Pregnancy in Sickle Trait: What we Know and What we Don’t

Summary

Sickle cell trait (SCT) is the carrier state for sickle cell disease that results from the HBB rs334 missense mutation (p.Glu6Val) in the β-globin chain of haemoglobin. While not associated with any impact on life expectancy, it has been established that SCT is associated with an increased risk of both venous thromboembolism (and in particular, pulmonary embolism) and chronic kidney disease. It is largely unknown what short- or long-term effect, if any, pregnancy has upon the risk or outcomes of these disorders. In addition, SCT has been linked with a variety of adverse outcomes in pregnancy, ranging from maternal complications such as elevated risk of bacteriuria to potentially life-threatening entities such as pre-eclampsia and prematurity. In these scenarios also, no clear association with SCT has been established. Given the high worldwide prevalence of SCT, further studies addressing these issues are warranted.

Introduction

Sickle cell trait (SCT), the heterozygous state for sickle cell disease (SCD), is a highly prevalent entity that is present in 7–8% of African-Americans and approximately 300 million individuals worldwide. It is particularly prevalent in sub-Saharan Africa (where carrier rates may be as high as 40% in certain regions), central India and the Arabian Peninsula. In 1954, Allison proposed that SCT conferred protection against malaria, both in terms of frequency of infection, as well as severity of disease (Allison, 1954). This sentinel observation has been largely upheld by later studies, although the protection appears to be more important in mitigating severity, rather than frequency of infection (Williams et al, 2005). Although SCT is generally regarded as a benign carrier state that has little to no impact on quality or duration of life, many studies over the past half century have set out to determine whether SCT is associated with a variety of medical complications. Exhaustive lists of such complications were reported (Sears, 1978; Johnson, 1982), but it is now generally accepted that many early studies were seriously flawed for a variety of reasons, including mis-diagnosis of sickle trait (which was often confused with sickle cell disease), publication bias, confounding, and grossly inadequate study design and/or statistical power (Gima & Bemis, 1975). The net result was a realization of the potential ethical and societal harm caused by much of this research in the African American community. The history and implications of this perspective have been discussed elsewhere (Naik & Haywood, 2015), and will not therefore be reiterated here. However, with the growing availability of suitable population cohort studies, and in response to an appeal from the National Institutes of Health (Goldsmith et al, 2012) and the Centers for Disease Control (Grant et al, 2011), a more focused and evidence-based approach to the study of possible disease associations with SCT has emerged. These studies have changed current perceptions regarding the complications that may or may not be associated with SCT, with greater evidence for some, and less evidence for others. The field is continuing to evolve with the publication of new, higher quality data, but with respect to the implications of SCT on pregnancy, it is clear that much remains to be done.

When is testing for sickle hemoglobinopathy performed and why?

Although newborn screening for sickle hemoglobinopathies targeted at the African-American population has been in place in more than 30 States since the early 1970s, universal newborn screening was not fully enacted throughout the United States until 2006. In Europe, national newborn screening programs are also operational in France, Great Britain, Spain, The Netherlands and Malta with several other countries, including developing nations, having regional or pilot programs underway (Daniel et al, 2019). A novel approach to newborn screening for SCD has been reported in several African countries by analysing neonatal blood spots intended to screen for HIV (Ndeezi et al, 2016). The primary purpose of newborn screening is to identify neonates with SCD in order to initiate proven life-saving antibiotic prophylaxis in the early years of life. Despite the fact that SCT is also ‘inadvertently’ detected by newborn screening, several studies have documented that the notification of individuals found to be heterozygous carriers of the sickle gene at birth is frequently sub-optimal and is often accompanied by very limited counselling as to the reproductive implications (Kavanagh et al, 2008; Kladny et al, 2011).

In the United States, targeted screening specifically aimed at detecting SCT also occurs in student athletes participating in competitive Collegiate sports (Eichner, 2010; Key et al, 2015) and in some, but not all branches of the armed services (Kark et al, 1987; Key et al, 2015). In these situations, the rationale for screening has been to prevent exertion-related injury and death, but this approach has been controversial. In the military, where screening for SCT was initiated in the 1970s, the risk for exertional complications has been identified to be greatest during basic military training. However, the necessity for SCT screening has been challenged by the demonstration that universal implementation of simple precautions aimed at preventing exertional rhabdomyolysis (including the allowance of time for heat acclimatization, mandatory hydration breaks and staff education about danger symptoms) can mitigate any increased risk of exertional sudden death related to SCT. In contrast, screening for sickle trait among college athletes is a relatively recent practice that was adopted by the National Collegiate Athletic Association (NCAA) in 2010 after a highly publicized death of a student athlete with SCT (Tarini et al, 2012). The need for this has also been criticized, both on the basis that it may lead to stigmatization of affected individuals, as well as the above-mentioned rationale that universal application of simple preventive measures can largely obviate the increased risk of adverse outcomes with extreme exertion (Bonham et al, 2010).

In addition to the above scenarios, the 2007 American College of Obstetrics and Gynecology (ACOG) practice guideline recommends pre- and post-conception screening for hemoglobinopathies in women of African, Southeast Asian, or Mediterranean descent. If both parents are then found to be hemoglobinopathy carriers, the guidelines recommend referral for genetic counselling to review prenatal testing and reproductive options (ACOG Committee on Obstetrics, 2007). Surveys of practicing obstetricians have reported that the vast majority (approximately 97%) of providers who answered the survey do screen individuals of African descent for SCT/SCD, but far fewer women of Mediterranean or Asian descent (50% and 30%, respectively) are screened for hemoglobinopathies (Azonobi et al, 2014).

Reviews addressing complications associated with SCT

Given the limited number of adequately designed and powered studies on pregnancy-related complications in SCT, we evaluated recent review articles pertaining to proposed medical complications of SCT. In 2009, Tsaras and colleagues published a narrative review in which they proposed that disease associations with SCT could be grouped into ‘Definite’, ‘Probable’, ‘Possible’ and ‘Unlikely’ (Tsaras et al, 2009). This categorization was analogous to the matrix used by Sears approximately 20 years earlier (Sears, 1978). More recently, several additional reviews have appeared in which the authors provided further literature updates and assessed the evidence linking various complications to SCT. In a few of these reviews, the conclusions were based on a systematic literature review with a defined methodologic approach (Jans et al, 2010; Noubiap et al, 2018; Naik et al, 2018), but in most instances, no formal evaluation strategy was described. The accompanying Table summarizes the major conclusions of these reviews, some of which were comprehensive in scope, while others focused on a limited number of complications. In many cases, we assigned a level of confidence to the strength of the association based on what was implied (but not always stated) in the narrative. Thus, in some instances, we may have misinterpreted author intent. Nonetheless, a few observations from the trends summarized in this fashion seem appropriate. First, there appears to have been an evolution in the perceived strength of evidence for some associations, such as venous thromboembolism and chronic kidney disease, while the association with other complications may have become less certain. Of course, it should be noted that many complications are not readily evaluable in a population cohort or case-control study. Second, relatively little attention has been paid to the development of an evidence-based consensus on the strength of association of pregnancy-related complications with SCT. Indeed, in a recent systematic literature review, pregnancy-related complications of SCT were notably omitted (Naik et al, 2018). Thus, the remainder of this article will focus on the ‘known unknowns’ surrounding pregnancy complications of SCT. As such, one purpose of this review is to suggest possible future research priorities.

Effects of pregnancy on complications known to be associated with SCT

While evaluating the complications that have been most convincingly associated with SCT (Table), we reasoned that in some cases, pregnancy might reasonably be expected to magnify the strength of the association (e.g. venous thromboembolism), whereas there may not be an a priori reason to expect the risk to be amplified in pregnancy for other complications.

Table 1:

Complications associated with sickle cell trait in review articles;

Complication	Definite/Strong Association	Probable/ModerateAssociation	Possible/Weak Association	Unlikely/Insufficient evidence
Renal
CKD*	(Naik et al, 2018; Xu & Thein, 2019)	(Naik & Haywood, 2015; Key et al, 2015; Naik et al, 2018)	(Key & Derebail, 2010)	(Sears, 1978)
ESRD*		(Naik 2018)	(Key & Derebail, 2010; Thoreson et al, 2015; Key et al, 2015; Xu & Thein, 2019)	(Sears, 1978)
Albuminuria	(Xu & Thein, 2019)	(Naik & Haywood, 2015; Naik et al, 2018)	(Key et al, 2015)
Renal medullary cancer	(Tsaras et al, 2009; Goldsmith et al, 2012; Xu & Thein, 2019)	(Key & Derebail, 2010; Naik et al, 2018)
Haematuria	(Tsaras et al, 2009; Goldsmith et al, 2012)	(Sears, 1978; Thoreson et al, 2015; Key et al, 2015; Naik et al, 2018; Xu & Thein, 2019)
Renal papillary necrosis	(Tsaras et al, 2009; Key et al, 2015)	(Key & Derebail, 2010; Thoreson et al, 2015; Key et al, 2015; Xu & Thein, 2019)	(Sears, 1978)
Hyposthenuria	(Tsaras et al, 2009; Key et al, 2015; Xu & Thein, 2019)	(Sears, 1978; Key & Derebail, 2010; Thoreson et al, 2015)
Vascular
Splenic infarction	(Tsaras et al, 2009)	(Sears, 1978; Naik et al, 2018)	(Xu & Thein, 2019)
Exertional rhabdomyolysis	(Tsaras et al, 2009; Naik et al, 2018)	(Key & Derebail, 2010; Key et al, 2015; Xu & Thein, 2019)	(Sears, 1978; Goldsmith et al, 2012; Naik & Haywood, 2015)
Exercise-related death	(Tsaras et al, 2009)	(Key & Derebail, 2010; Key et al, 2015)	(Goldsmith et al, 2012; Naik & Haywood, 2015; Naik et al, 2018; Xu & Thein, 2019)	(Sears, 1978)
Venous thromboembolism*	(Naik et al, 2018; Xu & Thein, 2019)	(Tsaras et al, 2009; Key & Derebail, 2010; Thoreson et al, 2015; Naik & Haywood, 2015; Naik et al, 2018)	(Goldsmith et al, 2012)	(Sears, 1978)
Proliferativeretinopathy			(Sears, 1978; Tsaras et al, 2009; Xu & Thein, 2019)
Stroke			(Thoreson et al, 2015)	(Sears, 1978; Tsaras et al, 2009; Goldsmith et al, 2012; Naik et al, 2018; Xu & Thein, 2019)
Pregnancy
Pre-eclampsia*			(Jans et al, 2010)	(Sears, 1978; Naik & Haywood, 2015; Xu & Thein, 2019)
Foetal loss/demise*		(Tsaras et al, 2009)		(Sears, 1978; Xu & Thein, 2019)
Low birth weight*		(Tsaras et al, 2009)	(Sears, 1978)	(Sears, 1978; Jans et al, 2010; Naik & Haywood, 2015)
Premature labour*				(Jans et al, 2010)
Asymptomatic bacteriuria inpregnancy*		(Sears, 1978; Jans et al, 2010)	(Tsaras et al, 2009; Goldsmith et al, 2012; Naik & Haywood, 2015)	(Xu & Thein, 2019)

CKD = chronic kidney disease; ESRD = end stage renal disease;

^{* =}complication addressed in this article

A]. Venous thromboembolism

The association between SCT and venous thromboembolism (VTE) has now been reported by many groups. Austin and colleagues reported a 2-fold increase in risk for venous thromboembolism, and a 4-fold risk specifically for pulmonary embolism (PE) among black patients in a prospective case-control study (Austin et al, 2007). A more recent prospective cohort study also demonstrated an approximately 2-fold increased risk for PE among individuals with SCT (Folsom et al, 2015). In a population-based cohort study of General Practices in the United Kingdom, Little et al reported an increased association with VTE (OR 1.78, 95% CI 1.18–2.69), and in particular PE (OR 2.27, 95% CI 1.17–4.39) (Little et al, 2017). In addition, a retrospective cohort study reported a modestly increased risk for PE with a relative risk of 1.37 (95% CI, 1.07–1.75) (Bucknor et al, 2014). The implication of these findings was summarized in recent systematic reviews which concluded that there is strong evidence demonstrating a positive association of PE with SCT (Naik et al, 2018; Noubiap et al, 2018).

An open question is whether the established risk of VTE in SCT is amplified by pregnancy or indeed, the administration of oestrogen-containing therapies. We reported that women carriers of the sickle gene on hormonal oral contraceptive pills (OCP) had an odds ratio for VTE of 12.1 (95% CI, 2.8–52) compared to non-carriers who were not taking OCP (Austin et al, 2009). Further analysis suggested that there may be a synergism in the risk for VTE in individuals with SCT and hormonal OCP use (Austin et al, 2009). However, this conclusion should be regarded as preliminary, given the small number of thrombotic events. To our knowledge, no other study has specifically addressed the interaction of OCP use and SCT with respect to any enhanced risk for VTE.

Pregnancy and the postpartum period are associated with a 6-fold increased risk for VTE, and it has been estimated that approximately 50% of pregnancy-related VTE are linked with an inherited thrombophilia (Marik & Plante, 2008). Several studies have evaluated whether pregnant women with SCT are at increased risk for VTE. A retrospective cohort study of 679 pregnant black women with SCT compared to 5,464 black non-carriers did not observe a higher incidence of VTE during pregnancy and the postpartum period. Specifically, the VTE incidence was 0.44% for pregnant women with SCT compared to 0.49% for pregnant non-carriers, with a RR of 0.94 (p = 0.92) (Pintova et al, 2013). The authors noted that based on the low incidence of VTE, they had about 79% power to detect a 3-fold increase in risk. A larger retrospective cohort study of 2,037 SCT subjects and over 20,000 controls showed a non-significant relative risk for pregnancy-related VTE of 1.6 (95 % CI, 0.5–5.5). The authors calculated that a sample size of 110,000 would be required for 80% power to detect a relative risk of 1.67 (Porter et al, 2014). A meta-analysis combining these two studies (and thus likely underpowered), showed a RR of 0.9 (95% CI, 0.28–2.89) for pregnancy-related VTE in patients with SCT compared to controls (Noubiap et al, 2018). Therefore, on the basis of the existing data, there is no conclusive evidence that SCT confers an enhanced increased risk for VTE in pregnancy. However, this topic warrants further investigation in adequately powered cohort studies.

B]. Chronic kidney disease (CKD)

A convincing body of data has demonstrated an association between SCT and renal disease. SCD has long been known to be associated with renal impairment, thought to be a result of decreased oxygen tension and acidosis in the renal medulla leading to intra-medullary erythrocyte sickling and vaso-occlusion. Renal complications associated with SCD include haematuria, impaired urinary concentration, hyperfiltration, albuminuria, glomerular hyperfiltration, CKD, and end stage renal disease (ESRD) (Naik & Derebail, 2017). There is longstanding evidence that SCT is associated with impaired urinary concentration, but of a lesser degree and presenting later in life compared to SCD (Cochran, 1963; Francis & Worthen, 1968). Also, analogously to SCD, SCT has been associated with haematuria, renal papillary necrosis, acute renal failure, renal medullary carcinoma, exertion-related rhabdomyolytic injury, CKD, and possibly ESRD (Ataga & Orringer, 2000; Naik & Derebail, 2017).

One retrospective cohort study showed a modestly increased risk of CKD among individuals with SCT with a RR of 1.13 (95% CI, 1.03–1.23), after controlling for comorbidities such as obesity and diabetes (Bucknor et al, 2014). In a much larger prospective cohort study combining data from multiple population studies, individuals with SCT were found to have an increased risk for CKD with an OR of 1.57 (95% CI,1.34–1.84) as well as increased risk for more rapid decline in eGFR per year (OR 1.32, 95% CI 1.07–1.61). Further subgroup analysis suggested that the risk for eGFR decline was primarily in SCT individuals without baseline CKD (Naik et al, 2014). In support of these data, several cohort studies have shown a 2-fold or greater increased prevalence of CKD in patients with SCT (Naik et al, 2017; Kramer et al, 2017). These studies are further bolstered by a 2011–2014 longitudinal study of 45,901 black U.S. army soldiers that reported a 2-fold increased risk for CKD and almost a 2-fold increased risk for acute kidney injury among individuals with SCT (Hu et al, 2019).

The evidence for an association between SCT and ESRD has been less clear. Derebail reported that SCT was twice as prevalent in black patients with ESRD on renal replacement therapy compared to the local community prevalence of SCT (15% vs 7%) (Derebail et al, 2010). In contrast, a case control study of patients with ESRD did not detect an association with SCT (Hicks et al, 2011). Subgroup analysis of the previously noted large multi-cohort population-based study did not detect an association between SCT and ESRD (Naik et al, 2014), although ESRD data were limited to only two of the included cohorts. However, the largest prospective cohort study to date detected a hazard ratio of 2.03 (95% CI, 1.44 to 2.84) in the incidence of ESRD per year in black patients with SCT compared to black noncarriers (Naik et al, 2017). Finally, the African American Study of Kidney Disease and Hypertension (AASK) (Sood et al, 2019), and the Chronic Renal Insufficiency Cohort (CRIC) study (Derebail et al, 2019) both recently reported that the presence of SCT did not adversely influence the rate of eGFR decline or adverse renal outcomes. Thus, even while allowing for differences in the cohorts and study design, the influence of SCT on the rate of GFR decline in those with established CKD has not been demonstrated.

Finally, there are no data whether pregnancy increases the risk for renal disease or accelerates the rate of progression of CKD in SCT. In pregnant women with CKD in the absence of SCT, there are conflicting data on the impact of pregnancy on renal function. Some studies have suggested that the risk of progression increases with each stage of CKD, while others show no impact. Overall however, data on pregnant women with advanced CKD or ESRD are limited (Hui & Hladunewich, 2019), and there is no information in the sub-population of women with SCT.

Pregnancy-specific complications in SCT

A]. Mortality

An analysis of pregnancy-related deaths in the United States between 2007 and 2016 by the Centers for Disease Control (CDC) reported a mortality ratio for black women that was 2.8–3.3 times that of non-Hispanic white women (Petersen et al, 2019). A systematic review of pregnancy complications in SCD demonstrated an increased risk for maternal mortality, as well as preeclampsia, stillbirth, preterm delivery, and small for gestational age infants (Oteng-Ntim et al, 2015). The effect of SCT on mortality related to pregnancy has not been studied, likely due to the relatively small number of perinatal maternal deaths (approximately 700 per year in the United States).

B]. Bacteriuria

A possible association between pregnancy-related bacteriuria and SCT was proposed years ago and has subsequently been studied by several groups. Early studies reported an increased prevalence of asymptomatic bacteriuria and pyelonephritis among individuals with SCT (Whalley et al, 1964; Baill & Witter, 1990). A more recent study was unable to detect an association between SCT and asymptomatic bacteriuria during pregnancy, but it reported more than a three-fold increase (2.4% vs 0.7%, p = 0.03) in the prevalence for pyelonephritis in SCT (Thurman et al, 2006). The most recent study on this topic failed to find an association between SCT and asymptomatic bacteriuria or pyelonephritis (Tita et al, 2007). Mechanistically, it has been proposed that chronic renal papillary necrosis due to microinfarction could lead to an increased susceptibility to genitourinary tract infection in SCT (Tsaras et al, 2009). In the past, the American College of Obstetrics and Gynecology (ACOG) recommended routine screening for asymptomatic bacteriuria during each trimester in pregnant SCT patients (ACOG Committee, 1996), but the most recent guidelines do not address this recommendation (ACOG Committee on Obstetrics, 2007). However, an informal survey of several large academic obstetric units suggests that the practice of screening for bacteriuria in pregnant women with SCT continues (unpublished data).

C]. Pregnancy-related hypertensive disorders

A prospective study of pregnant black women with SCT was the first to demonstrate a significantly increased incidence of preeclampsia (24.7% vs 10.1%, p < 0.0001) compared to black women noncarriers (Larrabee & Monga, 1997). Since then, two large retrospective cohort studies and one retrospective case-control study failed to show an association between SCT and an increased risk of preeclampsia (Stamilio et al, 2003; Tita et al, 2007; Taylor et al, 2006). In a broader scope retrospective cohort study of over 25,000 women in the military from 1993–2013, the incidence of pregnancy-related hypertensive disorder (PRHD, which included gestational hypertension, pre-eclampsia, and eclampsia) was compared in 5,004 individuals with SCT with 20,016 matched controls. The authors concluded that SCT is associated with an increased risk of PRHD with a hazard ratio of 1.43 (95% CI, 1.3–1.58), corresponding to an attributable risk of 30.6% (95% CI, 22.3%−37.6%) for PRHD from SCT. However, with pre-eclampsia and eclampsia comprising only a small portion (2%) of the PRHD outcomes, no association between SCT and pre-eclampsia or eclampsia was observed (O’Hara et al, 2019). These findings suggest that additional studies are needed to clarify the strength of the association between SCT and PRHD, as well as determine whether this association extends to pre-eclampsia and eclampsia. As previously discussed, given the association between SCT and acute kidney injury (Hu et al, 2019), as well as CKD (Hu et al, 2019; Bucknor et al, 2014; Naik et al, 2014) and possibly also ESRD (Naik et al, 2017), it is conceivable that pregnancy-related hypertensive disorders in SCT may contribute to an increased the risk of chronic renal disease, although no studies have directly addressed this possibility.

D]. Pregnancy loss

Perhaps the most worrisome association between SCT and pregnancy outcomes is the risk for pregnancy loss. An older retrospective cohort study of 500 individuals with SCT did not find an increased risk for perinatal mortality (Whalley et al, 1963). A more recent small retrospective case series of 131 pregnant patients with SCT showed an intrauterine foetal death (IUFD) rate of 8% (compared to 5% baseline at institution) and identified one neonatal death (Taylor et al, 2008). However, the significance of this finding is unclear. The same group, in a retrospective case-control study of 180 patients at a single institution, found an almost three-fold rate (9.7% vs 3.5%, p = 0.015) of foetal death after first-trimester viability amongst pregnant patients with SCT compared to ethnicity-matched noncarrier controls (Taylor et al, 2006). Histologic examination of the placentas of pregnant mothers (regardless of pregnancy outcome) showed evidence of more amniotic fluid infection (50% vs 18%) in mothers with SCT compared to noncarriers. In addition, red cell sickling in the intervillous and decidual vessels of the placentas was observed, with evidence of placental infarctions and retroplacental haemorrhages. In contrast, a more recent and larger retrospective cohort study of over 1,800 pregnant individuals with SCT did not find an association between maternal SCT and perinatal morality (RR 0.7, 95% CI 0.5–1.0) (Tita et al, 2007).

E]. Preterm delivery

Studies evaluating an association between SCT and other perinatal outcomes have generally been less controversial. Several have evaluated but failed to identify an association between SCT and preterm delivery (Tita et al, 2007; Baill & Witter, 1990; Bryant et al, 2007; Whalley et al, 1963). In fact, a possible protective effect of SCT on risk for preterm delivery has been reported (OR 0.15, 95% CI 0.05–0.49) (Bryant et al, 2007). The case-control study by Taylor et al did find a significantly shorter mean duration of pregnancy in individuals with SCT compared to noncarriers (233 vs 255 days; p < .001), suggesting increased risk of preterm delivery, although this was not measured directly and thus does not provide conclusive evidence for such an association (Taylor et al, 2006).

F]. Low birthweight

A few studies have evaluated risk for low birthweight and SCT. Taylor et al demonstrated that SCT was associated with decreased mean birth weight (2,114 g vs 2,672 g, p < 0.001) (Taylor et al, 2006). However, subsequent studies failed to confirm this association (Tan et al, 2008; Tita et al, 2007).

Overall, the weight of evidence currently suggests that there are no compelling data that adverse perinatal outcomes are associated with the presence of SCT.

Concluding remarks

Although the possible impact of the sickle cell trait on pregnancy and pregnancy-related outcomes has been the topic of many investigations over several decades, the quality of evidence generally remains low and precludes definitive statements about any associations. Two general approaches to conceptualizing this risk have been addressed in this review. First, studies of the widely accepted associations of SCT in the non-pregnant state, particularly VTE and renal disease, are needed to determine whether the risk is modified by pregnancy. Second, the effect of SCT on pregnancy outcomes, with a possible prioritization of pregnancy-related hypertensive disorders, seems to be warranted. Because these complications are relatively infrequent, only very large-scale population cohort studies are likely to be able to address the questions in a meaningful manner.