People who take sleeping pills often hope that sleeping pills will increase their sleep enough to make them more energetic in the day, and they may hope that sleeping pills will improve their long-term health. Unfortunately, nothing could be farther from the truth!
Because life and death are often our primary concern, I would like to first to discuss the darkest aspect of sleeping pills. I think that taking sleeping pills is like risking suicide. Later, I will discuss how sleeping pills fail to help us in the day.
Sleeping pill usage is associated with increased mortality
It is now over 26 years that I have been working to assess the risks of sleeping pills. I have learned that sleeping pills are associated with significantly increased mortality. This means that people who take sleeping pills die sooner than people who do not use sleeping pills.
I first became interested when I saw the work of Dr. E. Cuyler Hammond at the American Cancer Society. He was a leader of the Cancer Prevention Study I (CPSI). Their work had shown that people who reported long sleep had higher mortality, as well as (to a lesser extent) those with very short sleep. To understand further what this might mean, I went to visit The American Cancer Society, starting a collaboration with has extended over all of these years.
In the CPSI (this was the first of two large studies), in 1959-1960, the American Cancer Society asked its volunteers to give health questionnaires to participants whom the volunteers would be able to contact 6 years later. Because Cancer Society volunteers selected relatives and people whom they knew well, these volunteers accomplished the remarkable feat of collecting questionnaires from over 1 million Americans and then determining six years later (in over 98%) whether the participants had survived. As most people know, the main finding of this study was that people who smoked cigarettes had much higher rates of mortality from lung cancer and heart disease. The study had also asked about many other aspects of people's health which might cause cancer. Included were simple questions about reported insomnia, hours of sleep, and use of sleeping pills.
In 1964, Dr. Hammond had reported that participants in CPSI who said that they slept more than 7 hours or less than 7 hours had higher mortality than those who slept 7 hours . I wondered if this could be related to sleeping pills and worked with Dr. Hammond and Mr. Lawrence Garfinkel at the American Cancer Society to examine which CPSI participants had died after 6 years. We found that 50% more of those who said that they "often" took sleeping pills had died, compared to participants of the same age, sex, and reported health status who "never" took sleeping pills . This increased risk for those who reported taking sleeping pills was not influenced by how long participants reported sleeping. Whether or not participants reported insomnia did not explain the risk. Incidentally, one third of those who said that they often took sleeping pills also said that they never had insomnia, a matter I will discuss below.
We found that reported insomnia did not predict mortality independent of sleep duration. For example, people who reported sleeping 7 hours who said that they frequently had insomnia had similar mortality to people who said that they slept 7 hours and they had no insomnia. A finding in this study (and many others) was that people who complained of insomnia often reported sleeping as long (or longer) than people who reported no insomnia complaint. Similar paradoxes are found when sleep is measured with EEG (brain wave) recordings. Although people who complain of insomnia do sleep a bit less, on average, than people who report no insomnia, insomnia complaints are not closely related to sleeping less than average. In fact, people who sleep more than 8 hours a night report more insomnia than those who say they sleep 7-8 hours.
There were several limitations in the early CPSI study. The questionnaire had not asked the participants whether the "sleeping pills" they took were what we call prescription hypnotics, or whether they might be tranquilizers, antidepressants, or over-the-counter drugs of various kinds. Prescription hypnotics are those drugs which the U.S. Food and Drug Administration has approved for treatment to promote sleep. Because of computer limitations, there was only a very limited way that we could control the comparison of sleeping pill users with nonusers for other health factors which might cause deaths. Finally, by the time our study was published, people had largely switched from the barbiturate prescription sleeping pills (which were well known to cause thousands of overdose deaths each year) to other sleeping pills such as flurazepam (Dalmane), which were thought to be much safer in overdoses.
To reexamine these risks, the American Cancer Society agreed to ask new questions about sleeping pills to participants in a new study, called The Cancer Prevention Study II or CPSII. In 1982, American Cancer Society volunteers gave health questionnaires to 1.1 million new participants, and the survival of these people was ascertained in 1988. Many years were needed for the final data for this huge number of people to be put entered into computers, assembled into computer tapes, and loaned us by the American Cancer Society. Additional years were needed before the new Pentium computers became available to us, so that we could analyze this enormous amount of information. Even a Pentium needed literally months of constant work to complete all the analyses.
In the new study, we again found that people who said that they used sleeping pills had significantly higher mortality . In this study, we had indicated that we were interested mainly in prescription sleeping pills as distinct from tranquilizers or over-the-counter drugs. Because a reported frequency like "often" may be inexact, we had asked participants to estimate the number of sleeping pills which they took each month. When people were matched for age, sex, race, and education, and a total of 32 health risk factors, those who reported taking sleeping pills 30 or more times per month had 25% more mortality than those who said that they took no sleeping pills . However, we did not determine which particular sleeping pills were associated with this risk. The smaller risk of taking sleeping pills just a few times per month was 10-15% increased mortality, compared to those who took no sleeping pills. Sleeping pills appeared unsafe in any amount.
- 25% increased mortality among those taking sleeping pills nightly
- 10-15% increased mortality among those taking sleeping pills occasionally
To provide a perspective on this mortality risk, we noted that the risk of taking sleeping pills 30 or more times per month was not much less than the risk of smoking 1 pack of cigarettes a day, when the analyses were done in a similar manner.
If the association of increased mortality with sleeping pills represents sleeping-pill-caused deaths, then usage of sleeping pills may have shortened the lives of the people taking sleeping pills nightly by several years. Note that the "If" in this statement is important. The fact that usage of sleeping pills was associated with deaths does not prove that the sleeping pills were the cause, since possibly other factors (e.g., a painful cancer) might both cause people to die early and to use sleeping pills, without the sleeping pills having any relationship to the mechanism of death.
With the new computers, in CPSII we were able to examine in much more detail than in CPSI whether the risk of using sleeping pills could be explained by other factors. With these methods, we controlled for 32 different questionnaire responses which might have been related to sleeping pill use, for example, whether the participants said they had heart disease or cancer. In these cases, if people used sleeping pills because they had cancer, and it is the cancer (not the sleeping pills) which caused their increased mortality, the control method would remove overestimation of mortality which might not be due to sleeping pills but rather to cancer. On the other hand, if sleeping pills (or cigarettes) cause heart disease which in turn leads to deaths, controlling for heart disease would cause underestimation of the mortality which sleeping pills or cigarettes cause. Even with as much control for other factors as we thought possible in this study, the association of sleeping pill usage with increased mortality could not be explained by chance or by the other factors which we measured. This supports the likelihood that the association is causal.
One cause of death was especially increased. Among men, those who took sleeping pills 30 times a month had 7 times the risk of suicide! Women who took sleeping pills 30 times a month had 2 times the risk of suicide. Nevertheless, the suicides were only a small portion of deaths associated with using sleeping pills. Deaths from other common causes such as heart disease, cancer, and stroke were also increased among sleeping pill users.
To summarize, our new CPSII results in a second million participants confirmed that sleeping pill use is associated with excess mortality. It is probably impossible to design an epidemiologic study which would prove that sleeping pills cause the extra mortality associated with their use. The only way to be certain if the sleeping pills are directly causing the risk would be to randomly offer volunteers either sleeping pills or placebo pills for long-term trials. It is true that such studies would be quite hard to do and expensive, and that ethical concerns would have to be overcome. Nevertheless, people whose loved ones take sleeping pills have an urgent need to know if these pills are safe, which we will not know until random clinical trials are done. Until studies give us more clarity, my best guess is that taking sleeping pills shorten people's lives by increasing the risk of suicide and other causes of death. This is why I say that taking sleeping pills for a percentage of people may amount to doctor-assisted suicide.
The Dark Side of Sleeping
Pills, By Daniel F. Kripke, M.D.