The headline on the story in the HuffingtonPost immediately grabbed my attention: “Prescription Sleeping Pills May Increase Risk of Death Four-Fold.” (See FastLinks.) Based on an article published on the online journal BMJ Open, the piece made some sensational claims:
- “People who take certain prescription sleeping pills even once in a while may be up to five times more susceptible to early death.”
- The researchers found that those taking the lowest doses — four to 18 pills a year — were 3.6 times more likely to die early than those who took no sleeping pills. Those taking 18 to 132 pills a year had a 4.4 times higher risk of dying early.
- The authors concluded that hypnotics such as zolpidem (Ambien), temazepam (Restoril), and eszopiclone (Lunesta) “may have been associated with 320,000 to 507,000 excess deaths in the U.S. alone” in 2010. (Emphasis added.)
Holy moly! If sedatives and hypnotics taken at a dose of 1.5 pills per month can quadruple one's risk of death, they must be among the most toxic substances known to man. If sleeping pills kill at least four times as many people in the United States as gunshot wounds and automobile accidents combined, there is an unprecedented and unappreciated ongoing public health catastrophe.
But it seemed to me that these claims had no face validity; they were so extreme that I just did not believe them. Before dismissing them out of hand, I downloaded the BMJ Open paper so that I could analyze it in detail. (“Hypnotics' Association with Mortality or Cancer: A Matched Cohort Study,” BMJ Open 2012;2:e000850; see FastLinks.)
The authors used data from the Geisinger Health System in Pennsylvania. Using electronic health records, they retrieved information regarding outpatient encounters and prescriptions for primary care patients 18 and older from January 2000 through September 2006. They compared 10,531 patients who had been prescribed sedatives and hypnotics with 24,793 matched controls. (Important point: The authors had no way of telling whether the patients actually took the sleeping medicine or how frequently they took it. They were able only to track prescriptions written.)
These subjects were followed for an average of 2.5 years, with mortality data obtained from the Social Security Death Index, which the authors blithely claimed were “thought to be accurate” without referencing anything to establish that as a fact. The authors also attempted to control for possible confounders, and calculated hazard ratios for death for those who were and were not prescribed sedatives and hypnotic agents.
Their first finding was the most stunning: “Compared with non-users, patients prescribed 1–18 pills of any hypnotic each year had an HR [hazard ratio] for death of 3.60 [95% CI 2.92 to 4.44].” This is even more astounding than HuffPo suggested. It seems you don't necessarily even have to take 1–18 hypnotic pills a year to increase your risk of death. The mere fact that you are prescribed 18 or fewer pills per year means it would be prudent to start planning your funeral ASAP.
No need to get into detail about the other mortality data. Let's just say they suggest that the more sleeping pills you are prescribed, the more likely it is that you'll be meeting your maker before your time.
To be fair, the authors concede a major limitation of this study, that the data may show an association between being prescribed a sedative or hypnotic and increased risk of death but in no way established that taking sleeping pills caused death. The authors clearly believe, however, a causative link exists. One would have to join the authors in accepting two rather dubious arguments to endorse such a link:
- The study was successful in controlling for all possible confounding variables. The authors claim that they controlled for up to 116 comorbidity combinations. Call me skeptical, but I just don't believe this is possible, especially given the retrospective nature of the study. Data based on a retrospective review of electronic health records are certainly incomplete and not always accurate. Certainly sleep disturbances can be caused by all kinds of existing health problems, some of which may be undiagnosed at the time a sedative or hypnotic is prescribed.
- Insomnia does not increase mortality. The most important confounding variable in a study such as this is the presence or absence of baseline insomnia, and it is not possible to evaluate how pre-existing poor quality sleep affected the mortality results without a prospective study including a group of insomniacs not treated with sedatives or hypnotics. The authors try to minimize this problem by stating that “several large studies have reported that insomnia is not a significant mortality risk factor.”
The papers cited to support this claim have problems of their own. One of the studies referenced (Sleep 2005;28:965) analyzed data collected in a survey conducted for an entirely different purpose — to evaluate the risk of atherosclerosis. The methods used in that study were criticized in an editorial comment published in the same journal.
Another paper the authors reference (Arch Gen Psychiatry 2002;59:131) was written by the same lead author as the latest BMJ Open study. Again, the methods and conclusions of that study were challenged by a commentary in the journal Sleep Medicine. (2002;3:373.)
The study design had other flaws. The authors' inability to establish the causes of death in the patients they tracked made it impossible to establish a biologically convincing association between sleeping pills and increased death rates. Their speculation about possible causative mechanisms is also unconvincing. But, on the most basic level, extraordinary claims require extraordinary evidence. The contention that being prescribed 1–18 sleeping pills in a year drastically increases one's mortality risk is an extraordinary claim. What these authors provide is far from extraordinary evidence.
- Read the Huffington Post article at http://huff.to/JNhlRN.
- The BMJ Open article is available as free full text at http://bit.ly/IkUkc0.
- Visit Dr. Gussow's blog at www.thepoisonreview.com.
- Read all of Dr. Gussow's past columns in the EM-News.comarchive.
- Comments about this article? Write to EMN at email@example.com.
Click and Connect!Access the links in EMN by reading this issue on our website or in our iPad app, both available onwww.EM-News.com.