Background
In 2004, cardiovascular diseases (CVD) were the leading cause of death in the world, responsible for approximately 32% of all deaths in women and 27% in men. After infectious and parasitic diseases, cancer ranked third, causing some 12% of all deaths among women and 13% among men [
1]. Taking the age at death into account, however, CVD were responsible for only 14% and cancer for 8% of the total years of life lost (YLL) in 2004, surpassed by infectious and parasitic diseases, perinatal conditions and unintentional injuries.
Age-standardized mortality rates from CVD have been declining for several decades in Europe [
2,
3], while the downward trend in mortality from some common cancers is more recent, and varies with the type of cancer. The overall cancer mortality has been decreasing less steeply, and in some Eastern European countries is still increasing [
4]. While the proportion of all deaths from CVD was almost three-fold higher than from cancer in Europe, the YLL were less than two-fold higher for CVD than cancer, in 2004 [
1,
5].
In Portugal, CVD were the leading cause of death in 2006, followed by cancer, and these two groups of diseases were responsible for more than half of all deaths [
6]. Portugal is characterized by a very higher contribution of cerebrovascular disease to mortality relative to coronary heart disease [
1] and long-standing higher systolic blood pressure, as compared to most other European countries [
7]. The types of cancers that cause more deaths in Portugal are lung and stomach among men and breast among women [
4]. The relatively early stage of Portugal in the smoking epidemic results in an expected increase in the smoking-related burden of disease and deaths, particularly among women [
8,
9].
We aimed to describe time trends in the absolute number of deaths, death rates and YLL from CVD and cancer in the Portuguese population, during the period 1980–2010, and to quantify the contribution of the variation in the population’s size and age structure, and age-independent “risk” of death by cardiovascular or oncological causes to the change in the corresponding number of deaths.
Discussion
We observed a dramatic decrease in Portuguese CVD mortality rates from 1980 to 2010 in both sexes. However, this decrease was reflected in a fall in the total number of deaths only since the mid-1990s, due to the increases and ageing of the population. For cancer, we observed a much smaller decrease in the mortality rates from the 1990s until 2006 in both sexes, and an increase thereafter. These trends, together with increased premature mortality from cancer, meant that in men, YLL from cancer surpassed those from CVD after 2005.
In recent years, cancer has become the leading cause of death among men, in many other European countries, such as France, Spain, Netherlands and Italy [
18,
19]. Cardiovascular diseases and cancer interact as mutually competing risks. The observed decreasing trends in CVD mortality may have contributed to the recent increase in mortality rates from cancer, with more individuals surviving at risk of cancer. We also observed a decrease in the number of deaths caused by gastro-intestinal diseases, injuries and poisoning, and infectious and parasitic diseases in the last three decades, in Portugal [
10]. Since some of these causes of death are more common in the young and middle-aged population, these trends leave more at risk for the diseases associated with older age, such as diabetes, CVD or cancer [
20].
The mortality trends from CVD in Portugal paralleled steady mortality declines in Western Europe, North America, Japan and other developed areas of the world [
2,
3]. This reflects decreases in both cerebrovascular and ischaemic heart diseases, the two major contributors to this group and more frequent in the older population [
3,
21]. In Portugal, cerebrovascular disease has been the main cause of death during the last decades [
6,
22]. In developed countries, the mortality rates from stroke have been decreasing more steeply in countries with higher departure rates, which is the case of Portugal [
23]. This could reflect general improvements in living conditions, lower salt consumption thanks to improved methods for food preservation and recently, perhaps, lower case-fatality due to stroke units and better emergency care [
24]. Regarding ischaemic heart disease, modeling studies in many developed countries consistently suggest that 45% to 75% of the mortality declines can be attributed to decreases in major risk factors, with the remaining attributable to medical and surgical treatments [
25‐
28]. In Portugal, we likewise observed notable decreases in the mean systolic blood pressure and blood cholesterol, smoking consumption among men, and some dietary modifications, accompanied by the increased use of therapies for acute coronary syndromes [
7,
9,
29,
30]. Deaths from CVD in younger age groups probably reflected improved diagnostic capacities and successful early surgical interventions for congenital heart disease, plus better medical care for hypertension, cardiac arrhythmias and endocarditis [
31]. The pattern in the trends of YLL suggests that death by CVD is being increasingly delayed, clearly affecting mostly the elderly population.
Interpreting trends in cancer mortality rates can be challenging, because they reflect trends in several common cancers. In Portugal, the specific cancers that could particularly contribute to the slight decrease in overall cancer mortality rate between the 1990s and 2006 among adults include stomach cancer (in both sexes), prostate and lung among men and breast and uterus among women [
4]. Despite the trends in mortality rates, the total number of deaths and YLL have increased steadily. Among the more frequent cancers stomach, lung and breast cancer are associated with death at relatively younger ages [
20], contributing more to the increase in YLL due to cancer. On the other hand, we expected a lower contribution to YLL from prostate cancer, since this is a disease mainly of the elderly population [
20]. In younger ages, the decrease in leukaemia may also contribute [
4]. The few adult cancers increasing in more recent years differ between women and men. In women, the very recent increase in breast cancer mortality plus the long-term increase in the lung cancer mortality could help explain the turning point observed in the cancer mortality rate and YLL in 2006 [
32]. In men, colorectal cancer is the only common cancer which is increasing [
10]. However, further close analyses during the forthcoming years are clearly necessary to confirm that the trends observed in the 4 years since 2006 persist longer term. The steady decline in cancer mortality is likely to reflect improvements in risk factors such as male smoking and diet, compounded by better access to earlier and specialized diagnosis, staging and treatment [
33]. Worryingly, the mortality for some cancers has increased since 2006, which may reflect an increase exposure to other risk factors, such as obesity [
34]. The transition to a more Westernized diet has occurred faster in Portugal than in some other Mediterranean countries [
35]. The continued increases in smoking among women across the Eastern, Western and Southern parts of Europe [
36] are very concerning. Unless halted, these will inevitably increase cancer mortality among women in Portugal. Furthermore, the late implementation and the low participation rate in organised cancer screening programmes [
37] may delay the potential benefits. In 2008, the breast and cervical cancer screening did not cover the entire country and the screening program for colorectal cancer was still in its infancy [
37]. Apart from the variation in the population structure, we identify an increase in the risk of dying from cancer after 2006, which inverts the previous decrease observed in the 1990s.
The interpretation of the reported trends depends on the quality of the source data. Misattribution or miscoding can occur mainly because of incorrect diagnoses, incorrect or incomplete death certificates, misinterpretation of ICD rules for coding underlying causes, and variations in the use of categories for unknown and ill-defined causes [
38]. Like most developed countries Portuguese mortality data have high coverage, but have been considered to have low quality regarding the high proportion (21%) of deaths coded as ill-defined causes [
38]. Additionally, and more specifically, there is evidence of inaccuracies in attribution of frequent causes of death. For instance, in a validation study in the mid-2000s, stroke was not confirmed in a large proportion of death certificates indicating it as the underlying cause of death, and 5% of patients were identified only after the review of death certificates [
39]. The accuracy in attribution to causes of death varies over time due to quality of diagnoses and also due to changes in the coding rules. In Portugal, two revisions of the ICD were used in the calendar period considered in this study, with ICD-10 replacing ICD-9 in 2002. However, since the ICD-10 is more detailed than ICD-9 and we used two large groups of diseases, without focusing on very specific codes, errors might have been minimised.
Although we do not have information on CVD incidence and case-fatality in this population, we predict that death from CVD will be mainly associated with old age in Portugal, as in highly developed Western European countries. However, the prevalence of CVD is likely to increase and so is the related demand for healthcare. The observed trends in cancer mortality highlight the importance of primary and secondary prevention, demanding better planning in health services, in order to obtain improvements in earlier detection and/or increasing proportions of patients receiving new or more aggressive treatment.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MP, collaborated in the acquisition, analysis and interpretation of the data, and wrote the first draft of the article. BP, collaborated in the in the acquisition, analysis and interpretation of the data, and in the revision of the article. SC, collaborated in the interpretation of the data and reviewed the article critically for important intellectual content. KB, collaborated in the interpretation of the data and reviewed the article critically for important intellectual content. AA, designed the study, analysed and interpreted the data, and reviewed the article critically for important intellectual content. NL, designed the study, analysed and interpreted the data, and reviewed the article critically for important intellectual content. All authors read and approved the final manuscript.