WASHINGTON - Pregnant people who are vaccinated against the coronavirus are nearly twice as likely to get COVID-19 as those who are not pregnant, according to a new study that offers the broadest evidence to date of the odds of infections among vaccinated patients with different medical circumstances.

The analysis, based on medical records of nearly 14 million U.S. patients since coronavirus immunization became available, found that pregnant people who are vaccinated have the greatest risk of developing COVID among a dozen medical states, including being an organ transplant recipient and having cancer.

The findings come on top of research showing that people who are pregnant or gave birth recently and became infected are especially prone to getting seriously ill from COVID-19. And COVID has been found to increase the risk of pregnancy complications, such as premature births.

The Centers for Disease Control and Prevention has been urging people to get coronavirus shots before or during pregnancy, seeking to dispel fear - widespread in some communities, without scientific basis - that those vaccinations could be harmful. As of March, nearly 70% of people who were pregnant have been vaccinated before or during their pregnancy, according to federal data, though disparities persist among racial and ethnic groups.

The new study goes beyond what has previously been understood, suggesting that even pregnant people who are fully vaccinated tend to have less protection from the virus than many other patients with significant medical problems.

"If you are fully vaccinated, that's magnificent," said a lead author of the study, David R. Little, a physician who is a researcher at Epic, a Wisconsin company that maintains electronic patient records for nearly 1,000 hospitals and more than 20,000 clinics across the country. "But if you are fully vaccinated and become pregnant, you remain at higher risk of acquiring COVID."

Little said the findings buttress CDC recommendations that additional precautions against the virus should be taken during pregnancy, such as wearing masks and maintaining safe distances. He said the study also suggests that health-care workers should "be on the lookout" for symptoms and encourage testing to detect the virus early, when it is easier to treat.

The analysis was based on Epic medical records from 13.8 million patients between January 2021, when the first people in the United States were fully vaccinated and had enough time to develop immunity, and late January this year. Little and colleagues analyzed the risk from 12 comorbidities throughout that period. The study included the delta and omicron variant surges but did not differentiate the rate of breakthrough infections during those waves or other times.

The researchers measured the risk by analyzing the records of pairs of fully vaccinated patients from the same part of the country. In each pair, one patient had the condition that was being measured, and the other did not. The patients were not matched by age, and the pregnant people could have been matched in the analysis with a man or a woman.

The analysis found that the 110,000 pregnant individuals included in the study were 90% more likely to have been infected with coronavirus than the same number of people who were not pregnant. The next-highest risk - 80% greater - was among organ transplant recipients. The elevated risk among those two groups was higher than among patients with compromised immune systems, who had 60% greater odds of coronavirus infection.

The fact pregnancy appears to pose a greater risk than having a weakened immune system is striking, because public health officials have warned that being immunocompromised can render coronavirus vaccines significantly less effective. Federal regulators allowed people with immunity troubles to get a fourth shot substantially before guidance shifted this week to permit the extra vaccine doses for Americans who are 50 or older.

The study found that several conditions pose only a slightly greater risk of vaccinated people experiencing infections compared with people without those conditions. They include kidney, liver and blood disorders. Patients with lung diseases had a slightly higher risk - 30% greater than patients without those diseases. On the other hand, cardiovascular diseases appear to create no added risk, and patients with cancers had slightly lower odds of breakthrough cases than those who are cancer-free.

The findings do not explain the reason behind the risk levels. Denise Jamieson, a specialist in infectious diseases during pregnancy, called the high risk of infection among vaccinated pregnant people an "interesting and intriguing finding."

Jamieson, chair of gynecology and obstetrics at Emory University School of Medicine in Atlanta, said the finding could reflect that such people are more likely to get tested, because they are receiving prenatal care, putting them in unusually frequent contact with a health-care provider. Or it could suggest that people of childbearing age tend to be in jobs, such as teaching or nursing, that put them at uncommonly high risk of exposure to the virus. At the same time, Jamieson said, the apparently high risk of a breakthrough case during pregnancy could be a result of the way the study matched its pairs of patients.

Or, Jamieson said, the study could actually reveal that pregnant individuals are more susceptible to becoming infected with the coronavirus, even when vaccinated.

"It's definitely interesting," Jamieson said. "This study asks this question but doesn't answer it."