Intended for healthcare professionals

Letters

Trends in asthma mortality

BMJ 1997; 315 doi: https://doi.org/10.1136/bmj.315.7114.1012 (Published 18 October 1997) Cite this as: BMJ 1997;315:1012

Data on seasonality of deaths due to asthma were omitted from paper but editorial's author did not know

  1. M J Campbell, Professor of medical statisticsa,
  2. S T Holgate, Professor of immunopharmacologyb,
  3. S L Johnston, Senior lecturer in medicineb
  1. a School of Health and Related Research, University of Sheffield, Northern General Hospital, Sheffield S5 7AU
  2. b Department of Medicine, University of Southampton, Southampton General Hospital, Southampton SO16 6YD
  3. c National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA, USA 30341
  4. d West Cumberland Hospital, Whitehaven, Cumbria CA28 8JG
  5. e Public Health Research Unit, University of Glasgow, Glasgow G12 8RZ
  6. f Greater Glasgow Health Board, Glasgow G2 4JT
  7. g Department of Public Health Sciences, St Georges Hospital Medical School, London SW17 0RE

    Editor—Readers of Ann J Woolcock's editorial about our paper may have wondered at her comments on the seasonality of deaths due to asthma1 although data on this were not presented in our paper.2 This was because we had been asked to remove these results to shorten the paper and she had not been told. However, we wish to present them here. The dataset was the same as that referred to in our paper. Briefly, all deaths due to asthma (International Classification of Diseases code 493) in England and Wales from 1 January 1983 to 31 December 1995 were included. Deaths were divided into calendar months, an adjustment being made for the different lengths of the months. Age specific rates were calculated using the mid-period population (1988) and the age groups 0-4, 5-14, 15-44, 45-64, 65-74, 75-84, and 85. Analysis was by Poisson regression in STATA 5 (StataCorp, 1997). Seasonality was tested by including both sine and cosine terms with an annual period in the regression. Each age group, except for those who died aged under 5, showed significant seasonality. For those aged 45–64 the significance was marginal (2=8.4, df=2, P=0.015), but for all others the P value was <0.01. The phase of the seasonality was different for each age group (1). Deaths in the younger age groups peaked in the summer and deaths in the older age groups peaked in the winter. Deaths of people aged 5–14 peaked in August, and in those aged 15–44 the peak was slightly earlier, in July, but the rate in August was also high. Deaths in people aged 45–64 showed an intermediate pattern, with excess mortality from November to March and a separate peak in August. In those aged 65–74 there was a winter excess lasting from November to April but no summer peak.

    Figure1

    Death rates (per million population per year) by month of death and age group

    Age related seasonal trends offer clues to aetiology. It is widely accepted that atopy is the single most prominent risk factor for the development of asthma, and this is supported by strong epidemiological data linking exposure to environmental allergens to the chronic airway inflammation that characterises the disease. The peak in August in young people in England and Wales has been reported before.3 A possible explanation for the summer peak is exposure to outdoor aeroallergens. The peak pollen season is June and July, but certain fungal spores peak in August. It may take a month or two to build up sensitivity to pollen, but our results suggest that exposure to moulds is a more important factor than exposure to pollens for deaths due to asthma; this has also been suggested previously.4 Social factors such as compliance or difficulty in getting medical help during the August summer holiday may also play a part. Strong emphasis has been placed on allergens in the home, especially house dust mite, but exposure to these is likely to be year round, and peak exposure to house dust mite allergen is not seasonal.5 In the group aged 45 mortality is greater in the winter months. The most likely explanation is the presence of viral infections and diagnostic transfer from bronchitis (or chronic obstructive pulmonary disease) to asthma. There are strong epidemiological data linking viral infections with exacerbations of asthma, but adequate studies have not yet been carried out in exacerbations of chronic obstructive pulmonary disease.

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.

    Asthma mortality in United States has risen but is similar to that in England and Wales

    1. Carol A Johnson, Epidemiologistc,
    2. David M Mannino, Medical epidemiologistc,
    3. Annette Ashizawa, Epidemiologistc
    1. a School of Health and Related Research, University of Sheffield, Northern General Hospital, Sheffield S5 7AU
    2. b Department of Medicine, University of Southampton, Southampton General Hospital, Southampton SO16 6YD
    3. c National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA, USA 30341
    4. d West Cumberland Hospital, Whitehaven, Cumbria CA28 8JG
    5. e Public Health Research Unit, University of Glasgow, Glasgow G12 8RZ
    6. f Greater Glasgow Health Board, Glasgow G2 4JT
    7. g Department of Public Health Sciences, St Georges Hospital Medical School, London SW17 0RE

      Editor—M J Campbell and colleagues analyse the age specific trends in asthma mortality in England and Wales from 1983 to 1995.1 Their analysis indicates that there has been a downward trend in Britain's asthma mortality, especially among younger age groups. The authors suggest that this trend may be due to the increased use of prophylactic treatment.

      We analysed mortality files to characterise trends in mortality attributable to asthma in the United States during the same period. The results of this analysis differed from those presented by Campbell and colleagues. During the study period, asthma accounted for 54 455 deaths in the United States and the annual asthma mortality (adjusted for age to the 1980 American population)increased 39% (from 15.3/million to 21.3/million). Age specific death rates increased in every age stratum (2). By 1994, American age specific mortality from asthma among people younger than 65 was similar to reported rates in England and Wales, whereas among people aged 65 and older they were lower.

      Figure2

      Death rates from asthma (per million population) by age group from 1983 to 1994

      Asthma mortality in the United States has risen despite the increased use and sale of prophylactic drugs such as inhaled corticosteroids.2 Thus our results suggest that other factors, such as changing environmental exposures,3 recreational drug use,4 or socioeconomic factors,5 may be important in trends in asthma mortality over time. On the other hand, at the end of the study period the increasing age specific mortality in the United States was still similar to or lower than the decreasing rates reported for England and Wales, suggesting that there are other differences between these two study groups.

      Many gaps exist in our understanding of deaths related to asthma, which are thought to be largely preventable. Both of these studies highlight the need for improved surveillance of and education in asthma and the establishment of comprehensive prevention and intervention programmes.

      References

      1. 1.
      2. 2.
      3. 3.
      4. 4.
      5. 5.

      Death certification in asthma is inaccurate

      1. W T Berrill, Consultant physiciand
      1. a School of Health and Related Research, University of Sheffield, Northern General Hospital, Sheffield S5 7AU
      2. b Department of Medicine, University of Southampton, Southampton General Hospital, Southampton SO16 6YD
      3. c National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA, USA 30341
      4. d West Cumberland Hospital, Whitehaven, Cumbria CA28 8JG
      5. e Public Health Research Unit, University of Glasgow, Glasgow G12 8RZ
      6. f Greater Glasgow Health Board, Glasgow G2 4JT
      7. g Department of Public Health Sciences, St Georges Hospital Medical School, London SW17 0RE

        Editor—M J Campbell and colleagues analysed age specific trends in asthma mortality between 1983 and 1995, based on statistics derived from the Office of Population Censuses and Surveys.1 Their encouraging conclusion was that asthma mortality in England and Wales shows a downward trend. This conclusion may well be correct. However, there has been considerable uncertainty over the validity of statistics so derived,2 3 4 not because of any defect in the collection and recording of statistics by the Office of Population Censuses and Surveys (now the Office for National Statistics) but, in my view, because of three factors. The first of these is inexperience and uncertainty in the completion of death certificates, often by very junior hospital doctors insufficiently trained in death certification and inadequately guided in individual cases by senior medical staff. The second factor is overuse of the word asthma by both clinicians and patients, combined with failure to differentiate chronic obstructive pulmonary disease from genuine asthma. The final factor is the automated system with which the Office for National Statistics selects out the word asthma from enormously varied death certificates as being the primary cause of death. To test the last of these and gain some insight into the first two, I asked the Office for National Statistics to cooperate with me in looking at a small number of ambivalent death certificates that I had come across in the course of the Northern region's confidential asthma deaths inquiry (roughly 15% of such death certificates examined in which the word “asthma” appeared in part one of the certificate). The 1 shows details on the certificates, which related mostly to elderly patients. All of these deaths, I am assured, would have been taken into the statistics as deaths due to asthma despite what seems to be an unlikely and often conflicting combination of terms.

        Table 1

        Cause of death given on 15 death certificates; all deaths would have been coded as due to asthma

        View this table:

        From these death certificates, the likelihood that asthma was the primary cause of death seems unconvincing. I suggest, therefore, that one cannot draw conclusions on trends in asthma mortality from simply analysing such information unless individual cases are examined in more detail for accuracy of diagnosis.

        References

        1. 1.
        2. 2.
        3. 3.
        4. 4.

        Short term fluctuations may obscure more meaningful, longer term, changes

        1. Philip McLoone, Research fellowe,
        2. David S Morrison, Specialist registrar in public health medicinef
        1. a School of Health and Related Research, University of Sheffield, Northern General Hospital, Sheffield S5 7AU
        2. b Department of Medicine, University of Southampton, Southampton General Hospital, Southampton SO16 6YD
        3. c National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA, USA 30341
        4. d West Cumberland Hospital, Whitehaven, Cumbria CA28 8JG
        5. e Public Health Research Unit, University of Glasgow, Glasgow G12 8RZ
        6. f Greater Glasgow Health Board, Glasgow G2 4JT
        7. g Department of Public Health Sciences, St Georges Hospital Medical School, London SW17 0RE

          Editor—The significant downward trend in asthma mortality in England and Wales reported by M J Campbell and colleagues should be viewed from a longer term perspective.1 There are several precedents for doing so. When considering age specific annual mortality from asthma between 1951 and 1985 Alderson did not find any consistent trend but noted that fluctuations in rates could obscure more meaningful changes, such as the short lived increase in mortality during the 1960s.2 Burney, in a review of mortality from 1931 to 1985, stated that it “remained fairly constant,” although complex age, period, and cohort effects were found.3 In Scotland, trends in asthma mortality were reported to have been stable between 1970 and 1989.4 The 3 illustrates these data, updated to include the years 1990-5. Mortality is shown as age standardised three year moving averages for ages 5-44. This age range was chosen because of the difficulties in diagnosing asthma in small children and because of confusion with partially reversible chronic obstructive airways disease above middle age. Over the entire period these rates fluctuated without any consistent pattern, with relative rates in men and women changing throughout.

          Figure3

          Age standardised death rates from asthma per 100 000 population aged 5-44, 1974-95. Data are plotted as three year moving averages

          One might be tempted to argue that mortality fell among women during the late 1970s and early ‘80s, or that it increased during the late ‘80s. Like Campbell and colleagues we also found a decline in mortality in both sexes during the 1990s. In view of previous patterns, however, it is questionable whether this recent decline reflects improved prophylactic treatment of the disease or simply minor changes in a broadly constant secular trend.

          We agree with Ann J Woolcock's accompanying editorial, which raised important questions about the meaning of asthma at older ages, the relevance of sex differences, and the need to relate mortality to prevalence, severity, and treatment.5 We are puzzled, however, by her statement that “there was a big fall in death rates between 1983 and 1984”; this is not apparent in Campbell and colleagues' paper.

          If improved prophylactic treatment for asthma were responsible for decreasing mortality, a concomitant fall in hospital admission rates might be expected. Our research in Scotland has confirmed previous reports that hospital admissions for asthma are rising and continue to do so. This continuing steep rise suggests that we cannot yet say that better prophylactic treatment for asthma has shown a clear benefit.

          References

          1. 1.
          2. 2.
          3. 3.
          4. 4.
          5. 5.

          Asthma mortality is falling in most age groups in Scotland

          1. Balvinder Kaur, Clinical lecturer in public health medicine,
          2. Barbara Butland, Lecturer in medical statistics
          1. a School of Health and Related Research, University of Sheffield, Northern General Hospital, Sheffield S5 7AU
          2. b Department of Medicine, University of Southampton, Southampton General Hospital, Southampton SO16 6YD
          3. c National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA, USA 30341
          4. d West Cumberland Hospital, Whitehaven, Cumbria CA28 8JG
          5. e Public Health Research Unit, University of Glasgow, Glasgow G12 8RZ
          6. f Greater Glasgow Health Board, Glasgow G2 4JT
          7. g Department of Public Health Sciences, St Georges Hospital Medical School, London SW17 0RE

            Editor—Until 1989, asthma mortality in young adults in England and Wales remained similar to that occurring three decades ago.1 Ann J Woolcock's editorial, commenting on recent downward trends in asthma mortality in England and Wales,2 requested comparisons of asthma mortality between populations, preferably in relation to prevalence, severity, treatment, and allergic status.3

            We can make three contributions. The first is data from Scotland (source: Scottish Office, from the Lung and Asthma Information Agency) for 1890 deaths due to asthma (1983–95), in which Poisson regression analysis suggests that asthma mortality is falling in most age groups (4): by 7.3% (95% confidence interval 2.9% to 16.5%) a year in 5–14 year olds; 3.2%(0.1% to 6.3%) a year in 15–44 year olds; 3.4% (1.4% to 5.5%) a year in 45–64 year olds; 2.5% (0% to 4.9%) a year in 65–74 year olds; and 2.8% (0.2% to 5.3%) a year in 75–84 year olds. Only in people aged 85 is asthma mortality estimated to have increased, by 2.1% (2.5% to 6.9%) a year. Trends in asthma mortality do not differ significantly (P>0.05) between the sexes, except in 15–44 year olds, in whom the fall is concentrated in men.

            Figure4

            Death rates from asthma per million by age group in Scotland, 1983-95. No deaths occurred in 5–14 year age group in 1994

            The rate of fall in asthma mortality is less pronounced in the 15–64 year age groups in Scotland than in England and Wales, but, with the smaller number of Scottish deaths, the confidence intervals are wider and the rate of fall may be negligible or of the same magnitude. International comparisons of trends in asthma mortality will require larger populations—ones with higher asthma mortality than Scotland's or ones obtained by grouping several years' data. At smaller geographical levels few deaths occur annually and rates will be poorly estimated because of random variation, even if age groups or years are merged.

            Secondly, since deaths due to asthma are relatively uncommon, risk factors may be efficiently studied by case-control comparisons. This would necessarily include validation of certification of deaths due to asthma, which is particularly subject to diagnostic transfer in older age groups, and could be expedited by ongoing confidential inquiries into deaths due to asthma.

            Finally, the international study of asthma and allergies in childhood will be producing questionnaire based data on the prevalence and severity of asthma in children in 40 countries.4 Assessment of treatment and physical measurements remain to be done. The European Community respiratory health survey has already reported on the prevalence and treatment of asthma in adults across Europe.5 Data are emerging to link prevalence, severity, allergic status, and treatment to outcomes, including death, within international frameworks.

            References

            1. 1.
            2. 2.
            3. 3.
            4. 4.
            5. 5.