Background Reliable information on causes of death is essential to the
development of national and international health policies for prevent
ion and control of disease and injury. Medically certified information
is available for less than 30% of the estimated 50.5 million deaths t
hat occur each year worldwide. However, other data sources can be used
to develop cause-of-death estimates for populations. To be useful, es
timates must be internally consistent, plausible, and reflect epidemio
logical characteristics suggested by community-level data, The Global
Burden of Disease Study (GBD) used various data sources and made corre
ctions for miscoding of important diseases (eg, ischaemic heart diseas
e) to estimate worldwide and regional cause-of-death patterns in 1990
for 14 age-sex groups in eight regions, for 107 causes. Methods Prelim
inary estimates were developed with available vital-registration data,
sample-registration data for India and China, and small-scale populat
ion-study data sources, Registration data were corrected for miscoding
, and Lorenz-curve analysis was used to estimate cause-of-death patter
ns in areas without registration. Preliminary estimates were modified
to reflect the epidemiology of selected diseases and injuries. Final e
stimates were checked to ensure that numbers of deaths in specific age
-sex groups did not exceed estimates suggested by independent demograp
hic methods. Findings 98% of all deaths in children younger than 15 ye
ars are in the developing world. 83% and 59% of deaths at 15-59 and 70
years, respectively, are in the developing world. The probability of
death between birth and 15 years ranges from 22.0% in sub-Saharan Afri
ca to 1.1% in the established market economies. Probabilities of death
between 15 and 60 years range from 7.2% for women in established mark
et economies to 39.1% for men in sub-Saharan Africa. The probability o
f a man or woman dying from a non-communicable disease is higher in su
b-Saharan Africa and other developing regions than in established mark
et economies. Worldwide in 1990, communicable, maternal, perinatal, an
d nutritional disorders accounted for 17.2 million deaths, non-communi
cable diseases for 28.1 million deaths and injuries for 5.1 million de
aths. The leading causes of death in 1990 were ischaemic heart disease
(6.3 million deaths), cerebrovascular accidents (4.4 million deaths),
lower respiratory infections (4.3 diseases (2.9 million), perinatal m
illion), chronic obstructive pulmonary disease (2.2 million), tubercul
osis (2.0 million), measles (1.1 million), road-traffic accidents (1.0
million), and lung cancer (0.9 million). Interpretation Five of the t
en leading killers are communicable, perinatal, and nutritional disord
ers largely affecting children. Non-communicable diseases are, however
, already major public health challenges in all regions, Injuries, whi
ch account for 10% of global mortality, are often ignored as a major c
ause of death and may require innovative strategies to reduce their to
ll. The estimates by cause have wide CIs, but provide a foundation for
a more informed debate on public-health priorities.