MRI in Pregnancy and the Pregnant Worker — Fetal Exposure Evidence, Operating Mode and Field Strength, Consent, and Occupational Exposure of Pregnant Staff: Where Guidance Agrees and Differs

MRI in Pregnancy and the Pregnant Worker — Fetal Exposure Evidence, Operating Mode and Field Strength, Consent, and Occupational Exposure of Pregnant Staff: Where Guidance Agrees and Differs

MRIninja Knowledge Base | Child Page — MRI Safety (deep dive S10) Version 1.0 — October 2026

Parent page: MRI Safety — Master Page (hazard table in Section 3). Builds on: Gadolinium-Based Contrast Agent Safety (gadolinium in pregnancy and lactation) · Acoustic Noise in MRI (acoustic noise and fetal hearing) · RF Heating and Burns (RF heating) · Static Magnetic Field Contours (staff exposure in the fringe field)

1. Scope, Status and Boundaries

This page gathers the evidence and the guidance on two groups: the pregnant patient examined without gadolinium contrast, and the pregnant member of staff. It covers what the sources state about the physical agents (static field, heating, noise), the choice of operating mode and field strength, the epidemiological evidence on fetal and childhood outcomes after first-trimester MRI, consent and documentation, and the working arrangements for pregnant staff. Topics held elsewhere are not repeated: gadolinium in pregnancy and lactation, including the Ontario cohort results for contrast-enhanced scans (Gadolinium-Based Contrast Agent Safety); acoustic noise, hearing protection and the neonatal hearing studies (Acoustic Noise in MRI); the mechanism of RF heating and the SAR limits (RF Heating and Burns, SAR, Bioeffects, and MRI Safety Physics); the zones, the contours and staff exposure in the fringe field (MR Site Zones and Access Control, Static Magnetic Field Contours); and the pregnancy screening questions (Patient Preparation, Section 6.3).

Reading status. The page rests on the fifth edition (July 2026) of the UK MHRA guidelines (Sections 2.4.1, 2.6 and 3.4 as verbatim quotations returned by a summarising tool; the text ended in Section 4.1.4, so Chapters 4 and 5 and Appendix 2 were not read), the 2017 American College of Obstetricians and Gynecologists committee opinion (read in full), the 2022 Canadian Association of Radiologists recommendations (read through a summarising tool; references truncated), a large Ontario cohort (2016), a narrative review (2025, abstract), one 1993 survey of female MR workers (abstract), and two small items known only from a news report and a survey abstract. The ACR Manual on MR Safety, the standards (IEC 60601-2-33), the ICNIRP statements, the EU Directive 2013/35/EU text and the current versions of other national guidance were not read, and no systematic review or meta-analysis of MRI exposure in pregnancy was read.

What this page does not do. It does not give a policy for any facility or a clearance for any patient or worker. Statements that are the page’s own inference are marked Reasoning; where a source needed to complete a statement was not found they are marked (to be sourced).

2. Sources Read and What They State

Table 1. Documents used and how they were read.

3. Physical Agents and the Fetus: What the Guidelines State

The MHRA guidelines state that MRI is frequently employed in pregnancy because of the risks of ionising radiation imaging, but that it should be performed with caution because of the potential risks associated with exposure to the static magnetic field [1]. On the static field the guideline recommends caution above 4 T and states that there is no clear evidence that exposure to static or low-frequency magnetic fields adversely affects pregnancy outcomes [1]. On heating it states that excessive heating is considered a potential teratogen, that exposure should be minimised and standard operating levels used, that fetal temperature is typically about 0.5 °C higher than maternal temperature, that limiting the maternal core temperature rise to 0.5 °C helps ensure that the fetal rise stays below about 1 °C, that current SAR guidelines are primarily based on adult exposure, that SAR modelling does not include a dedicated model for pregnancy, that RF shimming should be avoided unless explicitly supported by safety data and robust electromagnetic modelling, and that for routine fetal or maternal MRI standard circularly polarised transmit modes remain the safest approach [1]. On noise and vibration it states that routine prenatal MRI at 1.5 T and 3 T with currently available gradient systems is safe, that the potential effects on the fetus are not fully established with respect to elastography vibration, and that caution should be applied [1]; the noise evidence is in Acoustic Noise in MRI. The ACOG opinion, quoting the American College of Radiology, states that tissue heating is proportional to the tissue’s proximity to the scanner and is negligible near the uterus, and that there is no evidence of actual harm, though theoretical concerns exist [2].

Reasoning. The two documents differ in how they describe heating: the MHRA text treats it as a potential teratogen to be managed by mode and by a temperature limit, with an explicit statement that the models were built for adults, while the ACOG text describes it as negligible near the uterus. The sources read give no measured fetal temperature rise in patients (to be sourced); the figures quoted are statements of the guidelines and not data.

4. Operating Mode and Field Strength

The MHRA guidelines state that pregnant patients should be scanned in normal operating mode whenever possible, with a preference for 1.5 T during the first trimester; that more conservative imaging is often adopted during early pregnancy because the embryo may be more susceptible during organogenesis; that scanning with the whole-body RF transmit coil should be limited to the normal operating mode; that procedures should be performed only after a careful risk-benefit assessment, particularly in the first trimester; that reduced-noise techniques should be employed whenever possible; that where scanning in First Level Controlled mode or the use of gadolinium is clinically necessary the decision should be made collaboratively by the referring clinician, the MR clinician, the patient and where necessary the MR safety expert; and that gadolinium should be avoided unless clinically essential [1]. The Canadian recommendations state that pregnant patients should be scanned in normal operating mode with a whole-body SAR limit of 2 W/kg, that 1.5 T is preferred in the first trimester if both field strengths are available, that non-contrast 1.5-T MRI is considered safe at any trimester, and that first-trimester fetal MRI at 1.5 T or below is generally deemed safe if clinically necessary [3]. The ACOG opinion, quoting the American College of Radiology, states that no special consideration is recommended for the first versus any other trimester and that there are no precautions or contraindications specific to the pregnant woman [2]. A 2025 review states that studies show no significant harm at 1.5 T or 3 T although caution is still advised in the first trimester [5].

Table 2. Position of each source on operating mode, field strength and the first trimester.

Reasoning. The documents agree that normal mode is the standard and that MRI is preferred to ionising radiation when imaging is needed; they differ on whether the first trimester deserves a separate rule and on whether 1.5 T should be chosen over 3 T when both are available. The Canadian text calls its preference “an abundance of caution”, which signals that the preference is precautionary and not a finding of a measured difference. None of the sources read reports a comparison of 1.5 T and 3 T outcomes in first-trimester patients.

5. Epidemiological Evidence on Fetal and Childhood Outcomes

The largest study read is a retrospective population-based cohort of 1,424,105 deliveries in Ontario (births after 20 weeks, April 2003 to March 2015): 1,737 pregnancies with first-trimester MRI were compared with 1,418,451 unexposed pregnancies. The adjusted risk ratio for stillbirth or neonatal death was 1.68 (95 per cent confidence interval 0.97 to 2.90; 19 exposed events), and the adjusted hazard ratios were 1.16 (0.96 to 1.40) for congenital anomaly, 1.50 (0.94 to 2.40) for vision loss, 1.04 (0.75 to 1.45) for hearing loss and 0.53 (0.08 to 3.67) for neoplasm (five or fewer exposed events); in a subgroup exposed at 5 to 10 weeks the hazard ratio for vision loss was 2.28 (1.09 to 4.77) [4]. The authors conclude that first-trimester MRI was not associated with a significant increase in the outcomes studied and that the study may not have detected rare adverse outcomes [4]. Figure 1 shows the estimates.

Figure 1 — First-trimester MRI exposure and fetal and childhood outcomes in an Ontario cohort (adjusted estimates with 95 per cent confidence intervals)

The authors state that the study was underpowered for uncommon outcomes, that the findings are exploratory and unadjusted for multiple comparisons, that pregnancies ending before 21 weeks were excluded so miscarriage after MRI is unknown, that pregnancy awareness at the time of MRI and the indication for MRI were not available, that no data were available on teratogenic medications or alcohol, that follow-up to age 4 was 61 per cent in the first-trimester cohort, and that most scans used 1.5-T scanners and the field and RF exposure could not be measured [4]. A conference abstract known from a news report compared 81 newborns exposed to 3-T MRI in utero with 162 unexposed newborns and found no significant difference in mean birth weight and no hearing impairment on neonatal screening in either group, without effect sizes, confidence intervals or gestational age at exposure [6]. The 2025 review states that there is no conclusive evidence of adverse effects of fetal MRI on development and calls for larger, prospective, long-term studies [5]. The neonatal hearing studies at 1.5 T are in Acoustic Noise in MRI.

Reasoning. The interval around the stillbirth or neonatal death estimate extends to nearly threefold, so the study can neither show that first-trimester MRI is free of effect on this outcome nor that it raises it; the same holds for vision loss, where the all-first-trimester estimate also includes 1, and the 5-to-10-week subgroup is one of several comparisons made without correction for multiplicity. The comparison group (no MRI) differs from the exposed group in the clinical reason for imaging, which the authors could not account for; the comparison can therefore not separate the effect of the scan from the effect of the condition that led to it. The estimates are not a measure of the effect of any one field strength or sequence.

6. Contrast, Noise and Consent

The evidence and guidance on gadolinium in pregnancy are in Gadolinium-Based Contrast Agent Safety (Section 9), and the evidence on noise and fetal hearing in Acoustic Noise in MRI (Section 7.2); they are not repeated. The MHRA, ACOG and Canadian texts all restrict gadolinium to situations of clear need, with differing wording: avoid unless clinically essential [1]; only if it significantly improves diagnostic performance [2]; not recommended unless absolutely necessary, with a Discussion section that calls it contraindicated [3].

On consent the sources differ. The Canadian panel reached a consensus that informed consent should be obtained before non-contrast 3-T MRI in the first trimester, left consent for other non-contrast scans to local institutional policy, recommended consent for contrast-enhanced scans, and states that the panel was divided on whether consent is needed for non-contrast MRI [3]. The ACOG opinion does not address consent [2]. The MHRA text read states that decisions on first-level mode or gadolinium should be made collaboratively with the patient [1]. The practice of documenting pregnancy status and obtaining written consent in the first trimester is on the Patient Preparation page and attributed there to another source [2]. The sources read do not report any study of consent or documentation practice or of its effect on outcomes (to be sourced).

7. Pregnant Staff

The MHRA guidelines state that MR staff may continue their normal duties during pregnancy; that it is considered prudent that pregnant staff do not remain within the MR examination room during image acquisition (the guidelines give fetal acoustic noise concerns as the reason in one passage); that employers are required to undertake a specific risk assessment for expectant mothers and, in Section 3.4.1, a particular requirement to carry out a risk assessment for pregnant staff; that employers may wish to limit pregnant staff to the exposure limits for members of the general public (the appendix to which this refers was not read); and that employees with implants or body-worn devices or who are pregnant must have a risk assessment in place [1]. The Canadian recommendations state that pregnant MR personnel are safe to work around and within the 1.5-T and 3-T environment and, as a precaution, should not stay in the bore or magnet room during data acquisition, although they may enter when needed to position patients or in an emergency [3]. A survey of United Kingdom units reported, in an abstract, that employers should carry out risk assessments for pregnant staff and that clearer guidance was needed [8].

The only occupational reproductive study read is a 1993 questionnaire survey of women working at more than 90 per cent of United States clinical MR facilities: of 1,915 completed questionnaires, 1,421 pregnancies were reported, 280 of them of MR workers (technologists or nurses) and 894 of employees in other jobs; relative risks for MR workers against other employees were 1.27 (95 per cent confidence interval 0.92 to 1.77) for spontaneous abortion, 0.90 (0.54 to 1.51) for conception taking longer than 12 months, 1.19 (0.76 to 1.88) for delivery before 39 weeks, 1.01 (0.50 to 2.04) for birth weight below 2.5 kg and 0.99 (0.80 to 1.22) for male sex of the offspring, with adjustment for age, smoking and alcohol not changing the pattern; the authors conclude that there is no substantial increase in these outcomes [7].

Table 3. Evidence on pregnant staff, source by source.

Reasoning. All confidence intervals in the 1993 survey include 1.0, and the point estimate for spontaneous abortion is above 1; the study can exclude a large increase but not a modest one, and it describes the work of 1990 and not the systems and practices of today. No dosimetry for staff in the fields of a magnet suite was read (to be sourced). Guidance on pregnant staff therefore rests on risk assessment and on precaution against noise exposure in the room, not on measured fetal exposure. The limits for staff exposure in the fringe field and the action levels are in Static Magnetic Field Contours, and the legal basis (the EU Directive 2013/35/EU, which names pregnant workers among workers at particular risk) is in the master page.

8. Practice: the Pathway the Sources Support

Reasoning. The following arrangement is the page’s synthesis of statements in the sources and is not itself a recommendation of any one source. For the patient: (1) establish pregnancy status and the gestation (Patient Preparation, Section 6.3); (2) confirm that MRI is expected to answer a relevant clinical question and that no non-ionising alternative would do, which both the MHRA and the ACOG text require in different words [1][2]; (3) use normal operating mode, and treat first-level mode as a collaborative decision [1][3]; (4) apply the local rule on field strength and consent, knowing that the sources differ (Table 2); (5) use reduced-noise techniques whenever possible (Acoustic Noise in MRI) and avoid gadolinium unless clinically essential (Gadolinium-Based Contrast Agent Safety) [1]; (6) document the decision. For staff: (7) a documented, individual risk assessment before and during the pregnancy [1]; (8) agreement on tasks that require presence in the room during acquisition, using the precaution stated in the MHRA and the Canadian text [1][3]; the Canadian text allows entry when needed to position patients or in an emergency [3]. The sources read do not say how the documentation should be held or who should sign it (to be sourced).

9. Where the Documents Differ and Open Questions

Table 4. Differences and unresolved questions.

10. Limits of the Evidence and What Was Not Read

The evidence on patients is one large retrospective cohort, one news-reported conference abstract, one narrative review and the guidelines themselves; the evidence on staff is a single survey of 1993 and one survey abstract. No randomised or prospective study and no meta-analysis was read, and the data on 3 T and on first-trimester scans are sparse. The ACR Manual on MR Safety, the IEC and ICNIRP documents, the EU Directive text, current national guidance outside the United Kingdom, the United States and Canada, and the MHRA appendix on exposure limits were not read, and the MHRA statements are quotations from a summarising tool. Absence of observed harm is bounded by the sample size and by what the studies could measure.

11. Topics Not Covered

Gadolinium and lactation (Gadolinium-Based Contrast Agent Safety); acoustic noise and fetal hearing studies (Acoustic Noise in MRI); the mechanism of RF heating and the temperature modelling of the fetus (RF Heating and Burns, SAR, Bioeffects, and MRI Safety Physics); staff exposure in the fringe field, the action levels and the legal basis (Static Magnetic Field Contours and the master page); the screening questionnaire (Patient Preparation); fetal MRI protocols and image quality; MRI-guided interventions in pregnancy; ultra-high-field systems; implants in pregnant staff (see the pages on implants); and the rules of individual jurisdictions (Incident Learning, Governance, Quality and Jurisdictions in MR Safety).

12. Evidence-Based References

No systematic review or meta-analysis of MRI exposure in pregnancy was read; the sections below follow the hierarchy of the sources actually read.

A. Guidelines, Consensus, Standards and Regulation

B. Cohort Study and Review

C. Survey of Exposed Workers

D. Items Known Only from a News Report or an Abstract

End of document — MRI in Pregnancy and the Pregnant Worker (S10) — MRIninja v1.0 — October 2026. Parent: MRI Safety master.