According to the U.S. Census Bureau, Hispanics are the largest and fastest growing ethnic group in the United States.1 By 2050, the percentage of Hispanic women at childbearing age is projected to increase 92%, compared to a 10% increase for African Americans.2 In recent years, there has been increasing evidence of a “Hispanic Paradox”3 concerning health outcomes in the United States: Hispanic groups that are mainly of low socioeconomic status have better health outcomes compared with non-Hispanic whites of similar socioeconomic status. A second part of the paradox, the “Acculturation Paradox,” is that health behaviors, risk factors, and health outcomes become unfavorable with greater acculturation.4 In Hispanics, investigators have shown that acculturation is a predictor of preterm birth5 and may account for some of the differences in the preterm birth statistics by regions in the United States. As the Hispanic population continues to grow and Hispanics become more acculturated to U.S. lifestyles, the effect of protective factors associated with maternal birth outside of the United States is likely to diminish, contributing to a projected increase in preterm births among Hispanics.6
Acculturation is defined as the development of new cultural traits when two previously culturally distinct groups interact with one another.7 As acculturation occurs, individuals choose some new behaviors based on their new cultural beliefs. These behaviors subsequently become factors affecting their health. As Hispanics are exposed to U.S. culture, they may obtain both positive and negative values and behaviors. It is erroneous to assume that higher levels of acculturation relate to more positive health outcomes.8
Progesterone and estriol are key reproductive hormones in pregnancy, and it is generally accepted that both hormones are important in the process of parturition.9 Progesterone is thought to maintain uterine quiescence; it promotes and maintains the pregnant state. Progesterone therapy is currently used for women who had a previous, spontaneous preterm birth.9 Functional progesterone withdrawal may be a key event in human parturition.10 In contrast, estriol opposes progesterone and may stimulate biochemical changes in the uterus leading to labor and delivery.11 However, the processes of progesterone withdrawal and estriol activation are not clearly understood, especially in human parturition. We included progesterone and estriol measured in the blood as key indicators related to the process of parturition, part of a larger study related to other stress hormones and immune factors in the blood related to parturition. Our study had the following purpose: we sought to examine the Hispanic acculturation paradox by identifying the effect of acculturation on serum progesterone and estriol levels, the progesterone/estriol ratio, and preterm birth.
MATERIALS AND METHODS
The sample consisted of 468 women. We determined the sample size by a power analysis to give a power of 0.80, capability to detect moderate effects, with a significance of .05. The inclusion criteria were 1) singleton intrauterine gestation at 22–24 weeks of gestation confirmed by accurate last normal menstrual period and ultrasonography, 2) self-identification as Hispanic ethnicity, 3) 14–40 years of age, and 4) ability to read either English or Spanish. Exclusion criteria were 1) inability to read either English or Spanish, 2) known uterine or cervical abnormalities, 3) multiple gestation, 4) kidney disease, 5) heart disease, 6) autoimmune disorders, 7) diabetes requiring medication, 8) asthma requiring use of steroid inhaler, 9) preeclampsia at the time of data collection, 10) oral steroid 1 month before the time of enrollment, 11) congenital anomalies as determined on fetal ultrasonography, 12) blood group isoimmunization, 13) active cervical vaginal bleeding or placenta previa, 14) major diagnosis of mental disorders, and 15) self-report of an ethnicity other than Hispanic. We included gestational diabetics after obtaining maternal and infant outcomes only if they were diet controlled. Those participants with preeclampsia were included in the sample if the preeclampsia occurred after data collection. The institutional review boards at the University of Texas Health Science Center at San Antonio and the University of Texas Medical Branch at Galveston gave approval to conduct the study, and all subjects gave informed consent.
We recruited participants from multiple sites in central, south, and the Gulf coast regions of Texas from 2003 to 2007. In Austin, we recruited women from five community clinics (n=279, 54%), primarily state funded and self-pay. In San Antonio, we recruited women from five private practice physicians (n=172, 33%), primarily Medicaid patients. In Pasadena (Houston area) we recruited women from one community clinic (n=68, 13%), primarily state funded and self-pay.
We screened prenatal charts of potential participants at a gestational age of less than 24 weeks, approaching participants before or after routine prenatal visits. The recruiter explained the study thoroughly, including obtaining information on the newborn, and she obtained informed consent in Spanish or English. At a separate scheduled data collection visit, the research nurse obtained demographic data, including years residing in the United States, from the prenatal chart, and she distributed a self-reported questionnaire (acculturation). We asked the participants to fill out a detailed family of origin chart indicating country of birth for themselves, as well as for their parents and grandparents. A research nurse measured height and weight, obtained the prepregnancy weight from the prenatal chart to calculate body mass index, and drew a venous blood sample for hormone measurements. The blood samples were drawn only from 1 pm to 3 pm to control for diurnal rhythms of the hormones. A research nurse also conducted a speculum examination and obtained vaginal discharge on a dry microscopic slide for later Gram staining for bacterial vaginosis. The prenatal chart was reviewed for ultrasound dating criteria. An ultrasonography at less than 20 weeks was used as preferred dating criteria. If the ultrasound examination was performed after 20 weeks, confirmation of dating was a combination of late ultrasonography and last menstrual period. The prepregnancy weight was obtained from the prenatal chart to calculate body mass index. After delivery, a research nurse reviewed the medical records of the participants at the delivery hospital for birth outcome data.
Acculturation was operationalized using a self-report instrument, the Bidimensional Acculturation Scale for Hispanics.12 The scale measures bidirectional changes in behavior that are important to a person in two cultural domains, Hispanic and non-Hispanic. We used the Language Proficiency subscale because it is a shorter version that provides a quick and efficient measure of acculturation. Previous testing with the Bidimensional Acculturation Scale showed high internal consistency and high validity coefficients, working effectively with Mexican-Americans and with Central Americans.13 We recognize that this scale measures only partially the many potential dimensions of acculturation; future work is needed in development of a more multi-dimensional measurement of acculturation. However, this scale does attempt to examine several domains of acculturation and currently is a valid choice of measurement. Psychometric analysis has shown that language items, as compared with other constructs, explain most of the variance of acculturation scales.14 A score of 2.5 was used as a cutoff score to indicate a low or high level of adherence to each cultural domain. The scores were dichotomized for analysis according to this high and low cutoff. This is based on the work of Marin, the author of the scale.12 We interpreted scores above 2.5 on both English and Spanish as indicating biculturalism of the subjects. There were three possibilities: English acculturated, Spanish acculturated, and bicultural. However, for purposes of analysis, we grouped the bicultural with the English acculturated. English and bicultural were considered in the more acculturated group compared with Spanish acculturated, which was considered in the less acculturated group. In addition, years in the United States was used as a linear measure of acculturation to validate relationships with the Language Proficiency subscale. Country of birth was another measure of acculturation. These measures assume that we can measure acculturation by examining the amount of exposure that individuals may have to the dominant culture (in this case, English-speaking culture).12
Blood samples were centrifuged at 2,000g (10 minutes) and the serum decanted into polypropylene tubes and stored at –80°C within 2–4 hours of collection. Progesterone was analyzed with an Enzyme Immunoassay technique (MP Biomedicals, Irvine, CA) and read spectrophotometrically at 450 nm with a μ-Quant Reader (Bio-Tek Instruments, Inc, Winooski, VT). Estriol was analyzed by EIA (Labor Diagnostika Nord, Nordhorn, Germany). The interassay coefficient of variation for progesterone was 9.04%, and the sensitivity of the assay was 0.3 ng/mL. The interassay coefficient of variation for estriol was 10.18%, and the sensitivity of the assay was 0.02 ng/mL. To control for potential differences in laboratory technique, the same laboratory technician conducted the analyses blinded to subject names and histories.
The research nurses conducted speculum examinations to obtain vaginal samples for Gram staining for bacterial vaginosis. Using the criteria defined as a standardized method,15 the three morphotypes shown to be most reliable for analyses of bacterial vaginosis included Lactobacillus spp, Gardnerella vaginalis, and Mobiluncus spp. A score from 1 to 10 was given summing the contributions of the individual morphotypes. Morphotypes were scored as the average number seen per oil immersion field. Total score=lactobacilli+G vaginalis and Bacteroides spp+curved rods. The three morphotypes shown to be the most reliable were used to create the scale. Normal was at the low end, and bacterial vaginosis at the upper end. Using the above scoring system, there was a correlation of 0.82 across data collection sites with very small mean differences in the category scored, indicating excellent agreement.
Maternal and infant outcomes were obtained by reviewing the medical record at delivery, using less than 37 weeks and 0 days of gestation as the definition of preterm birth. The data were analyzed with Wilcox-Mann-Whitney and Kruskal-Wallis tests, t tests, analysis of variance, logistic regression, and structural equation modeling. We analyzed progesterone and estriol for normal distribution and corrected with a log transformation. Normality was evaluated by histogram as well as skewness and kurtosis statistics.16 We used SAS 9.1 (SAS Institute Inc, Cary, NC).
The majority of the 468 women enrolled in the study were between 18 and 34 years old, not married, foreign-born, and had less than a high school education (Table 1). The percentages of Hispanic ethnicity were as follows: 28.63% (n=134) Mexican-American, 60.26% Mexican (n=282), 0.85% Puerto Rican (n=4), 6.84% South and Central American (n=32), and 3.42% (n=16) other or not definitely identified other than Hispanic. Yearly household income had a mean of $20,418 (standard deviation [SD] $11,966) with a median of $18,000 annual income. The gravidity in the sample was as follows: gravida 1 (n=127) 27.67%, gravida 2 or greater (n=332) 72.33%. The parity in the sample was as follows: para 0 (n=166) 36.4%, para 1 (n=144) 31.58%, para 3 (n=25) 5.48%, para 4 (n=11) 2.41%, para 5 (n=4) 0.88%, para 6 (n=1) 0.22%. The number of women having a previous preterm birth was 8.67% (n=37).
There were 36 preterm births in the sample, with 31 spontaneous and with five medically indicated due to preeclampsia. The incidence of preterm birth was 10.2% for the more acculturated women (English and bilingual) compared with 5% for the non–English-proficient women. The preterm birth rate was 13.9% for the women who were not proficient in Spanish compared with 5.8% for the women who were proficient in Spanish. Eight of the 36 preterm infants were delivered at less than 34 weeks of gestation.
In addition, we examined the influence of drug and cigarette use through urine samples analyzed by a commercial laboratory. Except for one infant of a mother who was using an opiate and who delivered a term low birth weight infant, the use of drugs did not influence preterm birth. The number of smokers was also very low (n=3) in this sample.
Of the 36 preterm births, there were no women with sexually transmitted diseases or documented chorioamnionitis on the medical record. Fifteen women had positive bacterial vaginosis scores per Nugent’s criteria on Gram staining, five with intermediate scores, and 16 with negative scores.
Progesterone levels were significantly higher among women who were foreign-born, with less than a high school education, having no medical insurance, low English proficiency, and living in the United States less than 10 years (Table 1). Factors with significant associations with the progesterone/estriol ratio included age (younger and older women had lower ratios), country of birth (U.S.-born had lower ratios than foreign-born). Other factors associated with the progesterone/estriol ratio were education (greater than high school had lower ratios) and medical insurance (those women with insurance had lower progesterone/estriol ratios). Women who had a positive bacterial vaginosis score had statistically significant higher progesterone levels than those with negative scores. However, there were not statistically different levels of estriol or the progesterone/estriol ratio by negative, positive, or intermediate bacterial vaginosis scores. Women who were proficient in English had lower progesterone/estriol ratios, whereas women who were proficient in Spanish had higher progesterone/estriol ratios. Women who were in the United States less than 10 years had higher progesterone/estriol ratios compared with those women who were in the United States more than 10 years.
The mean level for progesterone was 32.5 ng/mL (SD 11.44), ranging from 8.16 to 115.14 ng/mL. The estriol level mean was 10.36 (SD 4.66), ranging from 1.70 to 32.43. The progesterone/estriol ratio mean was 3.73 (SD 2.24), ranging from 0.84 to 17.67.
Table 2 gives the results of logistic regression of the risk for preterm birth of progesterone, estriol, and the progesterone/estriol ratio. We controlled for the following maternal characteristics potentially associated with preterm birth: marital status, age, education, prepregnancy body mass index, history of previous preterm birth, and gravidity. Marital status was dichotomized as single, married, or disrupted. Gravidity was also dichotomized as one or greater than one. History of previous preterm birth was dichotomized as yes or no. Education was dichotomized as less than high school or greater than high school.
The progesterone/estriol ratio is highly significant in predicting the risk of preterm birth at the lowest quartile, when adjusted for the aforementioned maternal characteristics and gestation at blood draw. With the higher quartiles, the odds ratios become nonsignificant. It is important to note that the results had nonsignificant changes when excluding the participants who had preterm birth due to preeclampsia. Table 3 gives the results of the risk ratios of English and Spanish proficiency on the risk of preterm birth. Higher English proficiency significantly positively predicted preterm birth, whereas Spanish proficiency actually negatively predicted preterm birth.
Figure 1 shows a plot of the probability of preterm birth by the raw levels of the progesterone/estriol ratio, indicating that the lower the ratio, the higher the probability of preterm birth. The lower the progesterone in the ratio and the higher the estriol, the greater the probability of preterm birth.
Figure 2 displays the path of the effect of English proficiency on gestational age. The method of analysis for this figure is structural equation modeling. Primary advantages of this approach include the ability to 1) assess fit of the overall model to data, 2) test hypothesized relationships among many variables simultaneously, 3) test indirect or mediating relationships, and 4) model complex combinations of binary and continuous data in the same model. Structural equation models are typically illustrated with path diagrams depicting variables and the hypothesized paths or coefficients connecting them. Nonstandardized and standardized path coefficients yield information about the size and the direction of the investigated paths and the relative importance of different paths, respectively.
The degree to which the model fits the data can be gauged by referring to fit indices, including the comparative fit index, nonnormed fit index, root mean square error of approximation, and the standardized root mean-squared residual. The comparative fit index and nonnormed fit index have been proposed as an alternative to the χ2 test. Values of comparative fit index range from 0 to 1, with values over 0.9 indicative of a good fit of the model.17,18
Greater English proficiency predicts decreased progesterone and higher estriol, (lower progesterone/estriol ratio) and directly predicts lower gestational age. English proficiency thus has a direct effect on decreasing gestational age (direct line from English proficiency to gestational age) as well as an indirect effect through the progesterone/estriol ratio (lines from English proficiency to progesterone and estriol). The higher the progesterone/estriol ratio was for a participant, the higher the participant’s gestational age. All fit indices indicated a good fit of the model using the above noted criteria. The progesterone/estriol ratio was analyzed by using a rank technique because the values did not have normal distribution. Gestational age was used as the outcome variable in the structural equation model because a continuous outcome variable is a more sensitive indicator than a binary outcome variable (preterm yes or no).
For each path in Figure 2, we conducted multiple regression analysis to adjust for other maternal and pregnancy variables. The path coefficients and their significant levels had minimal changes after adjusting for age, marital status, body mass index, and gestational age at which the blood was drawn. Because country of birth, years in the United States, medical insurance, and education were highly associated with acculturation, these variables were not adjusted for in the multiple regression analysis. The above results still held after excluding the five subjects with preterm birth due to preeclampsia.
Current evidence has established that preterm birth is multifactorial in origin.19,20 Improved understanding of the different mechanisms leading to preterm birth is essential to improve outcomes. Recent investigation21 has also illuminated the morbidities associated with late preterm birth as well as early preterm birth. Late preterm births make up 74% of all preterm births and are on the rise. Many of the preterm births from this sample were late preterm births.
Further understanding of the function of the hormones progesterone and estrogen is vital to understanding the mechanism of both term and preterm birth. Progesterone is an essential hormone in the regulation of human reproduction. One of the major functions of progesterone is to maintain pregnancy. It does this by supporting growth of the uterus and inhibiting uterine contractility.22 Researchers have postulated that the link between progesterone withdrawal and estrogen activation is part of an important mechanism for the endocrine control of human parturition.23,24 The results presented here validate that mechanism, even with levels drawn as early as 22–24 weeks of gestation.
The findings presented here represent a leading edge of new research. Acculturation in Hispanic women appears to affect the progesterone/estriol ratio, and this effect may partially explain the higher risk of preterm birth. Acculturation is directly related to preterm birth and affects the progesterone/estriol ratio and, thus, preterm birth. We speculate that one of the potential causes for the change in the ratio with greater acculturation may be the changes in the physiologic stress response resulting from increased acculturative stress. Of note is that progesterone levels differed significantly (decreased levels) in U.S.-born Hispanics as compared with foreign-born and differed depending on whether or not the women had medical insurance or were more educated. Country of birth (United States), Medicaid status, more than a high school education, and years in the United States all correlated with acculturation. In addition, younger and older women had lower progesterone/estriol ratios, consistent with previous evidence that younger and older women are more at risk for preterm birth.25
One limitation of this study is the lack of a more multi-dimensional measure of acculturation. Additionally, future work will also need to identify when in the acculturative process the deleterious response begins with maternal and fetal health. Also important for upcoming work is the delineation of which aspects of acculturation are most harmful to the health of the mother and fetus.
There is growing evidence that a higher acculturation level is related to worsening birth outcomes, specifically gestational age at delivery. Lara et al6 cite the fact that the effect of acculturation is the “cleanest” in Mexican-American women; the evidence agrees with this premise.26–28 Some literature reports mixed or no effects of acculturation on birth outcomes. This is consistent with the problem of measuring acculturation. Lara et al6 states that language, place of birth, and length of stay in the United States contribute to different effects on birth outcomes. Our results indicate that all three of these concepts have similar physiological effects on progesterone and estriol and the progesterone/estriol ratio. Heilemann et al29 found that place of birth, but not language preference, for Mexican-American women (ie, U.S.-born) most consistently predicted prenatal complications, excessive weight gain, cesarean births, and postpartum complications. Our results agree with those of Heilemann et al relating to U.S.-born compared with foreign-born women, namely that the U.S.-born women have the lowest progesterone/estriol ratios and subsequently more preterm births. It is essential that researchers elucidate the multi-dimensional aspects of acculturation to develop interventions that accurately target the effects of acculturation.
These findings demonstrate another possible aspect of obstetric risk that remains unconsidered in Hispanic women, that of acculturation. Further refinement of the risk of acculturation is essential to clarify how we can adjust our clinical care to prevent increasing preterm birth with the ever-increasing Hispanic population.
3. Franzini L, Ribble JC, Keddie AM. Understanding the Hispanic paradox. Ethn Dis 2001;11:496–518.
4. Abraido-Lanza AF, Chao MT, Florez KR. Do healthy behaviors decline with greater acculturation? Implications for the Latino mortality paradox. Soc Sci Med 2005;61:1243–55.
5. Coonrod DV, Day RC, Balcazar H. Ethnicity, acculturation, and obstetric outcomes: different risk factor profiles in low and high acculturation Hispanics and in white non-Hispanics. J Reprod Med 2004;49:17–22.
6. Lara M, Gamboa C, Kahramanian MI, Morales LS, Bautista DE. Acculturation and Latino health in the United States: a review of the literature and its sociopolitical context. Annu Rev Public Health 2005;26:367–97.
7. Hunt L, Schneider S, Comer B. Should “acculturation” be a variable in health research? A critical review of research on US Hispanics. Soc Sci Med 2004;59:973–86.
8. Acevedo MC. The role of acculturation in explaining ethnic differences in the prenatal health-risk behaviors, mental health, and parenting beliefs of Mexican-American and European American at-risk women. Child Abuse Negl 2000;24:111–27.
9. Strauss J, Barbieri RL. Yen and Jaffe’s Reproductive endocrinology: physiology, pathophysiology and clinical management. 5th ed. Philadelphia (PA): Saunders; 2004. p. 354–7.
10. Use of progesterone to reduce preterm birth. ACOG Committee Opinion No. 291. The American College of Obstetricians and Gynecologists. Obstet Gynecol 2003;102:1115–6.
11. Pepe G, Albrecht E. Actions of placental and fetal adrenal steroid hormones in primate pregnancy. Endocr Rev 1995;16:608–48.
12. Marin G, Gamba RJ. A new measurement of acculturation for Hispanics: the Bidimensional Acculturation Scale for Hispanics (BAS). Hisp J Behav Sci 1996;18:297–316.
13. Marin G. Issues in the measurement of acculturation among Hispanics. In: Geisinger KF, editor. Psychological testing of Hispanics. Washington (DC): American Psychological Association; 1992. p. 23–51.
14. Cabassa LJ. Measuring acculturation: where we are and where we need to go. Hisp J Behav Sci 2003;25:127–46.
15. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 1991;29:297–301.
16. Tabachnick BG, Fidell LS. Using multivariate statistics. 4th ed. Boston (MA): Allyn and Bacon; 2001. p. 73.
17. Browne MW, Cudeck R. Alternative ways of assessing model fit. In: Bollen KA, Long JS, editors. Testing structural equation models. Newbury Park (CA): Sage Publications; 1993. p. 136–62.
18. Hu L, Bentler PM. Cutoff criteria for fit indices in covariance structural analysis: conventional criteria versus new alternatives. In: Structural equation modeling. Vol 6. Newbury Park (CA): Sage Publications 1999. p. 1–55.
19. Romero R, Mazor M, Munoz H, Gomez R, Galasso M, Sherer DM. The preterm labor syndrome. Ann N Y Acad Sci 1994;734:414–29.
20. Challis JRG. Mechanism of parturition and preterm labor. Obstet Gynecol Surv 2000;55:650–60.
21. Davidoff MJ, Dias T, Damus K, Russell R, Bettegowda VR, Dolan S, et al. Changes in the gestational age distribution among U.S. singleton births: impact on rates of late preterm birth:1992 to 2002 [published erratum appears in Semin Perinatol 2006;30:313]. Semin Perinatol 2006;30:8–15.
22. Graham JD, Clarke CL. Physiological action of progesterone in target tissues. Endocr Rev 1997;18:502–19.
23. Thijssen JH. Progesterone receptors in the human uterus and their possible role in parturition. J Steroid Biochem Mol Biol 2005;97:397–400.
24. Mesiano S, Chan EC, Fitter JT, Kwek K, Yeo G, Smith R. Progesterone withdrawal and estrogen activation in human parturition are coordinated by progesterone receptor A expression in the myometrium. J Clin Endocrinol Metab 2002;87:2924–30.
25. Reedy NJ. Born too soon: the continuing challenge of preterm labor and birth in the United States. J Midwifery Womens Health 2007;52:281–90.
26. Cervantes A, Keith L, Wyshak G. Adverse birth outcomes among native born and immigrant women: replicating national evidence regarding Mexicans at the local level. Matern Child Health J 1999;3:99–109.
27. Crump C, Lipsky S, Mueller BA. Adverse birth outcomes among Mexican-Americans: are U.S. born women at greater risk than Mexico-born women? Ethn Health 1999;4:29–34.
28. Wolff CB, Portis M. Smoking, acculturation, and pregnancy outcome among Mexican-Americans. Health Care Women Int 1996;17:563–73.
29. Heilemann MV, Lee KA, Stinson J, Koshar JH, Goss G. Acculturation and perinatal health outcomes among rural women of Mexican descent. Res Nurs Health 2000;23:116–25.