Annals of Oncology 15:793-796, 2004
© 2004 European Society for Medical Oncology
Joinpoint regression analysis of lung cancer mortality, Andalusia 1975–2000
1 Research Department, Virgen del Rocío University Hospitals, Seville; 2 Public Health Center of Bollullos de la Mitación, Seville; 3 Medical-Surgical Unit of Respiratory Diseases, Virgen del Rocío University Hospitals, Seville, Spain
Received 2 June 2003; revised 11 November 2003; accepted 2 January 2004
 |
ABSTRACT |
|---|
 Top ABSTRACT IntroductionPatients and methodsResultsDiscussionREFERENCES |
|---|
Background:
Previous analyses of lung cancer mortality were based on models of death rates within one time period, assuming that rates increase or decrease with time at a constant rate. The aim of this work is to analyse recent changes in lung cancer mortality trends in Andalusia (Spain) during the period of 1975–2000 using joinpoint regression models.
Patients and methods:
Mortality data were obtained from the Death Registry of Andalusia. For each gender, age group-specific and standardised (overall and truncated) rates were calculated by the direct method (using the world standard population). The joinpoint analysis was used to identify the best-fitting points where a statistically significant change in the trend occurred.
Results:
Lung cancer is the leading cause of cancer mortality in men, with an increasing trend up to 1994. After that year, rates began to decrease significantly (–1.8% yearly from 1994 to 2000). Standardised rates in women exhibited a downward trend until the early 1990s, after which they levelled off (overall standardised rates) or increased significantly (truncated rates 35–64 years).
Conclusions:
An increase in lung cancer mortality has been observed in young women. There seems to be a relationship with the prevalence in smoking in men and women.
Key words: epidemiology, lung cancer, mortality, trends
|
Introduction |
|---|
|
TopABSTRACTIntroductionPatients and methodsResultsDiscussionREFERENCES |
|---|
Lung cancer is currently the most common malignant neoplasm in the world (12.8% of all new cancer cases and 17.8% of cancer deaths) [1]. In men, higher incidence rates are observed in Western Europe and North America. In women, the highest rates are found in North America and Northwestern Europe. In Europe, lung cancer is the leading cause of cancer mortality in men and the third in women [2].
In the Autonomous Community of Andalusia, in southern Spain, lung cancer is the leading cause of cancer mortality in men and the fifth in women. In a previous publication [3] we analysed lung cancer mortality trends in Andalusia and found that age-standardised rates increased considerably in men and decreased in women. These previous analyses were based on models of death rates within one time period (1975–1992). This study yielded the average percentage change assuming that rates increase or decrease with time at a constant rate, although the validity of this assumption has not been evaluated [4].
The recently developed joinpoint regression model is useful for identifying and describing the occurrence of changes in different time periods throughout trends in data [5]. Therefore, the aim of this study is to provide up-to-date information and to analyse recent changes in lung cancer mortality trends in Andalusia during the period of 1975–2000 using joinpoint regression models.
|
Patients and methods |
|---|
|
TopABSTRACTIntroductionPatients and methodsResultsDiscussionREFERENCES |
|---|
Data on population and lung cancer deaths in Andalusia during the period of 1975–2000 were obtained from official publications of the Instituto de Estadística de Andalucía. Populations were estimated on 1 July of each year, based on official census information. Since death from lung cancer is rare in persons <35 years of age, we have only included subjects
35 years of age. For each gender, age group-specific rates and standardised (overall and truncated) rates were calculated using the direct method (using the world standard population). These were expressed as rates/100 000 persons.
We used joinpoint regression analysis to identify points where a statistically significant change over time in linear slope of the trend occurred [5]. In joinpoint analysis, the best-fitting points where the rate changes significantly (increase or decrease) are chosen. The analysis starts with the minimum number of joinpoints, and tests whether one or more joinpoints are statistically significant and should be added to the model (up to four joinpoints). In the final model, each joinpoint indicates a statistically significant change in trend, and an annual percentage change (APC) is computed for each of those trends by means of generalised linear models assuming a Poisson distribution. Significant changes include changes in direction or in the rate of increase or decrease. The computation of mortality rates and their standard errors were performed using a spreadsheet (Microsoft Excel). Joinpoint analyses were performed using the ‘Joinpoint’ software from the Surveillance Research Program of the US National Cancer Institute [6].
|
Results |
|---|
|
TopABSTRACTIntroductionPatients and methodsResultsDiscussionREFERENCES |
|---|
The results are summarized in Figure 1 and Table 1. Figure 1 displays overall and truncated (35–64 years) standardised mortality rates due to lung cancer in both sexes. Table 1 demonstrates the results of the joinpoint regression analysis (i.e. the points in which rates changed significantly) and the APC for each trend in women and in men.
|
|
In the year 2000 there were 2622 lung cancer deaths in men and 265 in women in Andalusia. This represents 31% and 5%, respectively, of deaths due to malignant tumours. Lung cancer mortality rates adjusted for age in men changed during the study period from 34.1/100 000 males in 1975 to 49.9/100 000 in the year 2000, with an average annual increase of 1.6% (P <0.05). There was a marked increase from 1975 to 1994, after which rates began to decrease considerably (–1.8% per year; Figure 1). Age group analysis revealed that for the entire study period, rates increased significantly in all age groups >40 years old. However, joinpoint analysis makes it possible to distinguish in almost all of the age groups two time periods: an initial period of statistically significant increase (in all age groups) and a second period of a levelling off or decrease in rates.
Mortality rates are considerably lower for women than for men. Although no apparent trend is observed for entire study period (1975–2000) in standardised rates (overall and truncated), the recent trend starting in 1992 shows a statistically significant annual increase of 3.9% in truncated rates (Table 1). The age group analysis shows that the rates remain stable in all age groups except in the 75- to 79-year-old age group, which showed a significant decrease (P <0.05) over the length of the entire study period (1975–2000). Joinpoint analysis does not show the existence of statistically significant points of change in any of the age groups, with the exception of the 60–64 years group. In the year 1998, a change in trend occurred in which a significant decrease changed to a marked increase, although the change was not statistically significant.
|
Discussion |
|---|
|
TopABSTRACTIntroductionPatients and methodsResultsDiscussionREFERENCES |
|---|
Owing to the lack of availability of data on morbidity, we have used mortality data, since this is the only data source which is available in Andalusia that satisfies the criteria of continuity and globality [7]. In our case, we may consider mortality data to be a reliable indicator of lung cancer frequency because of the low survival rate of lung cancer patients (in spite of diagnostic and therapeutic advances, lung cancer is still highly lethal: <15% of patients survive beyond 5 years after diagnosis) [8], and because of the accuracy of death certificates from this condition in our region [9]. Furthermore, our study has analysed truncated rates (<65 years old), in which the quality of information on cause of death is better.
Our results indicate that lung cancer mortality in men in Andalusia peaked in the years 1993–1994, and began to decrease after that time (Figure 1). This trend is in agreement with the decrease in the prevalence of smokers observed during the period of 1987–2000. This phenomenon is similar to that found in Barcelona city, where the decrease in lung cancer mortality rates was preceded by a decrease in the prevalence of smoking [10]. If we take into consideration the changes in the prevalence of smoking in our area, this fact seems to indicate that this trend will persist in the next few years.
On the other hand, a definite increase was observed in younger women (35–64 years old), although their rates are still very low. This increasing trend could indicate the beginning of a lung cancer epidemic associated with the use of tobacco as a consequence of the continued increase in prevalence of smokers in recent decades [11], and this would be expected to extend to older age groups in the coming years. The number of adolescent smokers decreased in the 1960s, but unlike the situation in adults (where the number of smokers has progressively decreased), numbers of adolescent smokers levelled off during the 1980s. Furthermore, there has been a notable increase in women, to an even greater level than in men [12]. In other words, the formerly privileged situation of women may deteriorate precipitously in coming years. The social cost of this trend will become apparent when these cohorts of young women who have adopted smoking habits in great numbers reach the age where clinical disease associated with smoking becomes manifest.
The pattern of trends observed in Andalusia, which are similar to those observed in other regions of the Mediterranean, appears to be similar to those observed in Northern and Eastern Europe, although the changes in trends occurred almost a decade ago [13].
|
FOOTNOTES |
|---|
* Correspondence to: Dr A. Cayuela, Hospitales Universitarios Virgen del Rocío, Edificio de Laboratorios, Unidad de Apoyo a la Investigación, Avenida de Manuel Siurot, s/n, 41013 Sevilla, Spain. Tel: +34-9550-13293; Fax: +34-9550-13292; E-mail: aurelio.cayuela.sspa{at}juntadeandalucia.es
|
REFERENCES |
|---|
|
TopABSTRACTIntroductionPatients and methodsResultsDiscussionREFERENCES |
|---|
1. Hoffman PC, Mauer AM, Vokes EE. Lung cancer. Lancet 2000; 355: 479–485.[Web of Science][Medline]
2. Levi F, Lucchini F, La Vecchia C, Negri E. Trends in mortality from cancer in the European Union. Lancet 1999; 354: 742–743.[CrossRef][Web of Science][Medline]
3. Cayuela A, Rodríguez R (eds). Análisis de la Mortalidad en Andalucía 1975–1992. Seville: Instituto de Estadística de Andalucía 2000.
4. Esteve J, Benhamou E, Raymond L. Statistical Methods in Cancer Research, Vol. IV. Descriptive Epidemiology. IARC Scientific Publications No. 128. Lyon, France: IARC 1994.
5. Kim H-J, Fay MP, Feuer EJ, Midthune DN. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 2000; 19: 335–351.[CrossRef][Web of Science][Medline]
6. National Cancer Institute. Statistical Research and Applications Branch [On-line]. http://srab.cancer.gov/ (23 May 2003, date last accessed).
7. Glasser JH. Editorial: The quality and utility of death certificate date. Am J Public Health 1981; 71: 231–233.
8. Berrino F, Capocaccia R, Estève J et al. Survival of Cancer Patients in Europe: The EUROCARE-2 Study. IARC Scientific Publications No. 151. Lyon, France: IARC 1999.
9. Martínez C, Sánchez MJ, Rodríguez M et al. Exactitud del diagnóstico de cáncer en los certificados de defunción de la provincia de Granada. Rev Oncol 2000; 2: 245–252.
10. Villabí JR, Pasarín MI, Nebot M, Borrell C. Mortalidad por cáncer de pulmón: inicio de declive en los varones de Barcelona. Med Clin (Barc) 2001; 117: 727–731.[Medline]
11. López A. Cáncer de pulmón en mujeres. Arch Bronconeumol 2001; 37: 55–57.[Medline]
12. Barrueco M, Vicente M, Garavís JL et al. Prevención del tabaquismo en la escuela: resultados de un programa realizado durante 3 años. Arch Bronconeumol 1998; 34: 323–328.[Medline]
13. Borràs JM, Fernandez E, Gonzalez JR et al. Lung cancer mortality in European regions (1955–1997). Ann Oncol 2003; 14: 159–161.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J Zhang, I Dhakal, H Yan, M Phillips, and H Kesteloot Trends in pancreatic cancer incidence in nine SEER Cancer Registries, 1973 2002 Ann. Onc., July 1, 2007; 18(7): 1268 - 1279. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||