Heart failure mortality in Spain, 1977-1998

INTRODUCTION

As is already known, cardiovascular diseases are the primary cause of mortality in Spain and other developed countries, fundamentally due to ischemic heart disease and cerebrovascular disease.¹ However, heart failure is a condition that is acquiring increasing relevance, currently constituting the third cause of death due to cardiovascular disease, in addition to being an important cause of morbidity and the hospital workload. In Spain, heart failure is the first cause of hospitalization in persons over 65 years, which represents 5% of all hospital admissions in that age group.² In the U.S. heart failure is also the main cause of hospitalization in persons of advanced age, 70% of the persons affected being more than 60 years old. A prevalence of 1% at the age of 50-59 years is estimated, which increases to 10% at the age of 80-89 years. The most common cause of heart failure in 13 studies published in the last 10 years in the New England Journal of Medicine was coronary artery disease, which occurred in 70% of 20 000 patients studied.³ In spite of improvement in the treatment of heart failure, it continues to be a highly lethal disease, with a median survival of 1.7 years for men and 3.2 years for women.^4-6 Heart failure has been referred to as a public health problem of epidemic proportions, because it is responsible for major human and economic costs.⁷ If we add to this the decrease in the quality of life of affected patients, who are mainly persons of advanced age, heart failure is an important public health problem.

In this study we report the mortality due to heart failure in Spain in the last 20 years. The data are grouped in five 5-year periods and one 2-year period, corresponding to the periods 1977-1981, 1982-1986, 1987-1991, 1992-1996, and 1997-1998, the last years for which we have information from the National Institute of Statistics (INE, abbreviation in Spanish).

METHODS

The data on deaths were obtained from individual registries facilitated by the INE, and correspond to deaths of residents in Spain.

To limit the difficulties derived from changes introduced in the coding of heart failure in the ninth revision of the International Classification of Diseases (ICD) with respect to the eighth revision, more specific ICD codes have been selected. The fourth digit of the ICD-8 was used to facilitate study of the temporal series as a homogeneous unit. We included deaths whose code in 1975-1979 (eighth revision)⁸ was:

1. 427: symptomatic heart disease:

­- 427.0: congestive heart failure.

­- 427.1: left ventricular failure (codes 427.2, cardiac blockade, and 427.9, other cardiac rhythm disorders, are excluded).

2. 428: myocardial heart failure. Other forms of myocardial insufficiency.

For deaths that occurred after 1980, (ninth revision)⁹ the categories coded in section 428 have been selected for analysis: Heart failure, which includes codes 428.0: congestive heart failure, 428.1: left heart failure, and 428.9: unspecified heart failure.

The morbidity data were obtained from the National Minimum Data Set (NMDS), which is available on the web page of the Ministry of Health and Consumption. The NMDS code 428 corresponding to heart failure was collected specifically, which is not the case of the Hospital Morbidity Survey, in which it is grouped in codes 415 to 429.

The national annual rates were adjusted for 18 age groups ranging from 0-4 years to 85 years and over. The direct standardization method was used, taking as reference the standard European population. Rates are presented by 100 000 persons/year. The populations used were elaborated by the National Center of Epidemiology, using the population at mid-period for each 5-year period. The populations used for 1997-1998 were facilitated by the INE.

To analyze the statistical significance of the temporal tendency, linear log regression models were adjusted, assuming that the rates had a Poisson distribution and varied with age and time, using the «generalized linear model» procedure implemented in the S-Plus statistical program. Independent models were adjusted for men and women.

RESULTS

In 1998, heart failure caused induced the death of more than 21 000 people, equivalent to 4%-8% of the mortality due to all causes in men and women, respectively, and 12%-20% of cardiovascular mortality. It is the third cause of cardiovascular mortality after ischemic heart disease and cerebrovascular disease. In women, heart failure in general and cardiovascular mortality are practically twice as high as in men (Table 1).

The distribution by sex and age is shown in Table 2, where it is evident that rates increase with age in both men and women, remaining low until 60 years of age and with the greatest mortality concentrated in the elderly population. In absolute figures, twice as many deaths occur in women as in men; however, this is due to the different age distribution of the populations of each sex, with the feminine population being older. When adjusted for age, the rates become closer, 33 per 100 000 men and 34 per 100 000 women. When each age group is considered independently, the mortality is greater in men until the age of 75-79 years.

The study of the geographic distribution (Table 3; Figure 1), prepared for 1996-1998 to obtain more stable rates, shows that there were important interprovincial differences in Spain, with variations as high as 70%. The highest rates were found in Andalusia, Balearic Islands, Castellón, Girona, and Melilla and the lowest rates en Basque Country, Guadalajara, Segovia, and Soria. In Figure 1, a clear north-south geographic pattern is observed, as in the mortality due to ischemic heart disease or cerebrovascular disease.¹

Fig. 1. Geographic distribution of mortality rates due to heart failure. 1996-1998 period. Rates adjusted for age, 100 000 persons/year.

In Table 4 and Figure 2, the evolution of mortality due to heart failure since 1977 is shown. A decrease in rates is observed the last 20 years, which in men has represented a decrease of 40%. In women, the downward tendency is marked, but even so, a decrease of 30% was reached. The rates adjusted for age are slightly higher in men than in women in the first three 5-year periods; Nevertheless, from the 1992-1996 five-year period, rates in women exceeded those of men. The decrease in mortality was confirmed statistically by fitting a Poisson regression model. The data collected reveal that the downward tendency is greater in men than in women, with a (coefficient of -­0.14 in men and ­ -0.009 in women, a standard error of 0.002 in men and 0.001 in women, and statistical significance in both cases of P<.001 (Figure 2). The general decrease in mortality due to heart failure occurred in all the age groups, with two exceptions that we considered relevant. On the one hand, the mortality in the population over 85 years is almost the same as it was 20 years ago. From 1977 to 1991, mortality in the very elderly increased slightly in men (2.2%) and very significantly in women (11%), but since 1991 it has descended 5% in men. Among very elderly women, the variation in recent years is minimal (Table 4; Figure 3).

Fig. 2. Temporal evolution of mortality due to heart failure. Rates adjusted for age, 100 000 persons/year.

Fig. 3. Temporal evolution of mortality due to heart failure in older people. Specific rates by age, per 100 000 persons/year.

In spite of the general decrease in mortality rates, the absolute number of deaths (Table 4) in men has remained practically constant, whereas in women it has increased at the expense of the oldest age groups. If the morbidity data of the NMDS are also taken into consideration, which specify the number of hospital releases due to heart failure (Figure 4), an important increase in both sexes is observed.

Fig. 4. Temporal evolution of morbidity due to heart failure. 1997-1999 period. Number of hospital releases.

On the other hand, the mortality rates in young men 25 to 34 years-old increased strongly in 1980-1990, a phenomenon that is not observed in women (Table 4).

DISCUSSION

Before making any interpretation of these results, some methodological points must be considered. In spite of being a very frequent cause of death, few studies have been published on mortality due to heart failure.^2,10-12 It is likely that this is because the analysis of mortality due to heart failure is difficult due as much to the certification as to the coding of this cause of death. On the one hand, the certification of the basic cause of death in patients with heart failure is difficult because this condition is common in the terminal stage of many different diseases. On the other hand, the coding of heart failure was modified substantially in the ninth revision of the ICD, which came into effect in 1980, which represented a change in nomenclature that affected the homogeneity of the temporal series. These difficulties have made it impossible to know the characteristics of mortality due to heart failure. We think, however, that the present magnitude of the problem justifies its study in spite of these limitations, as long as methods are used to limit these difficulties and they are considered in the interpretation.

Different studies of tendencies in mortality due to heart failure use disparate criteria for avoiding these problems. For that reason, they cannot be compared directly (Table 5). In the present study, we used the criterion of studies referring to the U.S. and Catalonia,^11,12 selecting ICD-9 code 428 since 1980. Before 1980, codes 427.0 and 427.1 were added to code 428, a criterion that has been followed in part in another study.¹⁰ Our results show that this criterion resulted in a homogeneous series from 1975, one that did not present the typical disturbances due to changes in nomenclature,¹² which is why we propose that this method be used in the future.

The certification of heart failure as the cause of death, as noted, entails problems. In fact, in some countries, like the United Kingdom, it is expressly prohibited to list heart failure as the primary cause of death on death certificates and the underlying cause is specified instead.¹³ There is no universal agreement regarding the definition of heart failure even in clinical practice, and there is a widely recognized need for clinical, epidemiological, and investigational criteria to facilitate comparisons between different studies. In this sense, the European Society of Cardiology has published guidelines for the diagnosis of heart failure.¹⁴ In Spain, similar initiatives have been undertaken by the Heart Failure Working Group of the Spanish Society of Cardiology, which has developed guidelines for the diagnosis, classification, and treatment of heart failure and cardiogenic shock¹⁵, the aim of which is to improve the classification of heart failure in order to know the true incidence and prevalence of this disease and to better classify patients.

The results of our study reveal that mortality due to heart failure in Spain has similar epidemiological characteristics as in the U.S., Canada, and Catalonia,^10-12 although the characteristics have varied with respect to an earlier study of Spain in 1993.² The downward trend and distribution by sex and age have remained unchanged, but there are two new features in the pattern previously described. The first is that, due to the smaller decrease in women, the mortality of women now surpasses that of men. It is dangerous to conjecture as to reasons why the mortality due to heart failure has decreased less in women, but one possible reason is that death due to this cause now comes later. The second novelty is the stabilization of rates in persons over 85 detected from 1987-1991 in men and 6 years later in women. These findings suggest that therapeutic improvements in ischemic heart disease and in arterial hypertension have been able to delay the appearance of heart failure until more advanced ages.¹⁶ This effect could be more evident in women due to their greater longevity. It should be noted that in the U.S. mortality due to heart failure in persons over 65 years has also decreased at a rate of 1% a year in the 1990s.¹¹

Some authors have argued that the decrease in mortality might not be real, but due to the replacement of the diagnosis of heart failure by more specific diagnoses on death certificates. Nevertheless, the parallel decrease in mortality due to other cardiovascular causes and the finding of similar decreases when other categories are considered with heart failure,^2,10 such as hypertensive cardiorenal disease, chronic cardiopulmonary disease, myocardiopathy, poorly defined descriptions, complications of heart diseases, myocardial degeneration, and cardiomegaly, suggest that the decrease is real and not due to changes in the guidelines for death certificates.

The geographic differences have been examined widely in previous studies¹⁷ in which the geographic variation in hospitalization and mortality due to heart failure in Spain was analyzed for the period 1980-1993. In these years the geographic differences decreased, although the authors estimate that it is still possible to reduce hospitalization by 60% and mortality due to heart failure by 30% in persons over 45 years-old. In our study we also observed important regional differences that could be reduced by approximately 30%.

The phenomenon of increased mortality in young men in our study has not been described in other similar studies of mortality due to heart failure. This distribution by age, sex, and time period coincides with the distribution of drug use and the incidence of AIDS in Spain.^18,19 A possible explanation of this excess mortality can therefore be found in drug consumption, specifically cocaine. This can cause acute myocardial infarction, myocarditis and cardiomyopathy, and left ventricular hypertrophy, all diseases that can present with acute lung edema and heart failure.^15,20 On the other hand, is has been reported that some deaths due to an acute reaction to opiates are recorded as heart failure, probably due to the cardiovascular clinical manifestations of this reaction. In the municipality of Madrid, this percentage was 40% of overdoses in 1988.²¹ Although in later studies an improvement in the National Death Registry in detecting deaths due to acute reaction to psychoactive substances has been observed, such deaths continue to be underestimated.²²

Finally, we note that the decrease in mortality does not involve a decrease in the incidence. The number of deaths and hospital admissions has been increasing, which is explained only partially by the aging of populations. On the other hand, according to reviews of studies on heart failure made in the last 40 years, i n spite of advances in treatment the prognosis of heart failure does not seem to have improved. For that reason, aside from continuing to investigate new treatments to improve the quality of life of patients,²³ it is necessary to examine the prevention of heart failure by means of primar y and secondary prevention of ischemic heart disease and adequate control of arterial hypertension.²⁴

In addition, according to United Nations²⁵ population projections, in the year 2050 Spain will be the most elderly country in the world. The future scenario that these data sketch is an increasing number of older people with heart failure, which is why both healthcare and social needs may increase in the future.

CONCLUSIONS

In Spain, as in other countries, mortality due to heart failure continues to decrease and affect persons of more advanced ages. Nevertheless, an increasing number of deaths are detected in the population of very elderly women and an increased morbidity in both sexes. Given the demographic characteristics of the Spanish population, it is foreseeable that heart failure will become a condition demanding increasing attention.

ACKNOWLEDGMENTS

We would like to thank Drs. Luis de la Fuente, José Ramón Banegas, and Fernando Rodríguez Artalejo for their advice in the preparation of this study.

Ver editorial

---

Correspondencia: Dra. R. Boix.
Centro Nacional de Epidemiología.
Sinesio Delgado, 6. 28029 Madrid. Spain
Correo electrónico: [e-mail]