Background/Purpose: There are conflicting results about the effect pregnancy has on the health of systemic lupus erythematosus (SLE) women. The objective of the current analysis was to estimate the effect of pregnancy on disease flares in SLE using a stratified Cox model.

Methods: Data were prospectively collected in a university-based cohort from 1987-2015. Patients met ACR or SLICC criteria for SLE. All women aged 14-45 years with >1 measurement of disease activity were included, regardless of pregnancy status. Visits were classified into one of three exposure categories: pregnant, 1-year postpartum period, or non-pregnant/non-postpartum period (unexposed). Patients were allowed to switch from one category of exposure to another between any two clinic visits based on pregnancy status. At each visit, disease activity was measured by Physician Global Assessment (PGA) and SELENA-SLEDAI. Flares during follow-up were defined as: 1) change in PGA ≥1 from previous visit and 2) change in SELENA-SLEDAI ≥4 from previous visit. Patients with a >1-year gap in study visits were considered lost to follow-up, but were allowed to re-enter the cohort. Hazard ratios (HR) between 1) pregnant and unexposed periods and 2) postpartum and unexposed periods were estimated with a stratified Cox model, a model that adjusts baseline hazards based on the number of previous flares. Hydroxychloroquine (HCQ) was explored as a time-varying covariate, with effect measure modification defined by likelihood ratio test (α=0.20).

Results: There were 1349 patients, including 398 pregnancies in 304 patients. Median age at cohort entry was 31 years, and median follow-up was 4 years. The crude incidence of flares based on the PGA definition was 60.7 per 100 person-years (PY) during pregnancy compared to 40.2 per 100 PY during non-pregnant/non-postpartum periods (Table 1). Stratified Cox models estimated an increased rate of flare during pregnancy (HR: 1.59; 95% CI: 1.27, 1.96), with no evidence of an increased rate during the postpartum period. There was effect modification by HCQ use. Among periods of no HCQ use, the HR of flares in pregnancy compared to non-pregnant/non-postpartum periods was estimated to be 1.83 (95% CI: 1.34, 2.45) compared to 1.26 (95% CI: 0.88, 1.69) among periods with HCQ use (likelihood ratio p-value: 0.04). When flares were defined by SELENA-SLEDAI, results were similar, with an increased rate of flare during pregnancy but no observed increased rate of flare postpartum compared to non-pregnant/non-postpartum periods. No evidence of effect modification by HCQ use was found when flares were defined by SELENA-SLEDAI.

Conclusion: Our study supports and extends previous findings that the incidence of flare is increased during pregnancy. Continuing HCQ in pregnancy, however, appeared to mitigate the risk of flare during pregnancy.  We did not find evidence of an increased rate of flare postpartum.