Three studies published Wednesday by the Centers for Disease Control and Prevention show that protection against the coronavirus given by vaccines declined in the midsummer months when the more contagious delta variant rose to dominance in the United States.

At the same time, protection against hospitalization was strong for weeks after vaccination, indicating the shots will generate immune fighters that stave off the worst effects of the virus and its current variations.

Data from these studies persuaded the Biden administration to develop a plan for additional doses to bolster the immune systems of people vaccinated months earlier. The Biden administration will begin offering coronavirus booster shots to fully vaccinated adults who received the Pfizer-BioNTech and Moderna shots the week of Sept. 20, top health officials announced Wednesday, after concluding that a third shot is needed to fight off waning immunity.

“Examining numerous cohorts through the end of July and early August, three points are now very clear,” CDC director Rochelle Walensky said at a White House covid-19 news briefing Wednesday. “First, vaccine-induced protection against SARS-CoV-2 infection begins to decrease over time. Second, vaccine effectiveness against severe disease, hospitalization and death remains relatively high. And third, vaccine effectiveness is generally decreased against the delta variant.”

President Biden said on Aug. 18 that after Sept. 20, vaccinated Americans can get booster shots against coronavirus eight months after their second injection. (The Washington Post)

The trio of reports, published Wednesday in the Morbidity and Mortality Weekly Report, the CDC’s scientific digest, also reinforce the idea that vaccines alone will be unable to lift the nation out of the pandemic.

Masks and other precautions should be part of “a layered approach centered on vaccination,” wrote researchers from the New York State Department of Health and the University at Albany School of Public Health in their study of vaccine effectiveness across New York state.

All three reports measure vaccine effectiveness, which compares the rates of infection or hospitalization among vaccinated people with the rates among people who had not been vaccinated. Twenty percent of new infections and 15 percent of hospitalizations from covid-19, the disease caused by the virus, were among vaccinated people.

Until now, evaluations of vaccine effectiveness amid delta largely relied on observations from outside the United States. A recent New England Journal of Medicine study concluded the Pfizer vaccine was 88 percent effective against infections that caused symptoms in England.

Others, such as a study in Israel, found larger declines in protection against infection. One U.S. report that has not yet gone through peer review, collecting data from Mayo Clinic Health System facilities in five states, found a drop in the Pfizer-BioNTech vaccine’s effectiveness against delta infections to 42 percent. The other mRNA vaccine, made by Moderna, was 76 percent effective.

The new study from New York is the first to assess vaccine protection against coronavirus infection across the entirety of a U.S. state amid delta. The study authors found a modest drop in effectiveness: It descended from 92 percent in May to 80 percent in late July.

The second of the three studies found effectiveness against infection declined for nursing home residents after delta emerged. It dropped from 75 percent in March through May to 53 percent in June and July. Vaccination for visitors and staff is crucial, the study authors wrote, and “additional doses of COVID-19 vaccine might be considered for nursing home and long-term care facility residents.”

The third report, an analysis of patients at 21 hospitals in 18 states, found sustained protection against hospitalization. Effectiveness was steady at 86 percent, even in the midsummer months when delta outcompeted other variants of concern. For adults who do not have compromised immune systems, that effectiveness stood at 90 percent.

Many factors influence vaccine effectiveness and the changes the scientists observed in New York cannot be attributed to delta with certitude, they noted.

If vaccinated people behave in riskier ways, such as not wearing masks in crowded areas, that may influence vaccine effectiveness. So might waning immune protections.

To conduct their study, the researchers in New York linked multiple health reporting systems across the state. These included immunization registries, the statewide collection of coronavirus laboratory test results and the system that surveys New York’s inpatient facilities daily. Those databases allowed the study authors to connect vaccine status to every new case and hospitalization reported to the state from May 3 to July 25.

“The New York state data gives us a nice look at how we can link data together when you have comprehensive reporting across a number of systems,” said Robert A. Bednarczyk, an epidemiologist at Emory University Rollins School of Public Health, who has worked with the study authors in the past but was not involved with this research.

By the end of the study period, 66 percent of New Yorkers 18 and older were vaccinated. Vaccine effectiveness against hospitalization remained constant, above 90 percent. And of the more than 48,000 new infections from late spring into summer, 9,675 were in vaccinated people, or about 1 in 5 cases.

But Bednarczyk said breakthrough cases such as these do not mean the vaccines are failing. “The vaccine is doing what it’s supposed to do. It’s priming our immune system,” he said.

Immunized people may still get infected, because the vaccines aren’t perfect. But it is possible immune fighters will sweep the virus out the door much more quickly in a vaccinated person, Bednarczyk said, citing a not-yet-peer-reviewed paper from researchers in Singapore. In that report, vaccinated patients more swiftly defeated an infection compared with those who weren’t.

Though the results in New York may not easily translate to other communities. Maria Sundaram, an infectious-disease epidemiologist at the University of Toronto Dalla Lana School of Public Health, said it is difficult to make comparisons partly because this is an examination of an entire state, not a model that accounts for uncertainties in a population sample.

The change in vaccine effectiveness over time “to about 10 percent lower, I would take with a grain of salt,” Sundaram said, because there may be uncertainty from mismatches between databases or reporting lags.

Studies such as these show that, as valuable as coronavirus vaccines are, they have limits.

“As we’re releasing the brakes on these other non-pharmaceutical interventions” — [meaning masks and other precautions] — “we may see more cases,” Sundaram said. “Vaccines are very, very helpful but they’re not the end-all, be-all of covid-19 prevention.”

The booster shots that will become available in September are designed to be given eight months after people have received their second dose of the Pfizer or Moderna vaccine.

Surgeon General Vivek H. Murthy explained at Wednesday’s White House covid briefing how experts settled on the eight-month time frame, citing data showing that about six months after vaccination, mild to moderate infections started increasing.

Murthy said that while protection against hospitalization and death remained strong at that point, “our anticipation is that if the trajectory that we are seeing continues, that we will likely see in the future an increase in breakthrough hospitalizations and breakthrough deaths. And that’s why we use our judgment to see when to make a determination when that point may be. And that’s how we came to the eight-month mark.”

People who received Johnson & Johnson vaccines may need boosters, too, Murthy said, adding that officials will have more information about additional Johnson & Johnson shots in coming weeks.

The administration announcement drew mixed reaction, with some experts praising the plan and others saying it was premature and could subvert the process for considering vaccines and boosters.

Paul A. Offit, director of the Vaccine Education Center at Children’s Hospital of Philadelphia, questioned the premise of the decision — that a decline in immunity against mild or moderate illness would likely be followed be a decrease in protection against severe disease and hospitalization. He said he believes protection against serious disease might last a few years.

He also said he doubted that boosters would have a major impact on the pandemic, adding that such a change can be brought about only by getting more Americans vaccinated.

In a later news media call with vaccine experts, Offit said he would be more comfortable with the administration plan if data first had been submitted and scrutinized by the CDC and Food and Drug Administration and their outside advisers. “This seems to be a declaration without the vetting you would have liked to have seen,” Offit said.

Standard practice is for the FDA to assess the safety and efficacy of vaccines and for the CDC’s outside advisers, the Advisory Committee on Immunization Practices, to recommend what vaccines Americans should get, and when — not the surgeon general or the National Institutes of Health. After the advisory panel makes its recommendation, the CDC director decides whether to accept it.

Jesse Goodman, former chief scientist at the FDA and professor of medicine and infectious diseases at Georgetown, said he suspects that boosters will be needed and added, “being prepared for a booster campaign makes sense.” But he said it was important for the administration to avoid putting “the cart before the horse.”

Goodman added that it was repeatedly said in the covid-19 briefing that normal processes would prevail, “but I think it puts that process under a date-certain kind of pressure.”

Administration officials stressed in the covid-19 briefing and in subsequent interviews that a booster-shot campaign will not go forward without the go-ahead from the FDA and the CDC advisory committee.