Commonly prescribed sleeping pills lead to a fourfold increase in the chances of an early death, a study claims.
US experts warned the risks were present even for people who took less than 18 doses a year.
The paper, titled Hypnotics' Associations with Mortality or Cancer, published in the British Medical Journal Open, also found a significantly increased risk of cancer among those taking high doses.
The San Diego-based study analysed doctors tracing the survival of more than 10,500 people with a range of underlying conditions and an average age of 54 who were prescribed sleeping pills for an average of two-and-a-half years between 2002 and 2007.
Doctors compared the survival rates to more than 23,500 people with similar age, sex and health problems but who had not been prescribed any pills.
"Patients prescribed any hypnotic [sleeping pill] had substantially elevated hazards of dying compared to those prescribed no hypnotics," researchers said.
Although the actual number of deaths recorded was quite small, the authors calculated that there was a significant pattern in the results. It found that those patients with a prescription for sleeping pills had 4.6 times more chance of dying over an average observation period compared to non-users.
"Perhaps the mosts striking feature was that an increased hazard for death was present even in the lowest [third] of hypnotic use. Hypnotic drugs were associated with a 3.6-fold increased risk of dying for patients using less than or 18 hypnotic pills per year," the study said.
Those taking the most doses, more than or equal to 132 a year, were more than five times as likely to die as those prescribed none.
Drugs observed included benzodiazepines, such as temazepam, non-benzodiazepines, barbiturates and sedative antihistamines. The association was found for all drugs and every age group, but the greatest was among those aged 18 to 55.
There were 265 deaths within a group of 4,336 people taking zolpidem - a significantly greater percentage than the 295 deaths among 23,671 people who had not taken pills.
It was also found that those taking the highest number of doses were found to have a 35 percent greater risk of cancer, an association that was not explained by pre-eisting poor health.
Although the authors, doctors Daniel Kripke, Robert Langer and Lawrence Kline, believe their work backs up previous reseacth into the risks of sleeping pills, they point out that their studie does not guarantee cause and effect.
They do issue a stark warning, asking if it isn't time to reconsider the whether the use of sleeping pills is "sufficiently safe".
"Although the authors have not been able to prove that sleeping pills cause premature death, their analyses have ruled out a wide range of other possible causative factors," said BMJ Open editor-in-chief, Dr Trish Groves.
"These findings raise important concerns and questions about the safety of sedatives and sleeping pills."
In 2010, nearly one in 10 adults took a sleeping pill in the US alone.