Compared to never using sleeping pills, even using no more than 18 a year is tied to a more than threefold increased risk of death, according to researchers in the US who saw this result after controlling for every possible factor they could think of that might influence it. They also found a more than fourfold higher risk of death and a significant increase in cancer cases among regular pill users.

The findings are stark news for the growing number of people who rely on sleeping pills to get a good night's rest, especially as the results showed the link was the same for the newer, more popular sleeping pills such as zolpidem (Ambien) and temazepam (Restoril).

First author Dr Daniel F. Kripke, of the Viterbi Family Sleep Center at Scripps Health in San Diego, California, told the media:

"What our study shows is that sleeping pills are hazardous to your health and might cause death by contributing to the occurrence of cancer, heart disease and other ailments."

Kripke and colleagues write about their investigation in a paper published in the open-access online journal BMJ Open on 27 February.

Their work expands on previously published research linking higher mortality with use of sleeping pills.

However, Kripke said it is the first to show that eight of the most commonly used sleeping pills or hypnotic drugs, are linked to increased risk of death and cancer, including the popularly prescribed zolpidem (known as the brand Ambien) and temazepam (Restoril). These newer drugs were thought to be safer because their action is not as long-lasting as that of the older hypnotics.

Between 6 and 10% of adult Americans took a hypnotic drug for poor sleep in 2010.

This sector of the US pharmaceutical industry grew by 23% between 2006 and 2010, to an annual sales level of about $2 billion.

For their study, Kripke and colleagues examined data on nearly 40,000 patients of average age 54 cared for by a large integrated health system in the northeastern United States.

The data came from an electronic medical record that had been in place for over ten years. Participants included 10,529 patients who received hypnotic prescriptions, and 23,676 matched controls who received no hypnotic prescriptions. All were followed for an average of 2.5 years, from early 2002 to early 2007.

In their analysis, where they looked for links between sleeping pill intake, death by any cause, and cancer, compared to not taking sleeping pills, the researchers adjusted for the usual factors like age, gender, ethnicity, marital status, body mass index, smoking, alcohol use, and also took into account prior cancer and a large number of comorbidities, that is other illnesses and medical conditions that might influence the result.

The analysis split the participants into as many as 116 groups, which exactly matched cases and controls by 12 classes of comorbidity.

The results showed that, as expected, patients prescribed any hypnotic had "substantially elevated hazards of dying compared to those prescribed no hypnotics." write the authors.

They also found a dose-response effect, in that for those patients prescribed between 0.4 and 18 doses of hypnotics a year, the hazard ratio HR (95% confidence interval CI) was 3.60 (2.92 to 4.44); for those prescribed between 18 and 132 a year, it was 4.43 (3.67 to 5.36); and for those on more than 132 doses a year, it was 5.32 (4.50 to 6.30).

When they did separate analyses for the common hypnotics, including zolpidem, temazepam, eszopiclone, zaleplon, other benzodiazepines, barbiturates and sedative antihistamines, they found elevated HRs for each of these as well.

They also found a higher incidence of cancer cases among the patients in the upper third of hypnotic use (HR 1.35; 95% CI 1.18 to 1.55).

"Results were robust within groups suffering each comorbidity, indicating that the death and cancer hazards associated with hypnotic drugs were not attributable to pre-existing disease," write the authors, who conclude that:

"Receiving hypnotic prescriptions was associated with greater than threefold increased hazards of death even when prescribed <18 pills/year."

They write that the association was the same, even when they separately analyzed the commonly used types of sleeping pills, including the newer shorter acting drugs.

Also, when they took into account selective prescription for patients in poor health, this did not explain the higher risk of death.

Second author Dr Robert D. Langer, of the Jackson Hole Center for Preventive Medicine in Jackson, Wyoming, said:

"We tried every practical strategy to make these associations go away, thinking that they could be due to use by people with more health problems, but no matter what we did the associations with higher mortality held."

But third author, Dr Lawrence E. Kline, medical director of the Viterbi Family Sleep Center, said even though they tried to take as much into account as they could, we should note that the study is based on observational data (and not a clinical trial, for example), so "it's still possible that other factors explain the associations".

"We hope our work will spur additional research in this area using information from other populations," he urged.

Kline said he also hopes the study will nevertheless prompt doctors to consider alterantives to sleeping pills when they treat their patients.

He said at the Viterbi Family Sleep Center the clinicians use cognitive therapy to help patients understand more about sleep. For example, insomnia sufferers may not need the commonly recommended eight hours sleep every night.

Teaching good sleep habits and relaxation is another possible alternative, as is taking advantage of the body's natural rhythms, which follow the rising and setting of the sun.

"Understanding how to use the circadian rhythm is a very powerful tool that doesn't require a prescription," said Kline.

Sometimes sleep disorder stems from depression and emotional problems. Kripke said in such cases, doctors should treat those underlying causes and avoid prescribing sleeping pills that could make things worse.

Written by Catharine Paddock PhD