Serious Infections in Offspring Exposed in Utero to Vedolizumab

Introduction

Inflammatory bowel disease (IBD) peaks in incidence during reproductive years, and women with IBD are more likely to experience adverse pregnancy outcomes than the general population.¹ Therefore, finding ways of controlling IBD activity and maintaining remission during pregnancy is necessary to optimize pregnancy outcomes.

Biologics of several classes are highly efficacious in inducing and maintaining remission in IBD, including tumor necrosis factor inhibitors (TNFis), interleukin-12/23 inhibitors (ustekinumab), and integrin α₄β₇ inhibitors (vedolizumab). Limited evidence suggests these drugs do not increase the risk of adverse pregnancy outcomes including preterm birth, spontaneous abortion, or stillbirth.^2,3 These drugs, however, are typically immunoglobulins (IgGs), which are actively transported across the placenta during the second and third trimesters. Offspring exposed to vedolizumab in utero are born with lower blood drug (relative to maternal drug levels) compared with offspring exposed to TNFi or ustekinumab.^4,5 In addition, offspring exposed to vedolizumab are known to clear the drug quickly after birth.⁶ Tumor necrosis factor inhibitors and ustekinumab are known immunosuppresants with established risk of infection in exposed adults, raising concerns for potential immunosuppression in offspring exposed in utero. In contrast, vedolizumab acts locally on the gut and is not associated with increased risk of infection in adults,⁷ though risk in exposed offspring has not been dismissed. Limited research across 4 studies has so far detected no signal of increased risk (207 total pregnancies).^{5, 6, 8, 9}

Using a large population-based cohort of children born to mothers with IBD, we aimed to evaluate the risk of serious infections in offspring exposed in utero to vedolizumab.

Methods

We assembled a cohort of children born to mothers with IBD using the IBM MarketScan database (January 1, 2011, to December 31, 2018).¹⁰ We included pregnancies resulting in singlet live births using relevant International Classification of Diseases (ICD)-9/10 or procedure codes indicating delivery. From this cohort, we identified mothers who had been continuously enrolled in MarketScan for at least 12 months prior to delivery, with IBD diagnosed before delivery using previously validated case definitions.¹¹ Offspring were deterministically linked to their mother using family identifiers and delivery dates.^12,13

Drug exposure was defined as ≥1 filled prescription or infusion procedure code for a drug of interest during the pregnancy period. The main exposure of interest was vedolizumab; other exposure groups included exposure to TNFi, other biologics used in IBD (including ustekinumab), traditional systemic immunosuppressants (eg, azathioprine, mercaptopurine), or no drug exposure. If there were multiple exposures, the exposure category was mutually exclusive and assigned hierarchically, with vedolizumab > TNFi > other biologics > traditional immunosuppressants.

The outcome of interest was serious infection, defined as any inpatient ICD-9/10 code for infection of any type within the first year of life. Offspring were followed from birth until an event or censoring at end of follow-up (12 months of age), death, end of study period, or end of insurance eligibility.

Descriptive statistics were used to characterize the different exposure groups and crude infection risks. We performed logistic regression using generalized estimating equations (GEEs) to estimate crude and adjusted odds ratios (ORs) for serious infections with the drugs of interest compared with no use, adjusting for maternal age, preterm delivery, maternal preexisting or gestational diabetes, corticosteroid use, and concomitant use of another drug of interest.

Results

We identified 8507 offspring born to 7633 women with IBD, with descriptive statistics presented in Table 1. Forty-three offspring were exposed to vedolizumab. Of these, 7 were also exposed to TNFi and 13 to traditional immunosuppressants at some point during pregnancy.

Table 1.

Maternal characteristics of the study population (n = 8507).

	With Diseases of Interest
	Vedolizumab (n = 43)	TNFi (n = 1230)	Other Biologics (n = 17)	Traditional Immunosuppressants (n = 1822)	No Drug Exposure (n = 5395)
Maternal age (years), mean (SD)	31.6 (4.4)	31.7 (4.0)	32.6 (4.1)	32.6 (4.1)	32.4 (4.3)
Preexisting diabetes, n (%)	1 (2)	38 (3)	0 (0)	54 (3)	170 (3)
Gestational diabetes, n (%)	3 (7)	143 (12)	0 (0)	224 (12)	767 (14)
Pregnancy resulted in preterm birth, n (%)	2 (5)	133 (11)	2 (12)	191 (10)	569 (11)
Concomitant drug use, n (%)
Corticosteroid use, n (%)	14 (33)	220 (18)	2 (12)	263 (14)	361 (7)
Biologic, n (%)	7 (16)	1 (<1)	NA	NA	NA
Traditional systemic immunosuppresant, n (%)	13 (30)	240 (19)	2 (13)	NA	NA

Abbreviations: NA, not applicable

The cumulative incidence of serious infection at 1 year was 2.3% (95% confidence interval [CI], 0.4-12.0) in the vedolizumab group, similar to the rate in the TNFi (2.9%; 95% CI, 2.1-4.0), traditional immunosuppressants (2.5%; 95% CI, 1.9-3.3), and no drug exposure (3.0%; 95% CI, 2.6-3.6) groups. The risk appeared greater in the other biologic group (5.9%; 95% CI, 1.0-27.0), though confidence intervals were wide and overlapped for all categories. No cases of tuberculosis were detected.

Crude and adjusted ORs are presented in Table 2. Compared with unexposed children, we observed no clear excess risk of serious infections in offspring exposed in utero to vedolizumab (aOR 0.87; 95% CI, 0.13-6.15), although the CI was wide. The effect estimates also did not suggest an increased risk for those exposed to nonbiologic immunosuppressants or TNFi. There was, however, a potential trend towards an increased risk among children in the other biologic group (aOR 2.05; 95% 0.29-14.5), though the CI was imprecise.

Table 2.

Crude and adjusted odds ratios for serious infections across offspring in different exposure groups (n = 8507).

Exposure Group	Crude ORs (95% CI)	Adjusted ORs (95% CI)
No drug exposure	Ref	Ref
Vedolizumab	0.76 (0.11-5.52)	0.87 (0.13-6.15)
TNFi	0.96 (0.66-1.38)	0.95 (0.61-1.40)
Other biologic (ustekinumab, natalizumab)	2.11 (0.31-14.2)	2.05 (0.29-14.5)
Traditional immunosuppressants	0.80 (0.58-1.12)	0.78 (0.55-1.09)

Discussion

We did not detect a substantially increased risk of serious infection requiring an inpatient stay within the first 12 months of life in offspring exposed in utero to vedolizumab vs unexposed offspring born to mothers with IBD, though estimates were imprecise. This aligns with data from the PIANO study,⁵ a French retrospective cohort study,⁸ and the pan-European CONCEIVE study⁹; although study design and outcome definitions varied across the studies, and all differed from ours.

Our study has several strengths. We are the first to use a large administrative database to provide population-based estimates of the risk of serious infections in IBD offspring exposed in utero to vedolizumab. Our data source includes information on medical diagnoses and drugs in mother and offspring, allowing us to assess exposure, outcome, and covariates using established algorithms without the potential for recall bias. IBM MarketScan, which includes more than 230 million individuals, also allows for rare outcomes and exposures to be assessed and has been widely used to conduct similar pharmacoepidemiologic studies.

Our study also has potential limitations. There may have been residual confounding from disease activity. Though we attempted to adjust for this using corticosteroid and/or concomitant drug use as proxy for disease activity, these are not perfect. This said, disease activity is likely associated with serious infection through preterm birth, which we did account for, so any residual confounding likely only modestly affects the effect estimates. In administrative database research, imperfect case ascertainment is always a concern. To circumvent this, we used previously established case definitions with good validity.¹¹

In conclusion, we did not detect a clear excess risk for offspring exposed in utero to vedolizumab compared with unexposed offspring born to mothers with IBD. Ongoing caution and more research on short- and long-term effects are warranted for vedolizumab and other biologics that are actively transported across the placenta.

Author Contributions

J.G.: Designed study, collected and analyzed the data, drafted the manuscript

S.B.: Designed study, collected and analyzed the data, drafted the manuscript

W.A.: Provided important intellectual content and critically revised the manuscript

Y.S.P.: Collected and analyzed the data

K.F.: Designed study, provided important intellectual content, and critically revised the manuscript

E.V.: Designed study, collected and analyzed the data, drafted the manuscript

Funding

This work was supported by a Canadian Institutes of Health Research (CIHR) operating grant (419778). E.V. receives salary support from the Arthritis Society New/Mid Investigator: Stars Career Development Award (number STAR-19-0597) and the Fonds de recherche du Québec - Santé (FRSQ) Junior 2 Award (number 282178). K.F. is supported by a Senior Salary Support Award from the Fonds de recherche du Québec - Santé (FRQS; Quebec Foundation for Research—Health) and a William Dawson Scholar award from McGill University.

Conflicts of Interest

W.A. received consultancy fees from Abbvie, Amgen, Arena Pharmaceuticals, Dynacare, Janssen, Merck, Novartis, Pfizer, Sandoz, and Takeda. All other authors have nothing to declare.

Data Availability

Data was obtained from the IBM MarketScan commercial database. Access to the data can be obtained from the company for a fee.