Patient characteristics of pre-pregnancy
In total, 108 pregnancies in 74 patients were analysed, and the patients’ characteristics are shown in Table 1. Of these 74 patients, 48 became pregnant once, 21 twice, three patients three times, one patient four times, and two patients became 5 times, respectively, during the study period. Patient age at pregnancy was 33 (30–36) years. The SLEDAI score was 0 (0–4). One patient was complicated with active lupus nephritis, but there were no active major organ lesions in the other patients. In the past, 29 patients (26.9%) were diagnosed as having lupus nephritis. The serum C3 level before pregnancy was 81 (74–92) mg/dl. Anti-dsDNA antibody, anti-RO/SS-A antibody were positive in 73 (67.6%) and 38 patients (35.2%), respectively. Forty-six patients (43%) were also positive for lupus anticoagulant activity, anti-cardiolipin IgG antibody, or anti-β-2GP1 antibody. Ninety-one patients (84.3%) were treated with 10 (5–12) mg/dl of GC before pregnancy. GC pulse therapy was previously administered in 27 patients (25.0%). LDA was administered to 15 aPL-positive patients with a history of obstetrical abnormalities, and LMWH was co-administered from the first trimester.
Foetal outcome
Table 1 shows the foetal outcomes of all 108 pregnancies. Eighty-three pregnancies (76.9%) resulted in live births, 11 in miscarriage (at 5–11 weeks), 3 in stillbirth, and 11 in abortion. Two patients with APS experienced stillbirth at 22 and 25 weeks, respectively, and one patient had chromosome aberration at 15 weeks. The reasons for abortion were social background, poor control of the underlying disease, and taking medications that are contraindicated during pregnancy. The gestational age at delivery was 37.0 (33.5–37) weeks, and the foetal birth weight was 2760 (2289–3007) g. Twenty-seven (32.5%) of the 82 live births were premature, and 2 were terminated before 30 weeks of gestation due to maternal preeclampsia. Lupus nephritis relapsed in two patients, and their pregnancies ended before 32 weeks.
Comparison of patient backgrounds between preterm and full-term birth groups
Table 2 shows the pre-pregnancy clinical backgrounds of the preterm and full-term birth groups. There was no difference in age, duration of illness, or SLEDAI score between the two groups. The preterm birth group had a significantly higher history of lupus nephritis than the full-term birth group (P < 0.001). There was no difference in the positive rate and antibody titre of anti-dsDNA antibody, or those in the other autoantibodies, between the two groups. Pre-pregnancy serum C3 levels were significantly lower in the preterm birth group (77.5 mg/dl) than in the full-term birth group (87.5 mg/dl) (P = 0.027). In terms of treatment, there was no significant difference in the dose of GC administered before pregnancy, but the past use of GC pulse therapy was higher in the preterm birth group (P = 0.009).
Figure 1 shows the transition of each parameter before and during pregnancy between the full-term birth group and preterm birth group. In the preterm birth group versus full-term birth group, serum C3 levels were lower in the first, second, and third trimesters (81 vs 97 mg/dl, P <0.001; 88 vs 107 mg/dl, P < 0.001; and 100 vs 107 mg/dl, P = 0.024, respectively). However, there was no significant difference between the two groups in anti-dsDNA antibody titre and SLEDAI score during pregnancy. The dose of GC during pregnancy was significantly higher in the preterm birth group in the second and third trimesters (P = 0.01 and P = 0.001, respectively). Forty-six patients
Selection of risk factors for preterm birth before pregnancy
We focused on the serum C3 level as a predictor of preterm birth before pregnancy based on a comparison of patient backgrounds in the preterm and full-term birth groups. We performed a ROC analysis to determine a cut-off value for pre-pregnancy serum C3 level that could indicate risks for preterm birth. As a result, the cut-off value for the serum C3 level was set to 84 mg/dl (sensitivity: 0.76, specificity: 0.54, AUC: 0.653, P = 0.043). We then performed a multivariate analysis to select risk factors for preterm birth before pregnancy. Candidate risk factors obtained from the results of univariate analysis comparisons between the two groups included age at pregnancy, which is commonly cited as a risk of preterm birth, history of lupus nephritis, past use of GC pulse therapy, and serum C3 level < 85 mg/dl. The multivariate analysis using these variables identified a history of lupus nephritis (OR: 6.267, 95% CI: 1.750 - 22.443, P = 0.005) and pre-pregnancy serum C3 level < 85 mg/dl (OR: 4.754, 95% CI: 1.373 – 16.461, P = 0.014) as the risk factors for preterm birth (Table 3).
Figure 2 shows the relationship between the number of risk factors (a history of lupus nephritis and pre-pregnancy serum C3 level < 85 mg/dl) and the number of weeks of delivery. The preterm birth rate (preterm birth/total birth ratio) was 7.7% (2/26) in the patients without risk factors, 39.0% (16/41) in those with 1 risk factor, and 66.7% (8/12) in those with 2 risk factors. The greater the number of these risk factors, the higher was the preterm birth rate (P = 0.0007).
Relationship between the pre-pregnancy serum C3 levels and the number of weeks of delivery in the cases of SLEDAI score ≤ 4
In pregnancies with a SLEDAI score > 4, 5 of 7 (71.4%) had preterm birth, and with a SLEDAI score ≤ 4, 21 of 72 (29.1%) had preterm birth. Figure 3 shows the relationship between the pre-pregnancy serum C3 levels and the number of weeks of delivery in the cases of SLEDAI score ≤ 4. The preterm birth rate was significantly higher in the pre-pregnancy low C3 group (<85 mg / dl) (42.1%) than that in the high C3 group (≥ 85 mg / dl) (14.7%) (P = 0.018). The number of weeks of delivery was 37 (35-38) in the low C3 group, which was significantly earlier than that in the high C3 group 38 (37-39) (P = 0.044).
Maternal disease activity during pre-pregnancy and pregnancy
Among the 83 women delivering live babies, 7 (8.4%) had SLE disease activity (medium disease activity or higher) before pregnancy. The pre-pregnancy serum C3 level was 81 (74–92) mg/dl of all. In addition, the presence of proteinuria or cellular casts of 0.5 g/day or more was observed before pregnancy in one woman.
Twenty women (24.0%) had increased SLE disease activity (increase in the SLEDAI score of > 3) during pregnancy. Of the women who had undergone GC therapy before pregnancy, 43 (51.8%) had an increased level of GC (8.9 mg/day). GC was newly introduced in 4 women (4.8%) after their pregnancy was discovered. Due to SLE relapse during pregnancy, immunosuppressants were newly initiated during pregnancy in 5 women.
We compared SLE disease activity between the preterm and full-term birth groups (Figure 1). There was no significant difference in the SLEDAI score and anti-dsDNA antibody titres during either pre-pregnancy or pregnancy. Figure 4 shows a comparison of each parameter during pregnancy in the two groups [high (≥ 85 mg/dl) and low C3 group (< 85 mg/dl)] divided by serum C3 levels at pre-pregnancy. The serum C3 levels were significantly lower in the low C3 group than high C3 group during pre-pregnancy and pregnancy (pre-pregnancy: 95 vs 75 mg/dl, P < 0.001; first trimester: 103 vs 84 mg/dl, P < 0.001; second trimester: 114 vs 91 mg/dl, P < 0.001; and third trimester: 113 vs 99 mg/dl, P < 0.001). The SLEDAI score was lower in the high C3 group before and in the first trimester than those in the low C3 group, but no significant correlation was observed in the anti-dsDNA antibody titre in either group during pre-pregnancy pregnancy.