Background
Although all-cause mortality in Sri Lanka decreased significantly from 1950 to 1970, subsequent declines have been more modest with divergent trends by age and sex. This study investigates these trends through cause of death analysis for 1950–2006 in adults aged 15–64 years.

Methods
Deaths were obtained from the World Health Organisation (WHO) mortality database for 1950 to 2003, and the Department of Census and Statistics Sri Lanka for 1992–95 and 2004–06 where WHO data was unavailable. Adult deaths were categorised by age (15–34 and 35–64 years) and sex into: infectious diseases; external-causes; circulatory diseases; cancers; digestive diseases; respiratory diseases; pregnancy-related; ill-defined; and other-causes. Cause-specific mortality rates were directly age-standardised to the 2001 Sri Lankan Census population.

Results
Mortality declined in females aged 15–34 years by 85% over 1950–2006, predominantly due to sharp declines in infectious disease and pregnancy-related mortality over 1950–70. Among males aged 15–34 years the mortality decline was less at 47%, due to a rise in external-cause mortality during 1970–2000. In females aged 35–64 years mortality declined by 67% over 1950–2006, predominantly due to a sharp decline in infectious disease, ill-defined and other cause mortality over 1950–70. Among males aged 35–64 years, decline in mortality is evident to 1960 (19%) from decline in infectious disease mortality, followed by increased mortality from circulatory diseases and external cause mortality, despite continued decline in infectious disease mortality. All-cause mortality in males 35–64 years has stagnated since 1970, with fluctuating increases. Circulatory diseases were the leading cause of death among adults 35–64 years in 2002–06, with the male rate almost three times higher than females.

Conclusions
Significant disparities are demonstrated in Sri Lankan cause-specific adult mortality by sex and age group for 1950–2006. Female mortality progressively declined while male mortality demonstrated periods of increase and stagnation. Among males aged 15–34 years this coincides with periods of civil conflict over 1970–2000. Among males aged 35–64 years the increased mortality from non-communicable disease and external causes are the main reasons for stagnation in all-cause mortality since the 1970’s.