Pregnancies and Neonatal Outcomes in Patients with Sickle Cell Disease (SCD): Still a (High-)Risk Constellation?

Background: This monocentric study conducted at the University Hospital of Essen aims to describe maternal and fetal/neonatal outcomes in sickle cell disease (SCD) documented between 1996 to 2021 (N = 53), reflecting the largest monocentric analysis carried out in Germany. Methods/Results: 46 pregnancies in 22 patients were followed. None of the patients died. In total, 35% (11/31) of pregnancies were preterm. 15 pregnancies in eight patients were conceived on hydroxycarbamide (HC), of which nine had a successful outcome and three were terminated prematurely. There was no difference regarding the rate of spontaneous abortions in patients receiving HC compared to HC-naive patients prior to conception. In patients other than HbS/C disease, pregnancies were complicated by vaso-occlusive crises (VOCs)/acute pain crises (APCs) (96%, 23/24); acute chest syndrome (ACS) (13%, 3/24), transfusion demand (79%, 19/24), urinary tract infections (UTIs) (42%, 10/24) and thromboembolic events (8%, 2/24). In HbS/C patients complications included: VOCs/APCs (43%, 3/7; ACS: 14%, 1/7), transfusion demand (14%, 1/7), and UTIs (14%, 1/7). Independent of preterm deliveries, a significant difference with respect to neonatal growth in favor of neonates from HbS/C mothers was observed. Conclusion: Our data support the results of previous studies, highlighting the high rate of maternal and fetal/neonatal complications in pregnant SCD patients.


Introduction
Over the last years, sickle cell disease (SCD)-related maternal and neonatal mortality has decreased significantly. This can be ascribed to the availability of evidence-based clinical practice guidelines (CPGs) and preventive measures, antenatal counseling and implementation of newborn screening [1,2]. However, pregnancy in patients with SCD still remains associated with both maternal morbidity and fetal/neonatal morbidity and mortality [3,4]. The observation of increased maternal morbidity is supported by the fact that care in SCD throughout pregnancy and in the postpartum period is mainly restricted to supportive measures in addition to transfusion therapy. The reason is that therapeutic modalities require immediate discontinuation due to the fear of possible complications that could particularly endanger fetal well-being [5][6][7]. Therefore, it is not surprising that sickle cell-associated complications (e.g., anemia, vaso-occlusive crises (VOCs), including acute chest syndrome (ACS), pre-existing cardio-pulmonary and renal complications) often deteriorate in pregnant SCD patients, resulting in frequent inpatient admissions for medical care. In addition, infectious complications, thromboembolic events (TEs), and preeclampsia can further complicate pregnancy and the postpartum period [8].
Erythrocyte sickling may contribute to micro-vascular placental pathology, increasing the risk for decreased placental circulatory blood flow, (acute) fetal hypoxemia, and placental thromboses, all of which are closely associated with a high incidence of perinatal complications. Those can be spontaneous abortions, stillbirths, or an increased rate of preterm deliveries with low gestational weight in neonates. Often, there is an imperative need for cesarean deliveries [8].
The prevalence of SCD and the number of affected children born in Germany is presently unknown. Notwithstanding that reliable epidemiological data are lacking, it is estimated that approximately 3000 to 5000 affected children and adults currently live in Germany. According to conservative estimates, approximately 70 to 150 affected children are born in Germany each year. Furthermore, due to increased immigration within the past years, more children affected by SCD will be born within the next years [9].
Here we report pregnancy outcomes in patients with SCD registered at the Department of Hematology and Stem Cell Transplantation, University Hospital Essen, Germany. To the best of our knowledge, this is the first large German monocentric observational study addressing this clinically important topic in a considerable number of patients.

Study Design and Participants
This is a single-center observational study of pregnancies in women with homozygous SCD (HbSS) or compound heterozygous states, such as HbS/C, HbS/β-thalassemia, or HbS/O-Arab (±co-inheritance of heterozygous α-thalassemia, if tested) who were in part or entirely monitored at the Department of Hematology and Stem Cell Transplantation, University Hospital Essen, Germany, allowing a prospective follow-up. Pregnancies before referral to our department were recorded retrospectively, based on patients' information and medical reports. Data from patients delivering outside of Germany were generally not available. Carrier states for SCD were excluded. The study was approved by the Ethics Committee of the University of Duisburg-Essen and conducted in accordance with the Declaration of Helsinki.

Disease-Related Definitions, Methods, and Treatments
SCD was diagnosed according to international standards at the Hemoglobin Laboratory of the University Hospital Ulm and starting from 2017 at the Medical Care Center Dr. Eberhard and Partner Dortmund via molecular globin gene genetic analyses by polymerase chain reaction (PCR), sequencing and/or multiplex ligation-dependent probe amplification (MLPA).

Sickle Cell-Associated Definitions
An acute pain crisis (APC) or uncomplicated VOC was defined as an acute onset of pain for which there was no other medical explanation other than vaso-occlusion and which required analgetic treatment (enteral or parenteral), irrespective if the event required hospitalization or was managed at home. ACS or splenic sequestration were referred to as complicated VOCs. ACS was defined based on the finding of a new pulmonary infiltrate in the presence of other signs and symptoms: chest pain, a temperature of ≥38.5 • C, tachypnea, wheezing or cough. Splenic sequestration was defined on the basis of left upper quadrant pain, an enlarged spleen, and an acute decrease in hemoglobin (Hb) concentration (e.g., a decrease in Hb of 2 g/dL from baseline).

Pregnancy-Related Definitions
Spontaneous abortion and stillbirth were defined as unsuccessful pregnancy outcomes before or after gestational week (GW) 22. Preterm delivery was defined as delivery before GW 37. Preterm premature rupture of membranes (PPROM) was classified if PROM occurred before GW 37.

Treatments
Treatment with hydroxycarbamide (HC) was recommended according to international guidelines and prescribed, if desired. In all HC-treated patients, therapy was discontinued with establishment of pregnancy due to possible teratogenicity [10,11]. Prior to pregnancy, none of the patients were treated with crizanlizumab. In patients allowing for pre-conceptional care (PCC), HC was aimed to be discontinued at least three months before conception under close follow-up, if feasible.
In patients with past pregnancies not yet treated at our center, extended RBC phenotyping was performed on the date of their first visit. RBC transfusions were restricted to patients with symptomatic anemia, a peri-or antenatal Hb concentration <7 g/dL, or in pregnant patients with SCD-or fetal-related complications. A RBC exchange (RCE) was recommended for the treatment of severe VOCs (i.e., severe pain crises and/or ACS). Low-molecular-weight-heparins (LMWHs) were prescribed to patients with prior TEs, in the event of inpatient admission, and were routinely applied in the postpartum period. The duration of thromboprophylaxis was at the discretion of the treating physician, based on the individual risk profile and mode of delivery.

Statistical Analysis
Due to the small sample size, statistical analysis was not carried out except for maternal and neonatal outcomes using paired Student's t-test and two-tailed Mann-Whitney U test. Differences between groups were considered to be significant at a p value of < 0.05. Statistical analyses were performed with GraphPad Prism 6.
At first presentation, alloimmunization to RBC antigens was observed in two patients (Pat. 1: anti-K and 16: anti-M, and -E). Both of which had a medical history of repeated RBC transfusions.
Overall, patient age was independent with respect to pregnancy outcome (successful versus unsuccessful), irrespective (prior) HC exposure: overall: p = 0.7477; severe genotypes: p = 0.4697. Preterm deliveries (35% (11/31)) were secondary to premature rupture of membranes (PROM), HELLP syndrome (hemolysis, elevated liver enzymes, low platelets), placenta insufficiency or abruption, pathologic fetal cardiotocographs (CTGs) and doppler indices, amniotic infection syndrome (AIS), maternal urosepsis, fetal subdural hematoma, spontaneous onset of labor in a multigravida patient, or due to a combination of two causes. Medical data concerning the reasons for preterm deliveries in patients giving birth outside Germany were in general not available.
The histological evaluation of the placenta from the one patient who experienced a stillbirth was not available. In patients suffering from recurrent (≥2) miscarriages (N = 3), only one (Pat. 19) was tested for other causes of loss of pregnancy. In another patient (Pat. 11), the diagnosis of delta-storage pool disease was confirmed during her fourth pregnancy.

Gestational-, Peri-, and Postpartum Phase in Patients with Severe Genotypes (HbSS, HbS/β-Thalassemia, and HbS/O-Arab Disease)
In patients with severe genotypes and successful pregnancies (N = 24), APCs or VOCs were described by all patients during gestation, except for one carrying the HbSS genotype (96%, 23/24). These events predominately occurred during the second and third trimester, requiring in part (recurrent) inpatient admissions (54%, (13/24)) for (parenteral) analgetic treatment, and/or RBC transfusion support secondary to VOCs, or because of (infectiousrelated) hemolytic and/or symptomatic anemia. Throughout the observation time, two complicated VOCs (ACS) were observed in one HC-naive HbSS patient (Pat. 4) during her two successive pregnancies (GW 15 and 29 + 3) necessitating manual RCE. An additional ACS was observed in the postpartum period in a second HbSS patient (Pat. 11) (ACS rate in patients with severe disease-associated genotypes: 13% (3/24)).  RBC transfusions were necessary in 75% (18/24) of the patients. One HC-naive patient treated in Italy presenting with recurrent VOCs (including ACS) prior to pregnancy, received serial prophylactic (exchange) blood transfusions (SP(E)BTs) for four years which were continued throughout pregnancy and thereafter every three weeks (Pat. 15). In another pre-conceptionally HC-treated patient (Pat. 8), SPBTs (every three weeks, starting from GW 20) were indicated due to suspicious doppler indices and symptomatic/hemolytic anemia.
One possible delayed hemolytic transfusion reaction (DHTR) with diagnosis of an alloantibody against the S-antigen (DD hyper-hemolytic VOC) despite transfusion of extended-phenotyped and matched RBCs was observed in one HbSS patient (Pat. 16, GW 29) already expressing irregular antibodies against erythrocytic antigens (anti-E, anti-M). To avoid further RBC transfusions, resumption of HC (20 mg/kg BW) in combination with erythropoietin (EPO) in GW 34 prior to delivery in GW 35 + 2 was indicated. Due to transfusion-dependent anemia in her subsequent pregnancy despite continuation of EPO, resumption of HC (15 mg/kg BW; GW 30) was mandatory prior to delivery (GW 34 + 4).
In the presence of HC, breastfeeding was avoided in this patient.
Overall, thromboembolic complications were seen in two patients (left DVT (Pat. 19; GW 26) and pulmonary embolism in the context of ACS during puerperium (Pat. 11)). Infectious complications were restricted to (recurrent) urinary tract infections (UTIs) and documented in ten out of the 24 successful pregnancies (42%), including maternal urosepsis in third trimester (GW 28) in one patient (Pat. 11).
The peripartum complications included P(P)ROM in five pregnancies, possibly associated with or occurred independently to placental abruption, cervical insufficiency, or amniotic infection syndrome (AIS). One patient suffered from HELLP syndrome. VOCs/APCs and/or RBC transfusion demand in the peripartum period were observed in five of the pregnancies (21%). A caesarean section (CS) secondary to pregnancy-, including activephase arrest and/or pathologic doppler indices (N = 4), or SCD-related complications was mandatory in 42% (10/24) of patients for whom medical data were available, resulting in an overall CS rate of 67% (16/24). In the postpartum period, RBC transfusions were required in four patients. Postpartum cardiomyopathy was seen in one patient (Pat. 11).
Following delivery, HC was re-initiated, if breastfeeding was not desired. In breasfeeding patients, treatment was re-initiated as soon as weaning was ensured.

Gestational-, Peri-, and Postpartum Phase in Patients with HbS/C Disease
In the four patients with HbS/C disease, recurrent VOCs/APCs were observed in two patients. Both patients had a medical history of VOCs/APCs, however, none of them resulted in inpatient admission during pregnancy. One of the pregnancies was further complicated by transfusion demand for RBCs due to symptomatic anemia unrelated to a VOC/APC. Peripartum (SCD-associated) complications were observed in one of the pregnancies and required inpatient admission due to a VOC (pyelonephritis) in the third trimester of pregnancy which progressed to an ACS, requiring CS. AIS was seen in one of the four patients. In the postpartum period, only in one of the two patients with past VOCs/APCs during gestational phase were VOCs/APCs also documented, respectively. Throughout observation time, no serious infectious-or TE complications were observed.
Three of the overall seven successful pregnancies in the other two HbS/C patients were uneventful (43%, 3/7), despite documented VOCs/APCs prior to conception in one of them.

Discussion
Pregnancy in SCD is associated with a high incidence of adverse events as highlighted by past meta-analyses implicating the importance of an interdisciplinary therapeutic approach by national reference centers to ensure maternal and fetal/neonatal well-being [4,12]. The results of our single-center study are in partial agreement with previous studies.
Before commenting on the findings in our patients, we have to point out on the low number of patients with severe genotypes receiving HC, given the overall benefit regarding reduction in morbidity and mortality rates in HC-treated SCD patients [13][14][15]. The low number was based on the following notion: some of the patients rejected treatment because of individual concerns. Others were not taken care by a medical reference center prior to pregnancy and, thus, had only very limited access to antenatal care. These patients presented at our department at a late stage of pregnancy which may be attributed to their refugee status, often requiring direct hospitalization due to SCD-associated complications. Furthermore, despite the rising prevalence of SCD in Germany due to recent immigration, SCD is still considered a rare disease compared to other countries. Nevertheless, as the present analysis summarizes outcomes over a period of two decades, our results support the findings of other studies stating an overall improved maternal and neonatal outcome over the past decades [3,16,17].
In a meta-analysis published by Oteng-Ntim et al. (2015), the authors estimated a sixfold increased risk of maternal mortality in HbSS females compared to non-affected persons [4]. However, mortality rates vary depending on the particular analyses, the time of data collection, and may likely be associated with a poorer health care system as compared to Europe. Even though, no increased maternal mortality is observed in developing countries with substantial experience, as published by Babah et al. (2019) [18].
Although a low prevalence of HC-induced teratogenicity can be assumed following exposure during pregnancy, treatment is supposed to be withheld three months before conception, if feasible [19]. Of note, in all our documented pregnancies (except one) with prior maternal HC exposure (treatment duration: ≥1 year), treatment was stopped only after diagnosis of pregnancy within the 1st trimester, indicating that all fetuses were exposed to HC during organogenesis (3rd to 8th week after conception). In these patients no obvious difference with regard to fetal mortality was observed (i.e., spontaneous abortions compared to HC-naive mothers) and neonatal development was not associated with any adverse events throughout observation time. Furthermore, in one of our patients with known alloantibodies and severe transfusion-dependent anemia, HC was needed to be reinitiated in the 3rd trimester of pregnancy up to four weeks prior to delivery during her two consecutive pregnancies resulting in no adverse fetal outcome. Our observations are therefore in line with the results observed in the European non-interventional, multicentric, prospective Escort-HU study. Hence, in some patients not suitable for SP(E)BT, HC (±EPO) may remain the only therapeutic option to ensure maternal and fetal well-being. Noteworthy, HC might even be considered safe in some patients if treated throughout the entire pregnancy. However, decision should be based on a case-by-case discussion and cannot be generally recommended due to the potential harmful effects [20].
Whether or not SPBTs are beneficial in pregnant SCD patients remains a matter of debate [21][22][23][24]. Importantly, transfusion techniques and timing of SP(E)BT vary throughout the different studies probably explaining the contradictory results. Our study was not designed to answer this important topic. However, we did observe a high demand for RBC transfusions, especially in HbSS patients. Taking the risk of alloimmunization into account, transfusions were restricted to patients presenting either with symptomatic/hemolytic anemia or to those with other SCD-associated or obstetric complications.
Alloimmunization is a multifactorial process and presents one of the most clinically relevant complications in the context of pregnancy. The incidence of alloimmunization is highest in SCD patients, with a reported rate of up to 47% [25]. Most of the alloantibodies are formed against Rhesus-(Rh-) and Kell-(K-) antigens [1,26]. It is assumed that this is due to the much larger variety of Rh-antigens found outside of Europe, particularly Africa, which differ from the common variants of European blood donors [1]. We observed alloimmunization in two out of eight HbSS patients with a history of periodic blood transfusions. Our low rate may be related to extended red cell antigen typing which may lower the rate of alloimmunization, although this has not yet been clearly proven [27]. Nevertheless, even in patients with a history of low exposure to allogeneic RBCs, alloimmunization can still occur at a relatively high rate, as reported for Ugandan sickle cell patients (with 6.1% alloimmunization) [28]. Although extended RBC matching has not been formally proven to contribute to a reduction of alloimmunization rates, it can speed up identification of the specificity of alloantibodies in the case of a positive antibody screen and also helps to select compatible RBCs [27]. Furthermore, some clinically significant alloantibodies can be difficult to detect because they evanesce over time and drop below the detection threshold in the immunohematology laboratory [29]. This potentiates the risk for severe transfusion-associated complications after antigen-re-exposure by transfusion of antigen-positive red blood cells. Such a booster can particularly lead to DHTRs [30]. It also emphasizes the importance of extended RBC antigen characterization (serologically and/or genetically) of donors and recipients, especially in European countries with a low disease prevalence to decrease the risk of potential transfusion reactions [31]. Particularly intrauterine transfusions can efficiently trigger alloimmunization against RBC antigens and against HLA molecules. As a matter of fact, alloimmunization against additional RBC antigens has been reported to occur in up to 25% of women already expressing an alloantibody. In contrast to transfusions in adults, extended antigen matching of donor cells (including Duffy-, Kidd-, and S-antigens) prior to intrauterine transfusion was shown to decrease the formation of alloantibodies in the mothers by 60% [32,33].
In summary, pregnant mothers suffering from SCD who have experienced severe SCD-related complications, either preceding the current pregnancy or during a previous pregnancy, may be considered for SP(E)BTs. This is the case if a high-risk exists for both, the mother and the fetus (e.g., twin pregnancies), or in patients encountering SCD-associated complications during the current pregnancy, in patients receiving HC before the gravidity, or in patients already receiving a chronic transfusion program [27]. A switch to SP(E)BTs may be worth to be considered in patients receiving crizanlizumab prior to or in the context of pregnancy for prevention of recurrent VOCs. Whether outcome might be favored by early initiation of SPEBT as described by Vianello et al. (2018) may be answered by the TAPS-2 study, as placental damage in SCD might even occur early in pregnancy [34,35].
Because not all documented pregnancies were monitored at our department, we included a patient survey. Pregnancies in our patients with successful outcomes were accompanied by a high rate of vaso-occlusive events (VOEs), especially in HbSS patients (96%) compared to HbS/C disease, requiring in part, if severe, recurrent inpatient admissions for medical care. Of note, the patient query revealed that not all of them seeked advice of their treating physicians during a painful crisis. This may offer an explanation for the overall higher VOC/APC rates described in our patients unlike other studies. Nevertheless, routine counselling was associated with a sense of security according to the patient survey, as these patients could directly contact their attending physician in the event of impending complications, if desired.
Complicated VOEs (ACS) requiring inpatient admission for partial RCE and intravenous anti-infective treatment were observed for four pregnancies (13%, 4/31), including a patient with HbS/C disease. Thus, patients with HbS/C disease also are at increased risk for serious complications when being pregnant or in the postpartum period. Nevertheless, a higher rate of complications was seen in patients with genotypes mediating severe disease.
Special attention should also be paid to the high number of elective abortions in our study. Reasons for medical abortions were related to age at time of conception, socioeconomic factors, including uncertain partnerships, or disease-related stressors, implicating the importance of early psycho-social support in patients with SCD.
Funding: FA was supported as a Clinician Scientist within the University Medicine Essen Academy (UMEA) program, funded by the German Research Foundation (DFG; grant FU356/12-1) and the Faculty of Medicine, University of Duisburg-Essen, Germany.

Institutional Review Board Statement:
The study involving human participants was reviewed and approved by retrospective analysis and use of data was approved by the Ethical Committee of the Faculty of Medicine at the University Hospital of Duisburg-Essen (21-10031-BO; 5/6/2021) and the study was conducted in accordance to the Declaration of Helsinki.

Informed Consent Statement:
The patients/participants provided their written informed consent to participate in this study.

Data Availability Statement:
All datasets generated for this study are included in the article.