Maternal and infant outcomes in sarcoidosis pregnancy: a Swedish population-based cohort study of first births

Sarcoidosis is a multisystem inflammatory disease characterized by the formation of non-caseating granulomas of unknown etiology, primarily affecting the pulmonary and lymphatic system. About half of the sarcoidosis patient population is female, many of whom are diagnosed at a reproductive age [1]. Several systemic inflammatory disorders are associated with an increased risk of complications during pregnancy, labor and delivery compared to the general population [2‐5], however few studies have examined pregnancy in sarcoidosis.

In some case series, sarcoidosis symptoms remained unchanged or improved during pregnancy, and few adverse events were observed [6, 7]. However, other case reports have described severe adverse outcomes, including maternal mortality and cardiac arrest in patients with cardiac and neurosarcoidosis [8‐11]. Only one cohort study has been conducted which included 678 deliveries from the Nationwide Inpatient Sample in the U.S. [12]. Sarcoidosis was associated with an increased risk of preeclampsia, eclampsia, cesarean delivery, postpartum hemorrhage, and venous thromboembolism (VTE) and infants born to women with sarcoidosis were more likely to be premature and have intrauterine growth restriction [12]. Women with sarcoidosis were included in the study if they received a delivery-related discharge diagnosis of sarcoidosis, which may not capture all women with a history of sarcoidosis. Furthermore, parity was not taken into account, which is an important factor since several complications are more frequent in first pregnancy [13, 14].

Prospective investigations in a large, representative and contemporary sample of women with sarcoidosis are needed to develop and update recommendations for sarcoidosis pregnancy. Our aim was to estimate the risks of adverse maternal and infant outcomes in first-time pregnancies with and without sarcoidosis in a population-based cohort study in Sweden.

We identified 182 singleton first pregnancies in mothers with a history of a sarcoidosis diagnosis at conception and 6630 pregnancies in general population comparators without sarcoidosis between 2002 and 2013. Women with sarcoidosis were slightly older at delivery compared to women without sarcoidosis (32 vs. 31 years old) and were more likely to have a history of diabetes (2.2% vs. 1.2%) and hypertension (5.5% vs. 1.5%; Table 1). The median number of years since first diagnosis of sarcoidosis was 3.0 (interquartile range 1.4, 5.4).

There was no increased risk of gestational hypertension associated with sarcoidosis (RR 0.9; 95%CI 0.4, 2.1; Table 2). Nine percent of women with sarcoidosis had preeclampsia/eclampsia compared to 5% in the general population and the adjusted relative risk associated with sarcoidosis was 1.6 (95%CI 1.0, 2.6). The risk of cesarean delivery was increased by 30% in women with sarcoidosis (RR 1.3 95%CI 1.0, 1.6) and the increased risk was similar for elective and emergency cesarean deliveries. There was no increased risk of operative vaginal delivery nor postpartum hemorrhage. There were fewer than 5 cases of stillbirth, gestational diabetes and infection among women with sarcoidosis. There were no cases of maternal death, cardiac arrest, VTE and placental abruption.

The risk of preterm delivery for sarcoidosis pregnancies was 70% higher than the risk in the general population (Table 3; RR 1.7; 95%CI 1.1, 2.5). There were fewer than five newborns who had very preterm birth, a 5-min Apgar score < 7 or were small or large for gestational age. There was no increased risk of infection within three months after birth (RR 0.9; 95%CI 0.5, 1.7) associated with maternal sarcoidosis. There were no neonatal deaths in the sarcoidosis group. 6.1% of sarcoidosis pregnancies had a major non-chromosomal birth defect compared to 3.7% in the general population (RR 1.6; 95%CI 0.9, 2.8). There was a similar percentage of cardiac birth defects in newborns of mothers with sarcoidosis and the general population (1.7% vs. 1.8%, respectively) but a higher percentage of non-cardiac birth defects (5.0% vs. 2.1%, respectively) which were mainly defects of the eye, ear, face, neck or digestive system. In a post-hoc analysis excluding women with pregestational and gestational diabetes, the results did not change.

There were 140 sarcoidosis pregnancies and 3441 general population pregnancies with data on dispensed prescriptions between 3 months before and up to 1 year after pregnancy. There were no dispensations of azathioprine, methotrexate or leflunomide recorded < 3 months before or during any pregnancy with sarcoidosis. 7.9% of sarcoidosis pregnancies had at least one dispensation of oral glucocorticoids during pregnancy. In the three months before pregnancy, 5.7% of the sarcoidosis pregnancies had at least 1 dispensation of oral glucocorticoids. During the course of pregnancy, the proportion decreased to 4.3% in the 3rd trimester and was lowest in the 3 months after delivery (1.4%; Fig. 1).

When a less strict sarcoidosis definition was implemented, relative risks were similar to the main analysis (Additional file 2: Table S2). When missing values for the confounders BMI and smoking status were imputed with extreme values, results remained similar (Additional file 3: Table S3). In a post-hoc analysis, women were excluded who had a glucocorticoid dispensation < 9 months before LMP, a history of pregestational hypertension or diabetes (n = 29 women with sarcoidosis, n = 195 women without sarcoidosis) to determine whether these factors could explain the increased risks. The results were similar to the main results (Additional file 4; Table S4).

In this large population-based cohort study, sarcoidosis was associated with a 60% increased risk of preeclampsia/eclampsia and a 70% increased risk of preterm birth. Risk for cesarean delivery was 30% higher in pregnancies with sarcoidosis compared to the general population. We did not observe many severe complications during pregnancy or delivery for the mother. Newborns of mothers with sarcoidosis had an increased risk of birth defects, the majority of which were non-cardiac. In absolute numbers, most adverse outcomes in sarcoidosis pregnancies were uncommon, and the majority had uncomplicated pregnancies.

There were no cases of death, cardiac arrest, VTE or placental abruption linked to sarcoidosis. This is in contrast to a previous U.S. study which found a 6-fold increased risk of VTE [12]. The study from the U.S. also found a 70% increased risk of postpartum hemorrhage which we did not observe. Discrepancies in the relative risks between the two studies may be due to several differences. The study from the U.S. used data from the Healthcare Cost and Utilization Project Nationwide Inpatient Sample and identified sarcoidosis from delivery-related discharge diagnoses, potentially including women with a more active/severe disease at the time of delivery. In contrast, we identified sarcoidosis pregnancies by requiring at least two sarcoidosis health care visit prior to the pregnancy, which likely includes women with both active and resolved disease states. Definitions of pregnancy outcomes also vary between the two study populations. In Sweden, postpartum hemorrhage is defined as > 1000 ml blood loss (regardless of mode of delivery) compared to the WHO cut-off of > 500 ml [20]. There are also differences between the two countries in race, socioeconomic position and/or access to health care, which are related to disease severity and probably also adverse pregnancy outcomes [21]. The study from the U.S. was quite diverse, with Black Americans making up more than one third of the sarcoidosis cases. Furthermore, our study restricted to first births to account for differences in risk by parity because a selectively healthier group of women chose to have a subsequent pregnancy [13]. Despite these differences, we observed similarly increased risks of preeclampsia/eclampsia, cesarean delivery and preterm birth compared to the study from the U.S. Guidelines for care of mothers with sarcoidosis should address these risks, although they are not large; 9 out of 100 mothers with sarcoidosis had preeclampsia/eclampsia compared with 5 out of 100 mothers without sarcoidosis.

Newborns of women with sarcoidosis did not have more adverse outcomes compared to the general population in terms of size for gestational age, infection or neonatal death. However, we observed a higher risk of major birth defects associated with maternal sarcoidosis; 6.1% compared to 3.7% in the general population. There was no increased risk of cardiac birth defects, but an increased risk of non-cardiac birth defects. These congenital malformations might be more likely to be detected in infants born to women under surveillance for other comorbidities, thus causing a detection bias to inflate the risk in the sarcoidosis group. Another explanation for the higher risk of birth defects could be due to maternal characteristics such as diabetes or high BMI [19, 22], although we adjusted for BMI and excluded diabetes in a post-hoc analysis and results were similar. These results should be interpreted with caution due to the small number of events in the sarcoidosis group (n = 11).

Women with a history of sarcoidosis during pregnancy did not take many sarcoidosis-related medications; 7.9% were dispensed oral glucocorticoids at any time during pregnancy. Only 5.7% were dispensed oral glucocorticoids in the three months prior to pregnancy, indicating that few women had sarcoidosis in need of treatment. Interestingly, the lowest percentage of oral glucocorticoid use was in the first 3 months after delivery. This is consistent with some observations that sarcoidosis and other inflammatory diseases improve during pregnancy [6, 23, 24]. However, during the 4 to 6 months after delivery, the percentage increased again, potentially indicating a relapse. This should be investigated further in a larger study with information on disease activity.

Our findings are in line with studies in other inflammatory diseases which have shown an increased risk of preeclampsia, preterm birth and cesarean delivery compared to the general population [2‐4]. The percent of first pregnancies affected by these outcomes in sarcoidosis are similar to those reported in inflammatory arthritis and inflammatory bowel disease in Sweden (5–6% preeclampsia, 8–11% preterm birth, 26–29% cesarean delivery) [3, 25, 26]. These highly intertwined outcomes have multiple causes which are difficult to disentangle in small samples [27]. Pregnancy involves a complex process of dynamic immune regulation, with an anti-inflammatory stage in the 2nd trimester and later a switch to a pro-inflammatory environment needed to initiate labor [28]. In sarcoidosis, the immune system’s ability to regulate and adapt to the immune environment during pregnancy may be impaired, leading to pregnancy complications. Comorbidities associated with sarcoidosis and its treatment may play a role in the pregnancy complications we observed in this study. However, when excluding women with glucocorticoid dispensations, hypertension or diabetes before pregnancy, the relative risks did not markedly change.

Sarcoidosis diagnosis and maternal and infant outcomes in this study were based on ICD codes, which may be misclassified. By restricting to at least two visits listing sarcoidosis, we aimed to minimize this bias. The use of nationwide registers allowed for increased power to study rare outcomes in this patient population, however we still had limited power for some outcomes and could not study mediating factors such as medication. An important limitation was the lack of information on disease activity or phenotype, which is likely related to pregnancy outcome. This should be addressed in future studies and used to identify women at the highest risk of adverse outcomes. Our results may not be generalizable to other populations with a different quality and accessibility of maternity care and/or different rates of cesarean delivery. A key strength of this study is the use of population-based Swedish national registers in a setting with universal access to healthcare. The use of Sweden’s Birth Register which captures 99% of all deliveries since 1973, made it possible to assess the parity-specific risk for pregnancy outcomes and to adjust for confounding variables, such as BMI at first prenatal visit and self-reported smoking status [15].

Mothers with sarcoidosis and their treating physicians should be aware of the increased risks of adverse pregnancy outcomes associated with sarcoidosis, although it should be kept in mind that these outcomes are uncommon. Early detection of these conditions may prevent potential adverse effects on mothers and newborns. Besides its immediate consequences, preeclampsia is a predictor for later development of cardiovascular disease [29]. This highlights the need for adequate follow-up care of women with sarcoidosis who have experienced pre-eclampsia.

To conclude, our findings demonstrate that sarcoidosis is associated with increased risks of preeclampsia/eclampsia, cesarean delivery, preterm delivery and non-cardiac birth defects. Awareness of these increased risks can be used for pregnancy planning and counselling for women with sarcoidosis.