By carefully assembling and analyzing real-world evidence, FDA/CDER assessed potential serious risks of using gadolinium-based contrast agents for MRI during pregnancy.
Although clinical trials provide crucial evidence that a drug is safe and effective, drug toxicities sometimes become evident only after marketed use, when rare but serious adverse effects may be reported. To help ensure drugs are safe for patients, FDA continually monitors approved drugs under real-world conditions, and the Agency may update drug labeling or even withdraw a product from the market based on new findings. Much of the evidence pertaining to real-world experience with a drug comes from spontaneous reports to FDA’s Adverse Event Reporting System (FAERS), but CDER may also proactively address safety concerns by analyzing the data in patient medical records and information collected in the course of routine health care.
Safety concerns for gadolinium-based contrast agents
Gadolinium-based contrast agents (GBCAs) are intravenous (or IV) drugs that are given to an estimated 30-45% of patients undergoing magnetic resonance imaging (MRI) to provide detailed images. Because of its toxicity, gadolinium is complexed with a molecule called a chelator (figure 1) when used as a contrast agent. The chelator reduces toxicity by minimizing unbound gadolinium, which makes gadolinium less likely to interact with tissues in the body before the kidneys eliminate the contrast agent. However, chelators do not completely prevent patient exposure to gadolinium. Gadolinium complexed with other molecules can be found in very small amounts in various organs and tissues of patients who receive GBCAs. No long-term clinical consequences of gadolinium exposure in healthy individuals have been identified. As a precaution, FDA required medication guides for GBCAs. These guides outline what is known about the toxicities of these products and indicate that “people who get many doses of gadolinium medicines, women who are pregnant and young children may be at increased risk from gadolinium staying in the body.”
From clinical studies and initial marketing experience, adverse reactions attributed to GBCAs were described as short-term events and often categorized as allergic. However, in 2006, nephrogenic systemic fibrosis (NSF), a potentially fatal disease that occurs mainly in patients with chronic kidney disease or acute kidney failure who had undergone GBCA-MRI, was described. Currently, researchers believe that retention of gadolinium due to impaired renal function or other precipitating risk factors such as inflammation can lead to NSF, which is characterized by thickening and hardening (fibrosis) of the skin, subcutaneous tissues, and sometimes underlying skeletal muscle. There may also be serious damage to other organs, including the lungs and the heart. Boxed warnings were added to product labeling to restrict GBCA use in patients with severe renal impairment, and the number of reported NSF cases reported annually has decreased.
The scientific challenge
In 2016, a study based on health care data in Ontario, Canada, reported that GBCA-MRI exposure during pregnancy was associated with greater risk for fetal or neonatal death and rheumatological, inflammatory or infiltrative skin conditions. Subsequently, an FDA-led study identified one gadolinium-exposure in-utero for every 860 pregnancies (0.12% of all pregnancies). The majority of exposures were during the first few weeks of pregnancy, a time when individuals may not know they are pregnant. FDA-approved labeling for all GBCAs was updated to state that GBCA administration should be considered during pregnancy only if imaging was essential and should not be delayed. However, for pregnant women and the clinicians who treated them, serious uncertainties remained when deciding whether a GBCA-MRI was appropriate. Limitations of the Canadian study of GBCAs in pregnant women included insufficient sample size to support a statistical comparison of contrast MRI vs non-contrast MRI, and inadequate control for the reason MRI was administered (due to missing information).
To address this research question, FDA/CDER conducted a study in collaboration with researchers at the University of Florida using data from Medicaid insurance claims and linked medical records for more than 11 million pregnancies [3]. Pregnant women whose medical history indicated that their fetuses might be at greater risk for adverse effects (for example, those whose MRI was for pelvic examination) were excluded. The final cohort of patients consisted of 782 women who had received a GBCA-MRI while pregnant and 5,209 who were exposed to MRI without GBCA while pregnant. The primary endpoint was fetal death or infant death shortly after birth (information for the latter outcome was obtained from the National Death Index or Medicaid records). A secondary endpoint was admission to the neonatal intensive care unit within seven days of birth. The investigators used a statistical technique called propensity score matching to adjust for potential confounding patient variables, such as comorbidities, pregnancy characteristics, ultrasound history, and prenatal vitamin use.
Study results and practice implications
After adjustment for potential confounding variables, the study did not identify a difference in the risk of fetal or neonatal death associated with exposure to GBCA-MRI. The risk for the patients who received a GBCA appeared similar to those who received MRI without contrast (RR = 0.73, Table 1). Similar results were obtained for the secondary outcome measured in the study, i.e., GBCAs did not appear to increase the risk that newborns who had been exposed to GBCAs were admitted to a neonatal intensive care unit.
Adjusted risks of fetal or neonatal death or NICU admission with GBCA
| Outcome | Outcomes/pregnancies | Risk ratio (95% CI) |
|---|
| | MRI with GBCA | MRI, no GBCA | |
| Fetal or neonatal death | 11/782 (1.4%) | 73/5209 (1.4%) | 0.73 (0.34, 1.55) |
| NICU admission | 59/771 (7.7%) | 452/5136 (8.8%) | 1.03 (0.76, 1.39) |
Reassuringly, the risk estimates did not change substantially in sensitivity analyses. The study authors concluded that gadolinium use during pregnancy should be limited and used in accordance with professional society guidelines. They also pointed out that, because there is gadolinium retention in various tissues, the impact of exposure on subacute and chronic adverse outcomes in infants remains unknown and needs further study.
This study also highlights the critical importance of the medical data that are collected during patient care for monitoring the safety of medical products. CDER is working with interested parties to make sure that these data become more standardized, interoperable, transparent and are easily linked to other kinds of data in support of studies like the one described here.