The use of biologics for autoimmune rheumatic diseases in fertility and pregnancy

Short abstract

In an age where autoimmune rheumatic diseases are successfully managed with biologics, their discontinuation in pregnancy is inadvisable without careful forethought; maternal disease activity is associated with adverse pregnancy outcomes, which has long-term implications for both mother and offspring. We aim to provide clinicians with the necessary tools to facilitate decision-making – when a biologic should be used, when it can be discontinued in pregnancy if appropriate. The pathophysiology of these biologic molecules and their effect on fertility, pregnancy and parturition are discussed. A summary of the 2016 international guidelines (European League Against Rheumatism and British Society in Rheumatology) on biologics in pregnancy has been tabulated; more recent publications are discussed in depth. Data on transplacental-transfer ratios and breastmilk excretion rates are also included. Biologic effects on organogenesis, their implications for the exposed infant in terms of infection risks and vaccination requirements are included, and future directions for research proposed.

What are ‘biologics’?

Unlike chemically synthesised drugs with a known structure, biologics are complex molecules largely composed of immunoglobulin (Ig) G with modified Fc receptors or Fab fragments that bind and neutralise their target molecule. Suffixes to names of biologics convey specific information concerning their origin and structure, e.g. ‘-mab’ indicates monoclonal antibody (mAb), while ‘-cept’ denotes fusion of a receptor to the Fc portion of human IgG1, ‘-ximab’ chimeric mAb and ‘-zumab’ humanised mAb¹ (Table 1).

Table 1.

Biologic agents classified according to mode of action, used in the treatment of rheumatic diseases.

Mechanism of action	Drug	Structure
TNF inhibition	Infliximab	Chimeric mAb to TNFα
	Etanercept	Fusion protein with a modified Fc portion that binds to TNFα
	Adalimumab	Fully human monoclonal antibody to TNFα
	Certolizumab	mAb to TNFα antibody with a pegylated Fab portion without an Fc portion
	Golimumab	Fully human mAb to TNFα
IL-1 inhibition	Anakinra	Recombinant and modified human IL-1 receptor antagonist
	Canakinumab	Human mAb that binds specifically to IL-1ß
	Rilonacept	Fusion Fc receptor that binds IL-1ß
IL-6 inhibition	Tocilizumab	Humanised mAb against IL-6 receptors
	Sarilumab	Recombinant mAb to IL-6 receptors
IL-17 inhibition	Secukinumab	Human mAb neutralising the effects of IL-17A
	Ixekizumab	Humanised mAb against IL-17A
IL-12/23 inhibition	Ustekinumab	Fully human mAb binding specifically to IL-12/IL23 inhibiting its activity
Co-stimulatory blockade	Abatacept	Fusion Fc region to extra-cellular domain of CTLA-4 modifying T-cell response
B-cell depletion/inhibition	Rituximab	Chimeric mAb with murine anti-CD variable-sequence regions
	Belimumab	Fully humanised mAb directed against soluble B-lymphocyte stimulator

CTLA: cytotoxic T-lymphocyte-associated protein; CD: complementarity-determining; IL: interleukin; mAb: monoclonal antibody; TNF: tumour necrosis factor.

A recent survey in Europe and the USA revealed that clinicians concerned that anti-tumour necrosis factor (TNF) causes poor pregnancy outcomes have discouraged women from breastfeeding while on an anti-TNF.² With improved disease and pregnancy outcomes associated with ongoing biologic (especially anti-TNF) use during pregnancy, clinicians will therefore need to address these issues with women of childbearing potential; discontinuation of biologics during pregnancy is not advised without due consideration of the potential risks and benefits.^3,4

Pathophysiology of cytokines in fertility, pregnancy and delivery

In brief, pregnancy is an immune-tolerant state of the feto-maternal relationship in which T-helper (Th) cytokines play an important role. Normally, Th2 (humoral immunity) predominates over Th1 (cell-mediated immunity) in early gestation. This balance will shift to a Th1-predominant state later in pregnancy.^5,6

Inflammation can shift this balance towards a Th1 dominance, resulting in a stronger cell-mediated response and cytokine production. The Th1-predominant environment may lead to implantation failure during in vitro fertilisation (IVF), recurrent fetal loss and other features of placental insufficiency such as pre-eclampsia and/or fetal growth restriction, often cumulating in preterm deliveries.^7,8

TNF is associated with ovarian follicular and embryonic development, activation of apoptotic pathways in fetal and placental cells and is likely to play a role in stimulation of protective mechanisms to repair damages when structural anomalies are present or to trigger apoptotic processes to prevent the ongoing development of embryos with structural anomalies.^9,10 A marked increase of inflammatory cytokines occurs during infection or inflammation and together with other chemokines may play a role in preterm delivery.^11,12

Women with recurrent pregnancy loss had significantly higher Th1/Th2 ratios of interferon-γ/interleukin (IL)-4, TNF-α/IL-4 and TNF-α/IL-10 in CD3+/CD8− T helper cells than controls. Moreover, those who experienced recurrent implantation failure after IVF had a higher proportion of TNF-α producing CD3+/CD8− cells and higher Th1/Th2 ratios of TNF-α/IL-4 and TNF-α/IL-10 in CD3+/CD8− cells when compared with controls.¹³ In sub-fertile women, the preconception ratio of TNF-α: IL-10 elevation has been associated with high oocyte die-off ratio leading to lower success rates during IVF.^14,15 Adalimumab and immunoglobulins administration have been associated with a reduction in the oocyte die-off ratio and a greater chance of successful IVF.^14–16

Overall, these observations suggest that targeting TNF-α may be useful in treating women with recurrent pregnancy loss, recurrent implantation failure and early, severe pre-eclampsia due to excessive TNF-α production.¹⁷

During delivery, IL-1, IL-6, IL-8 and TNF levels peak. Hypothetically, preterm labour is driven by overwhelming inflammation.¹² Using an experimental lipopolysaccharide-induced preterm model, IL-6 inhibitors induced prolongation of the gestational period in mice (without adverse effects on pups) by significantly inhibiting IL-6-induced prostaglandin E₂ production.¹⁸ Therefore, it is possible that in the future preterm, labour might be manageable by a combination of biologic agents that specifically targeted these pathways of Th1 activation.

Clinical studies in men have reassuringly demonstrated that anti-TNF agents do not impair sperm quality, and their use has not been associated with teratogenicity in offspring.^19,20

Biologics in early pregnancy

Biologics are large molecular structures not easily transported across the placenta, particularly during the first trimester when fetal trophoblasts have yet to develop Fc receptors. Prior to 14 weeks’ gestation, these large molecules reach the fetus by passive diffusion; hence, only limited amounts reach the fetus during the period of embryogenesis (up to 12 weeks).²¹

An early and controversial study mining of the Food and Drug Administration database attempted to link anti-TNF use with VACTERL (vertebral defects, anal atresia, cardiac defects, tracheo-esophageal fistula, renal anomalies, and limb abnormalities)-association congenital abnormalities.²² Individual anomalies that form the VACTERL are not uncommon – hence, the need for three or more features is needed for the diagnosis. The cases reported in the study did not necessarily adhere to these criteria for the diagnosis of VACTERL-association.²³ Multiple studies subsequently failed to find any association. Therefore, many national and international guidelines support the use of anti-TNF agents in the first trimester.^24,25

Biologics in the second and third trimesters

From the second trimester onwards, transplacental passage of biologics is facilitated by Fc receptors on the fetal trophoblasts in tandem with increasing transplacental transfer of maternal IgG as pregnancy advances.^26,27 Fetal IgG1 levels reach 50% of maternal levels at 28–32 weeks’ gestation and increase exponentially afterwards with maximal transfer in the final 4 weeks prior to delivery.²⁸ Hence, at delivery, cord blood levels of neonates exposed to some of the biologics in late gestation are often much higher than maternal serum drug levels (Table 2). Cord blood levels probably depend on several factors: (i) The gestation in which the drug was used – the later in pregnancy, the higher the cord drug levels; (ii) structure of the Fc receptors – biologics with modified Fc receptors (i.e. the ‘-cepts’) or those lacking an Fc portion altogether (e.g. certolizumab) are less efficiently transported across the placenta and can often be used throughout pregnancy or at least until 34 weeks’ gestation (e.g. etanercept); (iii) half-life of the drug – the longer its half-life, the more likely its passage across the placenta. There are limited published studies on maternal to neonatal cord blood drug levels, with wide variations in the very small cohorts sampled (Supplementary Table 1). In addition to the factors just listed, drug levels are also likely influenced by the dosing interval (i.e. proximity of cord blood collection to administration of the last dose), dosages used (e.g. infliximab can be dosed at 3–10 mg/kg at six to eight weekly intervals) and quite possibly a multitude of other factors which remain poorly understood (e.g. drug antibodies). Apart from certolizumab – virtually none of which crosses the placenta – it remains unlikely that sufficient data will ever be available concerning the exact amount of drug transferred to the neonate in utero. Available published data on cord blood ratios have been summarised in Table 2. Furthermore, results from animal studies should not be extrapolated since Fc receptors and binding to neonatal trophoblasts may differ in animals. Additionally, doses tested in laboratory animals often exceed the recommended therapeutic doses used in human subjects^26,29 (full data in Supplementary Table 1).

Table 2.

The half-life of biologics and transplacental transfer ratio.

Drug	t½ (days)	References	Transplacental transfer ratio (cord blood: maternal drug level)a
Infliximab	8–10	30,31,32,33,34	2.23
Etanercept	4–5	35,36	0.06
Adalimumab	10–20	37,38,33,39,34	1.11
Golimumab	12–15	NA	–
Certolizumab	14	40,33,41	0.04
Canakinumab	22	42	2.11
Tocilizumab	8–14	NA	–
Ustekinumab	15–32	43	1.86
Abatacept	8–25	NA	–
Rituximab	18–22	44,45,46	1.69
Belimumab	19–20	NA

Note: NA: not available in human subjects.

^aMean value from cumulative studies.

Hence, most studies on biologics use in pregnancy have focused on clinical outcomes in exposed offspring.

Latest updates on biologic agents

While data remain limited, there are no signals of increased rates of malformations or adverse pregnancy outcomes from the various registries set up specifically to look at the use of biologics in pregnancy.

This review paper will endeavour to summarise the European League Against Rheumatism (EULAR) and British Society in Rheumatology (BSR) guidelines 2016 shown in Table 3^24,25,47 and focus on developments since publication of the guidelines. Additional information on breastmilk excretion is available in Supplementary Table 1.

Table 3.

Summary from EULAR points to consider and BSR guidelines for use during pregnancy.

Agent	Compatible for use in early pregnancy		What gestation to discontinue		Compatible with breastfeeding
	EULAR	BSR	EULAR	BSR	EULAR	BSR
Infliximab	✓	✓	Up to 20 weeksa	Stop at 16 weeks	✓	✓
Etanercept	✓	✓	Up to 32 weeksa	Not compatible with third trimester	✓	✓
Adalimumab	✓	✓	Up to 20 weeksa	Not compatible with third trimester	✓	✓
Certolizumab	✓	✓	Use throughout pregnancy	Compatible with third trimester	✓	No data
Golimumab	No increased malformations §	✗	Use another agent §	No data	✓	No data
Anakinra	✓	✗ ‡	Only if no other options	✗	✗ No data	No data
Tocilizumab	§	Stop three months prior ‡	Avoid §	✗	✗ No data	No data
Ustekinumab	No increased malformations §	–	Use another agent §	–	✗ No data	–
Abatacept	§	✗	✗ §	✗ No data	✗ No data	No data
Rituximab	✓ Can be used in exceptional circumstances.	Stop six months prior ‡	Not specified. †	✗	✗ No data	No data
Belimumab	No increased malformations §	✗ ‡	✗ §	✗	✗ No data	No data

Note: ✓: yes; ✗: avoid/not recommended; No data: no published data available; §: insufficient evidence; †: neonatal B cell depletion if used in late pregnancy; ‡: insufficient data to recommend drug, but unintentional exposure in the first trimester is unlikely to be harmful; NA: not applicable; – not covered by publication. EULAR: European League Against Rheumatism; BSR: British Society in Rheumatology.

^aCan be used throughout pregnancy if indicated.

IL-1 inhibitors: Canakinumab and anakinra

An international retrospective study of IL-1 inhibitors anakinra and canakinumab was collated by members of the International Society for Systemic Autoinflammatory Diseases.⁴⁸ Among the eight women exposed to canakinumab during pregnancy (plus five paternal exposures in the preconception period), no congenital abnormalities were observed. A single early miscarriage occurred at six weeks’ gestation in a woman with refractory Cogan’s syndrome. The seven infants exposed to canakinumab were delivered at term with normal birth weights. Maternal disease remained in remission during pregnancy, despite only four women continuing on canakinumab beyond the period of organogenesis (≥12 weeks’ gestation); others either stopped treatment or switched to anakinra (n = 2). Four infants were breastfed with no adverse effects. Follow-up time of infants varied from seven days to four years.

In the same study, 23 women (and 6 men) were exposed to anakinra. A single case of renal agenesis and ectopic neurohypophysis with growth hormone deficiency was noted in an infant whose mother was treated from 9 weeks’ gestation till delivery; miscarriage at 12 weeks was recorded in a woman with active Cogan’s syndrome.⁴⁸ An abstract summarising the experience of five women who received anakinra 100 mg into the third trimester found no malformations or neonatal problems – including the three infants breastfed while their mothers continued treatment with anakinra.⁴⁹

There were no reports of serious infections or developmental delay in a median follow-up of 18 months in neonates exposed to either canakinumab or anakinra.⁴⁸ However, data on vaccinations were lacking, especially for neonates exposed in the third trimester.⁴⁹

Canakinumab has a transplacental transfer ratio of 2.11 (Table 2) based on a single case where the cord canakinumab level was quantified in an infant exposed until 34 weeks’ gestation and delivered at 39 weeks’ gestation.⁴² Hitherto, despite multiple clinical cases published in the literature, anakinra has never been quantified in cord blood, and its transplacental transfer ratio is unknown.^49–52 Since anakinra has a half-life of 4–6 h compared to 22 days in canakinumab, it would seem reasonable to switch to anakinra in pregnancy if clinically indicated. Nevertheless, this short half-life entails daily administration – rather than the four- to eight-weekly dosing interval of canakinumab. Hence, quantification of anakinra cord, maternal and infant serum drug levels would be very helpful in informing clinicians and prospective mothers on the preferred IL-1 inhibitor to use in pregnancy. Although, cord serum quantification of anakinra (as a slight modification of the natural IL-1 receptor antagonist) may require a special test that prevents measuring both the native-occurring IL-1 and anakinra together.

Cases of rilonacept use during pregnancy have yet to be published.

IL-6 inhibitors: Tocilizumab and sarilumab

Since EULAR and BSR guidelines, there have been four additional publications including registry data from the Tocilizumab manufacturer's global safety database.^53,54,55,56 There have been 358 observed pregnancies; the miscarriage rate was 22.9%, though concurrent methotrexate (and occasionally leflunomide) use was prevalent in these cohorts.^55,56 The livebirth rate was only 60%, due to a high rate of miscarriage and elective termination of pregnancy (18.4%) in the cohorts studied. Six malformations were noted in the literature, but without any distinguishing pattern of malformation to suggest teratogenicity.^55,56 Tocilizumab use (>90% in four-weekly 8mg/kg infusions) occurred predominantly during pre-conception and the first trimester, the drug being discontinued upon confirmation of pregnancy for the majority of women; tocilizumab was used beyond the first trimester in only 15 pregnancies.^55,56 Of the 15 neonates exposed to tocilizumab later in pregnancy, none had their cord tocilizumab level quantified.

A case series of two women showed that breastmilk concentrations of tocilizumab peaked at day three post-administration but remained 1/500 to 1/1000 compared to maternal serum concentrations. Both breastfed infants remained well; they received all routine vaccinations, including Bacillus Calmette–Guérin (BCG) and rotavirus without problems.⁵⁷

Studies on sarilumab use in pregnancy have yet to be published. The manufacturer has set up an observational study (commencing 1 February 2018) to recruit at least 300 women (and infants) for the duration of pregnancy and one year follow-up post-partum (see: https://clinicaltrials.gov/ct2/show/NCT03378219).

IL-17A inhibitors: Secukinumab and ixekizumab

An abstract on secukinumab in 66 pregnancies from the manufacturer’s global safety database was recently presented.⁵⁸ There were 8 (12.1%) miscarriages, only 15 reported livebirths (without any congenital malformation) and 11 ongoing pregnancies. Close to half (48%) had either an elective termination of pregnancy (without any listed cause) or information on the pregnancy was missing. No information was provided on 18 pregnancies with paternal exposure. More recently, the experience of six women with psoriatic arthritis on secukinumab during pregnancy was published as an abstract.⁵⁹ Secukinumab was discontinued upon confirmation of pregnancy and no malformation, adverse maternal or neonatal outcomes occurred. Maternal psoriatic arthritis remained in remission throughout pregnancy even off secukinumab.⁵⁹

Similarly, data on ixekizumab from the manufacturer’s Safety Systems (from seven clinical trials) were limited.⁶⁰ Of 18 pregnancies, 8 (44.4%) resulted in livebirths and 5 (27.8%) in miscarriages or elective terminations. One third of cases (n = 6) had no recorded outcome; it was unclear if this represented missing information or any other adverse pregnancy outcome (e.g. stillbirth, late pregnancy loss, etc.). No congenital malformations were reported. Of the 34 cases with paternal exposure to ixekizumab, the vast majority (82.4%) resulted in livebirths, without any cases of congenital malformation.

For all studies, information on concurrent disease-modifying anti-rheumatic drugs (DMARD) use, duration of exposure to the IL-17A inhibitor, breastfeeding or reasons for termination of pregnancy were not available from the abstracts.

IL-12/23 inhibitor: Ustekinumab

Only 10 published papers and abstracts exist on the use of ustekinumab during pregnancy. Two of the larger studies were derived from Phase II studies on Crohn’s and psoriasis where ustekinumab was discontinued upon confirmation of pregnancy.^61,62 To date, 63 pregnancies with exposure to ustekinumab and known outcomes have been described.^61--70 Ustekinumab exposure occurred within three months pre-conceptually or in the first trimester in the majority of cases (59/63). Three women who were prescribed ustekinumab in the third trimester also had good neonatal outcomes.^63,68,69 While the dosage and dosing intervals for ustekinumab were different in the two of the women in the third trimester, their exposed offspring remained well, and one even had all his ‘routine’ vaccinations without problems.⁶⁹ Overall, the livebirth rate across all published cases was 57.1%, and miscarriage was low at 15.9%. Miscarriage occurred at 12 weeks in one mother who was a heavy smoker.⁶⁴ Only one case had concurrent DMARD documented (i.e. azathioprine 100mg daily). Since these were largely studies for Crohn’s and psoriasis, some women may have been co-prescribed methotrexate. No reasons were given for elective terminations, and no congenital malformations were reported. In the three cases involving third trimester exposure, neither cord drug levels nor drug excretion into breastmilk were quantified. In one case, despite ustekinumab use until 33 weeks’ gestation, the mother was advised not to breastfeed the infant once she resumed ustekinumab post-partum.⁶⁸ None of the published cases included paternal exposure to ustekinumab.

Human data from published studies do not appear to show an increased risk of congenital malformations or miscarriages in the exposed group. However, ustekinumab use in later pregnancy, transplacental transfer ratios, neonatal infections and responses to vaccinations were not recorded in the reports.

B-cell inhibitors: Belimumab

Most of the data on belimumab in pregnancy come from the pharmaceutical company’s registry (see: http://www.bprgsk.com, currently collecting information till 2021).⁷¹ Data on 66 pregnancies were presented as an abstract in EULAR in 2013 related to an open label Phase IV (long-term extension) clinical trial and showed a lower rate of fetal loss in women treated with belimumab compared to placebo and a livebirth rate of only 50%. Exact information on DMARDs was not available; there were two cases of malformations – one Dandy–Walker syndrome, and one bilaterally enlarged kidneys.⁷² Presumably belimumab was discontinued upon confirmation of pregnancy; a large number of women also elected to have their pregnancy terminated.⁷²

Two additional published case reports that described the use of belimumab beyond the first trimester of pregnancy did not demonstrate any adverse outcomes in either mother or neonate.^73,71 There is a single published case report of maternal use of belimumab till 26 weeks’ gestation for severe systemic lupus erythematosus (SLE). At 40 weeks, she delivered a healthy male infant who subsequently received all routine immunisations, including the rotavirus live vaccine at 6 weeks, without problems. He was also noted to mount a good response to the pneumococcal vaccine at seven months of age.⁷⁴

Abatacept

To date, 162 pregnancies with abatacept exposure in early pregnancy (152 maternal and 10 paternal) have been described.^75,76 The manufacturer’s registry revealed 151 maternal cases resulting in 86 (59.5%) live births, 49 miscarriages (including one late loss at 21 weeks’ gestation) and 19 elective terminations.⁷⁶ Methotrexate was co-administered in 13.2%. Seven cases (8.1%) of congenital malformations were recorded, including a single case of Down’s syndrome (terminated at 17 weeks’ gestation due to spontaneous rupture of membranes). The cases of malformation were isolated and did not reflect a specific pattern of abnormalities.⁷⁶ In all published cases of maternal exposure, abatacept was discontinued upon confirmation of pregnancy in the first trimester.^75,76 Paternal exposure to abatacept was not associated with any congenital anomalies in 9/10 ongoing pregnancies.⁷⁶ Information concerning the abatacept transplacental transfer ratio, use in late pregnancy, childhood infections and vaccinations has yet to be published, though follow-up in 16 of the offspring exposed in early pregnancy has not demonstrated any immune deficiencies.⁷⁶

Childhood infections and vaccinations

If a biologic transfers efficiently across the placenta, and maternal use has continued beyond 28 weeks, the exposed neonate is assumed to have cord drug levels exceeding those of maternal serum drug levels. The drug’s effects on the neonate will resemble those in the mother, i.e. potent immunosuppression. Hence, similar precautions are required as those taken in the case of the mother, e.g. prompt assessment and treatment of infections, avoidance of live vaccines (at least for the first six to seven months of the neonate’s life). These early vaccines to be avoided include rotavirus and BCG in certain countries. All other non-live vaccines are safe to be administered.

There is emerging evidence that rotavirus vaccine is safe even in infants exposed to maternal anti-TNF use in the third trimester.⁷⁷ Nevertheless, many will recall the solitary published case of the demise of a 4.5 month infant from disseminated non-caseating granulomatous disease (likely mycobacteria, though Ziehl–Neelsen staining for acid-fast bacilli and tuberculosis polymerase chain reaction methods were deemed equivocal) following BCG vaccination at 3 months of age; the mother had received infliximab therapy throughout pregnancy for her underlying Crohn’s, with her last infusion just two weeks prior to delivery.⁷⁸ Since TNF is integral to formation of the granuloma, the question then arises as to whether BCG vaccination would be safe in neonates exposed in the third trimester to other biologics (non-TNF-inhibitors).^79,80 While there has been no significant increase in neonatal infectious complications, carers should be advised to remain vigilant for possible infections. With dampened cytokine release in the presence of biologics, infants may not necessarily exhibit the same signs of infection that clinicians are accustomed to seeing.

The accumulation of biologics in offspring exposed in utero can persist for up to a year. Much depends on the timing of drug administration, the degree of transplacental transfer and the drug’s half-life. For instance, infliximab transfers efficiently starting from the second trimester (transplacental transfer 2.23) and remains in the infant for a median period of 7.3 months (although persisting up until 12 months of age), while adalimumab has a transplacental transfer of 1.11 and remains in the infant for a median of 4 months (Table 3).³⁴ It is postulated that ineffective clearance by the neonate’s immature reticuloendothelial system could contribute to the persistence of detectable drug levels in the infant’s serum.³³ Recent research has demonstrated that TNF remains in complex with adalimumab even six months after discontinuation of the drug in adults.⁸¹ Whether persistence of the drug has any adverse effect on the neonate has not been adequately studied.

Primate studies have shown that TNFs (particularly superfamily members lymphotoxin (LT)α and LTβ) are dispensable for the development of the neonatal immune system.¹¹ Golimumab in primate studies has not demonstrated a reduction in lymphoid tissue in macaque offspring exposed in utero.¹¹ This laboratory observation seems to be borne out by clinical findings – infants exposed in utero to anti-TNF agents are not at greater risk of infections, particularly if anti-TNF agents are used as monotherapy.^34,82

Nevertheless, there have been rare reports of neonatal neutropenia attributed to anti-TNF exposure in utero (women were treated with infliximab throughout pregnancy for inflammatory bowel disease (IBD)) – a known adverse effect of the drug.^83,84 However, when combined with other immunomodulators, in particular thiopurines, one international cohort study involving 80 women with IBD found that infants exposed to both thiopurines and anti-TNF agents demonstrated a three-fold increased risk of childhood infections, compared to those whose mothers were on monotherapy with an anti-TNF agent (relative risk 2.7; 95% confidence interval 1.09–6.78; p = 0.02).³⁴ The Groupe d’Étude Thérapeutique des Affections du Tube Digestif cohort (n = 232) found that the risk of infection and pregnancy complications was similar in two groups of pregnant women with IBD which, coincidentally, had similar rates of thiopurine use, although one group was on anti-TNF therapy and the other served as non-TNF control (n = 99).⁸⁵

Rituximab causes B-cell depletion, has a high transplacental transfer rate of 1.7 (Table 3) and is known to cause B-cell depletion and transient cytopenias in exposed infants.⁴⁴ In the global safety database for rituximab, 21 women receiving rituximab (for a variety of causes, including chemotherapy) had infants (n = 11) with recorded haematologic abnormalities that included B-cell depletion, lymphopenia, neutropenia, anaemias and thrombocytopenias.⁸⁶ Most spontaneously recovered in weeks to months post-partum, apart from one fatal case of cerebral haemorrhage in an exposed infant born with neonatal thrombocytopenia at 39 weeks’ gestation following the administration of rituximab to her mother at her seventh month of gestation for idiopathic thrombocytopaenic purpura. Suppressed B-cell development and depletion have not given rise to an increase in reported infectious complications amongst the affected neonates, and most of those studied continue to exhibit a robust immune response to vaccinations given.^44–46,86

Women on biologics planning pregnancy should be made aware that up until the present there have been no long-term outcome studies of children following in utero exposure. A theoretical risk has always existed of exposed offspring developing autoimmune diseases (e.g. anti-TNF-induced lupus), or being prone to allergies, malignancies, etc. However, fear of the unknown in the long term needs to be carefully balanced against the shorter term gain of maintaining maternal health and well-being, thereby ensuring an optimal pregnancy outcome. Mothers with active inflammatory/autoimmune diseases are at greater risk of disease flares leading to placental insufficiency clinically presenting as pre-eclampsia and fetal growth restriction – factors that in turn lead to preterm deliveries (and all the associated morbidity accompanying a preterm growth-restricted infant).^87,88 Moreover, the mothers themselves are also at risk since long-term population-based studies have demonstrated that women with autoimmune diseases (in particular SLE) who have pregnancies complicated by placental insufficiency or maternal-placental syndrome and preterm delivery prior to 34 weeks’ gestation are at a two-fold increased risk of accelerated development of cardiovascular events and premature cardiovascular death.^89–91

Breastfeeding

It is unlikely that comprehensive studies will ever be carried out on all biologics and the peak time for breastmilk excretion from time of drug administration (Supplementary Table 1). However, basic physiology indicates that breastmilk is predominantly IgA (not IgG1); the large structure of most biologic molecules means that they are less likely to be excreted in breastmilk. Additionally, such large proteinaceous molecules are denatured and digested by passage through the infant’s gastrointestinal tract and therefore do not attain significant levels in breastfed infants,⁴⁷ and hence why parenteral administration is necessary to achieve a therapeutic response. All breastmilk studies completed to date have borne this out by consistently showing a steady decline in breastfed infant drug levels despite ongoing maternal dosing. These data seem fairly reassuring should a woman desire to breastfeed even while taking a biologic.

Future directions

To date, published data seem largely reassuring: rates of congenital anomalies in offspring exposed in utero are similar to those in the general population, and they have achieved normal milestones in the first few years of life without an increased prevalence of allergies or life-threatening infections.

Although data concerning many newer biologics remain limited, the knowledge base is expanding. Pharmaceutical companies are encouraged to include pregnant women in drug trials, and pregnancy registries have been established for various drugs.⁹²

Table 4.

Practical points for consideration when facing a woman on a new biologic contemplating pregnancy.

1. The structure of the biologic and the likelihood of its passage across the placenta

2. When (or if) the biologic should be discontinued during pregnancy

3. Clinically significant and longer term effects on the neonate exposed in utero, the risks being balanced against the immediate short-term risk of a maternal disease flare or active disease in pregnancy resulting in an adverse outcome, preterm delivery or longer term maternal morbidity

It is highly desirable that pharmaceutical companies will in future invest in developing of biologics suitable for use in pregnancy. Perhaps biologics could be created from IgG2 molecules instead – the latter would give rise to minimal transfer compared to other IgG molecules. Alternatively, an expansion of the current repertoire of larger molecule biologics (e.g. certolizumab) whose size prohibits efficient transplacental transfer or modification of Fc receptors that are not conducive to transplacental transport would also be very welcome.^28,93 Companies could also offer complimentary maternal drug serum testing and cord blood testing for all infants exposed and such data should be shared in their registries and freely accessible to all clinicians caring for (and counselling) women who are pregnant or planning pregnancy. Additional studies looking at the combined effects of biologic agents and other immunomodulators such ashydroxychloroquine or sulfasalazine on the risk of infection in neonates are required to further delineate the contributory immunosuppressive effects of these drugs on the risk of infection in both mother and neonate.

Biosimilars are flooding the market. Studies looking at their effects are therefore necessary given that it remains unclear how (dis)similar these molecular structures are and what potential impact they might have on pregnancy.

The ongoing accumulation and sharing of information on biologic use in pregnancy are of paramount importance. It is vital to publish case reports of early biologic use in pregnancy to facilitate information sharing.

In addition, the reporting of pregnancy complications and their outcomes should be standardised. Little information can be gleaned from published statements such as ‘delivery of a normal infant at term’. Hence, the collaborations of rheumatologists, gastroenterologists, oncologists, neurologists, obstetricians, neonatologists and paediatricians are crucial for more comprehensive data collection that would benefit multiple subspecialties. Long-term studies to determine the safety of these drugs in exposed offspring are urgently needed. Would better disease control in pregnancy mean better long-term outcomes for both mother and child? Collaborative efforts and the establishment of long-term registries for both the disease and the drugs they are exposed to are essential for a holistic view of how these biologics influence pregnancy and perhaps even the next generation.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

Not applicable.

Informed consent

Not applicable.

Guarantor

MCS.

Contributorship

MM contributed to the sections on reproduction and fertility; MCS is responsible for the concept and initial drafts of the manuscript. Both authors have contributed to the drafting and approval of the final version of the manuscript and have agreed to be accountable for all aspects pertaining to the accuracy and integrity of the entire manuscript.