Biologics Use for Psoriasis during Pregnancy and Its Related Adverse Outcomes in Pregnant Women and Newborns: Findings from WHO Pharmacovigilance Study

Psoriasis is an autoimmune inflammatory disorder that affects approximately 2% of the global population, with approximately half of these patients being female [1]. Among female patients, 75% experience onset before the age of 40 years [2], indicating a significant impact on those of reproductive age [3]. During pregnancy, more than half of females with psoriasis experience a resolution of symptoms, while 23% report a worsening of symptoms [4]. The association between psoriasis in pregnant females and adverse pregnancy outcomes has been debated due to inconsistent findings [5]. However, a recent large-scale study has reported increased risks of gestational diabetes, gestational hypertension, preterm birth, and low birth weight, which correlate with the severity of the disease [6].

Therefore, while it is necessary to continue treatment during pregnancy for patients with extensive psoriasis, only 28% of dermatologists are familiar with the guidelines for psoriasis treatment during pregnancy [7], leading to a prevailing tendency to consider all biologics contraindicated during this period [8]. Available systemic treatments include tumor necrosis factor (TNF)-α inhibitors like infliximab, adalimumab, etanercept, and certolizumab pegol. Although findings on pregnancy-related adverse outcomes with TNF-α inhibitors have been inconsistent, a recent consensus suggests that, except for golimumab, these inhibitors do not result in adverse maternal or fetal outcomes [9, 10]. Certolizumab pegol, particularly, is considered the safest due to its inability to cross the placenta [11].

However, the safety of other biologics used for psoriasis, such as interleukin (IL)-17/IL-23 inhibitors (e.g., ustekinumab), IL-17 inhibitors (e.g., secukinumab, brodalumab, ixekizumab, bimekizumab), and IL-23 inhibitors (e.g., guselkumab, tildrakizumab, risankizumab), remains uncertain due to the paucity of data despite some reports [12]. Therefore, this study aimed to utilize World Health Organization (WHO) global pharmacovigilance data to investigate maternal, fetal, and neonatal outcomes associated with TNF-α inhibitors and other systemic treatments for psoriasis. By analyzing the adverse outcomes of each individual drug, this research will provide the most extensive dataset on biologic drug exposure to treat psoriasis during pregnancy. It will offer healthcare practitioners critical insights into the safety and potential adverse effects of various biologics that have not been comprehensively studied.

We conducted a case-non-case disproportionality analysis to assess associations between adverse maternal, fetal, or neonatal outcomes and the use of systemic psoriasis treatments compared to TNF-α inhibitors and certolizumab pegol in pregnant individuals. The systemic treatments for psoriasis examined, along with their Anatomical Therapeutic Chemical codes, include TNF-α inhibitors (etanercept, L04AB01; infliximab, L04AB02; adalimumab, L04AB04; and certolizumab pegol, L04AB05), IL-12/IL-23 inhibitor (ustekinumab, L04AC05), IL-17 inhibitors (secukinumab, L04AC10; brodalumab, L04AC12; ixekizumab, L04AC13; and bimekizumab, L04AC21), and IL-23 inhibitors (guselkumab, L04AC16; tildrakizumab, L04AC17; and risankizumab, L04AC18). We excluded non-biologic drugs used for psoriasis, such as methotrexate and acitretin, due to their known teratogenic effects [13].

This study used data from VigiBase, a global pharmacovigilance database managed by the Uppsala Monitoring Centre. As of 2024, VigiBase has collected more than 140 million individual case safety reports from over 170 countries and territories [14, 15]. All methods were performed in accordance with the relevant guidelines and regulations. The project was approved by the Uppsala Monitoring Centre and the Institutional Review Boards of Kyung Hee University. Since the data do not contain any identifiable information, the Institutional Review Board of Kyung Hee University waived the need for obtaining informed consent.

Medical Dictionary for Regulatory Activities (MedDRA) version 26.0 is integrated with VigiBase [14, 15]. We retrieved individual case safety reports from VigiBase, which include one or more pregnancy-related reactions and indicate systemic psoriasis treatment. Pregnancy-related reactions were defined by MedDRA categories as follows: pregnancy, puerperium, and perinatal conditions (system organ class); fetal and neonatal investigations (high level group term, HLGT); neonatal and perinatal conditions (HLGT); neonatal respiratory disorders (HLGT); exposures associated with pregnancy, delivery, and lactation (high level term, HLT); fetal therapeutic procedures (HLT); induced abortions (HLT); and obstetric therapeutic procedures (HLT) [16‒18]. The exact MedDRA terms utilized to identify each outcome are detailed in online supplementary Table S1 (for all online suppl. material, see https://doi.org/10.1159/000542217). We only retained reports mentioning terms primarily associated with pregnancy, excluding reports containing terms that are not always associated with pregnancy. Details are further explained in online supplementary Table S2.

We categorized exposure types into “exposure during pregnancy,” “exposure via other modality,” and “inadequate exposure timing.” Reports of exposure via breast milk or the father were classified as “exposure via other modality,” while those mentioning exposure before or after pregnancy, or without specifying a period within pregnancy, were categorized as “inadequate exposure timing.” Only “exposure during pregnancy” was included, and the other two categories were excluded. Further details are provided in online supplementary Table S3.

Our case-non-case disproportionality analysis included data on geographical reporting regions (Africa, America, Southeast Asia, Europe, Eastern Mediterranean, and Western Pacific), reporting years (1968–2024), patient outcomes (alive, fatal, and unknown), and types of reporters (health professionals, non-health professionals, and unknown) [14, 15].

Cases were defined as individual case safety reports mentioning pregnancy-related adverse outcomes, while non-cases were all other reports without mention of the adverse event. Pregnancy-related adverse outcomes, including maternal, fetal, and neonatal adverse outcomes, were grouped into seven categories: abortion or stillbirth (spontaneous abortion, stillbirth, and induced abortion), pregnancy complications (gestational diabetes mellitus; gestational hypertension; and fetal growth restriction), delivery complications, postpartum complications (postpartum hemorrhage), preterm birth, neonatal complications (low birth weight, neonatal respiratory disorders, neonatal metabolic-endocrine disorders, neonatal neurologic disorders, neonatal hematological disorders, neonatal hepatobiliary disorder [neonatal jaundice], congenital/neonatal infection, neonatal cardiovascular disorder, and neonatal gastrointestinal disorder), and congenital malformations (cardiovascular malformation, gastrointestinal malformation, musculoskeletal and connective tissue malformation, neurologic malformation, and renal dysplasia or agenesis) [16, 17]. Further details are available in online supplementary Table S4.

The reporting odds ratio (ROR), calculated based on the contingency table provided in online supplementary Table S5, is the ratio of the odds of adverse pregnancy, fetal, and neonatal outcomes associated with biologics for psoriasis of interest to the odds of the same outcomes associated with TNF-α inhibitors. Statistically significant disproportionality between cases and controls indicated a safety signal. Significant overreporting among reports suspecting a drug of interest was considered present when the lower limit of the ROR 95% confidence interval (CI) exceeded 1.00, indicating that the outcome of interest occurred more frequently in reports suspecting a drug of interest compared to reports suspecting certolizumab. A two-sided test with a p value threshold of <0.05 was utilized to determine statistical significance. All analyses were performed with SAS software (version 9.4; SAS Inc., Cary, NC, USA) [19].

Sensitivity analyses within reports submitted by health professionals were conducted to evaluate the credibility of our results. Additionally, we identified the individual’s age, the region of the report, reporter type, and reporting year as confounding variables. Multivariable logistic regression analyses were performed to adjust for these variables and minimize potential biases.

The overall flow of the study is shown in Figure 1; out of over 140 million reports from VigiBase, we extracted 38,849 reports of biologics-related adverse outcomes. Out of these reports, 8,350 reports were excluded by the exclusion criteria. Finally, 6,518 reports of pregnancy-related adverse outcomes were extracted from 30,499 reports. The characteristics of adverse outcomes related to biologics are reported in Table 1. The number of biologics-related adverse outcomes has been consistently rising, with TNF-α inhibitors taking up 20,325 reports, which is 66.64% of total reports (Table 1).

The numbers of reports of pregnancy-related adverse outcomes are detailed in online supplementary Table S6. For the TNF-α inhibitors (etanercept, infliximab, adalimumab, and certolizumab pegol), the reported pregnancy-related adverse outcomes were 634, 619, 1,337, and 469, respectively. For the IL-12/IL-23 inhibitor (ustekinumab), 140 pregnancy-related adverse outcomes were reported. For the IL-17 inhibitors (secukinumab, brodalumab, ixekizumab, and bimekizumab), the reported pregnancy-related adverse outcomes were 105, 2, 11, and 2, respectively. For the IL-23 inhibitors (guselkumab, tildrakizumab, and risankizumab), the reported pregnancy-related adverse outcomes were 18, 1, and 87, respectively (online suppl. Table S6). Out of a total of 6,518 biologics-related adverse outcomes, there were 3,425 reports related to abortion and stillbirth, 1,968 reports of pregnancy complications, 1,198 reports of delivery complications, 173 reports of postpartum complications, 1,060 reports of preterm birth, 476 reports of neonatal complications, and 207 reports of congenital malformations (online suppl. Table S6).

The frequency of the pregnancy-related adverse outcomes was compared with individual biologics to TNF-α inhibitors (Fig. 2-4). The overall frequencies of pregnancy-related adverse outcomes of biologics compared to TNF-α inhibitors were found to be low except brodalumab.

For the IL-12/IL-23 inhibitor ustekinumab, the overall frequency of pregnancy-related adverse outcomes was low (ROR, 0.27; 95% CI: 0.21–0.35). Specifically, the frequency of abortion and stillbirth was low (ROR, 0.79; 95% CI: 0.63–0.98), as were pregnancy complications (ROR, 0.59; 95% CI: 0.45–0.78), postpartum complications (ROR, 0.12; 95% CI: 0.02–0.82), and neonatal complications (ROR, 0.54; 95% CI: 0.31–0.94) (Fig. 2).

The disproportionality analysis of IL-17 inhibitors (secukinumab, brodalumab, ixekizumab, and bimekizumab) showed a generally lower frequency of pregnancy-related adverse outcomes, except for brodalumab. Specifically, secukinumab exhibited a low overall frequency of pregnancy-related adverse outcomes (ROR, 0.17; 95% CI: 0.13–0.22). In particular, adverse outcomes related to abortion and stillbirth, pregnancy complications, delivery complications, preterm birth, and neonatal complications were found to be low, with respective RORs of 0.74 (95% CI: 0.57–0.95), 0.47 (95% CI: 0.33–0.65), 0.35 (95% CI: 0.22–0.56), 0.38 (95% CI: 0.23–0.63), and 0.42 (95% CI: 0.21–0.85). Brodalumab’s overall frequency of pregnancy-related adverse outcomes was low (ROR, 0.20; 95% CI: 0.02–2.21). Ixekizumab also showed a low overall frequency of pregnancy-related adverse outcomes (ROR, 0.05; 95% CI: 0.03–0.08). Specifically, adverse outcomes related to abortion and stillbirth, pregnancy complications, delivery complications, and preterm birth were low, with respective RORs of 0.23 (95% CI: 0.12–0.44), 0.12 (95% CI: 0.04–0.40), 0.15 (95% CI: 0.04–0.63), and 0.09 (95% CI: 0.01–0.63). Lastly, bimekizumab’s overall frequency of pregnancy-related adverse outcomes was low (ROR, 0.10; 95% CI: 0.01–0.71) (Fig. 3).

The disproportionality analysis of IL-23 inhibitors (guselkumab, tildrakizumab, and risankizumab) showed a lower frequency of pregnancy-related adverse outcomes. Specifically, guselkumab exhibited a low overall frequency of pregnancy-related adverse outcomes (ROR, 0.09; 95% CI: 0.05–0.15). In particular, adverse outcomes related to abortion and stillbirth, pregnancy complications, and delivery complications were found to be low, with respective RORs of 0.54 (95% CI: 0.31–0.95), 0.37 (95% CI: 0.17–0.81), and 0.09 (95% CI: 0.01–0.63). Tildrakizumab also showed a low overall frequency of pregnancy-related adverse outcomes (ROR, 0.02; 95% CI: 0.00–0.14). Lastly, risankizumab’s overall frequency of pregnancy-related adverse outcomes was low (ROR, 0.38; 95% CI: 0.25–0.58). Specifically, adverse outcomes related to pregnancy complications, delivery complications, preterm birth, and neonatal complications were found to be low, with respective RORs of 0.48 (95% CI: 0.30–0.76), 0.07 (95% CI: 0.02–0.28), 0.08 (95% CI: 0.02–0.33), and 0.30 (95% CI: 0.09–0.94). However, adverse outcomes related to abortion and stillbirth showed higher frequency (ROR, 1.87; 95% CI: 1.32–2.63; Fig. 4).

Similar findings were observed when biologics were compared to certolizumab pegol, known to be the safest TNF-α inhibitor (online suppl. Table S7). Notably, infliximab exhibited an overall low frequency of pregnancy-related adverse outcomes (ROR, 0.71; 95% CI: 0.55–0.92). Etanercept showed a similar frequency of pregnancy-related adverse outcomes (ROR, 1.00; 95% CI: 0.77–1.31), while adalimumab exhibited a higher frequency of pregnancy-related adverse outcomes compared to certolizumab pegol (ROR, 1.42; 95% CI: 1.11–1.81). All TNF-α inhibitors showed a higher frequency of adverse outcomes related to abortion and stillbirth, with infliximab (ROR, 1.46; 95% CI: 1.24–1.72), etanercept (ROR, 1.29; 95% CI: 1.10–1.51), and adalimumab (ROR, 1.41; 95% CI: 1.23–1.62; online suppl. Table S7).

A multivariable analysis was conducted, adjusting for individuals’ age, reporting year, reporter type, and reporting region, to mitigate biases in comparisons between TNF-α inhibitors and specific biologics (online suppl. Table S8), as well as between certolizumab pegol and specific biologics (online suppl. Table S9). Additionally, a sensitivity analysis was conducted using 17,381 reports (56.99% of total reports) from health professionals (online suppl. Tables S10–15). Multivariable and sensitivity analyses showed results that were generally consistent with the primary analysis. However, etanercept and adalimumab had a lower or comparable frequency of pregnancy-related adverse outcomes compared to certolizumab pegol, which differed slightly from the primary result.

We conducted a disproportionality analysis using data from the WHO pharmacovigilance database to compare the frequency of pregnancy-related adverse outcomes in pregnant women and newborns associated with biologics used for psoriasis, specifically against TNF-α inhibitors and certolizumab pegol. All IL-12/IL-23 inhibitors, IL-17 inhibitors, and IL-23 inhibitors, except brodalumab, showed a lower overall frequency of these adverse outcomes in pregnant women and newborns compared to TNF-α inhibitors and certolizumab pegol. However, risankizumab had a higher reporting frequency of abortion and stillbirth. When comparing other TNF-α inhibitors to certolizumab pegol, infliximab had a lower frequency of pregnancy-related adverse outcomes in pregnant women and newborns, etanercept showed no significant difference, and adalimumab had a higher frequency. These biologics also exhibited a higher frequency of reports on abortion and stillbirth compared to certolizumab pegol, with the highest RORs for spontaneous abortions. Multivariable and sensitivity analyses largely corroborated the primary findings, aligning with the observed trends and associations. Our findings suggest that, except for brodalumab, all biologics have relative safety concerning pregnancy-related adverse outcomes in pregnant women and newborns compared to TNF-α inhibitors and certolizumab pegol. Additionally, the safety of using each TNF-α inhibitor during pregnancy is comparable to that of certolizumab pegol, consistent with previous findings.

TNF-α inhibitors have been extensively studied in systematic reviews and meta-analyses, resulting in conflicting findings regarding their use during pregnancy [12, 20‒25]. Even though several studies reported an association between TNF-α inhibitors and adverse pregnancy outcomes [20, 23], Meyer and colleagues reported no such associations [25]. Moreover, Ghalandari and colleagues found no difference in the prevalence of maternal, fetal, and neonatal adverse outcomes in pregnant individuals using TNF-α inhibitors compared to the general population [21]. It is believed that studies reporting adverse pregnancy outcomes related to TNF-α inhibitors did not adjust for underlying conditions in psoriasis patients, which are significant confounders [8, 22, 23]. A recent updated study by McMullan and colleagues reported that infliximab and adalimumab are tolerable during pregnancy, with no detected association between etanercept and adverse pregnancy outcomes in longitudinal studies. Furthermore, certolizumab pegol is considered safest to use during pregnancy, without the risk of fetal immunosuppression [12]. Currently, TNF-α inhibitors, except for golimumab, are generally considered safe for treating psoriasis during pregnancy [24]. Therefore, we analyzed the safety of other systemic treatments for psoriasis compared to TNF-α inhibitors [26]. Specifically, we compared the relative safety of each treatment to certolizumab pegol in our sub-analysis as it is known to be the safest among TNF-α inhibitors [27, 28].

The disproportionality analysis between specific biologics and TNF-α inhibitors revealed that these biologics exhibit a lower frequency of pregnancy-related adverse outcomes compared to TNF-α inhibitors. The IL-12/IL-23 inhibitor ustekinumab exhibited an overall low frequency of pregnancy-related adverse outcomes. Although there has been a case report of ustekinumab-induced spontaneous abortion [29], a recent prospective study found no association between ustekinumab and adverse pregnancy outcomes [30, 31], consistent with our findings.

The IL-17 inhibitors secukinumab and ixekizumab also showed an overall lower frequency of pregnancy-related adverse outcomes, consistent with the previous studies [32, 33]. Brodalumab had a comparable frequency of pregnancy-related adverse outcomes, whereas bimekizumab exhibited a lower frequency. However, due to the relatively recent approval of brodalumab and bimekizumab, the number of reports is insufficient to draw strong conclusions.

Among IL-23 inhibitors, guselkumab and tildrakizumab exhibited lower frequencies of pregnancy-related adverse outcomes, consistent with the previous studies [34, 35]. Risankizumab also showed a lower overall frequency but had a disproportionately higher reporting rate of abortion and stillbirth compared to TNF-α inhibitors. Due to the limited number of previous reports on risankizumab, a comparison with previous studies cannot be made. However, the number of reports of abortion and stillbirth associated with risankizumab is concerning, indicating a need for future studies focused specifically on these outcomes. These findings were consistent across comparisons with certolizumab pegol (online suppl. Table S7), multivariable analyses (online suppl. Tables S8, 9), and sensitivity analyses (online suppl. Tables S10–15), enhancing the validity of our data.

A notable observation was made when comparing specific TNF-α inhibitors to certolizumab pegol in terms of pregnancy-related adverse outcomes. In the crude disproportionality analysis, infliximab showed a lower frequency, etanercept showed a comparable frequency, and adalimumab showed a higher frequency of pregnancy-related adverse outcomes. However, in the multivariable and sensitivity analyses, etanercept, infliximab, and adalimumab all showed lower or comparable frequencies of pregnancy-related adverse outcomes, indicating their safety during pregnancy is comparable to that of certolizumab pegol, consistent with the previous studies [8, 21, 25, 36].

This study has several limitations. Despite the extensive dataset on biologic exposure during pregnancy, the indications for use were not restricted to psoriasis; many involved rheumatological and gastroenterological conditions, where dosages and administration differ [10]. Additionally, the specific trimester of biologics exposure is unknown. Most biologics, except for certolizumab pegol, are known to cross the placenta starting in the second trimester [37]. Detailed evaluations of the specific trimester during which exposure occurred would have significantly enhanced the analysis of biologics and their associated adverse pregnancy outcomes. The study by Sánchez-García and colleagues was also mostly limited to first-trimester exposures [8]. Recommendations to stop adalimumab at 20 weeks and etanercept at 30–32 weeks might have resulted in a lack of data on second- and third-trimester exposures [38], even in our database. Furthermore, comorbidities in psoriasis patients, disease activity, and maternal demographics were not fully assessed, preventing clear conclusions about adverse pregnancy outcomes due to biologic exposure. Moreover, small sample sizes for certain drugs, such as brodalumab, bimekizumab, and tildrakizumab, are not sufficient enough to draw strong conclusion. Moreover, with more than 40% of reports from non-healthcare professionals, the quality and reliability of the data may be affected. Lastly, this study compared the safety of biologics relative to TNF-α inhibitors, particularly certolizumab pegol, and found the safety regarding adverse pregnancy outcomes to be only described as frequency, necessitating a further study to elucidate the mechanisms and safety of each biologic.

Our study has several strengths. While previous research on the safety of biologics did not include newer anti-psoriatic biologics such as brodalumab, bimekizumab, and risankizumab [8], we included all biologics currently used in systemic therapy for psoriasis. Additionally, prior studies have primarily consisted of case reports and small samples focusing on specific biologics and their adverse outcomes. In contrast, our study utilized an extensive dataset and provided a comprehensive analysis of adverse pregnancy outcomes, categorized into maternal, fetal, and neonatal outcomes. Moreover, we have conducted additional multivariable and sensitivity analysis to validate our primary findings. This suggests that all of the current biologics for psoriasis during pregnancy are comparable to TNF-α inhibitors, particularly certolizumab pegol, regarding safety concerning pregnancy-related adverse outcomes.

This comprehensive, long-term global study found that biologics used for psoriasis, including IL-12/IL-23 inhibitors, IL-17 inhibitors, and IL-23 inhibitors (except for brodalumab), reported a lower overall frequency of pregnancy-related adverse outcomes in pregnant women and newborns compared to TNF-α inhibitors and certolizumab pegol, suggesting their relative safety during pregnancy. However, risankizumab was associated with a higher frequency of abortion and stillbirth, a profile not previously reported, necessitating further investigation. Additionally, infliximab, etanercept, and adalimumab exhibited comparable frequencies of pregnancy-related adverse outcomes in pregnant women and newborns to certolizumab pegol, aligning with previous studies on the safety of TNF-α inhibitors during pregnancy.

The authors express their gratitude to the Uppsala Monitoring Centre for providing access to the data used in this study. The findings and conclusions presented here are those of the authors and do not necessarily represent the views of the Uppsala Monitoring Centre or the World Health Organization.

Approval for the use of confidential and electronically processed patient data was granted by the Institutional Review Board of Kyung Hee University (KHUH 2022-06-042) and the Uppsala Monitoring Centre (WHO Collaborating Center). Due to the retrospective nature of the study, the Institutional Review Board of Kyung Hee University waived the need of obtaining informed consent.

All the authors declared no competing interests.

This research was supported by the National Research Foundation of Korea grant funded by the Korean Government (MSIT; RS-2023-00248157) and the MSIT (Ministry of Science and ICT), South Korea, under the ITRC (Information Technology Research Center) support program (IITP-2024-RS-2024-00438239) supervised by the IITP (Institute for Information & Communications Technology Planning & Evaluation). The funders played no role in the study design, data collection, data analysis, data interpretation, or manuscript writing.

Study concept and design, acquisition, analysis, or interpretation of data, drafting of the manuscript, and statistical analysis: Yi Deun Jeong, Hyesu Jo, Hanseul Cho, and Dong Keon Yon; critical revision of the manuscript for important intellectual content: Wonwoo Jang, Jaeyu Parks, Sooji Lee, Hayeon Lee, Kyeongmin Lee, Jiyeon Oh, and Xuerong Wen; and study supervision: Dong Keon Yon. Dr. Dong Keon Yon served as the guarantor, is the senior author, and had full access to all data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Yi Deun Jeong, Hyesu Jo, and Hanseul Cho contributed equally to this study as first authors. Dong Keon Yon and Lee Smith contributed equally as corresponding authors. The corresponding author attests that all listed authors meet the authorship criteria and others not meeting the criteria have been omitted. All authors have approved the final version of the manuscript before submission.

Yi Deun Jeong, Hyesu Jo, and Hanseul Cho contributed equally to this work.Edited by: H.-U. Simon, Bern.

The data are available upon request. Study protocol and statistical code: available from DKY ([email protected]). Dataset: their containing information that could compromise the privacy of research participants but are available from the Uppsala Monitoring Centre (WHO Collaborating Center) or WHO through a data use agreement upon reasonable request. Further inquiries can be directed to the corresponding author.