Background
Ischeamic heart disease (IHD) mortality trends have fallen in developed countries over recent decades; this is mainly due to a decrease in disease incidence through improved primary prevention measures and to a drop in case fatality reflecting advances in diagnosis, treatment and secondary prevention[
1]. Although downward trends in IHD mortality have been seen in Scotland [
2] these have been to a lesser extent than in other Western European countries, with the result that Scotland has one of the worst IHD mortality rates in the region[
3]. Scotland was found to have among the highest rates of fatal and non-fatal events when comparing the MONICA populations [
4]. However, survival rates from IHD in Scotland make for a more favourable comparison; results from the MONICA study showed that 28-day case fatality in Scotland was the same as or lower than the average across all populations. This suggests that high incidence has been driving Scotland's high IHD mortality.
Incidence rates of acute myocardial infarction (AMI), the most common form of IHD, have also declined substantially over recent years[
2]. This mainly reflects improvements in primary prevention of the disease; for example, there have been intensive lifestyle interventions to reduce levels of exposure to risk factors such as smoking[
5]. To bring Scotland's IHD mortality rates to a rate comparable with other European countries the downward trends in first time AMI events must continue, ideally at a faster pace.
Given the Scottish Government's commitment to tackling inequalities in health[
6] and the contribution inequalities make to overall mortality[
7], it is important to explore whether similar declines in AMI incidence are being experienced by all population groups within the country. Incidence is an important health indicator when examining the burden and patterning of a disease; however, to date, studies in Scotland have mainly focused on various inequalities in cardiovascular death [
8‐
10] or survival[
8,
11‐
13] and treatment[
14] of AMI or IHD. Little is known about social and geographical inequalities in AMI incidence and how they interact with age, gender and time. If such inequalities in first time AMI events have increased in Scotland it would suggest that improvements in primary prevention of the disease have not been experienced equally by all sections of Scottish society.
Population based studies of AMI incidence are uncommon but Scotland is in the advantageous position of having a data system permitting the linkage of information on hospital admissions with mortality data for the whole population (5.1 million in 2001[
2]). The aim of this work is to use these data to examine the changes in socioeconomic and geographic inequalities in incidence of AMI in Scotland between 1990–92 and 2000–02 and to explore how these inequalities interact with age and gender.
Discussion
This is the largest population based study to explore trends and inequalities in AMI incidence and their interaction with age and gender. Other studies have explored the associations between deprivation and IHD incidence [
22‐
25], the largest of which examined 52,360 individuals who developed IHD (from a study of 2.6 million men and women)[
24]. These studies found socioeconomic gradients in incidence but did not report interactions with age and gender or unexplained geographical variation. Within Scotland, most work investigating IHD or AMI to date has focused mainly on mortality [
8‐
10] and case fatality or survival patterns[
8,
9,
11,
26].
We have shown that AMI incidence rates have decreased substantially over recent decades; such decreases are likely to be due to improvements in primary prevention of the disease[
27,
28]. Male incidence rates for those aged 30–44 years old have decreased by 23%, the corresponding rates for those ages 45–59, 60–74 and 75+ years were 37%, 39% and 28%. The figures for women were 5%, 45%, 40% and 25%. IHD mortality also fell more sharply for women aged 45–59 during this time period [
10]; mortality rates for men decreased by 44% whilst the rate for women fell by 51%. At older ages the declines were comparable: 44% and 46% respectively for men and women aged 60–74 and 27% and 25% at ages 75+. Although comparing two slightly different disease groups – AMI incidence with IHD mortality – there is a suggestion that there has been a steeper downward trend from 1990–92 to 2000–02 in mortality than incidence. This could partly be due to the increasing use of more sensitive enzyme diagnostic testing, in particular troponin, from the beginning of the current decade resulting in higher hospitalisation rates[
29]. More up-to-date data are needed to explore the true effect of troponin on AMI incidence patterning. Regardless of this, reducing the incidence of AMI will have a substantial impact on mortality and, consequently, help reduce the rates in Scotland to a level that is comparable with the rest of Europe.
We have shown that socioeconomic deprivation has a marked effect on the risk of having a first AMI. Relative inequalities across Scotland are steepest in the youngest age groups and most pronounced in young women. Other work looking at IHD mortality has shown that, despite large reductions in rates, there remain strong socioeconomic variations in the disease[
30,
31]. The steepest inequalities were again found in the 45–59 year old age group in 2000–02 but, contrary to our findings for AMI incidence, these inequalities were higher in men than women[
31]. Studies which have examined short-term AMI case fatality have frequently shown stronger inequalities in the young and in particular women[
13,
32‐
35].
During the 10-year period under investigation in this paper, inequalities increased with the steepest changes being seen in the younger age groups (under 60). Increasing inequalities are likely to be due to higher levels of exposure of young people in deprived areas to risk factors such as smoking. In Scotland, in both men and women, cigarette smoking prevalence is highest in the 25–34 age group (39% of men and 35% of women), followed by those aged 35–44 years (34% of men and 33% of women)[
36]. Between 1995 and 2002 there was a small decrease in the proportion of men aged 16–64 years who smoked (34% to 31%) [
37‐
39] with a comparable decrease for women (36% to 32%). The number of cigarettes smoked per day is significantly associated with deprivation for both sexes with a steeper gradient for women than for men[
36], echoing the association between AMI incidence and area deprivation. In 2001, 70% of smokers were in lower socioeconomic groups[
37]. Capewell et al[
40] quantified the extent to which the fall in IHD mortality in Scotland was attributable to risk factor changes (and how much to medical and surgical treatment). They suggested that 51% of the reduction in mortality over a 20 year period was due to risk factor modifications. This reduction in mortality will largely be due to a reduction in incident events of the disease. They further estimated that, of this 51% reduction, smoking accounted for 36%, a secular fall in blood pressure and cholesterol reduction each contributed 6% and 3% was due to deprivation (susceptible to confounding by smoking, diet and blood pressure changes). Therefore trends and inequalities in smoking habits are likely to explain a large part of the changes in AMI incidence in Scotland over time.
Other traditional risk factors such as diet, physical inactivity, overweight, hypertension or cholesterol [
5] may also contribute to the changing patterns of AMI incidence, but to a lesser extent than smoking. For example, there have been substantial increases in obesity, which is associated with diet and physical inactivity, in more deprived areas in Scotland over recent years, with higher rates in women[
36]. Also, diabetes rates, which are associated with obesity levels, have increased in Scotland and are associated with deprivation with a suggestion of a stronger gradient in women[
36]. Findings from an English cohort study [
41], which examined whether deprivation status had an influence on changes in cardiovascular risk factors in middle-aged (35–55 years) men and women, suggested that there had been a widening deprivation gap in populations with high blood pressure. If the same were true in Scotland this could also contribute to the increasing inequalities in AMI incidence. That study also found no association between deprivation and change in cholesterol, and, perhaps somewhat surprisingly, a narrowing of inequalities in smoking in women. Other research has concluded that there is a narrowing of socioeconomic inequalities in high blood pressure rates in men, and in high body mass index and cholesterol in both sexes and a widening of inequalities in high blood pressure in women and smoking in both sexes [
42]. All such findings suggest that the relationship between cardiovascular risk factors and deprivation is complex and varies across sex, age and geographical groups. Further work within Scotland examining the relationship between such risk factors and AMI incidence will be important in developing our understanding of inequalities.
Another factor contributing to the inequalities in AMI incidence might be variations in follow-up contact rates with general practitioners (GP) following a diagnosis of angina. It is well known that patients often undergo episodes of unstable or stable angina prior to their first AMI[
43] and that young deprived patients (and in particular women) with angina have less frequent follow-up contacts with their GP. For example, one year GP contact rates in Scotland in 2001–02 for angina were higher in men than women, were extremely low in patients under 45 years and declined significantly with increasing socioeconomic deprivation[
44]; these factors may all contribute to the inequalities reported here. Although reducing exposure to risk factors is an important primary prevention strategy in all population groups, greater inequalities in AMI incidence at younger ages, and in particular among younger women, would suggest particular emphasis should be made on these sub-groups.
Much, but not all, of the geographic variation in AMI incidence in Scotland can be explained by a simple measure of area socioeconomic deprivation. After accounting for age, sex and deprivation a substantial proportion of the remaining variation in 2000–02 was attributable to differences between LCAs, particularly in women, suggesting strong regional patterning of AMI incidence and warranting investigation into the likely mechanisms linking area of residence to occurrence of first AMI events.
It may be misleading to present relative inequalities alone as these are related to underlying rates and the scale on which the inequalities are measured[
45,
46]. For example, the number of incident AMI events is much higher in older than younger age groups; therefore, risk ratios (comparing affluent to deprived areas) that are as high as in the younger age groups could only be achieved at older ages if the absolute differences between affluent and deprived areas were considerably larger. We have presented age standardised rates alongside risk ratio comparisons so that absolute differences can also be considered. Absolute inequalities (comparing the most deprived to the most affluent areas) were highest for men aged 60–74 years and have increased despite the decline in incidence rates at these ages between 1990–92 and 2000–02. There were modest decreases in inequalities in all other age groups. Absolute inequalities were highest for women aged 75+ years and the decreases over time in all age groups were small. We feel relative inequalities are more informative here, especially when examining the younger age groups; an absolute difference in AMI incidence at younger ages, which may seem small in comparison to the older age groups, may be of greater public health importance. IHD is a highly preventable disease in the young, so reducing the rates in the more deprived areas to a level similar to that in the more affluent areas should be achievable and is of primary importance.
Study limitations
One limitation of our study is that we only had an area based measure of deprivation as opposed to individual socioeconomic status. The Carstairs deprivation index is a commonly used measure and has been validated against individual socioeconomic status [
47]; however, the population size of the geographic area (postcode sector) for which our deprivation index is derived may be influencing our estimation of the socioeconomic gradient. Estimates of inequalities based on area deprivation have been shown to be diluted when the geographical units are large[
48]. The smaller areas in this study, postcode sectors, are still fairly large (mean population 5402) and therefore are likely to be heterogenous. However, the interpretation to be drawn from the analysis does not relate to individual socioeconomic status but to the area context. Previous work has shown the prevalence of cardiovascular disease in Scotland to be related to area deprivation (at the postcode sector level, based on the Carstairs score) and not individual occupational social class [
49]. It is unfortunate that the routine data used in this study do not permit adjustment for both individual and contextual measures. Secondly, on admission to hospital, MI is normally diagnosed as non-ST-elevated (non-STEMI) or ST-elevated (STEMI) and the treatment and prognosis differ according to this diagnosis [
50‐
52]. A further important avenue of research would therefore be to investigate how the trends in incidence of STEMI and non-STEMI contribute to the overall trends in MI incidence. Again, it is unfortunate that our routine data do not enable a distinction between STEMI and non-STEMI, and we therefore cannot comment on whether trends and inequalities reflect changes in one or the other. Moreover, we do not have routinely collected data on IHD risk factors for this population and therefore can only hypothesise as to why rates are decreasing and inequalities persisting or increasing. We can also only hypothesise about the contribution that reductions and inequalities in AMI incidence are having on the trends and patterns of AMI mortality in Scotland; further work incorporating case fatality from the disease is needed to answer such questions.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CD is the corresponding author and guarantor of this paper. CD formulated the research question, analysed and interpreted the data and wrote the paper. AL initiated the study and commented on the paper. RD helped with data analysis and commented on the paper.