Epidemiological studies have found that few cases of panic disorder ar
ise for the first time after the age of 40 years, and there is a stead
y decline in the prevalence of existing cases in the latter half of li
fe. The authors review these epidemiological findings and explore vari
ous hypotheses that might explain the decreased frequency of panic dis
order in old age. There is no available evidence to suggest that metho
dological factors have led to an underestimation of the prevalence of
this disorder in older rather than younger age groups. However, there
is evidence that disorder-associated mortality and age-related changes
in brain neurochemistry may contribute to the decreased frequency of
this illness in later life. A cohort effect also should be considered,
although currently there are no data available to support or refute t
his idea.