Discussion

This nationwide population-based study demonstrated that after adjusting for potential confounders, mothers with RA were 1.47, 1.20, 2.22 and 1.19 times more likely than unaffected mothers to have LBW, SGA babies, preeclampsia and undergo CS, respectively. Our findings were parallel with some previous studies that found that mothers with RA had an increased risk of preeclampsia,4 delivery by the caesarean method4 6 10 and LBW infants,4 13 compared with unaffected women. Nevertheless, a study by Reed et al6 failed to observe a significant effect of maternal RA on the risk of LBW. Furthermore, several studies also observed a significant relationship between preterm birth and RA in mothers,4 6 13 14 although we found mothers with RA had no higher odds of having preterm births than unaffected mothers, regardless of the definition of preterm cut-off points. Such inconsistent results may be due to differences in patient selection, the number of patients included in the series, methodological differences in study design, the presence or lack of a comparison group and the availability of advanced obstetric care. For example, Reed et al6 analysed inpatient data with predominantly non-Hispanic white women. Skomsvoll et al4 recruited women with various diagnoses, including RA, juvenile RA and ankylosing spondylitis. The relationship between maternal RA and adverse pregnancy outcomes cannot be specifically addressed.

To our knowledge, our study is the most complete nationwide population-based study ever conducted to assess the risk of adverse pregnancy outcomes among women with RA. Unlike earlier studies that included participants from diverse ethnic groups, over 98% of Taiwan's residents are of Chinese Han ethnicity. Whereas the homogenous population may exempt our study from potential confounding by race, it also limits the generalisability to other ethnic groups. In addition, we used nationwide population-based datasets linking the NHIRD with NBCR, which leaves little room for selection and non-response biases. The very large sample size in this study provides ample statistical power to detect differences in the risk of adverse birth outcomes comparing pregnant women with and without RA.

However, four limitations merit attention. First, the NHIRD lacks clinical information and therefore did not allow us to differentiate study participants according to the severity of their RA. Second, NHIRD uses discharge diagnoses provided by treating physicians, and no standardised criteria were used to define cases, which could leave some room for bias due to case misclassification. Furthermore, because women with RA might be miscategorised as unaffected mothers if they had been diagnosed long before being pregnant (ie, earlier than 1997 but not 1998–2003), our results would probably be biased towards the null. Third, multiple pregnancies may be a concern in the association investigated. However, due to the similar results observed among women in their first pregnancy and among all pregnancies, we chose to report all pregnancies to possess ample statistical power for examining a rarer outcome of preeclampsia. Finally, effects of pharmacotherapy in women with RA may be significant.15 Because the NHIRD does not include complete information regarding medications taken during pregnancy, it is not possible for us to assess the confounding role of medications in the relationship between RA and adverse birth outcomes.

Our findings suggest a need for more intensive prenatal care among pregnant mothers with RA. In addition, early intervention should be considered to counter potential adverse obstetric outcomes for pregnant women with RA. Future studies are needed to examine the potential mechanisms by which RA produces adverse pregnancy outcomes. Future studies exploring the association between adverse birth outcomes and the severity of RA and medications taken during pregnancy are also warranted, to enable more specific interpretation of these findings.