Months ago, U.S. government regulators called for COVID-19 vaccines to have an efficacy of at least 50%, meaning the drugs would reduce the risk of illness by half.

Manufacturers Pfizer and Moderna designed their vaccine trials on the assumption that the drugs would work a bit better than that, preventing disease in 60% of those exposed to the coronavirus. And during the summer, infectious-disease chief Dr. Anthony S. Fauci was even more optimistic, hoping for 70% or 75% efficacy.

So how is it that the Pfizer and Moderna vaccines seem to prevent disease more than 90% of the time?

The answer — with Pfizer applying Friday for emergency approval to distribute the drugs and Moderna said to be doing so soon — involves the “spike” protein: the little protrusions that form a “corona” on the surface of each virus particle.

The details require a bit of explanation, but here is the short version: Though it may seem that we’ve been living in a pandemic haze for ages, remember that the virus was unknown to science less than a year ago. And despite the extraordinary amount of research that took place in the months since then, no scientists can say how well a drug works in humans before they try it.

Much has been written about how the vaccines from Moderna and Pfizer, which collaborated with German firm BioNTech SE, make use of the genetic molecule RNA. Though that technology has been studied as a possible vaccine “platform” for years, no such drug has yet made it to market.

But the end goal is the same as with traditional vaccines, said Michele Kutzler, a vaccine researcher and associate professor at the Drexel University College of Medicine. All such drugs work by exposing the recipient to a microbe in such a way that they do not get sick, enabling the immune system to develop customized antibodies and other defenses should it ever encounter a real infection.

Of the dozens of COVID-19 vaccines in development worldwide, many, including the Pfizer and Moderna products, seek to provoke this response by giving the immune system a harmless taste of the coronavirus’ “spike” proteins.

Early in the year, laboratory studies suggested that the spike might be the right agent to include in a vaccine, as the virus uses it to penetrate human cells. If the immune system were given a sneak preview of this attack, it could learn to make antibodies that would bind to the spike, almost like gumming up a key so it no longer fits in a lock.

The early results from Pfizer and Moderna suggest that relying on the spike was the right call, and that other spike-based vaccines, whether made with newer platforms or traditional technology, could work as well, Kutzler said.

“I think all of the different types of vaccines focusing on the spike have the potential to be protective with the same mechanism,” she said.

Note, however, that she said “potential.” It is one thing to demonstrate that a vaccine has efficacy — defined as its ability to prevent illness in the controlled confines of a clinical trial, in which people who get the vaccine are compared with those who get a placebo. A vaccine’s effectiveness, on the other hand, is how well it works in the real world.

For a variety of reasons, the vaccines may not prevent disease quite as well when administered to millions in the general public, said Craig Shapiro, a pediatric infectious-disease physician at Nemours/Alfred I. duPont Hospital for Children in Wilmington.

For example, the Pfizer and Moderna products both require two injections, spaced several weeks apart, and likely will not work as well in people who fail to return for the second dose, he said.

The vaccines’ effectiveness also could suffer if they are not handled properly. The Pfizer product, for instance, must be stored in special freezers at an average temperature of negative 94 degrees Fahrenheit. But once doses are delivered to a pharmacy or other administration site, they can be kept for up to five days at 18 to 28 degrees, within range of standard freezers, the company says.

And it remains unclear how well the drugs will prevent disease in the elderly and other at-risk populations, though early signs are encouraging. This week, scientists reported that another COVID-19 vaccine, developed at the University of Oxford and made by British firm AstraZeneca, seems to trigger a protective immune response in older people. But those preliminary results, published in the Lancet, describe only the presence of “neutralizing” antibodies in the recipients’ bloodstreams, not whether the drug prevents actual illness.

Then there is the question of safety. But so far, so good: No serious side effects have been identified in the tens of thousands who have received either the Pfizer or Moderna vaccines.

“This is all very promising, absolutely,” Shapiro said. “But we still have a little ways to go.”

Still another issue is how well the drugs reduce the spread of disease. It is possible that in some cases, the vaccines will prevent COVID-19 symptoms but not infection, meaning that such people could still unwittingly transmit the virus to others. Researchers are keeping a watchful eye to see if that happens.

One final puzzle: How long will protection from vaccines last? The answer depends on a variety of factors, said Kutzler, the Drexel scientist. The level of protective antibodies in a person’s bloodstream may decline after a few months, but that might be OK, provided the immune system retains a supply of “memory” cells that can make more antibodies in a hurry, she said.

Another arrow in the immune system’s quiver is something called a T-cell, which might also help with longer-term protection, she said. Human beings can produce these cells both in response to an infection and to a vaccine. One type of “helper” T-cell assists in coaxing the production of antibodies, while another type, called killer T-cells, destroys infected cells directly.

Kutzler predicts that a particularly strong T-cell response may result from another type of high-tech vaccine that she studies, in which DNA, rather than RNA, prompts human cells to make the spike protein. Such a vaccine is in development at Inovio Pharmaceuticals, in Plymouth Meeting, though the final “phase 3″ tests of that drug have not yet begun, pending questions from the U.S. Food and Drug Administration.

In her lab, Kutzler is working on an adjuvant — a type of additive that could boost the staying power of DNA vaccines.

“Durability is going to be a huge question,” she said. “Our goal is to have it last more than a year.”

But in the short term, with COVID-19 surging across the country, Kutzler and Shapiro agree that preventing 90% of new cases seems like a win.