Reply

To the Editor: We read with considerable interest the findings presented in the study by Dr. Remaeus et al, who reported on the outcomes of antirheumatic treatment in pregnancy (1). Remaeus et al concluded that women with psoriatic arthritis (PsA) receiving antirheumatic treatment had a higher risk of their pregnancy resulting in preterm birth and caesarean delivery compared with women without PsA. The authors attributed this increased risk to antirheumatic treatment, especially treatment with biologic disease-modifying antirheumatic drugs (bDMARDs) during pregnancy. The authors made some assumptions that we would like to challenge. We were particularly surprised that bDMARD therapy was used as a proxy for disease activity. Previous studies have demonstrated that disease burden improved with the use of bDMARDs. Ursin et al reported significantly lower disease activity in pregnant women with PsA who took tumor necrosis factor inhibitors (2). We also reported similar findings in a cohort of PsA patients (3). Furthermore, most of the published evidence in relation to pregnancy outcomes in patients with inflammatory arthritis point to better outcomes when the disease is well controlled at the time of conception and throughout pregnancy. The authors suggested that the risk of spontaneous preterm birth could be up to 4-fold higher in pregnant women who took bDMARDs than in pregnant women not exposed to bDMARDs. However, this finding may reflect a selection bias in the study design that the authors have not recognized, as the underlying mechanisms, risk factors, and etiology of preterm birth are not completely understood; therefore, many confounders may have contributed to their results. It is unclear what the psoriasis status is in their study population and whether bDMARDs were prescribed for active psoriasis rather than PsA. Findings from a previous study suggested that psoriasis is associated with adverse maternal and fetal outcomes (4). Given the lack of data on disease activity and on other patient factors, such as metabolic syndrome, that are significantly associated with psoriasis, it is unclear to what extent preterm birth could be directly attributed to the effects of bDMARD treatment. Finally, the factors involved in rates of caesarean section that Remaeus et al described are evenmore varied and uncontrolled; thus, these data are virtually unusable. In conclusion, we agree that disease activity and severity are important factors that should be optimally controlled in patients with PsA before they consider pregnancy. We disagree with the assumption that treatment with bDMARDs in PsA patients during pregnancy is a good surrogate for the clinical metrics reported in the study from Remaeus et al. The conclusion that bDMARDs are associated with adverse outcomes in PsA pregnancies is counterintuitive and potentially a dangerous message to this population. Author disclosures are available at https://onlinelibrary.wiley.com/action/ downloadSupplement?doi=10.1002%2Fart.42253&file=art42253-sup0001-Disclosureform.pdf.


To the Editor:
We read with considerable interest the findings presented in the study by Dr. Remaeus et al, who reported on the outcomes of antirheumatic treatment in pregnancy (1). Remaeus et al concluded that women with psoriatic arthritis (PsA) receiving antirheumatic treatment had a higher risk of their pregnancy resulting in preterm birth and caesarean delivery compared with women without PsA. The authors attributed this increased risk to antirheumatic treatment, especially treatment with biologic disease-modifying antirheumatic drugs (bDMARDs) during pregnancy.
The authors made some assumptions that we would like to challenge. We were particularly surprised that bDMARD therapy was used as a proxy for disease activity. Previous studies have demonstrated that disease burden improved with the use of bDMARDs. Ursin et al reported significantly lower disease activity in pregnant women with PsA who took tumor necrosis factor inhibitors (2). We also reported similar findings in a cohort of PsA patients (3). Furthermore, most of the published evidence in relation to pregnancy outcomes in patients with inflammatory arthritis point to better outcomes when the disease is well controlled at the time of conception and throughout pregnancy.
The authors suggested that the risk of spontaneous preterm birth could be up to 4-fold higher in pregnant women who took bDMARDs than in pregnant women not exposed to bDMARDs. However, this finding may reflect a selection bias in the study design that the authors have not recognized, as the underlying mechanisms, risk factors, and etiology of preterm birth are not completely understood; therefore, many confounders may have contributed to their results. It is unclear what the psoriasis status is in their study population and whether bDMARDs were prescribed for active psoriasis rather than PsA. Findings from a previous study suggested that psoriasis is associated with adverse maternal and fetal outcomes (4). Given the lack of data on disease activity and on other patient factors, such as metabolic syndrome, that are significantly associated with psoriasis, it is unclear to what extent preterm birth could be directly attributed to the effects of bDMARD treatment. Finally, the factors involved in rates of caesarean section that Remaeus et al described are even more varied and uncontrolled; thus, these data are virtually unusable.
In conclusion, we agree that disease activity and severity are important factors that should be optimally controlled in patients with PsA before they consider pregnancy. We disagree with the assumption that treatment with bDMARDs in PsA patients during pregnancy is a good surrogate for the clinical metrics reported in the study from Remaeus et al. The conclusion that bDMARDs are associated with adverse outcomes in PsA pregnancies is counterintuitive and potentially a dangerous message to this population. disease activity and/or severity during pregnancy are not perfect measures; instead, information on a validated measure of disease activity would be ideal. Nevertheless, given the heterogeneity of the PsA diagnosis and the spectra of treatment strategies, we believe that assessing the impact of treatment exposure in detail provides important insights that augment the understanding of risks associated with a diagnosis of PsA. In our study, we found that adverse outcomes were more common among pregnant women with PsA than among pregnant women without PsA and that the highest risks were seen among those who were treated with antirheumatic medications, especially bDMARDs, during pregnancy. Although the use of bDMARDs during pregnancy has lately become more widespread, in our cohort of PsA pregnancies, 11% of women took bDMARDs during pregnancy, indirectly suggesting that women continuing this treatment most probably had severe disease.
We do not attribute the increased risk of preterm birth and caesarean delivery to bDMARD treatment as such; however, we described an increased risk of adverse pregnancy outcomes in PsA patients who took bDMARDs compared with pregnancy outcomes in those without PsA. Our interpretation, as also expressed in our study, is that the increased risk of adverse outcomes in pregnancies with antirheumatic treatment is probably attributed to disease severity rather than an effect of the treatment itself.
We agree that the underlying mechanisms of spontaneous preterm birth in women with PsA are not fully understood. We also endorse that metabolic factors might be of importance, as explored in a previous study from our group (2). In our current study, we were able to take known risk factors for spontaneous preterm birth into account. Very few of the 921 pregnant women with PsA in our cohort had hypertension or diabetes before pregnancy, although more women with PsA than without PsA were categorized as obese. However, in distribution of women with PsA in terms of obesity, the lowest proportion of obesity was among those who were taking bDMARDs, indicating that metabolic syndrome was not the main explanation for the increased risk of preterm birth in this cohort.
In conclusion, our findings support that the risk of adverse pregnancy outcomes varies with presence, timing, and also type of antirheumatic treatment in women with PsA. The outcomes should not be interpreted as caused by the medications. It is of great importance to further elucidate the impact of maternal disease activity on pregnancy outcomes in PsA. We read with great interest the article by Dr. Xue and colleagues (1) on the discovery of the expansion of ficolin-1-positive (FCN-1+) monocyte-derived dendritic cells (mo-DCs), which were reported to be correlated with the severity of skin disease in patients with diffuse cutaneous systemic sclerosis (dcSSc). Advances in single-cell technologies have enabled investigators to use high-resolution gene profiling and gene identification to understand cellular heterogeneity in tissue. In their study, Xue et al used single-cell RNA-sequencing analysis to identify a gene expression signature of the disease-specific FCN-1+ myeloid cell type in dcSSc skin. They illustrated that FCN-1+ cells belong to a single cluster that was best characterized by gene expression of FCN1, EREG, S100A8, and THBS1. In addition, the frequency of perivascular FCN-1+ myeloid cells was higher in dcSSc patients who had elevated modified Rodnan skin thickness (MRSS) scores than in patients with low MRSS scores or in skin from healthy donors. Thus, the authors proposed that FCN-1+ cells represent a dcSSc-specific myeloid population. Noteworthy, this cell population was mostly expanded in selected dcSSc patients (i.e., 2 of 12 patients), displaying a proinflammatory gene profile (2).
Interestingly, the gene expression profile of FCN-1+ cells was reminiscent of the inflammatory monocyte-like (Inf-Mo-like) cells reported in the inflamed colon of patients with Crohn's disease (CD) (3). Frequency of Inf-Mo-like cells was positively correlated with disease severity, as measured in CD patients by the simple endoscopic score for CD. Gene set enrichment analysis (4) further indicated that cluster E, which included Inf-Mo-like cells reported in inflamed CD colon (see Figure 6 in Chapuy et al [3]) was enriched in genes expressed by cluster 5 that comprise FCN-1+ cells in the skin of dcSSc patients. The normalized enrichment score was 1.650 with a false discovery rate q value of 0.0275, and the core genes enriched included EREG, S100A9, THBS1 FCN1, TIMP1, IL1RN, SAMSN1, C5RA1, OLR1, CD300E, APOBEC3A, FPR1, SERPINA1, and GK. Thus, we note the