Imaging in Pregnancy: Safety and Decision-Making

Pregnancy imaging decisions are built on two facts. Ultrasound and MRI use no ionising radiation, so they are preferred whenever they can answer the clinical question. Most diagnostic radiographs and many justified CT studies still deliver fetal doses well below the level associated with deterministic effects such as malformation or neurodevelopmental injury.

The task. Match the modality to the question, optimise dose when ionising radiation is used, and counsel the patient honestly. Orthopaedic practice meets this most often in trauma, suspected pelvic or spinal injury, infection, and severe back pain with neurological compromise.

The maternal emergency comes first. In trauma, suspected cauda equina syndrome, septic arthritis, an unstable fracture or suspected visceral injury, the correct principle is to image decisively. A missed maternal diagnosis is usually a bigger threat to mother and fetus than the radiation from a justified diagnostic study.

The doses. Everything is compared against natural background over the whole pregnancy, and the figures examiners reward are the realistic ones:

The thresholds. Deterministic effects (malformation, growth restriction, neurological injury) have a practical threshold on the order of 100 to 200 mGy and are not expected below about 50 mGy. ICRP states that a fetal dose below about 100 mGy is not a reason to terminate a pregnancy. Every study in the table except iodine-131 sits below that 50 mGy figure, which is why the honest answer to "is it safe?" is almost always yes for a single justified examination: essentially every diagnostic orthopaedic radiograph or extremity CT is orders of magnitude below the threshold, and even a CT of the abdomen and pelvis typically stays well under it.

View larger

Dose reduction is about protocol, not refusal. When radiographs or CT are the right test:

• Justify the study
• Collimate to the region of interest
• Use a single phase and avoid multiphase acquisitions
• Apply low-dose or iterative-reconstruction CT protocols
• Avoid repeats

The shielding update. Routine external lead "fetal" or "gonadal" shielding is no longer recommended by many bodies (the AAPM among them). It can obscure anatomy, trigger the automatic exposure control to increase the dose, and does little against internal scatter. Optimising the protocol, not a lead apron over the abdomen, is what lowers fetal dose.

Iodinated contrast. It may be used when contrast-enhanced CT is the correct study, and the decision should rest on clinical necessity rather than reflex avoidance. It does cross the placenta, and the theoretical concern is transient suppression of the fetal thyroid by the iodine load; that concern has never been borne out as clinically significant. A single check of neonatal thyroid function in the first week is reasonable and is what the ACR suggests, but in most countries the routine newborn heel-prick screen already covers it, so no extra test is usually needed. Do not withhold a needed contrast-enhanced CT over this.

Breastfeeding. It does not need to be interrupted after contrast, for either iodinated or gadolinium agents: the amount excreted into breast milk and then absorbed by the infant gut is negligible (ACR and ESUR). Telling a mother to "pump and dump" after contrast is a common, outdated myth.

Diagnostic imaging is only half of the orthopaedic problem. If the pregnant patient has a fracture that needs fixing, the image intensifier will be on for the duration of the operation, and this is the part of the pathway that is most under a surgeon's own control.

The most useful finding here is counter-intuitive. In a series of eight pregnant women undergoing open reduction and internal fixation of an acetabular fracture between 5 and 26 weeks' gestation, intraoperative pelvic fluoroscopy averaged 39 seconds and conferred the least fetal exposure of any imaging in the pathway, while the pre-operative pelvic CT conferred the most, exceeding 50 mGy in every case that required one. All eight pregnancies reached 36 weeks with normal Apgar scores and seven of eight reductions were anatomic. Evidence level 4: a small retrospective case series with calculated rather than measured fetal doses.

What follows. The place to economise on fetal dose is the pre-operative CT protocol, not the fixation: hobbling the fluoroscopy to protect the fetus buys almost nothing and risks a malreduced acetabulum, which is a lifelong problem for the mother. The figure above 50 mGy sits above the 10 to 35 mGy usually quoted for CT abdomen and pelvis because these were trauma-protocol pelvic CTs concentrated directly on the fetus rather than dose-optimised modern acquisitions. Both figures are real; quote the one that matches the protocol being discussed.

Practical dose reduction with the C-arm. The habits that lower fetal dose:

• Pulsed rather than continuous screening, at the lowest frame rate that lets you work, and last-image-hold instead of re-screening to look at what you just did
• Collimate tightly to the field of interest and avoid magnification, which raises dose rate
• Keep the fetus out of the primary beam wherever the anatomy allows, and remember that scatter falls with the square of distance
• Position the intensifier close to the patient and the tube far from it; never operate with the tube above the table when it can be below
• Use a mini C-arm for extremity work: its dose is a fraction of a standard unit and the fetus is nowhere near the beam
• Brief the radiographer before the incision, and record screening time and dose-area product so a physicist can calculate an actual fetal dose afterwards if anyone asks
• For non-urgent surgery, the second trimester is conventionally preferred; after about 20 weeks use left lateral tilt to avoid aortocaval compression, which is a circulatory rather than a radiation consideration but belongs in the same plan

The Pregnant Member of Theatre Staff

The question is asked at least as often about the surgeon, scrub nurse or radiographer as about the patient, and the answer is reassuring. Modelling scattered dose in orthopaedic theatre during fluoroscopically assisted spinal surgery found that even if the annual occupational dose limits were fully exhausted every year for ten years, the resulting risk of congenital malformation would be at least two orders of magnitude below the spontaneous background rate (level 5: phantom simulation with scatter measurement).

Practically. Declare the pregnancy so that monitoring applies, wear an additional dosimeter under the lead at waist level, and observe the conventional limit of about 1 mSv to the fetus over the declared remainder of the pregnancy, a figure that routine orthopaedic practice with proper lead and sensible positioning does not come close to. Standing on the intensifier side rather than the tube side, and taking one step back during screening, does more than anything else.

View larger

Gestational age matters because the fetus passes through phases of differing radiosensitivity. Counselling and protocol decisions should reflect this, while remembering that diagnostic doses rarely reach the relevant thresholds: the window changes the counselling, not the indication.

View larger

Counselling with real numbers. "Counsel honestly" only works if the risk is anchored to the baseline risks the patient already carries. In the general population the risk of a major congenital anomaly is about 3 per cent (roughly 1 in 33); a single diagnostic study adds a negligible increment to that and to the baseline childhood-cancer risk in the table, dwarfed by the baseline. Even the highest common diagnostic exposures do not reach the level at which termination would be discussed.

Threshold versus no threshold. Deterministic effects have a practical dose threshold, whereas stochastic (cancer) risk is conventionally modelled as having none. Presenting the linear-no-threshold model without the tiny absolute numbers can frighten a patient into refusing necessary imaging or seeking inappropriate termination. Balanced counselling gives both the very low absolute risk and the benefit of an accurate diagnosis.

Practical consent. Document the indication and the discussion. For the highly anxious patient, or one who has had multiple or high-dose studies, ask a medical physicist for an estimated fetal dose. If a study was already done before the pregnancy was known, reassure with the actual estimated dose; the vast majority are far below any threshold.

Not every musculoskeletal complaint in pregnancy is mechanical. Several pregnancy-associated conditions can mimic trauma or degenerative disease, and recognising them changes both imaging choice and management.

Transient osteoporosis of the hip. The radiograph is often unremarkable, and an unremarkable radiograph neither excludes the diagnosis nor justifies withholding MRI. The MRI is the film that keeps these patients non-weight-bearing until the oedema settles.

View larger

View larger

Sacral insufficiency fracture. Peripartum sacral pain with an empty radiograph is the pattern; the next test is MRI, not another film and not a shrug that the pain is mechanical.

View larger

View larger

Cauda equina syndrome. Saddle anaesthesia and a litre of residual urine are not mechanical back pain. Urgent non-contrast MRI, then decompression.

View larger

View larger

Global Epidemiology

Trauma complicates roughly 6 to 8 percent of pregnancies and is a leading non-obstetric cause of maternal death, which is precisely why decisive maternal imaging is emphasised worldwide. Musculoskeletal complaints are near-universal in pregnancy: pelvic girdle pain and low back pain affect a large proportion of pregnancies, the majority mechanical and self-limiting, but a minority masking insufficiency fracture, transient osteoporosis, infection or neural compression.

Side-by-Side Guidance

High- vs Limited-Resource Practice

In well-resourced settings, urgent MRI and dose-optimised multidetector CT are usually available around the clock, so the safest adequate test is rarely out of reach. In limited-resource settings, MRI may be unavailable or delayed; here the justified use of radiographs and CT becomes more important, and ultrasound (including FAST in trauma) carries even greater weight as a radiation-free, bedside, low-cost first-line tool. The underlying principle is universal: match the modality to the question, optimise dose, and never withhold imaging that will change maternal management.

How strong is the gadolinium signal? The same 2016 Ontario cohort found a statistically significant association between gadolinium MRI and stillbirth or neonatal death, and with a broad rheumatological, inflammatory or infiltrative skin outcome. The absolute numbers were small and confounding by indication is hard to exclude. Most authorities therefore counsel avoidance of gadolinium unless essential, while acknowledging that the evidence is observational rather than causal.

Is first-trimester MRI safe? Theoretical concerns about tissue heating and acoustic noise persist, and many centres still prefer to defer non-urgent MRI beyond the first trimester out of caution. Set against the cohort's finding of no increase in fetal or early-childhood harm, that is a precautionary convention rather than an evidence-based prohibition.