Association of sickle cell trait with adverse pregnancy outcomes in a population‐based cohort

Abstract

Introduction

Sickle cell trait (SCT) is common in African descendants. Its association with several adverse pregnancy outcomes (APOs) has been reported but remains inconsistent. The objectives of this study are to test associations of SCT with APOs in non‐Hispanic Black women, including (1) validate the associations of SCT with previously reported APOs, (2) test novel associations of SCT with broad spectrum of APOs, and (3) estimate the attributable risk of SCT for implicated APOs.

Material and methods

This is a retrospective analysis of a prospectively designed population‐based cohort. Women/participants were self‐reported non‐Hispanic Black women from the UK Biobank (UKB). SCT status was determined based on heterozygous Glu6Val in the HBB gene. Several APOs were studied, including four previously reported SCT‐associated APOs (preeclampsia, bacteriuria, pregnancy loss, and preterm delivery), and broad conditions related to pregnancy, childbirth, and the puerperium. APOs were curated by experts’ peer review and consensus processes. Associations of SCT with APOs were tested by estimating its relative risk and 95% confidence interval (95% CI), adjusting for number of live births and age at first birth. Attributable risk proportion (ARP) and population attributable risk proportion (PARP) of SCT to APOs were estimated.

Results

Among the 4057 self‐reported non‐Hispanic Black women with pregnancy records in the UKB, 581 (14.32%) were SCT carriers. For four previously reported SCT‐associated APOs, two were confirmed at a nominal P < 0.05; relative risk (RR) was 2.39 (95% CI 1.09–5.23) for preeclampsia, and 4.85 (95% CI 1.77–13.27) for bacteriuria. SCT contributed substantially to these two APOs among SCT carriers, with attributable risk proportion estimated at 61.00% and 68.96% for preeclampsia and bacteriuria, respectively. SCT also contributed substantially to these two APOs in the population (self‐reported Black UK women), with population attributable risk proportion estimated at 18.30% and 24.14% for preeclampsia and bacteriuria, respectively. In addition, novel associations were found for seven other APOs (nominal P < 0.05).

Conclusions

SCT is significantly associated with APOs in this study and contributes substantially to APOs among self‐reported Black women in the UK. Confirmation of these findings in independent study populations is required.

In this population‐based cohort, sickle cell trait is significantly associated with adverse pregnancy outcomes (APOs), including previously known sickle cell trait‐associated APOs (preeclampsia and bacteriuria) and seven other APOs to be confirmed in future studies. Sickle cell trait contributes substantially to APOs among self‐reported Black women in the UK.

Abbreviations

APO

adverse pregnancy outcomes

ARP

attributable risk proportion

CI

confidence interval

PARP

population attributable risk proportion

RR

relative risk

SCD

sickle cell disease

SCT

sickle cell trait

Key message.

Sickle cell trait, determined by genetic data, is significantly associated with and contributes substantially to adverse pregnancy outcomes among Black women in a UK population‐based cohort.

1. INTRODUCTION

Adverse pregnancy outcomes (APOs), including preeclampsia, are major health risks for pregnant individuals during pregnancy and throughout their lifespan. ¹ Within these outcomes lies a racial disparity in which non‐Hispanic Black women are more likely to experience numerous APOs. ² While major differences in access to care and other social and economic factors likely play leading roles to the inequality, ³ host factors may also contribute to the disparity.

Sickle cell disease (SCD), a recessive disease caused by two mutations of the hemoglobin subunit beta gene (HBB), is one of the host factors that has been associated with APOs. ⁴ SCD affects approximately 100 000 Americans and is significantly more prevalent in the Black population, occurring approximately among 1 out of every 365 Black or African‐American births. ⁵ Compared with the relatively rare SCD, individuals who inherit only one mutated HBB gene, referred to as sickle cell trait (SCT) carriers, are considerably more common and occur in approximately one of 13 Black or African‐American births. ⁵ , ⁶ The estimated number of neonates born annually with SCT globally in 2010 was around 5.79 million. ⁷ Because only 30–45% of hemoglobins are affected (>80% are affected in SCD), SCT was typically considered asymptomatic previously. ⁸ However, with the availability of large and well characterized cohorts with both extensive disease and genetic information, SCT has now been associated with multiple common diseases, including venous thromboembolism (especially pulmonary embolism), diabetes, and chronic kidney disease. ⁹ , ¹⁰ In addition, several studies also reported associations of SCT with APOs, including preeclampsia, bacteriuria, pregnancy loss, preterm delivery, and low birthweight. ¹¹ , ¹² , ¹³ , ¹⁴ Nevertheless, the available evidence for the associations remains inconsistent. ¹⁵ , ¹⁶ Based on a recent and the most comprehensive review of studies on associations of SCT with APOs, ¹⁷ none of the associations was categorized as “definite/strong”, whereas several were categorized as “probable/moderate”, “possible/weak”, and “unlikely/insufficient.

The objective of this study is to provide additional pieces of evidence for associations of SCT with APOs in self‐reported Black women from the UK Biobank (UKB), with three specific aims: (1) validate the associations of SCT with previously reported APOs, (2) test novel associations of SCT with broad spectrum of APOS, and (3) estimate the attributable risk of SCT for implicated APOs in non‐Hispanic Black women. Several advantages of the UKB improve the quality of the evidence, including accurately defined SCT status by genetic data and all women (both SCT carriers and non‐carriers) were evaluated in a same manner for broad‐spectrum of APOs.

2. MATERIAL AND METHODS

Participants in this study were self‐reported Black women from the UKB, a population‐based study of 500,000 volunteers from the UK, 40–69 years old at recruitment. Extensive phenotypic and genomic information is available for each participant in the UKB, including disease diagnosis, questionnaire, biological measurements, lifestyle indicators, and biomarkers in blood and urine. ¹⁸ The availability of the International Classification of Diseases‐10 (ICD‐10) codes from inpatient and outpatient records, self‐reports, and genetic information make it possible to determine the status of SCT, diagnosis of APOs, as well as genetic background for each woman.

SCT carrier status was determined by a combination of ICD‐10 codes (ie D57.3) and heterozygous Glu6Val in the HBB gene of the whole exome sequencing (WES) data. ⁹ Because the purpose of this study is to test association of SCT with APOs, woman with ICD‐10 codes for β‐thalassemia (D56.1) and SCD (D57.0, D57.1, D57.2, D57.8) were excluded. In addition, woman with double heterozygous (heterozygous E6V and another known mutation in the HBB gene) and homozygous E6V, were excluded from this analysis. ¹⁹

For APOs, pregnancy‐related conditions were based on the “First Occurrence of Health Outcomes Defined by 3‐character ICD10 code”, which were curated data‐fields by experts’ peer review and consensus processes. These data‐fields were generated by mapping a set of diagnostic codes for a wide range of health outcomes across self‐report, primary care, hospital inpatient data, and death data. Specifically, for previous reported SCT‐associated APOs, ICD‐10 codes for preeclampsia (O14), bacteriuria (O23), pregnancy loss (O03), and preterm delivery (O60) were used. In addition, broad conditions related to or aggravated by the pregnancy, childbirth, or by the puerperium (O00–O99) were evaluated. The counts of data‐fields for conditions originating in the perinatal period (P00–P96) are limited in the current UKB data release and therefore were not included in the analysis.

The primary association tests in this study were performed in self‐reported Black women with pregnancy records. The frequency of APOs in SCT carriers and non‐carriers were counted. Associations of SCT with APOs were tested by estimating its relative risk (RR) and 95% confidence interval (95% CI) using the R stats package (v4.0.5; R Core Team 2021). RRs were estimated adjusting for number of live births, age at first birth and genetic background (top 10 principal components provided by the UKB). Attributable risk for APOs by SCT was estimated using the twoxtwo R package (v0.1.0; Nagraj 2021), including both the attributable risk proportion (ARP), which is the attributable fraction among the exposed (SCT carriers), and population attributable risk proportion (PARP), which is the proportion of cases overall in the entire population (regardless of exposure status) that can be attributed to the exposure. ²⁰ For both ARP and PARP, estimates of RR and SCT frequency from the current study were used.

2.1. Ethics statement

The UK Biobank was approved by North West – Haydock Research Ethics Committee (REC reference: 16/NW/0274; IRAS project ID: 200778) on May 10, 2016. All participants included in the study provided informed consent (https://www.ukbiobank.ac.uk/media/t22hbo35/consent‐form.pdf). Data from the UK Biobank was accessed through a Material Transfer Agreement under Application Reference Number: 50295. This study was performed in accordance with the Declaration of Helsinki.

3. RESULTS

Among the 231,029 women with a pregnancy record in the UKB cohort, 729 (0.32%) women were SCT carriers (Table 1). As expected, the SCT carrier rate was considerably and significantly higher in self‐reported Black women (14.32%) than in self‐reported White women (0.01%), Asians (0.19%) and others (2.66%), all P < 0.001. Self‐report Black women also had a significantly higher number of live births and younger age at first live birth compared with other ancestral groups.

TABLE 1.

Characteristics of study participants related to sickle cell trait (SCT) and pregnancy in the UK Biobank.

	All	Black	White	Asian	Other
Women/participants n (%)	231,029 (100%)	4,057 (1.76%)	217,755 (94.25%)	4,826 (2.09%)	4,391 (1.90%)
SCT carriers, n (%)	729/231,029 (0.32)	581/4,057 (14.32)	22/217,755 (0.01)	9/4,826 (0.19)	117/4,391 (2.66)
Number of live births, mean (95% CI)	2.14 (2.14–2.15)	2.54 (2.49–2.59)	2.13 (2.13–2.14)	2.2 (2.16–2.24)	2.37 (2.33–2.4)
SCT non‐carriers	2.14 (2.14–2.15)	2.52 (2.46–2.57)	2.13 (2.13–2.14)	2.37 (2.34–2.41)	2.18 (2.14–2.22)
SCT carriers	2.69 (2.57–2.8)	2.67 (2.54–2.8)	2.45 (2.1–2.81)	1.67 (0.51–2.82)	2.86 (2.56–3.17)
Age at first live birth, mean (95% CI)	25.91 (25.89–25.93)	23.95 (23.76–24.14)	25.94 (25.92–25.96)	26.1 (25.92–26.29)	25.95 (25.79–26.1)
SCT non‐carriers	25.91 (25.89–25.94)	23.9 (23.7–24.1)	25.94 (25.92–25.96)	25.95 (25.8–26.11)	26.18 (25.99–26.37)
SCT carriers	24.1 (23.67–24.53)	25.5 (23.24–27.76)	24.22 (23.74–24.71)	23.17 (19.89–26.45)	23.28 (22.22–24.33)

The associations of SCT with APOs were tested in self‐reported Black women with pregnancy records. For the four previously reported SCT‐associated APOs (Table 2), two were confirmed in our study at a nominal P < 0.05. SCT was associated with increased risk for preeclampsia (RR = 2.39, 95% CI 1.09–5.23) and bacteriuria (RR = 4.85, 95% CI 1.77–13.27). Higher rates of premature labor in SCT carriers than non‐carriers were also found, although the difference was not statistically significant (nominal P > 0.05).

TABLE 2.

Association of sickle cell trait (SCT) with adverse pregnancy outcomes (APOs) in Black women (UK Biobank).

	No. (%) of women			RR a (95% CI)
APOs, ICD‐10 code	Total, n = 4,057	SCT, n = 581	Non‐SCT, n= 3,476
Previous reported APOs
Fetal loss, O03	121 (2.98)	15 (2.58)	106 (3.05)	0.86 (0.47–1.58)
Preeclampsia, O14	30 (0.74)	9 (1.55)	21 (0.6)	2.39 (1.09–5.23)
Bacteriuria, O23	20 (0.49)	7 (1.2)	13 (0.37)	4.85 (1.77–13.27)
Premature labor, O60	50 (1.23)	10 (1.72)	40 (1.15)	1.71 (0.84–3.50)
Novel associated APOs
Excessive vomiting, O21	24 (0.59)	10 (1.72)	14 (0.4)	4.95 (2.04–12.00)
Malpresentation of fetus, O32	38 (0.94)	12 (2.07)	26 (0.75)	3.29 (1.64–6.62)
Maternal care for known or suspected abnormality of pelvic organs, O34	90 (2.22)	19 (3.27)	71 (2.04)	1.79 (1.08–2.96)
Fetal distress, O68	166 (4.09)	38 (6.54)	128 (3.68)	1.80 (1.25–2.60)
Perineal laceration during delivery, O70	163 (4.02)	35 (6.02)	128 (3.68)	1.88 (1.32–2.70)
Complications of the puerperium, not elsewhere classified, O90	11 (0.27)	5 (0.86)	6 (0.17)	5.88 (1.67–20.64)
Other maternal diseases classifiable elsewhere but complicating pregnancy, childbirth and the puerperium, O99	97 (2.39)	29 (4.99)	68 (1.96)	2.74 (1.75–4.30)
Any of conditions related to or aggravated by the pregnancy, childbirth or by the puerperium, O00‐O99	861 (21.22)	133 (22.89)	728 (20.94)	1.20 (1.03–1.41)

Note: RRs are bolded if Bonferroni‐corrected P < 0.05.

^a Adjusted for number of live births (no. of missing = 8), age at first birth (no. of missing = 339), and genetic background (no. of missing = 262).

Among other conditions related to pregnancy, childbirth and the puerperium, significant associations with SCT status were found for seven APOs (nominal P < 0.05), all associated with increased risk, with RR > 1 (Table 2). The strongest association was found for complications of the puerperium (RR = 5.88, 95% CI 1.67–20.64), followed by excessive vomiting (RR = 4.95, 95% CI 2.04–12.00), malpresentation of fetus (RR = 3.29, 95% CI 1.64–6.62), other maternal diseases complicating pregnancy, childbirth and the puerperium (RR = 2.74, 95% CI 1.75–4.30), perineal laceration during delivery (RR = 1.88, 95% CI 1.32–2.70), fetal distress (RR = 1.80, 95% CI 1.25–2.60), and maternal care for known or suspected abnormality of pelvic organs (RR = 1.79, 95% CI 1.08–2.96).

Attributable risks of SCT to APOs in self‐reported Black women were estimated for two confirmed APOs by both ARP and PARP (Table 3). The attributable fractions of SCT to the APOs among SCT carriers were large: ARP was 61.00% and 68.96% for preeclampsia and bacteriuria, respectively. The attributable fractions of SCT to these two APOs in the self‐reported Black UK women were also substantial: PARP was 18.30% and 24.14% for preeclampsia and bacteriuria, respectively.

TABLE 3.

Attributable risk of sickle cell trait (SCT) for adverse pregnancy outcomes (APOs) in Black women (UK Biobank).

	No. women in SCT carriers		No. women in non‐SCT carriers		ARP, % (95% CI)	PARP, % (95% CI)
Pregnancy outcome, ICD 10	APO+	APO−	APO+	APO−
Previous reported APOs
Preeclampsia, O14	9	572	21	3,455	61.00 (30.74–91.26)	18.30 (0.00–37.35)
Bacteriuria, O23	7	574	13	3,463	68.96 (40.57–97.35)	24.14 (0.00–48.46)

Abbreviations: ARP, attributable risk proportion; PARP, population attributable risk proportion.

4. DISCUSSION

In this population‐based cohort of self‐reported Black women where SCT carrier status can be accurately defined by genetic data and their broad spectrum of APOs can be comprehensively followed, we found 14.32% of women were SCT carriers, considerably higher than previously estimated (~ 8%). ⁵ , ¹⁶ , ²¹ We confirmed the associations of SCT with two of four previously reported SCT‐associated APOs (preeclampsia and bacteriuria). Notably, we showed SCT contributes substantially to these two APOs among SCT carriers and in the population (self‐reported Black UK women). These substantial attributable risk estimates are informative, considering that many other risk factors may contribute to APOs. Furthermore, our data support two previously reported SCT‐associated APOs (preeclampsia and bacteriuria) and identified seven novel APOs potentially associated with SCT. ¹⁷ The established association from this study provides evidence to the understudied research field where the association between SCT and APOs was inconclusive. Screening for SCT carriers in Black women and offering targeted prevention and intervention during pregnancy may reduce APOs and racial disparity in maternal care. ²²

A novel finding was a considerably higher SCT carrier rate among self‐reported Black women in the study than previously reported in African descendants. Most prior studies estimated an SCT carrier rate of ~ 8% in African ancestry. ¹⁶ , ²¹ The Center for Disease Control (CDC) described that about one in 13 Black or African‐American babies is born with SCT. ⁵ Several factors in detection methods, study design, and population genetics may jointly contribute to the difference. The SCT status in our study was determined primarily based on the DNA sequencing data of HBB where there are various mutations in the gene (the definitive method for diagnosing SCT and SCD). In contrast, the SCT status from many prior studies was based on hemoglobin analyses using electrophoresis, high‐performance liquid chromatography (HPLC), and isoelectric focusing (IEF). While these methods are valid and cost‐effective tests for screening SCT in the clinic, confirmation using DNA analysis is required. ²³ Various study designs may also affect the estimates of SCT carrier rate. Under‐diagnosis of SCT is possible in observational and retrospective studies where adult women did not have a hemoglobin analyses, cannot recall the results or did not collect the inpatient/outpatient records. This bias can be demonstrated by comparing results of ICD‐10 codes with the DNA test in our study. The vast majority of genetically defined SCT carriers (62%) did not have an ICD‐10 code for SCT. Finally, the underlying SCT carrier rate may differ considerably among various subregions of Africa due to a combination of selection for carriers through their survival advantage in malaria‐endemic regions and subsequent migration. ²⁴ The subregional difference in heterozygous carrier rates of Glu6Val (rs334) among African descendants is also supported by several population‐based databases (8.58% in African/African American women of gnomAD but 17.97% in African women of the 1000 Genomes Project).

Different from previous understanding, recent studies have demonstrated that SCT is not a benign condition and is associated with multiple common diseases, including pulmonary embolism, diabetes, and chronic kidney disease. ⁹ , ¹⁰ For pregnancy‐related complications, several small cohort studies have also found an increased risk for APOs in SCT, including preeclampsia, bacteriuria, pregnancy loss, preterm delivery, and low birthweight. ¹¹ , ¹² , ¹³ , ¹⁴ However, numerous additional cohort studies, all with larger sample sizes and adjustment for confounders, have failed to demonstrate an increased risk for APOs among SCT carriers. ¹⁵ , ¹⁶ , ²⁵ To date, the consensus concludes that overall evidence for the association of SCT with APOs is “weak” or “inconclusive”, and calls for large and well‐designed studies in the field. ¹⁰ , ¹⁷ , ²² Our study, characterized by accurately defined SCT status (based on HBB mutations) and comprehensive assessment of broad‐spectrum of APOs, provides an additional piece of supporting evidence for the association (confirmed two previously reported APOs and identified seven novel SCT‐associated APOs). The prospective nature of our study, where women with or without the exposure (SCT) were objectively evaluated for APOs, minimizes potential biases such as information bias, recall bias and selection bias. Furthermore, the fact that the SCT carrier status is not known for most carriers (not classified by the ICD‐10 code) further reduces the likelihood of observer bias.

It is worth noting that the increased risk for APOs among SCT carriers is in addition to the already elevated risk for APOs in women of African descent compared with other ancestral populations. In the UKB, compared with self‐reported White women, self‐reported Black women have significantly higher risk for many APOs at a nominal P < 0.05, including preeclampsia (RR = 2.89, 95% CI 1.94–4.32) and bacteriuria (RR = 3.23, 95% CI 1.95–5.37) (Table S2). The established association and substantial attributable risk of SCT to APOs in Black women have public health implications for prevention of APOs and reducing racial disparity. As an important step toward this preventive effort, a screening program for SCT among Black women may be considered to identify SCT carriers. This is necessary because, although the national newborn screening program for SCD and SCT is widely implemented in the UK and USA, the vast majority of adults may not be aware of their carrier status. The screening for SCT among childbearing adults was also proposed by Taylor et al. in a viewpoint article in JAMA in 2014. ²⁵ Prevention of APOs and their clinical management may differ for different complications. For example, low‐dose aspirin can be considered during pregnancy to prevent or delay the onset of preeclampsia. ²⁶

Several limitations in this study are noted. First, despite the large sample size of the UKB (~ 500,000), the number of SCT carriers is relatively small due to the fact that only ~ 1.6% of the UKB women are self‐reported Black. Therefore, caution should be exercised when interpreting null associations, especially for the rare complications such as venous thromboembolism (VTE).

Secondly, several reported associations in the study may not be statistically significant if multiple tests are adjusted. Considering four and 68 tests were performed for previously reported APOs and other novel APOs (Table 2, Table S1), P < 0.0125 and P < 0.007, respectively, are required at the 5% Type I error using the Bonferroni correction. Similarly, a Bonferroni‐corrected P < 0.007 is required for 72 tests of different APOs between Black and White women (Table S2). Confirmation of our findings in independent populations is needed, especially for the novel SCT‐associated APOs.

Thirdly, limited clinical information and laboratory measurements related to each APO event are available in the UKB. Similarly, other comorbidity and non‐genetic risk factors for APOs such as socioeconomic status were not adjusted in this study. Additional hospital‐based cohorts with detailed clinical and laboratory measurements for APOs among women with and without SCT are needed.

Lastly, it is noted that the prevalence rates of APOs in this study, including preeclampsia and bacteriuria, were lower than expected. While this may reflect under‐diagnosis and under‐reporting of APOs in the UKB, it unlikely to be biased toward positive associations because both SCT carriers and non‐carriers were evaluated for APOs in the same manner.

5. CONCLUSION

Data from this population‐based cohort provide suggestive evidence that SCT is significantly associated with APOs and contributes substantially to APOs among self‐reported Black women in the UK. If confirmed in other studies, these findings may support the need for screening SCT carriers in Black women for targeted prevention and intervention to reduce APOs and racial disparity in maternal care.

AUTHOR CONTRIBUTIONS

JH: Drafting of manuscript and statistical analysis. ASR: Drafting of manuscript and administrative, technical, or material support. ZS: Statistical analysis. MSC: Conception and design, obtaining funding, supervision. JX: Conception and design, acquisition of data, drafting of manuscript, obtaining funding, and supervision. All authors: Analysis and interpretation of data and critical revision of the article for important intellectual content.

CONFLICT OF INTEREST STATEMENT

None declared.

Supporting information

Tables S1S2

Hulsizer J, Rifkin AS, Shi Z, et al. Association of sickle cell trait with adverse pregnancy outcomes in a population‐based cohort. Acta Obstet Gynecol Scand. 2023;102:1100‐1105. doi: 10.1111/aogs.14622

DATA AVAILABILITY STATEMENT

The data used in this study are available in the UK Biobank, a publicly available repository. Data was accessed through a Material Transfer Agreement under Application Reference Number: 50295. For additional information, please feel free to contact the corresponding author, Jianfeng Xu, DrPH.