Are sleeping pills harmful in the long term?

Benzodiazepines, the classic sleeping pills, have a proven favourable balance between benefits and risks when used short-term for a maximum of two to four weeks. Once that window has passed, that balance has not been established. A well-known problem is that short-term use can easily grow into long-term use: in a large Japanese cohort study of nearly 4,000 new users, 1 in 10 people were still taking the drugs after twelve months. Older age, concurrent use of antidepressants, and a higher starting dose were the strongest predictors of that long-term use and, with it, of an increased risk of dependence.

Physical dependence with long-term use of benzodiazepines is a well-supported risk. A distinction exists here between low-dose dependence that arises from the treatment itself, also known as iatrogenic dependence, and high-dose misuse. Both problems have been present for fifty years, but according to the available literature they are underestimated or ignored by many prescribers and official bodies.

In the area of thinking and reaction, long-term use of benzodiazepine sleeping pills shows measurable deficits in attention, reaction speed, and higher brain functions such as planning and decision-making. These deficits persist, even in people who have been using the drugs for years. In an on-road driving test, users of benzodiazepine hypnotics drove with significantly less stability than people not taking this medication, with a significant increase in lateral lane deviation. Notably, with use of more than three years the driving effect in the road test appeared to diminish, presumably due to tolerance, yet the neurocognitive deficits remained fully present. A smaller driving simulator study found no demonstrable reduction in simulator driving performance in long-term users, but the groups were small enough for this to be considered limited evidence.

There is still insufficient research on possible lasting changes in the brain from long-term benzodiazepine use to draw firm conclusions. This is acknowledged as a serious research question, but has not yet been answered.

Newer sleeping pills, the so-called z-drugs such as eszopiclone and the pyrazolopyrimidines, represent a step forward compared with classic benzodiazepines. Eszopiclone shows measurable sleep improvements compared with placebo: on average twelve minutes faster time to sleep onset, seventeen minutes less time awake during the night, and twenty-eight minutes more total sleep per night, based on moderate-quality evidence. Side effects do occur: an unpleasant taste (18% more often than with placebo), dry mouth, daytime drowsiness, and dizziness. Extra caution is warranted in older adults with cognitive or motor vulnerability. The broader group of newer agents appears to produce less tolerance and fewer withdrawal symptoms than the older benzodiazepines, but that advantage has only been described in a narrative literature review, not demonstrated in direct comparative studies.