Sanfra Anastine had surgery at age 42 and couldn’t speak for about 12 hours afterward. The next time she was operated on she was 56 and it took three months for her speech to return. Now 61, Anastine says that she doesn’t have difficulty forming words anymore but is still more forgetful than before her second surgery. She’s afraid of what will happen if she has to go under anesthesia again.

It is common to hear that an elderly patient “just isn’t the same” after surgery, says Roderic Eckenhoff, an anesthesiologist at the University of Pennsylvania.* Many people wonder if anesthesia—which is designed to make people groggy and temporarily rob them of their mental faculties—is to blame. Elderly patients often exhibit a condition called postoperative cognitive decline in which they experience lapses in memory and attention, but it usually does not last for more than a few weeks.

Most evidence suggests that receiving general anesthesia during the course of surgery does not increase the likelihood of developing lasting dementia. Yet it is clear something is going on: Recent experiments on animals and human cells show that anesthesia can increase the buildup of the proteins thought to underlie Alzheimer’s disease, especially in high doses.

So what do we know about the relationship between general anesthesia, which is typically inhaled and completely knocks people unconscious, and dementia, a permanent, debilitating condition?

Despite being a well-established component of modern medicine, much of how anesthesia works is a mystery. Evidence indicates that drug molecules bind to sites on the surface of neurons and deactivate different proteins important in a wide range of cognitive functions, including sleep, attention, learning and memory. Most recently research has suggested that in addition to targeting specific areas involved in sleep and arousal general anesthesia works by knocking out the neural networks that enable communication between brain regions.

Because anesthesia affects so many diverse brain processes and areas, some researchers worry that it may have unforeseen consequences. The molecules in anesthesia “can trigger other mechanisms that have nothing to do with anesthesia itself,” says Maria Lioudyno, a neuroscientist at the University of California, Irvine. Including “processes that may be linked to neurodegeneration.”

At the cellular level, Lioudyno has found that anesthesia can set off a chemical cascade triggering the release of microglia, immune cells normally deployed to fight infections in the brain. When microglia are activated for long periods of time they can inflame brain tissue, which is thought to contribute to the cognitive problems associated with Alzheimer’s.

Recent research on animals has also shown that anesthesia can induce brain changes like those thought to underlie dementia. In 2004 and 2007 studies, for instance, Eckenhoff and his colleagues found that exposing mice to inhaled anesthetics, especially at high doses, accelerated the buildup and toxicity of amyloid beta, a protein implicated in the development of Alzheimer’s. Other studies have shown a similar effect with tau, another Alzheimer’s-linked protein.

These studies are very preliminary, Eckenhoff cautions. Humans are vastly more complex than mice and just because something shows up in a petri dish does not mean it leads to the development of Alzheimer’s. Although changes are showing up on the cellular and tissue levels, “the effect on what we really care about—cognition, memory, the ability to learn—seems to be really minimal,” Eckenhoff says.

Furthermore, epidemiological evidence casts doubt on a link between receiving anesthesia and developing dementia. In a 2013 Mayo Clinic study doctors compared medical records of 900 people over 45 who had developed dementia with a similar group who did not develop the disorder and found that both received anesthesia at similar rates, making it unlikely to be a risk factor.

But although anesthesia does not appear to increase the risk of developing dementia, there is no denying that some people seem more deeply affected by it than others. One possible explanation is that its effects may be amplified in patients who are already genetically predisposed to dementia or have other risk factors. “Human beings come in many different sizes and have different preexisting conditions that may put them at higher risk,” says Robert Whittington,  professor of clinical anesthesiology at Columbia University Medical Center. Postoperative cognitive decline has been found to be especially prevalent in patients who have had cardiac surgery, for instance, as well as among people with diabetes and hypertension, conditions which have also been tied to Alzheimer’s.

It is difficult to untangle the effects of anesthesia from those caused by the operation itself, however. Surgery is a traumatic experience that is known to provoke inflammation. Eckenhoff believes neuroinflammation from surgery rather than anesthesia is the true culprit in cognitive decline, which can “interact with pathology that is sort of smoldering along in somebody with incipient Alzheimer’s disease” and accelerate it, he says. “We don’t think that anesthesia and surgery actually cause Alzheimer’s or cause dementia,” he adds. “We think that it interacts with individual vulnerabilities where if you’re already predisposed to getting something like this, this speeds it up.”

Scientists are working on ways to identify populations that might be more susceptible to dementia via biomarkers and other tests, and eventually hope to use that information to make surgery safer for them. This could potentially include smarter, biologically targeted anesthetics, along with drugs to counteract the stress involved in surgery. For instance, statins—commonly used to treat cardiovascular disease—have been found to reduce cognitive decline in mice when given before an operation.

For now, however, Eckenhoff recommends that susceptible people to avoid cosmetic or other elective surgical procedures.

*Correction (10/23/14): This sentence was edited after posting to correct Roderic Eckenhoff's first name.