Objective Understanding the scale of social inequalities in coronary heart disease (CHD) mortality and differential trends in these by age, gender and social position is crucial for assessing the impact on health inequalities of concomitant changes in risk factors, healthcare, legislation and policy on health inequities.

Previous analyses of trends in Scotland suggest that falling CHD mortality rates may have recently begun to level in younger age-groups. Furthermore, the flattening of mortality rates was confined to the most socially deprived groups. Given the uncertainty attached to the estimates, the authors advised caution and suggested confirmatory research of these potentially important findings.

We therefore update previous analyses by examining socio-economic trends over a longer time-span, 1982–2006, and for England (to overcome the previous technical limitation of small counts).

Setting and methods Small areas (c.1500 people) in England were aggregated into equal quintiles by ascending scores on a composite index of multiple deprivation (with quintile 5 the most deprived group). CHD deaths and populations from 1981 to 2007 for each quintile group were used to calculate three-year moving averages of overall (age-standardised) and age-specific mortality rates for each social group, separately by gender for those aged 35 and over.

Trends by quintiles were analysed using joinpoint regression modelling to calculate annual percentage change estimates and to identify turning points in each series.

Results Social gradients in mortality were large and persistent. The rates of decline in age-standardised CHD mortality were socially patterned with the steepest fall in the most affluent quintile. Thus, while absolute inequalities narrowed over the period, relative inequalities increased slightly. For similar age and deprivation groups, rates fell faster for men than women. From 2000, mortality rates levelled off in the most affluent groups (q1 and q2) for women aged 45–54. For men aged 35–44, rates of fall were significantly lower in the most recent period compared to the early 1990s for all except q3, but are still either higher or at the same level as in the late 1980s.

Conclusions Comparisons across and within countries provide valuable insights into potential drivers of inequality for CHD—a leading cause of death in the UK, and worldwide. The differing results for England and Scotland are intriguing. These differences may be real or may reflect data artefact arising due to higher levels of selective health migration in deprived areas and uncertainties in accurately estimating the population structure.