By Ahizechukwu C. Eke, MD, MPH

Assistant Professor in Maternal Fetal Medicine, Division of Maternal Fetal Medicine, Department of Gynecology & Obstetrics, Johns Hopkins University School of Medicine, Baltimore

Dr. Eke reports no financial relationships relevant to this study.

SYNOPSIS: In this population-based cohort study of 1,027 infants born to women treated with anti-TNF-α biologic therapy, there was an increased prevalence of preterm birth (adjusted odds ratio [aOR], 1.61; 95% confidence interval [CI], 1.29-2.02), cesarean delivery (aOR, 1.57; 95% CI, 1.35-1.82), and small for gestational age neonates (aOR, 1.36; 95% CI, 0.96-1.92) when treatment with anti-TNF was compared to non-biologic systemic treatment. Since disease processes varied greatly in these pregnant women, it was difficult to rule out confounding by disease severity (confounding by indication).

SOURCE: Bröms G, Kieler H, Ekbom A, et al. Anti-TNF treatment during pregnancy and birth outcomes: A population-based study from Denmark, Finland, and Sweden. Pharmacoepidemiol Drug Saf 2020;29:316-327.

Anti-tumor necrosis factor-alpha (anti-TNF-α), biologic therapy (adalimumab, certolizumab-pegol, golimumab, etanercept, and infliximab) is increasingly being used in managing inflammatory arthritides (ankylosing spondylitis, psoriatic arthritis, and rheumatoid arthritis) and Crohn’s disease during pregnancy.^1,2 These inflammatory arthritides are known to be associated with adverse pregnancy outcomes.³ In this population-based study in Denmark, Finland, and Sweden, Bröms and colleagues described their findings in pregnant women treated with anti-TNF-α biologic therapy compared to those treated with non-biologic systemic therapy (anti-malarials, azathioprine, corticosteroids, methotrexate, and sulfasalazine).¹ Women were eligible if they gave birth to a singleton infant between January 2006 and December 2013 in Denmark; January 2006 and December 2012 in Finland; and July 2006 and December 2013 in Sweden; and had their medical records completed in the national medical birth registers of these three countries.¹ Outcomes assessed included preterm birth at less than 37 weeks of gestation, small for gestational age (moderate and severe), and cesarean delivery (planned and emergent). For statistical analyses, descriptive statistics was performed to describe the study sample. Univariate logistic regression analyses then were done to estimate the probability of preterm birth, cesarean delivery, and small for gestational age for participants on an anti-TNF-α biologic therapy vs. a non-biologic systemic regimen. To control for confounding, the authors used multivariable logistic regression to define the models that best predicted the probability of outcomes (preterm birth, cesarean delivery, and small for gestational age) as a function of covariates (country, maternal age, parity, smoking, body mass index, previous surgeries, and maternal disease). Sensitivity analyses were done to identify pertinent associations between inflammatory arthritides/Crohn’s disease observations, anti-TNF-α biologic therapy, model inputs, and predictions, leading to the development of better models.

Among 1,633,909 live births included in this population-based cohort study, 1,027 pregnant women were managed with anti-TNF biologic therapy, while 9,393 pregnant women were treated with non-biologic systemic treatment during the course of pregnancy. Rheumatoid arthritis was the most common indication for anti-TNF-α biologic therapy (52.3%), followed by Crohn’s disease (23.1%). The most common anti-TNF-α biologic therapies used were etanercept (49.7%), adalimumab (25.0%), and infliximab (20.3%).

Anti-TNF-α biologic therapy increased the prevalence of preterm birth (adjusted odds ratio [aOR], 1.61; 95% confidence interval [CI], 1.29-2.02), cesarean delivery (aOR, 1.57; 95% CI, 1.35-1.82), and small for gestational age neonates (aOR, 1.36; 95% CI, 0.96-1.92) when treatment with anti-TNF-α biologic therapy was compared to non-biologic systemic treatment. The risk of preterm birth was somewhat reduced in Danish and Swedish women after controlling for previous surgery and hospitalization in multivariable analysis (aOR, 1.41; 95% CI, 1.10-1.80). However, sensitivity analyses had minimal effect on the results. Since disease processes varied greatly in these pregnant women, it was difficult to rule out confounding by disease severity (confounding by indication).

COMMENTARY

Anti-TNF-α biologic therapy works by blocking the actions of TNF-α and neutralizing its biologic effects.⁴ TNF-α is an important pro-inflammatory and immunomodulatory cytokine, and it is critical for differentiation of dentritic cells, B-cells, and T-cells during pregnancy.⁵ The levels of TNF-α increase throughout gestation, especially in women with Crohn’s disease and inflammatory arthritides. Because of the unique structure of TNF-α inhibitors (monoclonal antibodies composed of human immunoglobulin G), trans-placental passage of these immunoglobulins increases as pregnancy progresses, primarily during the second half of pregnancy.⁶ Since these drugs cross the placenta, they are assumed, theoretically, to affect the mother and the developing fetus. However, with limited data on the safety of anti-TNF-α therapy use during pregnancy, no harmful embryotoxic effects have been reported with use of anti-TNF-α therapy.⁶

Trans-placental passage of anti-TNF-α therapy occurs by passive diffusion via binding to Fc receptors in placental syncytiotrophoblasts, and the rate of placental transfer can be influenced by molecular weight, lipid solubility, degree of ionization, and half-life.⁶ The half-lives of adalimumab, certolizumab-pegol, and etanercept are 10 to 20 days, 11 to 14 days, and four to six days, respectively, while those for golimumab and infliximab are 10 to 18 days, and 7.7 to 9.5 days, respectively. This means it would take approximately four to six weeks for these drugs to be eliminated completely from the body after the last dose.

Since these medications have long half-lives, the major issue with their use in pregnant women and neonates is preterm birth and immunosuppression (and potentially increased risk for infection), respectively, although these have not been confirmed in prospective cohort studies. Anti-TNF-α therapy also is known to be secreted in breast milk.

As described by Bröms and colleagues, anti-TNF-α biologic therapy, when used in women with inflammatory arthritides and Crohn’s disease, was associated with adverse pregnancy outcomes (preterm birth, small for gestational age, and cesarean deliveries). However, it is difficult to tease out if these adverse outcomes were the result of disease severity or the anti-TNF-α biologic therapy, or both. This special kind of confounding is known as confounding by indication. Confounding by indication is said to have occurred when a disease — the indication for drug use (or exposure) — independently affects the outcome.⁷ Confounding by indication constitutes a major challenge in observational studies, and is one of the most difficult types of confounding to control. Although there are several studies that have used multivariable regression models to control for confounding by indication, propensity score-based analysis remains one of the best techniques to control for confounding by indication, since it provides a more precise estimate of treatment effect where confounding by indication is presumed to be present.

In conclusion, it is unclear how the findings of this study by Bröms and colleagues would translate to clinical practice, since confounding by indication was not adequately controlled for using appropriate statistical methods. When used during pregnancy, it is advised to discontinue anti-TNF-α therapies approximately four to six weeks prior to delivery to allow time for fetal recovery from any immunosuppression.

Anti-TNF-α biologic therapy can theoretically interfere with an infant’s response to live vaccines, and, therefore, such vaccines should be avoided in these circumstances. In addition, anti-TNF-α therapies may increase the risk of systemic maternal and neonatal infections, reactivation of chronic diseases (such as tuberculosis), and development of lymphomas when used in pregnant women.

REFERENCES

1. Bröms G, Kieler H, Ekbom A, et al. Anti-TNF treatment during pregnancy and birth outcomes: A population-based study from Denmark, Finland, and Sweden. Pharmacoepidemiol Drug Saf 2020;29:316-327.
2. Genest G, Spitzer KA, Laskin CA. Maternal and fetal outcomes in a cohort of patients exposed to tumor necrosis factor inhibitors throughout pregnancy. J Rheumatol 2018;45:1109-1115.
3. Aljary H, Czuzoj-Shulman N, Spence AR, Abenhaim HA. Pregnancy outcomes in women with rheumatoid arthritis: A retrospective population-based cohort study. J Matern Fetal Neonatal Med 2020;33:618-624.
4. Lis K, Kuzawinska O, Balkowiec-Iskra E. Tumor necrosis factor inhibitors — state of knowledge. Arch Med Sci 2014;10:1175-1185.
5. Bränn E, Edvinsson Å, Rostedt Punga A, et al. Inflammatory and anti-inflammatory markers in plasma: From late pregnancy to early postpartum. Sci Rep 2019;9:1863.
6. Yockey LJ, Iwasaki A. Interferons and proinflammatory cytokines in pregnancy and fetal development. Immunity 2018;49:397-412.
7. Brookhart MA, Stürmer T, Glynn RJ, et al. Confounding control in healthcare database research: Challenges and potential approaches. Med Care 2010;48:S114-120.