The beginning of the end of the lung cancer epidemic in Dutch women?
Article first published online: 17 JUN 2008
Copyright © 2008 Wiley-Liss, Inc.
International Journal of Cancer
Volume 123, Issue 6, pages 1472–1475, 15 September 2008
How to Cite
Karim-Kos, H. E., Janssen-Heijnen, M. L.G., van Iersel, C. A., van der Meer, R. M., de Vries, E. and Coebergh, J. W. W. (2008), The beginning of the end of the lung cancer epidemic in Dutch women?. Int. J. Cancer, 123: 1472–1475. doi: 10.1002/ijc.23681
- Issue published online: 16 JUL 2008
- Article first published online: 17 JUN 2008
- Manuscript Accepted: 7 APR 2008
- Manuscript Received: 30 JAN 2008
- Dutch Cancer Society. Grant Number: 715401
- lung neoplasms;
In some European countries, female lung cancer mortality and incidence have started to decrease or flatten out, whereas they are still rising in The Netherlands. We present recent mortality and incidence trends of lung cancer and smoking trends in The Netherlands to show the end of the lung cancer epidemic in Dutch women. Lung cancer mortality and incidence rates by gender were analyzed for 4 age groups (20–44, 45–49, 50–54 and 55–59), and smoking prevalence rates were examined for women using joinpoint regression and birth cohort analysis. Data on mortality were collected for the period 1960–2006, incidence for the period 1989–2003 and smoking prevalence for the period 1988–2007. Because of decreasing lung cancer mortality and incidence rates among males and dramatically increasing rates among females, rates of young males were surpassed by those of females after the mid-1990s. However, although in young women (20–49) mortality increased with 4–5% per year, it flattened out (no significant in- or decreases) since 1999. Among older women, mortality rates were still increasing markedly. Mortality rates and smoking prevalence tended to decrease in women born after the 1950s. This is the first report suggesting that the lung cancer epidemic in Dutch women is coming to an end. Although the increase in lung cancer incidence and mortality among Dutch women has been one of the most dramatic in Europe, the recent decrease in young women is expected to be followed by a future leveling off or a slight decrease in overall female lung cancer rates. © 2008 Wiley-Liss, Inc.
In The Netherlands, the increase in female lung cancer mortality has been one of the most dramatic in Europe1 and lung cancer has become the second cause of death from cancer among women since 2000. Lung cancer was responsible for 17% of all female cancer deaths in 2006 and it is likely to become the first within 5 years.2 In 2006, 3,172 women died from lung cancer, almost 9 deaths per day. In contrast to males, with declining rates since the 1980s, the age-standardized (European standard) mortality rate of lung cancer in women has increased dramatically between 1970 and 2006 from 5 to 30 per 100,000 in The Netherlands.2 Dutch female lung cancer incidence and mortality rates are among the highest in Europe.3 As smoking prevalence decreased among Dutch women from 40% in the 1970s to 25% in 2007 (Fig. 1), the rising incidence and mortality rates are expected to flatten out or decrease, as already observed in Iceland, Ireland, UK1 and USA.5
In this short report, we present recent age-specific incidence and mortality trends of lung cancer and smoking trends at young and middle age in The Netherlands and suggest the beginning of the end of the lung cancer epidemic in Dutch women.
Material and methods
Mortality from malignant neoplasms of the trachea, bronchus and lung for the period 1960–2006 were derived from Statistics Netherlands.2 Four revisions of the International Classification of Diseases (ICD) were used during this period: the seventh (1958–68), eighth (1969–78), ninth (1979–95) and tenth revision since 1996. For the whole period, cancer deaths were recoded according to ICD-codes C33/C34 of the tenth revision of ICD.6
Data on lung cancer incidence for the period 1989–2003 were obtained from the nation-wide Netherlands Cancer Registry (www.ikcnet.nl), which consists of 9 regional cancer registries since 1989. The cancer registries receive lists of newly diagnosed cases on a regular basis from the pathology departments, all participating in a nation-wide pathology network (PALGA). In addition, the medical records departments of hospitals provide lists of diagnoses of outpatients and hospitalized patients with a suspected cancer diagnosis. Following this notification, the necessary information of newly diagnosed tumors is abstracted from the medical records by trained tumor registration clerks. Topography is coded according to the International Classification of Diseases of Oncology.7
Annual age-specific mortality and incidence rates were calculated for 4 age groups: 20–44, 45–49, 50–54 and 55–59. Rates at ages 20–44 were standardized to the European standard population using the direct method.
Joinpoint regression analysis was used to identify years where a significant change in the mortality trend occurred.8 The estimated annual percent change (EAPC) and the corresponding 95% confidence interval was calculated for each of those trends by fitting a regression line to the natural logarithm of the rates, using calendar year as regressor variable [i.e., y = mx + b where y = ln(rate) and x = calendar year, then EAPC = 100 × (em − 1)]. The joinpoint regression models were performed using the Joinpoint Regression Program (version 3.0) from the Surveillance Research Program of the US National Cancer Institute (http://srab.cancer.gov/joinpoint/).
To estimate the effect of birth cohort on trends in mortality of female lung cancer, mortality rates for ages 20–59 were calculated for birth cohorts of 10 years. These “synthetic” birth cohorts were created based on the year and age of death, using 5-year age and 5-year calendar period analysis.
The effect of birth cohort on smoking trends in women was examined by calculating the smoking prevalence rates for ages 15–64 years by 10-year birth cohorts. Smoking prevalence data were available for the period 1988–2007 and were collected by STIVORO, the Dutch national expert center on tobacco prevention.
Figure 2 shows the dramatic increase of lung cancer mortality among women aged 20–59 in The Netherlands since the 1960s. The same increase was observed in female lung cancer incidence (national incidence data only available since 1989) (Fig. 3). As a result of the decreasing male rates and the increasing female rates, male mortality and incidence rates were even surpassed by female rates since the mid-1990s, except for those over age 50. The male-to-female (M:F) mortality rate ratio decreased from 5.0 in 1970 to 0.7 in 2006 for ages 20–44 and from 11 to 0.9 at ages 45–49. Incidence M:F rate ratios were 0.8 for ages 20–44 and 45–49 in 2003.
In young women (ages 20–44), lung cancer mortality increased by 4.3% annually, from 0.5 per 100,000 in 1960 to 3.5 in 1999. After 1999 however, rates dropped to 3.0 per 100,000 in 2006. The same reversal of the increasing mortality trend started in 2004 in women aged 45–49 (from 31 per 100,000 in 2004 to 23 in 2006). Among women aged 50–54 and 55–59 rates kept rising by 4–6% annually (Table I).
|Age||Age-specific mortality rate per 100,000||Joinpoint analysis|
|Trend 1||Trend 2|
|1960||2006||Period||EAPC1 (95% CI)||Period||EAPC1 (95% CI)|
|20–44||0.5||3.0||1960–1999||4.3 (3.8, 4.9)||1999–2006||−1.5 (−8.4, 5.8)|
|45–49||1.5||22.5||1960–2004||5.2 (4.7, 5.7)||2004–2006||−14.1 (−35.4, 14.3)|
|50–54||5.6||49.6||1960–2006||5.0 (4.7, 5.3)|
|55–59||6.1||67.1||1960–1990||5.8 (5.0, 6.7)||1990–2006||4.0 (3.0, 5.0)|
Figure 4 gives age-specific mortality rates of female lung cancer by birth cohort and showed that mortality tended to decrease in women born after the 1950s. The same pattern was observed for smoking prevalence among women by birth cohort (Fig. 5). Women born after the 1950s started smoking less.
We demonstrated the beginning of the end of the lung cancer epidemic among Dutch women. Although overall female lung cancer mortality and incidence rates continued to rise markedly in The Netherlands, reaching levels exceeding those of males, we observed a decline in the mortality and incidence trend among young women, particularly for the generation born after the 1950s. These findings are in line with the decreasing smoking prevalence in women born after the 1950s.
Although lung cancer is uncommon among young age groups, these lung cancer trends are important, particularly since they give information on recent changes in risk-factor prevalence (e.g., smoking) and thereby information on the likely future trends in middle and elderly age.9, 10 The end of the lung cancer epidemic among Dutch men was also first observed among young men in the early 1980s.11
The decrease in smoking prevalence among women occurred mostly between 1970 and 1989 (−10%). From the 1990s, the smoking prevalence became more or less stable and between 2000 and 2007 it decreased with another 5%. Based on this information, we expect first a slight decrease in the overall female lung cancer mortality and incidence followed by a leveling off.
In this study, we focused mainly on lung cancer mortality, because mortality data was available for a longer time period than incidence data. As the case-fatality of lung cancer is high and therefore mortality trends closely follow incidence trends, this main focus on mortality is justified. Furthermore, impressive changes in lung cancer survival did not take place in the Netherlands,12, 13 which implies that mortality is mainly influenced by incidence and therefore by changes in risk-factor prevalence. This is also confirmed by our findings that lung cancer mortality and smoking prevalence among women started to decrease in the same generation.
At the beginning of the 21st century, in Europe there was a general tendency for M:F rate ratios for lung cancer mortality to converge toward 1.0.14 Convergence of male and female lung cancer rates can be caused by declining male rates and rising female rates as observed in this study and in other European countries like Finland. In countries like Denmark, Sweden and Ireland, convergence of the M:F rate ratios were only because of decreasing male rates.14 Jemal et al.15 showed that in the USA, smoking prevalence converged to 1.0 among young men and women born after 1960, resulting in converging male and female lung cancer rates. From the available recent data on smoking prevalence by birth cohort in The Netherlands, we found only an M:F rate ratio of smoking prevalence smaller than 1.0 among young adults aged 20–24 and born in the 1960s, which increased up to 1.2 for those born in the 1980s (data not shown). This finding is confirmed by the data presented in Figure 1; in the 1980s, the smoking prevalence rates among men were surpassed by those of women aged 20–34. This is a plausible explanation for the observation that male lung cancer rates were surpassed by female rates in The Netherlands. However, from the 1990s, again men started smoking more than women in this age group, which might result in an increasing M:F ratio for lung cancer trends among young adults in the future.
Despite the fact that smoking prevalence became equal among young men and women in the past, this is not the only explanation for the female lung cancer incidence and mortality rates exceeding those of men. Possible other explanations include a higher female susceptibility to tobacco smoke,16 a different smoking pattern17 or more passive smoking18 among women.
On average, 57% of lung cancer is avoidable by reducing smoking in Europe,19 underlining the importance of antismoking interventions to attain lower lung cancer incidence and mortality among men and women. Such interventions should focus on adolescents and young adults to prevent that they start smoking, particularly since there was a slight increase of the smoking prevalence among both boys and girls aged 15–19 between 1990 and 2000 (Figs. 1 and 5). This increase was not observed among women aged 20–34, their smoking prevalence even continued to decrease during this period. This means that many of the 15- to 19-year-old girls who started smoking, stopped before reaching the age of 20. From the annual smoking monitor among youth in The Netherlands it is known that about 45% of smoking girls aged 15–19 quitted.20 The slight increase of smoking prevalence among girls aged 15–19 between 1990 and 2000 is therefore not expected to have a major influence on overall female lung cancer mortality and incidence in the future.
In the light of the recently increased smoking prevalence among adolescents, preventing smoking uptake must remain a main public health issue. Besides the importance of antismoking interventions (i.e., quit smoking campaigns, smoke free public places and increasing tax on cigarettes), we should also focus on further research to early detection (e.g., screening), better diagnostics and the role of estrogens21 and genetics22 in lung cancer to optimize lung cancer treatment and thereby reducing lung cancer mortality.
The work on this article was conducted within the project “Progress against cancer in The Netherlands since 1970s? Epidemiological interpretation of changes in survival, incidence and mortality.”
- 2Statistics Netherlands. Available at: http://statline.cbs.nl.
- 4STIVORO. Annual Report 2006. The Hague: STIVORO, 2007; p 70.
- 6World Health Organization. International statistical classification of disease and related health problems: 10th revisioned. Geneva: World Health Organization, 1992.
- 7FritzA,PercyC,JackA,ShanmugaratnamK,SobinL,ParkinDM,WhelanS, eds. Classification of diseases for oncology,3rd edn. Geneva: World Health Organization, 2000.
- 20Smoking habits of young people 2007. Amsterdam: TNS NIPO, 2007; p 373.