A preview of some highlights of the study, which was co-ordinated by the Epidemiology Unit of the Istituto Nazionale Tumori, Milan, Italy and is due to be published by Annals of Oncology shortly - was presented today (Thursday 25 September) at the final day of ECCO 12 - The European Cancer Conference.
The report, which is a follow up to EUROCARE-2, published in 1999, has analysed data from registries covering all or part of 22 countries1 and 42 types of cancer. It examined five-year survival after diagnosis for 1.8 million adults and 24,000 children diagnosed during 1990-94 and followed to the end of 1999.
One of the report's authors, Professor Michel Coleman, Professor of Epidemiology and Vital Statistics at the London School of Hygiene and Tropical Medicine, told the conference that the aim of exploring differences in cancer survival was not to establish league tables or excite national rivalries, but to estimate the range of survival rates and to identify regions or countries in which survival could be improved. "There is increasing evidence that international survival differences are at least partly due to factors that can be changed, such as stage at diagnosis, access to optimal treatment and investment in health care," he said.
[The proportion of gross domestic product devoted to health care in the contributing countries in 1995 ranged from 6% in Poland to 10.6% in Germany, and the per capita spend varied from $420 in Poland to $2,555 in Switzerland]2.
Looking at Europe overall, average five-year survival rates varied from a high of 94% for lip cancer to less than 4% for pancreatic cancer.
There was an average of 80%, or better, survival for cancers of the lip, testis and thyroid and for melanoma and Hodgkin's disease - but these cancers accounted for only 4% of adult malignancies.
About a third of cancers averaged survival rates of 60-79%. These included cancers of the breast, prostate, bladder, cervix, uterus and larynx.
For one-fifth of all cancers, the European average five-year survival was in the range 40-59% - this group included colon, rectum and kidney cancers and non-Hodgkin lymphoma.
About 10% of all cancers had poor prognosis, with an average of 20-39% five-year survival. These included stomach and ovarian cancer and multiple myeloma.
However, about a quarter of all adult cancers fell into the worst prognosis category - including cancers of the lung, pancreas, oesophagus, brain and liver. These had survival rates of less than 20%.
Childhood cancer survival was generally good with two-thirds of all affected children (13 of the 24 malignancies analysed) achieving 75% five-year survival or better.
Survival for most cancers declined steeply with age at diagnosis. Two exceptions were breast cancer, where the best survival was for women aged 45-54 and prostate cancer where it was among men aged 55-64 years.
But, even after the investigators had adjusted for differences in background mortality by age and sex between regions and countries they found large variations between countries in overall cancer survival and survival for specific cancers, as well as between men and women.
Overall survival was generally below the European average in five eastern European countries (Czech Republic, Estonia, Poland, Slovakia and Slovenia, with Poland the lowest) and in Denmark, England, Scotland, Wales, Malta and Portugal among participating western European countries. Sweden tended to have the highest survival rates among the five Nordic countries, while the areas of France and Switzerland covered by the study often had the highest survival rates among western European populations.
For men, the all-cancers survival index ranged from 25 to 32% in eastern European populations and from 40 to 47% in most of the Nordic and western European populations. For England, Scotland, Wales and Denmark the range was from 33 to 37%, slightly below the European average of 38%.
For women, the range was from 41 to 47% in eastern European countries and in the narrow range of 55 to 58% in ten of the Nordic and western European countries. For England, Scotland, Wales and Denmark the range was from 47 to 51%, slightly below the European average of 52%.
However, while survival in the UK and Denmark was generally below the European average, these countries had either above or about average survival rates for melanoma, testicular cancer and Hodgkin's disease - illustrating how countries performed differently for different cancers.
Some cancers showed particularly wide differences in survival. Lung cancer varied more than two-fold although the best rates were still under 15%. In some cancers there was evidence that the nature of the cancer was contributing to differences. For example, stomach cancer survival was higher in western Europe where incidence is high than in the Nordic countries and the UK where incidence is lower. Survival was relatively high in southern Europe because of a higher proportion of tumours in which either the location of the tumour or the specific type of tumour conferred a better prognosis.
For 30 of the 35 cancers examined that occur in both sexes, survival was higher in women. According to the researchers the advantage among women was most likely to arise from sex differences in tumour biology, immune system mechanisms, symptom awareness, stage of disease at diagnosis or access to effective treatment.
Turning to time trends since the mid-1980s, one of the most worrying aspects was the widening gap in cancer survival between eastern Europe and other parts of Europe. Breast cancer survival improved steadily in all European countries, with evidence that some western European countries with previously poorer survival were catching up. But, improvements were less marked for eastern Europe, so the gap between east and west has actually increased
For colorectal cancer the gap between the higher Nordic survival figures and western Europe is also closing, but again the gap between west and east is widening.
Prostate cancer survival trends vary widely, with survival increasing in most countries. Five-year survival increased in the Nordic countries, except Denmark, and there were sharp improvements in France, Germany, Italy, England, Scotland and Wales. Trends were more diverse in eastern Europe, with a rapid increase in Estonia but a decline in Poland and Slovakia. The trends appear mostly to reflect international differences in the use of Prostate-Specific Antigen (PSA) as a diagnostic test, and the increases cannot all be interpreted as a better outcome for men with this cancer.
Professor Coleman said that although European survival trends for all cancers combined were not adjusted for differences in the case-mix between countries, they were remarkably consistent and appeared to offer an overall indicator of the performance of health care systems in each country.
He sounded a warning note for the future on the east-west divide. "The survival trends reflect a substantial and increasing gap in the overall prognosis of cancer between eastern and western Europe. Since all five eastern European countries participating in EUROCARE will probably join the EU in 2004, this raises a major new problem of inequity in health within the EU.
Professor Coleman added that "The wide differences in survival trends are likely to reflect differences in both stage at diagnosis and the availability of and access to health resources, both of which are amenable to intervention. They represent a benchmark for reduction in inequalities in cancer survival across Europe in the future. While we do not yet have a fully satisfactory interpretation of these differences, we have few alternatives to this type of study and we believe this is a vital contribution towards the ongoing efforts to improve survival in Europe. The EUROCARE team is actively inviting independent clinical commentary on the study. Generating an appropriate level of concern to address this problem would be an important outcome of the EUROCARE project."
(Thursday 25 September, 10.45 hrs CET - How to improve cancer care in Europe session).
1For 11 of the 22 participating countries the registries contributing data covered 100% of the population. In others it covered varying proportions. Overall, the EUROCARE-3 study covered a quarter of the total population of the 22 countries (100 million out of 400 million).
2The per capita spend on health is in constant purchasing power parity US dollars.