Why We Sleep

particular examination of health relationships that exist with sleeping pills—good
or bad.
In  the  early  2000s,  insomnia  rates  ballooned  and  sleeping  pill  prescriptions
escalated dramatically. It also meant much more data was available. Kripke began
examining these large epidemiological databases. He wanted to explore whether
there was a relationship between sleeping pill use and altered disease or mortality
risk. There was. Time and again, the same message emerged from the analyses:
individuals  taking  sleeping  pills  were  significantly  more  likely  to  die  across  the
study periods (usually a handful of years) than those who were not, the reasons for
which we will soon discuss.
Often,  however,  it  was  tricky  to  conduct  a  well-matched  comparison  with
these early databases, as there were not enough participants or measured factors
that  he  could  control  for  to  really  tease  out  a  pure  sleeping  pill  effect.  By  2012,
however,  there  were.  Kripke  and  his  colleagues  set  up  a  well-controlled
comparison, examining more than 10,000 patients taking sleeping pills, the vast
majority of whom were taking zolpidem (brand name Ambien), though some were
taking  temazepam  (brand  name  Restoril).  He  contrasted  them  with  20,000  very
well  matched  individuals  of  similar  age,  race,  gender,  and  background,  but  who
were  not  taking  sleeping  pills.  In  addition,  Kripke  was  able  to  control  for  many
other factors that could inadvertently contribute to mortality, such as body mass
index, history of exercise, smoking, and drinking. He looked at the likelihood of
disease and death across a two-and-a-half-year window, shown in
Figure 15
.
V
,
VI
Those  taking  sleeping  pills  were  4.6  times  more  likely  to  die  over  this  short
two-and-a-half-year period than those who were not using sleeping pills. Kripke

---

further discovered that the risk of death scaled with the frequency of use. Those
individuals  classified  as  heavy  users,  defined  as  taking  more  than  132  pills  per
year, were 5.3 times more likely to die over the study period than matched control