Talk about sleepless nights.

Patients taking prescription sleep aids on a regular basis were nearly five times as likely as non-users to die over a period of two and a half years, according to a recent study. Even those prescribed fewer than 20 pills a year were at risk, the researchers found; heavy users also were more likely to develop cancer.

Unsurprisingly, the findings, published online in the journal BMJ, have caused a quite a stir. Americans filled some 60 million prescriptions for sleeping pills last year, up from 47 million in 2006, according to IMS Health, a health care services company. Panicked patients have been calling doctors’ offices seeking reassurance; some others simply quit the pills cold turkey.

Some experts were quick to point out the study’s shortcomings. The analysis did not prove that sleeping pills cause death, critics noted, only that there may be a correlation between the two. And while the authors suggested the sleeping pills were a factor in the deaths, those who use sleep aids tend as a group to be sicker than those who don’t use them. The deaths may simply be a reflection of poorer health.

Still, the findings underscore concern about the exploding use of sleeping pills. Experts say that many patients, especially the elderly, should exercise more caution when using sleep medications, including the non-benzodiazepine hypnotics so popular today, like zolpidem (brand name Ambien), eszopiclone (Lunesta) and zaleplon (Sonata).

“If someone comes to me on a sleeping pill, usually my tactic is to try to take them off it,” said Dr. Nancy A. Collop, president of the American Academy of Sleep Medicine and director of the Emory Sleep Center in Atlanta, who was an investigator in a clinical trial of Lunesta five years ago.

The non-benzodiazepine sedative hypnotics, on the market since the late 1980s, are believed to be safer and less likely to be abused than benzodiazepines or barbiturates. But many people take them for years, even though most are approved only for short-term use and generally their safety and effectiveness have not been evaluated beyond several weeks in clinical trials. (One exception is Lunesta, which was tested for up to six months.)

Some data suggest that the medications do not even do what they promise all that well, said Dr. Steven Woloshin, a professor of medicine at the Dartmouth Institute for Health Policy and Clinical Practice.

“The definition of insomnia is that you get less than six and a half hours of sleep, and it takes you 30 minutes or more to fall asleep,” said Dr. Woloshin, who is advocating for new labeling that spells out how well a drug actually works compared with a placebo. “But even when the drugs work better than placebos, and they don’t always, people still don’t fall asleep in less than 30 minutes, and they still don’t sleep much longer than six hours.”

And, the clinical trials show, some still feel groggy the next day, a side effect of the drugs.

Data from the largest trial reviewed by the Food and Drug Administration when it approved Sonata found the drug did not extend sleeping time, Dr. Woloshin pointed out; subjects on a 10-milligram dose and those taking a placebo both slept 6 hours 20 minutes on average. And those on the drug took 36 minutes on average to fall asleep during the first week of the trial, only 14 minutes less than the control group needed to fall asleep.

One in 20 said they felt sleepy the next day, and some reported memory problems.

Data from a large trial of Lunesta, also reviewed by the F.D.A., showed that subjects slept 37 minutes longer than a control group, but got only 6 hours 22 minutes of sleep, and it still took them 30 minutes to fall asleep; it took the subjects on placebo 15 minutes longer.

Even setting aside the controversial new study, the risks of popular sedative hypnotics can be significant. They can trigger severe allergic reactions in rare cases and may increase breathing problems; they can worsen depression, often a cause of sleeplessness.

They can contribute to confusion that affects judgment and behavior and have been implicated in falls that can be fatal for older adults. They may be involved in overdoses in combination with alcohol, and their use raises the risk of car accidents.

Patients who use hypnotics also have reported engaging in complex and bizarre behaviors while they are in a half-awake, half-asleep state, from raiding the fridge in the middle of the night to having sex, talking on the phone and even driving.

Drug company officials say the sedative hypnotics have been shown to be both safe and effective, though like other prescription sleeping pills, they may carry some risks.

The study published last month, by researchers from the Viterbi Family Sleep Center at Scripps Clinic in San Diego, was not the first to tie sleeping pills to an increased risk of death. A 2007 study in Norway of nearly 15,000 people found men were 1.5 times as likely to die and women 1.7 times as likely to die if they used sleeping pills. A 2010 Canadian study of more than 14,000 people found users were 1.36 times as likely to die as non-users.

A 2009 Swedish study that followed several thousand men and women for 20 years found that men who were regular users of hypnotics were 4.5 times as likely to die, while women were twice as likely to die.

In the latest study, researchers used health system medical records to compare survival rates of 10,529 patients who used hypnotic drugs with 23,671 people who used none during the same period. Non-users were a much healthier group, but the researchers were able to compare users and non-users with the same chronic conditions. They did not, however, have access to critical information about psychiatric diagnoses, nor did they know the causes of death.

“We think there are many things going on, not just one,” said Dr. Daniel F. Kripke, the paper’s lead author. He believes the drugs may worsen depression and may damage DNA, leading to cancer.

Representatives of drug companies making popular sleeping pills were critical of Dr. Kripke’s study. For one thing, they said, the observation period was too short to conclude that the drugs could contribute to the development of cancer. Sanofi-Aventis, maker of Ambien, issued a statement calling the new study “highly questionable.”

So what is an insomniac to do?

Nothing rash. Quitting sleeping pills abruptly can result in serious withdrawal symptoms and agitation, not to mention sleepless nights. Patients must taper off the medication over many weeks, experts say.

If sleeplessness is a chronic problem, talk to your physician. Ask for a thorough evaluation to determine if an underlying medical or psychiatric condition may be the cause; sleeplessness can also be a side effect of some medications. Consider asking for a referral to a doctor who specializes in sleep medicine.

Practice good “sleep hygiene” — habits that are conducive to sleeping well. Have a set bedtime and adopt a pre-sleep ritual, like reading. Do not nap during the day. Do not watch television or use the computer before going to sleep. Avoid exercise, nicotine, alcohol, caffeine and spicy food for several hours before going to sleep. Block out light and noise from the bedroom.

Some studies suggest that cognitive behavioral therapy can be effective in combating insomnia. If your insurance does not cover face-to-face counseling, there are online programs that have been found to be successful for both insomnia and anxiety.