Papers

Regression analysis of recent changes in cardiovascular morbidity and mortality in the netherlands

BMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7083.789 (Published 15 March 1997) Cite this as: BMJ 1997;314:789
  1. Luc Bonneux (bonneux{at}mgz.fgg.eur.nl), medical epidemiologista,
  2. Caspar W N Looman, biostatisticiana,
  3. Jan J Barendregt, mathematical economista,
  4. Paul J Van der Maas, professor in public healtha
  1. a Department of Public Health Erasmus University Rotterdam PO Box 1738 3000 DR Rotterdam Netherlands
  1. Correspondence to: Dr Bonneux
  • Accepted 16 January 1997

Abstract

Objectives: To test whether recent declines in mortality from coronary heart disease were associated with increased mortality from other cardiovascular diseases.

Design: Poisson regression analysis of national data on causes of death and hospital discharges.

Setting and subjects: Population of the Netherlands, 1969-93.

Main outcome measures: Annual changes in mortality from coronary heart disease, stroke, and other cardiovascular diseases and annual changes in hospital discharge rates for acute coronary events, stroke, and congestive heart failures.

Results: Patterns of cardiovascular mortality changed abruptly in 1987-93. Annual decline in mortality from coronary heart disease increased sharply for women and men: from −1.9% (95% confidence interval −2.2% to −1.6%) and −1.7% (−1.9% to −1.4%) respectively in 1979-86 to −3.1% (−3.5% to −2.6%) and −4.2% (−4.6% to −3.9%) in 1987-93. The longstanding decline in mortality from stroke levelled off: from annual change of −3.3% (−3.7% to −2.8%) and −3.2% (−3.7% to −2.8%) in 1979-86 to −0.1% (−0.7% to 0.4%) and −1.1% (−1.7% to −0.5%) in 1987-93. Mortality from other cardiovascular diseases, however, started to increase: from −2.0% (−2.4% to −1.6%) and −0.2% (−0.5% to 0.2%) in 1979-86 to 1.5% (1.0% to 2.0%) and 1.9% (1.5% to 2.3%) in 1987-93. Hospital discharge rates for acute coronary heart disease, congestive heart failure, and stroke increased during 1980-6. During 1987-93 discharge rates for stroke and coronary heart disease stabilised but rates for congestive heart failure increased.

Conclusion: Improved management of coronary heart disease seems to have reduced mortality, but some of the gains are lost to deaths from stroke and other cardiovascular diseases. The increasing numbers of patients with coronary heart disease who survive will increase demands on health services for long term care.

Key messages

  • In the Netherlands, mortality from coronary heart disease has decreased in recent years, but mortality from other cardiovascular diseases has increased

  • The longstanding decline in mortality from stroke has stopped, and hospital discharge rates of patients with a diagnosis of congestive heart failure have increased

  • The most parsimonious hypothesis explaining these changes is that increasing numbers of survivors of coronary heart disease are boosting the numbers of patients at high risk of other cardiovascular disorders

  • Health services will have to cope with more patients disabled by chronic cardiovascular disease, with their high needs for care

Introduction

In the early 1970s mortality from cardiovascular diseases started to decline in many industrialised countries.1 Despite considerable debate, most observers would agree that reductions in risk factors, particularly smoking and hypertension, was more effective than improvements in treatments in achieving this decline in the 1970s and early 1980s.2 3 4 5 6 7 In the mid-1980s, however, management of acute myocardial infarction was revolutionised, particularly by thrombolytic treatment, causing steep decreases in mortality from coronary heart disease.8 9 10

Coronary heart disease is not the only cardiovascular disease, however, and other cardiovascular diseases, such as stroke and congestive heart failure, share many of the same risks.5 11 12 The improved prognosis for coronary heart disease caused by improved management should increase the number of surviving patients at high vascular risk. We present a time series analysis of Dutch nationwide statistics to illustrate the relation between mortality from coronary heart disease and other cardiovascular diseases.

Methods

Source of data

For our mortality analysis, we used the registered numbers of death by cause, age (from 25 to 84), sex, and calendar year from Statistics Netherlands.13 We took account of only primary causes of death and considered three causes of cardiovascular related death: coronary heart disease; stroke; and all other causes, including unknown, of sudden death.14 Table 1) shows the ICD codes (international classification of diseases) that we searched for.

Table 1

ICD codes used in search for changes in cardiovascular morbidity and mortality in the Netherlands during 1969-93. ICD-8 and ICD-9 (international classification of diseases, eighth and ninth revisions) used for analysis of deaths registered in Statistics Netherlands,13 and ICD-9-CM (international classification of diseases, ninth revision, clinical modification) used for analysis of diagnosis at discharge listed in hospital register

View this table:

The second database we used was the hospital register. This provides nationwide coverage and is complete since 1980. The register includes hospital patients' diagnosis at discharge as classified by the treating physician and codified by local staff. For every discharge, the patient's vital status is registered. Patients dying during the ambulance ride or at entry in the emergency room are not included and are considered “dead out of hospital.” Again, we considered only primary diagnoses.

Statistical analysis

We estimated trends over time by Poisson regression analysis.15 We used mid-year populations as person-years, and we specified five year age groups (categorical) and calendar year (continuous) as independent variables and mortality as a dependent variable. We analysed trends in mortality for 1969-78, 1979-86, and 1987-93 and trends from the hospital register for 1980-6 and 1987-93. We chose 1987 as the cut off point because of the apparent rupture in trend in that year.

Results

Figure 1) and table 2) show changes in mortality from cardiovascular diseases and in discharge rates from hospital for stroke, acute coronary heart events, and congestive heart failure.

Fig 1
Fig 1

Age standardised (Dutch population of 1980-93 as direct standard) annual mortalities and hospital discharge rates for cardiovascular diseases in the Netherlands

Table 2

Age standardised rates (per 100 000 population) and annual percentage changes in mortality and hospital discharge rates for population of the Netherlands during 1969-93. (Data on national mortality from Statistics Netherlands13 and data on hospital mortality and morbidity from hospital register)

View this table:

For women, all cardiovascular related death rates declined from 1969 until 1986, suggesting a change in common risk factors of most cardiovascular diseases. During 1987-93, however, changes in cardiovascular related mortality levelled off, from −2.3% a year in 1969-86 to −0.7% a year in 1987-93. Mortality from coronary heart disease decreased steeply, but other vascular related death rates started to increase in 1987. Mortality from stroke levelled off after a long period of decline.

For men, changes in cardiovascular related mortality during 1969-78 were limited to a decline in mortality from stroke. During 1979-86 the rate of decline in cardiovascular related mortality increased, driven by declining death rates from both coronary heart disease and stroke; rates for other cardiovascular causes of death changed little. In 1987-93 mortality from coronary heart disease declined steeply, but death rates from stroke declined less steeply and death rates from other cardiovascular causes seemed to increase. These upward changes are similar to, but less clear than, those seen among women.

The hospital register showed an increase in discharge rates for acute coronary heart disease, congestive heart failure, and stroke in 1980-6. In 1987-93 the discharge rates stabilised for stroke and coronary heart disease, though rates for coronary heart disease still increased among women. The age adjusted case fatality ratio for an acute coronary event decreased from 13.6% (95% confidence interval 13.3% to 13.9%) for men and 18.4% (17.8% to 19.0%) for women in 1980-1 to 7.9% (7.7% to 8.1%) for men and 9.9% (9.5% to 10.2%) for women in 1992-3. The case fatality ratios for stroke remained constant at about 17.7% for both sexes since 1987. For congestive heart failure, both sexes showed an age dependent increase in discharges by calendar date. In more recent years, more patients were hospitalised at higher ages.

Discussion

Our study provides circumstantial evidence that the sharp drop in mortality from coronary heart disease between 1985 and 1992, the levelling off of mortality from stroke, and the increase in mortality from congestive heart failure are causally linked by the same process: the increased survival of patients with coronary heart disease. Cardiovascular diseases share many of the same risk factors, and having one disease increases the risk for others: a history of ischaemic heart disease increases the risk for other heart diseases, notably heart failure and dysrhythmia, and these increase the risks for cardiogenic stroke.16 As the prognosis for patients with coronary heart disease improves the increasing numbers of surviving patients will increase the pool of people at high risk of other heart diseases and stroke. This will result in increasing death rates from stroke and other cardiovascular diseases.

Validity of study

The changes we observed might have been caused by changing diagnostic habits, policies for referral, or rules of classification. The validity of the Dutch register of causes of death is reasonable.17 Any changes in coding between cardiovascular and non-cardiovascular causes of death are likely to have been small and unable to bias seriously trends over time. The possibility of misclassification between different causes of cardiovascular related death is high, but the observed patterns in both the hospital register and the mortality statistics were consistent (see table 2).

General practice registers have shown that nearly all patients suspected of having an acute myocardial infarction are hospitalised in the Netherlands.18 Death rates both outside hospital and in hospital showed substantial reductions, which makes it unlikely that deaths outside hospital were exchanged for deaths in hospital. Moreover, the secular trend of improving prognosis has been documented before, in the Netherlands as elsewhere, and has been linked to improved management.4 8 9 10 19

Part of the increased rates of discharge of patients with stroke and the decrease in case fatality in the early 1980s was caused by the introduction of computed tomography, which ascertained more benign lesions.6 7 20 Since 1987, the incidence of and mortality from stroke have remained constant, suggesting a steady state in survival. The sharp decrease in mortality from coronary heart disease and the concomitant levelling off of mortality from stroke after a long period of decline has been observed in the prospective, population based Minnesota heart survey.7 9 20 The age dependent increase in mortality from congestive heart failure has also been documented before.21 22 23 24

Secular changes in risk factors might explain the observed changes, but these have been modest (at best) in the period under study in the Netherlands.25

Conclusion

Improvements in treating coronary heart disease seem undeniable, but some of the gains made are lost again to deaths from stroke and congestive heart failure. This has important consequences for public health, as increasing numbers of surviving but disabled patients with chronic cardiovascular disease are boosting demand for health care.

Acknowledgments

We thank J B Reitsma for his support in extracting the data from the hospital register.

Footnotes

  • Funding This study was funded by the Netherlands Heart Foundation and the Dutch Foundation for Scientific Research.

  • Competing interests None.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
View Abstract