135 million adults with diabetes for 1995,
Murray and Lopez estimated 118 million
cases at all ages for 1990. McCarty and
Zimmet estimated 110 million cases at
all ages for 1994. Most recently, Amos et al. in a revision of the earlier McCarty
and Zimmet report—estimated 118 million
cases at all ages for 1995. Comparing the
1995 and 2000 country estimates of Amos
et al. and the present study, in most cases
there is good agreement, although there are
a small number of important differences that
bear further examination.
McCarty and Zimmet, and subsequently
Amos et al., also attempted to make
projections of diabetes frequency into the
future. However, they did not incorporate
projected changes in patterns of urbanization
directly into their models, as was done
in the present study. Instead, they hypothesized
that in general, populations would
move toward progressively higher prevalences
of diabetes in the future. They first
examined all available prevalence rates for an
ethnic group, and then for each country,
they selected the “most appropriate” higher
rate based on predicted future gross national
product and urbanization of each country.
The higher age-specific prevalence estimates
were then applied to a projected national age
distribution. Using this procedure, Amos et
al. estimated a world total of 147 million
cases for the year 2000 (as compared with
our estimate of 154 million) and 221 million
cases for the year 2010. Thus, the projections
of Amos et al. yield an average annual
increment to the world’s diabetic population
of 5.8 million people for the years
1995–1999 and of 7.4 million people for the
years 2000–2010. Our apparently more
c o n s e rvative approach yields an average
annual growth of the world’s diabetic population
of 3.8 million people for the years
1995–1999 (starting from a higher 1995
baseline figure) and of 5.8 million people for
the years 2000–2025. These alternatives ,
essentially the result of adopting “pessimistic”
or “optimistic” scenarios, probably
form useful boundaries for the likely burden
of diabetes during the first years of the 21st
century. Thus, they may be viewed as comp
lementary to one another.
The failure to include subjects 20
years of age in the present study should not
be taken as ignoring the importance and
severity of diabetes in childhood and adolescence.
It is simply a re flection of the fact
that their inclusion would have had little
effect on the numerical estimates. Of the 118
million cases estimated by Murray and
Lopez for 1990, only 214,000, or 0.2%,
were 15 years of age. This suggests that
frequency of diabetes in adults is a close and
efficient proxy for global frequency of diabetes
at all ages.
The division of countries into “developed”
and “developing” groups is somewhat
arbitrary and is unduly simplistic,
given that many developing countries have
complex infrastructures and considerable
economic potential. However, it does serve
to differentiate those countries that have
been industrialized for some time from the
newly emerging economies and those yet to
experience such changes. Clearly, there are
close links between socioeconomic transition
and epidemiological transition.
The much greater predicted number of
middle-aged (45–64 years) than elderly
( 65 years) people with diabetes in the
developing countries is important. Such
subjects will have to endure the condition
during some of the most productive years