Tuesday February 28 2012
The study cannot prove sleeping pills directly raise death risk
Sleeping pills are linked to a raised risk of death, according to several high-profile news reports in today’s newspapers. Several front page stories have covered the link, reporting a four-fold increased risk of death among users of the drugs, medically known as ‘hypnotics’.
The news is based on the results of a large US study that compared the medical records of more than 10,000 people prescribed sleeping pills and 23,000 similar people who had never been prescribed them. It followed them for an average of 2.5 years and found that people prescribed hypnotics, even at very low doses, were more likely to die than those not prescribed hypnotics. It also found that people prescribed high doses (more than 132 pills a year) were more likely to develop any cancer.
Although a significant association was found, the most important thing to note is that this study cannot tell us that the drugs themselves are the direct cause of higher rates of death and cancer. This is because both the drugs and the risk of death could be associated with other factors, such as lifestyle, alcohol or smoking, which the researchers cannot rule out. While the researchers did make adjustments to account for behaviour and other health problems, these may not fully compensate for their influence.
This research raises an important issue that requires further research, but this cohort study alone cannot show that hypnotics are responsible for the increased mortality or cancer. It should be noted that UK guidelines recommend the use of hypnotics only under specific circumstances and only then in short courses.
Where did the story come from?
The study was carried out by researchers from the Scripps Clinic Viterbi Family Sleep Center and the Jackson Hole Center for Preventive Medicine, USA. The study received funding from the Geisinger Center for Health Research and the Scripps Clinic Academic Fund. The study was published in the peer-reviewed medical journal BMJ Open.
This story was given prominent coverage in a number of different publications. Most of the coverage was accurate although some sources incorrectly implied that the use of hypnotic medicines was found directly to cause death. In actuality, the research found a link between the two, but does not support a direct relationship. A common factor, such as underlying health problems, may make individuals both more likely to use sleeping pills and more likely to die. Few media sources explained that this study cannot prove direct causation.
What kind of research was this?
This was a large US cohort study that aimed to examine death and cancer rates associated with using a class of sleeping drugs called hypnotics. Death and cancer rates in patients using the drugs were compared to rates in a matched group of patients who had never used them. To do this the research used electronic medical records to identify patients from these two groups and look at any cancer diagnoses and deaths over a period of up to five years. Commonly prescribed hypnotics include temazepam, flurazepam and loprazolam.
Cohort studies can show association, but they cannot demonstrate a direct link. This study cannot show that hypnotic medications directly cause cancer or death. To do this, a randomised controlled trial (RCT) would have to be performed. However, as the researchers point out, an RCT addressing this question would be unethical, as an association between hypnotics and cancer and death has already been suggested from the results of previous studies.
What did the research involve?
The researchers examined electronic medical records between 2002 and 2007 for 10,529 patients who had received hypnotic prescriptions and 23,676 matched patients who had not received a hypnotic prescription. All were outpatients in the Geisinger Health System in the USA. Patients receiving hypnotics were those who had been prescribed a hypnotic drug at least once by their family doctor and who were above the age of 18. Two controls matched for sex, age and smoking history were selected for each person who had been prescribed a hypnotic. The records were examined to see if patients subsequently died or were diagnosed with cancer.
The researchers then calculated hazard ratios (HRs) for death and cancer associated with the prescription of hypnotic medication. A hazard ratio expresses the ratio of event rates between two groups of people. When calculating HR values they adjusted their analyses for the following potential confounders: age, gender, smoking, body mass index, ethnicity, marital status, alcohol use and prior cancer.
What were the basic results?
The average age of all people in this cohort was 54 years. Over an average follow-up period of 2.5 years there were 295 deaths among people who did not use hypnotics (1.2%) and 638 among those who did (6.1%).
Patients prescribed any hypnotic had increased risk of dying compared with patients who had never been prescribed hypnotics. The risk of dying increased in line with the amount of hypnotics they were prescribed:
- Patients who were prescribed 1-18 pills of any hypnotic a year were 3.6 times more likely to die than matched patients not using the drugs (HR 3.60, 95% CI 2.92 to 4.44).
- Patients who were prescribed 18-132 pills a year were almost 4.5 times more likely to die (HR 4.43 95% CI 3.67 to 5.36).
- Patients who were prescribed more than 132 pills a year were 5.3 times more likely to die (HR 5.32, 95% CI 4.50 to 6.30).
The researchers then examined eight different types of hypnotic separately (zolpidem, temazepam, eszopiclone, zaleplon, triazolam, flurazepam, barbiturates and antihistamines). Each of these hypnotics was associated with an increased risk of death.
Heavy use of any hypnotic (more than 132 pills a year) was also significantly associated with cancer (HR 1.35, 95% CI 1.18 to 1.55). When different types of cancer were analysed separately, it was found that the risk associated with hypnotic usage was greater than the risk of lymphoma, lung, colon and prostate cancers posed by current smoking.
The researchers estimate that hypnotics may have been associated with 320,000 to 507,000 excess deaths in the USA.
How did the researchers interpret the results?
The researchers concluded that in this study “receiving hypnotic prescriptions was associated with greater than threefold increased hazards of death even when prescribed <18 pills/year”.
This cohort study has found that receiving a prescription of any hypnotic (sleeping pill) was associated with increased mortality. For those prescribed high doses, cancer incidence also increased. However, due to the study design, the researchers have identified several limitations:
- Although the researchers did make adjustments to account for many variations including age, gender, smoking, body mass index, ethnicity, marital status, alcohol use, prior cancer, and a number of other concurrent conditions, it is difficult to adjust for these completely. Their influence may not have been fully accounted for and, in addition, other unknown factors may influence the relationship.
- There is the possibility that people who do and do not take hypnotics differ in other medical factors that may account for their different mortality and cancer incidence (for example, chronic diseases). Of note is the fact that the study was unable to control for depression, anxiety and other emotional factors as these diagnoses are confidential in America. As such, mental health illnesses may be an important confounding factor.
- The study was done on the basis of prescriptions. The researchers did not monitor how many prescriptions were filled, if the medication was taken or if the medication was taken correctly.
- Cohort studies can only show association and, therefore, this study cannot demonstrate that hypnotics are directly responsible for the increased mortality shown. A randomised trial would be required for this. However, the researchers make the point that, as cohort studies have demonstrated a hazard of hypnotic medications, it may not be ethical to perform a randomised controlled trial.
This study raises an important issue and further research into the safety of these medications is required. However, it is also important to view the use of hypnotics in the context of UK guidelines, which may differ from the way they are used in the US. The British National Formulary recommends that:
- hypnotics should not be prescribed indiscriminately
- hypnotics should be reserved for short courses in the acutely distressed
- hypnotics should be used to alleviate acute conditions after their cause has been established
- hypnotics should be avoided in the elderly due to their greater risk of falls if they become confused
There are also several circumstances and conditions that make the use of hypnotics inadvisable, which are medically known as ‘contraindications’. These contraindications are taken into account when considering the use of hypnotics in UK patients.
Analysis by Bazian