By Rebecca H. Allen, MD, MPH, Editor
Synopsis: In this retrospective cohort study among 5,142,339 pregnancies in Ontario, Canada, exposure to preconception computed tomography (CT) was weakly associated with spontaneous pregnancy loss (adjusted hazard ratio [aHR], 1.08; 95% confidence interval [CI], 1.07 to 1.08 for one CT scan; aHR, 1.14; 95% CI, 1.12 to 1.16 for two CT scans; and aHR, 1.19; 95% CI, 1.16 to 1.21 for three or more CT scans). For the 3,451,968 live births, there was a similar weak association with congenital anomalies (aHR, 1.06; 95% CI, 1.05 to 1.08 for one CT scan; aHR, 1.11; 95% CI, 1.09 to 1.14 for two CT scans; and aHR, 1.15; 95% CI, 1.11 to 1.18 for three or more CT scans).
Source: Simard C, Fu L, Odugbemi T, et al. Exposure to computed tomography before pregnancy and risk for pregnancy loss and congenital anomalies: A population-based cohort study. Ann Intern Med. 2025; Sep 9. doi: 0.7326/ANNALS-24-03479. [Online ahead of print].
Computed tomography (CT) commonly is used for radiologic diagnosis. The authors of this study sought to estimate the risk of CT scans to women of reproductive age who underwent CT scans prior to pregnancy. They theorized that the ovarian follicles may be damaged by CT scan radiation, which would lead to adverse effects months or years later.
This was a retrospective cohort study that used the validated health administrative databases for the province of Ontario, Canada, where all residents are insured by the Ontario Health Insurance Plan. The database stores diagnostic, procedural, and sociodemographic data, including all inpatient and outpatient radiology studies for Ontario residents. All pregnancies (spontaneous abortion, ectopic pregnancy, induced abortion, stillbirth at 20 weeks or more, or hospital live birth) of Ontario female residents aged 16 to 45 years with a valid Ontario Health Insurance Plan number were identified between April 1, 1992, and March 31, 2023. Women were excluded if they had prothrombotic risk factors, such as history of cancer, venous thromboembolism, inflammatory diseases, nephrotic syndrome, cirrhosis, sickle cell disease, or thrombophilia, or if they had CT scans during pregnancy.
The authors collected the number of CT scans (one, two, or three or more) any time up to four weeks before the estimated date of conception when the gestational age of the pregnancy was known, as well as the number of abdominal, pelvic, lumbar spine, or sacral spine CT scans that occurred. Head CT only served as a negative control. The primary outcome was spontaneous pregnancy loss (spontaneous abortion or ectopic pregnancy before 20 weeks of gestation or stillbirth at 20 weeks of gestation or more). The secondary outcome was any chromosomal or congenital anomaly among live births diagnosed between 0 and 365 days after birth.
The authors then evaluated the association between the number of preconception CT scans and each study outcome using Cox regression and adjusted hazard ratios. Hazard ratios were adjusted for year, age, residential area income quintile, rural residence, and comorbid conditions within two years of the date of conception, including obesity, diabetes, sexually transmitted infection, pelvic inflammatory disease, endometriosis, thyroid disorder, mental health condition, and smoking.
The authors found that among 5,142,339 pregnancies, 687,692 (13.4%) patients had a CT scan at a median of 50 months (interquartile range, 21-96 months) before conception. Of the 1,670,526 pregnancies that ended in spontaneous or induced abortion, gestational ages were known from the database for 155,276 (9.3%) of them; the rest were estimated from known median values. There were 535,165 (10.4%) pregnancies that ended in spontaneous pregnancy loss, including spontaneous abortion, ectopic pregnancies, and stillbirths. The adjusted hazard ratio for spontaneous pregnancy loss was 1.08 (95% confidence interval [CI], 1.07-1.08) for one CT scan, 1.14 (95% CI, 1.12 to 1.16) for two CT scans, and 1.19 (95% CI, 1.16 to 1.21) for three or more CT scans. These rates were slightly higher for CT scans of the abdomen, pelvis, and lower spine, and slightly lower for head CTs only.
The peak risk was at four to eight weeks prior to conception, with an adjusted hazard ratio of 1.24 (95% CI, 1.18 to 1.30). Among 3,451,968 live births, 225,002 (6.5%) congenital anomalies were diagnosed within the first year of life. The adjusted hazard ratios were 1.06 (95% CI, 1.05 to 1.08) for one CT scan, 1.11 (95% CI, 1.09 to 1.14) for two CT scans, and 1.15 (95% CI, 1.11 to 1.18) for three or more CT scans. There were no differences when the location of the CT scans was compared. There was no effect of timing of the CT scan on the risk of congenital anomalies.
Commentary
In this large, retrospective cohort study from the health system of Ontario, the authors found a weak but dose-dependent association with preconception CT scans and the subsequent risk of spontaneous pregnancy loss (spontaneous abortion, ectopic pregnancy, and stillbirth) and congenital anomalies. The authors found that the location of the CT scan and how recently the CT scan was performed influenced the risk of spontaneous pregnancy loss but not congenital anomalies. The authors surmised that the CT radiation could have caused deoxyribonucleic acid (DNA) damage in ovarian follicles. Although the authors did not suggest that CT scans be withheld when needed for diagnostic purposes, they did advise judicious use of CT scans in children and young adults, with consideration for magnetic resonance imaging (MRI) and ultrasound instead if possible.
The major strength of the study was the large database available to the investigators with validated radiology and pregnancy encounters and diagnoses for every patient. However, if there were patients who had spontaneous abortions but never presented for care, they would have been missed. Furthermore, any patient with a congenital anomaly that resulted in a spontaneous abortion or induced abortion would have been missed, since anomalies were not reliably documented in those instances as compared to the time of live birth and the 365 days after birth. In addition, combining ectopic pregnancies with spontaneous abortions and stillbirths was not helpful (in my opinion) because they have very different etiologies, not always related to the health of the ovarian follicle.
An abdominal CT scan exposes a patient to approximately 1.3 mGy to 35 mGy of radiation and a pelvic CT scan exposes a patient to 10 mGy to 50 mGy depending on the CT protocol.1 After fertilization, but before implantation, it is estimated that a dose of 50 mGy to 100 mGy would be required to either cause death of the embryo or no effect (the all or nothing hypothesis). For organogenesis, congenital anomalies are estimated to occur at 200 mGy. The American College of Obstetricians and Gynecologists recommends that abdominal and pelvic CT scans only be used in pregnancy if absolutely needed for a diagnosis where an MRI or ultrasound would not provide the necessary information.1 Oral CT contrast is safe to use in pregnancy. Intravenous contrast also is safe and is not known to cause birth defects, but it does cross the placenta, so it is advised to be used only when absolutely needed.
Very little is known about the effect of radiation on pregnancy outcomes when it occurs prior to conception, which is what this study was attempting to elucidate. I think this study raises intriguing questions but, as an epidemiologic study, it only can suggest correlations rather than causation. I agree with the recommendation for judicious use of CT scans, since it does expose patients to radiation. This already is the common practice of pediatric emergency rooms, in my experience.
Rebecca H. Allen, MD, MPH, is Professor, Department of Obstetrics and Gynecology, Warren Alpert Medical School of Brown University, Women & Infants Hospital, Providence, RI.
Reference
1. American College of Obstetricians and Gynecologists. Guidelines for Diagnostic Imaging During Pregnancy and Lactation. Committee Opinion No. 723. October 2017. Reaffirmed 2021. https://www.acog.org/clinical/clinical-guidance/committee-opinion/articles/2017/10/guidelines-for-diagnostic-imaging-during-pregnancy-and-lactation
