Predictors of Pregnancy Outcomes in Patients With Lupus

Context Systemic lupus erythematosus (SLE) primarily affects women of childbearing age. Current recommendations about pregnancy in women with SLE are based on studies with methodological limitations. Contribution This longitudinal study enrolled and followed a multiethnic group of pregnant women with inactive or stable mild/moderate SLE at conception. Clinical and disease activity assessments and laboratory and serologic studies were performed at baseline and in each trimester. Caution The timing of patient enrollment precluded ascertainment of first-trimester losses. Implication Most pregnant patients with SLE had favorable pregnancy outcomesfew experienced severe flares. Certain clinical and laboratory variables can identify patients at high risk for adverse pregnancy outcomes. Systemic lupus erythematosus (SLE) primarily affects women of childbearing age. Without treatment with cytotoxic agents, SLE does not adversely affect fertility (1, 2), but fetal and maternal health during pregnancy are concerns. Advice about safety and timing of pregnancy requires identification of clinical and laboratory variables that predict outcomes. It has been suggested that pregnancies in women with SLE result in higher rates of preterm birth, preeclampsia, and fetal loss than in healthy women (310). Previous studies have identified active disease, hypocomplementemia, presence of antidouble-stranded DNA (dsDNA) antibodies, prior nephritis, and presence of antiphospholipid antibodies (aPLs) (68, 1013) as risk factors for adverse pregnancy outcomes (APOs). The effects of pregnancy on SLE activity and the contribution of disease activity to APOs are unclear (10, 1418). Currently, patients with SLE are advised to consider pregnancy during periods of minimal and stable disease (19). However, data supporting this advice are based on retrospective or prospective single-center studies involving few patients, have limited generalizability to multiethnic populations, and are controversial (310). To develop more robust data to inform patients and their physicians about pregnancy in women with SLE, we leveraged the PROMISSE (Predictors of Pregnancy Outcome: Biomarkers in Antiphospholipid Antibody Syndrome and Systemic Lupus Erythematosus) study. PROMISSE is the largest multicenter, multiethnic, and multiracial study to prospectively assess the frequency of APOs, clinical and laboratory variables that predict them, and pregnancy-associated flare rates in women with inactive or mild/moderate SLE at conception. Methods Study Design PROMISSE is a multicenter, prospective observational study of pregnancies in women with SLE (meeting 4 revised American College of Rheumatology criteria) (20), women with SLE and aPLs, women with aPLs alone, and healthy women at low risk for APOs (1 successful pregnancy, no prior fetal death, and <2 miscarriages at <10 weeks' gestation). Criteria for the healthy control participants were designed to minimize factors unrelated to SLE that might affect outcomes. This article focuses on patients with SLE with or without aPLs (Appendix Figure). Those with aPLs with or without SLE were previously reported (21). Appendix Figure. Study flow diagram. Pregnant women at <12 weeks' gestation with aPL positivity and SLE and healthy control participants were screened. Exclusion criteria included multifetal pregnancy, prednisone use >20 mg/d, blood pressure >140/90 mm Hg, urinary protein-creatinine ratio >1000 mg/g on 24-h or spot urine collection, serum creatinine level >1.2 mg/dL, screening too late in pregnancy, and previous enrollment in PROMISSE. Healthy pregnant women were enrolled if they had 1 successful pregnancy, no history of fetal death, <2 miscarriages at <10 weeks' gestation, and no aPLs. The current study included all PROMISSE patients who met American College of Rheumatology criteria for SLE. aPL = antiphospholipid antibody; APO = adverse pregnancy outcome; PROMISSE = Predictors of Pregnancy Outcome: Biomarkers in Antiphospholipid Antibody Syndrome and Systemic Lupus Erythematosus; SLE = systemic lupus erythematosus. * Did not meet criteria for the primary study outcome (elective termination [n = 1], incompetent cervix [n = 2], and premature preterm rupture of membranes [n = 1]). See Table 1. Did not meet criteria for the primary study outcome (indicated delivery for poor obstetric history [n = 1], SLE flare [n = 1], bleeding placental abruption [n = 1], termination for complete heart block [n = 1], fetal death due to trisomy 18 [n = 1], and premature preterm rupture of membranes and/or premature labor [n = 15]). See Table 1. Patient Population Pregnant patients were enrolled between September 2003 and December 2012 at 8 sites in the United States and 1 in Canada. Institutional review boards approved the protocol and consent forms, and written, informed consent was obtained from patients. Consecutive pregnant women meeting inclusion criteria were recruited up to 12 weeks' gestation, which precluded ascertainment of first-trimester losses. Only 1 pregnancy for each patient was included. Inclusion criteria for enrollment were singleton intrauterine pregnancy, age 18 to 45 years, and hematocrit greater than 26%. Because the overall goal of PROMISSE was to identify risk factors for and mechanisms of APOs specifically attributable to lupus, aPLs, or both, exclusion criteria were based on the following potential causes of APOs: prednisone use greater than 20 mg/d, urinary proteincreatinine ratio greater than 1000 mg/g, erythrocyte casts on urinalysis, serum creatinine level greater than 1.2 mg/dL, diabetes mellitus, and blood pressure greater than 140/90 mmHg at screening. Definition of SLE Disease Activity and Flares During Pregnancy Investigators used the Systemic Lupus Erythematosus Pregnancy Disease Activity Index (SLEPDAI), an instrument incorporating history, physical examination, and laboratory measures to gauge lupus activity. The SLEPDAI was modified to discount physiologic changes in pregnancy that mimic disease activity to ensure attribution to lupus (19, 22, 23). A composite was used to define flares as mild/moderate or severe. The composite was similar to that used in the SELENA (Safety of Estrogens in Lupus Erythematosus, National Assessment) trial, except SLEPDAI was substituted for the SLEDAI (24) instrument used in the SELENA trial. The composite included SLEPDAI score; assessment of new or worsening disease activity, medication changes, and hospitalizations not captured in the SLEPDAI score; and Physician's Global Assessment (PGA) score (range of 0 [inactive disease] to 3 [severe disease]). Study investigators were trained with paper pregnant patients with SLE and case report forms prepared by one of the authors (gold standard). The average correlation among investigator responses with the gold standard was 0.89 (95% CI, 0.83 to 0.95), and mean scores were within15% of the gold standard. Interrater reliability estimated by the intraclass correlation coefficient was also high (0.78 [CI, 0.61 to 0.89]). In some situations, the SLEPDAI and PGA were not completed because required serologic or complete blood count data were unavailable. In those cases, flare status was based on review of medical records and evidence of a clinical change or addition of new medications. Adverse Pregnancy Outcomes Adverse pregnancy outcomes included 1 or more of the following: 1) fetal death after 12 weeks' gestation unexplained by chromosomal abnormalities, anatomical malformation, or congenital infection; 2) neonatal death before hospital discharge due to complications of prematurity, placental insufficiency (for example, abnormal fetal surveillance test results, abnormal Doppler flow velocimetry waveform analysis suggestive of fetal hypoxemia, or oligohydramnios [25]), or both; 3) preterm delivery or termination of pregnancy at less than 36 weeks due to gestational hypertension, preeclampsia, or placental insufficiency; and 4) small-for-gestational-age (SGA) neonate, defined as one with a birthweight below the fifth percentile without anatomical or chromosomal abnormalities. Screening and Follow-up Visits Screening evaluations included history, assessment of American College of Rheumatology criteria, physical examination, complete blood count, comprehensive metabolic panel, urinalysis, and random or 24-hour urine collection for proteincreatinine ratio (if dipstick grade was >1+). Serologic profiles (anti-dsDNA antibodies, anti-SSA/Ro antibodies, anti-SSB/La antibodies, and C3 and C4 levels) were determined at local laboratories. Tests for aPLs included anticardiolipin antibodies (IgG, IgM, and IgA), anti2-glycoprotein I antibodies (IgG and IgM), and lupus anticoagulant (LAC) and were performed at core laboratories (21). Tests were repeated each trimester. The SLEPDAI and PGA were scored at the screening, second-trimester (20 to 23 weeks), and third-trimester (32 to 35 weeks) visits. Statistical Analysis We evaluated bivariate associations of APO status and each predictor variable with chi-square and Fisher exact tests for categorical variables and the t test for continuous variables. Multivariable analyses were conducted using logistic regression models. Three separate models were fit to allow for greater flexibility in modeling the potentially time-varying relationship between predictors that were measured repeatedly during pregnancy and APOs. Model 1 included variables measured at screening to identify baseline characteristics that are predictive of an APO occurring at any time. Model 2 considered these baseline variables plus variables measured at 20 to 23 weeks to predict APOs after 23 weeks, and model 3 considered additional variables measured at 32 to 35 weeks to predict late APOs. Decisions on variable selection at each step of model development were based on both clinical factors and statistical significance. For example, change in C3 level is routinely monitored in patients with SLE and was therefore prioritized