Background/Purpose:

Prior reports suggest that disease activity in rheumatoid arthritis (RA) improves in the setting of pregnancy.  However, prior studies report half of patients having moderate-severe disease activity throughout pregnancy; these are levels of activity unacceptable in our current RA treatment paradigm.  The use of TNF-inhibitors remains controversial in pregnancy with many clinicians more comfortable with prednisone. The purpose of this study was to prospectively evaluate RA disease activity and medication use as predictors of pregnancy outcomes.

Methods:

RA pregnancies were followed prospectively in a single clinic. An assessment of each patient was performed at all pregnancy and postpartum visits, and included demographic data and measures of disease activity, such as the physician global assessment (PGA) and 28 joint DAS-CRP. A DAS-CRP <2.6 was considered remission.  Medication use was determined in collaboration between the physician and patient. Outcome measures included the presence of a miscarriage, preterm birth or preeclampsia. Non-parametric testing was used for the analysis.

Results:

The 31 pregnancies in this cohort had a mean age of delivery of 32.5 years; 23 (74.2%) of mothers were Caucasian. Nine (29%) were exposed to TNF-inhibitors, while 15 (48.4%) were exposed to prednisone. 2 pregnancies (6.5%) ended in early miscarriages, neither with prednisone or TNF-inhibitor exposure and both with low RA activity.  Of the 29 live births, preterm delivery occurred in 5 (17.2%) and preeclampsia in 3 (10.9%). Remission was achieved in 54% of women in the 1^st trimester, 77% in the 2^nd trimester, 67% in the 3^rd trimester, and 61% postpartum. There was a strong correlation between DAS-CRP and PGA (R=0.86). Higher disease activity in the 1^st trimester was associated with an increased risk of preterm birth (DAS-CRP = 5.58 preterm vs. 2.20 term, p=0.028; PGA= 30.0 preterm vs. 75.2 term, p=0.019). All pregnancies fell into one of three 1^st trimester medication groups (table). Hydroxycholoroquine and sulfasalazine were evenly distributed in each group and did not appear to impact pregnancy outcomes.  The patients with neither TNF-inhibitor nor prednisone exposure had a mean 1^st trimester DAS-CRP of 1.63, while patients exposed to prednisone or TNF-inhibitor had a mean DAS-CRP of 3.92 or 3.19, respectively. Over half of patients (54.2%) exposed to prednisone had preterm birth and/or preeclampsia compared to 14.3% of those with TNF-inhibitor exposure.  

Conclusion:

In the 1^st trimester, high RA activity and prednisone use are both associated with an increased risk of preterm birth and/or preeclampsia. While both the prednisone and TNF-inhibitor groups had similar disease activity in the 1^st trimester, prednisone use was significantly associated with a higher risk of adverse outcome. This study suggests that TNF-inhibitors may be preferable to prednisone in RA pregnancies.

Table: Medication group and pregnancy outcome