Editorials

Learning from asthma deathsNeed to relate deaths to prevalence, severity, and treatment in different populations

BMJ 1997; 314 doi: http://dx.doi.org/10.1136/bmj.314.7092.1427 (Published 17 May 1997) Cite this as: BMJ 1997;314:1427

Need to relate deaths to prevalence, severity, and treatment in different populations

There seems little doubt that the prevalence of asthma in children and adolescents is increasing all over the world,1 and several epidemics of deaths in the past 30 years have aroused vigorous debate about the possible causes. Awareness of the importance of decreasing death rates in the 1980s led to the development of asthma management plans or guidelines, first at national level, such as those for Australia2 and England,3 then at international level, and now globally.4 The aim of these plans was to achieve better care. One of the signs of this is decreasing death rates, especially in young people, in whom deaths are mostly avoidable (as shown by the falling rates in the presence of increasing prevalence of the disease). Since it is widely stated in the lay press that death rates from asthma are rising worldwide, the paper by Campbell et al in this week's BMJ (p 1439), showing that the mortality from asthma is decreasing in those under 65 years in England and Wales, is timely.5

Caution is needed in interpreting data from death certificates, especially in elderly people,6 as fashions for diagnosing airway diseases change. However, in people under 35 death certificates are thought to more accurately reflect asthma deaths, and records over the past 100 years show periodic swings in mortality in this age group. The recent pattern has been of an increase in mortality between 1975 and 1983 and subsequently a decrease. This has been seen in England and Wales, Australia (where mortality peaked in 1986),7 Germany, and most dramatically New Zealand (peaking in 1979 with a high of 4.1/100 000 in 5-34 year olds).8 In Japan and the United States mortality is continuing to increase.9 In the United States the absolute rates are much lower than elsewhere, although the reasons for this, and the continuing rise, are not clear.

It is hard to interpret changes in mortality without knowing about changes in prevalence and severity of asthma in the different age groups within the population. However, the fact that the death rates in young people are decreasing as the prevalence of asthma in young people in England is increasing10 suggests either that the new asthmatic patients have mild disease or that treatment has improved. Overall it seems likely that improvement in long term management has occurred, although direct evidence for this from studies of severity of the disease in the community is lacking.

Campbell et al's data on seasonality are also of interest.5 They show that there is an increase in death rates in young people in the summer. It is unlikely that this is due to viral infections, and in fact viral infections rarely cause death from asthma in young people. It seems more likely that the deaths are due to sudden severe attacks and, as the authors suggest, are the result of exposure to allergen. The authors are correct in drawing attention to this possibility, which is overlooked by many pulmonary physicians. Death from asthma is a rare event, and individual doctors are unlikely to notice that there are more deaths during the summer. Many groups have studied the risk factors for death, but there has not been much interest in seasonality. This is an important risk in addition to the well described risks of psychiatric caseness and non-compliance,11 12 which are year round phenomena.

Campbell et al did not examine sex differences, which may be relevant. Deaths have been increasing in young males in Japan and in young black people in the United States. In Australia deaths in children below the age of 15 years are predominantly among boys, whereas after the age of 15 most deaths occur in young women.7 The authors also did not comment on the fact that there was a big fall in death rates between 1983 and 1984, not only for asthma but for other respiratory deaths. No explanation is offered for this. It is not clear from the data whether 1983 was a peak year or whether something happened to reduce the number of deaths drastically in 1984.

What can be learnt from studying asthma mortality? It seems that in order to interpret the figures we need to know prevalence, severity, treatment regimens, and allergic status of the population. Furthermore, we need this information from different populations. Already there is some indication that the high death rate among black people in America's inner cities results from difficulties with providing appropriate care rather than from a higher incidence of the disease. A study of the deaths in Saskatchewan showed that a small dose of inhaled corticosteroids protected against death,13 and the data from Campbell et al support this theory by showing that there has been a proportionate increase in the use of inhaled steroids as the death rates have fallen.5 The debate about the place of inhaled β agonists in treatment could be helped by more data that relate deaths to prevalence, severity, and treatment. Eventually it may be possible to set up registers of severe asthmatic patients (such as those who have ever needed treatment at a hospital) so that more information is available to compare populations. However, most asthma is mild, fluctuates with time, and is hard to define. As a result, most patients and many doctors are unaware that it can be fatal and that those at risk need to be identified, the severity of their disease documented, and appropriate treatment given until the disease is controlled. It is important therefore to keep examining the trends in mortality, to report them, and to continue to emphasise the fact that this is a fatal disease.

Campbell et al suggest that part of the fall in mortality might be due to the increased use of inhaled corticosteroids.5 It is not clear what proportion of the drugs prescribed for asthma inhaled corticosteroids should represent because there is a large group of people with mild disease who do not require steroids and most asthmatic patients need more than one drug. Nevertheless, criteria for the use of inhaled corticosteroids are becoming clearer, and the Saskatchewan study showed clearly that those who took more than 400 μg of beclomethasone were protected from death.13