Systemic lupus erythematosus (SLE) is a multi-organ autoimmune disease that primarily affects women of childbearing age. The recommendation in the past was to avoid pregnancy due to maternal and foetal complications1. However, better outcomes are now being achieved2,3. Factors that contribute to complications have been identified as lupus activity at the time of conception, the presence of previous lupus nephritis (LN), antiphospholipid antibodies (APA), anti-SSA/Ro and anti-SSB/La antibodies and maternal hypertension (HTN) prior to pregnancy4,5.

Pre-existing LN when planning a pregnancy carries an increased risk of developing pre-eclampsia, delayed intrauterine growth, stillbirth, small for gestational age babies, and preterm delivery5,6. HTN and kidney failure are predictors of poor maternal and foetal prognosis, and it is sometimes difficult to determine whether an increase in proteinuria or onset of HTN is due to pregnancy or a renal flare. This may explain the incidence of renal flares in pregnancy and the puerperium at between 8% and 70% in different studies, while reactivations accompanied by impaired kidney function are between 0% and 23%5–10.

It is important to highlight that the presence of moderate and severe kidney failure at conception increases the risk of developing severe HTN and impaired kidney function. In normotensive patients with mild and stable impaired kidney function, the risk of progression to irreversible kidney failure is low. In contrast, the prognosis is worse if kidney function is moderately impaired (serum creatinine 1.4–3.0 mg/dl), and impaired kidney function will persist in one third of these patients10–12.

The aim of this study was to evaluate maternal and foetal outcomes in pregnant patients with SLE with and without a previous diagnosis of LN and its influence on prognosis.

Using ACR 1997 classification criteria13, we retrospectively studied all pregnant patients with a diagnosis of SLE with and without previous LN, attended in the autoimmune diseases and pregnancy clinic of the Maternal and Neonatal Hospital of the city of Córdoba, Argentina, from January 2015 to April 2017. LN was confirmed by biopsy and classified according to NL ISN/RPS 2004 criteria14. All patients were assessed in the clinic by a rheumatologist and an obstetrician under high-risk pregnancy follow-up. Demographic data, duration of SLE, clinical and laboratory data were recorded. Antiphospholipid syndrome (APS) was classified according to the Sydney criteria15.

Routine laboratory work-up was performed at each trimester visit, including complete cytology, serum albumin, urea, creatinine, plasma ionogram, hepatogram, uraemia, complete urine and 24 h proteinuria. Antinuclear antibodies (ANA) were determined by indirect immunofluorescence assay using Hep-2, anti-dsDNA by indirect immunofluorescence test with Crithidia Luciliae, anti-ENA antibodies (Sm, RNP, Ro, La) by ELISA and then confirmed by Immunoblot assay, complement C3 and C4 levels by immunoturbidimetry, and anti-dsDNA and C3 and C4 levels were tested during each trimester.

Anti-cardiolipin (aCL) IgG and IgM and anti-B2 glycoprotein I antibodies were measured by ELISA and moderate and high titres were considered positive. The presence of lupus anticoagulant (LA) was assessed according to the criteria of the Subcommittee on Lupus Anticoagulant/Antiphospholipid Antibody of the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis16.

All patients underwent a first trimester ultrasound. Doppler ultrasound to evaluate the uterine arteries was performed at 20 and 24 weeks to assess the presence of a protodiastolic notch that could predict pre-eclampsia or delayed intrauterine growth. Pre-eclampsia laboratory profile was performed every 3 weeks from the second trimester, glycaemia, and an oral glucose tolerance test (OGTT) at 28 and 32 weeks of pregnancy, toco/gynaecological ultrasound every 3 weeks to evaluate the growth curve, and Doppler ultrasound of the umbilical artery and middle cerebral artery every 2 weeks after the 30th week of pregnancy, as necessary.

The patients were divided according to renal function into mild kidney failure (KF) with a serum creatinine level <1.5 mg%, moderate KF with a creatinine level between 1.5 mg% to 2.4 mg%, and severe KF with a creatinine level >2.5 mg%. Lupus activity was assessed in all patients with SELENA-SLEDAI adapted to pregnancy at conception17, in each trimester and in the puerperium. Lupus reactivation was considered a change of more than 4 on the SELENA SLEDAI compared to the previous visit. Persistent activity was defined by a score of 4 (excluding serology alone) at two or more consecutive visits. Pre-eclampsia was differentiated from renal lupus flare by urinary sediment and the presence of other characteristics of lupus flare18.

Renal flares in pregnancy were defined by the presence of active urinary sediment, increased proteinuria >2 g/24 h if baseline was <3.5 g/24 h or double the proteinuria if baseline was >3.5 g/24 h and an increase in serum creatinine to 30% from baseline19. Complete remission was considered when serum creatinine was normal (according to gender and body weight), proteinuria less than .2–.5 g/24 h, and urinary sediment inactive; partial remission: normal serum creatinine, proteinuria greater than .2–.5 g/24 h but less than 2 g/24 h, or a decrease in proteinuria (greater than 50% from baseline with urinary protein excretion less than 3.5 g/24 h)20.

Clinical remission on treatment was defined according to DORIS (prednisone less than or equal to 5 mg/day, immunosuppressive therapy allowed; SLEDAI (clinical without serology) = 0 PGA less than 518.

Adverse pregnancy outcome was defined as any maternal or foetal complication during pregnancy and in the immediate postpartum period (up to 6 weeks); the maternal complications assessed were pre-eclampsia, eclampsia, HELLP syndrome (haemolysis, elevated liver enzymes, plateletopenia)21, gestational diabetes22, spontaneous premature rupture of membranes, chorioamnionitis, arterial and venous thrombosis, reactivation of disease and LN, infections, and maternal mortality.

Foetal outcome was assessed as live birth, preterm delivery (birth before 37 weeks gestation), spontaneous abortion (spontaneous foetal loss before 20 weeks gestation), stillbirth (intrauterine foetal death after 20 weeks), low birth weight (live birth less than 2500 g [LBW]), and neonatal death (neonatal death up to 28 days after birth).

The route of delivery, whether vaginal, forceps or caesarean section, and the indication for caesarean section, as well as the treatment received during gestation and puerperium, were determined.

Seventy-nine medical records of pregnant women with a diagnosis of SLE were reviewed, two of which were excluded as they were lost to follow-up. Seventy-nine patients were included in the analysis, 39 without LN and 40 with a previous diagnosis of LN, with a total of 121 pregnancies (52 without LN and 69 with LN). Seventy percent were residents of the city of Córdoba and 77.2% had no social security. Fifty-eight percent were Caucasian. The mean maternal age at first pregnancy was 26.8 years (16–38 years) in the group without LN and 27.2 years (15–39 years) for the LN group; disease duration in years was similar in both SLE groups, as was the number of previous abortions. Maternal characteristics are described in Table 1.

The kidney biopsies were: 5 (12.5%), class II; 9 (22.5%) class III; 19 (47.5%) class IV; 5 (12.5%) class V, 1 (2.5%) class III + V, and 1 with no data.

Table 2 shows treatment at baseline and during pregnancy. Nephritis was the main reason for steroid pulses in the patients with LN, and in the 3 patients without LN the indications were thrombocytopenia in 2, and serositis in one.

During follow-up, the patients with LN had more frequent changes to their therapy with increased doses of prednisone, methylprednisolone pulses and increased immunosuppression.

The demographic characteristics of our series, age and disease duration are like those described by Janardana et al.4, Aggarwal et al.23, Chen et al.24, and Buyon et al.25. However, in our study the percentage of patients with LN was higher due to the design. Of note, 55% of patients in the nephritis group had disease activity at the start of pregnancy, whereas only 10% in the group without nephritis. However, in other series reviewed, 90% of patients were in remission. When compared with the Janardana et al.4 and Aggarwal et al.23 series in India, 67% and 47%, respectively, had live births despite a better start to pregnancy. Our foetal outcomes resemble Buyon’s series25, perhaps similar ethnicity could explain this. This coincides with the literature, which describes the average number of successful pregnancies in different series as between 65%–90%, with a less favourable prognosis when the patient became pregnant with active disease4,26. In our sample, the rates of baseline disease activity were significantly higher in the LN group from the start, which correlated with higher reactivation, higher doses of immunosuppressants and higher emergency caesarean section outcomes when compared with the lupus group without nephritis. However, foetal outcomes remained the same in both groups.

In line with other studies, a significantly higher rate of lupus flares was observed here in patients with previous nephritis compared to those without renal involvement26. Despite the renal reactivations in this group, few developed kidney failure, as in other series2,6. LN, mainly diffuse proliferative (class IV), has been associated with various complications in pregnancy, especially if active27. In this study, most of the patients with LN were class IV, and despite being on immunosuppressive therapy with azathioprine and steroids, four patients in this class had reactivations. Only one reactivation with kidney involvement was documented in a patient in the group without nephritis.

In terms of treatment received, the LN group required higher doses of steroids and immunosuppressive therapy for disease control, like that shown in other series26. Although no significant differences were found, probably due to the number of patients, patients in the LN group received more antithrombotic prophylaxis with heparin.

Pre-eclampsia is a frequent complication compared to the general population and notably higher in patients with previous nephritis27,28. In this study we found significant differences in the frequency of pre-eclampsia in patients with previous LN, in line with other studies11,25,29. This could be related to the higher disease activity at the start of gestation and the presence of APA. Despite receiving ASA, the rate of pre-eclampsia remains high, in agreement with other similar series: 11.9%–18.1%4,24.

Preterm delivery is a common complication in pregnant women with lupus, estimated to occur in about one third of pregnancies27. A high rate of preterm delivery was recorded in both groups in this study, slightly higher in the LN group, and this could be related to the development of pre-eclampsia, greater disease activity and the need for steroids to control it28.

Regarding foetal prognosis, the most important predictors are disease activity prior to pregnancy, the presence of kidney failure and the presence of APA3,15,28. In this series, the LN group has a greater history of associated APS and APA, and despite this, favourable results were found in more than 90% of pregnancies, with similar numbers of live births in both groups, which contrasts with other series4,23, as mentioned above. This could be due to the treatment received with the combination of heparin and aspirin prophylaxis. More foetal deaths were documented in the LN group, which may be related to the abovementioned factors.

Caesarean section was the main route of delivery in both groups, and its emergency indication was significantly higher in the LN group. This was due to the coexistence of pre-eclampsia, as described by Saavedra21 and Rodrigues et al.30.

However, compared with Asian, North American, or Canadian series, the number of caesarean sections is unusually high, and there could be some cultural bias23–25.

A strength of our study is that we could compare patients with a major manifestation such as LN with lupus patients without nephritis. We consider the retrospective design to be a weakness, as it did not allow us to reliably collect data on previous disease and treatment, and therefore we do not have data on preconception counselling and family planning, or the number of patients.

The patients with LN experienced more maternal complications such as lupus flare and pre-eclampsia. However, LN would not lead to worse obstetric or foetal outcomes. Patients should be closely monitored before and after pregnancy.

The study was undertaken with a grant from SECYT, Secretaría de Ciencia y Técnica,Facultad de Ciencias Médicas,Universidad Nacional de Córdoba.

The authors have no conflict of interests to declare.