González-Diego P, López-Abente G, Pollán M, Ruiz M
Cancer Epidemiology Department, National Centre for Epidemiology, Carlos III Institute of Health, Sinesio Delgado 6, 28029, Madrid, Spain.
Eur J Cancer. 2000 Sep;36(14):1816-24. doi: 10.1016/s0959-8049(00)00184-2.
The time trend in ovarian cancer mortality in the European Union over the period 1955-1993, and the age, period-of-death and birth cohort components underlying the trend's evolution were analysed using log-linear Poisson models to quantify risk of dying from ovarian cancer in the different countries and regions of Europe, and ascertain the relative annual trend for each country. Furthermore, age-period-cohort models were fitted for each country in order to ascertain the effect on time trend exerted by the respective age, period-of-death and birth cohort components. Ovarian cancer mortality proved 2.77-fold (95% confidence interval (CI) 2.60-2.95) higher in northern versus southern Europe over the period 1955-1993. Denmark registered the highest adjusted rates, namely, 14.3 per 100000 person-years for the 1989-1993 5-year period, the last studied, with Portugal (4.5 per 100000) and Greece (4.5 per 100000) being the countries with the lowest rates. Spain and Greece, with annual rises of 5.8% (95% CI 5.3-6.3) and 5.1% (95% CI 4.2-6.0) respectively, were the countries that displayed the greatest increase in ovarian cancer mortality. Risk of death associated with the birth cohort effect declined in all northern countries from 1920 to 1930. In the south, Italy and France recorded a decline in risk from 1930. Women in Spain and Greece registered an increase in birth cohort-associated mortality, which became less pronounced after 1930. Ovarian cancer mortality in Europe evinces a south-north distribution pattern. The mortality risk for women cohorts born in northern Europe witnessed a gradual decline from 1920 to 1930. In the southern region: (1) Italy and France display a cohort effect of decreased risk from 1930; and (2) Greece and Spain show a cohort effect of increased risk among the different generations of women, though this became less pronounced from 1930 onwards.
利用对数线性泊松模型分析了1955 - 1993年期间欧盟卵巢癌死亡率的时间趋势，以及该趋势演变背后的年龄、死亡时期和出生队列成分，以量化欧洲不同国家和地区死于卵巢癌的风险，并确定每个国家的相对年度趋势。此外，为每个国家拟合了年龄-时期-队列模型，以确定各个年龄、死亡时期和出生队列成分对时间趋势的影响。结果表明，1955 - 1993年期间，北欧的卵巢癌死亡率比南欧高2.77倍（95%置信区间（CI）为2.60 - 2.95）。丹麦的调整率最高，即在最后研究的1989 - 1993年5年期间为每10万人年14.3例，而葡萄牙（每10万人年4.5例）和希腊（每10万人年4.5例）的发病率最低。西班牙和希腊的卵巢癌死亡率年增长率分别为5.8%（95% CI为5.3 - 6.3）和5.1%（95% CI为4.2 - 6.0），是卵巢癌死亡率增长幅度最大的国家。在所有北欧国家，与出生队列效应相关的死亡风险在1920年至1930年期间有所下降。在南方，意大利和法国的风险在1930年后有所下降。西班牙和希腊的女性出生队列相关死亡率有所上升，1930年后这种上升趋势变得不那么明显。欧洲的卵巢癌死亡率呈现出南-北分布模式。北欧出生队列的女性死亡率风险在1920年至1930年期间逐渐下降。在南部地区：（1）意大利和法国自1930年起显示出风险降低的队列效应；（2）希腊和西班牙在不同代女性中显示出风险增加的队列效应，不过从1930年起这种效应变得不那么明显。
