Background Coronary heart disease (CHD) mortality has steadily declined since the early 1970s in the Netherlands. However, in some Western countries the rate of decline in younger groups may be starting to plateau or even rise.
Objective To examine trends in age-specific CHD mortality rates among Dutch adults from 1972 to 2007, with a particular focus on recent trends for the younger age groups
Methods Data for all CHD deaths (1972–2007) in the Netherlands were grouped by year, sex, age. A joinpoint regression was fitted to each age-sex-group to detect points in time at which significant changes in the trends occur. For every time period, the linear slope of the trend, p value, observed number of deaths, CHD mortality rates and change in the CHD mortality rate were calculated.
Results Between 1972 and 2007, the age-adjusted CHD mortality rates decreased overall by 76% in both men and women. In men (35–54 years), the change in CHD mortality rate in the period 1980–1993 was −0.53 but attenuated in period 1993–1999: −0.16. In women (35–54 years) the decline likewise attenuated to −0.44 in period 1979–1989: and −0.05 in period 1989–2000. After 1999–2000, CHD mortality rate further declined in both men (period 1999–2007: −0.46) and women (period 2000–2007: −0.38).
Conclusions Evidence from several Western countries suggests that among young adults (<55 years), CHD mortality rates are levelling out. In this study, similar attenuation of the decline in CHD mortality among young adults in the Netherlands has been observed. Furthermore, this is the first study to observe a subsequent increase in the pace of decline after a period of flattening. In order to better explain these encouraging changes in CHD mortality rates, a detailed analysis of recent changes in cardiovascular risk factors and treatments is now urgently required.
Statistics from Altmetric.com
There has been a consistent decline in coronary heart disease (CHD) mortality in the Netherlands since the early 1970s. This reflects both a decrease in incidence as well as improvements in population risk factors and better treatments for patients with CHD.1 However, CHD remains a dominant cause of death and disability and claims a heavy economic toll. A recent report on the cost of diseases in the Netherlands ranked CHD as one of the top 10 most expensive diseases.2
Evidence from Australia, America and the UK3–5 suggests some levelling out of CHD mortality rates among younger men and women and with warnings that CHD mortality rates in these age groups may even be starting to increase. These unfavourable mortality rates may reflect adverse trends in major cardiovascular risk factors, with a dramatic increase in obesity and diabetes,6 a flattening of blood pressure falls7 and persistent smoking in young adults 8 in most industrialised countries.
The Dutch population has shown favourable trends for physical inactivity,9 cholesterol and dietary intake of fat.10 However, recent substantial rises in obesity9 and diabetes11 have also been observed, which might compromise the previously observed decline in cardiovascular mortality rates, especially within younger age groups. Our objective, therefore, was to examine trends in age-specific CHD mortality rates among Dutch adults from 1972 to 2007, with a particular focus on recent trends for the younger age groups.
Data on all deaths in the Netherlands from 1972 to 2007, were provided by Statistics Netherlands, by year, sex, age at death and contemporaneous International Classification of Diseases (ICD) code for cause of death, along with population estimates for each year-age-sex-group. To make results comparable with other national studies, the overall rates (number of deaths/population ×100 000) were standardised to the European standard population.12
The overall rates from 1980 to 2007 for cancer, CHD, stroke and other causes of death were depicted by graphs. Rates were standardised to the Dutch population in 2007.
To identify points at which a significant change of the age- and gender-specific trend in mortality occurred, we performed joinpoint regression analysis with the software (Joinpoint version 3.4.0) provided by the Surveillance Research Program of the US National Cancer Institute (2008). With this analysis, it is possible to identify years when a significant change in the linear slope of the trend (on a log-scale) is detected over the study period.13 The best fitting points, called ‘joinpoints’, are chosen when the rate changes significantly.
The analysis starts with the minimum number of joinpoints (eg, 0 joinpoints, which is a straight line), and tests whether one or more joinpoints (up to three, corresponding to four distinct periods identified by trend 1 to trend 4) are significantly different and must be added to the model. In the final model, each joinpoint (if any) informs of a significant change in the slope. We used a Bayesian information criterion approach to select the most parsimonious model that best fitted the data.13 14
For every period the linear slope of the trend and p value of the final model of the joinpoint regression analysis was tabulated. In these tables also the minimum and maximum observed number of deaths and the minimum and maximum observed CHD mortality rates per 100 000 (number of deaths/population)×100 000)) were presented for every identified period. Furthermore, we calculated the change in observed CHD mortality rate per 100 000 per period ((CHD mortality rate last year of period—CHD mortality rate first year of period)/CHD mortality rate first year of period). Finally, 10 age- and sex-specific graphs showing the observed and modelled CHD mortality rates per 100 000 from 1972 to 2007 are presented. The dots in the age- and sex-specific graphs represent the observed number of deaths per 100 000 from 1972 to 2007. The line in the age- and sex-specific graphs represents the final model from the joinpoint regression analysis. All analyses were performed in accordance with privacy legislation in the Netherlands.15
Age-adjusted trends in CHD mortality
Between 1972 and 2007, the age-adjusted (European population) CHD mortality rates decreased overall by 75.6% in men and 75.7% in women (figure 1).
Compared with other causes of death the decline in CHD mortality rates from 1980 to 2007 was more pronounced (figure 2). In men, age-adjusted (Dutch population 2007) mortality rates for cancer, stroke and other causes of death decreased overall by 18.0%, 59.3% and 28.0%, respectively. CHD decreased overall by 76.3%. In women, age-adjusted (Dutch population 2007) mortality rates for cancer, stroke and other causes of death decreased overall by 13.1%, 58.3% and 33.1%, respectively. CHD mortality rates decreased overall by 74.4%.
Age- and sex-specific CHD mortality trends
Age- and sex-specific trends in CHD mortality showed a more complex picture (figure 3). In men aged between 55 and 74 years, we observed a constant decline of the change in CHD mortality rate over time (table 1). However, men aged 35–54 showed clear attenuation of the decline in CHD mortality rate in the period 1993–1999 (change in CHD mortality rate in period 1980–1993: -0.53 and in period 1993–1999: −0.16). After 1999, CHD mortality rate declined further (change in CHD mortality rate in period 1999–2007 was −0.46). In men aged 75–94 years, the change in CHD mortality rate attenuated significantly in the last period; however, this decrease was less pronounced (age group 75–84: from −0.32 in period 1996–2003 to -0.031 in period 2003–2007, age group 85–94: from −0.39 in period 1972–1998 to −0.30 in period 1998–2007).
In women aged between 55–64 years and 85–94 years, we observed a constant decline of the change in CHD mortality rate over time. However, in women aged 35–54, 65–74 and 75–84 years, attenuation of the decline in CHD mortality rate was seen (table 2). In women aged 35–54 years, attenuation was the most pronounced (change in CHD mortality rate in period 1979–1989: −0.44 and in period 1989–2000: −0.05). However, after 2000, CHD mortality rates declined further (change in CHD mortality rate in period 2000–2007 was −0.38).
Summary of the study findings
This analysis of mortality data from CHD is consistent with earlier analyses demonstrating substantial declines in CHD since the early 1970s. However, we observed for the first time that after a period where the decline in CHD mortality attenuated, mortality then substantially declined further. Furthermore, there is a clear disparity in trends in CHD mortality rates between the different age-sex-groups. Slowing or flattening of the decline in CHD mortality in young adults has been reported in England and Wales,4 the USA,3 New Zealand and Australia.5 16 In this study, attenuation of the decline in CHD mortality among men and women aged 35–54 years was also seen.
The underlying causes of the complex changes in CHD mortality rates are far from clear. A variety of CHD policy models have therefore been developed17 18 to better explain observed trends. Modelling studies in many developed countries consistently suggest that 55–75% of demonstrated, substantial CHD mortality decline can be attributed to decreases in major risk factors, the remaining part being attributable to medical and surgical treatments.19–21
Attenuation of the decline in CHD mortality among young Dutch adults in the 1990s is alarming, because it occurred despite the increasingly wide use of lipid-lowering drugs,22 increased physical activity,9 decreased cholesterol levels and decreased dietary intake of saturated fat10 during this period. This suggests that adverse trends in unfavourable risk factors, especially obesity and diabetes, compounded by persistent smoking, may powerfully contribute to changes in CHD mortality rates, as both obesity, particularly among young adults, and diabetes increased during the 1990s in the Netherlands,9 11 as in England and Wales.4
In contrast to the other studies describing mortality flattening, this is the first study to report a further decline in CHD mortality rates since 1999 among young men (and since 2000 among young women). Data on changes in risk factors in the Netherlands since 2000 are limited, but indicate that the prevalence of obesity may have stabilised.23 Furthermore, smoking prevalence further decreased after being relatively stable during the 1990s.24 These changes may have contributed to the more recent downward trend in CHD mortality. However, explanation of the changes in CHD mortality is difficult, as the available data for age- and gender-specific risk factors and treatments is limited for the period since 2000 and thereafter.
CHD mortality trends cannot always be described accurately above the country level, as countries in the same geographical or administrative region might have different peaks in the mortality rate which are reached at different time points.25 Crucially, our study and previous studies26 27 on age-specific CHD mortality trends suggest that the pace of change both in magnitude and direction also might be different among countries that are experiencing otherwise similar epidemiological patterns. Furthermore, there is growing evidence that such changes can occur very soon after alterations in risk factors. For example, in Poland, a rapid decline in mortality due to CHD was seen between 1991 and 1994, following dramatic dietary changes in 1989 and 1990 with increases in the ratio of polyunsaturated fat to saturated fat and fruit consumption. Similar changes were observed in neighbouring countries, including the Czech Republic and East Germany.27
Message for policy makers
There is growing evidence that slowing or flattening of the decline in CHD mortality and relatively quickly changes in CHD mortality may rapidly follow changes in risk factors. This is an important public health message. Therefore closer monitoring of CVD mortality rate is now required. Furthermore, it is critical to understand what is driving these rapid changes in the Netherlands. A detailed analysis of recent changes in cardiovascular risk factors and treatment is urgently needed, in order to better inform policy.
Policy makers might appreciate that CVD mortality rates can change rapidly, in the Netherlands, as in diverse other population as it might suggest that prevention interventions provide a return on the investment surprisingly quickly (in years rather than decades). However, the CHD trends in the Netherlands do not necessarily imply a similar trend for other countries in the same geographical or administrative region. It is therefore still important to take country-specific data into account when considering policy implications.
Strengths and limitations
We were able to perform trend analysis over a large (35 year) time frame. Furthermore, joinpoint regression analysis is able to identify periods of similar annual percentage change in an objective manner. This avoids the need to prespecify time periods (which may bias the way in which the trends are analysed).
Any study that uses mortality data across multiple revisions of the International Statistical Classification of Diseases (ICD) will suffer attribution bias owing to both the change between versions of ICD and the procedures used to code deaths. However, it is very unlikely that attenuation of the decrease in CHD mortality is a result of such bias, because in men and women flattening of the data was demonstrated in different time periods, while quality of coding in CHD has shown to be similar for men and women.28
Attenuation of the decrease in CHD mortality is not an artefact of low rates, as mortality substantially declined further in the period thereafter and it has been also shown among age groups with high rates. In addition, a corresponding decline in hospitalisation for acute myocardial infarction was observed between 1995 and 2000 in all age groups, except for men younger than 40 years and women younger than 50 years.29
In conclusion, recent age-specific changes in risk factor trends in the Netherlands are encouraging. However, in order to inform future cardiovascular prevention strategies, a better understanding of such trends is necessary. A more detailed analysis of recent changes in cardiovascular risk factors and treatments is therefore now urgently required.
Funding IV was funded by the Netherland Heart Foundation (grant NHF project Facts and Figures). The funding source had no involvement in the study design, collection, analysis and interpretation of data, writing of the report and decision to submit the paper for publication.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.