Bryant, Allison S. MD, MPH1; Cheng, Yvonne W. MD, MPH1; Lyell, Deirdre J. MD2; Laros, Russell K. MD1; Caughey, Aaron B. MD, PhD1
From the 1Department of Obstetrics, Gynecology, and Reproductive Sciences, University of California, San Francisco; and 2Department of Gynecology and Obstetrics, Stanford School of Medicine, Palo Alto, California.
Dr. Caughey is supported by the National Institute of Child Health and Human Development, Grant HD01262 as a Women’s Reproductive Health Research Scholar.
Dr. Bryant is supported by the National Institutes of Health Grant 1 KL2 RR024130-01 as a UCSF Multidisciplinary Clinical Research K12 Scholar.
Corresponding author: Allison S. Bryant, MD, MPH, Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, 505 Parnassus Avenue, Box 0132, San Francisco, CA 94143-0132; e-mail: email@example.com.
OBJECTIVE: To estimate the association between presence of the sickle cell trait and preterm delivery among African-American women.
METHODS: A retrospective cohort study was conducted to study all deliveries by African-American women at one institution between 1976 and 2001. The primary predictor of interest was the presence of the sickle cell trait, and the primary outcome was preterm delivery. Post hoc analyses were conducted to explore the relationship between presence of sickle cell trait and multiple gestations, an apparent modifier of the relationship between sickle cell trait and preterm delivery.
RESULTS: Of the 5,028 African-American women eligible for inclusion, 6.5% carried the sickle cell trait. In adjusted analyses, the presence of sickle cell trait was associated with a lower risk of preterm delivery at less than 32 weeks (adjusted odds ratio 0.15, 95% confidence interval 0.05–0.49), and was found to be associated with an increased odds of multiple gestations (adjusted odds ratio 1.94, 95% confidence interval 1.22–2.09). A significant interaction exists between the presence of multiple gestation and sickle cell trait and the odds of preterm delivery, such that the protective effect of sickle cell trait on preterm delivery risk is greater among those with multiple gestations than those with singletons.
CONCLUSION: Among African-American women, the presence of the sickle cell trait was associated with lower odds of early preterm delivery despite adjustment for potential confounders. An increased odds of multiple gestations was also noted among these women, possibly suggesting a persistence of selection for the hemoglobin S trait in nonmalarial settings.
LEVEL OF EVIDENCE: II
Sickle cell trait is relatively prevalent among African Americans in the United States, with carrier frequencies of 8–10%1 while rates as high as 25–30% in parts of West Africa have been recorded.2 Carriers of the sickle cell trait have hemoglobin that is composed mostly of hemoglobin A, with a smaller fraction of the abnormal hemoglobin S. Sickle cell trait is considered to be a benign carrier state with only mild, if any, hematologic derangements. While the genetic implications of carriage of the sickle cell trait are well established, the effect, if any, on pregnancy outcomes remains debated. An increased risk of asymptomatic bacteriuria and urinary tract infections during pregnancy has been clearly demonstrated.3–5 Some studies, including many conducted in non–U.S. populations, have also shown an increased risk of pregnancy, postpartum, and neonatal complications among women with sickle cell trait,6–9 whereas others have not.3,10–12 In a Nigerian population, carriage of the sickle cell trait was associated with significantly fewer malarial attacks among pregnant women, but the effect of sickle cell trait on pregnancy in nonmalarial environments remains unclear. Few studies have examined the association between sickle cell trait and preterm delivery, in particular among U.S. African Americans, and those that have involved relatively few subjects.
We hypothesized that there is no association between carriage of the sickle cell trait and preterm delivery risk among African-American women. We sought to test this hypothesis within a large cohort of African-American women delivering at a single U.S. institution.
MATERIALS AND METHODS
We conducted a retrospective cohort study of all African-American, non-Hispanic women delivered at more than 24 weeks of gestation at the University of California at San Francisco between 1976 and 2001. Women who were transported to University of California at San Francisco from other institutions for delivery and thus had limited information on early pregnancy screens were excluded, as were women with sickle cell disease.
The primary outcomes of interest were preterm delivery before 37 weeks of gestation and very preterm delivery before 32 weeks of gestation. Using χ2 tests for categorical variables and one-way analysis of variance tests for continuous predictors, we conducted univariable analyses to identify predictors of preterm deliveries in this population. The primary predictor of interest was carriage of sickle cell trait. Hemoglobinopathies were screened for using hemoglobin electrophoresis in all women included in these analyses.
Other covariates were examined as possible confounders or effect modifiers of sickle cell trait upon preterm delivery risk. These covariates included maternal age, parity, receipt of public health insurance, prior preterm delivery, multiple gestations, year of delivery, and maternal comorbidities such as preeclampsia, gestational diabetes, and anemia. Interactions were tested in multivariable models by creating and including selected interaction terms and determining Wald statistics and their associated P values.13 Significant interactions were considered when the associated P values were less than .20. Multiple logistic and linear regression models were created using backward selection to model the odds of preterm delivery and gestational age at delivery, adjusting for significant confounders and interaction terms.
To further explore the effects of the sickle cell trait upon preterm birth risk, we performed post-hoc univariable and multivariable analyses to examine the association between sickle cell trait and multiple gestations. Cochran-Mantel-Haenszel tests of homogeneity were used to test for interactions in these relationships; P<.20 was used as the cutoff.
Institutional review board approval was obtained from the Committee on Human Research at University of California at San Francisco. All analyses were performed using the STATA 8.0 software package (Stata Corporation, College Station, TX).
Of the 5,392 African-American women delivered at our institution between 1976 and 2001, 5,028 (93.2%) were eligible for study inclusion. Three hundred twenty-six of these (6.5%) were carriers of the sickle cell trait. Term deliveries occurred in 4,318 women (85.8%), whereas 494 (9.8%) were delivered between 32 and 37 weeks of gestation, and 216 (4.3%) were delivered at less than 32 weeks of gestation. Table 1 displays the relationship between sickle cell trait and various risk factors for preterm delivery. In these unadjusted analyses, carriers of sickle cell trait had more multiple gestations when compared with noncarriers (3.5% and 1.8% respectively, P=.04). Sickle cell trait carriers were less likely to be nulliparous and had lower mean hematocrits on hospital admission at the time of delivery, which were of no clinical significance. Among selected subgroups of women in our population, presence of the sickle cell trait seemed to be associated with a lower risk of preterm and very preterm deliveries (Table 2).
When we adjusted for those covariates identified in bivariate analyses to be significantly associated with the risk of preterm delivery, the presence of the sickle cell trait was associated with a lower risk of preterm delivery less than 32 weeks (adjusted odds ratio [AOR] 0.15, 95% confidence interval [CI] 0.05–0.49), as presented in Table 3. Given the possibility that women who presented late to care or directly to Labor and Delivery with a preterm delivery and no prenatal care may not have been screened for sickle cell trait but could have been miscoded as not having the trait, we performed a sensitivity analysis excluding all women with first prenatal visits after 28 weeks gestation. Our findings of a protective effect of sickle cell trait against preterm delivery less than 32 weeks became robust with an AOR of 0.15 (95% CI 0.03–0.62). Similarly, adjustment of our models for year of delivery to account for temporal changes in practice did not change the relationship between sickle cell trait and preterm delivery.
When we added interaction terms to the model, an interaction between sickle cell trait and multiple gestations was suggested by this term’s P value of .10. After adjustment for the interaction term, the effects of sickle cell trait on preterm delivery remained statistically significant. When we stratified by the presence of multiple gestations (Table 3), we found that sickle cell trait was not significantly associated with preterm delivery at less than 37 weeks among singletons, but was associated with a lower risk of preterm delivery at less than 32 weeks with an AOR of 0.26 (95% CI 0.08–0.84). Among women with multiple gestations who also carried the sickle cell trait (n=22), there was a nonsignificant reduction in the risk of delivery at less than 37 weeks as compared with women without sickle cell trait (AOR 0.32, 95% CI 0.09–1.08). None of these women had a delivery earlier than 32 weeks gestation; thus we were unable to calculate an AOR for sickle cell trait. Figure 1 shows the proportion of women with preterm deliveries at less than 37 and less than 32 weeks, stratified by presence of the sickle cell trait and multiple gestations, and suggests a significant mitigation of the effect of multiple gestations on preterm delivery by the presence of the sickle cell trait. When gestational age at delivery was treated as a continuous variable, presence of the sickle cell trait among patients with singletons was associated with a statistically insignificant increase in weeks of gestation (0.26 weeks 95% CI –0.06 to 0.60) after adjustment for other predictors. Among patients with multiple gestations (n=164), sickle cell trait was associated with an increase in gestation of 2.72 weeks (95% CI 0.88–4.55).
We also performed a post-hoc analysis to explore the effect of the sickle cell trait on the odds of multiple gestations among African-American women. We found an increased odds of multiple gestation among women who carry the sickle cell trait (AOR 1.94, 95% CI 1.22–2.09), even after adjustment for the use of assisted reproductive technologies.
Last, there were 42 women identified as having sickle cell disease in our overall population. These women were excluded for the purposes of the original analyses; however, a post-hoc analysis revealed that there were no differences in risk of preterm or very preterm deliveries or multiple gestations between those African-American women with sickle cell disease and those without. Detection of statistically significant differences may have been hindered by small numbers.
Among our cohort of African-American women, we found that the presence of the sickle cell trait is associated with a reduced risk of early preterm delivery, which we have not seen reported in the literature. Women with sickle cell trait had an 85% reduction in the risk of preterm delivery less than 32 weeks gestation, after adjustment for other associated risk factors for preterm delivery. Interestingly, the reduction in preterm birth seemed to be modified by pregnancy plurality, with a more pronounced effect seen among women with multiple gestations. We were also surprised to find that sickle cell trait seemed to be associated with an increased risk of multiple gestations in our population of African-American women.
A recent retrospective case–control study demonstrated shorter mean pregnancy duration among women with sickle cell trait compared with women with normal hemoglobin in a population of U.S. African-American women.14 We found contrary results, particularly with regard to gestational age at delivery and odds of early preterm delivery. Our study involves a much larger cohort of African-American women cared for at a single institution over three decades. Our database allowed for adjustment for multiple potential confounders which may influence the risk of preterm delivery, and allowed us to explore potential interactions between these variables. We also chose to exclude patients transported to our institution in an attempt to minimize information bias.
In a retrospective study conducted among African-American women in Alabama, investigators found that women with sickle cell trait had higher mean gravidity and more live births than women with hemoglobin AA.15 Another study in the United Kingdom showed a higher mean parity among women with sickle cell trait as compared with women without any hemoglobinopathies.12 Both groups of authors suggest that this apparent increase in fertility among women with sickle cell trait may demonstrate evidence of ongoing balancing selection for the hemoglobin S variant in nonmalarial environments. Our findings of increased risk of multiple gestations, greater proportion of multiparity, and decreased risk of preterm delivery in women with sickle cell trait lend credence to this theory.
Our study was somewhat limited by its retrospective nature. Although our practice throughout this time frame was to perform hemoglobin electrophoreses on all African-American women receiving prenatal care in our institution, we were unable to audit all medical records in our database to confirm this. We excluded women transported to our institution because our database lacked their early pregnancy screens and performed a sensitivity analysis limiting the population to women with prenatal care initiated in the first or second trimester, both of which help to allay concerns regarding information bias. We also acknowledge changes in reproductive health practices over the period during which data for this study was collected, which may have altered both the risk of preterm delivery and of multiple gestations. We are unaware that any of these practice changes were applied in a systematic fashion with regard to sickle cell trait status, however; adjustment for time period of delivery in our models did not change the conclusions with respect to the association between sickle cell trait and outcomes studied.
In conclusion, we found both an increased risk of multiple gestations and a decreased risk of very preterm delivery among African-American women with sickle cell trait. The latter risk was even more pronounced among the subgroup with multiple gestations. These findings can be used to counsel African-American women with sickle cell trait regarding pregnancy risk. Future studies should explore possible biologic causes and teleologic significance of this relationship.
1. Motulsky AG. Frequency of sickling disorders in U.S. blacks. N Engl J Med 1973;288:31–3.
2. Fleming AF. The presentation, management and prevention of crisis in sickle cell disease in Africa. Blood Rev 1989;3:18–28.
3. Baill IC, Witter FR. Sickle trait and its association with birthweight and urinary tract infections in pregnancy. Int J Gynaecol Obstet 1990;33:19–21.
4. Pastore LM, Savitz DA, Thorp JM Jr. Predictors of urinary tract infection at the first prenatal visit. Epidemiology 1999;10:282–7.
5. Miller JM Jr. Sickle cell trait in pregnancy. South Med J 1983;76:962–96, 965.
6. Larrabee KD, Monga M. Women with sickle cell trait are at increased risk for preeclampsia. Am J Obstet Gynecol 1997;177:425–8.
7. Manzar S. Maternal sickle cell trait and fetal hypoxia. Am J Perinatol 2000;17:367–70.
8. Pastorek JG 2nd, Seiler B. Maternal death associated with sickle cell trait. Am J Obstet Gynecol 1985;151:295–7.
9. Roopnarinesingh S, Ramsewak S. Decreased birth weight and femur length in fetuses of patients with the sickle-cell trait. Obstet Gynecol 1986;68:46–8.
10. Blattner P, Dar H, Nitowsky HM. Pregnancy outcome in women with sickle cell trait. JAMA 1977;238:1392–4.
11. Stamilio DM, Sehdev HM, Macones GA. Pregnant women with the sickle cell trait are not at increased risk for developing preeclampsia. Am J Perinatol 2003;20:41–8.
12. Tuck SM, Studd JW, White JM. Pregnancy in women with sickle cell trait. Br J Obstet Gynaecol 1983;90:108–11.
13. Hosmer D, Lemeshow S. Applied logistic regression. New York (NY): Wiley & Sons; 1989.
14. Taylor MY, Wyatt-Ashmead J, Gray J, Bofill JA, Martin R, Morrison JC. Pregnancy loss after first-trimester viability in women with sickle cell trait: time for a reappraisal? Am J Obstet Gynecol 2006;194:1604–8.
15. Hoff C, Thorneycroft I, Wilson F, Williams-Murphy M. Protection afforded by sickle-cell trait (Hb AS): what happens when malarial selection pressures are alleviated? Hum Biol 2001;73:583–6.
© 2007 by The American College of Obstetricians and Gynecologists.