal morbidity than their Caucasian and Asian counterparts. Native Americans had the highest rates of preterm delivery and poor neonatal outcomes among the women with placenta previa. Our study highlights the need for more research in health disparities with special attention to Native Americans as a high-risk population.
The relationship we found between Asian race and increased risk of placenta previa is consistent with three prior studies.14, 15, 16 Taylor et al.14 conducted a smaller case–control study of 810 women with placenta previa, using the Washington State birth certificate data, but only reported data on Asian and Caucasian women and showed a lower overall prevalence of placenta previa (3.3/1000), compared with our study (4/1000). The studies by Yang et al.15 and Shen et al.16 were larger retrospective cohort studies using the US national-linked birth/infant mortality database and the National Inpatient Sample, respectively. However, the Yang et al.15 study was limited by the absence of a Hispanic ethnic category, and their results differed from ours in that they showed African-American women had a lower frequency of placenta previa (3.0/1000) then Caucasian women (3.3/1000), compared with our study which found African-American women had a higher frequency of placenta previa (4.4/1000) compared with Caucasian women (3.6/1000). Although the Shen et al.16 study did include Hispanic women, as did our study, they reported that African-American women were at a higher risk of placenta previa then Asian women, in contrast to our result that Asian women had the highest risk of placenta previa followed by African-American women. Furthermore, the overall lower prevalence of placenta previa in the Taylor et al.14 and Yang et al.15 studies suggest that the information they derived from birth certificate data may have underreported placenta previa.
The relationship between race and ethnicity and the development of placenta previa is poorly understood. The pathophysiology of placenta previa may be multifactorial and explained by a combination of genetic, cultural and socioeconomic factors. One proposed mechanism of placenta previa is abnormal placentation due to prior endometrial or myometrial damage such as in the setting of a previous cesarean delivery, myomectomy or abortion. Other proposed theories include the failure of normal placental migration away from the cervical os during gestation, poor uteroplacental perfusion as seen with tobacco use and increased placental surface area as in the setting of multiple gestation.14 Research is needed to elucidate the genetic and non-genetic reasons why Asian and African-American women are at an increased risk of placenta previa.
One of the strengths of our study is its large size, with a sufficiently diverse population to be able to detect significant differences in the prevalence of placenta previa by racial and ethnic subgroups. The validity of our data is strong in that each diagnosis of placenta previa was confirmed with a chart review by a board certified obstetrician who was blinded to race and ethnicity, and that the random chart review of cesarean deliveries without a placenta previa ICD-9 code did not reveal any cases of unreported placenta previa. An additional strength of our study is that we had adequate medical information to control for important confounders such as prior cesarean delivery, maternal age and multiple gestation.
Our study is also unique in that we differentiated between the major racial and ethnic groups, whereas numerous previous studies either grouped Asian and African-American women together in an umbrella ‘minority’ group,2, 17, 23 or combined non-Caucasian non-African-American women into one category.8, 11, 13, 18, 21, 22 Other studies failed to include a separate category for Asian17 or Hispanic women,15 or had limited comparisons of only Asian and Caucasian women, but not African-American or Hispanic women.14 (See Appendix 1, Supplementary Information, for a summary of the literature).
Asian Americans are a heterogeneous group with significant differences in perinatal outcomes.25 However, because we lacked information on the country of origin among Asians in our cohort, we were unable to assess whether the associated risk of placenta previa varied by sub-categories of Asian country of origin. We were also limited in our data analysis by a lack of information regarding parity, history of spontaneous and/or induced abortion and tobacco use, all of which have been identified as risk factors for placenta previa.9, 11, 12, 13 We were therefore unable to control for these potential confounders in our logistic regression model. Despite this, the c-statistic of our multivariable model (0.69) is relatively high, given the limitations of our predictors. Furthermore, we were limited by the inability to demonstrate causality due to the retrospective cohort study design.
In conclusion, we found that Asian women, followed by African-American women, have a significantly increased risk of pregnancy complicated by placenta previa, compared with Caucasian women. Based on these results, we suggest that clinicians should have a higher index of suspicion for placenta previa in these racial groups and must exclude its presence with detailed antenatal ultrasound evaluation. Furthermore, these results may help inform future research of possible genetic and health disparities-related contributions to the etiology of placenta previa (Supplementary Information).
This study was supported by Kaiser Foundation Research Institute Community Benefits Grant.
About this article
Supplementary Information accompanies the paper on the Journal of Perinatology website (http://www.nature.com/jp)
Journal of Racial and Ethnic Health Disparities (2018)