Pregnancy testing before MRI for neuroimaging research: Balancing risks to fetuses with risks to youth and adult participants

While conducting research, it is crucial to protect our participants from undue risk. For human neuroimaging, many institutions require excluding pregnant people from completing magnetic resonance imaging (MRI) for research to minimize potential risk. This often entails mandatory urine pregnancy testing for all people assigned female at birth who have started to menstruate. This restriction stemmed from early concerns about fetal health from animal studies that have not been substantiated through human research. Yet, some institutions do not require this and many researchers are actively pursuing human fetal MRI.

Animal studies conducted in the 1980–90s raised concerns that MRI during pregnancy might confer negative effects on offspring (1-5), with acoustic noise and tissue heat absorption being typical mechanisms of concern-rather than radiation or radioactive tracers with X-ray, CT, or PET. However, several animal studies employed long-duration scanning (e.g., 4 or 8 hours) not realistic or typical for human neuroimaging studies, e.g., ~30-90-minute head MRI. Furthermore, environmental noise is damped in utero and the specific absorption rate (SAR) is tightly limited under normal scanning parameters to avoid heating (6).

More recent studies examining (medically indicated) prenatal MRI exposure in humans have found no significant effects on offspring outcomes (7-16). A 2020 review concluded that non-contrast MRI with typical field strength and acquisition protocols appears safe during all trimesters of pregnancy (17). Findings from this review were adapted and updated in Table 1 (17). As scientists, we often observe noisy or inconsistent results. Here, the pattern of results is unequivocal: no study has identified harm to children from MRI exposure in utero.

Table 1.

Results from studies examining the safety of MR imaging during pregnancy, adapted and updated from Lum & Tsiouris, 2020 (8)

Study (citation)	N Exposed	N Unexposed	Field Strength	Trimester	Outcomes
Zvi et al. 2020 (8)	131	771	1.5 T	all	No difference in birth weight or adaptive behavior composite scores at 2.5-6 years
Chartier et al. 2019 (9)	81	162	3 T	all	No difference in neonatal hearing impairment or birth weight
Jaimes et al. 2019 (10)	62	62	1.5 T v 3 T	2nd-3rd	No difference in neonatal hearing impairment between 1.5 T and 3 T
Ray et al. 2016 (11)	1737	1.4 M	Not recorded	1st	No difference in perinatal death rate, congenital anomalies, vision loss, hearing loss, or tumor
Bouyssi-Kobar et al. 2015 (12)	72	0	1.5 T	2nd-3rd	Typical birth weight, No neonatal or preschool-age hearing loss or functional impairment
Choi et al. 2015 (13)	15	0	1.5 T	1st	No negative health effects attributable to MRI at birth
Strizek et al. 2015 (14)	751	10,042	1.5 T	2nd-3rd	No difference in neonatal hearing impairment or birth weight
Reeves et al. 2010 (15)	103	0	1.5 T	2nd-3rd	No increased risk of neonatal hearing impairment
Kok et al. 2004 (16)	35	0	1.5 T	2nd-3rd	No negative health effects attributable to MRI at birth or 3m follow-up

Whereas most studies in Table 1 examined exposure during the 2^nd-3^rd trimesters (12, 14-16), outcomes following MRI during the first trimester were examined in a study of >1.4 million pregnancies (11). This study concluded that “Exposure to MRI during the first trimester of pregnancy, compared with nonexposure, was not associated with increased risk of harm to the fetus or in early childhood” (11). As such, the safety of MRI during pregnancy is supported by numerous organizations including the American College of Obstetricians and Gynecologists, the American College of Radiology, the Canadian Association of Radiologists, and the Medicines and UK Healthcare products Regulatory Agency (18-21).

Critically, many participants experience urine pregnancy testing as uncomfortable, stressful, invasive, or burdensome–even to the point of refusing participation in research. Mandatory testing is also invalidating of participant self-report and can feel paternalizing. Many participants (e.g. asexual, queer, and lesbian) are able to accurately report that they are not pregnant as they are not sexually active with men (22). Non-binary or trans participants–e.g., those assigned female at birth who identify as male–may find mandated pregnancy testing additionally aversive or invasive. We should also acknowledge the unmitigated risk of stress from false positive tests.

In pediatric research, testing is often required after menarche, yet the age of initiating sexual intercourse is often years later (23-25). Anecdotally, parents of youth in our studies have expressed disbelief that staff cannot rely on their report of their daughters’ pregnancy status. Furthermore, one is generally required to report a minor’s positive test to their caregiver, a policy out of step with state regulations. For example, New York provides minors with the right to confidential pregnancy testing, recognizing that not all parents provide safe environments, and therefore, disclosure can be destructive or harmful (26).

Lastly, on June 24, 2022, the United States Supreme Court overturned Roe v Wade, ending the constitutional right to abortion. While some states, e.g., New York (27), have moved to protect the right to abortion, others have moved to remove that right and to criminalize abortion. A woman who miscarried at a Texas hospital was arrested before the overturning of Roe v Wade after telling a hospital employee that she had tried to induce an abortion (28). Though charges in this case were dropped, it is possible in a post-Roe U.S.A. that documentation of first trimester pregnancy could be used as legal evidence against women charged for having an abortion. Thus, collecting pregnancy data in places where abortion may become criminalized could constitute a perceived or actual threat for participants, particularly if they choose to later terminate the pregnancy. This is especially salient given the current level of concern about data privacy, e.g., from period tracking phone apps (29), that could deter participation to avoid unnecessary pregnancy testing.

More broadly, it is important to consider what it means to be collecting and storing first trimester pregnancy data. For participants who live in places seeking to criminalize abortion, we could potentially be gathering “incriminating” data. Furthermore, it is yet to be tested whether an NIH Certificate of Confidentiality is sufficient to prevent pregnancy-related information from being released in the case of a lawsuit in all states (30). Thus, with rapidly shifting legal landscapes, collecting pregnancy-related information is increasingly dangerous for both researchers and participants.

Overall, pregnancy testing before research MRI was mandated when the risks to fetuses were unclear. These risks have not been substantiated, and the risk-benefit ratio of pregnancy testing has clearly shifted. There is recognition among experts that MRI is safe to conduct in fetuses and thus among participants who may be pregnant. Many institutions, including Emory University, Rutgers University, University of California Berkeley, University of Maryland, and the University of Pennsylvania, already do not require pregnancy testing before research MRI (31, 32) and provide a framework for waiver-based alternatives (31, 33). For example, researchers can present the risks (i.e., though MRI is safe, we cannot rule out all possibility of risk), allow participants to sign a release and provide optional pregnancy testing.

We conclude that for brain MRI studies at typical field strength (1.5T, 3T) with standard acquisition parameters (managing SAR appropriately, etc.) without contrast (concerns about gadolinium are substantiated (34)) pregnancy testing should be voluntary rather than mandated. We hope that this information will be useful in weighing the risks and benefits of pregnancy testing, and in opening conversations with MRI centers and institutional review boards to reconsider current practices.