Patterns of coronary heart disease mortality over the 20^th century in England and Wales: Possible plateaus in the rate of decline

Cardiovascular disease appears set to continue as the dominant cause of death and disability in the UK. Age-standardised coronary heart disease (CHD) rates in England and Wales were high between 1950 and the mid 1970s. After this time CHD mortality rates began to fall and have been falling since [1]. Age-standardisation of CHD mortality rates allows for direct comparison between different time periods, when the age structure of England and Wales were significantly different. However, little work has been done on long term trends in age-stratified CHD mortality rates. One recent study suggested a period effect: that patterns in CHD mortality rates between 1920 and 1986 were similar for diverse age groups and, while rates were higher in higher age groups, all age groups experienced a peak in the mid 1970s [2]. Closer examination of age-stratified rates is very important because, when only age-standardised rates are considered, reductions in the mortality rate in older age groups may obscure less positive trends in younger men and women.

The recent trends in cardiovascular risk factor prevalence in younger people are disturbing. Obesity levels in people aged 45 and under have been rising for over ten years, and there has been a similar increase in type 2 diabetes. Physical activity levels in younger people decreased slightly over the last fifteen years, and smoking levels within this age group have remained high, with approximately 25% of men and 20% of women under 45 smoking [3]. These adverse trends could begin to slow, or even reverse, the previously observed decline in mortality rates within younger age groups. Evidence from Australia [4] and America [5] suggests some levelling out of CHD mortality rates among younger men and women and warn that CHD mortality rates in these groups may be starting to increase.

Studying the patterns of mortality over time for chronic disease is important. Firstly, to assess whether current disease reduction strategies are proving to be successful. Secondly, to allow inferences to be made about the contribution of different population factors to disease risk. It is useful to examine long term trends as this allows for recent trends to be viewed in perspective as well as providing a history of the disease burden. As well as recording age-stratified mortality rates over time, it is also useful to record patterns in the number of CHD deaths experienced by different age groups. Such patterns allow for an estimate of the burden of CHD that incorporates the dynamics of an ageing population which are otherwise obscured by declining mortality rates.

A useful alternative to recording trends by calendar year is to examine the experiences of different birth cohorts. This approach allows for a direct comparison of CHD mortality rates within different generations. Such a method can be used to compare the effect of environmental exposures over the whole life course, or those which are typical for a given generation [6].

Recording long-term CHD mortality trends requires use of different definitions for CHD drawn from different International Classification of Diseases (ICD) revisions. Inconsistencies in definitions between revisions effectively limits the potential time period under consideration; it is generally agreed that reliable coding of CHD did not begin until the fourth ICD revision (in 1931) [7]. Reasonably consistent definitions have existed since the seventh ICD revision in 1967 [8]. Because of discontinuities in trends due to definition revisions the patterns in CHD mortality from the first half of the twentieth century should be treated with caution.

This paper therefore examines the variation in CHD mortality by calendar year and by birth cohort for all registered deaths between 1931 and 2005 in England and Wales with a particular focus on recent trends for the younger age groups.

Data on all deaths in England and Wales in the twentieth century by year, sex, age at death and contemporaneous ICD code for cause of death were provided by the Office for National Statistics, along with population estimates for each year-age-sex group. These data were accompanied by data on population estimates, and the number of CHD deaths by age and sex for 2001 to 2005. The data used in this study were subject to previous ethical review within the Office for National Statistics. Separate ethics approval was not considered necessary for this analysis.

A review of literature considering long term trends in CHD produced a framework for the coding of CHD over different revisions of the ICD [6, 7, 9, 10]. Where there was ambiguity over ICD codes one of the authors (SC), a researcher with extensive experience in heart disease, provided a final decision on CHD codes. A coding frame for CHD incorporating ICD revisions 4 to 10 was constructed and is shown in Table 1. This coding frame was used to reduce the twentieth century data to deaths from CHD since 1931.

The age-stratified number of CHD deaths, and the age-stratified CHD mortality rate (number of deaths/population) for men and women by calendar year were tabulated and graphed. The rate of change in CHD mortality rates was tracked by calculating the percentage difference between two sequential years, using moving five year averages for smoothing. The rate of change was tabulated and graphed by calendar year.

Ten year birth cohorts were constructed by subtracting the mean age of each five year age band from the year of death. The population for each birth cohort was derived within gender and age groups from annual population figures. Age-stratified CHD mortality rates were estimated as the number of CHD deaths within these groups, divided by the population. Confidence intervals for rates were calculated using the method described by Altman [11]. It should be noted that mortality rates for ages 30–34 and 35–39 within the 1895–1904 birth cohort were calculated using CHD mortality deaths that occurred from 1931 onwards. Age-stratified population levels for this cohort were calculated in a similar way as for other birth cohorts.

Age-stratified CHD mortality rates for men and women by calendar year are shown in Figure 1. They show a consistent pattern, both within age groups and gender, of a sharp increase until the mid 1970s, and then a steady decline. A sharp rise in the CHD mortality rate for both men and women is evident in 1967, the first year of the ICD-7 revision.

Figure 1 also shows the number of CHD deaths in older age groups of men and women by calendar year. The pattern of a higher burden of CHD mortality for older age groups has only recently emerged in men, whereas it has been established in women for far longer. As an example the CHD death rate among women aged 80–84 exceeded that for women age 60–64 for the first time in 1949, a phenomenon which did not occur in the same age groups in men until 1985 (Table 2).

In 1945 there were 1,823 and 178 less CHD deaths in the older age group (aged 80 to 84) than in the younger age group (aged 60 to 64) among men and women respectively. Ten years later, in 1955, the older age group among women experienced 1,608 more CHD deaths than the younger age group compared with 2,409 less CHD deaths among men. Another fifty years on in 2005 there were 6,114 and 7,839 more CHD deaths in the older age group among men and women respectively.

The rate of change of CHD mortality rates for each age-sex group is shown in Figures 2 and 3. For younger women (under 50), the rate of change in CHD mortality has reversed in the last 20 years. The line in these charts crossing zero suggests a future plateau and possible reversal of previous improvement in CHD mortality rates. Among younger men, the rate of change in CHD mortality has been more consistent than for women over the past 15 years. The line in these charts remaining below zero indicates that CHD mortality rates in this group have continued to fall steadily. In men aged 45 to 49 and 50 to 54 there appears to be a small reverse in the most recent years included in this series. This pattern is very different within older people where the rate of change in CHD mortality rates has continued to improve over the past 20 years.

Variations in CHD mortality rates between those birth cohorts with deaths up to age 59 shows that for both men and women the cohort born in 1925 to 1934 experienced higher rates of coronary heart disease than the 1895 to 1904 cohort (Figures 2, 3 and 4). This trend was observed for all ages from 30–34 to 55–59. As an example of one extreme the CHD rate ratio for 1925 – 1934 cohort using the 1895–1904 cohort as a reference shows that among younger men the rate was as much as ten times higher than the earlier cohort (for men aged 30–34) and four times higher for women (aged 30–34). Rate ratios (and 95% confidence intervals) suggest that these comparisons represent significant differences in the rates between the two time periods. Similarly the rate ratio (and 95% CIs) for those aged 60 and over for both men and women are less than 1.0 suggesting that the trend is for lower mortality rates in older birth cohorts. Table 3 presents mortality rates and rate ratios for ten year time periods. Among men in the youngest age groups the rate ratio comparing CHD mortality in recent calendar years with the referent time period (1936–45) shows that CHD mortality rates were consistently higher in all but the oldest male age group. The rate ratios for women suggest that the progress in relative reduction of CHD mortality rates has been more consistent over time.

In this paper we set out to examine the variation in CHD mortality by calendar year and birth cohort for all registered deaths in England and Wales between 1931 and 2005, a longer time period than previous studies.

Although CHD mortality rates continue to drop in older age groups the actual burden of coronary heart disease is increasing due to the ageing of the population. The rate of improvement in CHD mortality appears to be beginning to decline and maybe even reversing among younger women. CHD mortality rates of those younger than 60 are worse for those born in the early 20^th century than for those born in the 19^th. If this trend in younger age groups is not halted the burden of coronary heart disease is likely to increase.