Pfizer and Moderna shots are less effective in people with weak immune systems, a study finds.

Coronavirus vaccines were significantly less effective in protecting people with weakened immune systems than they were for other people, the Centers for Disease Control and Prevention reported on Tuesday, buttressing the agency’s call for immunocompromised adults to receive third or fourth doses of vaccines.

Two doses of either the Pfizer-BioNTech or Moderna vaccines were 77 percent effective against Covid-related hospitalization for immunocompromised people. That was a significant degree of protection, the agency said, but far lower than the shots’ benefit to people without immune deficiencies: In those people, the agency said, the vaccines were 90 percent effective against Covid hospitalizations.

The Moderna vaccine also offered more protection to people with weakened immune systems than did the Pfizer shot, mirroring results seen across American adults. And certain people with immune deficiencies — especially organ or stem cell transplant recipients, who often take drugs to suppress their immune systems and prevent rejection of the transplant — showed weaker responses to Covid vaccines than other categories of immunocompromised people did.

The study did not examine recipients of the Johnson & Johnson vaccine.

To help immunocompromised people mount a more aggressive immune response, the C.D.C. suggests that they be given three doses of Pfizer or Moderna vaccines, plus an additional booster shot six months after the third dose. In addition, the agency’s scientists wrote, they should take precautions like wearing masks, and be considered for treatments like monoclonal antibody therapy as early as possible after a Covid diagnosis.

The study released on Tuesday used a circuitous experimental design. The researchers examined roughly 20,000 immunocompromised adults and 70,000 people without immune deficiencies hospitalized this year with Covid-like illness. Of the immunocompromised patients in the study, 43 percent were fully vaccinated. Of the other participants, 53 percent were vaccinated.

The researchers then determined how many of those hospitalized patients were indeed infected with the coronavirus, and compared the odds of a positive test results between fully vaccinated and unvaccinated patients.

The immunocompromised patients included those with cancer, inflammatory disorders, organ or stem cell transplants and other immune deficiencies.

The study’s authors cautioned that there could have been cases in which patients were misclassified as immunocompromised, and that there could have been biases in which patients sought out coronavirus tests.