Researchers today call for official death registries in all countries to record whether the dead person was a smoker, in a research article published in The Lancet.
New analyses of nearly half a million death records in South Africa – the first, and so far the only, country to record smoking on death registration forms – show that the death rate from tobacco is more than twice as great in the coloured (mixed ancestry) as in the white population.
The study, which is the first large-scale analysis of mortality from smoking in any African country, found that in the coloured population smoking causes one in four of all deaths in middle-aged men and one in six of all deaths in middle-aged women.
At ages 35 – 64 years, the excess risk of death among smokers was greater in the coloured than in the white population (men 14.2% vs 7.6%, women 11.0% vs 7.7%), as was the proportion who had smoked (men 68% vs 47%, women 46% vs 28% at these ages). These findings imply that, for both men and women, the death rate from smoking is more than twice as great in the coloured as in the white population.
The black (African) population already accounts for more than half of all deaths from smoking in South Africa, due to its larger size. At present, the death rate from smoking is not yet as high in the black as in the white population, but the researchers warn that this is likely to change if the large numbers of young black adults who now smoke continue to do so. The system of recording smoking on death notification forms will, however, automatically monitor any future changes in the death rate from smoking.
South Africa modified its national death notification form in 1998 to ask a simple yes/no question about whether the dead person had been a smoker five years earlier. An international team of researchers has analysed the answers about smoking on the death notification forms of nearly half a million (481 640) adults in South Africa who died during 1999 – 2007. The excess of smokers among people who died of diseases such as lung cancer, chronic lung disease, tuberculosis, heart attack, and stroke indicated for each disease the proportion of the deaths that had been caused by smoking.
According to Professor Debbie Bradshaw, lead author of the study, "There is already a high death rate from smoking in the mixed-ancestry coloured population of South Africa, and there will be major increases in tobacco-attributed mortality in many other African populations where young adults now smoke, unless there is widespread cessation."
Professor Sir Richard Peto, another study author, adds that "Death registries around the world should routinely ask whether the dead person was a smoker. This would help assess national death rates from smoking and would help countries discover whether deaths from smoking are increasing or decreasing. There will be hundreds of millions of tobacco deaths this century if current smoking patterns continue."
Professor Freddy Sitas, another study author, who in 1994 first proposed including the question about smoking in South African death registries, adds, "Our results show that in 1999 – 2007, smoking caused many deaths from cancer and heart disease, but the main way it killed, particularly in the black population, was by increasing mortality from TB and other lung diseases. All countries have heterogeneous populations, and need to know how disease-specific mortality from smoking varies with cultural background and socio-economic status."
Writing in a linked Comment, Professor Lionel Opie, at the University of Cape Town, South Africa, adds that "This remarkable Article assesses the lethal effects of smoking from the answers to one simple—but key—yes or no question on the South African death registration form: smoker or not, 5 years before death? This vital question has led to new understanding in South Africa, and could do likewise elsewhere. Even in countries with sophisticated health systems, it should become part of the routine death registration process."
Explore further: Stroke risk similar among men and women smokers worldwide
www.thelancet.com/journals/lan … (13)61610-4/abstract