Pregnant women taking tumor necrosis factor inhibitors at conception experienced a higher rate of spontaneous abortion than did patients who did not.
The data, culled from the British Society for Rheumatology Biologics Registry, is from the "largest detailed prospective collection of pregnancy outcomes in women with arthritis-related diseases exposed to anti-TNF therapy" to date, according to the authors.
Despite its relatively large size, however, the study was unable to control for the looming possibility that disease severity itself plays a role in adverse pregnancy outcomes, as the authors readily admitted.
Nevertheless, "no firm conclusions can be drawn about the safety of anti-TNF therapy during pregnancy and, without further evidence, guidelines which suggest these drugs should be avoided at the time of conception must remain," recommended the authors.
The study was led by Dr. Suzanne M. M. Verstappen of the University of Manchester’s Arthritis Research UK Epidemiology Unit and was published online in Annals of the Rheumatic Diseases.
Dr. Verstappen and her colleagues looked at women in the registry who received adalimumab, etanercept, or infliximab either at conception or at any time prior to conception. A subset was also exposed to methotrexate and/or leflunomide at time of conception in addition to the anti-TNFs, two drugs with a "known risk of adverse pregnancy outcomes," according to the authors.
A fourth cohort with active rheumatoid arthritis had no history of anti-TNF use, but rather received nonbiological disease-modifying antirheumatic drugs (DMARDs), excluding methotrexate and leflunomide (Ann. Rheum. Dis. 2011 [doi:10.1136/ard.2010.140822]).
Among cohort Ia, which included 20 women (21 pregnancies) who took anti-TNFs plus either methotrexate and/or leflunomide at conception, there were 10 live births, 4 terminations, and 7 (33%) spontaneous abortions (miscarriages occurring prior to 20 weeks or to viability outside the womb).
Among cohort Ib, which included 44 women who took anti-TNFs at conception, but not methotrexate or leflunomide, there were 50 pregnancies. They included 32 live births among this cohort, 4 terminations, 12 spontaneous abortions (24%) and 2 intrauterine deaths (occurring post-20 weeks).
There was also one neonatal death registered.
The women who had taken anti-TNFs in the past, but not at the time of conception (cohort II), did have seemingly better outcomes: the 59 pregnancies (54 women) resulted in live births in 46 cases (including one of two twins), terminations in 2, and spontaneous abortions in 10 (17%). There were two intrauterine deaths, including the twin death.
Finally, among the 10 pregnancies in 10 women who had no history of anti-TNF use (cohort III), there were zero terminations and one spontaneous abortion (10%).
In all cohorts, the majority of patients had RA, and the majority took etanercept.
The baseline disease activity score-28 (DAS28) was significantly higher in the anti-TNF cohorts, vs. cohort III: 6.5, 6.1, and 6.0 in cohorts Ia, Ib, and II, respectively, vs. 5.1 in cohort III.
Dr. H. Michael Belmont, medical director of New York University’s Hospital for Joint Diseases, as well as the director of the lupus clinic at Bellevue Hospital, New York, commented that, until more data are available, "The default choice would be to avoid these drugs during pregnancy. Certainly starting their use in a women contemplating conception should be delayed based on the data on hand."
Dr. Belmont, who was not affiliated with this study, joined the authors in pointing to a 2009 study by Carter et al., which found an increased observed: expected incidence of VACTERL defects (vertebral abnormalities, anal atresia, and/or cardiac, trachea, esophageal, renal, and limb abnormalities) in children born to women taking anti-TNFs (J. Rheumatol. 2009;36:635-41).
However, "The total observed number [of children born with VACTERL abnormalities] was small, 19 newborns, and the actual rate could not be calculated, as the authors relied on the [Food & Drug Administration] base of adverse events but did not know the [actual] number of exposed pregnancies, having to rely instead on historical controls to determine the expected ratio," he conceded.
On the other hand, Dr. Arthur Kavanaugh, a rheumatologist and director of the at the University of California San Diego’s Center for Innovative Therapy pointed to flaws in the current study’s analysis.
"The other explanation [for the finding of increased spontaneous abortion among anti-TNF users] is that it’s the disease more than the drug, so that people who have worse RA are going to have pregnancy outcomes that are not necessarily as good, and you just don’t know that from this because there’s only 10 people in the DMARD-only arm," he said in an interview.
"That’s just incredibly small. If you have one different outcome in that group, it would make a humongous difference in how you look at the data."