Management of pregnancy in autoimmune rheumatic diseases: maternal disease course, gestational and neonatal outcomes and use of medications in the prospectiveItalian P-RHEUM.it study

Introduction

Autoimmune rheumatic diseases (ARD) frequently affect female individuals of childbearing age; therefore, reproductive issues are of fundamental importance in the management of women living with these chronic diseases. Treatment strategies for ARD have significantly improved over the last few decades, facilitating permissive conditions for a pregnancy,1 2 but also posing new challenges such as the use of novel antirheumatic drugs in pregnancy and lactation.3 Adequate knowledge about the implications of ARD on reproductive health is essential for physician–patient communication, hence scientific societies produced recommendations and guidelines to assist healthcare professionals.4–8 However, these sets of guidance were mostly based on low-quality evidence and expert opinion, due to the paucity of randomised controlled trials and prospective studies of adequate sample size involving pregnant patients with ARD.

The need for evidence-based answers to the most common questions of women with ARD who are planning a pregnancy9 prompted researchers to set up prospective cohorts of pregnant patients and national Pregnancy Registries in several Countries worldwide.3 10 In order to facilitate harmonisation and standardisation in collected variables, a EULAR Task Force was convened to define a core data set for registries and observational studies that prospectively collect information about pregnant women with ARD and their infants.11

In 2017, the Italian Society for Rheumatology (SIR) promoted a prospective cohort study called P-RHEUM.it (The ITalian registry of Pregnancy in the RHEUMatic diseases). Its purpose was to address several primary and secondary objectives regarding maternal disease course during pregnancy, pregnancy outcomes and maternal/infant outcomes after delivery, with a focus on patient-reported outcomes (PROs) by means of validated questionnaires.

We report an ad-interim descriptive analysis to provide a general overview of the real-world experience of pregnancy in ARD regarding maternal disease flares, adverse pregnancy outcomes (APO), neonatal complications and use of medications according to the current standard of care in rheumatology.

Patients and methods

Statistical analysis

The selection of pregnancies to be analysed was performed on all pregnancies potentially ended by at least 90 days from the data extraction date (8 November 2023), on fulfilment of inclusion criteria, singleton status and complete data entry at least up to the visit at 30–60 days after delivery. The estimated date of delivery was calculated based on the first day of the last menstrual period (which could be the actual one or the estimated one based on fetal ultrasound). The flowchart is depicted in figure 1.

Figure 1

Request permissions

Flowchart showing the selection of pregnancies included in the analysis.

Data were descriptively analysed and reported as numbers and percentages for categorical variables and as median values (IQR) for continuous variables. Proportions were calculated according to the appropriate denominator, considering different groups such as patients, pregnancies or neonates. Missing data were reported when appropriate. All data processing was performed using R V.4.2.2 (Foundation for Statistical Computing, Vienna, Austria).

Results

Over the study period, 1298 pregnancies were enrolled. We analysed 866 singleton pregnancies in 851 patients with ARD enrolled in 27 centres. The distribution of maternal diseases is shown in figure 2. A diagnosis of inflammatory arthritis (IA) was present in 276 (32.4%) patients, while 557 (65.5%) belonged to the group of connective tissue diseases (CTD) - systemic vasculitis (SV) (including antiphospholipid antibodies (aPL) carriers); 18 (2.1%) women were asymptomatic carriers of anti-Ro/SSA antibodies.

Pregnancy and neonatal outcomes

Patients were enrolled at a median gestational age of 11 weeks (IQR 8–15); 355 (41.7%) were primigravida. More than one pregnancy in the same woman was registered in 14 cases (1.6%) (figure 3). Maternal disease flares were observed in 135 (15.6%) pregnancies, ranging from 8.8% in undifferentiated connective tissue disease to 26.1% in psoriatic arthritis (PsA) (table 3). No maternal deaths occurred during the study period.

Figure 3

Request permissions

Illustration of number of patients, pregnancies and neonates included in the analysis, presented on maternal diagnosis of autoimmune rheumatic disease. Non-viable pregnancies: A: pregnancy loss ≤10 GW; B: pregnancy loss 11–20 GW; C: pregnancy loss >20 GW; D: elective termination of pregnancy. anti-Ro/SSA, asymptomatic carriers of anti-Ro/SSA antibodies; GW, gestational week; IIM, idiopathic inflammatory myopathies; JIA, juvenile idiopathic arthritis; PAPS - aPL, primary antiphospholipid syndrome and antiphospholipid antibody carriers; PsA, psoriatic arthritis; pSS, primary Sjögren’s syndrome; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; SpA, spondyloarthritis; SSc, systemic sclerosis; SV, systemic vasculitis; UA, undifferentiated arthritis; UCTD, undifferentiated connective tissue disease.

Table 3

View inline•Open as popup

Characteristics, complications and outcomes of pregnancies. Categorical variables are expressed as numbers (percentages) and continuous variables as median (IQR).

Out of 866 pregnancies (table 3), 61 (7%) ended in miscarriage, and 11 (1.3%) underwent elective termination (1 due to malformation); 53 (6.1%) had been induced by assisted reproduction techniques (ARTs). Obstetrical complications occurred in 261 (30.1%) pregnancies, including 20 (2.3%) pregnancies complicated by pre-eclampsia (PE), 61 (7%) by gestational diabetes and 35 (4%) by gestational hypothyroidism (online supplemental table 5). Fetal growth restriction (FGR) was detected in 69 (8%) of pregnancies, 15 (5.3%) in 281 IA pregnancies, 52 (9.2%) in 567 CTD-SV pregnancies and in 2 pregnancies (11.1%) in anti-Ro/SSA carriers. Congenital heart block (CHB) was diagnosed in two fetuses of patients carrying anti-Ro/SSA antibodies, namely one patient with systemic lupus erythematosus (SLE) and one asymptomatic carrier. In our cohort, the incidence of CHB was 2 cases out of 157 pregnancies (1.3%) in women with anti-Ro/SSA.

Live births were 794 (91.7%); 4 perinatal deaths (0.5%) occurred, all in newborns of patients with SLE (3 cases due to very severe preterm delivery caused by PE; 1 case due to myocarditis associated with CHB). Most neonates were born at term (84.9%); severe preterm birth (PTB) before 34 gestational weeks (GWs) occurred in 26 (3.3%) cases. Vaginal delivery was carried out in 463 (53.5%) cases; caesarean section (C-section) was performed in 312 (39.3%) cases, 61.5% elective and 38.5% emergent.

Among 790 newborns who were alive at 28 days after delivery, 31 (3.9%) were classified as small-for-gestational-age (SGA), 16 (2%) displayed congenital malformations, and 169 (21.4%) suffered from complications at birth (table 4). Exclusive maternal breast feeding was received by 404 (46.7%) neonates in the first month of life. EPDS was compiled by 414 women (52.4%) at 1–2 months after delivery and 89 (21.5%) scored positive for emotional disturbances/depressive symptomatology.

Table 4

View inline•Open as popup

Neonatal outcomes (excluding perinatal deaths). Categorical variables are expressed as numbers (percentages) and continuous variables as median (IQR).

Discussion

This nationwide prospective cohort study was created in 2017 with the goal of systematically collecting data on pre-specified outcomes of pregnancy in patients with ARD. With regular follow-up visits and detailed information on maternal disease activity, obstetrical complications, and use of medications at several time points, prospective pregnancy registers can provide a unique perspective that is complementary and additional to the information derived from administrative databases or global safety registers. This inaugural paper of the P-RHEUM.it study aims to provide an overview of the cohort and to serve as a reference for future publications focusing on the inferential analysis within single diseases.

By enrolling nearly 1300 pregnancies in 29 centres over 5 years, the P-RHEUM.it study can provide insight into several aspects of the current management of ARD during pregnancy and their impact on obstetrical and neonatal outcomes. Nearly two-thirds of enrolled patients had a diagnosis within the spectrum of CTD-SV (including rare diseases), with being SLE the most represented disease in this cohort (19.6%). Although IA such as rheumatoid arthritis (RA), PsA and spondyloarthritis (SpA) are more prevalent diseases than CTD-SV, the composition of our cohort reflects the nature of the enrolling centres, which are second-level or third-level, hospital-based rheumatology units; many centres run joint pregnancy clinics. Therefore, we may speculate that the referral of pregnant patients to specialised centres is more likely for those with CTD-SV who are perceived as at increased risk of pregnancy complications. Since 38% of the patients with IA in our cohort were treated with bDMARDs, we may also speculate that patients with IA are more likely to be referred when severe and treated with second-line drugs.

Preconception counselling and shared-decision making are the key elements when taking care of women living with ARD who wish for a pregnancy.3 In our cohort, this approach was reflected by the condition of remission or low disease activity for most of the patients, as described by patient-reported and physician-reported outcomes. Only one-fifth of patients had a disease flare during the year prior to conception, supporting the fact that most of the patients have planned their pregnancy while on stable inactive disease, as also suggested by a median disease duration of 6.4 years prior to conception. Along with this, disease flares during pregnancy occurred in a minority of patients (15%), showing that good disease control prior to conception and during the first trimester is a protective factor against flares later in pregnancy, as already described in RA19 20 and SLE.21

The use of pregnancy-compatible medications before conception is one of the pillars of preconception counselling and aims at keeping the maternal disease under control during pregnancy; inactive disease is indeed a major asset in favouring both maternal and fetal well-being.3 Active disease was demonstrated to be strongly associated with APO; therefore, good disease control prior to and during pregnancy is the best strategy to minimise the impact of maternal disease on pregnancy outcomes.22

When comparing outcomes in our cohort with those of pregnancies in the Italian general obstetric population (GOP) during the same calendar years (online supplemental table 7), the frequency of spontaneous miscarriage was apparently not significantly different (7% in ARD vs 11% in GOP). However, we cannot exclude under-reporting of early miscarriages in our cohort. Conversely, there is a striking difference in the frequency of previous miscarriages, which was 56.9% and 17.5% in our cohort and in GOP, respectively. The fact that one out of two non-primigravida patients had experienced at least one previous miscarriage may reflect the influence of subclinical/undiagnosed and/or untreated ARD on pregnancy outcomes.23 Chronic disease and/or its treatment may also have influenced fertility, as pregnancies induced by ARTs were 6.1%, while they were 3.7% in GOP. On the other hand, a normal time-to-pregnancy (TTP) of 3 (IQR 1–8) months was observed, suggesting that infertility/subfertility was not a concern in our cohort. The TTP reported in the literature was longer than ours3 and we cannot exclude underestimation, since we enrolled patients who were already pregnant and not during the preconception period.

In the present study, we restricted the analysis to singleton pregnancies, in order to assess obstetrical complications without the bias of multifetal pregnancies which are at increased risk of APO such as PE, FGR and PTB.24 PE deserves a special comment, because it carries a high risk of irreversible maternal and fetal damage and has been reported to occur more frequently in pregnant women with ARD than GOP.3 LDASA was proved to be an effective preventative treatment in GOP at risk of PE.25 Therefore, rheumatology scientific societies elaborated the recommendation of offering LDASA to women with ARD at increased risk of PE, particularly those with lupus nephritis, antiphospholipid syndrome or carriers of a high-risk aPL profile.5 6 26 LDASA seems to be also beneficial in preventing other complications such as FGR and PTB27; therefore, the use in pregnant women with ARD may go beyond the current recommendations, as suggested by the fact that nearly 6 out 10 patients in our cohort were on LDASA during pregnancy. The widespread use of LDASA in our cohort may have contributed to the overall low frequency of PE (2.3%) that is in line with GOP data reported in the literature (5.3%, range 1.8–9.3)28; however, it must be acknowledged that the assessment of the real incidence of PE is hindered by many factors, and no nationwide Italian data are available up to now. The heterogeneous approach to the use of LDASA in our cohort reflects the current debate in rheumatology on whether all patients should be treated or only selected groups.29–32

One of the highlights of this study is the high proportion of live births (91.7%), the majority of whom were born at term (84.9%). The frequency of congenital malformations and SGA newborns was not substantially different from GOP. This information is relevant to patients because they usually want to know if there is a good chance of having ‘a baby in hands’ who is not affected by prematurity and other major problems. We may speculate that this high proportion of live births, reported also by other European pregnancy registries for SpA,33 SLE34 and primary Sjögren’s syndrome,35 is the tangible output of modern rheumatology management applied to pregnant patients. The use of HCQ during pregnancy was widespread in most ARD included in our cohort, reaching the highest frequency in SLE (80.5%), as per international recommendations.5 6 csDMARDs were taken during pregnancy by 1 out 10 patients, particularly those with SLE and idiopathic inflammatory myopathies. This suggests that pregnant women were treated as if they were not pregnant with regard to pregnancy-compatible medications.36 The biggest change in management can actually be seen for TNFi. As a consequence of the initial cautious approach of discontinuing TNFi at the beginning of pregnancy, an increased risk of maternal flares and APO was observed in patients with SpA and RA.37 38 Based on increasing safety data about the use of TNFi during pregnancy,6 7 clinicians have progressively gained confidence using TNFi during pregnancy more liberally.36 39 40 Particularly, CPZ was the most used TNFi in our cohort, due to its demonstrated lack of transplacental passage,41 that allows maternal treatment throughout pregnancy without any potential impact on the immunocompetence of the neonate. Another advantage of using TNFi during pregnancy is the chance for reducing or stopping glucocorticoids, which may in turn facilitate PTB42 and serious maternal and neonatal infections.43 Glucocorticoids were used in less than one-third of our patients, which is a testament to both the awareness about the risk of adverse effects during pregnancy and the practice of aiming at steroid-free disease remission.

Despite the encouraging results in terms of live births, a higher frequency of PTB (13.5%) as compared with GOP (6.3%) must be acknowledged. Severe PTB (3.2%) was also more frequent than in GOP (1.6%) and accounted for three out of four perinatal deaths. The frequency of spontaneous vaginal delivery was much lower than in GOP (31.2% vs 63.1%), probably due to the common practice of inducing delivery at 40 GW in women with ARD (20.3% in our cohort) to minimise the risk of oligohydramnios and/or sudden fetal death due to acute placental failure in the post-term. Reassuringly, the proportions of both elective and emergent C-section were similar to GOP.

The puerperium is a delicate period, especially for women with ARD who may experience intense disease flares.44 Breast feeding is universally encouraged and women with ARD should be allowed to carry it out by receiving compatible medications.7 Nearly half of the patients in our cohort breast fed in the first month after delivery (similarly to GOP), meaning that they had received adequate counselling and were able to make an informed decision about breast feeding.

The P-RHEUM.it study was the first national registry in which emotional well-being after delivery was investigated by means of EPDS, a simple and rapid-to-score tool that can signal whether a woman reported symptoms related to anxiety and depression that deserve further assessment. In fact, it was estimated that scoring positive at EPDS is associated with a pooled risk of postpartum depression (PPD) of 27.5% (95% CI 17.8 to 37.3).45 In our cohort, one-fifth of patients who compiled the EPDS at the postpartum visit scored positive, showing that PPD should be ruled out in a non-negligible proportion of patients with ARD. Similarly, a Mexican study found that 26.9% of women with ARD scored positive at EPDS.46 A population-based study in the USA estimated a slightly increased risk of PPD in women with SpA/PsA/RA as compared with controls, with PPD being diagnosed in 17.2% of patients.47 Emotional distress and PPD should not be overlooked in patients with ARD and measures should be implemented in order to minimise the impact on baby–mother bonding and parenting ability.

Conclusion

Over decades, ‘Reproductive Rheumatology’ has tackled research in a sensitive field. It has been highly challenging to perform randomised clinical trials in pregnant women, because of regulatory issues, difficulty in getting dedicated funding and the need for a multicentre international approach to reach sufficient numbers. As a consequence, there is a lack of unbiased, rigorous studies that can drive clinical decisions. Prospective studies and registries have contributed to filling the gap, showing that pregnancy is possible in women with ARD and discussion about it should be embraced rather than discouraged.

The P-RHEUM.it study has captured the real-world experience in the management of pregnancy in a large cohort of women with different ARD. Maternal and neonatal outcomes were overall good, suggesting that a set of measures can work as a ‘toolkit’ in the approach to pregnancy. These measures include multidisciplinary preconception counselling and pregnancy monitoring, and individual risk stratification for a tailored approach that aims at inactive disease while on pregnancy-compatible medications. The most sensitive indicator of the effectiveness of these interventions is the alignment of the frequency of most complications and outcomes to that of the GOP. This can be an important reassuring message to women with ARD, as well as a strong incentive to comply with treatment and monitoring plan in order to minimise the impact of the disease on pregnancy outcomes.

However, there is still much to be addressed to improve the reproductive journey of women with ARD. The P-RHEUM.it study will dissect disease-specific issues to better understand the impact of current rheumatology management and to plan for additional interventions and/or changes in practice. It will also bring the emotional well-being of the patient into the spotlight by analysing the QoL and the risk of mental health distress, along with data about children’s conditions in the first 2 years of life, which is a frequently-asked-question by the patients.